Browse the corpus

1. Introduction Although Staphylococcus aureus is a commensal of humans [1], it is also a frequent cause of human infections which may become serious if caused by antimicrobial resistant strains [2]. Antibiotic resistant S. aureus, especially MRSA, are equally adopted to hospitals and outer environments evolving as major pathogens of public health concern [3, 4]. Shortly after the introduction of methicillin in clinical world to treat infections caused by penicillinase producing S. aureus in 1960, MRSA emerged and spread worldwide [5, 6]. The high rate of methicillin resistance among Staphylococcus aureus has resulted into the increased interest for the use of clindamycin for treatment of infections caused by S. aureus [7]. But recently, increasing numbers of strains of S. aureus are acquiring resistance toward clindamycin [7]. Vancomycin is regarded as the drug of choice for treatment of infections caused by MRSA [8]. But emergence of VISA and VRSA has been reported by many authors [8]. Further, there are reports of treatment failure of the infections caused by MRSA having MIC of vancomycin just below cutoff value [8]. High vancomycin MIC for MRSA which are susceptible to vancomycin may indicate the drug resistance to many antibiotics [8]. MRSA is resistant to entire classes of β-lactams including cephalosporins and carbapenems and has higher risk of development of resistance to quinolones, aminoglycosides, and macrolides [9–12].

Vancomycin is regarded as the drug of choice for treatment of infections caused by MRSA [8]. But emergence of VISA and VRSA has been reported by many authors [8]. Further, there are reports of treatment failure of the infections caused by MRSA having MIC of vancomycin just below cutoff value [8]. High vancomycin MIC for MRSA which are susceptible to vancomycin may indicate the drug resistance to many antibiotics [8]. MRSA is resistant to entire classes of β-lactams including cephalosporins and carbapenems and has higher risk of development of resistance to quinolones, aminoglycosides, and macrolides [9–12]. Methicillin resistance in S. aureus is mediated through an altered protein called low-affinity penicillin binding protein (PBP2a). PBP2a is encoded by mecA gene which is present in chromosomal mobile genetic element called Staphylococcal cassette chromosome mec (SCCmec) [13, 14]. Due to possible association of MRSA with multiple antibiotic resistance and relatively difficult and higher cost of treatment, the accurate and rapid identification of MRSA is crucial in clinical world for timely management of the infections caused by this superbug [15]. Detection of methicillin resistance in Nepal is based on cefoxitin and oxacillin disc diffusion methods with limited reports on MIC determination and detection of mecA gene by polymerase chain reaction (PCR) [16, 17]. In present study, we evaluated the performance of cefoxitin disc diffusion and oxacillin broth microdilution methods for detection of MRSA taking presence of mecA gene as reference. Further, we also studied the rates of inducible clindamycin resistance and beta-lactamase production among the strains of S. aureus and we determined the minimum inhibitory concentration of vancomycin for S. aureus isolated from pus/wound swab samples.

ds for detection of MRSA taking presence of mecA gene as reference. Further, we also studied the rates of inducible clindamycin resistance and beta-lactamase production among the strains of S. aureus and we determined the minimum inhibitory concentration of vancomycin for S. aureus isolated from pus/wound swab samples. 2. Materials and Methods 2.1. Study Site and Population The present study was carried out among the patients (inpatients and outpatients) attending Shree Birendra Hospital, Kathmandu, Nepal, from July 2013 to January 2014. A total of 711 nonrepeated pus/wound swab samples from different anatomic locations received from the patients for bacteriological culture were included in the study. 2.2. Isolation and Identification of Staphylococcus aureus The specimens were inoculated on blood agar and mannitol salt agar (HiMedia laboratories private limited, India) and incubated aerobically at 37°C for 48 hours. The strains of Staphylococcus aureus were identified on the basis of colony morphology, Gram's stain, and different biochemical tests [18].

aphylococcus aureus The specimens were inoculated on blood agar and mannitol salt agar (HiMedia laboratories private limited, India) and incubated aerobically at 37°C for 48 hours. The strains of Staphylococcus aureus were identified on the basis of colony morphology, Gram's stain, and different biochemical tests [18]. 2.3. Antimicrobial Susceptibility Testing The antimicrobial susceptibility testing was performed by modified Kirby-Bauer disc diffusion technique using Mueller-Hinton agar (HiMedia laboratories private limited, India) following Clinical and Laboratory Standards Institute (CLSI) guidelines [19]. Antibiotic discs used were ciprofloxacin (5 μg), clindamycin (2 μg), chloramphenicol (30 μg), erythromycin (15 μg), gentamicin (10 μg), tetracycline (30 μg), cotrimoxazole (25 μg), rifampin (5 μg), mupirocin (200 μg), and penicillin G (10 units). 2.4. Detection of Strains of MRSA by Cefoxitin Disc Diffusion Method Susceptibility of Staphylococcus aureus isolates to cefoxitin (30 μg) was determined by modified Kirby-Bauer disc diffusion method following CLSI guidelines [19]. The strains of Staphylococcus aureus which were found to be resistant to cefoxitin were screened as MRSA (Table 1).

f MRSA by Cefoxitin Disc Diffusion Method Susceptibility of Staphylococcus aureus isolates to cefoxitin (30 μg) was determined by modified Kirby-Bauer disc diffusion method following CLSI guidelines [19]. The strains of Staphylococcus aureus which were found to be resistant to cefoxitin were screened as MRSA (Table 1). 2.5. Determination of Minimum Inhibitory Concentrations (MICs) of Oxacillin and Vancomycin MICs of oxacillin (Table 1) and vancomycin for all isolates of Staphylococcus aureus were determined by broth microdilution method as described by Andrews [20] and CLSI M07-A9 guidelines [21]. The results were interpreted according to CLSI guidelines [19]. The concentrations of oxacillin used were 0.0125 μg/mL to 128 μg/mL and the concentrations of vancomycin used were 0.06 μg/mL to 32 μg/mL. 2.6. Detection of β-Lactamase Production β-lactamase production in isolated S. aureus was detected by iodometric method as described by Samant and Pai [22]. 2.7. Detection of Inducible Clindamycin Resistance Erythromycin resistant isolates were tested for inducible clindamycin resistance by D-test as per CLSI guidelines [19].

2.6. Detection of β-Lactamase Production β-lactamase production in isolated S. aureus was detected by iodometric method as described by Samant and Pai [22]. 2.7. Detection of Inducible Clindamycin Resistance Erythromycin resistant isolates were tested for inducible clindamycin resistance by D-test as per CLSI guidelines [19]. 2.8. Detection of mecA Gene by Polymerase Chain Reaction (PCR) Conventional phenol: chloroform method [23] was employed for extraction of chromosomal deoxyribonucleic acid (DNA) from the isolates. After optimization, the extracted DNA was subjected to PCR (Figure 1) for detection of mecA gene using PCR profiles described by Abu Shady et al. [24] (Table 1). The primer mecAF (5′-aaaatcgatggtaaaggttggc-3′) and the reverse primer mecAR (5′-agttctggagtaccggatttgc-3′) supplied by Eurogentec were used. 2.9. Quality Control For quality control, Escherichia coli ATCC 25922, S. aureus ATCC 25923, S. aureus ATCC 29213 (mecA negative), and S. aureus ATCC 700699 (mecA positive) were used. 2.10. Data Analysis The data obtained were analyzed with the help of statistical package for social sciences version 16.0. Chi-square test was used to analyze association between two variables and P value less than 0.05 was considered statistically significant. 3. Results Among 711 pus/wound swab samples processed during the study, 110 (15.47%) showed culture positivity for S. aureus. Out of 110 S. aureus, 39 (35.50%) isolates were MRSA by cefoxitin disc diffusion method.

2.10. Data Analysis The data obtained were analyzed with the help of statistical package for social sciences version 16.0. Chi-square test was used to analyze association between two variables and P value less than 0.05 was considered statistically significant. 3. Results Among 711 pus/wound swab samples processed during the study, 110 (15.47%) showed culture positivity for S. aureus. Out of 110 S. aureus, 39 (35.50%) isolates were MRSA by cefoxitin disc diffusion method. 3.1. Antibiotic Susceptibility Patterns of S. aureus Among the methicillin resistant strains, highest rate of susceptibility was seen toward chloramphenicol (100%) followed by mupirocin (97.40%). Similarly, among methicillin sensitive S. aureus (MSSA) strains, highest rate of susceptibility was seen to rifampin and tetracycline (100%) followed by chloramphenicol and mupirocin (98.60%) (Table 2). 3.2. β-Lactamase Production among MRSA and MSSA Beta-lactamase production was observed in 79 (71.82%) isolates of total 110 S. aureus. Of which 52 (65.82%) isolates were MSSA and 27 (34.18%) isolates were MRSA. Statistically, there was no significant association between methicillin resistance and β-lactamase production (P value > 0.05). 3.3. Inducible Clindamycin Resistance among MSSA and MRSA The inducible clindamycin resistance was observed in 11 isolates. Among which, 6 were MSSA and 5 were MRSA. Statistically, there was no significant association between methicillin resistance and inducible clindamycin resistance (P value > 0.05).

3.2. β-Lactamase Production among MRSA and MSSA Beta-lactamase production was observed in 79 (71.82%) isolates of total 110 S. aureus. Of which 52 (65.82%) isolates were MSSA and 27 (34.18%) isolates were MRSA. Statistically, there was no significant association between methicillin resistance and β-lactamase production (P value > 0.05). 3.3. Inducible Clindamycin Resistance among MSSA and MRSA The inducible clindamycin resistance was observed in 11 isolates. Among which, 6 were MSSA and 5 were MRSA. Statistically, there was no significant association between methicillin resistance and inducible clindamycin resistance (P value > 0.05). 3.4. Minimum Inhibitory Concentration of Oxacillin and Vancomycin A total of 35 (31.82%) S. aureus isolates were found to be MRSA by broth microdilution method with MIC cutoff value of 4 μg/mL. Among them, 11 (31.43%) isolates had MIC of >128 μg/mL (high level oxacillin resistant strains). The MIC of oxacillin for S. aureus isolates ranged from 0.032 μg/mL to 256 μg/mL. Only 4 out of 39 MRSA screened by cefoxitin disc diffusion method were found to be susceptible to oxacillin by broth microdilution method. Spearman's correlation between the two phenotypic methods was significant (0.922) at the 0.01 level (2-tailed). Similarly, all S. aureus had MIC of vancomycin below 2 μg/mL (0.016 μg/mL to 1 μg/mL) that is susceptible to vancomycin irrespective to methicillin resistance.

ible to oxacillin by broth microdilution method. Spearman's correlation between the two phenotypic methods was significant (0.922) at the 0.01 level (2-tailed). Similarly, all S. aureus had MIC of vancomycin below 2 μg/mL (0.016 μg/mL to 1 μg/mL) that is susceptible to vancomycin irrespective to methicillin resistance. 3.5. Detection of mecA Gene A total of 32 (29.1%) S. aureus isolates were found to contain mecA gene. All of the mecA containing strains of S. aureus were MRSA by both phenotypic methods, that is, cefoxitin disc diffusion method and oxacillin broth microdilution method. Four out of 39 MRSA screened by cefoxitin disc diffusion method, which were found to be susceptible to oxacillin by broth microdilution method, were not found to contain mecA gene. Further, the gene was found absent on MSSA detected by any of two phenotypic methods. 3.6. Evaluation of Cefoxitin Disc Diffusion and Oxacillin Broth Microdilution Methods in Reference to Presence of mecA Gene MecA gene was found to be absent in 7 of the MRSA detected by cefoxitin disc diffusion method and 3 of the MRSA detected by oxacillin broth microdilution method. The sensitivity of both methods was 100% but the specificity of oxacillin broth microdilution method was greater (96.15%) than that of cefoxitin disc diffusion method (91.03%).

d to be absent in 7 of the MRSA detected by cefoxitin disc diffusion method and 3 of the MRSA detected by oxacillin broth microdilution method. The sensitivity of both methods was 100% but the specificity of oxacillin broth microdilution method was greater (96.15%) than that of cefoxitin disc diffusion method (91.03%). 4. Discussion In our study 35.50% of the isolates were found to be MRSA by cefoxitin disc diffusion method, which was comparable with the findings by Kshetry et al. (37.6%) [8] and Sanjana et al. (39.6%) [25]. But lower prevalence was reported by Subedi and Brahmadathan (15.4%) [26] and Baral et al. (26%) [27] and higher prevalence was reported by Khanal and Jha (68%) [16] and Tiwari et al. (69.1%) [28]. The difference in rates of isolation of MRSA in different studies might be due to the difference in locations and time periods of the studies, difference in hygienic conditions maintained in different hospitals [8], healthcare facilities provided by the hospital, implementation of infection control program, and rational use of antibiotics, which may vary from hospital to hospital [29]. No resistance of MRSA to older drug, chloramphenicol, in our study indicates routine exposure of bacteria to newly developed antibiotics and reversal of susceptibility to outdated antibiotic [30]. The low incidence of mupirocin resistance signifies low usage of the antibiotic [31].

4. Discussion In our study 35.50% of the isolates were found to be MRSA by cefoxitin disc diffusion method, which was comparable with the findings by Kshetry et al. (37.6%) [8] and Sanjana et al. (39.6%) [25]. But lower prevalence was reported by Subedi and Brahmadathan (15.4%) [26] and Baral et al. (26%) [27] and higher prevalence was reported by Khanal and Jha (68%) [16] and Tiwari et al. (69.1%) [28]. The difference in rates of isolation of MRSA in different studies might be due to the difference in locations and time periods of the studies, difference in hygienic conditions maintained in different hospitals [8], healthcare facilities provided by the hospital, implementation of infection control program, and rational use of antibiotics, which may vary from hospital to hospital [29]. No resistance of MRSA to older drug, chloramphenicol, in our study indicates routine exposure of bacteria to newly developed antibiotics and reversal of susceptibility to outdated antibiotic [30]. The low incidence of mupirocin resistance signifies low usage of the antibiotic [31]. In the present study, inducible clindamycin resistance was found in 10% of S. aureus isolates, which was in agreement with the result reported by Ansari et al. (12.4%) [32]. In our study, the occurrence of inducible clindamycin resistance was not significantly different among MRSA and MSSA. However, differentiation of inducible clindamycin resistant phenotypes from others is crucial for therapeutic implication of clindamycin. As use of clindamycin for treatment of the infections caused by such bacteria may result into treatment failure [7], clindamycin should not be used for treatment of such infections; rather it should be used only for the treatment of the infections caused by bacteria which are negative for inducible clindamycin resistance. Clindamycin susceptible strains which are erythromycin resistant may show inducible clindamycin resistance (D-test positive) and it has been suggested that inducible clindamycin resistant strains should be reported as clindamycin resistant [19]. Avoiding the use of clindamycin for the treatment of infections caused by erythromycin resistant strains also omits the chances of treatment failure [33].

ble clindamycin resistance (D-test positive) and it has been suggested that inducible clindamycin resistant strains should be reported as clindamycin resistant [19]. Avoiding the use of clindamycin for the treatment of infections caused by erythromycin resistant strains also omits the chances of treatment failure [33]. In the present study, 71.1% of isolates were beta-lactamase producers by iodometric method. This is low in comparison to finding by Shrestha and Rana in nosocomial S. aureus isolates in Kathmandu and Lalitpur based hospitals [34]. This may be due to high rate of drug resistance among nosocomial isolates. Globally, beta-lactamase production rate lies between 55.7% and 92.6% for Staphylococci [22]. In our study, all the beta-lactamase producers were also resistant to penicillin G. In case of MSSA, penicillin is considered superior to oxacillin to treat S. aureus infections if they are penicillinase nonproducers [35]. Since most of the resistance in S. aureus is secondary to beta-lactamase production and high level production of the enzyme results in development of borderline methicillin resistant Staphylococcus aureus, detection of beta-lactamase in S. aureus is always crucial [36].

ions if they are penicillinase nonproducers [35]. Since most of the resistance in S. aureus is secondary to beta-lactamase production and high level production of the enzyme results in development of borderline methicillin resistant Staphylococcus aureus, detection of beta-lactamase in S. aureus is always crucial [36]. In this study, the sensitivity of both the cefoxitin disc diffusion method and oxacillin broth microdilution method was found to be 100% but specificity of oxacillin broth microdilution method was found to be better. However, cefoxitin disc diffusion is preferred over MIC determination because it is easy to perform and requires no special equipment [37]. MecA gene was not present in some of the strains of MRSA screened by cefoxitin disc diffusion method or oxacillin broth microdilution method. But CLSI guidelines regard the isolates as MRSA if they are found resistant to either cefoxitin or oxacillin or both regardless of the presence of mecA gene [19].

equipment [37]. MecA gene was not present in some of the strains of MRSA screened by cefoxitin disc diffusion method or oxacillin broth microdilution method. But CLSI guidelines regard the isolates as MRSA if they are found resistant to either cefoxitin or oxacillin or both regardless of the presence of mecA gene [19]. Interestingly, isolates (n = 7) which had no mecA gene but were found to be methicillin resistant by phenotypic methods were observed to be beta-lactamase producers. Those isolates (n = 4) which were MRSA by cefoxitin method, but MSSA by oxacillin MIC method, had MIC value of 2 μg/mL. However, the oxacillin MIC value of isolates (n = 3) which were MRSA by both phenotypic methods but had no mecA gene was 4 μg/mL. The possible reason for methicillin resistance in absence of mecA gene may be hyperproduction of β-lactamase [38, 39]. Besides, in a recent study by Ballhausen et al. [40], mecC, a mecA homologue, has also been found to confer methicillin resistance in S. aureus in which mecA gene was absent. Though more research is needed, questions can be raised in considering mecA as sole genetic marker for methicillin resistance. But we could not check the presence of mecC as a possible reason for the phenotypic expression of methicillin resistance in absence of mecA gene. The presence of mecA gene in plasmid of S. aureus isolates has also been reported [41]. Since our study was completely dependent on the detection of mecA on chromosomal DNA, plasmid encoded mecA may have contributed for methicillin resistance in phenotypic tests. Therefore, all the genotypic possibilities should be analyzed for the phenotypic expression of methicillin resistance in S. aureus in order to discover appropriate epidemiological marker of methicillin resistance [42].

osomal DNA, plasmid encoded mecA may have contributed for methicillin resistance in phenotypic tests. Therefore, all the genotypic possibilities should be analyzed for the phenotypic expression of methicillin resistance in S. aureus in order to discover appropriate epidemiological marker of methicillin resistance [42]. In the global scenario, 13 VRSA isolates have been isolated since its first detection in 2002 in USA with scanty reports from India and Iran [43, 44]. The vanA gene responsible for reduced susceptibility of S. aureus toward vancomycin has been found to be transferred from Enterococcus faecalis and E. faecium [44].

osomal DNA, plasmid encoded mecA may have contributed for methicillin resistance in phenotypic tests. Therefore, all the genotypic possibilities should be analyzed for the phenotypic expression of methicillin resistance in S. aureus in order to discover appropriate epidemiological marker of methicillin resistance [42]. In the global scenario, 13 VRSA isolates have been isolated since its first detection in 2002 in USA with scanty reports from India and Iran [43, 44]. The vanA gene responsible for reduced susceptibility of S. aureus toward vancomycin has been found to be transferred from Enterococcus faecalis and E. faecium [44]. In Nepal, there are limited literatures regarding MIC of vancomycin for S. aureus isolated from clinical samples. We reported the MICs of vancomycin for S. aureus to be 0.016 μg/mL to 1 μg/mL. Similarly, Kshetry et al. reported the MICs of vancomycin to MRSA to be 0.125 μg/mL to 1 μg/mL [8]. Slightly higher MICs were reported by Amatya et al. (i.e., 0.5 μg/mL to 2 μg/mL) [45]. Till now no strains of S. aureus resistant to vancomycin have been reported from Nepal [46]. However, four VISA isolates have been reported by Pahadi et al. with MICs of vancomycin to MRSA ranging from 0.5 μg/mL to 4 μg/mL [46]. VISA and VRSA have been reported by many other authors from different countries [8]. Exposure of the S. aureus to vancomycin may be responsible for its reduced susceptibility to the reserve drug and it is attributed to the selective pressure [8]. It is difficult to treat the infections caused by VRSA due to limited antibiotics available for its treatment [8] and it is emerging as a serious public health problem.

f the S. aureus to vancomycin may be responsible for its reduced susceptibility to the reserve drug and it is attributed to the selective pressure [8]. It is difficult to treat the infections caused by VRSA due to limited antibiotics available for its treatment [8] and it is emerging as a serious public health problem. 5. Conclusions On the basis of our findings, both phenotypic methods (cefoxitin disc diffusion and oxacillin broth microdilution) could be used for routine diagnosis of MRSA; however cefoxitin disc diffusion might be preferred over MIC method considering time and labour factor. MRSA and inducible clindamycin resistant S. aureus are emerging as a serious threat to public health in Nepal. Vancomycin can still be used as the drug of choice for treatment of infections caused by MRSA. Acknowledgments The authors would like to thank the Goldengate International College, Kathmandu, Nepal, and Shree Birendra Hospital, Kathmandu, Nepal, for providing opportunity to conduct this study. The authors would also like to thank all the technical staff and the patients for their help during the study. Finally, the authors would like to extend their gratitude to Microbiology Department of Dhulikhel Hospital for providing needed primer for the study. Abbreviations MRSA:Methicillin resistant S. aureus MIC:Minimum inhibitory concentration VRSA:Vancomycin resistant S. aureus VISA:Vancomycin intermediate sensitive S. aureus PBP2a:Low-affinity penicillin binding protein SCCmec:Staphylococcal cassette chromosome mec PCR:Polymerase chain reaction CLSI:Clinical and Laboratory Standards Institute

Abbreviations MRSA:Methicillin resistant S. aureus MIC:Minimum inhibitory concentration VRSA:Vancomycin resistant S. aureus VISA:Vancomycin intermediate sensitive S. aureus PBP2a:Low-affinity penicillin binding protein SCCmec:Staphylococcal cassette chromosome mec PCR:Polymerase chain reaction CLSI:Clinical and Laboratory Standards Institute ATCC:American type culture collection MSSA:Methicillin sensitive S. aureus DNA:Deoxyribonucleic acid. Competing Interests The authors declare that there is no conflict of interests regarding the publication of this paper. Figure 1 Gel electrophoresis showing the PCR products (lane 1 and lane 9: DNA ladder, lane 2: positive control, lane 3: negative control, lane 4: P18, lane 5: P36, lane 6: P53, lane 7: P78, and lane 8: P104). Table 1 Comparison of the phenotypic and genotypic methods for detection of MRSA. Different methods used for detection of MRSA Cefoxitin disc diffusion Oxacillin broth microdilution Polymerase chain reaction Methods to identify MRSA strains Strains of S. aureus having zone of inhibition of  ≤21 mm to cefoxitin disc (30 μg) Strains of S. aureus having oxacillin MIC of ≥4 μg/mL Strains of S. aureus harboring mecA gene Table 2 Antibiotic susceptibility patterns of MSSA and MRSA.

n broth microdilution Polymerase chain reaction Methods to identify MRSA strains Strains of S. aureus having zone of inhibition of  ≤21 mm to cefoxitin disc (30 μg) Strains of S. aureus having oxacillin MIC of ≥4 μg/mL Strains of S. aureus harboring mecA gene Table 2 Antibiotic susceptibility patterns of MSSA and MRSA. Antibiotics MSSA MRSA P value Susceptible (%) Susceptible (%) Erythromycin 33 (46.5) 7 (17.9) 0.003 Clindamycin 57 (80.3) 25 (64.1) 0.062 Gentamicin 64 (90.1) 14 (35.9) 0.000 Ciprofloxacin 37 (52.1) 9 (23.1) 0.003 Chloramphenicol 70 (98.6) 39 (100) 0.457 Cotrimoxazole 30 (42.2) 12 (30.8) 0.236 Mupirocin 70 (98.6) 38 (97.4) 0.664 Rifampin 71 (100) 35 (89.7) 0.006 Tetracycline 71 (100) 34 (87.2) 0.002 Penicillin G 19 (26.8) 0 (0) 0.000

1. Introduction Epididymo-orchitis consists of inflammation of the epididymis and testes. There are nearly 600,000 cases of epididymo-orchitis per year in the United States, accounting for roughly 1/144 (0.69%) outpatient visits among 18- to 50-year-old men, with the majority of those patients aged 18 to 35 years [1]. Mittemeyer et al. conducted a study of 610 cases among soldiers in the United States army and their family members and found that most of the patients were aged 20 to 29 years. However, almost all ages were involved (with a range of 4 months to 76 years) [2]. Acute epididymitis is characterized by acute testicular pain and most commonly involves the symptom of acute scrotum as well. Orchitis usually originates from inflammation of the epididymis [3]. The symptoms of epididymo-orchitis include swelling and tenderness of the epididymis or scrotum in about 75% patients. Some patients even presented with bloodstream infections, sepsis, or septic shock due to the underlying pathogens [4].

otum as well. Orchitis usually originates from inflammation of the epididymis [3]. The symptoms of epididymo-orchitis include swelling and tenderness of the epididymis or scrotum in about 75% patients. Some patients even presented with bloodstream infections, sepsis, or septic shock due to the underlying pathogens [4]. Acute scrotum is a symptom that consists of an acute painful swelling of the scrotum. Epididymo-orchitis and testicular torsion should be considered in making a differential diagnosis when the symptom of acute scrotum occurs [5, 6]. Seasonal variations are known to occur with respect to some infectious diseases and may be due to the epidemiology of the prevalent pathogens, changes in environmental and meteorological factors, and alterations in human behavior. For example, seasonal fluctuations in urinary tract infections have been reported in several studies [7, 8]. The present study sought to evaluate the effect of various meteorological indicators on the incidence of epididymo-orchitis. More specifically, we conducted a 14-year population-based study to evaluate the relationship between meteorological indicators and epididymo-orchitis in Taiwan.

ported in several studies [7, 8]. The present study sought to evaluate the effect of various meteorological indicators on the incidence of epididymo-orchitis. More specifically, we conducted a 14-year population-based study to evaluate the relationship between meteorological indicators and epididymo-orchitis in Taiwan. 2. Methods 2.1. Data Source This study is a nationwide population-based investigation that utilized data from Taiwan's National Health Insurance Research Database (NHIRD). The data in the NHIRD comes from the National Health Insurance (NHI) program, which began in 1995 and covered 99.9% of Taiwan's 23 million residents as of the end of 2013 [9]. All the medical claims data of inpatients and outpatients are included in the NHIRD. More specifically, this study utilized the Longitudinal Health Insurance Database 2000 (LHID2000) [10], a subdataset of the NHIRD. The LHID2000 includes data from January 2000 to December 2013 for a randomly selected sample of one million people out of the 23 million people included in the NHIRD in the year 2000. The sample of patients included in the LHID2000 has a similar demographic distribution and origin to the broader population included in the NHIRD [11]. All the clinical diagnoses in this study were made according to the International Classification of Diseases, 9th revision, Clinical Modification (ICD-9-CM).

year 2000. The sample of patients included in the LHID2000 has a similar demographic distribution and origin to the broader population included in the NHIRD [11]. All the clinical diagnoses in this study were made according to the International Classification of Diseases, 9th revision, Clinical Modification (ICD-9-CM). The meteorological data utilized in this study was provided by Taiwan's Central Weather Bureau (CWB) and consists of data collected from 27 CWB weather stations distributed across various territories of Taiwan (i.e., the islands of Taiwan, Penghu, Kinmen, and Lienchiang). The monthly meteorological data included temperature (measured in degree Celsius), relative humidity (measured in percentage), total rainfall amount (measured in millimeters), total rain days (measured in days), and total sunshine hours (measured in hours) [12]. According to the weather typical of Taiwan, the months of March, April, and May constitute the spring season; June, July, and August constitute the summer; September, October, and November are the fall; and December, January, and February comprise winter. 2.2. Ethics Statement This study was conducted after we received approval from the Institutional Review Board of Chang Gung Memorial Hospital at Linkou (CGMH IRB 103-2071B). As this was a retrospective study and all data was anonymous, the Institutional Review Board department agreed with the authors that it was not necessary to obtain patient consent.

as conducted after we received approval from the Institutional Review Board of Chang Gung Memorial Hospital at Linkou (CGMH IRB 103-2071B). As this was a retrospective study and all data was anonymous, the Institutional Review Board department agreed with the authors that it was not necessary to obtain patient consent. 2.3. Study Subjects The study subjects consisted of patients included in the LHID2000 who were newly diagnosed with epididymitis and orchitis (ICD-9-CM: 604), received a prescription for antibiotic medication, and received testicular sonography examination between January 2000 and December 2013 from the LHID2000 (Figure 1). More specifically, data for patients from administrative regions without CWB weather stations were excluded. In addition, patients who were diagnosed before December 31st, 2000, or after January 1st, 2013 (n = 438), and patients with incomplete demographic data (n = 1,645) were also excluded. Finally, data for a total of 7,223 patients with epididymo-orchitis were collected and analyzed in this study. The diagnoses of epididymo-orchitis were based on detailed clinical examinations, with typical symptoms including painful swelling of the scrotum that may radiate to the lower abdomen, fever, frequency and urgency in voiding, and dysuria. The diagnoses of epididymo-orchitis were made by urologists, infectious disease physicians, or licensed physicians. All the patients included in this study were under antibiotic treatment.

luding painful swelling of the scrotum that may radiate to the lower abdomen, fever, frequency and urgency in voiding, and dysuria. The diagnoses of epididymo-orchitis were made by urologists, infectious disease physicians, or licensed physicians. All the patients included in this study were under antibiotic treatment. 2.4. Statistical Analysis Descriptive statistics for the characteristics of study subjects and meteorological data were first calculated by Student's t-test and Chi-square test; Spearman's rank correlation was used to examine the relationship between the meteorological factors and the monthly incidence rates of epididymo-orchitis. The linear regression model was also used to estimate the relationship between the meteorological factors and monthly incidence rates of epididymo-orchitis. All the tests were two-sided, with p value < 0.05 being regarded as statistically significant. All the statistical analyses were performed with SAS 9.2 software. 3. Results Data for a total of 7,223 patients with epididymo-orchitis was collected and analyzed in this 14-year nationwide population-based study. The demographic characteristics of those patients are listed in Table 1. The mean age of male patients with epididymo-orchitis was 43.46 ± 20.03 years, with a major proportion being aged 20–49 years. Most of the patients lived in Northern and suburban areas of Taiwan.

in this 14-year nationwide population-based study. The demographic characteristics of those patients are listed in Table 1. The mean age of male patients with epididymo-orchitis was 43.46 ± 20.03 years, with a major proportion being aged 20–49 years. Most of the patients lived in Northern and suburban areas of Taiwan. The average monthly incidence rate of epididymo-orchitis was 9.09 per 100,000 population. A comparison of the average monthly epididymo-orchitis incidence rate with the monthly average meteorological factors is shown in Table 2. During the study period from 2000 to 2013, the hottest month was July, with an average temperature of 28.72 degrees Celsius, while the coolest month was January, with an average temperature of 16.90 degrees Celsius. In addition, the month with the most sunshine hours was July, with an average of 223.84 hours, while the month with the least sunshine hours was February, with an average of 125.81 hours. The relative humidity was highest in June, with an average of 79.31%, and the lowest in December, with an average of 73.73%. The highest average total rainfall was in August, with 347.25 mm, while the lowest average total rainfall was in January, with 62.23 mm. The highest average number of rain days was in June, with 13.80 days, and the lowest average number of rain days was in December, with 7.39 days. The monthly incidence of epididymo-orchitis was highest in May and lowest in February (Table 2). The monthly epididymo-orchitis incidence rate and corresponding monthly meteorological factors during the study period are shown in Figure 1.

days, and the lowest average number of rain days was in December, with 7.39 days. The monthly incidence of epididymo-orchitis was highest in May and lowest in February (Table 2). The monthly epididymo-orchitis incidence rate and corresponding monthly meteorological factors during the study period are shown in Figure 1. The correlations between the meteorological factors and the incidence rate of epididymo-orchitis are listed in Table 3. The average temperature, total rainfall amount, and total sunshine hours had statistically significant correlations with the incidence of epididymo-orchitis. However, only average temperature (ß = 0.11, P < 0.001) was found to have a significant linear correlation with the incidence of epididymo-orchitis after regression model analysis (Table 4). 4. Discussion This is the first study to investigate the relationships between different meteorological indicators and the incidence of epididymo-orchitis in Taiwan. The results of this large nationwide population-based study demonstrate that the incidence of epididymo-orchitis is positively correlated with average temperature.

sion This is the first study to investigate the relationships between different meteorological indicators and the incidence of epididymo-orchitis in Taiwan. The results of this large nationwide population-based study demonstrate that the incidence of epididymo-orchitis is positively correlated with average temperature. In this study, average temperature (ß = 0.11, P < 0.001) was found to have a significant linear correlation with the incidence of epididymo-orchitis. Lyronis et al. conducted a study of acute scrotum and found that epididymo-orchitis was the most common cause of acute scrotum and that the incidence of epididymo-orchitis was higher in summer [13]. Our study also revealed that the incidence rates of epididymo-orchitis were higher in May and July. The retrograde ascent of pathogens is the usual cause of epididymo-orchitis. Sexually transmitted Neisseria gonorrhoeae or Chlamydia trachomatis pathogens are the most common causes of epididymo-orchitis in men aged 14 to 35 years, whereas Escherichia coli pathogens along with urinary tract infection constitute the most common cause in men younger than 14 years and older than 35 years of age [14–17].

ly transmitted Neisseria gonorrhoeae or Chlamydia trachomatis pathogens are the most common causes of epididymo-orchitis in men aged 14 to 35 years, whereas Escherichia coli pathogens along with urinary tract infection constitute the most common cause in men younger than 14 years and older than 35 years of age [14–17]. Potential mechanisms for the positive correlation between temperature and epididymo-orchitis infection are discussed below. Increased temperatures increase people's perspiration and total body water loss [18]. In turn, the resulting phenomena of relative dehydration and more concentrated urine with less frequent voiding increase the rate of urinary tract infections. Increased temperatures also cause peripheral vasodilation and the pooling of blood in the skin, leading to a decrease in the effective blood volume. A hot environment may also cause peripheral vasodilation and pooling of blood in the skin which means that exposure to heat can cause a decrease in effective blood volume. In contrast, exposure to cold has also been found to cause changes in the vasopressin system leading to diuresis and the increased clearance of potential pathogens of the urinary tract [19]. In addition, seasonal variations in gonorrhea and Chlamydia infections in adolescents have previously been reported, including higher positive tests rates in the summer and fall [20].

ause changes in the vasopressin system leading to diuresis and the increased clearance of potential pathogens of the urinary tract [19]. In addition, seasonal variations in gonorrhea and Chlamydia infections in adolescents have previously been reported, including higher positive tests rates in the summer and fall [20]. Epididymo-orchitis and testicular torsion are both accompanied by acute scrotal pain or swelling of the scrotum. Differential diagnosis between these two diseases is important due to possibility of torsion resulting in testicular infarction, which constitutes a surgical emergency. The seasonality of testicular torsion has previously been documented in several studies. Chiu et al. conducted a 10-year cohort study with 1,782 testicular torsion patients that found that January had a significantly higher incidence rate of torsion. In addition, the incidence rate of testicular torsion was negatively associated with temperature [21]. Relatedly, Mabogunje reported an increased incidence of testicular torsion between November and February [22]. Several studies have also reported increased incidences of testicular torsion during low temperature periods in Japan, the United States, Greece, and Nigeria [13, 22–24]. In contrast, the current study found that the incidence of epididymo-orchitis was increased in May and July, while a positive correlation between epididymo-orchitis and temperature was also noted. These opposing seasonality characteristics of epididymo-orchitis and testicular torsion may be helpful in distinguishing the two diseases from each other.

found that the incidence of epididymo-orchitis was increased in May and July, while a positive correlation between epididymo-orchitis and temperature was also noted. These opposing seasonality characteristics of epididymo-orchitis and testicular torsion may be helpful in distinguishing the two diseases from each other. A study conducted in the United State found that nearly 60,000 male adults made visits to office-based physicians for epididymo-orchitis, accounting for 1/350 (0.29%) of all the visits in the period under consideration [25]. In another nationwide study in the United States, epididymo-orchitis accounted for 1/144 (0.69%) of outpatient visits among 18- to 50-year-old men [1]. The 9.09 per 100,000 population incidence of epididymo-orchitis found in the current study was lower than previous studies. The population resident in the area without CWB weather stations was excluded for having no meteorological data available. Some epididymo-orchitis patients among this population may be excluded also. This may affect the calculated incidence of epididymo-orchitis. The general climate is Marine Subtropical or tropical climate in Taiwan with a relative small change of temperature between summer and winter. Ratkowsky et al. demonstrated a linear relationship between the growth rate of bacterial culture and temperature even within a small range of temperature in the laboratory [26]. Thus, temperature, even a relative small range of temperature, plays an important role in growth of bacteria.

ature between summer and winter. Ratkowsky et al. demonstrated a linear relationship between the growth rate of bacterial culture and temperature even within a small range of temperature in the laboratory [26]. Thus, temperature, even a relative small range of temperature, plays an important role in growth of bacteria. The strength of the current study is that it utilized a longitudinal nationwide database with large number of subjects and a long follow-up period. However, the study also had several limitations that should also be discussed. First, the diagnosis of epididymo-orchitis was recognized according to the ICD-9 code. The ICD-9-CM code 604.90 is used for both epididymitis and orchitis. Thus, it is difficult to distinguish epididymitis from orchitis using the ICD-9 code. Second, data from diagnostic laboratory tests that can help confirm diagnoses of epididymitis and orchitis, such as urinalysis, urine culture, and C-reactive protein (CRP) level data, are not included in the NHIRD. Therefore, epididymo-orchitis caused by sexually transmitted diseases pathogens may be included in this study. The detailed reports of color Doppler ultrasonography are also not included. Third, the data regarding the meteorological indicators came from 27 weather stations monitored by Taiwan's CWB, but these 27 stations did not cover every territory of Taiwan. Finally, this study utilized a retrospective study design. Further prospective studies are thus warranted to investigate the relationship between weather and epididymo-orchitis.

eorological indicators came from 27 weather stations monitored by Taiwan's CWB, but these 27 stations did not cover every territory of Taiwan. Finally, this study utilized a retrospective study design. Further prospective studies are thus warranted to investigate the relationship between weather and epididymo-orchitis. 5. Conclusion In conclusion, various meteorological indicators, especially average temperature, appeared to significantly affect the incidence of epididymo-orchitis in Taiwan. A significant positive linear correlation was discovered between temperature and patients with epididymo-orchitis. This result may serve as a useful clinical clue to help physicians distinguish between patients with epididymo-orchitis and testicular torsion in hot or cold weather. Acknowledgments The authors acknowledge support from the Taiwan Typhoon and Flood Research Institute, National Applied Research Laboratories, which provided data from the Data Bank for Atmospheric & Hydrologic Research. This work was also supported by grants from the Taoyuan General Hospital, Ministry of Health and Welfare (10322), and Ministry of Science and Technology, Taiwan (MOST 104-2320-B-016-012-MY3). Competing Interests The authors declare that they have no conflict of interests.

Acknowledgments The authors acknowledge support from the Taiwan Typhoon and Flood Research Institute, National Applied Research Laboratories, which provided data from the Data Bank for Atmospheric & Hydrologic Research. This work was also supported by grants from the Taoyuan General Hospital, Ministry of Health and Welfare (10322), and Ministry of Science and Technology, Taiwan (MOST 104-2320-B-016-012-MY3). Competing Interests The authors declare that they have no conflict of interests. Authors' Contributions Jui-Ming Liu wrote the manuscript and conducted the data analysis. Fung-Wei Chang and See-Tong Pang wrote the proposal and designed the manuscript. Ying-Hsu Chang and Te-Wei Ho contributed to the conception of the study. Ren-Jun Hsu made data analysis. Po-Hung Lin revised the manuscript. All authors contributed to data analysis, drafting, and critically revising the paper, read and approved the final manuscript, and agreed to be accountable for all aspects of the work. Ren-Jun Hsu and Po-Hung Lin contributed equally to this work. Figure 1 The monthly incidence of epididymo-orchitis and the corresponding monthly weather data during the study period. (a) Epididymo-orchitis incidence and monthly average temperature. (b) Epididymo-orchitis incidence and monthly total sunshine hours. (c) Epididymo-orchitis incidence and monthly relative humidity. (d) Epididymo-orchitis incidence and monthly total rain fall. (e) Epididymo-orchitis incidence and monthly total rain days. Table 1 Demographic characteristics of study subjects of epididymo-orchitis from 2000 to 2013 in Taiwan.

Figure 1 The monthly incidence of epididymo-orchitis and the corresponding monthly weather data during the study period. (a) Epididymo-orchitis incidence and monthly average temperature. (b) Epididymo-orchitis incidence and monthly total sunshine hours. (c) Epididymo-orchitis incidence and monthly relative humidity. (d) Epididymo-orchitis incidence and monthly total rain fall. (e) Epididymo-orchitis incidence and monthly total rain days. Table 1 Demographic characteristics of study subjects of epididymo-orchitis from 2000 to 2013 in Taiwan. Characteristics n (%) Number of cases 7,223 Mean age (years) (standard deviation) 43.46 (20.03) Age at diagnosis (years) <20 751 (10.40) 20–29 1,281 (17.74) 30–39 1,371 (18.98) 40–49 1,240 (17.17) 50–59 922 (12.76) 60–69 657 (9.10) ≥70 1,001 (13.85) Insured region Northern 3,811 (52.76) Central 1,153 (15.96) Southern 2,056 (28.46) Eastern 203 (2.82) Urbanicity Urban 2,623 (36.31) Suburban 4,105 (56.84) Rural 484 (6.70) Missing 11 (0.15) Table 2 Average epididymo-orchitis incidence rate and meteorological factors according to month.

<20 751 (10.40) 20–29 1,281 (17.74) 30–39 1,371 (18.98) 40–49 1,240 (17.17) 50–59 922 (12.76) 60–69 657 (9.10) ≥70 1,001 (13.85) Insured region Northern 3,811 (52.76) Central 1,153 (15.96) Southern 2,056 (28.46) Eastern 203 (2.82) Urbanicity Urban 2,623 (36.31) Suburban 4,105 (56.84) Rural 484 (6.70) Missing 11 (0.15) Table 2 Average epididymo-orchitis incidence rate and meteorological factors according to month. Month  Monthly incidence rate Temperature (°C) Total sunshine hours Relative humidity (%) Total rainfall (mm) Total rain days Mean 95% CI Mean (°C) Max (°C) Minimum (°C) Jan 8.09 (7.54, 8.64) 16.90 21.90 10.60 128.48 75.95 62.23 8.29 Feb 7.36 (6.53, 8.20) 18.17 23.50 10.80 125.81 77.64 66.78 7.78 March 9.78 (8.77, 10.79) 19.79 24.30 12.80 139.23 75.42 77.97 9.50 April 9.46 (8.43, 10.50) 22.94 26.80 17.10 128.83 77.24 108.26 10.84 May 10.27 (9.23, 11.31) 25.86 28.70 20.50 160.26 77.85 213.51 12.09 June 9.07 (8.34, 9.82) 27.52 29.80 21.60 167.20 79.31 315.79 13.80 July 9.51 (8.37, 10.66) 28.72 30.80 22.50 223.84 77.14 291.61 11.28 Aug 9.58 (8.53, 10.62) 28.48 30.10 22.30 196.28 78.31 347.25 13.52 Sep 8.88 (7.96, 9.80) 27.39 30.00 21.20 171.50 77.44 283.08 11.52 Oct 9.46 (8.71, 10.20) 24.92 28.00 19.00 175.30 73.97 101.41 6.51 Nov 8.87 (8.16, 9.58) 22.06 25.60 16.90 139.00 75.03 95.65 7.87 Dec 8.73 (7.99, 9.48) 18.41 22.90 11.60 137.63 73.73 71.05 7.39 Table 3 Spearman's rank correlations between meteorological factors and monthly epididymo-orchitis incidence rates. Variable r Temperature 0.28∗∗∗ Total sunshine hours 0.17∗ Relative humidity 0.08 Total rainfall 0.23∗∗ Total rain days 0.14

Month  Monthly incidence rate Temperature (°C) Total sunshine hours Relative humidity (%) Total rainfall (mm) Total rain days Mean 95% CI Mean (°C) Max (°C) Minimum (°C) Jan 8.09 (7.54, 8.64) 16.90 21.90 10.60 128.48 75.95 62.23 8.29 Feb 7.36 (6.53, 8.20) 18.17 23.50 10.80 125.81 77.64 66.78 7.78 March 9.78 (8.77, 10.79) 19.79 24.30 12.80 139.23 75.42 77.97 9.50 April 9.46 (8.43, 10.50) 22.94 26.80 17.10 128.83 77.24 108.26 10.84 May 10.27 (9.23, 11.31) 25.86 28.70 20.50 160.26 77.85 213.51 12.09 June 9.07 (8.34, 9.82) 27.52 29.80 21.60 167.20 79.31 315.79 13.80 July 9.51 (8.37, 10.66) 28.72 30.80 22.50 223.84 77.14 291.61 11.28 Aug 9.58 (8.53, 10.62) 28.48 30.10 22.30 196.28 78.31 347.25 13.52 Sep 8.88 (7.96, 9.80) 27.39 30.00 21.20 171.50 77.44 283.08 11.52 Oct 9.46 (8.71, 10.20) 24.92 28.00 19.00 175.30 73.97 101.41 6.51 Nov 8.87 (8.16, 9.58) 22.06 25.60 16.90 139.00 75.03 95.65 7.87 Dec 8.73 (7.99, 9.48) 18.41 22.90 11.60 137.63 73.73 71.05 7.39 Table 3 Spearman's rank correlations between meteorological factors and monthly epididymo-orchitis incidence rates. Variable r Temperature 0.28∗∗∗ Total sunshine hours 0.17∗ Relative humidity 0.08 Total rainfall 0.23∗∗ Total rain days 0.14 ∗∗∗ P < 0.001, ∗∗P < 0.01, ∗P < 0.05. Table 4 Regression model of meteorological factors and monthly epididymo-orchitis incidence rates. Variable ß Temperature 0.11∗∗∗ Total sunshine hours 0.01∗ Relative humidity 0.04 Total rainfall 0.002∗ Total rain days 0.06 ß: coefficient of regression analysis. ∗∗∗ P < 0.001, ∗P < 0.05.

1. Introduction Gram-positive bacteria are the most predominant microorganisms associated with sepsis in healthcare settings and the most prevalent cause of bacteremia in patients with hematopoietic stem cell transplantation [1, 2]. Enterococcal bacteremia is associated with increased risk of mortality in patients with hematopoietic stem cell transplantation, irrespective of susceptibility to vancomycin [3, 4]. In cardiology units, infective endocarditis, infectious aneurysm, catheter-related bloodstream infections, or surgical site infections after cardiac surgeries are the major infections of which the most common causative microorganisms are Gram-positive cocci [5–10]. In both medical units, early detection of Gram-positives and resistant markers is very critical in managing patient care, antibiotic stewardship, and preventing spread of resistant microorganisms.

ardiac surgeries are the major infections of which the most common causative microorganisms are Gram-positive cocci [5–10]. In both medical units, early detection of Gram-positives and resistant markers is very critical in managing patient care, antibiotic stewardship, and preventing spread of resistant microorganisms. As early intervention with antimicrobial therapy is associated with improved prognosis with each hour of delay associated with increased mortality, rapid diagnosis is critical [11]. The Verigene Gram-positive blood culture assay (BC-GP) (Nanosphere, Inc., Northbrook, IL) is a sample to result microarray system for identification of common Gram-positive bacteria and major resistance markers directly from positive blood culture. Although a number of studies have previously evaluated BC-GP reporting performance ranging from 92 to 99% agreement with conventional methodology [12–16], one limitation of previous reports has been that many of the studies have been from the United States as well as countries outside of Japan with only one published report of limited scope in Japan [17]. Genetic variation among bacterial lineages circulating in different geographical regions of the world can affect the sensitivity of molecular assays based on oligonucleotide probes to detect organisms or resistance markers. Studies in Hong Kong and Belgium have reported lower BC-GP performance [18, 19].

As early intervention with antimicrobial therapy is associated with improved prognosis with each hour of delay associated with increased mortality, rapid diagnosis is critical [11]. The Verigene Gram-positive blood culture assay (BC-GP) (Nanosphere, Inc., Northbrook, IL) is a sample to result microarray system for identification of common Gram-positive bacteria and major resistance markers directly from positive blood culture. Although a number of studies have previously evaluated BC-GP reporting performance ranging from 92 to 99% agreement with conventional methodology [12–16], one limitation of previous reports has been that many of the studies have been from the United States as well as countries outside of Japan with only one published report of limited scope in Japan [17]. Genetic variation among bacterial lineages circulating in different geographical regions of the world can affect the sensitivity of molecular assays based on oligonucleotide probes to detect organisms or resistance markers. Studies in Hong Kong and Belgium have reported lower BC-GP performance [18, 19]. The purpose of this study was to determine the potential impact on patient management and patient outcome of BC-GP in specialized hospitalized settings within the Japanese healthcare environment, which faces a number of challenges. An increasingly aging population and the accompanying burden of increasing overall healthcare costs have challenged the healthcare infrastructure. Despite methicillin-resistant Staphylococcus aureus (MRSA) accounting for over 90% of the hospital-acquired infections caused by resistant bacteria in Japan [20], outsourcing of clinical microbiology testing or no weekend coverage is common in many healthcare facilities as part of cost-containment. This paper represents the first comprehensive evaluation of BC-GP in Japan to validate its clinical performance. The simulated blood culture study includes a library of well-characterized healthcare-associated MRSA (HA-MRSA), community-acquired MRSA (CA-MRSA), and vancomycin-resistant enterococci (VRE) strains circulating in Japan.

presents the first comprehensive evaluation of BC-GP in Japan to validate its clinical performance. The simulated blood culture study includes a library of well-characterized healthcare-associated MRSA (HA-MRSA), community-acquired MRSA (CA-MRSA), and vancomycin-resistant enterococci (VRE) strains circulating in Japan. 2. Methods BC-GP was evaluated at Toranomon Hospital (TH) and Sakakibara Heart Institute (SHI), in accordance with site-specific institutional review board approved study protocols, during June 26, 2012, to March 6, 2013. TH is an 1168-bed general teaching hospital with a 123-bed hematological care unit for hematopoietic stem cell transplantation performing 140 to 160 hematopoietic stem cell transplants per year. The microbiology laboratory at the hospital operates daily during day hours. SHI is a 320-bed teaching hospital specializing in cardiovascular diseases with a caseload of over 1,500 open-heart surgeries per year. The microbiology laboratory in the hospital is operated by an outside commercial reference laboratory. The microbiology laboratory operates during the day shift on weekdays and closed on weekends and holidays. Blood cultures were performed at SHI using BacT/ALERT FA bottles and monitoring with BacT/ALERT 3D (bioMérieux, Marcy l'Etoile, France). TH used BACTEC Plus bottles and monitoring with BACTEC 9240 and FX (Becton Dickinson, Franklin Lakes, and NJ). Only one positive blood culture bottle containing Gram-positive cocci or bacilli per patient was included in the study. Two ml of positive blood culture medium was stored at −85C for retesting.

toile, France). TH used BACTEC Plus bottles and monitoring with BACTEC 9240 and FX (Becton Dickinson, Franklin Lakes, and NJ). Only one positive blood culture bottle containing Gram-positive cocci or bacilli per patient was included in the study. Two ml of positive blood culture medium was stored at −85C for retesting. Routine microbiological identification and susceptibility testing of isolates were performed using conventional identification tests such as bile solubility, optochin disk susceptibility, and the MicroScan WalkAway system (Beckman Coulter, Pasadena, CA) at TH and the Vitek 2 system (bioMérieux, Marcy l'Etoile, France) at SHI. Cefoxitin screening for methicillin resistance was performed according to CLSI guidelines [21]. A latex agglutination test for detection of penicillin-binding protein PBP2a was also utilized [22]. BC-GP testing was performed on positive blood culture showing Gram-positive organisms according to the manufacturer's instructions. Briefly, a well-mixed 350 μl sample of the blood culture media was pipetted into the sample well of the BC-GP nucleic extraction tray, placed onto the Verigene Processor SP for processing and analysis by the Verigene Reader.

ive blood culture showing Gram-positive organisms according to the manufacturer's instructions. Briefly, a well-mixed 350 μl sample of the blood culture media was pipetted into the sample well of the BC-GP nucleic extraction tray, placed onto the Verigene Processor SP for processing and analysis by the Verigene Reader. An assessment was performed to determine the difference in time between reporting of results using BC-GP and culture-based identification and antimicrobial susceptibility results for 139 positive blood culture broths. For culture-based results, the time required until generation of the final report was the time between the Gram stain reading and entering of final identification and susceptibility results into the laboratory information system. For BC-GP, the time to result was the time between the Gram stain reading and entering of BC-GP results into the laboratory information system.

neration of the final report was the time between the Gram stain reading and entering of final identification and susceptibility results into the laboratory information system. For BC-GP, the time to result was the time between the Gram stain reading and entering of BC-GP results into the laboratory information system. A challenge set of 208 simulated blood cultures was constructed using type, reference strains, and clinical strains from different geographical regions in Japan to evaluate BC-GP. The clinical strains included organisms that presented challenges for commercial identification systems in previous clinical studies at TH and SHI. In addition, a library of well-characterized HA-MRSA, CA-MRSA, and VRE strains from Japan was also tested. Simulated blood culture studies were performed at Miroku Medical Laboratory (Saku City, Nagano Prefecture, Japan). Two hundred and eight strains were adjusted to a turbidity of approximately 100 CFU/ml in sterile saline. Three hundred μl was inoculated into BACTEC Plus Aerobic/F bottles containing 8 to 10 ml of human whole blood (blood type O, Tennessee Blood Services, Memphis, TN) for a final inoculum of 30 CFU/bottle. A BACTEC Plus Anaerobic/F bottle was also inoculated for S. pneumoniae and the S. anginosus group. Each bottle was incubated in the BACTEC system until a positive signal was generated. If BC-GP generated negative results, 11-fold dilution of the blood culture medium using sterile distilled water was retested.

U/bottle. A BACTEC Plus Anaerobic/F bottle was also inoculated for S. pneumoniae and the S. anginosus group. Each bottle was incubated in the BACTEC system until a positive signal was generated. If BC-GP generated negative results, 11-fold dilution of the blood culture medium using sterile distilled water was retested. Each of the positive blood culture isolates at the two hospital sites was stored in 10% skim milk (Difco) at −85C. Species identification was confirmed using matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) (Microflex LT with Biotyper ver. 3.0 software; Bruker Daltonik GmbH, Bremen, Germany) for all strains [23]. If the organism was not identified to the species level by MALDI-TOF MS (score value < 2.0) or identified as Micrococcus, Listeria, Staphylococcus other than S. aureus, Streptococcus other than S. pyogenes, and S. agalactiae, confirmatory testing by PCR-direct sequencing of 16S rDNA or sodA was performed at Juntendo University or Tokyo Women's Medical University [24–26]. Specific-PCR for detecting mecA [27] in all staphylococci and vanA, vanB in all enterococci was performed [28]. The methods used for SCCmec typing of community-acquired or healthcare-associated MRSA used to characterized strains have been previously described [29, 30].

rsity or Tokyo Women's Medical University [24–26]. Specific-PCR for detecting mecA [27] in all staphylococci and vanA, vanB in all enterococci was performed [28]. The methods used for SCCmec typing of community-acquired or healthcare-associated MRSA used to characterized strains have been previously described [29, 30]. Concordance was determined in comparison to results of the reference methods. There was agreement if BC-GP target detection agreed with the reference method at either the genus or the species level. The ninety-five percent confidence intervals (95% CI) and the paired t-test were determined using GraphPad StatMate (GraphPad Software Inc., San Diego, CA). 3. Results In this study, the overall identification agreement between BC-GP and the reference method was 327/347 (94%) for prospective blood cultures at two clinical sites and simulated blood cultures. The combined agreement between PCR and BC-GP for mecA detection was 71/73 (97%) for prospective blood cultures and simulated blood cultures. The combined identification accuracy from both hospital sites was 129/139 (93%). For monomicrobial cultures, the identification accuracy was 121/124 (98%). Agreement between PCR and BC-GP for mecA positivity was 51/53 (96%). Table 1 shows the results from TH. Overall, 96/104 (92%) organisms were correctly identified by BC-GP to the species or genus level including detection of resistance genes. As shown in Table 2, total agreement of BC-GP with the reference method at SHI was 33/35 (94%) organisms.

C-GP for mecA positivity was 51/53 (96%). Table 1 shows the results from TH. Overall, 96/104 (92%) organisms were correctly identified by BC-GP to the species or genus level including detection of resistance genes. As shown in Table 2, total agreement of BC-GP with the reference method at SHI was 33/35 (94%) organisms. BC-GP reports the presence of mecA for only S. aureus and S. epidermidis. In this study, of the 102 staphylococcal strains, 72 were either S. aureus or S. epidermidis. Discordant results were due to undetectable mecA S. epidermidis organisms in polymicrobial cultures containing both mecA positive and mecA negative S. epidermidis. Of the 30 Staphylococcus spp., other than S. epidermidis and S. aureus, 21 (70%) were positive for mecA, including 2 S. lugdunensis, which could not be reported as mecA positive by BC-GP.

undetectable mecA S. epidermidis organisms in polymicrobial cultures containing both mecA positive and mecA negative S. epidermidis. Of the 30 Staphylococcus spp., other than S. epidermidis and S. aureus, 21 (70%) were positive for mecA, including 2 S. lugdunensis, which could not be reported as mecA positive by BC-GP. Table 3 shows the difference in time between the generation of BC-GP results and culture-based final identification and antimicrobial susceptibility results at both hospitals. BC-GP results were available at a mean of 28.2 to 51.0 hours before culture-based final identification and susceptibility results at TH. At SHI, BC-GP generated results at a mean of 34.5 to 196.6 hours earlier. In comparing the time to final culture-based identification and susceptibility results at TH and SHI, with the exception of S. aureus, results for S. epidermidis and coagulase-negative staphylococci other than S. epidermidis, enterococci, and streptococci required significantly (P < 0.05) longer time at SHI (83.3, 123.6, 159.1, and 199.1 hours, resp.) compared to TH (40.8, 53.9, 36.1, and 53.5 hours, resp.).

H and SHI, with the exception of S. aureus, results for S. epidermidis and coagulase-negative staphylococci other than S. epidermidis, enterococci, and streptococci required significantly (P < 0.05) longer time at SHI (83.3, 123.6, 159.1, and 199.1 hours, resp.) compared to TH (40.8, 53.9, 36.1, and 53.5 hours, resp.). Using simulated Gram-positive blood cultures, BC-GP correctly identified 198/208 (95%) of the organisms (Table 4). Six streptococci (3%) were either incorrectly identified or identified at the genus level only by BC-GP. With respect to the 4 BC-GP false-negative blood culture bottles, 1 S. pyogenes was correctly identified and 1 S. mitis generated a positive Streptococcus genus/S. pneumoniae signal following 11-fold dilution of the blood culture medium. The mecA gene was detected in 20/20 (100%) MRSA organisms representing community-acquired (SCCmec type IIa, IV, and V) and healthcare-associated (SCCmec type I, IIb, III, and nontypeable) strains by BC-GP. BC-GP detected 14/14 (100%) vanA and 20/20 (100%) vanB genes in well-characterized VRE strains from previous studies in Japan. 4. Discussion The performance of BC-GP observed in our study was similar to previous reports [12–16, 31–33]. As clinical microbiology services are outsourced or not operating during off-shifts during weekdays and closed on weekends and holidays in many hospitals in Japan, rapid diagnostic testing has significant potential to impact patient care by dramatically decreasing the time to organism identification and antimicrobial susceptibility results.

gy services are outsourced or not operating during off-shifts during weekdays and closed on weekends and holidays in many hospitals in Japan, rapid diagnostic testing has significant potential to impact patient care by dramatically decreasing the time to organism identification and antimicrobial susceptibility results. BC-GP detects mecA in all staphylococci based on measurement of signal intensity; however, reporting is restricted to S. aureus and S. epidermidis based on an algorithm in which mecA is only reported when S. aureus or S. epidermidis is detected by BC-GP. Future versions of BC-GP should consider modification of the algorithm to allow reporting of mecA detection for staphylococci other than S. aureus or S. epidermidis as 70% of the 30 non-S. aureus and S. epidermidis strains in our study were methicillin-resistant. Although S. epidermidis is the major coagulase-negative staphylococcal pathogen, other staphylococci such as S. lugdunensis and S. haemolyticus are important pathogens in the healthcare environment [34, 35].

. epidermidis as 70% of the 30 non-S. aureus and S. epidermidis strains in our study were methicillin-resistant. Although S. epidermidis is the major coagulase-negative staphylococcal pathogen, other staphylococci such as S. lugdunensis and S. haemolyticus are important pathogens in the healthcare environment [34, 35]. Polymicrobial positive blood cultures generated the majority of the discordance. In contrast, the performance of BC-GP was 121/124 (98%) in monomicrobial clinical blood cultures and 198/208 (95%) in simulated blood cultures. In this study, polymicrobial blood cultures accounted for 14/106 (13%) and 3/33 (9%), respectively, of the positive blood cultures at TH and SH. Combined, polymicrobial cultures represented 17/139 (12%) of the prospective clinical blood cultures which is consistent with previous studies reporting 6 to 20% of all bloodstream infections to be polymicrobial [36–38]. BC-GP correctly identified all of the organisms in 12/17 (70%) of polymicrobial cultures. In previous BC-GP studies, the correct identification rates in polymicrobial cultures ranged from 57 to 86% [14–16, 31, 32]. As misleading information can affect clinical diagnosis, resulting in inappropriate selection of antimicrobial agents, there is a need to understand the limitations of BC-GP.

ymicrobial cultures. In previous BC-GP studies, the correct identification rates in polymicrobial cultures ranged from 57 to 86% [14–16, 31, 32]. As misleading information can affect clinical diagnosis, resulting in inappropriate selection of antimicrobial agents, there is a need to understand the limitations of BC-GP. An additional concern with polymicrobial cultures is the clinical interpretation of mecA detection and staphylococcal detection. In our 17-polymicrobial cultures, 5 samples yielded 2 or 3 staphylococcal strains with or without mecA. This may lead to unnecessary use of vancomycin or underestimating infection caused by methicillin-resistant staphylococci. Repeating BC-GP on another set of blood cultures may lessen the risk of this problem. For monomicrobial cultures or simulated cultures, all of the discrepancies were observed with streptococci except for 1 S. caprae strain. BC-GP misidentification for streptococci included 2 S. mitis identified as S. pneumoniae, no detection of 2 S. pneumoniae, 2 S. anginosus group, and 2 S. pyogenes. Previous reports have also reported similar results for S. mitis, S. oralis, and S. pneumoniae [14–16, 39]. As genetic relatedness among S. mitis, S. oralis, and S. pneumoniae is well known based on >99% homology of 16S rRNA gene sequences [24], BC-GP results for S. pneumoniae or Streptococcus with alpha-hemolysis without any positive species-specific signals should be carefully interpreted and confirmed by conventional methods such as optochin sensitivity or the bile solubility test. Interestingly, 11-fold dilution of the original blood culture medium can lead to detection of the Streptococcus signal and species signal (Table 4). A range of detection depending on the organism by BC-GP has been reported [40].

confirmed by conventional methods such as optochin sensitivity or the bile solubility test. Interestingly, 11-fold dilution of the original blood culture medium can lead to detection of the Streptococcus signal and species signal (Table 4). A range of detection depending on the organism by BC-GP has been reported [40]. The major benefit of utilizing BC-GP is the earlier time for reporting identification and resistance determinants from positive blood cultures allowing for earlier selection of appropriate antimicrobial therapy and implementation of infection control measures such as isolation and contact precaution [39, 41]. This has the potential of making a significant impact in the Japanese healthcare delivery system. The difference in time between BC-GP results and final culture-based identification and susceptibility results shown in Table 3 at TH is consistent with previous reports [14–16, 18, 32, 41]. On the other hand, earlier results of 80.7 to 196.6 hours for organisms other than S. aureus using BC-GP at SHI reflects the unavailability of clinical microbiology services during weekends and holidays. As clinical microbiology services are outsourced in many hospitals in Japan or limited to one shift during weekdays or not offered during weekends, the potential cost-benefits of retaining blood culture services in hospitals are significant. Furthermore, training of laboratory personnel in the general laboratory to recognize Gram-positive cocci and bacilli will allow BC-GP testing over the weekend as well as evenings/nights. BC-GP provides the opportunity for hospitals to retain in-house a very critical laboratory service.

ces in hospitals are significant. Furthermore, training of laboratory personnel in the general laboratory to recognize Gram-positive cocci and bacilli will allow BC-GP testing over the weekend as well as evenings/nights. BC-GP provides the opportunity for hospitals to retain in-house a very critical laboratory service. In conclusion, BC-GP provided accurate identification and detection of resistance markers compared with routine culture-based laboratory methods for Gram-positive organisms including CA-MRSA, HA-MRSA, and VRE strains circulating in Japan. Minimizing the time to optimizing antimicrobial therapy using BC-GP may contribute to reduced costs and improved patient care. In 2016, BC-GP received regulatory approval in Japan becoming the first multitarget molecular test for positive blood cultures approved as an in vitro diagnostic device to aid in the diagnosis of bacterial bloodstream infections. Additional studies will be necessary to validate the cost-effectiveness of BC-GP within the context of the Japanese healthcare delivery system. Acknowledgments This study was supported, in part, by a Grant-in-Aid (S0991013) from the Ministry of Education, Culture, Sport, Science, and Technology, Japan (MEXT), for the Foundation of Strategic Research Projects in Private Universities. Ethical Approval This study was approved by the TH and SHI internal review boards. Competing Interests All the authors declare no competing interests.

Acknowledgments This study was supported, in part, by a Grant-in-Aid (S0991013) from the Ministry of Education, Culture, Sport, Science, and Technology, Japan (MEXT), for the Foundation of Strategic Research Projects in Private Universities. Ethical Approval This study was approved by the TH and SHI internal review boards. Competing Interests All the authors declare no competing interests. Authors' Contributions Ken Kikuchi designed and carried out the study and drafted the manuscript. Mari Matsuda, Shigekazu Iguchi, Tomonori Mizutani, Kaori Sansaka, Kenta Negishi, Kimie Shimada, Shigeyuki Notake, Hideji Yanagisawa, and Reiko Yabusaki carried out the laboratory works. Keiichi Hiramatsu, Michiru Tega-Ishii, Jun Umemura, Hiroshi Takahashi, Hideki Araoka, and Akiko Yoneyama supervised the data collection and coordinated and participated in designing the study. Table 1 Performance of the BC-GP assay at Toranomon Hospital. Organism BC-GP (total) BC-GP (monomicrobial cultures) Number of organisms Number (%) of isolates Number of organisms Number (%) of isolates Correctly identified Not detected Incorrectly identified Correctly identified Not detected Incorrectly identified

Authors' Contributions Ken Kikuchi designed and carried out the study and drafted the manuscript. Mari Matsuda, Shigekazu Iguchi, Tomonori Mizutani, Kaori Sansaka, Kenta Negishi, Kimie Shimada, Shigeyuki Notake, Hideji Yanagisawa, and Reiko Yabusaki carried out the laboratory works. Keiichi Hiramatsu, Michiru Tega-Ishii, Jun Umemura, Hiroshi Takahashi, Hideki Araoka, and Akiko Yoneyama supervised the data collection and coordinated and participated in designing the study. Table 1 Performance of the BC-GP assay at Toranomon Hospital. Organism BC-GP (total) BC-GP (monomicrobial cultures) Number of organisms Number (%) of isolates Number of organisms Number (%) of isolates Correctly identified Not detected Incorrectly identified Correctly identified Not detected Incorrectly identified Staphylococcus 78 73 (94) 4 (5) 1 (1) 71 70 (99) 1 (2) S. aureus 16 16 (100)     16 16 (100) Methicillin-sensitive 9 9 (100)     9 9 (100) Methicillin-resistant 7 7 (100)     7 7 (100) S. epidermidis 37 34 (92) 2 (5) 1 (3) 34 34 (100) Methicillin-sensitive 2 1 (50) 1 (50)a   1 1 (100) Methicillin-resistant 35 33 (94) 1 (3)b 1 (3)c 33 33 (100) S. lugdunensis 1 1 (100)     1 1 (100) Other CNS 24 22 (92) 2 (8)   20 19 (95) 1 (6) S. caprae 9 8 (89) 1 (11)   7 6 (67) 1 (33) S. hominis 8 8 (100)     7 7 (100) S. haemolyticus 4 3 (75) 1 (25)d   3 3 (100) S. capitis 1 1 (100)     1 1 (100) S. schleiferi 1 1 (100)     1 1 (100) S. simulans 1 1 (100)     1 1 (100)

0) S. lugdunensis 1 1 (100)     1 1 (100) Other CNS 24 22 (92) 2 (8)   20 19 (95) 1 (6) S. caprae 9 8 (89) 1 (11)   7 6 (67) 1 (33) S. hominis 8 8 (100)     7 7 (100) S. haemolyticus 4 3 (75) 1 (25)d   3 3 (100) S. capitis 1 1 (100)     1 1 (100) S. schleiferi 1 1 (100)     1 1 (100) S. simulans 1 1 (100)     1 1 (100) Streptococcus 9 7 (78) 1 (11) 1 (11) 6 5 (83)   1 (27) S. agalactiae 1 1 (100)     1 1 (100) S. anginosus group 1 1 (100) S. constellatus 1 1 (100)e Other streptococci 7 5 (72) 1 (14) 1 (14) 5 4 (80) S. mitis 2 1 (50)   1 (50)f 2 1 (50)   1 (50)f S. infantis 2 1 (50) 1 (50)g   1 1 (100) S. tigurinus 2 2 (100)     1 1 (100) S. oralis 1 1 (100)     1 1 (100) Enterococcus 17 16 (94) 1 (6)   16 16 (100) E. faecalis 2 1 (50) 1 (50)h   1 1 (100) Vancomycin-sensitive 2 1 (50) 1 (50)   1 1 (100) E. faecium 15 15 (100)     15 15 (100) Vancomycin-sensitive 15 15 (100)     15 15 (100) Total 104 96 (92) 6 (6) 2 (2) 93 91 (98) 1 (1) 1 (1) Other nontarget Gram-positives 15       10 Bacillus 3       2 B. subtilis 2       1 B. cereus 1       1 Corynebacterium 12       8 C. striatum 9       5 C. jeikeium 3       3 Total isolates 119       103 aPolymicrobial culture of methicillin-sensitive and methicillin-resistant S. epidermidis. bPolymicrobial culture with E. faecium. cCorrectly identified as Staphylococcus, but not as S. epidermidis, S. aureus, or S. lugdunensis. dPolymicrobial culture with S. tigurinus. e“S. anginosus group” identified by the BC-GP assay is defined as “correctly identified” for each species. fIdentified as S. pneumoniae.

aPolymicrobial culture of methicillin-sensitive and methicillin-resistant S. epidermidis. bPolymicrobial culture with E. faecium. cCorrectly identified as Staphylococcus, but not as S. epidermidis, S. aureus, or S. lugdunensis. dPolymicrobial culture with S. tigurinus. e“S. anginosus group” identified by the BC-GP assay is defined as “correctly identified” for each species. fIdentified as S. pneumoniae. gPolymicrobial culture with methicillin-resistant S. epidermidis. hPolymicrobial culture with Escherichia coli. Table 2 Performance of the BC-GP assay at Sakakibara Heart Institute. Organism BC-GP (total) BC-GP (monomicrobial cultures) Number of organisms Number (%) of isolates Number of organisms Number (%) of isolates Correctly identified Not detected Incorrectly identified Correctly identified Not detected Incorrectly identified Staphylococcus 24 24 (100)     22 22 (100) S. aureus 7 7 (100)     7 7 (100) Methicillin-sensitive 6 6 (100)     6 6 (100) Methicillin-resistant 1 1 (100)     1 1 (100) S. epidermidis 12 12 (100)     11 11 (100) Methicillin-sensitive 2 2 (100)     2 2 (100) Methicillin-resistant 10 10 (100)     9 9 (100) S. lugdunensis 1 1 (100)     1 1 (100) Other CNS 4 4 (100)     3 3 (100) S. hominis 2 2 (100)     2 2 (100) S. haemolyticus 1 1 (100) S. capitis 1 1 (100)     1 1 (100)

1 1 (100) S. epidermidis 12 12 (100)     11 11 (100) Methicillin-sensitive 2 2 (100)     2 2 (100) Methicillin-resistant 10 10 (100)     9 9 (100) S. lugdunensis 1 1 (100)     1 1 (100) Other CNS 4 4 (100)     3 3 (100) S. hominis 2 2 (100)     2 2 (100) S. haemolyticus 1 1 (100) S. capitis 1 1 (100)     1 1 (100) Streptococcus 8 8 (100)     8 7 (87)   1 (13)b S. pyogenes 1 0 (0)   1 (100)a 1 0 (0)   1 (100)b S. agalactiae 1 1 (100)     1 1 (100) S. anginosus group 2 2 (100)     2 2 (100) S. anginosus 2 2 (100)b     2 2 (100)c Other streptococci 4 4 (100)     4 4 (100) S. oralis 2 2 (100)     2 2 (100) S. sanguinis 1 1 (100)     1 1 (100) S. parasanguinis 1 1 (100)     1 1 (100) Enterococcus 2 1 (50) 1 (50) E. faecalis 1 0 (0) 1 (100)c Vancomycin-sensitive 1 0 (0) 1 (100)c E. faecium 1 1 (100) Vancomycin-sensitive 1 1 (100) Listeria spp. 1 1 (100)     1 1 (100) Total 35 33 (94) 1 (3) 1 (3) 31 30 (97)   1 (3) Other nontarget Gram-positives 1       1 Corynebacterium striatum 1       1 Total isolates 36       32 aIdentified to the genus level, but not to the species level. b“S. anginosus group” identified by the BC-GP assay is defined as “correctly identified” for each species. cPolymicrobial culture with S. epidermidis. Table 3 Difference in time to final identification and antimicrobial susceptibility report. Organism Sakakibara Heart Institute Toranomon Hospital Mean (h) Range Mean (h) Range S. aureus 34.5 (P < 0.05) 21.5–46.8 28.2 (P < 0.05) 19.6–47.0 S. epidermidis 80.7 (P < 0.05) 23.9–160.9 38.3 (P < 0.05) 21.6–72.5 Coagulase-negative staphylococci 121.1 (P < 0.05) 25.9–217.4 51.4 (P < 0.05) 21.4–72.2

Table 3 Difference in time to final identification and antimicrobial susceptibility report. Organism Sakakibara Heart Institute Toranomon Hospital Mean (h) Range Mean (h) Range S. aureus 34.5 (P < 0.05) 21.5–46.8 28.2 (P < 0.05) 19.6–47.0 S. epidermidis 80.7 (P < 0.05) 23.9–160.9 38.3 (P < 0.05) 21.6–72.5 Coagulase-negative staphylococci 121.1 (P < 0.05) 25.9–217.4 51.4 (P < 0.05) 21.4–72.2 Enterococcus spp. 156.6 (P < 0.05) 95.9–217.4 33.6 (P < 0.05) 23.4–47.7 Streptococcus spp. 196.6 (P < 0.05) 42.3–502.6 51.0 (P < 0.05) 23.4–69.2 aThe difference in time between BC-GP result and final culture-based identification and susceptibility results. Table 4 Detection of Gram-positive bacteria and resistance genes in simulated blood cultures by BC-GP. Organism Total no. of strains No. (%) of isolates correctly identified not detected incorrectly identified Staphylococcus 54 54 (100) S. aureus 35 35 (100) Methicillin-sensitive, mecA− 15 15 (100) Methicillin-resistant, mecA+ 20 20 (100) Health-care associated 10 10 (100) Community acquired 10 10 (100) S. epidermidis 1 1 (100) Methicillin-sensitive, mecA− 1 1 (100) S. lugdunensis 8 8 (100) Other CNS 10 10 (100) S. hominis 2 2 (100) S. haemolyticus 2 2 (100) S. saprophyticus 2 2 (100) S. capitis 1 1 (100) S. cohnii 1 1 (100) S. warneri 1 1 (100) S. pseudintermedius 1 1 (100)

Staphylococcus 54 54 (100) S. aureus 35 35 (100) Methicillin-sensitive, mecA− 15 15 (100) Methicillin-resistant, mecA+ 20 20 (100) Health-care associated 10 10 (100) Community acquired 10 10 (100) S. epidermidis 1 1 (100) Methicillin-sensitive, mecA− 1 1 (100) S. lugdunensis 8 8 (100) Other CNS 10 10 (100) S. hominis 2 2 (100) S. haemolyticus 2 2 (100) S. saprophyticus 2 2 (100) S. capitis 1 1 (100) S. cohnii 1 1 (100) S. warneri 1 1 (100) S. pseudintermedius 1 1 (100) Streptococcus 87 77 (88) 4 (5) 6 (7) S. pyogenes 9 8 (89) 1 (11)a S. agalactiae 8 8 (100) S. dysgalactiae 8 8 (100) S. anginosus groupb 10 8 (80)   2 (20) S. anginosus 3 2 (66)   1 (34)c S. constellatus 4 3 (75)   1 (25)c S. intermedius 3 3 (100) S. pneumoniae 22 20 (91)   2 (9)d Other streptococci 30 25 (86) 3 (10) 2 (7) S. mitis 12 9 (75) 1 (8) 2 (17)e S. oralis 4 4 (100) S. infantis 2 2 (100) S. mutans 2 1 (50) 1 (50) S. sobrinus 1 0 (0) 1 (100) S. sanguinis 1 1 (100)   ∖ S. parasanguinis 1 1 (100) S. peroris 1 1 (100) S. australis 1 1 (100) S. tigurinus 1 1 (100) S. cristatus 1 1 (100) S. gordonii 1 1 (100) S. gallolyticus 1 1 (100) S. lutetiensis 1 1 (100) Enterococcus 57 57 (100) E. faecalis 32 32 (100) Vancomycin-sensitive 18 18 (100) Vancomycin-resistant, vanA+ 4 4 (100) Vancomycin-resistant, vanB+ 10 10 (100) E. faecium 25 25 (100) Vancomycin-sensitive 5 5 (100) Vancomycin-resistant, vanA+ 10 10 (100) Vancomycin-resistant, vanB+ 10 10 (100) Listeria spp. 5 5 (100) Micrococcus spp. 5 5 (100) Total 208 198 (95) 4 (2) 6 (3) aNot detected initially, but positive using 11-fold diluted blood culture sample.

Enterococcus 57 57 (100) E. faecalis 32 32 (100) Vancomycin-sensitive 18 18 (100) Vancomycin-resistant, vanA+ 4 4 (100) Vancomycin-resistant, vanB+ 10 10 (100) E. faecium 25 25 (100) Vancomycin-sensitive 5 5 (100) Vancomycin-resistant, vanA+ 10 10 (100) Vancomycin-resistant, vanB+ 10 10 (100) Listeria spp. 5 5 (100) Micrococcus spp. 5 5 (100) Total 208 198 (95) 4 (2) 6 (3) aNot detected initially, but positive using 11-fold diluted blood culture sample. b Streptococcus sp. belonging to the S. anginosus group is identified as S. anginosus group by BC-GP. cPositive signal for Streptococcus, but no signal for S. anginosus group. dPositive signal for Streptococcus, but no signal for S. pneumoniae. eMisidentified as S. pneumoniae.

1. Introduction Urinary tract infections (UTIs) are common and, as such, are associated with substantial socioeconomic implications [1]. At least one-half of women report experiencing one or more UTIs in their lifetime [2, 3]. Most UTIs are uncomplicated bacterial infections of the lower urinary tract (uUTIs) [2, 4]. Escherichia coli is the uropathogen responsible for 75% or more of uUTIs [5–8]. Many patients with uUTIs experience considerable impact on their health-related quality of life (HRQoL), including impairments in functioning at work, school, and home [9, 10]. Treatment of uUTIs is necessary not only to improve patient HRQoL, but also to reduce the risk of progression to pyelonephritis, a more serious condition [5]. Several antibiotics are used in Canada for treatment of uUTIs, including trimethoprim, trimethoprim–sulfamethoxazole (TMP–SMX), fluoroquinolones such as ciprofloxacin, nitrofurantoin, β-lactam antibiotics such as amoxicillin–clavulanate (AMC), and fosfomycin, a phosphoric acid derivative that represents its own class of antibiotics [11–14].

e used in Canada for treatment of uUTIs, including trimethoprim, trimethoprim–sulfamethoxazole (TMP–SMX), fluoroquinolones such as ciprofloxacin, nitrofurantoin, β-lactam antibiotics such as amoxicillin–clavulanate (AMC), and fosfomycin, a phosphoric acid derivative that represents its own class of antibiotics [11–14]. In addition to an already high burden of illness, standard treatment options are being challenged by rising bacterial resistance rates. Common uropathogens, in particular display variable resistance patterns over time and by region [5]. This requires practitioners to be aware of current local resistance rates and to be vigilant in tracking regional changes before empirically prescribing antibiotics for the treatment of UTIs [5]. For decades, TMP–SMX was considered first-line therapy for uUTIs [3], but rising rates of TMP–SMX resistance among uropathogens, and consistent evidence that in vitro TMP–SMX resistance correlates with bacterial and clinical failure, have forced a reappraisal of its first-line use for uUTIs [5]. Today, the E. coli resistance rate to TMP–SMX in Canada exceeds the 20% threshold at which guidelines recommend against TMP–SMX use for uUTI treatment [5, 11, 15–17]. E. coli resistance to the commonly used fluoroquinolone ciprofloxacin was 19% in Canada in 2007–2009 [16], approximately twice the threshold above which the Infectious Diseases Society of America/European Society for Microbiology and Infectious Diseases (IDSA/ESCMID) clinical practice guidelines no longer recommend use of fluoroquinolones for uUTIs [5]. Bacterial resistance has several detrimental consequences for patients and society, including treatment failure and risk of complications, increased length of stay in hospital facilities, increased expenses, and increased mortality [18].

cal practice guidelines no longer recommend use of fluoroquinolones for uUTIs [5]. Bacterial resistance has several detrimental consequences for patients and society, including treatment failure and risk of complications, increased length of stay in hospital facilities, increased expenses, and increased mortality [18]. As bacterial resistance rates to current first-line antibiotics continue to rise, there will be a need for options that are effective, safe, and not likely to further promote community resistance. Although there is limited experience with fosfomycin for treating uUTIs in Canada [3, 19], it has been used for decades in other countries [20]. A meta-analysis of seven randomized controlled trials involving 1272 female patients with uUTIs showed no difference in cure rate between fosfomycin and a variety of comparators (including TMP–SMX, ciprofloxacin, and nitrofurantoin) [1]. Furthermore, the prevalence of E. coli resistance to fosfomycin is low (with an average resistance rate of 1.2% in a recent international survey) and has not shown a significant increase over time [21, 22]. To date, there is no published economic analysis that provides evidence on the relative cost-effectiveness of fosfomycin compared to recommended empirical therapy for uUTIs in Canada. The objective of this study was to examine the economic impact of using fosfomycin for first-line empirical treatment of uUTIs in adult women in Ontario, taking into account current antibiotic prices, treatment algorithms, event probabilities, and resistance rates.

to recommended empirical therapy for uUTIs in Canada. The objective of this study was to examine the economic impact of using fosfomycin for first-line empirical treatment of uUTIs in adult women in Ontario, taking into account current antibiotic prices, treatment algorithms, event probabilities, and resistance rates. 2. Methods 2.1. Decision-Tree Model Based on evidence of equivalent efficacy and safety in head-to-head trials [38–43] and a meta-analysis [1], we conducted a cost-minimization analysis (CMA) using a decision-tree model (TreeAge Pro 2015 software®, Williamstown, MA, USA). The model is based on a previous model of uUTI treatment caused by E. coli in women older than 18 years [27] and uses a payer perspective, specifically that of the Ontario Ministry of Health and Long-term Care. Only direct medical care costs are considered. The time horizon of the model included 2 treatment courses. The model assumed that first-line empirical treatment of uUTIs would consist of one of the antibiotics currently reimbursed by the Ontario Drug Benefit Program (ODBP) [44] and recommended by recent guidelines [5, 13, 14] for the treatment uUTIs: sulfonamides (TMP/SMX DS 800 mg/160 mg, TMP 100 mg, or TMP 200 mg), fluoroquinolones (Cipro® XL™ 500 mg, ciprofloxacin 250 mg [generic], or norfloxacin 400 mg), or nitrofurantoins (MacroBID® 100 mg, nitrofurantoin 50 mg [generic], or nitrofurantoin 100 mg [generic]).

cent guidelines [5, 13, 14] for the treatment uUTIs: sulfonamides (TMP/SMX DS 800 mg/160 mg, TMP 100 mg, or TMP 200 mg), fluoroquinolones (Cipro® XL™ 500 mg, ciprofloxacin 250 mg [generic], or norfloxacin 400 mg), or nitrofurantoins (MacroBID® 100 mg, nitrofurantoin 50 mg [generic], or nitrofurantoin 100 mg [generic]). The decision-tree model incorporates all possible outcomes of a uUTI caused by E. coli, Enterococcus species, Klebsiella pneumoniae, Proteus mirabilis, or all other bacterial species (combined) treated with either fluoroquinolones, sulfonamides, nitrofurantoin, or fosfomycin (Figure 1). For simplicity, the model does not include the probability of rare complications associated with specific antibiotics (e.g., Stevens-Johnson syndrome). The probability of vaginal yeast infections is also not included because this complication is generally treated with over-the-counter drugs that are not reimbursed by the ODBP. The model compares all four treatments and considers the possibility that the bacterial species would be susceptible or resistant to each antibiotic. In the latter case, the UTI either spontaneously resolves or persists. Patients with persistent UTIs either progress to pyelonephritis (with outpatient or inpatient treatment) or, in the absence of pyelonephritis, are treated with a different antibiotic. The model was internally validated. All codes and programming were validated by a third party who was not involved in creating the model.

The model compares all four treatments and considers the possibility that the bacterial species would be susceptible or resistant to each antibiotic. In the latter case, the UTI either spontaneously resolves or persists. Patients with persistent UTIs either progress to pyelonephritis (with outpatient or inpatient treatment) or, in the absence of pyelonephritis, are treated with a different antibiotic. The model was internally validated. All codes and programming were validated by a third party who was not involved in creating the model. 2.2. Costs Province-level data on drug costs were used in the model. For the base case scenario, the cost of treatment for each antibiotic used to treat uUTIs was computed based on prices in the Ontario Drug Benefit Formulary (ODBF) [44] and the recommended dosage for uUTIs, as outlined by recent clinical guidelines [5, 13, 14]. The cost of each comparator was obtained by calculating an average cost of treatment for the class of antibiotic weighted by the volume of claims reimbursed by the ODBP as reported by Pharmastat data (Table 1).

lary (ODBF) [44] and the recommended dosage for uUTIs, as outlined by recent clinical guidelines [5, 13, 14]. The cost of each comparator was obtained by calculating an average cost of treatment for the class of antibiotic weighted by the volume of claims reimbursed by the ODBP as reported by Pharmastat data (Table 1). Data on healthcare resource use were obtained from multiple public sources including the Physician Services under the Health Insurance Act (April 1, 2016), Ontario Case Costing Initiative, and from Pharmastat provincial data [24, 45]. Healthcare resource use varied by outcome. For uUTIs that resolved with first-line empirical treatment, a single clinic visit was included in the model. Urinalysis and urine cultures were not performed prior to first-line treatment. For uUTIs that persisted and required treatment with another antibiotic, the model included two clinic visits (one initial visit and one following failure of the first-line antibiotic) and a urine culture. Progression to pyelonephritis was assumed to occur during or after first-line treatment; thus, patients with pyelonephritis requiring outpatient treatment consumed the following resources: one initial clinic visit, one emergency room visit, two follow-up clinic visits (one 48 hours after initiation of treatment for pyelonephritis and one at the end of treatment), and one urine culture at the end of treatment to confirm resolution of infection. For patients with pyelonephritis requiring inpatient treatment, the model included an initial clinic visit, hospitalization for pyelonephritis, and a follow-up clinic visit with a urine culture at the end of treatment to confirm resolution of infection.

ulture at the end of treatment to confirm resolution of infection. For patients with pyelonephritis requiring inpatient treatment, the model included an initial clinic visit, hospitalization for pyelonephritis, and a follow-up clinic visit with a urine culture at the end of treatment to confirm resolution of infection. First-line empirical treatment could be any one of the following: sulfonamide, fluoroquinolone, nitrofurantoin, or fosfomycin. Second-line treatment for persistent infections was either Cipro® XL™ 500 mg once per day for 3 days (if first treatment was not ciprofloxacin) or MacroBID® 100 mg twice per day for 7 days (if first treatment was ciprofloxacin). Pyelonephritis was treated with either Cipro® XL™ 500 mg/1 g once per day for 7 days (if first treatment was not ciprofloxacin) or Clavulin® 500/125 g three times per day for 14 days (if first treatment was ciprofloxacin). Costs associated with productivity loss are not included in the analysis; only direct medical care costs are considered.

First-line empirical treatment could be any one of the following: sulfonamide, fluoroquinolone, nitrofurantoin, or fosfomycin. Second-line treatment for persistent infections was either Cipro® XL™ 500 mg once per day for 3 days (if first treatment was not ciprofloxacin) or MacroBID® 100 mg twice per day for 7 days (if first treatment was ciprofloxacin). Pyelonephritis was treated with either Cipro® XL™ 500 mg/1 g once per day for 7 days (if first treatment was not ciprofloxacin) or Clavulin® 500/125 g three times per day for 14 days (if first treatment was ciprofloxacin). Costs associated with productivity loss are not included in the analysis; only direct medical care costs are considered. 2.3. Probabilities Probabilities were obtained from published estimates. Probabilities of resistance for fluoroquinolones, sulfonamides, and nitrofurantoin were obtained from the CANWARD surveillance study 2007–2011 [15] and the CANWARD surveillance study 2007–2009 (Table 2) [16]. Resistance rates for fosfomycin were obtained from the ARESC (Antimicrobial Resistance Epidemiological Survey on Cystitis) study [22]. Probabilities for effectiveness were assumed to be equal for all antibiotics based on the meta-analysis published by Falagas et al. (2010) [1]. Effectiveness was assumed to be reduced by 20% for resistant isolates [33]. Probabilities of pyelonephritis (i.e., pyelonephritis due to initial treatment failure and hospitalization for pyelonephritis) were obtained from different sources [27, 28, 34–36].

ed on the meta-analysis published by Falagas et al. (2010) [1]. Effectiveness was assumed to be reduced by 20% for resistant isolates [33]. Probabilities of pyelonephritis (i.e., pyelonephritis due to initial treatment failure and hospitalization for pyelonephritis) were obtained from different sources [27, 28, 34–36]. 2.4. Sensitivity Analyses Variables suspected to have the most influence on the expected value of cost-per-patient were subjected to a deterministic one-way sensitivity analysis (DSA). Where possible, ranges for variables were extracted from the literature or other public sources. In the absence of a reliable source, a hypothesis of the plausible range was used. The base case values and the ranges tested in the sensitivity analyses are shown in Tables 1 and 2. 3. Interpretation This study uses a cost-minimization analysis (CMA) to identify the most affordable options for the treatment of uUTI in Ontario, taking into account both cost-per-dose and costs associated with antibiotic resistance. 3.1. Analysis and Results With the current cost of fosfomycin of $17.51, the cost-per-patient for treating uUTIs with this drug was determined to be $105.12. This is comparable to the cost-per-patient for each of the other currently reimbursed antibiotics (Figure 2).

3. Interpretation This study uses a cost-minimization analysis (CMA) to identify the most affordable options for the treatment of uUTI in Ontario, taking into account both cost-per-dose and costs associated with antibiotic resistance. 3.1. Analysis and Results With the current cost of fosfomycin of $17.51, the cost-per-patient for treating uUTIs with this drug was determined to be $105.12. This is comparable to the cost-per-patient for each of the other currently reimbursed antibiotics (Figure 2). The weighted average cost-per-patient for treating uUTI in Ontario was also estimated based on 2 scenarios: with and without fosfomycin. In this analysis, cost-per-patient was obtained by replacing the decision node of the tree with a chance node which identifies probabilities of events. The probabilities were the market share of each alternative treatment. Thus, the weighted average cost takes into account the antibiotics currently reimbursed by the ODB and their respective market shares. The weighted average cost-per-patient for treating uUTI in Ontario is $98.29 (Figure 3). Market shares are based on Ontario Pharmastat data. Incorporating fosfomycin into the mix of treatments reimbursed in Ontario results in a weighted average cost-per-patient of $98.41 (Figure 4). For this analysis, the estimated market share of fosfomycin was 1.77% as reported by Pharmastat. The impact of reimbursing fosfomycin in the Ontario treatment landscape has a negligible impact on the weighted cost-per-patient and provides physicians with an effective alternate treatment option.

nt of $98.41 (Figure 4). For this analysis, the estimated market share of fosfomycin was 1.77% as reported by Pharmastat. The impact of reimbursing fosfomycin in the Ontario treatment landscape has a negligible impact on the weighted cost-per-patient and provides physicians with an effective alternate treatment option. 3.2. Results of Uncertainty Analysis Sensitivity analyses were performed based on the decision and chance nodes of the decision tree. The tornado diagram revealed that the probability of developing pyelonephritis has the largest potential effect on the expected value (cost-per-patient) of the model (Figure 5). Other variables that had a potential effect on cost-per-patient were as follows:Probability of hospitalization for pyelonephritis Probability of resistance of E. coli to TMP–SMX Cost of hospitalization for pyelonephritis Cost of nitrofurantoin Probability of being infected by E. coli Variables accounting for 93.18% of the potential variation in cost-per-patient were the probability of progressing to pyelonephritis (55.13%), having to be hospitalized for pyelonephritis (33.17%), and contracting an infection with TMP–SMX-resistant E. coli (4.87%). Detailed results of the one-way sensitivity analyses are presented in Appendix.

3.18% of the potential variation in cost-per-patient were the probability of progressing to pyelonephritis (55.13%), having to be hospitalized for pyelonephritis (33.17%), and contracting an infection with TMP–SMX-resistant E. coli (4.87%). Detailed results of the one-way sensitivity analyses are presented in Appendix. Figure 6 compares the cost-per-patient for treating uUTI with a fluoroquinolone, sulfonamide, nitrofurantoin, or fosfomycin with various probabilities of progression to pyelonephritis. At all levels of risk of progression to pyelonephritis, fosfomycin is associated with the highest cost-per-patient. As the risk of pyelonephritis increases, the expected values of other antimicrobial treatments approach that of fosfomycin. Note that the base case value for progression to pyelonephritis is 4% [27, 28, 34, 35]. Figure 7 compares the cost-per-patient for treating uUTI with one of the 4 antimicrobial treatments with the probability of hospitalization due to pyelonephritis. At all probabilities, fosfomycin is associated with the highest cost. Note that the base case value for the probability of hospitalization due to pyelonephritis is 20% [36].

s the cost-per-patient for treating uUTI with one of the 4 antimicrobial treatments with the probability of hospitalization due to pyelonephritis. At all probabilities, fosfomycin is associated with the highest cost. Note that the base case value for the probability of hospitalization due to pyelonephritis is 20% [36]. Figure 8 compares the expected cost-per-patient for the treatment of a uUTI with a fluoroquinolone, sulfonamide, nitrofurantoin, or fosfomycin with the prevalence of TMP–SMX-resistant E. coli. The cost-per-patient when using a sulfonamide steadily increases with increasing resistance probabilities. The cost-per-patient of the other antibiotics is not affected by this variable. Note that the base case value for E. coli resistance to sulfonamides is 26.7% [15]. A tornado diagram was also developed using the one-way sensitivity analyses of the chance node instead of the decision node of the decision tree (Figure 9). The major difference in the results of this sensitivity analysis was that the cost of nitrofurantoins was the third highest contributor to the potential variation in cost-per-patient—meaning that the cost of nitrofurantoins has a great influence on the cost of treating patients with uUTI covered by the ODB. Recall that, in the decision node analysis, the third highest contributor was probability of contracting an infection with TMP–SMX-resistant E. coli.

the potential variation in cost-per-patient—meaning that the cost of nitrofurantoins has a great influence on the cost of treating patients with uUTI covered by the ODB. Recall that, in the decision node analysis, the third highest contributor was probability of contracting an infection with TMP–SMX-resistant E. coli. 4. Discussion Although there is limited experience with fosfomycin for treating uUTI in Canada [3], it has been used successfully for decades in other countries [19, 20]. Fosfomycin is a phosphoric acid derivative that represents its own class of antibiotics; no other member of this class is currently approved by regulatory agencies worldwide [19, 20].

mited experience with fosfomycin for treating uUTI in Canada [3], it has been used successfully for decades in other countries [19, 20]. Fosfomycin is a phosphoric acid derivative that represents its own class of antibiotics; no other member of this class is currently approved by regulatory agencies worldwide [19, 20]. Resistance rates for ciprofloxacin and TMP–SMX (INESSS first-line recommendations for treatment of uUTI) were much lower a decade ago than they are now. The ECO-SENS survey, conducted in 1999-2000, showed E. coli resistance rates of 0% for ciprofloxacin and 12% for TMP–SMX in Canada (n = 166, urinary isolates) [6]. As of 2009, the CANWARD surveillance initiative revealed that E. coli resistance to TMP–SMX in urinary isolates was 24.3% and had undergone a significant increase over just a 3-year period—the resistance rate was 18.6% in 2007 (data derived from urine samples of 2943 outpatients and inpatients, P = 0.02) [16]. Recent data from the CANWARD surveillance initiative (2011) focusing on resistance rates regardless of source (i.e., not specific to uropathogens) revealed that E. coli resistance to ciprofloxacin and TMP–SMX is quite high in Canada (26.9% for ciprofloxacin and 29.3% for TMP–SMX, N = 646). Ontario resistance rates in 2011 were similar to those seen across Canada (24.6% for ciprofloxacin and 28.1% for TMP–SMX, N = 167) [46]. It is clear that antibiotic resistance is on the rise, particularly for ciprofloxacin and TMP–SMX.

MX is quite high in Canada (26.9% for ciprofloxacin and 29.3% for TMP–SMX, N = 646). Ontario resistance rates in 2011 were similar to those seen across Canada (24.6% for ciprofloxacin and 28.1% for TMP–SMX, N = 167) [46]. It is clear that antibiotic resistance is on the rise, particularly for ciprofloxacin and TMP–SMX. In contrast, the ECO-SENS survey (conducted in 1999-2000 and then again in 2007-2008) showed that the initial low resistance rate of E. coli to fosfomycin did not increase significantly over time. In E. coli isolates from 5 European countries, the resistance rate to fosfomycin was 0.4% in 1999-2000 and 1.2% in 2007-2008 [21]. This consistently low resistance rate may be due to the fact that fosfomycin resistance comes at a biological cost to microorganisms [19]. Furthermore, the low resistance of E. coli in the ECO-SENS survey could be explained by the exclusion of hospitalized patients during the two weeks before the onset of symptoms and data from the CANWARD surveillance initiative were based on samples from outpatients and inpatients. Microbiological studies have shown that fosfomycin-resistant strains of E. coli are “biologically impaired”—they take up to 39% longer to replicate and they have a 38% reduction in the ability to adhere to uroepithelial cells. These resistant strains are unable to compete with susceptible strains in the environment, possibly contributing to the low rates of fosfomycin resistance observed over time [19, 47].

gically impaired”—they take up to 39% longer to replicate and they have a 38% reduction in the ability to adhere to uroepithelial cells. These resistant strains are unable to compete with susceptible strains in the environment, possibly contributing to the low rates of fosfomycin resistance observed over time [19, 47]. Fosfomycin also has a low propensity for cross-resistance, which is likely due to the fact that it differs from other antibiotics in its general chemical structure and site of action [19, 48]. The unique mechanism of action of fosfomycin appears to underlie its low rate of cross-resistance. This mechanism of action, which inhibits the early stages of cell wall synthesis [49], reduces the risk of cross-resistance with other antibacterial agents, making fosfomycin a good choice for the treatment of multidrug-resistant strains and leaving options available for further therapies [19, 47]. In the ECO-SENS survey, cross-resistance in E. coli was observed with ampicillin and SMX, alone or linked with resistance to TMP and TMP–SMX. A multiple-resistant phenotype involving fluoroquinolone resistance was also noted. The ECO-SENS authors concluded that resistance to any agent was associated with increased resistance to all other agents tested, fosfomycin being the only exception [50].

llin and SMX, alone or linked with resistance to TMP and TMP–SMX. A multiple-resistant phenotype involving fluoroquinolone resistance was also noted. The ECO-SENS authors concluded that resistance to any agent was associated with increased resistance to all other agents tested, fosfomycin being the only exception [50]. In addition to the biological aspect of the low fosfomycin resistance rate, single-dose antibiotics are associated with a high degree of compliance and greater convenience, which may help prevent resistance from developing [19, 51, 52]. Maintenance of 100% adherence to optimized dosing regimens is critical for ensuring antimicrobial efficacy of the drug and prevention of resistance. Experts have recommended that underdosing be prevented to “suppress or decrease the potential amplification of resistant mutant subpopulations” and that regimens should be as short as possible in order to maximally suppress resistance [52]. Other potential benefits of single-dose therapy include directly observed therapy (i.e., compliance is not an issue) and fewer side effects [53].

s or decrease the potential amplification of resistant mutant subpopulations” and that regimens should be as short as possible in order to maximally suppress resistance [52]. Other potential benefits of single-dose therapy include directly observed therapy (i.e., compliance is not an issue) and fewer side effects [53]. Since resistance patterns of E. coli strains causing uUTI vary considerably across regions and can be well beyond the resistance threshold for certain antibiotics, the IDSA/ESCMID guidelines note that a specific treatment recommendation may not be universally suitable [5, 46]. The guidelines now include fosfomycin as one of the recommended first-line options for the treatment of uUTIs along with nitrofurantoin, TMP–SMX, and pivmecillinam [5]. In May 12, 2016, the FDA made recommendations resulting from a November 5, 2015, meeting, (Joint Meeting of the Antimicrobial Drugs Advisory Committee [formerly known as the Anti-Infective Drugs Advisory Committee] and Drug Safety and Risk Management Advisory Committee Meeting) that, based on risk versus benefit, fluoroquinolones (such as ciprofloxacin and levofloxacin) should not be used first line for the treatment of uUTI; rather nitrofurantoin, TMP–SMX, and fosfomycin should be used instead [54]. In this new world of rapidly advancing microbial resistance, an additional antibiotic in the armamentarium, especially one with a low resistance profile, may be of substantial benefit to Ontario practitioners and patients.

veolar lavage fluid (BALF), endotracheal aspiration (ETA), and exhaled breath condensate (EBC), where it can be assayed by ELISA using commercial immunoassay kits [5]. Some clinical studies have proved that sTREM-1 did have the ability to identify patients with sepsis while others come to an opposite conclusion [6, 7]. However, the real effect of sTREM-1 on diagnosis of VAP is still unknown and has not been well evaluated yet. Furthermore, the majority of the studies about sTREM-1 levels in VAP were heterogeneous due to including different diseases. What is the most important is that some of these studies did not exclude the patients with extrapulmonary infection which might lead to misinterpretation of the study results. The subjects of our study were patients with ischemic stroke, who had a high incidence of VAP and were without any infectious complications at the time of invasive mechanical ventilation. Trauma, malignant neoplasms, and acute inflammatory response could all lead to the raise of sTREM-1 in serum due to its high sensitivity and low specificity which resulted in the poor performance in VAP diagnosis. BALF was partly secreted by lung which was infected as the target organ. So we measured sTREM-1 levels in serum, BALF, ETA, and EBC samples from patients who underwent bronchoscopy for a clinical suspicion of VAP and we hypothesized that sTREM-1 in different body fluid might have different values in diagnosing VAP.

Since resistance patterns of E. coli strains causing uUTI vary considerably across regions and can be well beyond the resistance threshold for certain antibiotics, the IDSA/ESCMID guidelines note that a specific treatment recommendation may not be universally suitable [5, 46]. The guidelines now include fosfomycin as one of the recommended first-line options for the treatment of uUTIs along with nitrofurantoin, TMP–SMX, and pivmecillinam [5]. In May 12, 2016, the FDA made recommendations resulting from a November 5, 2015, meeting, (Joint Meeting of the Antimicrobial Drugs Advisory Committee [formerly known as the Anti-Infective Drugs Advisory Committee] and Drug Safety and Risk Management Advisory Committee Meeting) that, based on risk versus benefit, fluoroquinolones (such as ciprofloxacin and levofloxacin) should not be used first line for the treatment of uUTI; rather nitrofurantoin, TMP–SMX, and fosfomycin should be used instead [54]. In this new world of rapidly advancing microbial resistance, an additional antibiotic in the armamentarium, especially one with a low resistance profile, may be of substantial benefit to Ontario practitioners and patients. 4.1. Summary of Results This analysis revealed that the cost-per-patient for treating uUTI in Ontario is similar for all 4 antibiotics, ranging from $96.19 for sulfonamides to $105.12 for fosfomycin. The average cost-per-patient for treating uUTI in Ontario with all 4 reimbursed antimicrobial agents (fosfomycin, fluoroquinolones, sulfonamides, or nitrofurantoin), weighted by market share, is $98.41 (versus $98.29 without fosfomycin).

all 4 antibiotics, ranging from $96.19 for sulfonamides to $105.12 for fosfomycin. The average cost-per-patient for treating uUTI in Ontario with all 4 reimbursed antimicrobial agents (fosfomycin, fluoroquinolones, sulfonamides, or nitrofurantoin), weighted by market share, is $98.41 (versus $98.29 without fosfomycin). Most of the potential variability in cost-per-patient (93.18%) is associated with the probability of progressing to pyelonephritis (55.13%), having to be hospitalized for pyelonephritis (33.17%), and contracting an infection with TMP–SMX-resistant E. coli (4.87%). 4.2. Study Limitations There were limitations to this study. First, efficacy was assumed to be equal based on the meta-analysis conducted by Falagas et al. (2010) [1]. However, this analysis was only focused on comparing fosfomycin with other antibiotics. Meta-analyses providing efficacy comparisons between the other antimicrobial agents were not identified in the literature. Similarly, we were unable to locate sources for the efficacy of each agent in susceptible versus resistant infections. Secondly, the only published data we had for the probability of developing pyelonephritis was from a US source. As the model was the most sensitive to this variable, we cannot definitively determine its reliability in Canada. 4.3. Generalizability This model has been constructed to specifically address the empirical treatment of uUTI in women older than 18 years of age. This population is aligned with the current approved indication of fosfomycin in Canada.

Secondly, the only published data we had for the probability of developing pyelonephritis was from a US source. As the model was the most sensitive to this variable, we cannot definitively determine its reliability in Canada. 4.3. Generalizability This model has been constructed to specifically address the empirical treatment of uUTI in women older than 18 years of age. This population is aligned with the current approved indication of fosfomycin in Canada. The data used to populate the model were derived from Ontario and/or Canadian sources wherever possible. Costs and market share of currently reimbursed antibiotics were derived from provincial data, providing an accurate view of the present market. Consultation with experts on clinical practice patterns and costs provided an additional layer of real-world validity.

The data used to populate the model were derived from Ontario and/or Canadian sources wherever possible. Costs and market share of currently reimbursed antibiotics were derived from provincial data, providing an accurate view of the present market. Consultation with experts on clinical practice patterns and costs provided an additional layer of real-world validity. 4.4. Health Services Impact Since the availability of fosfomycin does not significantly increase the weighted average cost of treatment per patient, reimbursement of this antibiotic in Ontario has a negligible effect on the provincial healthcare budget. The societal implications are, however, substantial. Providing access to an additional antibiotic, especially one with a low resistance profile like fosfomycin, may benefit both Ontario practitioners and their patients. With rising rates of resistance to fluoroquinolones and sulfonamides, these “staple” antibiotics are no longer broadly considered appropriate first-line empirical therapies by international experts [5]. This shift of recommended treatments in the armamentarium leaves physicians in Ontario with fewer options for treating uUTIs. Reimbursement of fosfomycin by the ODBP would allow physicians to effectively treat patients with uUTIs while ensuring that community resistance to important antibiotics (fluoroquinolones and sulfonamides) is not further increased in Ontario.

is. BALF was partly secreted by lung which was infected as the target organ. So we measured sTREM-1 levels in serum, BALF, ETA, and EBC samples from patients who underwent bronchoscopy for a clinical suspicion of VAP and we hypothesized that sTREM-1 in different body fluid might have different values in diagnosing VAP. 2. Methods 2.1. Research Briefs The present study was a multicenter (155 ICU beds in total) prospective observational trial which included samples of serum, BALF, ETA, and EBC from eligible patients. The study was approved by the Review Board and Ethics Committee of Shanghai Jiaotong University School of Medicine (No: 2013-Clinical-Res-085) and informed consent was obtained for all patients, from either the patient or the next of kin. The study was performed from January 2013 to December 2015. One month before the start of this study, a standardized sampling, processing, analysis, and statistics procedure was set by 9 investigators from these centers after 3 days' learning and discussion. All of the centers carefully followed this procedure during the study time. There was no significant difference in compliance among the centers.

mentarium leaves physicians in Ontario with fewer options for treating uUTIs. Reimbursement of fosfomycin by the ODBP would allow physicians to effectively treat patients with uUTIs while ensuring that community resistance to important antibiotics (fluoroquinolones and sulfonamides) is not further increased in Ontario. 4.5. Future Research Further research into the efficacy of fluoroquinolones, sulfonamides, and nitrofurantoin in resistant infections would be beneficial for future pharmacoeconomic evaluations, especially since the rates of resistance for some of these agents have risen drastically in a short period of time. Additionally, head-to-head studies of antibiotic efficacy and safety in current uUTI patient populations would provide much needed data for developing accurate probabilities of treatment success/failure. 5. Conclusions Fosfomycin in addition to being a safe and effective agent to treat uUTI has a low resistance profile, offers a single-dose treatment regimen, and is comparable in cost to other reimbursed antibiotics. Acknowledgments Medical writing and editorial support were provided by Sabrina Smith, Church Point, NS, Canada, funded by International Market Access Consulting and Triton Pharma (acquired by Paladin Labs Inc. in January 2014). Appendix See Table 3. Competing Interests Louise Perrault, Sybil Dahan, George G. Zhanel, and Jacques LeLorier have served as consultants to Triton Pharma and received funding from Triton Pharma for their contributions to this analysis.

Acknowledgments Medical writing and editorial support were provided by Sabrina Smith, Church Point, NS, Canada, funded by International Market Access Consulting and Triton Pharma (acquired by Paladin Labs Inc. in January 2014). Appendix See Table 3. Competing Interests Louise Perrault, Sybil Dahan, George G. Zhanel, and Jacques LeLorier have served as consultants to Triton Pharma and received funding from Triton Pharma for their contributions to this analysis. Authors' Contributions Louise Perrault and Sybil Dahan conceived of the economic model, developed it, and produced the analyses. George G. Zhanel and Jacques LeLorier provided microbiology and clinical expertise. All authors contributed to interpretation of the data, helped to draft the report, and read and approved the final version. Figure 1 Decision-tree diagram showing intermediate and final outcomes for an uncomplicated E. coli UTI treated with a sulfonamide. Figure 2 Cost-per-patient for treating uUTI in Ontario. Figure 3 Weighted average cost-per-patient for treating uUTI in Ontario without fosfomycin. Figure 4 Weighted average cost-per-patient for treating uUTI in Ontario with fosfomycin. Figure 5 Tornado diagram comparing the results of the one-way sensitivity analyses of variables potentially affecting the expected value of the cost-per-patient (decision node). Figure 6 Comparison of the expected value of the cost-per-patient when treating uUTI with fluoroquinolones, sulfonamides, nitrofurantoin, and fosfomycin versus the probability of developing pyelonephritis.

Figure 5 Tornado diagram comparing the results of the one-way sensitivity analyses of variables potentially affecting the expected value of the cost-per-patient (decision node). Figure 6 Comparison of the expected value of the cost-per-patient when treating uUTI with fluoroquinolones, sulfonamides, nitrofurantoin, and fosfomycin versus the probability of developing pyelonephritis. Figure 7 Comparison of the expected value of the cost-per-patient when treating uUTI with fluoroquinolones, sulfonamides, nitrofurantoin, and fosfomycin versus the probability of hospitalization for pyelonephritis. Figure 8 Comparison of the expected value of the cost-per-patient when treating uUTI with fluoroquinolones, sulfonamides, nitrofurantoin, and fosfomycin versus the probability of a uUTI with TMP–SMX-resistant E. coli. Figure 9 Tornado diagram comparing the results of the one-way sensitivity analyses of variables potentially affecting the expected value of cost-per-patient (chance node). Table 1 Cost of antibiotics and healthcare resources used for the treatment of uUTIs and the ranges tested in the sensitivity analyses.

Figure 8 Comparison of the expected value of the cost-per-patient when treating uUTI with fluoroquinolones, sulfonamides, nitrofurantoin, and fosfomycin versus the probability of a uUTI with TMP–SMX-resistant E. coli. Figure 9 Tornado diagram comparing the results of the one-way sensitivity analyses of variables potentially affecting the expected value of cost-per-patient (chance node). Table 1 Cost of antibiotics and healthcare resources used for the treatment of uUTIs and the ranges tested in the sensitivity analyses. Variable Base case (CAD $) Range tested (CAD $) References Fosfomycin 17.51 13.00–18.00 [23], Pharmastat Sulfonamides 0.95 0.88–2.79 [23], Pharmastat Fluoroquinolones 4.25 3.71–11.78 [23], Pharmastat Nitrofurantoin 10.54 4.68–11.12 [23], Pharmastat Hospitalization for pyelonephritis 6,918.15 4,842.71–8,993.60 OCCI, hypothesis of plausible range (±30% of base case) Initial clinic visit 77.20 — [24] Follow-up clinic visit 45.90 — [24] Urine culture 45.00 — Laboratoires Biron, oral communication, May 30 2016 Emergency visit 287.60a — [24–26] aEmergency MD cost ($97.60) + total cost for an emergency visit in Quebec in 2009, MD cost excluded ($190). Table 2 Probability values used in the model for decision-tree nodes and ranges used in the sensitivity analyses.

art of this study, a standardized sampling, processing, analysis, and statistics procedure was set by 9 investigators from these centers after 3 days' learning and discussion. All of the centers carefully followed this procedure during the study time. There was no significant difference in compliance among the centers. 2.2. Study Population Consecutive sampling was used to recruit critical ill patients. Patients were eligible for enrolment if they were diagnosed with ischemic stroke by magnetic resonance imaging (MRI), underwent bronchoscopy while mechanically ventilated for clinically suspected VAP, and aged between 18 and 80 years. Patients were excluded from the study if they were moribund or not expected to survive 3 days because of an underlying irreversible medical condition, had active pneumonia when admitted to ICU or had extrapulmonary infection during ICU stay, were immunocompromised, were with malignancies, and were pregnant or if informed consent could not be obtained.

Variable Base case (CAD $) Range tested (CAD $) References Fosfomycin 17.51 13.00–18.00 [23], Pharmastat Sulfonamides 0.95 0.88–2.79 [23], Pharmastat Fluoroquinolones 4.25 3.71–11.78 [23], Pharmastat Nitrofurantoin 10.54 4.68–11.12 [23], Pharmastat Hospitalization for pyelonephritis 6,918.15 4,842.71–8,993.60 OCCI, hypothesis of plausible range (±30% of base case) Initial clinic visit 77.20 — [24] Follow-up clinic visit 45.90 — [24] Urine culture 45.00 — Laboratoires Biron, oral communication, May 30 2016 Emergency visit 287.60a — [24–26] aEmergency MD cost ($97.60) + total cost for an emergency visit in Quebec in 2009, MD cost excluded ($190). Table 2 Probability values used in the model for decision-tree nodes and ranges used in the sensitivity analyses. Variable Base case Range tested References Sensitive isolate: UTI resolution (clinical cure) 0.94 — [27–32] Resistant isolate: UTI resolution (clinical cure) 0.74 — [33] Pyelonephritis due to initial treatment failure 0.04 0–0.08 [27, 28, 34, 35], hypothesis of plausible range Hospitalization for pyelonephritis 0.20 0–0.40 [36], hypothesis of plausible range Empirical extended treatment due to initial treatment failure (switch from initial treatment to another antibiotic) 0.75/0.25 (0.25/0.75 for ciprofloxacin) — [27, 28, 35] Pathogens present in UTI E. coli 0.841 0.767–0.918 [7, 22, 37] Enterococcus spp. 0.028 — [37] K. pneumoniae 0.038 — [37] P. mirabilis 0.026 — [37] Treatment with fosfomycin 0.0177 0–0.250 Pharmastat, hypothesis of plausible range Treatment with sulfonamides 0.2566 — Pharmastat Treatment with nitrofurantoins 0.5410 — Pharmastat Treatment with fluoroquinolones 0.1848 — Pharmastat

7] Enterococcus spp. 0.028 — [37] K. pneumoniae 0.038 — [37] P. mirabilis 0.026 — [37] Treatment with fosfomycin 0.0177 0–0.250 Pharmastat, hypothesis of plausible range Treatment with sulfonamides 0.2566 — Pharmastat Treatment with nitrofurantoins 0.5410 — Pharmastat Treatment with fluoroquinolones 0.1848 — Pharmastat E. coli resistance to fosfomycin 0.006 0.004–0.012 [6, 21, 22] E. coli resistance to sulfonamides (TMP–SMX) 0.267a 0.141–0.294 [15, 16] E. coli resistance to fluoroquinolones (ciprofloxacin) 0.218a 0.191–0.245 [6, 15, 22] E. coli resistance to nitrofurantoin 0.011a 0.0016–0.019 [15, 16, 22] TMP–SMX, trimethoprim–sulfamethoxazole; UTI, urinary tract infection. aDerived from urinary, respiratory, wound, and blood isolates. Table 3 Results of tornado diagram (decision node).

E. coli resistance to fluoroquinolones (ciprofloxacin) 0.218a 0.191–0.245 [6, 15, 22] E. coli resistance to nitrofurantoin 0.011a 0.0016–0.019 [15, 16, 22] TMP–SMX, trimethoprim–sulfamethoxazole; UTI, urinary tract infection. aDerived from urinary, respiratory, wound, and blood isolates. Table 3 Results of tornado diagram (decision node). Variable Range Low EV High EV Spread Spread2 Risk (%) Cumulative risk (%) Variable index Probability of developing PN 0.0 to 0.08 88.93 103.46 14.530 211.121 0.553 0.553 0 Probability of hospitalization for PN 0.0 to 0.4 90.56 101.83 11.270 127.013 0.332 0.885 1 Probability of resistance of E. coli to TMP–SMX 0.141 to 0.294 92.64 96.96 4.320 18.662 0.049 0.934 2 Cost of hospitalization for PN 4842.71 to 8993.60 94.41 97.98 3.570 12.745 0.033 0.967 3 Cost of nitrofurantoin 4.68 to 11.12 93.51 96.20 2.690 7.236 0.019 0.986 4 Probability of being infected by E. coli 0.767 to 0.918 95.56 96.86 1.300 1.690 0.004 1 5 Probability of resistance of E. coli to nitrofurantoin 0.0016 to 0.019 96.20 96.20 0.000 0.000 0.000 1 6 Probability of resistance of E. coli to fosfomycin 0.004 to 0.012 96.20 96.20 0.000 0.000 0.000 1 7 Probability of resistance of E. coli to ciprofloxacin 0.191 to 0.245 96.20 58.69 0 0 0.000 1 8 Cost of ciprofloxacin 3.71 to 11.78 96.20 96.20 0.000 0.000 0.000 1 9 Cost of fosfomycin 13.0 to 18.0 96.20 58.69 0 0 0.000 1 10 EV, expected value; PN, pyelonephritis.

1. Introduction Ventilator-associated pneumonia (VAP) which constitutes a frequent infection in intensive care unit (ICU) patients consumes vast healthcare resources and increases proportionally to the duration of ICU stay and mortality [1]. The incidence of VAP varies depending on the study population. For example, in patients with ischemic stroke who are characterized by advanced age, depressed level of consciousness, immune suppression, and long-term bed rest, the incidence of VAP can increase to approximately 40%, which is associated with a less favorable neurologic and functional outcome [2]. It is recognized that one-third to half of all VAP-related deaths are directly attributable to pneumonia. Diagnosis of VAP with clinical suspicion is overly sensitive with low specificity, leading to unnecessary antibiotics use [3], and the incidence of VAP among ICU ischemic stroke patients has not been thoroughly investigated.

gnized that one-third to half of all VAP-related deaths are directly attributable to pneumonia. Diagnosis of VAP with clinical suspicion is overly sensitive with low specificity, leading to unnecessary antibiotics use [3], and the incidence of VAP among ICU ischemic stroke patients has not been thoroughly investigated. The triggering receptor expressed on myeloid cells-1 (TREM-1) is a member of the immunoglobulin superfamily. Its expression on phagocytes is upregulated by exposure to bacteria and fungi. A soluble form of TREM-1 (sTREM-1) was proposed as a new biomarker which had been tested for acute infections with different diagnostic and prognostic value [4]. sTREM-1 can be found in different body fluids, such as serum, bronchoalveolar lavage fluid (BALF), endotracheal aspiration (ETA), and exhaled breath condensate (EBC), where it can be assayed by ELISA using commercial immunoassay kits [5]. Some clinical studies have proved that sTREM-1 did have the ability to identify patients with sepsis while others come to an opposite conclusion [6, 7].

tudy if they were moribund or not expected to survive 3 days because of an underlying irreversible medical condition, had active pneumonia when admitted to ICU or had extrapulmonary infection during ICU stay, were immunocompromised, were with malignancies, and were pregnant or if informed consent could not be obtained. 2.3. Diagnosis, Treatment, and Prevention of VAP VAP was suspected if the patient had a radiographic infiltrate that was new or progressive, together with clinical findings that were suggestive of infection, such as the onset of fever (temperature ≥38.3°C), leukocytosis (≥10 × 109/L or ≤4 × 109/L), purulent sputum, and decline in oxygenation. In addition, a positive BAL culture (≥104 colony-forming units/mL, cfu/mL) was required to confirm the diagnosis of VAP [8]. The protocol for VAP treatment and prevention followed standard protocols in all institutions also based on accepted guidelines. Care bundles for preventing VAP were also employed [3, 8]. 2.4. Clinical Assessment Baseline assessment included the evaluation of demographic data (age and gender), medical history, stroke subtype, ratio of partial oxygen to fraction of inspired oxygen (PaO2/FiO2), acute physiology and chronic health evaluation (APACHE II), Glasgow Coma Scale (GCS), modified clinical pulmonary infection score (mCPIS), and the level of inflammatory biomarkers.

of demographic data (age and gender), medical history, stroke subtype, ratio of partial oxygen to fraction of inspired oxygen (PaO2/FiO2), acute physiology and chronic health evaluation (APACHE II), Glasgow Coma Scale (GCS), modified clinical pulmonary infection score (mCPIS), and the level of inflammatory biomarkers. 2.5. Samples Processing and Measurement Bronchoscopy was performed on the day when VAP was suspected and all samples were collected for measurement at the same day. The diagnostic flexible bronchoscopy guideline of British Thoracic Society was also followed during the performance [9]. Selection of the segment for bronchoalveolar lavage (BAL) was guided by chest X-ray changes. The right middle lobe or lingual lobe was selected when diffuse infiltrates were present. Five aliquots of 20 mL of sterile saline were instilled and aspirated gently. The first aliquot was discarded and the subsequent four aliquots were pooled for analysis. Part of the first BALF specimen was sent to the laboratory immediately after collection for measurement of sTREM-1 levels; the remainder of the sample was sent to the microbiology laboratory for quantitative culture. Quantitative culture of BALF retrieved by direct bronchoscopic methods yields the best sensitivity and specificity to diagnose VAP and can differentiate true infection from colonization or inflammation. Culture-positive BALF was defined as a count ≥104 cfu/mL. Thus, the diagnosis of pneumonia was confirmed by a positive BALF culture [10].

of BALF retrieved by direct bronchoscopic methods yields the best sensitivity and specificity to diagnose VAP and can differentiate true infection from colonization or inflammation. Culture-positive BALF was defined as a count ≥104 cfu/mL. Thus, the diagnosis of pneumonia was confirmed by a positive BALF culture [10]. The EBC fluid was collected in a plastic container located in the center of the exhaled portion of the ventilator tubing (mid-way between the ventilator and the patient's endotracheal tube). EBC was centrifuged and cell-free supernatants were aliquoted into 2 polypropylene tubes and stored at −80°C. ETA fluid was collected using a feeding tube, which was introduced down the endotracheal tube until resistance was felt. Then, using a syringe, 1 ml/kg of normal saline (maximum 20 ml) was instilled through the feeding tube followed by a small amount of air to clear the dead space of the endotracheal tube and suction was applied to obtain the fluid [11]. Blood samples (5 mL) from median basilic vein were separated by centrifugation (TDL-60C, 6000 revolutions per minute, China) and stored at −80°C in anticoagulative tubes. Commercially available ELISA assays were used for detecting sTREM-1. The detection limit of the methods for sTREM-1 was 0.05~1600 pg/mL (SEA571Hu96 Tests, USA). All samples were run in duplicate. To minimize variations in samples collection, the procedure was performed by the same investigator.

Blood samples (5 mL) from median basilic vein were separated by centrifugation (TDL-60C, 6000 revolutions per minute, China) and stored at −80°C in anticoagulative tubes. Commercially available ELISA assays were used for detecting sTREM-1. The detection limit of the methods for sTREM-1 was 0.05~1600 pg/mL (SEA571Hu96 Tests, USA). All samples were run in duplicate. To minimize variations in samples collection, the procedure was performed by the same investigator. 2.6. Statistical Analysis Statistical analysis was performed using SPSS version 19.0 (IBM for Windows). Data were initially assessed for normality and log transformed where appropriate. Data between the VAP positive and VAP negative were compared using Chi-square test for equal proportion or Fisher exact test where numbers were small with results presented as percentages (n). Continuously normally distributed variables were compared using Student's t-test and presented as means (standard deviations), whereas nonnormally distributed data was compared using Wilcoxon rank-sum test and reported as medians (interquartile range). The statistical tests performed were two-sided. All analysis was performed on an intention-to-treat basis and a two-sided p < 0.05 was considered to be statistically significant. Figures were drawn using Graphpad prism version 5.0 and Medcalc11.4.2.

Wilcoxon rank-sum test and reported as medians (interquartile range). The statistical tests performed were two-sided. All analysis was performed on an intention-to-treat basis and a two-sided p < 0.05 was considered to be statistically significant. Figures were drawn using Graphpad prism version 5.0 and Medcalc11.4.2. 3. Results 3.1. Characteristics of the Patients Over the study period, 328 patients with acute ischemic stroke and requiring mechanical ventilation were admitted to ICU. Thirty-two patients were excluded because informed consent was not obtained. One hundred and sixty-four patients who had active pneumonia when admitted or had extrapulmonary infection during ICU stay were also excluded. One hundred and thirty-two patients clinically suspected of VAP were included in the study. Among them 76 (57.58%) were diagnosed of VAP while 56 were not. The baseline characteristics of the 132 patients were shown in Table 1. The demographic data, position of stroke, PaO2/FiO2, APACHE II, GCS, and modified CPIS were not significantly different between patients with or without VAP (p > 0.05).

d in the study. Among them 76 (57.58%) were diagnosed of VAP while 56 were not. The baseline characteristics of the 132 patients were shown in Table 1. The demographic data, position of stroke, PaO2/FiO2, APACHE II, GCS, and modified CPIS were not significantly different between patients with or without VAP (p > 0.05). 3.2. Bacteria Detection BALF culture was positive (considering the cut-off of >104 cfu/mL) in all the 76 patients with VAP, with the growth of the following agents: Gram-negative bacterium (n = 62, 81.58%). The top three Gram-negative bacteria were Pseudomonas aeruginosa (n = 28), Klebsiella pneumoniae (n = 15), and Acinetobacter baumannii (n = 11). Five Methicillin-resistant Staphylococcus aureus were also detected in 14 strains of Gram-positive bacterium. Blood culture was positive in 4 patients with VAP, with the growth of Staphylococcus aureus methicillin-resistant (n = 3) and Staphylococcus aureus methicillin-sensitive (n = 1). 3.3. sTREM-1 Detection and Comparison sTREM-1 was detected in all the 132 patients at the day of VAP suspicion. The levels of sTREM-1 in serum, BALF, ETA, and EBC between two groups were compared. BALF sTREM-1 concentrations were higher in patients with VAP than patients without VAP [32.35 (IQR, 30.08–41.72) versus 18.92 (11.89–31.72)] pg/mL as well as EBC sTREM-1 concentrations [1.57 (IQR, 1.02–2.61) versus 0.41 (0.19–1.61)] pg/mL (p < 0.05). sTREM-1 levels were consistently higher in BALF than EBC in all patients. Differences of sTREM-1 levels in serum and ETA between two groups were not statistically significant (p > 0.05) (Figure 1).

3.3. sTREM-1 Detection and Comparison sTREM-1 was detected in all the 132 patients at the day of VAP suspicion. The levels of sTREM-1 in serum, BALF, ETA, and EBC between two groups were compared. BALF sTREM-1 concentrations were higher in patients with VAP than patients without VAP [32.35 (IQR, 30.08–41.72) versus 18.92 (11.89–31.72)] pg/mL as well as EBC sTREM-1 concentrations [1.57 (IQR, 1.02–2.61) versus 0.41 (0.19–1.61)] pg/mL (p < 0.05). sTREM-1 levels were consistently higher in BALF than EBC in all patients. Differences of sTREM-1 levels in serum and ETA between two groups were not statistically significant (p > 0.05) (Figure 1). There were 56 patients who were not VAP among the 132 patients included in our study. In these 56 patients, the average time of suspected VAP was about 76 h. We then also included another 56 patients who needed invasive mechanical ventilation during our study period to be a negative control group. None of the 56 patients in the control group was with infectious diseases. We compared the sTREM-1 levels in 76 h of our 56 non-VAP patients with the negative control group and found that the sTREM-1 concentrations in serum [322.94 (IQR, 276.62–427.72) versus 308.69 (284.77–418.93)] pg/mL, BALF [18.92 (IQR, 11.89–31.72) versus 17.98 (10.92–33.60)] pg/mL, ETA [680.53 (IQR, 642.87–718.72) versus 673.98 (657.82–715.21)] pg/mL, and EBC [0.41 (IQR, 0.19–1.61) versus 0.46 (0.17–1.73)] pg/mL were similar. The results proved that the 56 patients in our study did not have any infectious disease during our study period indeed.

IQR, 11.89–31.72) versus 17.98 (10.92–33.60)] pg/mL, ETA [680.53 (IQR, 642.87–718.72) versus 673.98 (657.82–715.21)] pg/mL, and EBC [0.41 (IQR, 0.19–1.61) versus 0.46 (0.17–1.73)] pg/mL were similar. The results proved that the 56 patients in our study did not have any infectious disease during our study period indeed. 3.4. sTREM-1 Sensitivity and Specificity and Diagnostic Values for VAP Receiver-operating characteristic curve (ROC) was shown in Figure 2. A cut-off value of 23.61 pg/mL for sTREM-1 in BALF resulted in a sensitivity of 85.5% and specificity of 73.1% which corresponded to the area of 0.831 under the ROC. A cut-off value of 0.31 pg/mL for sTREM-1 in EBC resulted in a relative high sensitivity of 98.3% and low specificity of 48.1% which corresponded to the area of 0.752 under the ROC. The area under the ROC of serum sTREM-1 concentrations was 0.528 and 0.551 of ETA, which were both having low sensitivity and specificity (Table 2). 3.5. Correlation of sTREM-1 in Different Samples A scatter plot was employed to determine whether there were correlations of sTREM-1 between different samples. The results showed that there was a significant correlation between BALF and EBC sTREM-1 levels (R2 = 0.78, Y = −0.16 + 0.046X, p < 0.05). No statistically significant correlation was identified between the ranks of BALF and serum sTREM-1 values (Y = 300.81 + 1.483X) or the ranks of BALF and ETA sTREM-1 values (Y = 647.01 + 0.023X) , p > 0.05, Figure 3.

significant correlation between BALF and EBC sTREM-1 levels (R2 = 0.78, Y = −0.16 + 0.046X, p < 0.05). No statistically significant correlation was identified between the ranks of BALF and serum sTREM-1 values (Y = 300.81 + 1.483X) or the ranks of BALF and ETA sTREM-1 values (Y = 647.01 + 0.023X) , p > 0.05, Figure 3. 4. Discussion 4.1. Key Findings In this prospective multicenter study, all the 132 patients with acute ischemic stroke and suspected VAP underwent bronchoscopy detection. Samples of serum, BALF, ETA, and EBC of these patients were collected and analyzed. We found that the concentration of sTREM-1 in BALF and EBC could effectively categorize patients as VAP positive or VAP negative when using direct bronchoscopic quantitative culture samples as the comparison standard and they had good correlation (R2 = 0.78). To the best of our knowledge, this is the first study which simultaneously detected concentration of sTREM-1 in serum, ETA, BALF, and EBC in patients who were suspected of VAP. 4.2. Relationship to Previous Studies Despite the major advances that have been achieved in prevention and diagnosis of VAP in the last few years, there are still a large number of unanswered questions. The true incidences of VAP in high risk patients, especially those with ischemic stroke in the ICU, are yet to be determined.

ationship to Previous Studies Despite the major advances that have been achieved in prevention and diagnosis of VAP in the last few years, there are still a large number of unanswered questions. The true incidences of VAP in high risk patients, especially those with ischemic stroke in the ICU, are yet to be determined. What is more, the diagnostic value of sTREM-1 as a biomarker in determining VAP was controversial. A recent study confirmed that sTREM-1 levels did have diagnostic and prognostic values in VAP [12]. However, other studies reported poor sensitivity and specificity of sTREM-1 to detect patients with pneumonia [13, 14]. An important explanation for the disparity lies in the difference in study populations. Some studies did not exclude patients who were immunocompromised which might lead to false negative results of sTREM-1 assay. Others did not exclude the patients who had extrapulmonary infection during ICU stay which increased the sTREM-1 concentration that was not the result of VAP. The value of BALF sTREM-1 in pneumonia was also controversial because of the noninfective influence factors, such as tuberculosis and connective tissue diseases [15]. Additionally, an earlier study indicated that it was hard to classify patients as VAP positive or VAP negative by using CPIS and sTREM-1 [16].

lt of VAP. The value of BALF sTREM-1 in pneumonia was also controversial because of the noninfective influence factors, such as tuberculosis and connective tissue diseases [15]. Additionally, an earlier study indicated that it was hard to classify patients as VAP positive or VAP negative by using CPIS and sTREM-1 [16]. Procalcitonin (PCT) is a prohormone secreted into serum most likely from neuroendocrine cells in the lungs or intestine as part of the systemic inflammatory response. The rapid release and long half-life of procalcitonin make it potentially useful as a diagnostic indicator of VAP. Four studies report the use of serum PCT as a biomarker in diagnosing VAP [17–20]. These studies revealed that sensitivities ranged between 41 and 100% with lower sensitivities indicating the potential to miss many positive VAP patients. Specificity was higher in two of four studies with a range of 97–100%. Variable cut-off values and dissimilar study designs across the studies contribute to the difficulty in interpreting the results. The different patient populations across studies may have contributed to elevated PCT levels that were unrelated to VAP. The most important thing is that some interference factors like previous antibiotics use did affect the mean serum PCT level and sensitivity and specificity which suggested that serum PCT was not a good biomarker for VAP [21].

ations across studies may have contributed to elevated PCT levels that were unrelated to VAP. The most important thing is that some interference factors like previous antibiotics use did affect the mean serum PCT level and sensitivity and specificity which suggested that serum PCT was not a good biomarker for VAP [21]. Therefore, new biomarkers are urgently needed to diagnose VAP especially for patients with ischemic stroke due to its high mortality. So we detected the concentrations of sTREM-1 and expected that it had higher sensitivity and specificity than PCT. 4.3. Study Significance Biomarkers can be detected in any biological sample including serum, ETA, BALF, and EBC. A biomarker for VAP should be low or absent when infection is not present and elevated in the presence of infection. sTREM-1 belongs to the immunoglobulin superfamily, which is expressed in neutrophils, monocytes, and macrophages in the course of acute inflammatory response. sTREM-1 triggers the secretion of proinflammatory mediators through a signaling pathway (DAP12) and functions as an amplifier of the inflammatory response. This characteristic leads to its high sensitivity and poor specificity in infectious diseases. In response to infection, sTREM-1 is either secreted or shed and can be measured in body fluids and is almost undetectable in patients with nonmicrobial inflammation. So in patients with pneumonia, it is a better choice to detect the concentration of sTREM-1 in BALF than in serum which leads to a relative higher specificity.

ponse to infection, sTREM-1 is either secreted or shed and can be measured in body fluids and is almost undetectable in patients with nonmicrobial inflammation. So in patients with pneumonia, it is a better choice to detect the concentration of sTREM-1 in BALF than in serum which leads to a relative higher specificity. One previous study reported that sTREM-1 increased in serum samples of patients with pneumonia and in EBC samples of patients with VAP [15]. Moreover, the raise of sTREM-1 level in pleural effusion, cerebrospinal fluid, urine, and synovial fluid also has important clinical significance in identifying the infectious diseases [22]. Therefore, sTREM-1 as a biomarker for the diagnosis of inflammatory diseases is of high sensitivity and low specificity. According to the characteristics of sTREM-1 expression, it is concluded that analyses of serum sample alone were not able to distinguish between the systemic acute inflammatory response and local infection accurately, although in some patients it could reflect the severity of the infection especially in neonatal sepsis [23]. In our study, the specificity of serum sTREM-1 in diagnosing VAP was only 46.2%. We acquired ETA samples via the artificial airway (endotracheal tube) which could inevitably be contaminated by the bacterial colonization, leading to misinterpretation of the study results. Therefore, our study also analyzed sTREM-1 levels in BALF and EBC in order to achieve more accurate evaluation.

P was only 46.2%. We acquired ETA samples via the artificial airway (endotracheal tube) which could inevitably be contaminated by the bacterial colonization, leading to misinterpretation of the study results. Therefore, our study also analyzed sTREM-1 levels in BALF and EBC in order to achieve more accurate evaluation. Damage of the tracheal mucous membrane and stress could all lead to the raise of sTREM-1 in serum due to its high sensitivity and low specificity which resulted in the poor performance in VAP diagnosis. The area under the ROC is only 0.528. The incidence of VAP will increase along with the incidence of bacterial colonization when mechanical ventilation was prolonged which will lead to the false positive result of sTREM-1 in ETA. In our study, the specificity of sTREM-1 in ETA was only 28.8%, which could not diagnose VAP accurately.

the ROC is only 0.528. The incidence of VAP will increase along with the incidence of bacterial colonization when mechanical ventilation was prolonged which will lead to the false positive result of sTREM-1 in ETA. In our study, the specificity of sTREM-1 in ETA was only 28.8%, which could not diagnose VAP accurately. sTREM-1 as an inflammatory biomarker could be detected in epithelial lining fluid which could also be detected in BALF indirectly that was secreted by lung which was infected as the target organ. Detection of sTREM-1 in BALF was a more effective way to effectively categorize patients as VAP positive or VAP negative and it was less intrusive to extrapulmonary infection. Our study also confirmed that on the day of clinically suspected VAP sTREM-1 in BALF had high sensitivity and specificity with 0.831 area under the ROC which indicated that it had higher diagnostic accuracy of VAP. An increasing number of researches on EBC in the diseases of chronic obstructive pulmonary disease, asthma, and other airway inflammatory diseases were published [24–26]. However, studies on EBC in pneumonia or VAP were relatively rare. What is more, the results of these studies were still controversial because the methods of EBC acquisition, preservation, and detection have not been standardized.

pulmonary disease, asthma, and other airway inflammatory diseases were published [24–26]. However, studies on EBC in pneumonia or VAP were relatively rare. What is more, the results of these studies were still controversial because the methods of EBC acquisition, preservation, and detection have not been standardized. EBC fluid was collected in a plastic container located at the center of the exhaled portion of the ventilator tubing that was far away from the artificial airway. Although the level of EBC was relatively low because it was composed of expiratory air, it had a high sensitivity (98.3%) in diagnosing VAP. However, detection and analysis of EBC were inevitably affected by the bacterial colonization in the artificial airway, which resulted in an unsatisfactory specificity (48.1%). At all events, the area under ROC of sTREM-1 in EBC could reach 0.752, and the concentration of sTREM-1 in EBC had good correlation with sTREM-1 in BALF (R2 = 0.78), indicating that it might be one of the accessory biomarkers in diagnosing VAP. As the collection of EBC is noninvasive, it is promising in clinical practice in the future with the standardized methods. This study has several limitations. We included a homogeneous population of ischemic stroke which might minimize the heterogeneity and avoided misinterpreting the study result. However, its application to routine clinical practice remains uncertain. Whether the detection of BALF sTREM-1 in general may translate into clinical benefits outside this setting deserves further explorations.

ation of ischemic stroke which might minimize the heterogeneity and avoided misinterpreting the study result. However, its application to routine clinical practice remains uncertain. Whether the detection of BALF sTREM-1 in general may translate into clinical benefits outside this setting deserves further explorations. 5. Conclusion In conclusion, sTREM-1 levels in BALF and EBC accurately distinguished patients of acute ischemic stroke with or without VAP when using direct bronchoscopic quantitative culture samples as the reference standard. We can expect the routine use of sTREM-1 levels in BALF or EBC to diagnose patients with VAP in the future with the standardization of samples improvement. Acknowledgments The authors appreciate Zhongheng Zhang from the Department of Critical Care Medicine in Jinhua Municipal Central Hospital for revising the language of the manuscript. Abbreviations APACHE II:Acute physiology and chronic health evaluation BAL:Bronchoalveolar lavage BALF:Bronchoalveolar lavage fluid CFU:Colony-forming units CPIS:Clinical pulmonary infection score EBC:Exhaled breath condensate ELISA:Enzyme-linked immunosorbent assay ETA:Endotracheal aspiration GCS:Glasgow Coma Scale ICU:Intensive care unit IQR:Interquartile range PCT:Procalcitonin POCI:Posterior circulation infarcts ROC:Receiver-operating characteristic curves sTREM-1:Soluble triggering receptor expressed on myeloid cells-1 TACI:Total anterior circulation infarcts VAP:Ventilator acquired pneumonia. Competing Interests The authors have declared that no competing interests exist.

IQR:Interquartile range PCT:Procalcitonin POCI:Posterior circulation infarcts ROC:Receiver-operating characteristic curves sTREM-1:Soluble triggering receptor expressed on myeloid cells-1 TACI:Total anterior circulation infarcts VAP:Ventilator acquired pneumonia. Competing Interests The authors have declared that no competing interests exist. Authors' Contributions Yuetian Yu and Cheng Zhu both conceived and designed the experiments. Yuetian Yu, Cheng Zhu, Chunyan Liu, and Rong Yin performed the experiments. Yuetian Yu and Rong Yin analyzed the data. Cheng Zhu and Rong Yin contributed reagents/materials/analysis tools. Yuetian Yu and Cheng Zhu both helped to draft and edit the article. Yuan Gao and Jianguo Cao revised and approved the final manuscript. Yuetian Yu and Cheng Zhu contributed equally to this work. Figure 1 Median soluble triggering receptor expressed on myeloid cells-1 (sTREM-1) levels in serum, endotracheal aspiration (ETA), bronchoalveolar lavage fluid (BALF), and exhaled breath condensate (EBC) samples in 76 patients with and 56 patients without ventilator-associated pneumonia (VAP). In the data bars, the mid-lines represent medians; the tops and bottoms of the bars represent the 25th and 75th percentiles; ∗p < 0.05. Figure 2 Receiver-operating characteristic curves (ROC) of soluble triggering receptor expressed on myeloid cells-1 (sTREM-1) in serum, endotracheal aspiration (ETA), bronchoalveolar lavage fluid (BALF), and exhaled breath condensate (EBC) samples.

Figure 1 Median soluble triggering receptor expressed on myeloid cells-1 (sTREM-1) levels in serum, endotracheal aspiration (ETA), bronchoalveolar lavage fluid (BALF), and exhaled breath condensate (EBC) samples in 76 patients with and 56 patients without ventilator-associated pneumonia (VAP). In the data bars, the mid-lines represent medians; the tops and bottoms of the bars represent the 25th and 75th percentiles; ∗p < 0.05. Figure 2 Receiver-operating characteristic curves (ROC) of soluble triggering receptor expressed on myeloid cells-1 (sTREM-1) in serum, endotracheal aspiration (ETA), bronchoalveolar lavage fluid (BALF), and exhaled breath condensate (EBC) samples. Figure 3 Correlation of sTREM-1 in different samples. (a) Correlation of BALF and EBC sTREM-1 concentration. (b) Correlation of BALF and serum sTREM-1 concentration. (c) Correlation of BALF and ETA sTREM-1 concentration. Table 1 Characteristics of the study groups (mean ± SD/%/IQR). Characteristics All patients (n = 132) Patients with VAP (n = 76) Patients without VAP (n = 56) p

Figure 3 Correlation of sTREM-1 in different samples. (a) Correlation of BALF and EBC sTREM-1 concentration. (b) Correlation of BALF and serum sTREM-1 concentration. (c) Correlation of BALF and ETA sTREM-1 concentration. Table 1 Characteristics of the study groups (mean ± SD/%/IQR). Characteristics All patients (n = 132) Patients with VAP (n = 76) Patients without VAP (n = 56) p Age, yrs 70.27 ± 11.43 69.39 ± 10.84 70.89 ± 11.36 0.447 Male gender, n (%) 63 (47.7) 37 (48.7) 26 (46.4) 0.861 Previous medical illness Cardiovascular disease, n (%) 70 (53.03) 39 (51.31) 31 (55.35) 0.725 Cerebrovascular disease, n (%) 41 (31.06) 23 (30.26) 18 (32.14) 0.851 Position of stroke TACI, n (%) 48 (36.36) 29 (38.16) 19 (33.93) 0.715 POCI, n (%) 84 (63.64) 47 (61.84) 37 (66.07) 0.715 Smoking history 39 (29.55) 24 (31.58) 15 (26.78) 0.571 Hypoproteinemia 32 (24.24) 17 (22.37) 15 (26.79) 0.682 APACHE II 17.92 ± 4.94 18.21 ± 5.48 17.52 ± 4.88 0.442 GCS 5.71 ± 2.27 5.82 ± 2.35 5.69 ± 2.78 0.085 mCPIS 3.42 ± 1.45 3.37 ± 1.28 3.45 ± 1.61 0.751 PO2/FiO2 (mmHg) 112.67 ± 50.11 115.75 ± 48.58 108.92 ± 52.78 0.443 TACI, total anterior circulation infarcts; POCI, posterior circulation infarcts; APACHE II, acute physiology and chronic health evaluation; GCS, Glasgow Coma Scale; mCPIS, modified clinical pulmonary infection score. Table 2 Best cut-off values of sTREM-1 that were obtained from ROC curves for VAP diagnosis.

Age, yrs 70.27 ± 11.43 69.39 ± 10.84 70.89 ± 11.36 0.447 Male gender, n (%) 63 (47.7) 37 (48.7) 26 (46.4) 0.861 Previous medical illness Cardiovascular disease, n (%) 70 (53.03) 39 (51.31) 31 (55.35) 0.725 Cerebrovascular disease, n (%) 41 (31.06) 23 (30.26) 18 (32.14) 0.851 Position of stroke TACI, n (%) 48 (36.36) 29 (38.16) 19 (33.93) 0.715 POCI, n (%) 84 (63.64) 47 (61.84) 37 (66.07) 0.715 Smoking history 39 (29.55) 24 (31.58) 15 (26.78) 0.571 Hypoproteinemia 32 (24.24) 17 (22.37) 15 (26.79) 0.682 APACHE II 17.92 ± 4.94 18.21 ± 5.48 17.52 ± 4.88 0.442 GCS 5.71 ± 2.27 5.82 ± 2.35 5.69 ± 2.78 0.085 mCPIS 3.42 ± 1.45 3.37 ± 1.28 3.45 ± 1.61 0.751 PO2/FiO2 (mmHg) 112.67 ± 50.11 115.75 ± 48.58 108.92 ± 52.78 0.443 TACI, total anterior circulation infarcts; POCI, posterior circulation infarcts; APACHE II, acute physiology and chronic health evaluation; GCS, Glasgow Coma Scale; mCPIS, modified clinical pulmonary infection score. Table 2 Best cut-off values of sTREM-1 that were obtained from ROC curves for VAP diagnosis. Variables AUC 95% CI Cut-off (pg/mL) Sensitivity% Specificity% sTREM-1 in serum 0.528 0.438～0.617 358.93 72.4 46.2 sTREM-1 in ETA 0.551 0.461～0.638 641.94 85.5 28.8 sTREM-1 in BALF 0.831 0.755～0.892 23.61 85.5 73.1 sTREM-1 in EBC 0.752 0.668～0.824 0.32 100 48.1 AUC = area under the curve.

1. Introduction Diarrhea is the third leading cause of deaths in the developing countries. Each year, about 1.8 million people die from diarrheal diseases worldwide, and 90% of the victims are children under the age of five [1, 2]. Cryptosporidiosis is a type of the diarrheal illness caused by the infection of Cryptosporidium. Cryptosporidium is especially dangerous for infants, children, older people, pregnant women, and individuals with Acquired Immune Deficiency Syndrome. There has been an increased prevalence of cryptosporidiosis in densely populated areas in China, such as Beijing, Shanghai, Jiangsu, and Anhui Province [3, 4]. Most infections of Cryptosporidium stem from recreational water bodies [5]. Considering the risk of different ways of Cryptosporidium infection (i.e., drinking unboiled tap water and swimming and diving in the water bodies), swimming poses the highest risk of infection [6]. However, this exposure pathway for Cryptosporidium infection has received little attention in China. Recreational water bodies can be contaminated by animal excretions, runoff from Cryptosporidium-contaminated soil, and discharge from wastewater treatment facilities and sewage systems [7]. A single infected animal excretion could contain billions of oocysts [5]. These oocysts are very resistant to many environmental stresses and can survive for a long period of time. Due to their small size and low settling velocity, they are very slow to settle out of watercourses [8]. For the above-mentioned reasons, swimmers health is highly jeopardized when they swim in recreational water bodies.

ocysts are very resistant to many environmental stresses and can survive for a long period of time. Due to their small size and low settling velocity, they are very slow to settle out of watercourses [8]. For the above-mentioned reasons, swimmers health is highly jeopardized when they swim in recreational water bodies. The primary purpose of this paper is to show the distribution of Cryptosporidium in the Yunlong Lake and to discuss the potential health risks while swimming or diving in this lake.

ocysts are very resistant to many environmental stresses and can survive for a long period of time. Due to their small size and low settling velocity, they are very slow to settle out of watercourses [8]. For the above-mentioned reasons, swimmers health is highly jeopardized when they swim in recreational water bodies. The primary purpose of this paper is to show the distribution of Cryptosporidium in the Yunlong Lake and to discuss the potential health risks while swimming or diving in this lake. 2. Methods Yunlong Lake is located in the southwest of Xuzhou City and has a surface area of 6.8 km2 and perimeter of 12 km. It is a recreational aquatic venue used for swimming and boating. A total of 120 water samples and 60 sediments samples were collected from 12 sampling sites along the lake (as shown in Figure 1). The sampling sites included restaurants, docks, fishing areas, boating training bases, and areas around living quarters. In each sampling site, duplicate water samples were collected from 0.1–2 meters below the water surface. Approximately 10 L of water was collected in a 10 L sterile plastic tank. Sediment samples were taken by a piston cylindrical sampler (KHT0204, Yinhua, Shangyu). Water and sediment samples were collected five times in January, March, July, August, and November 2015. The months for sampling, July and August, are the highest flow months of the year as well as the peak summer vacation months in China. Therefore, the proportion of people swimming in this two-month period is very high in comparison to the rest of the year. January and November were chosen because they are the driest months reported in this region.

and August, are the highest flow months of the year as well as the peak summer vacation months in China. Therefore, the proportion of people swimming in this two-month period is very high in comparison to the rest of the year. January and November were chosen because they are the driest months reported in this region. Water samples were filtered through membrane filters with a pore size of 2 μm by a vacuum device and concentrated using two centrifugation steps at 4,000 ×g at 4°C for 15 minutes, followed by two microcentrifugation steps at 15,000 ×g at 4°C for 5 minutes. The entire resuspended pellet was transferred into the slide wells. Sediment samples (20 g) taken from each site were mixed with distilled water and concentrated using two centrifugation steps at 2,000 ×g at 4°C for 20 min, discarding the supernatant and mixing the pellet with distilled water. The mixed solution was taken and transferred into a microcentrifuge tube and centrifuged at 2000 ×g at 4°C for 10 min [9]. Cryptosporidium oocysts were numerated using an epifluorescence microscope under 200x magnification (Environmental Protection Agency (EPA) Method 1622) [10].

mixing the pellet with distilled water. The mixed solution was taken and transferred into a microcentrifuge tube and centrifuged at 2000 ×g at 4°C for 10 min [9]. Cryptosporidium oocysts were numerated using an epifluorescence microscope under 200x magnification (Environmental Protection Agency (EPA) Method 1622) [10]. Exposure assessments consist of identifying the pathways (i.e., exposure routes: swimming and diving in the lake) and numbers of Cryptosporidium oocysts reaching a person and potentially leading to infection [6]. It has been reported about 18–37 mL of water is swallowed by swimmers per swimming event [11]. Additionally, 10.5–21.6% of people swim twenty times per year, and people swim an average of once a week during the summer [12]. Further, approximately 2.8–13 mL of water is swallowed for each dive and the estimated risk of diving was 2.1% for divers [13]. The studied area, Xuzhou, has a similar situation when compared with the previous research areas [6]. Therefore, the exposure frequency and population proportion were adopted for risk assessment in this study.

ately 2.8–13 mL of water is swallowed for each dive and the estimated risk of diving was 2.1% for divers [13]. The studied area, Xuzhou, has a similar situation when compared with the previous research areas [6]. Therefore, the exposure frequency and population proportion were adopted for risk assessment in this study. The exponential dose response model was used to estimate the infection risk of a single exposure [6], and the model is shown below: (1) Pi=1−e−rN, where “Pi” is the probability of infection and “r” shows the probability that Cryptosporidium can reach and infect a person. According to United States Environmental Protection Agency (USEPA), rCryptosporidium = 0.09 [14]. N is the dose, calculated by (2) N=CVrec, where “C” is the detected Cryptosporidium concentration in water samples, with unit of number of Cryptosporidium per liter (n/L), rec is the recovery rate (rec was 0.3), and “V” is individual consumption of water in liters (L) [6, 15, 16]. An individual annual risk is defined as the risk associated with an individual's exposure to Cryptosporidium over a period of the year. The calculation of annual risk of infection, “Piy,” is shown as below: (3) Piy=1−1−Pin, where “n” is the annual exposure number. Beta probability density distribution was used to model the probability of annual illness that a healthy consumer is infected by Cryptosporidium [17]. The annual illness risk is shown as follows: (4) Pily=Piy×Pil, where “Pily” is probability of annual illness and “Pil” is probability of illness given infection.

An individual annual risk is defined as the risk associated with an individual's exposure to Cryptosporidium over a period of the year. The calculation of annual risk of infection, “Piy,” is shown as below: (3) Piy=1−1−Pin, where “n” is the annual exposure number. Beta probability density distribution was used to model the probability of annual illness that a healthy consumer is infected by Cryptosporidium [17]. The annual illness risk is shown as follows: (4) Pily=Piy×Pil, where “Pily” is probability of annual illness and “Pil” is probability of illness given infection. Table 1 gives the model parameters for risk assessment of Cryptosporidium in Yunlong Lake. 3. Results Figures 2 and 3 show the level of Cryptosporidium oocysts in water and sediment samples, respectively. Cryptosporidium oocysts were 0–8/10 L (with a mean value of 2.6/10 L) in water samples and 0–260/g (with a mean value of 129 oocysts/g) in sediment samples. The detection results of the samples are shown in Table 2. A higher incidence of Cryptosporidium positive samples was found in the rainy season (i.e., July-August) as compared to the results obtained in other periods. The seasonal pattern of Cryptosporidium in Yunlong Lake is similar to the Three Gorges Reservoir, China [6]. Cryptosporidium in water samples of sampling sites 9 and 12 was detected positive for the whole year. It was also observed that water at sampling sites 9 and 12 was very cloudy.

esults obtained in other periods. The seasonal pattern of Cryptosporidium in Yunlong Lake is similar to the Three Gorges Reservoir, China [6]. Cryptosporidium in water samples of sampling sites 9 and 12 was detected positive for the whole year. It was also observed that water at sampling sites 9 and 12 was very cloudy. Table 4 shows the risk of infection and illness by swimming and diving exposure in the lake according to the measured oocysts number per liter of the lake water. Swimming in July has the highest infection risk of 1.85 × 10−3 per time, while swimming in March has the lowest infection risk of 5.55 × 10−4 per time. Swimming in July has the highest yearly infection (3.64 × 10−2) and illness risks (1.82 × 10−2). Table 5 shows the risk of infection and illness caused by diving exposure in the lake according to measured oocysts number per liter of the lake water. Diving in July has the highest infection risk of 6.51 × 10−4 per time, while diving in March has the lowest infection risk of 1.95 × 10−4 per time. Diving in July has the highest yearly infection (1.29 × 10−2) and illness risks (6.45 × 10−3). 4. Discussions

Table 5 shows the risk of infection and illness caused by diving exposure in the lake according to measured oocysts number per liter of the lake water. Diving in July has the highest infection risk of 6.51 × 10−4 per time, while diving in March has the lowest infection risk of 1.95 × 10−4 per time. Diving in July has the highest yearly infection (1.29 × 10−2) and illness risks (6.45 × 10−3). 4. Discussions Cryptosporidium have been detected in source water bodies worldwide [15, 18–20]. However, cryptosporidiosis outbreak in recreational water is more frequent than in source water [5]. This study provides information on Yunlong Lake contamination by Cryptosporidium. The incidence of Cryptosporidium was not found to be higher than other water bodies reported in China (Table 3). Cryptosporidium was more frequently detected in sediments than that of surface water in Yunlong Lake (with positive rate of 47% and 35% for sediments and surface water, resp.) (Table 2). In the present study, Cryptosporidium concentration in sediment samples was not strongly correlated with those of water samples with a correlation coefficient of only 0.23. Oocysts cannot be homogeneously distributed throughout the water [21], which could help to understand the weak correlation between Cryptosporidium concentration in water and in sediments. An obvious seasonal contamination pattern was observed, with higher Cryptosporidium frequency in summer compared to the other seasons, which was similar to Xiao's report [6].

d throughout the water [21], which could help to understand the weak correlation between Cryptosporidium concentration in water and in sediments. An obvious seasonal contamination pattern was observed, with higher Cryptosporidium frequency in summer compared to the other seasons, which was similar to Xiao's report [6]. Modeling results indicated a potential health risk for people swimming or diving in Yunlong Lake. Compared to diving, swimming has the higher infection risk (Table 4). Similarly, An and coworkers reported that swimming is the primary risk of Cryptosporidium infection in rivers [22]. In regard to an acceptable risk level, an annual individual infection risk of 10−4 was suggested by the USEPA. The average infection risks of Cryptosporidium for swimmers and divers in Yunlong Lake were all greater than the USEPA's acceptable risk level (Tables 4 and 5). A compilation of Cryptosporidium concentrations in water samples from other study areas is listed in Table 3. Average concentration of Cryptosporidium oocysts in the present study (2.6/10 L water) is comparable to the others. Watersheds from most of China have similar or higher Cryptosporidium positive rates in comparison with Yunlong Lake.

osporidium concentrations in water samples from other study areas is listed in Table 3. Average concentration of Cryptosporidium oocysts in the present study (2.6/10 L water) is comparable to the others. Watersheds from most of China have similar or higher Cryptosporidium positive rates in comparison with Yunlong Lake. 5. Conclusions On the basis of the above results and discussions, it was concluded that swimming in Yunlong Lake has a higher risk of Cryptosporidium infection than the acceptable risk level set by the USEPA. Multiple-site monitoring of recreational water quality and timely reporting of information regarding infection risk are strongly recommended based on the data showing that this lake is indeed contaminated by Cryptosporidium. Acknowledgments The authors wish to thank the “the Fundamental Research Funds for the Central Universities” (2017QNA35), National Natural Science Foundation of China (41403090), and the Foundation of JiangSu Collaborative Innovation Center for Building Energy Saving and Construct Technology (SJXTY1507). Abbreviations USEPA:United States Environmental Protection Agency. Conflicts of Interest The authors declare that there are no conflicts of interest regarding the publication of this paper. Figure 1 Sample collection sites. Figure 2 Cryptosporidium oocysts in water samples. Figure 3 Cryptosporidium oocysts in sediment samples. Table 1 Model parameters for risk assessment of Cryptosporidium.

Abbreviations USEPA:United States Environmental Protection Agency. Conflicts of Interest The authors declare that there are no conflicts of interest regarding the publication of this paper. Figure 1 Sample collection sites. Figure 2 Cryptosporidium oocysts in water samples. Figure 3 Cryptosporidium oocysts in sediment samples. Table 1 Model parameters for risk assessment of Cryptosporidium. Variable description (unit) Mean January average oocysts concentration, sites 1–12 (per 100 L) 25 March average oocysts concentration, sites 1–12 (per 100 L) 15 July average oocysts concentration, sites 1–12 (per 100 L) 33 August average oocysts concentration, sites 1–12 (per 100 L) 30 November average oocysts concentration, sites 1–12 (per 100 L) 26 Water ingestion volume by swimming in the lake (mL/time) 37 Water ingestion volume by diving in the lake (mL/time) 13 Percentage of exposure swimmers (%) 16.05 Percentage of exposure divers (%) 2.1 Infectivity of annual exposures for swimming (times) 20 Infectivity of annual exposures for diving (times) 19.82 Probability of illness given infection 0.5 Table 2 Sample detection results for the 12 sites in Yunlong Lake, Xuzhou, 2015. Number of the positive samples detected Positivity rate in percent January Water 4 33.3% Sediment 5 41.7% March Water 4 33.3% Sediment 6 50.0% July Water 6 50.0% Sediment 6 50.0% August Water 4 33.3% Sediment 6 50.0% November Water 3 25.0% Sediment 5 41.7% Table 3 Occurrence of Cryptosporidium in water collected from various sites in China.

samples detected Positivity rate in percent January Water 4 33.3% Sediment 5 41.7% March Water 4 33.3% Sediment 6 50.0% July Water 6 50.0% Sediment 6 50.0% August Water 4 33.3% Sediment 6 50.0% November Water 3 25.0% Sediment 5 41.7% Table 3 Occurrence of Cryptosporidium in water collected from various sites in China. Location of sampling Mean (oocysts/10 L) Positive rate Year of sampling Reference The Three Gorges Reservoir 1.92 100% 2011 [6] River network system, Tongxiang 5.1 78.7% 2009 [23] Water samples in Qinghai N/A 25.4% 2011-2012 [24] Waterworks in 33 major cities 0.7 N/A 2009–2011 [25] Huangpu River N/A 37.6% 2013-2014 [26] Wastewater treatment plants in Harbin N/A 31.3% 2009-2010 [27] Source water, Shanghai 5.2 32% 2009-2010 [28] Recreational water, Xuzhou 0.97 35% 2015 This manuscript Table 4 Simulated risks by exposure events in Yunlong Lake (swimming). Exposure routes Probability of infection per time Probability of infection per year Probability of illness per year Swimming in January 9.21 × 10−4 1.83 × 10−2 9.15 × 10−3 Swimming in March 5.55 × 10−4 1.10 × 10−2 5.51 × 10−3 Swimming in July 1.85 × 10−3 3.64 × 10−2 1.82 × 10−2 Swimming in August 1.11 × 10−3 2.20 × 10−2 1.10 × 10−2 Swimming in November 7.43 × 10−4 1.48 × 10−2 7.40 × 10−3 Table 5 Simulated risks by exposure events in Yunlong Lake (diving).

1 × 10−4 1.83 × 10−2 9.15 × 10−3 Swimming in March 5.55 × 10−4 1.10 × 10−2 5.51 × 10−3 Swimming in July 1.85 × 10−3 3.64 × 10−2 1.82 × 10−2 Swimming in August 1.11 × 10−3 2.20 × 10−2 1.10 × 10−2 Swimming in November 7.43 × 10−4 1.48 × 10−2 7.40 × 10−3 Table 5 Simulated risks by exposure events in Yunlong Lake (diving). Exposure routes Probability of infection per time Probability of infection per year Probability of illness per year Diving in January 3.24 × 10−4 6.46 × 10−3 3.23 × 10−3 Diving in March 1.95 × 10−4 3.89 × 10−3 1.95 × 10−3 Diving in July 6.51 × 10−4 1.29 × 10−2 6.45 × 10−3 Diving in August 3.90 × 10−4 7.77 × 10−3 3.89 × 10−3 Diving in November 2.61 × 10−4 5.21 × 10−3 2.61 × 10−3

1. Introduction Mycobacterium ulcerans is the leading cause for Buruli ulcer disease (BUD), the third most widespread mycobacteriosis after tuberculosis and leprosy [1]. The disease has been reported in over thirty countries worldwide, including Ghana [2, 3], where occurrence rates of up to 22% have been reported in certain areas [4]. The disease has been recognized by World Health Organization (WHO) as an important emerging infectious disease of public health concern [5]. BUD, a skin condition, is characterized by preulcerative and ulcerative lesions [6]. The observed clinical presentations include nodules, plaques, papules, oedemas, and ulcers. These presentations have been observed to be the cytotoxic effect of the toxin mycolactone produced by the growing MU in the host [6, 7].

[5]. BUD, a skin condition, is characterized by preulcerative and ulcerative lesions [6]. The observed clinical presentations include nodules, plaques, papules, oedemas, and ulcers. These presentations have been observed to be the cytotoxic effect of the toxin mycolactone produced by the growing MU in the host [6, 7]. Hitherto, the treatment of Buruli ulcer disease relied on surgical excision and repair [8]. A study by Etuaful et al. in Ghana showed that the oral administration of rifampicin and intramuscular administration of streptomycin were effective at the preulcerative and the early ulcerative stages [9]. Currently, Buruli ulcer is managed on a regimen of streptomycin and rifampicin for an eight-week period as recommended by the WHO [9, 10]. This was a shift from an initial wide surgical excision of the infected tissue, followed by skin grafting [11, 12]. In severe cases, patients are treated with antibiotics in addition to surgery. Other treatment options include topical applications such as phenytoin [13], hyperbaric oxygen, and heat [14]. Meanwhile, disease recurrence of up to 16% has been documented in Ghana [15].

infected tissue, followed by skin grafting [11, 12]. In severe cases, patients are treated with antibiotics in addition to surgery. Other treatment options include topical applications such as phenytoin [13], hyperbaric oxygen, and heat [14]. Meanwhile, disease recurrence of up to 16% has been documented in Ghana [15]. Similar to most infectious diseases, the administration of antibiotics is usually recommended after laboratory confirmation of the condition. This is challenging, as the disease usually occurs in rural endemic communities with limited access to healthcare facilities [16, 17]. Even in areas where the facilities are available, polymerase chain reaction (PCR), the most sensitive technique for disease confirmation, is mostly not performed. The implication is possible misdiagnosis leading to misuse of antibiotics, an occurrence that can lead to the generation of mutant strains. Furthermore, the intramuscular administration of streptomycin can also lead to patient's noncompliance. The use of streptomycin may be inconveniencing particularly in paediatrics, since they have been identified to form a greater percentage of BUD patients [12].

rrence that can lead to the generation of mutant strains. Furthermore, the intramuscular administration of streptomycin can also lead to patient's noncompliance. The use of streptomycin may be inconveniencing particularly in paediatrics, since they have been identified to form a greater percentage of BUD patients [12]. Several in vitro studies have been conducted on the susceptibility of Mycobacterium ulcerans to antimicrobials, leading to in vivo studies and clinical trials [18–20]. However, the performance of more in vitro investigations on antimicrobials at various locations is crucial. Such investigations may bring out new antimicrobial agents useful for the management of Buruli ulcer disease in addition to what is already available. This study therefore investigated the in vitro activity of eight antimicrobials against isolates of M. ulcerans.

tions on antimicrobials at various locations is crucial. Such investigations may bring out new antimicrobial agents useful for the management of Buruli ulcer disease in addition to what is already available. This study therefore investigated the in vitro activity of eight antimicrobials against isolates of M. ulcerans. 2. Methods 2.1. Study Site and Collection of Bacterial Isolates The study was carried out at the Animal Experimental Department of the Noguchi Memorial Institute for Medical Research (NMIMR) in Ghana, after obtaining approval from Ethical and Protocol Review Committee of NMIMR and the Ghana Health Service (GHS), [Study # 040/09-10]. Two M. ulcerans reference strains (ATCC 970321, Ghana D19F9, and ATCC 990826, Benin D27D14) kindly donated from the laboratories of Professor Francois Portaels, Belgium, and ten clinical M. ulcerans isolates from Animal Experimentation and Bacteriology Departments of NMIMR, Ghana, were used for the study. The 10 isolates comprised six (6) from Amasaman and four (4) from Nsawam, both communities in Ghana. Nsawam is a town in south Ghana and is the capital of the Akuapim South Municipal District, a district in the Eastern Region of Ghana. Amasaman on the other hand is a small town and the capital of Ga West District of the Greater Accra region, located about 20 kilometers west of Accra, the capital city of Ghana [21]. A Community Health Centre is situated in the Amasaman community that cares for all Buruli ulcer patients in the Ga District and its surrounding communities [21]. Swab specimens having exudates from lesions were collected from the 10 different BU patients from Amasaman and Nsawam communities and cultured for 8–12 weeks at 32°C on Lowenstein-Jensen (L-J) medium. Growth characteristic M. ulcerans colonies were noticed after the incubation period [16]. M. ulcerans colonies were confirmed by conventional laboratory methods [17], by subjecting the cultures with growth to Ziehl-Neelsen staining and a confirmatory IS2404 PCR [17]. Briefly, for the Ziehl-Neelsen staining procedure, smears were prepared from the cultures with growth and allowed to dry, after which they were heat fixed and flooded with carbol fuchsin stain. After application of heat beneath the slide until steam appeared, the slide was then left for 5 minutes while applying heat intermittently. Excess carbol fuchsin that resulted after the flooding of the slide with the stain was washed off with tap water and the excess water was drained off.

ed with carbol fuchsin stain. After application of heat beneath the slide until steam appeared, the slide was then left for 5 minutes while applying heat intermittently. Excess carbol fuchsin that resulted after the flooding of the slide with the stain was washed off with tap water and the excess water was drained off. Decolorization was then done by covering the slide with 20% H2SO4 and left to stand for 2–5 minutes. The slide was subsequently counterstained with 3% methylene blue and left for 2 minutes after which it was washed with plenty of water, drained, and left to dry. The stained slide was then observed with Olympus BX40 light microscope (New York Microscope Co., USA) under oil immersion. The isolates from these two towns (Nsawam and Amasaman) in addition to the reference strains were tested against eight selected antimicrobial agents, namely, amikacin, azithromycin, dapsone, ciprofloxacin, clofazimine, ofloxacin, rifampicin, and streptomycin. All the antibiotics were obtained from Fluka Biochemika (St. Louis, MO) through World Health Organization.

Amasaman) in addition to the reference strains were tested against eight selected antimicrobial agents, namely, amikacin, azithromycin, dapsone, ciprofloxacin, clofazimine, ofloxacin, rifampicin, and streptomycin. All the antibiotics were obtained from Fluka Biochemika (St. Louis, MO) through World Health Organization. 2.2. Determination of Anti-M. ulcerans Activities After authentication of the MU isolates by IS2404 PCR, the anti-M. ulcerans activities were determined by the agar dilution method by incorporating 100 μg/mL stock solutions prepared by diluting 1 mg of antibiotic in 10 mL of appropriate diluents. This was incorporated into Lowenstein-Jensen medium at a dilution ratio of 1 : 20 to obtain a final concentration of 5 μg/mL. M. ulcerans suspensions were prepared by adjusting turbidity to MacFarland's standard solution 1. This was prepared in sterile phosphate buffered saline and matching with the standard solution equivalent to a concentration of 1 mg/mL of bacilli.

at a dilution ratio of 1 : 20 to obtain a final concentration of 5 μg/mL. M. ulcerans suspensions were prepared by adjusting turbidity to MacFarland's standard solution 1. This was prepared in sterile phosphate buffered saline and matching with the standard solution equivalent to a concentration of 1 mg/mL of bacilli. The anti-M. ulcerans activities of the antimicrobials were determined using the agar dilution method [22]. Briefly, labelled media test tubes containing antimicrobial-incorporated and antimicrobial-free Lowenstein-Jensen media were inoculated with 100 μL of the test isolates. The tubes were gently rolled to ensure even distribution of inoculum on surface of L-J medium. The tests were done in triplicate. They were incubated at 32°C for 8 weeks under microaerophilic conditions [23]. An antimicrobial-incorporated L-J medium that had no colonies growing after 8 weeks on all the test triplicates was categorized as having anti-M. ulcerans activity, in vitro. 2.3. Determination of Minimum Inhibitory Concentrations (MICs) The MICs of the antimicrobials were determined by Canetti's agar proportion method, modified, and used by Thangaraj et al. [19]. Briefly, the antimicrobials with anti-M. ulcerans activities were incorporated into L-J media at two-fold serial dilutions (1–8 dilutions). This was distributed into screw-capped tubes filled with required volumes and concentrations of antimicrobials before inspissation. The final test concentrations ranged between 0.04 μg/mL and 5 μg/mL in L-J medium.

nti-M. ulcerans activities were incorporated into L-J media at two-fold serial dilutions (1–8 dilutions). This was distributed into screw-capped tubes filled with required volumes and concentrations of antimicrobials before inspissation. The final test concentrations ranged between 0.04 μg/mL and 5 μg/mL in L-J medium. Triplicates of each dilution of the antimicrobial-incorporated and antimicrobial-free media were inoculated with 100 μL of the prepared inoculum containing various isolates (Ghana and Benin reference strains and the 10 clinical isolates from Amasaman and Nsawam). Noninoculated triplicates of antimicrobial-incorporated media and antimicrobial-free media were incubated alongside the inoculated ones to serve as drug and L-J media controls, respectively. All tubes were incubated at 32°C for eight weeks under microaerophilic conditions. The tubes were read weekly by observing for growth. The lowest concentration of antimicrobial that completely inhibited growth (zero colonies) was taken as the MIC. 2.4. Statistical Analyses Data obtained was stored in Microsoft Excel and the analysis was conducted by the Statistical Package for the Social Sciences (SPSS), version 12.0.1. Data was summarized by determining the means, median, minimum, and maximum values of the MICs of the antibiotics. Specific determinations included differences in the MICs of the antibiotics and differences in the susceptibilities of the M. ulcerans isolates to the antibiotics.

he Social Sciences (SPSS), version 12.0.1. Data was summarized by determining the means, median, minimum, and maximum values of the MICs of the antibiotics. Specific determinations included differences in the MICs of the antibiotics and differences in the susceptibilities of the M. ulcerans isolates to the antibiotics. 3. Results All twelve M. ulcerans isolates were susceptible to all eight antimicrobials tested. This shows that all antimicrobials had an anti-M. ulcerans activity. Generally, the MICs ranged between 0.16 μg/mL and 2.5 μg/mL, with azithromycin and rifampicin showing the highest inhibitory activity of 0.16 μg/mL (Table 1). For azithromycin, the MIC of 0.16 μg/mL was observed in four of the isolates (one reference and three clinical isolates). All of those isolates that were inhibited by azithromycin MIC of 0.16 μg/mL were from the Amasaman community (Table 1). Meanwhile, the MIC of 0.16 μg/mL for rifampicin was observed in two of the isolates, also all from Amasaman community. The isolates (A9 and A13) that were inhibited by MIC of 0.16 μg/mL for rifampicin were also inhibited by the same MIC value (0.16 μg/mL) for azithromycin. Clofazimine had the lowest inhibitory activity of 2.5 μg/mL for nine out of twelve isolates tested (Table 1).

two of the isolates, also all from Amasaman community. The isolates (A9 and A13) that were inhibited by MIC of 0.16 μg/mL for rifampicin were also inhibited by the same MIC value (0.16 μg/mL) for azithromycin. Clofazimine had the lowest inhibitory activity of 2.5 μg/mL for nine out of twelve isolates tested (Table 1). The best mean MIC was recorded by azithromycin (0.39 μg/mL), followed by rifampicin (0.81 μg/mL). Meanwhile, clofazimine performed very poorly with a mean MIC of 2.19 μg/mL (Table 2). The MIC ranges for the individual antibiotics are the following: Az (0.16–0.63), Rif (0.16–1.25), Dap (0.31–1.25), Str (0.63–1.25), Amk (0.63–2.50), Cip (0.63–1.25), Ofl (0.63–2.50), and Clo (1.25–2.50) (Table 2). 4. Discussions From this study, all twelve M. ulcerans isolates were susceptible to azithromycin, rifampicin, streptomycin, amikacin, ciprofloxacin, ofloxacin, dapsone, and clofazimine at a screening dilution of 5 μg/mL. The MICs obtained for eight antimicrobials against isolates of M. ulcerans ranged between 0.16 μg/mL and 2.5 μg/mL. Azithromycin exhibited the highest inhibitory activity of 0.16 μg/mL against four of the isolates while clofazimine had the lowest MIC of 2.5 μg/mL against nine isolates out of the twelve tested. This observation supports outcomes of studies that have indicated that azithromycin has some level of antimycobacterial activity in vitro and in vivo [24, 25].

inhibitory activity of 0.16 μg/mL against four of the isolates while clofazimine had the lowest MIC of 2.5 μg/mL against nine isolates out of the twelve tested. This observation supports outcomes of studies that have indicated that azithromycin has some level of antimycobacterial activity in vitro and in vivo [24, 25]. Between the two antileprosy drugs included in this study, dapsone had better activity than clofazimine, even though clofazimine is a well-known second-line antileprosy drug [26]. In a study by Espey et al., dapsone was administered in combination with rifampicin and it reduced the size of BU lesions impressively [27]. The fluoroquinolones, ciprofloxacin and ofloxacin, exhibited activities slightly better than clofazimine, with results comparable to results by Thangaraj et al., who also studied Ghanaian isolates [19]. However, fluoroquinolones in general are not recommended for pregnant women and children because of a possible negative impact on articular cartilage [12]. Among the four known anti-TB drugs investigated in this study (rifampicin, streptomycin, amikacin, and ciprofloxacin), the highest activity was registered by rifampicin at 0.81 μg/mL, followed by streptomycin at 1.10 μg/mL. The minimum inhibitory concentrations of the eight antibiotics were also in consonance with values obtained by other investigators [19, 28, 29]. This observation only confirms WHO's recommendation for the use of rifampicin combination with streptomycin as the provisional drugs for the management of Buruli ulcer [10].

. The minimum inhibitory concentrations of the eight antibiotics were also in consonance with values obtained by other investigators [19, 28, 29]. This observation only confirms WHO's recommendation for the use of rifampicin combination with streptomycin as the provisional drugs for the management of Buruli ulcer [10]. Although Buruli ulcer affects all age groups, children have been identified to represent the majority of patients [12]. Therefore, it is important to assure compliance despite effects such as nausea, stomach pain, and diarrhea that patients may associate with a particular drug [12]. Streptomycin, although recommended, has a lot of limitations in terms of its mode of administration. The mode of administration makes its use in paediatrics discouraging. Rifampicin on the other hand also has the characteristics of coloring of body fluids [30].

a that patients may associate with a particular drug [12]. Streptomycin, although recommended, has a lot of limitations in terms of its mode of administration. The mode of administration makes its use in paediatrics discouraging. Rifampicin on the other hand also has the characteristics of coloring of body fluids [30]. Azithromycin has added properties such as its oral administration, rapid absorption from the gastrointestinal tract, wide distribution, its ability to penetrate and concentration in host tissue, and an efficient excretion rate [29]. In addition, it has a high peak serum concentration and a high therapeutic index. These properties of azithromycin would make it an important therapeutic agent for the disease, after its activity in vivo has been established [29]. Therefore, azithromycin would provide a more user-friendly alternative. The in vivo activity of azithromycin in combination with amikacin against M. ulcerans infection has been studied by Bentoucha et al. [31]. In a study conducted by Addo et al., grasscutters experimentally infected with M. ulcerans developed lesions that resolved and recurred after a period. The pathognomonic indicators were similar to those shown in humans [32]. This corroborated rates of recrudescence reported on the disease and the possibility of paying less attention to a seemingly resolved BU lesion. Recurrence is an important reality in BU management and antimicrobials such as azithromycin can play a very important role, particularly in the preulcerative and postsurgical periods in BU disease due to their stability and their penetrative capacity [33–35]. Additionally, multidrug regimens containing azithromycin may help control secondary bacterial infections sometimes seen in BU patients. Even though all the M. ulcerans isolates exhibited variable susceptibility to all the antimicrobials tested, the Nsawam isolates showed relatively lower susceptibility to the antimicrobials as compared to the Amasaman isolates, as well as the control strains. The ranges were however narrow. Exposures to antimicrobials are most likely to create decreased susceptibility as observed in some of the isolates. These are only observations in susceptibility patterns exhibited by the M. ulcerans isolates used in this study.

d to the Amasaman isolates, as well as the control strains. The ranges were however narrow. Exposures to antimicrobials are most likely to create decreased susceptibility as observed in some of the isolates. These are only observations in susceptibility patterns exhibited by the M. ulcerans isolates used in this study. 5. Conclusions Azithromycin has been identified as an effective drug for controlling M. ulcerans isolates in Ghana. This study therefore elucidates the need to seriously and continuously conduct more susceptibility studies involving antimicrobials against isolates of M. ulcerans, particularly synergistic investigation of antimicrobials. This is against the backdrop that multidrug therapy tends to be more effective in antimycobacterial therapy. Acknowledgments The authors would like to thank all staff of the Animal Experimental Department of the Noguchi Memorial Institute for Medical Research in Ghana. Abbreviations BU:Buruli ulcer BUD:Buruli ulcer disease M. ulcerans: Mycobacterium ulcerans NMIMR:Noguchi Memorial Institute for Medical Research WHO:World Health Organization L-J:Lowenstein-Jensen Rif:Rifampicin Str:Streptomycin MTB: Mycobacterium tuberculosis TB:Tuberculosis MIC:Minimum inhibitory concentration BN:Benin GH:Ghana A:Amasaman N:Nsawam Az:Azithromycin Dap:Dapsone Amk:Amikacin Cip:Ciprofloxacin Ofl:Ofloxacin Clo:Clofazimine. Conflicts of Interest The authors declare that they have no conflicts of interest.

WHO:World Health Organization L-J:Lowenstein-Jensen Rif:Rifampicin Str:Streptomycin MTB: Mycobacterium tuberculosis TB:Tuberculosis MIC:Minimum inhibitory concentration BN:Benin GH:Ghana A:Amasaman N:Nsawam Az:Azithromycin Dap:Dapsone Amk:Amikacin Cip:Ciprofloxacin Ofl:Ofloxacin Clo:Clofazimine. Conflicts of Interest The authors declare that they have no conflicts of interest. Authors' Contributions Enid Owusu conceived and designed the experiments. Kwesi K. Addo and Phyllis Addo participated in collection and analyses of the data. Enid Owusu and Mercy J. Newman jointly drafted the manuscript. Enid Owusu, Kwesi K. Addo, and Phyllis Addo contributed to the interpretation of the data. Phyllis Addo and Mercy J. Newman jointly developed the structure and arguments for the manuscript. Kwesi K. Addo and Enid Owusu made critical revisions and approved final version. All authors read and approved the final manuscript. Table 1 Minimum inhibitory concentration (μg/mL) of the antimicrobials. Antimicrobials MIC of M. ulcerans isolates (in μg/mL)

Authors' Contributions Enid Owusu conceived and designed the experiments. Kwesi K. Addo and Phyllis Addo participated in collection and analyses of the data. Enid Owusu and Mercy J. Newman jointly drafted the manuscript. Enid Owusu, Kwesi K. Addo, and Phyllis Addo contributed to the interpretation of the data. Phyllis Addo and Mercy J. Newman jointly developed the structure and arguments for the manuscript. Kwesi K. Addo and Enid Owusu made critical revisions and approved final version. All authors read and approved the final manuscript. Table 1 Minimum inhibitory concentration (μg/mL) of the antimicrobials. Antimicrobials MIC of M. ulcerans isolates (in μg/mL) GH BN A5 A9 A13 A61 A68 A69 N1 N2 N5 N8 Az 0.16 0.63 0.63 0.16 0.16 0.31 0.31 0.16 0.63 0.63 0.31 0.63 Rif 1.25 0.63 0.63 0.16 0.16 1.25 0.63 0.63 1.25 0.63 1.25 1.25 Dap 0.31 0.63 1.25 1.25 0.63 1.25 1.25 0.63 0.31 1.25 1.25 1.25 Str 0.63 1.25 0.63 1.25 1.25 1.25 1.25 1.25 1.25 0.63 1.25 1.25 Amk 0.63 1.25 0.63 2.50 1.25 2.50 2.5 1.25 0.63 1.25 0.63 1.25 Cip 1.25 1.25 1.25 1.25 0.63 1.25 1.25 0.63 1.25 1.25 1.25 1.25 Ofl 1.25 0.63 1.25 1.25 0.63 1.25 2.50 2.50 1.25 2.5 1.25 2.50 Clo 2.50 2.50 2.50 1.25 1.25 1.25 2.50 2.50 2.50 2.50 2.50 2.50 BN: Benin, GH: Ghana, A: Amasaman, N: Nsawam, Az: azithromycin, Rif: rifampicin, Dap: dapsone, Str: streptomycin, Amk: amikacin, Cip: ciprofloxacin, Ofl: ofloxacin, and Clo: clofazimine. GH and BN are lab strains. Table 2 Mean, minimum, and maximum values of MIC (μg/mL).

GH BN A5 A9 A13 A61 A68 A69 N1 N2 N5 N8 Az 0.16 0.63 0.63 0.16 0.16 0.31 0.31 0.16 0.63 0.63 0.31 0.63 Rif 1.25 0.63 0.63 0.16 0.16 1.25 0.63 0.63 1.25 0.63 1.25 1.25 Dap 0.31 0.63 1.25 1.25 0.63 1.25 1.25 0.63 0.31 1.25 1.25 1.25 Str 0.63 1.25 0.63 1.25 1.25 1.25 1.25 1.25 1.25 0.63 1.25 1.25 Amk 0.63 1.25 0.63 2.50 1.25 2.50 2.5 1.25 0.63 1.25 0.63 1.25 Cip 1.25 1.25 1.25 1.25 0.63 1.25 1.25 0.63 1.25 1.25 1.25 1.25 Ofl 1.25 0.63 1.25 1.25 0.63 1.25 2.50 2.50 1.25 2.5 1.25 2.50 Clo 2.50 2.50 2.50 1.25 1.25 1.25 2.50 2.50 2.50 2.50 2.50 2.50 BN: Benin, GH: Ghana, A: Amasaman, N: Nsawam, Az: azithromycin, Rif: rifampicin, Dap: dapsone, Str: streptomycin, Amk: amikacin, Cip: ciprofloxacin, Ofl: ofloxacin, and Clo: clofazimine. GH and BN are lab strains. Table 2 Mean, minimum, and maximum values of MIC (μg/mL). Antimicrobials Minimum value of MIC Maximum value of MIC Mean MIC Az 0.16 0.63 0.39 Rif 0.16 1.25 0.81 Dap 0.31 1.25 0.94 Str 0.63 1.25 1.10 Amk 0.63 2.50 1.36 Cip 0.63 1.25 1.15 Ofl 0.63 2.50 1.26 Clo 1.25 2.50 2.19 MIC: minimum inhibitory concentration, Az: azithromycin, Rif: rifampicin, Dap: dapsone, Str: streptomycin, Amk: amikacin, Cip: ciprofloxacin, Ofl: ofloxacin, and Clo: clofazimine.

1. Background Malaria is a protozoal disease. It is a parasitic infection of red blood cells. In humans it is generally caused by five different species of Plasmodium, namely, P. vivax, P. falciparum, P. malariae, P. knowlesi, and P. ovale. According to an estimate about 40% of the world population lives in high malaria zone [1]. Pakistan is a malaria endemic country and it is the second most prevalent disease in Pakistan. Major causative agents of malaria in Pakistan are P. vivax and P. falciparum with P. vivax being more common [2]. Malaria caused by P. falciparum is more severe and may often lead to cerebral malaria and death especially in children. Initially malaria due to P. vivax was generally considered as milder and manageable compared to P. falciparum infection, but recent global reports suggest that P. vivax malaria may cause complications leading to death. The global mortality rate for P. vivax is documented as 0.1–1.6% [3]. Hence, beside P. falciparum the P. vivax malaria should also be closely monitored to avoid complications and mortality. Thus, timely diagnosis of malaria in endemic areas is vital for early treatment and prevention of fatal outcomes in cases of P. falciparum, P. vivax, or mixed P. falciparum and P. vivax malaria.

[3]. Hence, beside P. falciparum the P. vivax malaria should also be closely monitored to avoid complications and mortality. Thus, timely diagnosis of malaria in endemic areas is vital for early treatment and prevention of fatal outcomes in cases of P. falciparum, P. vivax, or mixed P. falciparum and P. vivax malaria. Due to the limitation of local health service resources, imprecise clinical diagnosis remains the basis of therapeutic care for the majority of febrile patients in remote malaria endemic areas of Pakistan, where laboratory diagnostics is often out of reach. Diagnosis based on clinical features alone has very low specificity and results in overtreatment [4] and emergence of drug resistant strains. In order to avoid this, the WHO recommends confirmed diagnosis of all malaria suspected cases before giving treatment [5, 6]. In Pakistan laboratory diagnosis of malaria is indispensable to avoid misdiagnosis as per national guidelines. For precise malaria diagnosis, several diagnostic approaches are employed in labs including microscopy, immune-florescence technique, immune-chromatographic testing (ICT), PCR, and use of hematological analyzers [7–10]. The microscopic detection of malarial parasite is generally considered as a gold standard in malaria diagnosis due to low cost and accessibility. Although cheap, specific, and sensitive this procedure requires an expert microscopist and may become unreliable, time-consuming, and laborious at low parasite densities of <1000 parasites/µl [11].

on of malarial parasite is generally considered as a gold standard in malaria diagnosis due to low cost and accessibility. Although cheap, specific, and sensitive this procedure requires an expert microscopist and may become unreliable, time-consuming, and laborious at low parasite densities of <1000 parasites/µl [11]. Malaria caused by P. falciparum may become complicated and fatal if misdiagnosed or left untreated. In order to detect low level parasitemia and also to detect mono- or coinfection of different parasite species immune-chromatographic/rapid diagnostic testing devices were developed. These devices based on parasite antigens or panspecific aldolases are simple and easy to use. The rapid diagnostic tests (RDTs) for malaria detection are usually based on principle of sandwich ELISA. These are inexpensive and easily available tests and require no prior training. Tests can be performed and results can be interpreted following manufacturer's instructions. Several reports on sensitivity and specificity of various commercial ICT devices are available [12, 13]. Devices like NOW Malaria® and ICT Malaria Combo can simultaneously detect histidine-rich protein 2 (HRP2) of P. falciparum and aldolase of all the Plasmodium species. Reports on poor specificity of these devices for few species like P. ovale led to developing devices based on detection of panspecific parasite lactate dehydrogenase (pLDH) enzyme [14]. OptiMal® is an ICT device of choice as it has a sensitivity of about 100% and is more specific 95% [15].

f all the Plasmodium species. Reports on poor specificity of these devices for few species like P. ovale led to developing devices based on detection of panspecific parasite lactate dehydrogenase (pLDH) enzyme [14]. OptiMal® is an ICT device of choice as it has a sensitivity of about 100% and is more specific 95% [15]. Automated malaria detection by hemoanalyzer is another approach to suspect malaria in febrile patients. Abnormalities in scattergrams of flow-cytometry-based hemoanalyzers like Sysmex XE-2100 and Cell Dyn have been reported as an aid in diagnosing malaria followed by microscopic confirmation [16]. Molecular diagnostic technique like PCR has an edge over the manual microscopy and serodiagnosis by RDTs. It is a reliable technique and can be used for malaria diagnosis. Beside genus specific PCR, species-specific multiplex and nested PCR have been developed for malarial parasite (MP) detection at a threshold of even 1 parasite/μl [17]. The PCR can be used as an internal quality control rather than being used as part of routine diagnosis as it is expensive and time-taking and needs trained individuals [18].

PCR, species-specific multiplex and nested PCR have been developed for malarial parasite (MP) detection at a threshold of even 1 parasite/μl [17]. The PCR can be used as an internal quality control rather than being used as part of routine diagnosis as it is expensive and time-taking and needs trained individuals [18]. According to the national statistical survey in 2007 malaria results in ~30,000 annual deaths in Pakistan [19]. The disease may be fatal especially in children and nonimmune individuals so high sensitivity of diagnosis in malaria endemic areas is particularly important. Misdiagnosis due to poor specificity of diagnostic modalities may be another issue leading to increased drug pressure causing antimalarial drug resistance. Aside from the vector control; the malaria-related morbidity and mortality may also be controlled by timely and accurate diagnosis of infection [20]. In this study performance of Sysmex XE-2100, ICT Malaria Combo, and First Response Malaria for early detection of MP was evaluated with microscopy and PCR as gold standard and internal quality control, respectively. 2. Materials and Methods This cross sectional study was conducted at National Institute of Blood Diseases and Bone Marrow Transplantation (NIBD), Karachi, Pakistan. The patients and controls were recruited after approval by ethical review committee of NIBD. The study protocol adhered to the tenets of the Declaration of Helsinki.

ls and Methods This cross sectional study was conducted at National Institute of Blood Diseases and Bone Marrow Transplantation (NIBD), Karachi, Pakistan. The patients and controls were recruited after approval by ethical review committee of NIBD. The study protocol adhered to the tenets of the Declaration of Helsinki. 2.1. Study Population Blood specimens (6000 µl) were collected in EDTA tubes from patients admitted at NIBD with clinical suspicion of malarial infection. Following national/WHO guidelines of precise diagnosis prior to treatment for malaria management 128 patients with clinical suspension of malaria were selected for this study. As a control group, sampling from 150 healthy individuals without clinical symptoms of malaria or any other infection or disease was also performed. The control group was confirmed as “malaria negative” by microscopy and RDT at the time of enrollment. The study was conducted over a period of about 9 months from October 2013 to July 2014. Signed informed consent and detailed questionnaire were obtained from the study population. 2.2. Laboratory Procedures The complete blood count (CBC) data from Sysmex XE-2100 was recorded following standard machine operating protocol. The data was then analyzed and compared with morphological data to set standard design for automated malaria diagnosis. Sensitivity and specificity of this machine were also evaluated.

boratory Procedures The complete blood count (CBC) data from Sysmex XE-2100 was recorded following standard machine operating protocol. The data was then analyzed and compared with morphological data to set standard design for automated malaria diagnosis. Sensitivity and specificity of this machine were also evaluated. The microscopic examination of Giemsa/Leishman stained thick and thin blood smear for malaria diagnosis is the gold standard. Thick smear can detect parasite even in low densities since high volume of infected specimen is screened while thin smear helps in species differentiation. For microscopy, both thin and thick smears were prepared immediately in duplicate to avoid any discrepancies in morphological detection of malarial parasites in blood. The smears were stained with 4% Giemsa's/Leishman's stain and observed according to WHO standard guidelines by three independent observers. The immunochromatographic testing was performed using two different RDTs, that is, ICT Malaria Combo and First Response Malaria, on fresh blood samples (not more than three hours old) as per supplier's instructions. Molecular detection of malarial infection based on polymerase chain reaction (PCR) was performed using previously designed primers by Padley et al. 2003. Parasite DNA was extracted from fresh EDTA containing blood using QIAamp DNA Mini Kit (Qiagen, USA, Cat. number 51306). The extracted DNA was amplified and species were identified. The recorded results by agarose gel electrophoresis were then used as quality control to countercheck the data obtained by other diagnostic tests.

e DNA was extracted from fresh EDTA containing blood using QIAamp DNA Mini Kit (Qiagen, USA, Cat. number 51306). The extracted DNA was amplified and species were identified. The recorded results by agarose gel electrophoresis were then used as quality control to countercheck the data obtained by other diagnostic tests. 2.3. Data Analysis The data was analyzed by SPSS version 17. The PL was calculated by multiplying number of asexual stages of parasite observed by microscopy with absolute RBCs count per 2500 RBCs. To assess sensitivity and specificity, results of microscopy, automated hemoanalyzer, and RDT were compared with PCR results. The sensitivity was calculated as the proportion of positive test results obtained among samples scored as containing malaria parasites by PCR; the specificity was the proportion of negative test results obtained among samples whose PCR results were negative. Positive and negative predictive values were also calculated as the proportion of true positive or true negative results among all samples scored as positive or negative by PCR, respectively. Youden's J-index was also calculated for comparative performance analysis of different tests.

s whose PCR results were negative. Positive and negative predictive values were also calculated as the proportion of true positive or true negative results among all samples scored as positive or negative by PCR, respectively. Youden's J-index was also calculated for comparative performance analysis of different tests. 3. Results During the present study about 126 samples were found to be parasite positive by microscopy. Of these 126 samples, about 91 (71.094%) were infected with Plasmodium vivax and 19 (14.844%) with P. falciparum while mixed infection of P. vivax and P. falciparum was observed in 18 (14.062%) samples. Parasite load/µl (PL) was also estimated by microscopy to evaluate the degree of severity of malaria. Only parasite load greater than 350 parasites/µl was observed; none of the patients had very low parasitemia (Table 1). The mean parasite load for P. vivax and P. falciparum was 14496/µl and 24410/µl, respectively. Like other diseases commercially prepared immune-chromatographic rapid diagnostic tests (RDTs) are available to detect malaria. The ICT Malaria Combo and First Response Malaria were compared for their efficacy. Among the two tested devices First Response Malaria seemed to be better with a sensitivity of 91.52% (95% CI: 87.52–95.52; Table 2). The positive predictive value of this device was 93.90% (95% CI: 91.10–96.70). The results are comparable with the gold standard microscopy. Based on their Youden's J-index of above 0.8 both the devices fall in category of very good diagnostic modalities.

ter with a sensitivity of 91.52% (95% CI: 87.52–95.52; Table 2). The positive predictive value of this device was 93.90% (95% CI: 91.10–96.70). The results are comparable with the gold standard microscopy. Based on their Youden's J-index of above 0.8 both the devices fall in category of very good diagnostic modalities. The Sysmex XE-2100 is generally used to record the routine hematological parameters as first-line screening test for any febrile patient. In the present study abnormal scattergrams on this analyzer were used for presumptive diagnosis of malaria. A total of about 126 cases were categorized as malaria suspected cases, on the basis of abnormal scattergrams in DIFF, WBC/Baso, IMI, and RET-EXT channels. Youden's J-index of automated hematological analyzer Sysmex XE-2100 for malaria detection was 0.98. The test seemed comparable with gold standard microscopy (Table 2). The pseudoeosinophilia and graying of neutrophil cluster and double neutrophil and eosinophil populations in DIFF channel were observed (Figure 1). In WBC/BASO channel, more than seven dots along the x-axis between first and third vertical marking were observed in case of P. vivax infection only. Increased signals (dots) in basophil region were also observable in approximately all cases of P. vivax, whereas there were no basophils in peripheral film. In case of malaria infection, multiple gray dots in the middle area were observed in the IMI channel despite the absence of immature granulocytes (myelo- and metamyelocyte) and any fluorescent signals above neutrophils in DIFF channel. Furthermore, the presence of gray dots along right side of the box extending vertically down and moving horizontally towards the y-axis in RET-EXT channel was indicative of P. falciparum infection (Figure 1). The abnormalities in DIFF channel were observed in 95% cases while the percentages of positive cases for WBC/Baso, IMI, and RET-EXT channels were 79%, 59%, and 83%, respectively. None of the samples from the control group showed abnormalities in any of these channels (Figure 2).

e of P. falciparum infection (Figure 1). The abnormalities in DIFF channel were observed in 95% cases while the percentages of positive cases for WBC/Baso, IMI, and RET-EXT channels were 79%, 59%, and 83%, respectively. None of the samples from the control group showed abnormalities in any of these channels (Figure 2). The species-specific PCR, being the internal quality control, was the most sensitive and specific test. It detected 127 malaria positive cases altogether (Table 2). None of the patients had infection caused by P. malariae and P. ovale. The product size for P. vivax and P. falciparum was 300 bp and 276 bp, respectively (Figure 3). The PCR also detected a case of P. vivax otherwise missed by microscopy. Furthermore, a case of P. vivax with low parasitemia (PL: 386/µl) was missed by PCR. It can thereby be suggested that even PCR may overlook infection with low levels of parasite in blood. This sample was found to be positive by microscopy on third reexamination by expert microscopist. On comparison the specificity and sensitivity of microscopy and automated hemoanalyzer were similar (Table 2). The species differentiation/identification by hemoanalyzer was not as obvious as with microscopy. So the results on hemoanalyzer may predict malaria but needs further confirmation by the gold standard microscopy. Hence, microscopy of thick and thin film remains the gold standard. Rapid diagnostic tests have acceptable sensitivity and specificity.

identification by hemoanalyzer was not as obvious as with microscopy. So the results on hemoanalyzer may predict malaria but needs further confirmation by the gold standard microscopy. Hence, microscopy of thick and thin film remains the gold standard. Rapid diagnostic tests have acceptable sensitivity and specificity. 4. Discussion The present study was designed to evaluate the utility in terms of efficiency of existing routine malaria diagnostic tests compared with gold standard microscopy and PCR as internal quality control. Microscopy being the gold standard was the only test producing quantitative results in the present study. The sensitivity of microscopy by thick smear is 5–10 parasites/µl. It is cheaper when compared with other methods. The only limitation is the risk of human error and thus observer's expertise is required [21]. In the present study all the slides were observed by three microscopists independently but a clinically unapparent case of P. vivax malaria was missed. This is in line with other studies where cases were either missed or misdiagnosed by microscopy [22].

isk of human error and thus observer's expertise is required [21]. In the present study all the slides were observed by three microscopists independently but a clinically unapparent case of P. vivax malaria was missed. This is in line with other studies where cases were either missed or misdiagnosed by microscopy [22]. Two commercially prepared rapid test devices, that is, ICT Malaria Combo and First Response Malaria, were compared for their efficiency with microscopy as gold standard. Both the test devices were accurate with the accuracy of 93.16% (95% CI: 90.16–96.16) and 93.81% (95% CI: 89.7–97.7), respectively. The First Response Malaria was found to be more efficient device with a sensitivity of 91.52% (95% CI: 87.52–95.52). The sensitivity is comparable with an earlier study by Bharti et al. [23] on First Response Malaria (sensitivity: 93%). On the other hand the sensitivity for ICT Malaria Combo was found to be lower (90.83%; 95% CI: 86.83–94.83) than the reported value of 95.7% by Grobusch et al. [9] (Table 2).

2% (95% CI: 87.52–95.52). The sensitivity is comparable with an earlier study by Bharti et al. [23] on First Response Malaria (sensitivity: 93%). On the other hand the sensitivity for ICT Malaria Combo was found to be lower (90.83%; 95% CI: 86.83–94.83) than the reported value of 95.7% by Grobusch et al. [9] (Table 2). Recent research is directed to detect malaria on the basis of the abnormal scattergrams of flow-cytometry-based automated hemoanalyzers [16, 24, 25]. First few reports were from Cell Dyn and Sysmex XE-2100. Extensive studies were done by Korean scientists explaining the utility of abnormal DIFF scattergrams in detecting malaria on Sysmex XE-2100. Huh et al. [26] reported unclassified spots extending from neutrophils towards eosinophil area, two eosinophil populations, two neutrophil populations, and overlapping of neutrophil and eosinophil populations as the most common abnormalities observed in WBC scattergrams on Sysmex XE-2100. The abnormalities may be caused by hemozoin containing particles interfering with the machine's WBC detection system resulting in abnormal counting of hemozoin containing neutrophils as well as due to their detection as eosinophils near the neutrophil cluster. Yoo et al., in 2010 [27], found 15.7% cases with abnormal WBC scattergram like two neutrophil and two eosinophil populations in assessment of 413 malaria cases.

etection system resulting in abnormal counting of hemozoin containing neutrophils as well as due to their detection as eosinophils near the neutrophil cluster. Yoo et al., in 2010 [27], found 15.7% cases with abnormal WBC scattergram like two neutrophil and two eosinophil populations in assessment of 413 malaria cases. About 95% malaria cases had abnormal WBC scattergram during the present study. Most common abnormality was found to be graying of eosinophil and neutrophil populations (53.12%). Other common abnormalities were overlapping of eosinophil and neutrophil populations (20.30%) and two eosinophil populations (32.80%). A rightward shift of RBC ghost in WBC/BASO and DIFF scattergrams was also very commonly found in most malaria positive cases. This can be attributed to the presence of extracellular pigment and RBC lysis which are reflected in that area [28]. Pseudoeosinophilia by machine compared with manual differential count was observed in 6.25% cases which is comparatively lower than previous report of 39% of cases of pseudoeosinophilia [16, 29].

cases. This can be attributed to the presence of extracellular pigment and RBC lysis which are reflected in that area [28]. Pseudoeosinophilia by machine compared with manual differential count was observed in 6.25% cases which is comparatively lower than previous report of 39% of cases of pseudoeosinophilia [16, 29]. The nucleic acid based detection of malarial parasites by PCR is a more sensitive and specific approach than the gold standard microscopy. During this study the species-specific qualitative PCR detected P. vivax in one sample which was MP negative by microscopy. Additionally, this particular sample was also negative by both RDTs while clear abnormal signals suggesting presence of MP were observed in scattergrams on Sysmex XE-2100. This observation is in line with another study conducted in Pakistan where real-time PCR detected 3 samples missed by microscopy [30]. Thus, we recommend use of PCR for accurate diagnosis of malaria in public reference centers involved in WHO guided malaria control program in Pakistan. Coleman and colleagues conducted a detailed surveillance study in Thailand on comparison of PCR and microscopy for the detection of asymptomatic malaria in P. falciparum/vivax endemic area [31]. They suggested that PCR is a more precise and reproducible test for the MP species identification and detection but its performance decreased markedly at low parasite densities, that is, <500/µl. The influence of low PL on performance of PCR was also recorded during the present study where MP positive case of P. vivax infection (PL: 386/µl) was found to be false negative (Table 2).

for the MP species identification and detection but its performance decreased markedly at low parasite densities, that is, <500/µl. The influence of low PL on performance of PCR was also recorded during the present study where MP positive case of P. vivax infection (PL: 386/µl) was found to be false negative (Table 2). Our study was time bound (9 months). Observer's training is required for identification of abnormal signals in scattergram. It must also be borne in mind that abnormal scattergrams may be observed in other conditions like dengue, basophilic stippling, thalassemia, and chronic myeloid leukemia. Observed scattergram abnormalities can only depict the presence of MP and cannot be used to differentiate between species of Plasmodia.

scattergram. It must also be borne in mind that abnormal scattergrams may be observed in other conditions like dengue, basophilic stippling, thalassemia, and chronic myeloid leukemia. Observed scattergram abnormalities can only depict the presence of MP and cannot be used to differentiate between species of Plasmodia. 5. Conclusion Thus, after comparison it can be concluded that the microscopy of thick and thin films remains the gold standard for malaria diagnosis despite chances of human error. The microscopy may be confirmed with PCR since the specificity and sensitivity for PCR are the highest. In remote endemic areas where microscopy due to absence of expert microscopists seems impossible, automated hemoanalyzers can serve as a useful adjunct to timely clinical diagnosis of malaria. The positive signals on hemoanalyzers need further confirmation by microscopy and PCR up to species level at the nearest reference lab to avoid unnecessary treatment, leading to development of drug resistant strains of MP. The comparative high cost of PCR limits its applicability in most diagnostic labs in developing countries, and we hereby recommend it to be added as a mandatory confirmatory test at least at all national reference labs. Acknowledgments This work was funded by Pakistan Medical Research Council under Project reference no. 4-22-10/11/RDC/NIBD, Karachi. Authors would like to acknowledge Iffat Shamim and Danish Zahid for their technical assistance. Abbreviations MP:Malarial parasite PL:Parasite load CBC:Complete blood count RDTs:Rapid diagnostic tests HRP2:Histidine-rich protein 2.

Acknowledgments This work was funded by Pakistan Medical Research Council under Project reference no. 4-22-10/11/RDC/NIBD, Karachi. Authors would like to acknowledge Iffat Shamim and Danish Zahid for their technical assistance. Abbreviations MP:Malarial parasite PL:Parasite load CBC:Complete blood count RDTs:Rapid diagnostic tests HRP2:Histidine-rich protein 2. Conflicts of Interest The authors have no conflicts of interest to declare. Authors' Contributions Samina Naz Mukry designed the project, secured funding, and supervised this project along with finalizing the results, statistical analysis, and manuscript drafting. Madiha Saud carried out the CBC on Sysmex XE-2100; performed RDTs based malarial parasite screening and confirmation by gold standard microscopy; and was involved in data management. Gul Sufaida optimized and performed species-specific PCR. Kashif Shaikh provided expert support for microscopy and was involved in manuscript preparation. Arshi Naz and Tahir Sultan Shamsi reviewed the manuscript.

rial parasite screening and confirmation by gold standard microscopy; and was involved in data management. Gul Sufaida optimized and performed species-specific PCR. Kashif Shaikh provided expert support for microscopy and was involved in manuscript preparation. Arshi Naz and Tahir Sultan Shamsi reviewed the manuscript. Figure 1 Parasitemia (P. falciparum and P. vivax) associated abnormalities in Sysmex XE2100 channels compared with normal sample at the top. (a) DIFF channel: the blue arrow indicates abnormal events depicting pseudoeosinophilia and graying of neutrophil cluster and double neutrophil and eosinophil populations; (b) WBC/BASO channel: yellow box shows parasite containing ghost RBCs; (c) IMI channel: comparatively more immature cells in area below yellow line in malaria positive cases; and (d) RET- EXT channel: presence of cluster of abnormal cells in any of the sectors extending vertically down and moving horizontally towards the y-axis. Figure 2 Abnormal scattergram in Sysmex XE2100 of malaria positive cases. Figure 3 Agarose gel electrophoresis of amplified product obtained by 16S rRNA PCR using Plasmodium species-specific primers (Padley et al.). Lane 1: P. vivax; Lane 2: Mixed P. vivax + P. falciparum; Lane 3: P. falciparum; Lane 4: negative; Lane 5: P. falciparum; Lane 6: P. vivax; Lane 7: 100 bp ladder. Table 1 Parasite load/µl as estimated by microscopy. Parasites/µl P. falciparum P. vivax <500 1 1 >500 3 25 >5000 23 67 >50,000 2 5 Table 2 Performance analysis of different tests with species-specific PCR as internal control.

Figure 3 Agarose gel electrophoresis of amplified product obtained by 16S rRNA PCR using Plasmodium species-specific primers (Padley et al.). Lane 1: P. vivax; Lane 2: Mixed P. vivax + P. falciparum; Lane 3: P. falciparum; Lane 4: negative; Lane 5: P. falciparum; Lane 6: P. vivax; Lane 7: 100 bp ladder. Table 1 Parasite load/µl as estimated by microscopy. Parasites/µl P. falciparum P. vivax <500 1 1 >500 3 25 >5000 23 67 >50,000 2 5 Table 2 Performance analysis of different tests with species-specific PCR as internal control. Variables ICT Malaria Combo First Response Malaria Microscopy Automated hemoanalyzer (Sysmex XE-2100) PCR True positive 109 108 126 126 127 True negative 150 150 150 150 150 False positive 8 7 0 0 0 False negative 11 10 2 2 1 Negative Predictive Value (95% CI) 93.16 (90.16–96.16) 93.75 (90.85–96.65) 98.68 (96.31–99.60) 98.68 (96.31–99.60) 99.33 (99.24–99.42) Positive predictive value (95% CI) 93.16 (90.16–96.16) 93.90 (91.10–96.70) 99.20 (98.20–100.20) 99.20 (98.20–100.20) 100 Sensitivity (95% CI) 90.83 (86.83–94.83) 91.52 (87.52–95.52) 98.41 (96.40–100.40) 98.41 (96.40–100.40) 99.21 (98.20–100.20) Specificity (95% CI) 94.90 (92.30–97.50) 95.54 (93.14–97.94) 100 100 100 Accuracy (95% CI) 93.16 (90.16–96.16) 92.80 (89.7–97.7) 99.28 (97.00–101.00) 99.28 (97.00–101.00) 99.64 (99.57–99.71) Youden's J- index 0.86 0.87 0.98 0.98 0.99

1. Introduction Bloodstream infections (BSIs) are defined as the presence of viable infectious microorganism in the bloodstream causing clinical illness [1]. They are among the leading causes of mortality and morbidity worldwide [2]. The term bloodstream infection and bacteremia are synonymously used which generally refer to the significant growth of a microorganism in a blood culture obtained from the patient with clinical signs of infection [3]. In clinical practice, bacteremia may range from self-limiting infections to life threatening septicemia that requires prompt and rational antimicrobial treatment [4]. However, in the developing countries, changing epidemiology, lack of standard antimicrobial guidelines in locality, emergence of antimicrobial resistance, and paucity of good diagnostic facilities are major denominators for surge in BSI associated morbidity and mortality [5].

t and rational antimicrobial treatment [4]. However, in the developing countries, changing epidemiology, lack of standard antimicrobial guidelines in locality, emergence of antimicrobial resistance, and paucity of good diagnostic facilities are major denominators for surge in BSI associated morbidity and mortality [5]. Alongside, incidence rates of BSI have been found to be bimodal. Increased rates have been observed in extreme ages of life due to poor immune competency as well as the presence of comorbid conditions [6, 7]. Differences in the diagnostic approaches, antimicrobial therapy, and clinical management of BSI among pediatric and adult patients have been described elsewhere [5, 8]. Furthermore, treatment of bloodstream infections in developing world is often empirical, primarily due to the lack of standard therapeutic guidelines and unavailability of susceptibility pattern of the local isolates [3]. Over the years, there has been dramatic shift in the etiology of bloodstream infections with gram negative bacterial dominance. These agents are continuously evolving with novel drug resistant determinants resulting in the poor therapeutic outcomes in BSI [2]. Reports regarding higher prevalence of gram negative isolates causing BSI in South Asian region producing extended spectrum beta-lactamases (ESBL) and carbapenemases are of great concern, as it has major impact on selecting and prescribing the antimicrobial therapy [9–11]. In the relevance of the global studies in the trends of bloodstream infections and increasing antimicrobial resistance, there is a strong need of evaluation of such trends in Nepalese scenario. Therefore, a systematic study was carried out among the pediatric and adult patients to investigate the etiology and trends of bacterial pathogens as well as the role of drug resistant isolates in these infections.

asing antimicrobial resistance, there is a strong need of evaluation of such trends in Nepalese scenario. Therefore, a systematic study was carried out among the pediatric and adult patients to investigate the etiology and trends of bacterial pathogens as well as the role of drug resistant isolates in these infections. 2. Materials and Methods 2.1. Study Setup This was a hospital based cross sectional study carried out between March 2015 and August 2016 (over a period of 18 months) at the Department of Microbiology, Manmohan Memorial Teaching Hospital, a tertiary care referral center with 300 patient beds, in Kathmandu, Nepal. During the study, a total of 3,088 patients (pediatric and adults), clinically suspected of bloodstream infections (BSIs), were enrolled. Patients already on antibiotics and repeated samples from the same patients were excluded.

ching Hospital, a tertiary care referral center with 300 patient beds, in Kathmandu, Nepal. During the study, a total of 3,088 patients (pediatric and adults), clinically suspected of bloodstream infections (BSIs), were enrolled. Patients already on antibiotics and repeated samples from the same patients were excluded. 2.2. Laboratory Investigations Patients visiting outpatient departments (pediatric and general medicine) and those admitted in the inpatient units were investigated for bloodstream infections by respective unit physicians. At the onset of fever (>37°C) or in the presence of any clinical symptoms compatible with infection, a blood culture specimen was taken with aseptic technique by cleansing of the collection site with 70% alcohol and subsequently followed by povidone iodine. One mL (for neonates), 5 mL (for children), and 10 mL (for adults) of blood specimen were collected and they were inoculated into brain heart infusion (BHI) broth at the blood to broth ratio of 1 : 10. After incubation, at 37°C for 24, 48, and 72 hours, blind subcultures were made on MacConkey agar and blood agar plates (HiMedia Laboratories, India). The plates were observed for bacterial growth after 24 hrs of aerobic incubation at 37°C. Identification of significant isolates was done by using standard microbiological techniques which involved morphological appearance of the colony, gram's staining reactions, catalase test, coagulase test, and oxidase test with other biochemical and serological properties [12]. Samples were considered sterile, if no growth was observed on subculture after 7 days of aerobic incubation at 37°C. Laboratory confirmed BSI was considered when a bacterial pathogen was recovered from at least one blood culture with promising clinical symptoms.

t with other biochemical and serological properties [12]. Samples were considered sterile, if no growth was observed on subculture after 7 days of aerobic incubation at 37°C. Laboratory confirmed BSI was considered when a bacterial pathogen was recovered from at least one blood culture with promising clinical symptoms. 2.3. Antimicrobial Susceptibility Testing The susceptibility of bacterial isolates against different antibiotics was tested by the disk diffusion method [modified Kirby-Bauer method] on Mueller Hinton agar (HiMedia Laboratories, India) following standard procedures recommended by the Clinical and Laboratory Standards Institute (CLSI), Wayne, USA [13]. Antibiotics that were tested in this study include ampicillin (Amp 10 μg), amikacin (30 μg), gentamycin (Gen 10 μg), ciprofloxacin (CIP 5 μg), levofloxacin (5 μg), trimethoprim-sulfamethoxazole/cotrimoxazole (COT30 μg), cloxacillin (5 μg), cefixime (CFM 5 μg), cefotaxime/ceftriaxone (CTX/CTR 30 μg), ceftazidime (CAZ 30 μg), chloramphenicol (C 30 μg), azithromycin (AZM 15 μg), piperacillin-tazobactam (PIT 100/10 μg), imipenem (IPM 10 μg), meropenem (MRP 10 μg), teicoplanin (30 μg), and polymyxin B (PB 300 U) from HiMedia Laboratories, India. Interpretations of antibiotic susceptibility results were made according to the guidelines of interpretative zone diameters of CLSI [13]. Escherichia coli ATCC 25922, Staphylococcus aureus ATCC 25923, and Pseudomonas aeruginosa ATCC 27853 were used as the control organisms for antibiotic sensitivity.

boratories, India. Interpretations of antibiotic susceptibility results were made according to the guidelines of interpretative zone diameters of CLSI [13]. Escherichia coli ATCC 25922, Staphylococcus aureus ATCC 25923, and Pseudomonas aeruginosa ATCC 27853 were used as the control organisms for antibiotic sensitivity. 2.4. Identification of Multidrug Resistant (MDR) Isolates Multidrug resistant (MDR) bacterial isolates were identified according to the criteria recommended by joint committee of international experts from European Centre for Disease Prevention and Control (ECDC) and the Centers for Disease Control and Prevention (CDC) [14]. In this study, the isolate resistant to at least one antimicrobial from three different groups of first-line drugs tested was regarded as multidrug resistant (MDR).

ttee of international experts from European Centre for Disease Prevention and Control (ECDC) and the Centers for Disease Control and Prevention (CDC) [14]. In this study, the isolate resistant to at least one antimicrobial from three different groups of first-line drugs tested was regarded as multidrug resistant (MDR). 2.5. Phenotypic Detection of Extended Spectrum β-Lactamase (ESBL) The initial screening test for the ESBL production was performed by using ceftriaxone (CRO 30 μg), ceftazidime (CAZ 30 μg), and cefotaxime (CTX 30 μg) disks (HiMedia, Mumbai, India). If the zone of inhibition (ZOI) was ≤25 mm for CRO, ≤22 mm for CAZ, and/or ≤27 mm for CTX, the isolate was considered a presumptive ESBL producer as recommended by CLSI [13]. Combination disk test (CDT) was used for the phenotypic confirmation of potential ESBL producing strains in which ceftazidime (CAZ) and cefotaxime (CTX) (30 μg each) alone and in combination with clavulanic acid (CA) (10 μg) was used. An increase in zone of inhibition of more than or equal to 5 mm for antimicrobial agent in combination with CA versus its zone when tested alone was considered as ESBL producer [13]. For ESBL standardization, Escherichia coli ATCC 25922 and Klebsiella pneumoniae ATCC 700603 were used as negative and positive controls.

An increase in zone of inhibition of more than or equal to 5 mm for antimicrobial agent in combination with CA versus its zone when tested alone was considered as ESBL producer [13]. For ESBL standardization, Escherichia coli ATCC 25922 and Klebsiella pneumoniae ATCC 700603 were used as negative and positive controls. 2.6. Phenotypic Test for Metallo-β-Lactamase (MBL) Isolates that were found nonsusceptible to third-generation cephalosporins (ceftazidime), imipenem, or meropenem in Kirby-Bauer disk diffusion method were presumptively considered MBL producers and confirmed by the combined disk method. Briefly, the test inoculums (comparable to 0.5 McFarland standards) were prepared and transferred onto the Mueller Hinton agar plates. In the combination disk test for MBL, two imipenem (IPM) disks (10 μg), one containing 10 μL of 0.1 M (292 μg) anhydrous EDTA (Sigma Chemicals, St. Louis, MO), were placed 25 mm apart. An increase in the zone size of more than or equal to 7 mm for imipenem-EDTA disk compared to imipenem disk alone indicated MBL producer strain as described by Yong et al. [15].

imipenem (IPM) disks (10 μg), one containing 10 μL of 0.1 M (292 μg) anhydrous EDTA (Sigma Chemicals, St. Louis, MO), were placed 25 mm apart. An increase in the zone size of more than or equal to 7 mm for imipenem-EDTA disk compared to imipenem disk alone indicated MBL producer strain as described by Yong et al. [15]. 2.7. Phenotypic Test for Methicillin Resistant (MRSA) and Inducible Clindamycin Resistant (iMLSB) Staphylococcus aureus Methicillin resistant Staphylococcus aureus (MRSA) isolates were detected by cefoxitin disk (30 μg) method of CLSI. S. aureus isolates were deemed methicillin resistant when the ZOI for cefoxitin was ≤21 mm [13]. Similarly, inducible macrolide-lincosamide-streptogramin-B (iMLSB) resistance was detected in S. aureus by disk approximation using clindamycin (2 μg) and erythromycin (15 μg) on MHA plates. After overnight incubation, isolates with flattened zone of inhibition adjacent to the erythromycin disk (referred to as a “D” zone) were considered to exhibit inducible clindamycin resistance [13]. 2.8. Data Analysis Patient information regarding patient name, age, sex, ward/bed number (if admitted), brief clinical history, duration of hospital stay, history of antibiotic use, and bacterial isolates and their antimicrobial susceptibilities was taken and entered into a computer program. Data analysis was carried out using the Statistical Package for Social Sciences [SPSS™] version 20.0 [IBM, Armonk, NY, USA] and presented in percentage base distribution. Data with p value of less than 0.05 (CI-95%) was regarded as significant.

timicrobial susceptibilities was taken and entered into a computer program. Data analysis was carried out using the Statistical Package for Social Sciences [SPSS™] version 20.0 [IBM, Armonk, NY, USA] and presented in percentage base distribution. Data with p value of less than 0.05 (CI-95%) was regarded as significant. 3. Results Overall, a total of 3,088 blood cultures specimens from patients suspected with bloodstream infections were processed. Specimens were from infants (n = 386), children (n = 574), adults (n = 1949), and the elderly (n = 179). Male comprised 50.6% of total patients with male to female ratio of 1.02. Out of 3,088 blood cultures during the period, 231 (7.48%) were positive for significant growth of bacterial pathogen suggesting bloodstream infection. 3.1. Trends of BSI among Pediatric and Adult Patients More pediatric patients (9.3%) were found with BSI as compared to the adult patients (6.6%) and this trend was statistically significant (p = 0.008, CI-99%). Among all cases, the highest proportion of culture confirmed bloodstream infections were observed among the infants (n = 50, 12.9%) followed by the children (n = 40, 6.9%), adults (n = 131, 6.72%), and elderly (n = 10, 5.5%) patients. (Table 1). In this study, the rate of blood stream infections varied significantly (p value = 0.00) between inpatient (22.1%) and outpatient (7.5%) groups. Furthermore, among various inpatient wards, higher blood culture positivity was observed among critical care patients (pediatric ICU: 26.08% and adult: 27.2%) (data not presented). 3.2. Trends of Bacterial Isolates

In this study, the rate of blood stream infections varied significantly (p value = 0.00) between inpatient (22.1%) and outpatient (7.5%) groups. Furthermore, among various inpatient wards, higher blood culture positivity was observed among critical care patients (pediatric ICU: 26.08% and adult: 27.2%) (data not presented). 3.2. Trends of Bacterial Isolates Table 2 illustrates the common bacterial isolates causing bloodstream infections. Gram negative bacteria (65.8%) were the leading pathogenic agents compared to the gram positive bacteria (34.2%) in this study (p < 0.05). Also, there was significant difference between the bacterial etiology of BSI among pediatric and adult patients. (Table 1). Staphylococcus aureus was the leading pathogen involved in pediatric cases (n = 41, 45.6%), while Salmonella enterica (n = 40, 28.3%) was the common pathogen involved in adult cases of BSI. Furthermore, Staphylococcus aureus was isolated among 45 (31%) of 145 blood culture isolates from inpatients which was followed by Pseudomonas (n = 39) and Acinetobacter species (n = 30). Among 86 blood culture isolates from outpatients, Salmonella enterica was isolated from 46 (53.5%) cases, followed by Staphylococcus aureus (n = 21) and Escherichia coli (n = 9).

d among 45 (31%) of 145 blood culture isolates from inpatients which was followed by Pseudomonas (n = 39) and Acinetobacter species (n = 30). Among 86 blood culture isolates from outpatients, Salmonella enterica was isolated from 46 (53.5%) cases, followed by Staphylococcus aureus (n = 21) and Escherichia coli (n = 9). 3.3. Trends of Antimicrobial Susceptibilities of Gram Negative Isolates (excluding Salmonella) In this study, susceptibilities of beta-lactam antibiotics, fluoroquinolones, aminoglycosides, and carbapenems towards gram negative isolates from pediatric patients and adult patients were poles apart. Escherichia coli isolates from pediatric patients were completely susceptible (100%) to ciprofloxacin, levofloxacin, amikacin, imipenem, and meropenem, but those from adult patients were resistant to ciprofloxacin and levofloxacin (83.3% each), ampicillin (66.7%), cefixime and ceftazidime (66.7% each), gentamycin and amikacin (33.3% each), and imipenem (16.7%), respectively. Similarly, isolates of Klebsiella spp. from pediatric patients were highly susceptible (100%) to ceftazidime, amikacin, and carbapenems while those from adult patients were highly resistant (75%) to cephalosporins and fluoroquinolones, 50% resistant to carbapenems, piperacillin-tazobactam, and chloramphenicol, respectively. Similarly, susceptibilities of Pseudomonas spp. against ciprofloxacin (25.0% versus 96.7%), levofloxacin (12.5% versus 87%), imipenem (0% versus 22.5%), and chloramphenicol (0% versus 12.9%) also differed considerably among pediatric and adults patients. However, susceptibilities of Acinetobacter spp. were also found in the similar trend (Table 3).

Pseudomonas spp. against ciprofloxacin (25.0% versus 96.7%), levofloxacin (12.5% versus 87%), imipenem (0% versus 22.5%), and chloramphenicol (0% versus 12.9%) also differed considerably among pediatric and adults patients. However, susceptibilities of Acinetobacter spp. were also found in the similar trend (Table 3). 3.4. Trends of Susceptibilities of Salmonella enterica Isolates Salmonella enterica isolates from pediatric patients were highly susceptible to ampicillin whereas 11.7% isolates from adult patients were found resistant to it. In addition, entire isolates were susceptible to cotrimoxazole, cefixime, ceftriaxone, chloramphenicol, and azithromycin in both groups (Table 4). However, more than 70% of the Salmonella Typhi and almost all (up to 100%) Salmonella Paratyphi A isolates were resistant to fluoroquinolones. 3.5. Trends of Susceptibilities of Gram Positive Isolates

Salmonella enterica isolates from pediatric patients were highly susceptible to ampicillin whereas 11.7% isolates from adult patients were found resistant to it. In addition, entire isolates were susceptible to cotrimoxazole, cefixime, ceftriaxone, chloramphenicol, and azithromycin in both groups (Table 4). However, more than 70% of the Salmonella Typhi and almost all (up to 100%) Salmonella Paratyphi A isolates were resistant to fluoroquinolones. 3.5. Trends of Susceptibilities of Gram Positive Isolates Staphylococcus aureus revealed high level of resistance among tested antimicrobials (Table 5). Isolates from pediatric patients were highly resistant to ampicillin (80%), erythromycin (51.1%), and clindamycin (51.1%) and less susceptible to cephalexin (37.7%), cloxacillin (37.7%), and imipenem (37.7%), respectively. Similarly, isolates from adult patients were resistant to ampicillin (100%), cotrimoxazole (84.0%), ciprofloxacin (88.0%), erythromycin (84.0%), clindamycin (80.0%), and levofloxacin (68.0%). Teicoplanin, amikacin, and cloxacillin were effective antimicrobials for both groups of patients with staphylococcal BSI. Furthermore, entire isolates (100%) of Enterococcus spp. were susceptible to teicoplanin and about 50% susceptible to amikacin, ciprofloxacin, and levofloxacin whereas only 20% and 75% were susceptible to ampicillin (in adult and pediatric patients, resp.).

icrobials for both groups of patients with staphylococcal BSI. Furthermore, entire isolates (100%) of Enterococcus spp. were susceptible to teicoplanin and about 50% susceptible to amikacin, ciprofloxacin, and levofloxacin whereas only 20% and 75% were susceptible to ampicillin (in adult and pediatric patients, resp.). 3.6. Trends of Resistance Determinants in Gram Negative Isolates In this study, high proportions of the gram negative isolate were multidrug resistant (34.8%) (Table 6). About two-thirds (71.4%) of Klebsiella spp. and half (50%) of the Escherichia coli were multidrug resistant (MDR). Similarly, among gram negative nonfermenters, 73.3% of Acinetobacter spp. and 41% of Pseudomonas spp. were MDR. About 42.8% of Klebsiella spp., 30% of E. coli, 16.7% of Acinetobacter, and 12.8% of Pseudomonas spp. were ESBL producers. In addition to this, 10% of E. coli, 28.5% of Klebsiella spp., and 6.6% of Acinetobacter spp. were MBL producers. The proportion of MDR isolates was significantly varied in the pediatric (13.3%) and adult (29.0%) patients (p < 0.005, CI-95%) (Table 7). 3.7. Trends of Resistance Determinants in Gram Positive Isolates Drug resistance was also common among gram positive isolates. About 60% of S. aureus and 44.4% of Enterococcus spp. were multidrug resistant. Methicillin resistance and inducible clindamycin resistance (iMLSB) were observed in 31.4% and 10% of Staphylococcus aureus isolates, respectively. However, incidence of multidrug resistant gram positive bacteria causing BSI among pediatric and adult patients was not significantly different (Table 8).

tidrug resistant. Methicillin resistance and inducible clindamycin resistance (iMLSB) were observed in 31.4% and 10% of Staphylococcus aureus isolates, respectively. However, incidence of multidrug resistant gram positive bacteria causing BSI among pediatric and adult patients was not significantly different (Table 8). 4. Discussion Bloodstream infections remained a challenge for the infectious disease physicians due to the changing bacterial etiology and emergence of antimicrobial resistance. Early detection of causative organism and determination of its antimicrobial susceptibility profile are necessary to help clinicians decide appropriate empirical therapy, which ultimately decreases the emergence of resistance [16]. Our study evaluates the incidences of bloodstream infections, bacterial etiology, and antimicrobial susceptibilities among the pediatric and adult group of patients and confirms that there is significant variation among these parameters within these groups of patient. Our data emphasizes the dominance of antibiotic resistant bacteria causing BSIs in both groups of patients and this trend is ever increasing.

microbial susceptibilities among the pediatric and adult group of patients and confirms that there is significant variation among these parameters within these groups of patient. Our data emphasizes the dominance of antibiotic resistant bacteria causing BSIs in both groups of patients and this trend is ever increasing. In this study, overall incidence of bloodstream infection based on significant bacterial growth in the blood cultures obtained from suspected patients was 7.48%. Comparatively, similar rates of BSI have been reported from the studies of Sharma et al. (6.9%), Pandey et al. (12.6%), and Shrestha et al. (13.3%) from nearby hospitals in Kathmandu, Nepal [17–19]. Similar rates of BSI were also documented in other studies of this region, particularly by Easow et al. (10.2%) from Pokhara, Singh et al. (10.16%) and Gupta and Kashyap (16.5%) from India, and Qureshi and Aziz (16.6%) from Pakistan [10, 11, 20, 21]. Our findings of BSI are lower when compared to the reported rates by Amatya et al. (23.1%), Alam et al. (20.9%), Arora and Devi (20.02%), and Fayyaz et al. (20.0%), respectively [22–25]. The variation in the BSI rates among these studies may be attributable to sampling volume of blood culture, culture system, and medium formulation as well as type of patients enrolled in the study. Furthermore, lower rates of BSI may be ascribed to the injudicious use of antibiotics not only by clinicians before referring to the tertiary care center but also by patients themselves.

utable to sampling volume of blood culture, culture system, and medium formulation as well as type of patients enrolled in the study. Furthermore, lower rates of BSI may be ascribed to the injudicious use of antibiotics not only by clinicians before referring to the tertiary care center but also by patients themselves. Bloodstream infections varied significantly within age groups, where the highest prevalence was recorded among patients at the lower extreme of ages: infants (12.9%), children (6.9%), adults (6.72%), and elderly (5.5%). Pradhan et al. also reported the similar rates of BSI in pediatric population in their study [26]. In this study, more pediatric patients (9.3%) were found with BSI as compared to adult patients (6.6%) and this difference was statistically significant (p < 0.05). Unfortunately, we could not evaluate these trends with other previous studies due to unavailability of published literature from Nepal comparing these trends. Furthermore, disparity between the BSI rates among various age groups could not be justified because the risk factors associated with BSI were not evaluated. However, it is generally accepted that infants are more likely to have BSI compared with patients of higher ages. Infants have poor skin integrity with developing immune system and they frequently participate in activities that predispose them to external environment making them prone to bacterial infections [5].

ated. However, it is generally accepted that infants are more likely to have BSI compared with patients of higher ages. Infants have poor skin integrity with developing immune system and they frequently participate in activities that predispose them to external environment making them prone to bacterial infections [5]. Gram negative bacteria (65.8%) were found to be the predominant cause of bloodstream infection in this study, but the majority of bacterial findings were Staphylococcus aureus. Gram positive bacteria were significantly associated with pediatric BSI, while gram negative isolates were significant in adult BSI cases (p < 0.05). Staphylococcus aureus (n = 41, 45.6%) was the most common among pediatric patients while Salmonella enterica (n = 40, 28.3%) was the most common among adult patients (p < 0.05). This finding is consistent with the previous findings of Pradhan et al. [26], Amatya et al. [22], Sharma et al. [17], and Easow et al. [20]. However, Shrestha et al. [19] from Nepal, Arora and Devi from India [24], and Moyo et al. from Tanzania [27] reported the gram positive bacterial dominance in blood stream infections. Over the time, etiology of BSI is continuously changing and BSI in our study was largely due to gram negative bacteria, but significant contribution from gram positive isolates, mostly S. aureus, was remarkable. We believe that the variation might be due to difference in geographical location, nature of patient population, endemicity of the etiological agents, and seasonal variation.

our study was largely due to gram negative bacteria, but significant contribution from gram positive isolates, mostly S. aureus, was remarkable. We believe that the variation might be due to difference in geographical location, nature of patient population, endemicity of the etiological agents, and seasonal variation. Salmonella enterica associated bloodstream infections were the most common gram negative bacteremic cases in our study, which suggests enteric fever is still a major problem in our region [17, 18, 28]. This trend is also consistent with previous studies from Nepal [18, 20], India [10], Pakistan [21], and Africa [5]. In our study, all the cases of enteric fever were community acquired. We isolated Salmonella spp. from 21.2% of the patients with BSIs, while a previous study from western Nepal reported Salmonella associated blood culture positivity in 51.6% of suspected cases [17]. Notably, more (55.1%) S. Paratyphi were responsible for enteric fever cases than S. Typhi in this study which may have future implication on the vaccine strategies, as the polyvalent vaccines against typhoid and paratyphoid fever are unavailable in this region [29, 30]. Escherichia coli, Klebsiella spp., and Pseudomonas and Acinetobacter spp. were other common gram negative isolates in this study which is similar to the findings of previous studies from this region [18, 19, 26]. A high prevalence of nonfermenters in this study, particularly Pseudomonas spp. (16.8%) and Acinetobacter spp. (12.9%), highlights the significant concern of hospital acquired BSI in our region.

mmon gram negative isolates in this study which is similar to the findings of previous studies from this region [18, 19, 26]. A high prevalence of nonfermenters in this study, particularly Pseudomonas spp. (16.8%) and Acinetobacter spp. (12.9%), highlights the significant concern of hospital acquired BSI in our region. The antibiotic susceptibility spectrum of bacterial pathogens significantly varies according to the patient population, their age, and the strains. In this study, Salmonella enterica isolates were noted to be susceptible to most of the routinely used antibiotics. However, increased resistance of Salmonella enterica isolates towards ampicillin (14.2%) and fluoroquinolones (pediatric 77.7% and adults 87.5%) has compromised the therapeutic choices. Nalidixic acid resistance (NAR) was documented in 85.7% of the Salmonella isolates which might be the principal cause of fluoroquinolones resistance in this study [19]. In developing countries, including Nepal, ampicillin and fluoroquinolones are still the drugs of choice for treatment of enteric fever. The occurrence of drug resistance among these isolates is of great concern. Although there were no MDR Salmonella isolates in our study, previous reports of Pokharel et al. (5%) and Khanal et al. (26%) have documented the existence of MDR Salmonella in Nepal [9, 31].

the drugs of choice for treatment of enteric fever. The occurrence of drug resistance among these isolates is of great concern. Although there were no MDR Salmonella isolates in our study, previous reports of Pokharel et al. (5%) and Khanal et al. (26%) have documented the existence of MDR Salmonella in Nepal [9, 31]. Broad spectrum antibiotics, notably cephalosporins and quinolones, are the mainstay for therapy for undifferentiated febrile illness in Nepalese hospitals citing their low toxicity and higher effectiveness [18, 19]. Irrational and increased use of these antibiotics resulting in upsurge of multiple drug resistance in microorganisms is an emerging problem [32]. Most of the isolates (42.8%) found in our study were multidrug resistant (MDR) including gram positive bacteria (58.2%) and gram negative bacteria (34.8%), respectively, and this trend was significantly high among adult patients. Similar rates of multiple drug resistant bacteria were reported from an Indian study [10]. Among Staphylococcus aureus, increased resistance was observed in ampicillin (80% and 100%), cotrimoxazole (26.6% and 84.0%), erythromycin (51.1% and 84.0%), and ciprofloxacin (25.0% and 96.7%) among pediatric and adult patients, respectively. Methicillin resistant Staphylococcus aureus (MRSA) was found in significant frequency (31.4%), almost similar to the findings of previous studies [10, 33, 34]. Isolation of MRSA strains, particularly from pediatric cases of BSI, is a serious issue as therapeutic choices become very limited in these cases [33]. Further, Enterococcus spp. in BSI cases in this study were comparatively few but a higher number (44.4%) of them were MDR which is similar to an Indian study [10].

34]. Isolation of MRSA strains, particularly from pediatric cases of BSI, is a serious issue as therapeutic choices become very limited in these cases [33]. Further, Enterococcus spp. in BSI cases in this study were comparatively few but a higher number (44.4%) of them were MDR which is similar to an Indian study [10]. Among gram negative isolates, overall antibiotic susceptibility pattern suggests a high proportion of MDR organism in our hospital. In this study, we found 34.5% of gram negative isolates to be MDR. Higher proportion (48.2%) of MDR organisms was responsible for BSI cases in adults as compared to pediatric patients (30.0%) and this trend was statistically significant (p < 0.05). Higher resistance was observed among the ampicillin (up to 100%), fluoroquinolones (up to 83%), and broader spectrum cephalosporins (up to 66.7%). The fact that cephalosporins are one of the most commonly used antibiotics for inpatients as well as for outpatients could be the reason for such high level of resistance being observed in the developing countries. Similar patterns of susceptibilities were found in other studies from South Asian region [10, 25]. Gram negative isolates resistant to broad spectrum cephalosporins and carbapenems were also documented in this study. Beta-lactam antibiotic resistance among gram negative bacteria to antibiotics is often associated with the production of hydrolytic enzymes particularly extended spectrum β-lactamases (ESBLs), class C cephalosporinase (AmpC), and carbapenemases (including Metallo-beta-lactamases) [35]. The high level of cephalosporin resistance observed in our study has been supported by higher rates of beta-lactamase producing bacterial strains. Among total 152 gram negative isolates in this study, 12.5% were producing ESBL and 3.9% producing MBL. The higher rates of ESBL and MBL documented in Klebsiella spp. (42.8% and 28.5%) were similar to the findings of a previous study [36]. Furthermore, previous reports of beta-lactamase producing organisms associated with various bacterial infections from Nepal have also highlighted the rising scenario of dissemination of these superbugs in hospital as well as in the community [37, 38]. The greatest threat with MDR, ESBL, and carbapenemase producing gram negative bacteria is that bacterial infections including bacteremia are becoming untreatable due to the limited therapeutic options of the antibiotics available, resulting into increased mortality and health care resources [39].

he community [37, 38]. The greatest threat with MDR, ESBL, and carbapenemase producing gram negative bacteria is that bacterial infections including bacteremia are becoming untreatable due to the limited therapeutic options of the antibiotics available, resulting into increased mortality and health care resources [39]. Therefore, we believe this report would be helpful in encouraging the physicians to discontinue the irrational use of antibiotics and controlling the occurrence and the spread of resistance. 5. Limitations The study has some limitations. We could not rule out the prior use of antimicrobial drugs among patients; instead the study relied on information provided by the patients or their guardians. In addition, the conventional blood culture system and single blood culture specimen could have produced in less sensitivity. Risk factors and associated clinical outcomes were also not evaluated. Molecular characterization of the resistant phenotypes and their epidemiology would be more significant in public health perspective.

n, the conventional blood culture system and single blood culture specimen could have produced in less sensitivity. Risk factors and associated clinical outcomes were also not evaluated. Molecular characterization of the resistant phenotypes and their epidemiology would be more significant in public health perspective. 6. Conclusion This study provides some insight into the local trends and bacterial etiology of bloodstream infections among pediatric and adult patients. Gram negative bacteria are the major contributors of BSI in our patients. A higher rate of antimicrobial resistance among gram negative and gram positive organisms is an alarming issue. Exact contributing factors for the bloodstream infections (BSI) within these groups need to be further elucidated. Rational use of antibiotics, formulation of antibiotic policy, and prompt therapy of bloodstream infections for the effective management and prevention of drug resistance are urgently needed in our setting. Acknowledgments The authors are grateful to all the staff of Microbiology Laboratory of the Manmohan Memorial Teaching Hospital, for their support. In addition, They are extremely thankful to the patients and their guardians for providing necessary information without whom this research would be languishing in limbo. Abbreviations ASM:American Society for Microbiology ATCC:American type culture collection BHI:Brain heart infusion BSI:Bloodstream infection CDT:Combined disk test CLSI:Clinical and Laboratory Standards Institute CAZ:Ceftazidime CTX:Cefotaxime iMLSB:Inducible macrolide-lincosamide-streptogramin-B

Acknowledgments The authors are grateful to all the staff of Microbiology Laboratory of the Manmohan Memorial Teaching Hospital, for their support. In addition, They are extremely thankful to the patients and their guardians for providing necessary information without whom this research would be languishing in limbo. Abbreviations ASM:American Society for Microbiology ATCC:American type culture collection BHI:Brain heart infusion BSI:Bloodstream infection CDT:Combined disk test CLSI:Clinical and Laboratory Standards Institute CAZ:Ceftazidime CTX:Cefotaxime iMLSB:Inducible macrolide-lincosamide-streptogramin-B ESBL:Extended spectrum β-lactamase MBL:Metallo-β-lactamase MRSA:Methicillin resistant Staphylococcus aureus MDR:Multidrug resistant ZOI:Zone of inhibition. Additional Points Availability of Data and Materials. The primary raw data will be made available to the interested researchers by the corresponding author of this article if requested. Ethical Approval This research was approved by the Institutional Review Committee of Manmohan Memorial Institute of Health Sciences (IRC-MMIHS), Kathmandu, Nepal. Letter of approval (Ref. number 004/MMIHS/2071) was obtained after submitting and presenting the proposal to the committee. Consent Informed consent was taken from the patients or their parents before participating in the study. Data regarding personal information and infectious disease were coded and kept confidential. Conflicts of Interest There is nothing to be declared in this work.

Ethical Approval This research was approved by the Institutional Review Committee of Manmohan Memorial Institute of Health Sciences (IRC-MMIHS), Kathmandu, Nepal. Letter of approval (Ref. number 004/MMIHS/2071) was obtained after submitting and presenting the proposal to the committee. Consent Informed consent was taken from the patients or their parents before participating in the study. Data regarding personal information and infectious disease were coded and kept confidential. Conflicts of Interest There is nothing to be declared in this work. Authors' Contributions Narayan Prasad Parajuli conceived the design of the study, participated in case identification by laboratory investigations, and drafted the manuscript. All remaining authors contributed towards laboratory investigations, data analysis, and drafting and revising the paper and agreed to be accountable for all aspects of the work. Table 1 Trends of bloodstream infection among various group of patients (n = 3,088). Patients Age group Number of patients (%) Number of patients with BSI (%) p Pediatric Infants (<1 year) 386 (12.5%) 50 (12.9%) 0.008 Children (1–14 years) 574 (18.6%) 40 (6.9%) Adults Adults (15–59 years) 1949 (63.2%) 131 (6.7%) Elderly (60 years or above) 179 (5.7%) 10 (5.5%) Total 3,088 231 (7.48%) Table 2 Trends of bacterial isolates associated with bloodstream infections among pediatric and adult patients (n = 231). Bacterial isolates Number (%) Pediatric patients (n = 90) Adult patients (n = 141) p

Children (1–14 years) 574 (18.6%) 40 (6.9%) Adults Adults (15–59 years) 1949 (63.2%) 131 (6.7%) Elderly (60 years or above) 179 (5.7%) 10 (5.5%) Total 3,088 231 (7.48%) Table 2 Trends of bacterial isolates associated with bloodstream infections among pediatric and adult patients (n = 231). Bacterial isolates Number (%) Pediatric patients (n = 90) Adult patients (n = 141) p Outpatients (n = 28) Inpatients (n = 62) Outpatients (n = 50) Inpatients (n = 91) Gram positive isolates 79 (34.2) 14 (50.0) 35 (56.4) 8 (16.0) 22 (24.1) Staphylococcus aureus 70 (30.3) 13 (46.5) 32 (51.6) 8 (16.0) 17 (18.6) 0.000 Enterococcus spp. 9 (3.9) 1 (3.5) 3 (4.8) 0 (0.0) 5 (5.4) 0.492 Gram negative isolates 152 (65.8) 14 (50.0) 27 (43.6) 42 (84.0) 69 (75.9) Salmonella enterica 49 (21.2) 9 (32.1) 0 (0.0) 37 (74.0) 3 (3.2) 0.001 Escherichia coli 20 (8.6) 4 (14.2) 4 (6.4) 5 (10.0) 7 (7.6) 0.550 Klebsiella pneumoniae 7 (3.0) 0 (0.0) 3 (4.8) 0 (0.0) 4 (4.3) 0.559 Pseudomonas aeruginosa 39 (16.8) 0 (0.0) 8 (12.9) 0 (0.0) 31 (34.0) 0.007 Acinetobacter spp. 30 (12.9) 0 (0.0) 9 (14.5) 0 (0.0) 21 (23.0) 0.191 Citrobacter spp. 4 (1.7) 1 (3.5) 1 (1.6) 0 (0.0) 2 (2.1) 0.507 Enterobacter spp. 3 (1.3) 0 (0.0) 2 (3.3) 0 (0.0) 1 (1.0) 0.336 Table 3 Trends of antimicrobial susceptibilities of gram negative isolates (excluding Salmonella) among pediatric and adult patients (n = 96). AMP COT CIP LEV CFM CAZ GEN AK IPM MRP PIT C PB

Acinetobacter spp. 30 (12.9) 0 (0.0) 9 (14.5) 0 (0.0) 21 (23.0) 0.191 Citrobacter spp. 4 (1.7) 1 (3.5) 1 (1.6) 0 (0.0) 2 (2.1) 0.507 Enterobacter spp. 3 (1.3) 0 (0.0) 2 (3.3) 0 (0.0) 1 (1.0) 0.336 Table 3 Trends of antimicrobial susceptibilities of gram negative isolates (excluding Salmonella) among pediatric and adult patients (n = 96). AMP COT CIP LEV CFM CAZ GEN AK IPM MRP PIT C PB Escherichia coli (n = 20) Pediatric (n = 8) 4 (50.0) 2 (25.0) 0 (0.0) 0 (0.0) 2 (25.0) 2 (25.0) 4 (50.0) 0 (0.0) 0 (0.0) 0 (0.0) 2 (25.0) 0 (0.0) — Adults (n = 12) 8 (66.7) 4 (33.3) 10 (83.3) 10 (83.3) 8 (66.7) 8 (66.7) 4 (33.3) 4 (33.3) 2 (16.7) 2 (16.7) 2 (16.7) 4 (33.3) — Klebsiella spp. (n = 7) Pediatric (n = 3) 3 (100) 2 (66.7) 2 (66.7) 1 (33.3) 1 (33.3) 0 (0.0) 1 (33.3) 0 (0.0) 0 (0.0) 0 (0.0) 1 (33.3) 0 (0.0) — Adults (n = 4) 4 (100) 3 (75) 3 (75.0) 3 (75.0) 3 (75.0 3 (75.0 3 (75.0 3 (75.0) 2 (50.0) 2 (50.0) 2 (50.0) 2 (50.0) — Pseudomonas spp. (n = 39) Pediatric (n = 8) — — 2 (25.0) 1 (12.5) — 1 (12.5) 2 (25.0) 1 (12.5) 0 (0.0) 0 (0.0) 1 (12.5) 0 (0.0) 0 (0.0) Adults (n = 31) — — 30 (96.7) 27 (87.0) — 6 (19.3) 21 (67.7) 18 (58.0) 7 (22.5) 7 (22.5) 9 (29.0) 4 (12.9) 0 (0.0)

Klebsiella spp. (n = 7) Pediatric (n = 3) 3 (100) 2 (66.7) 2 (66.7) 1 (33.3) 1 (33.3) 0 (0.0) 1 (33.3) 0 (0.0) 0 (0.0) 0 (0.0) 1 (33.3) 0 (0.0) — Adults (n = 4) 4 (100) 3 (75) 3 (75.0) 3 (75.0) 3 (75.0 3 (75.0 3 (75.0 3 (75.0) 2 (50.0) 2 (50.0) 2 (50.0) 2 (50.0) — Pseudomonas spp. (n = 39) Pediatric (n = 8) — — 2 (25.0) 1 (12.5) — 1 (12.5) 2 (25.0) 1 (12.5) 0 (0.0) 0 (0.0) 1 (12.5) 0 (0.0) 0 (0.0) Adults (n = 31) — — 30 (96.7) 27 (87.0) — 6 (19.3) 21 (67.7) 18 (58.0) 7 (22.5) 7 (22.5) 9 (29.0) 4 (12.9) 0 (0.0) Acinetobacter spp. (n = 30) Pediatric (n = 9) — 5 (55.5) 4 (44.4) 2 (22.2) 4 (44.4) 4 (44.4) 5 (55.5) 3 (33.3) 1 (11.1) 1 (11.1) 1 (11.1) 3 (33.3) — Adults (n = 21) — 11 (52.3) 10 (47.2) 8 (38.1) 14 (66.7) 14 (66.7) 8 (38.1) 8 (38.1) 5 (23.8) 5 (23.8) 5 (23.8) 8 (38.1) — AMP = ampicillin, COT = cotrimoxazole (trimethoprim + sulfamethoxazole), CIP = ciprofloxacin, LEV = levofloxacin, CFM = cefixime, CAZ = ceftazidime, GEN = gentamycin, AK = amikacin, IPM = imipenem, MRP = meropenem, PIT = piperacillin + tazobactam, C = chloramphenicol, PB = polymyxin. Table 4 Trends of antimicrobial susceptibilities of Salmonella enterica clinical isolates (n = 49).

Acinetobacter spp. (n = 30) Pediatric (n = 9) — 5 (55.5) 4 (44.4) 2 (22.2) 4 (44.4) 4 (44.4) 5 (55.5) 3 (33.3) 1 (11.1) 1 (11.1) 1 (11.1) 3 (33.3) — Adults (n = 21) — 11 (52.3) 10 (47.2) 8 (38.1) 14 (66.7) 14 (66.7) 8 (38.1) 8 (38.1) 5 (23.8) 5 (23.8) 5 (23.8) 8 (38.1) — AMP = ampicillin, COT = cotrimoxazole (trimethoprim + sulfamethoxazole), CIP = ciprofloxacin, LEV = levofloxacin, CFM = cefixime, CAZ = ceftazidime, GEN = gentamycin, AK = amikacin, IPM = imipenem, MRP = meropenem, PIT = piperacillin + tazobactam, C = chloramphenicol, PB = polymyxin. Table 4 Trends of antimicrobial susceptibilities of Salmonella enterica clinical isolates (n = 49). Antimicrobial agents Salmonella enterica serotype Typhi (n = 22) Salmonella enterica serotype Paratyphi (n = 27) Resistance (%) Resistance (%) Pediatric (n = 5) Adults (n = 17) Pediatric (n = 4) Adults (n = 23) Ampicillin 0 (0.0) 2 (11.7) 1 (25.0) 4 (17.3) Cotrimoxazole 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) Ciprofloxacin 4 (80.0) 12 (70.5) 3 (75.0) 23 (100.0) Cefixime 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) Ceftriaxone 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) Chloramphenicol 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) Azithromycin 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) Table 5 Trends of antimicrobial susceptibilities of gram positive isolates (n = 79).

0.0) Ciprofloxacin 4 (80.0) 12 (70.5) 3 (75.0) 23 (100.0) Cefixime 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) Ceftriaxone 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) Chloramphenicol 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) Azithromycin 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) Table 5 Trends of antimicrobial susceptibilities of gram positive isolates (n = 79). Antimicrobial agents Staphylococcus aureus (n = 70) Enterococcus (n = 9) % resistance % resistance Pediatric (n = 45) Adults (n = 25) Pediatric (n = 4) Adults (n = 5) Ampicillin 36 (80.0) 25 (100) 1 (25.0) 4 (80.0) Cotrimoxazole 12 (26.6) 21 (84.0) — — Ciprofloxacin 11 (24.4) 22 (88.0) 2 (50.0) 2 (40.0) Cephalexin 17 (37.7) 11 (44.0) — — Cloxacillin 17 (37.7) 9 (36.0) — — Erythromycin 23 (51.1) 21 (84.0) — — Levofloxacin 9 (20.9) 17 (68.0) 1 (25.0) 3 (60.0) Amikacin 4 (8.9) 11 (44.0) 2 (50.0) 2 (40.0) Clindamycin 23 (51.14) 20 (80.0) — — Piperacillin-tazobactam 8 (17.7) 6 (24.0) 1 (25.0) 3 (60.0) Imipenem 17 (37.7) 9 (36.0) — — Teicoplanin 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) Table 6 Trends of multidrug resistance of bacterial isolates among pediatric and adult patients (n = 231). Bacterial isolates Total number (%) Pediatric patients (n = 90) Adult patients (n = 141) p  value Total MDR isolates (GNB) 53 (34.8) 12 (29.2) 41 (36.9) 0.346 Total ESBL isolates (GNB) 19 (12.5) 8 (19.5) 11 (7.8) 0.136 Total MBL isolates (GNB) 6 (3.9) 1 (2.4) 5 (4.5) 0.433 Total MDR isolates (GPC) 43 (54.4) 17 (34.6) 26 (86.6) 0.000 Total MDR isolates 96 (41.5) 29 (32.2) 67 (47.5) 0.020 GNB = gram negative bacilli, GPC = gram positive cocci.

Bacterial isolates Total number (%) Pediatric patients (n = 90) Adult patients (n = 141) p  value Total MDR isolates (GNB) 53 (34.8) 12 (29.2) 41 (36.9) 0.346 Total ESBL isolates (GNB) 19 (12.5) 8 (19.5) 11 (7.8) 0.136 Total MBL isolates (GNB) 6 (3.9) 1 (2.4) 5 (4.5) 0.433 Total MDR isolates (GPC) 43 (54.4) 17 (34.6) 26 (86.6) 0.000 Total MDR isolates 96 (41.5) 29 (32.2) 67 (47.5) 0.020 GNB = gram negative bacilli, GPC = gram positive cocci. Table 7 Trends of antimicrobial resistance mechanisms in gram negative isolates among pediatric and adult patients (n = 231). Bacterial isolates Total number (%) Pediatric patients (n = 90) Adult patients (n = 141) p value Salmonella enterica n = 49 MDR 0 (0.0) 0 (0.0) 0 (0.0) — NARS 42 (85.7) 7 (77.7) 35 (87.5) 0.381 Escherichia coli n = 20 MDR 10 (50.0) 3 (37.5) 7 (58.3) 0.325 ESBL 6 (30.0) 2 (25.0) 4 (33.3) 0.545 MBL 2 (10.0) 0 (0.0) 2 (16.7) 0.347 Klebsiella spp. n = 7 MDR 5 (71.4) 1 (33.3) 4 (100) 0.143 ESBL 3 (42.8) 1 (33.3) 2 (50.0) 0.629 MBL 2 (28.5) 0 (0.0) 2 (50.0) 0.286 Pseudomonas spp. n = 39 MDR 16 (41.0) 2 (25.0) 14 (45.1) 0.269 ESBL 5 (12.8) 2 (25.0) 3 (9.6) 0.268 MBL 0 (0.0) 0 (0.0) 0 (0.0) — Acinetobacter spp. n = 30 MDR 22 (73.3) 6 (66.6) 16 (76.1) 0.453 ESBL 5 (16.7) 3 (33.3) 2 (9.5) 0.143 MBL 2 (6.6) 1 (11.1) 1 (4.7) 0.517 GNB: gram negative bacteria. Table 8 Trends of antimicrobial resistance mechanisms in gram positive isolates among pediatric and adult patients (n = 79). Bacterial isolates Number (%) Pediatric patients (n = 90) Adult patients (n = 141) p value Staphylococcus aureus n = 70 MDR 42 (60.0) 16 (35.5) 23 (92.0) 0.000

Acinetobacter spp. n = 30 MDR 22 (73.3) 6 (66.6) 16 (76.1) 0.453 ESBL 5 (16.7) 3 (33.3) 2 (9.5) 0.143 MBL 2 (6.6) 1 (11.1) 1 (4.7) 0.517 GNB: gram negative bacteria. Table 8 Trends of antimicrobial resistance mechanisms in gram positive isolates among pediatric and adult patients (n = 79). Bacterial isolates Number (%) Pediatric patients (n = 90) Adult patients (n = 141) p value Staphylococcus aureus n = 70 MDR 42 (60.0) 16 (35.5) 23 (92.0) 0.000 MRSA 22 (31.4) 12 (17.1) 10 (14.3) 0.188 iMLSB 7 (10.0) 3 (4.2) 4 (5.8) 0.201 Enterococcus spp. n = 9 MDR 4 (44.4) 1 (11.1) 3 (33.3) 0.357

1. Introduction Coagulase-positive and coagulase-negative staphylococci (CoNS) are commonly isolated from blood cultures. Staphylococcus aureus (S. aureus) is responsible for a serious bacteremia requiring immediate antibiotic treatment. Delay in optimal therapy is associated with prolonged hospitalization and higher rates of mortality [1]. Vancomycin initiation is considered the standard of care for empiric therapy when Gram-positive cocci in clusters (GPCC) are reported in suspected blood stream infections. Furthermore, vancomycin continues to be first line for the treatment of methicillin-resistant Staphylococcus aureus (MRSA) bacteremia with minimum inhibitory concentrations of ≤1.5 mcg/mL. However, this agent has been found to be inferior to antistaphylococcal beta-lactams for methicillin-susceptible Staphylococcus aureus (MSSA) strains [2, 3]. A previous 6-month retrospective chart review was performed at our teaching hospital in an effort to characterize the most common Staph etiology. In this review, 16 cases were due to MSSA and 26 were due to MRSA. The time to optimal coverage for MSSA with an antistaphylococcal beta-lactam was 3.86 days. Patients with S. aureus bacteremia on average were hospitalized for 12.3 days and received 29.7 days of antibiotic therapy.

erize the most common Staph etiology. In this review, 16 cases were due to MSSA and 26 were due to MRSA. The time to optimal coverage for MSSA with an antistaphylococcal beta-lactam was 3.86 days. Patients with S. aureus bacteremia on average were hospitalized for 12.3 days and received 29.7 days of antibiotic therapy. CoNS are common blood culture contaminants of which only about 20% are considered true opportunistic pathogens [4]. A typical contaminated blood culture investigation will reveal only one positive culture out of multiple bottle draws, no signs, or symptoms of endovascular infection and an immunocompetent host. However, true CoNS bloodstream infections are frequently associated with foreign bodies such as vascular catheters and implantable medical devices [5, 6]. Consequences of contaminated blood cultures include increased cost, prolonged hospitalization, overuse of antibiotics with associated adverse events, and additional work for members of the healthcare team [7]. A recent retrospective review at our teaching hospital identified a positive blood culture contamination rate of 35%.

ences of contaminated blood cultures include increased cost, prolonged hospitalization, overuse of antibiotics with associated adverse events, and additional work for members of the healthcare team [7]. A recent retrospective review at our teaching hospital identified a positive blood culture contamination rate of 35%. Conventional methods of bacterial identification which differentiate MRSA from MSSA and CoNS typically require 24–48 hours from the time blood cultures turn positive [8]. Rapid diagnostics are capable of identifying these organisms in as little as 30 minutes from the positive blood culture report [6]. Several studies have demonstrated the following: a reduction in time to optimal antibiotic therapy by 1.7–2.5 days, a decrease in length of hospital stay by around 2–6 days, and a median reduction of hospital costs for up to $20,000 per patient [5, 7, 9]. Previous studies have also identified that a direct microbiology notification to an infectious disease (ID) pharmacy team results in facilitation of this intervention [6]. Patients with more than one infection pending microbiologic workup of other sites may require a continuation of empiric vancomycin and would therefore not be expected to maximally benefit from rapid blood culture identification.

an infectious disease (ID) pharmacy team results in facilitation of this intervention [6]. Patients with more than one infection pending microbiologic workup of other sites may require a continuation of empiric vancomycin and would therefore not be expected to maximally benefit from rapid blood culture identification. There are several molecular FDA approved methods for rapid blood culture identification. Testing platforms include GeneXpert, MALDI TOF mass spectrometry, and QuickFISH. Molecular methods range in testing sophistication, need for equipment acquisition, and reagent cost. QuickFISH GPCC BC (QF GPCC) is a peptide nucleic acid fluorescence in situ hybridization (PNA-FISH) test capable of differentiating staphylococci in positive blood cultures. QuickFISH targets the 16S rRNA of S. aureus and CoNS directly from positive blood cultures with a sensitivity and specificity exceeding 98% [10, 11]. Fluorescently labeled probes bind to a specified region of the bacterial RNA to form a distinct green or red color when visualized under the microscope. Organism identification is based on color emission. Our teaching hospital uses culture based technology to identify Staph species from positive blood cultures. The goal of this study is to evaluate whether introducing rapid diagnostic testing in conjunction with implementing a stratification algorithm for testing eligibility would be an appropriate clinical and cost saving approach.

hospital uses culture based technology to identify Staph species from positive blood cultures. The goal of this study is to evaluate whether introducing rapid diagnostic testing in conjunction with implementing a stratification algorithm for testing eligibility would be an appropriate clinical and cost saving approach. 2. Methodology An internal concurrent 4-month observational study was performed. Due to platform availability at the time, QuickFISH was selected for this quality improvement project. For the purpose of performing this validation study, test kits and other equipment were provided by the manufacturer free of charge. This study was approved by the Hospital Integrity Board. During the study period, positive blood cultures continued to be worked up in accordance with the standard of care. An additional call to the ID pharmacy service occurred for all positive blood cultures with GPCC during the hours of 0700-1530. The ID pharmacy service investigated each case to determine if rapid identification of the Staphylococcus species would impact the use of antibiotics. If the ID pharmacy service deemed there is a need to continue empiric vancomycin, the rapid test would not be performed. However, if the Staph bacteremia was considered to be a monomicrobic infection or a potential contaminant and the patient was on vancomycin, the blood culture isolate would be rapidly identified.

cs. If the ID pharmacy service deemed there is a need to continue empiric vancomycin, the rapid test would not be performed. However, if the Staph bacteremia was considered to be a monomicrobic infection or a potential contaminant and the patient was on vancomycin, the blood culture isolate would be rapidly identified. The target population for testing was identified using a stratified approach (Figure 1) and included age > 18 years, immunocompetent, no foreign devices in place such as central lines or cardiac devices, no additional infectious disease indications requiring continuation of empiric vancomycin, and no-risk factors for MRSA such as recent hospitalization, recent antibiotic exposure, prior MRSA infection, or nasal colonization. If the patient met criteria, the ID pharmacy service would notify microbiology to run the QuickFISH test. Results were not made available for patient intervention since the test results were used for research purposes only.

recent hospitalization, recent antibiotic exposure, prior MRSA infection, or nasal colonization. If the patient met criteria, the ID pharmacy service would notify microbiology to run the QuickFISH test. Results were not made available for patient intervention since the test results were used for research purposes only. 3. Definitions We have defined optimal antibiotic therapy as the switch from empiric vancomycin to cefazolin or nafcillin in patients with MSSA bacteremia or discontinuation of unnecessary vancomycin in patients with CoNS if deemed to be a contaminant. Time to optimal antibiotic therapy for culture based methods was defined as the time microbiology reported the GPCC after performing the Gram stain to the time of antibiotic optimization. For rapid diagnostic testing, the expected time to optimal treatment for CoNS was defined as 2 hours from the time when GPCC was identified by Gram stain and 4 hours for S. aureus identification including mecA gene determination. We calculated the expected time that could have been avoided in hours to optimize antibiotic therapy using the new technology by subtracting 2 hours (if CoNS) or 4 hours (if MSSA) from the time to optimal antibiotic therapy using culture based methodology for microbiological identification.

ing mecA gene determination. We calculated the expected time that could have been avoided in hours to optimize antibiotic therapy using the new technology by subtracting 2 hours (if CoNS) or 4 hours (if MSSA) from the time to optimal antibiotic therapy using culture based methodology for microbiological identification. 4. Results A total of 43 patients with GPCC were screened for QuickFISH test eligibility. Nine patients met the inclusion criteria for testing and therefore underwent QuickFISH (Table 1). Time to optimal antibiotic therapy may be decreased by approximately 20 to 74 hours (M = 35, SD = 16) for a total of 314 hours. Among the nine patients, 7 patients had CoNS and one had micrococcus. Among patients with CoNS, call back to the emergency department (ED) would have been avoided for 2 patients with contaminated bottles. One patient with MSSA bacteremia would have been expected to receive optimal therapy 74 hours earlier. A total of 34 patients did not meet the criteria for QuickFISH testing (Table 2). All the 34 patients received therapy with vancomycin for a range of 3–14 days (median = 6 days). The majority of these patients had CoNS (N = 28), 5 patients had MRSA, and 1 patient had MSSA. Of the 6 patients who were excluded due to presence of MRSA risk factors, 4 cultures were deemed to be contaminated and therefore we would have benefited from knowing the identification rapidly to prevent unnecessary antibiotic exposure. Projected hospital cost avoidance in the 9 cases that met the inclusion criteria for testing if rapid testing implemented is as follows: one patient with MSSA bacteremia: using primary literature cost data by Bauer and colleagues, the cost avoidance would have been $21,387; two ED patients with contaminated blood culture bottles: cost avoidance for ED call back would have been $718 × 2 = $1436 (this cost represents the ED visit only and does not include any interventions, doctors, pharmacists, and nurses time, medications, imaging studies, cultures, or laboratory tests); one patient had a contaminated blood culture bottle and discharge held by one day with a cost avoidance for extra day being approximately $1,024; and 5 patients for which vancomycin was initiated due to the panic call of GPCC in blood culture. The average drug acquisition cost of vancomycin per day is approximately $20.

ne patient had a contaminated blood culture bottle and discharge held by one day with a cost avoidance for extra day being approximately $1,024; and 5 patients for which vancomycin was initiated due to the panic call of GPCC in blood culture. The average drug acquisition cost of vancomycin per day is approximately $20. Therefore, $100 in drug acquisition cost could have been avoided (this cost does not include the cost for therapeutic drug monitoring or additional case workup for positive blood culture investigation). If we apply Gander and colleagues' estimates to all 8 contaminated blood cultures, the expected cost avoidance would total (8 × $8,720) $69,760 [12]. During the study period, 43 patients had GPCC reported in blood culture bottles. Additionally, the projected cost for performing this test is $130 per patient. This includes both the QuickFISH and mecA gene tests as well as labor cost. If the QuickFISH test had been indiscriminately implemented for all cases, the cost for performing this test would have been $5,590. However, using the stratified approach, only 9 patients were tested for a projected cost of $1,170 and the cost avoidance for untested patients, not including the 4 patients who were inappropriately excluded because of MRSA risk factors, would have been $3900.

, the cost for performing this test would have been $5,590. However, using the stratified approach, only 9 patients were tested for a projected cost of $1,170 and the cost avoidance for untested patients, not including the 4 patients who were inappropriately excluded because of MRSA risk factors, would have been $3900. 5. Discussion Time to optimal antibiotic therapy may be decreased by approximately 1–3 days if interventions have been made based on the results of the rapid diagnostic assays for the 9 cases that met the inclusion criteria for testing. Additionally, earlier discharge may be achieved and call back to the ED following discharge may be avoided. Stratification based on eligibility for the rapid identification test minimizes the use of this test for which early identification of GPCC is not likely to impact empiric antibiotic therapy. Therefore, a stratification algorithm for rapid molecular identification of GPCC from blood cultures was created accordingly (Figure 1). Among the 34 patients who did not meet the criteria for rapid testing, 30 patients were appropriately excluded, while the cultures for the remaining 4 patients were deemed to be contaminated and therefore we would have benefited from knowing the identification rapidly in order to prevent unnecessary antibiotic exposure. Therefore, we concluded that the presence of MRSA risk factor does not appear to be appropriate exclusion criteria and was consequently removed from the proposed stratification algorithm.

d and therefore we would have benefited from knowing the identification rapidly in order to prevent unnecessary antibiotic exposure. Therefore, we concluded that the presence of MRSA risk factor does not appear to be appropriate exclusion criteria and was consequently removed from the proposed stratification algorithm. Even though molecular assays have been shown to reduce the time to optimal antibiotic therapy as well as length of hospital stay and cost of stay in previously published literature, in this observational study, we tested the appropriateness of a stratified approach to minimize the use of the rapid diagnostics to include only cases where early identification of GPCC is likely to impact empiric antibiotic therapy. However, this study has several limitations. First, it represented a small sample size because the ID PharmD is not available 24 hours per day for 7 days to receive all the notifications for GPCC. Second, it is a single center, 4-month observational trial. Third, rapid diagnostic was tested for validation only and the intervention was not allowed. Therefore, actual time to optimal antibiotic therapy for rapid diagnostic test was unable to be measured which could otherwise be longer than 2 hours for CoNS and 4 hours for S. aureus. This time included the expected microbiology workup time following the Gram stain and did not include the lag time between the microbiology reporting and the antibiotic optimization. A two-arm, prospective study design could overcome this limitation which is the potential future direction of this trial. Fourth, the presence of MRSA risk factors was not an appropriate exclusion criteria and was subsequently removed from the hospital's proposed rapid diagnostic stratification algorithm. Finally, some of the cost avoidance data were extrapolated from previously published cost avoidance trials.

re direction of this trial. Fourth, the presence of MRSA risk factors was not an appropriate exclusion criteria and was subsequently removed from the hospital's proposed rapid diagnostic stratification algorithm. Finally, some of the cost avoidance data were extrapolated from previously published cost avoidance trials. 6. Conclusion Introducing rapid molecular testing in conjunction with implementing patient stratification algorithm for rapid molecular identification of GPCC from blood cultures in addition to the ID pharmacy intervention will provide a positive impact on the clinical and economic outcomes in our health care setting. Disclosure Thamer A. Almangour is a Former PGY2 Infectious Diseases Pharmacy Resident at Columbus Regional Health, Midtown Medical Center. Abdullah A. Alhifany is a Former PGY1 Pharmacy Practice Resident at Columbus Regional Health, Midtown Medical Center. Conflicts of Interest QuickFISH test kits and equipment were provided by the manufacturer free of charge for the purpose of performing this validation study (Thamer A. Almangour). The authors declare that there are no additional conflicts of interest regarding the publication of this paper. Figure 1 Criteria for rapid diagnostic testinga. Table 1 Patients who met inclusion criteria for rapid diagnostic testinga.

Conflicts of Interest QuickFISH test kits and equipment were provided by the manufacturer free of charge for the purpose of performing this validation study (Thamer A. Almangour). The authors declare that there are no additional conflicts of interest regarding the publication of this paper. Figure 1 Criteria for rapid diagnostic testinga. Table 1 Patients who met inclusion criteria for rapid diagnostic testinga. Age Gender Empiric therapy Isolated microorganism Plan made Time to optimal antibiotic therapy (hours) using culture based method Expected time avoided (hours) to optimal antibiotic therapy using the molecular diagnostic test 61 F Vancomycin MSSA Changed to nafcillin 78 74 68 M Vancomycin CoNS Vancomycin discontinued 44 42 32 F Vancomycin and levofloxacin CoNS Discharged from ED n/a 24∗ 78 F Vancomycin CoNS Vancomycin discontinued 30 28 74 F Vancomycin and doxycycline CoNS Vancomycin discontinued and patient discharged 49 47 49 M Vancomycin CoNS Discharged from ED n/a 24∗ 80 F Vancomycin and ceftriaxone Micrococcus Vancomycin discontinued 28 26 85 F Vancomycin and piperacillin/tazobactam CoNS Vancomycin discontinued 22 20 29 M Vancomycin CoNS Vancomycin discontinued 31 29 aF = female, M = male, MSSA = methicillin-susceptible Staphylococcus aureus, CoNS = coagulase-negative staphylococci, and ED = emergency department. ∗Patients were called to return for emergency department evaluation, to assess clinical status and repeat blood cultures. Table 2 Patients who did not meet inclusion criteria for rapid diagnostic testinga.

aF = female, M = male, MSSA = methicillin-susceptible Staphylococcus aureus, CoNS = coagulase-negative staphylococci, and ED = emergency department. ∗Patients were called to return for emergency department evaluation, to assess clinical status and repeat blood cultures. Table 2 Patients who did not meet inclusion criteria for rapid diagnostic testinga. Reason for the exclusion Number of patients (n = 34) Additional infectious disease indications requiring continuation of empiric vancomycin 20 Immunocompromised host 3 Foreign device in place 5 Risk factor for MRSA 6 aMRSA = methicillin-resistant Staphylococcus aureus.

1. Introduction Timely surveillance of enteric diseases is necessary to identify and control cases and outbreaks to prevent further transmission. Previous work conducted in British Columbia (BC), Canada, showed that 2.8 foodborne disease outbreaks per million population are reported on average every year [1] and if outbreaks are identified early, they are more likely to be solved [2]. This led us to launch a project to assess the timeliness of surveillance. Timeliness of surveillance is defined as the time from onset of illness to the reporting of case and laboratory information to public health (PH) authorities. If this process is measured in a consistent way, we can improve our understanding of surveillance and identify steps requiring improvement. Several regions and countries have conducted this work which led to changes to improve timeliness [3–8]. In BC (population 4.6 million), Salmonella, shiga-toxin producing E. coli (STEC), Shigella, and Listeria infections are reportable to regional health authorities. Salmonella and STEC account for the largest number of bacterial enteric disease outbreaks in BC [1]. STEC, Shigella, and Listeria infections can cause severe acute illness and complications [9]. Our objective was to evaluate the timeliness of enteric disease surveillance in BC and to compare our results to similar work conducted in other jurisdictions internationally in order to improve the timeliness of surveillance in BC.

In BC (population 4.6 million), Salmonella, shiga-toxin producing E. coli (STEC), Shigella, and Listeria infections are reportable to regional health authorities. Salmonella and STEC account for the largest number of bacterial enteric disease outbreaks in BC [1]. STEC, Shigella, and Listeria infections can cause severe acute illness and complications [9]. Our objective was to evaluate the timeliness of enteric disease surveillance in BC and to compare our results to similar work conducted in other jurisdictions internationally in order to improve the timeliness of surveillance in BC. 2. Methods In BC, local laboratories report cases of Salmonella, STEC, Shigella, and Listeria infection to the regional health authority via fax or electronically. Cases are interviewed by regional public health officials using standard case report forms (CRF) (http://www.bccdc.ca/health-professionals/professional-resources/surveillance-forms). Although cases are reported provincially using an electronic public health information system, in 2012-13, most of the dates were kept on paper at each regional health authority.

health officials using standard case report forms (CRF) (http://www.bccdc.ca/health-professionals/professional-resources/surveillance-forms). Although cases are reported provincially using an electronic public health information system, in 2012-13, most of the dates were kept on paper at each regional health authority. All isolates are sent to the BC Centre for Disease Control (BCCDC) Public Health Laboratory (PHL) for further typing which includes speciation, serotyping, and pulse-field gel electrophoresis (PFGE), as required. Isolates which are difficult to serotype are sent to the National Microbiology Laboratory (NML) for serotyping. Results and dates are entered into laboratory information systems (Sunquest and Bionumerics). Partial data (confirmation, speciation, and serotyping) can be viewed by BCCDC provincial epidemiologists. All Salmonella Enteritidis and Heidelberg isolates are sent to the NML for phage type (PT) analysis. Results are returned by fax to the PHL and entered into a database. PFGE and PT results for Salmonella Enteritidis, Heidelberg, Typhimurium, E. coli O157:H7, Shigella, and Listeria are emailed weekly to the provincial epidemiologists for surveillance purposes.

isolates are sent to the NML for phage type (PT) analysis. Results are returned by fax to the PHL and entered into a database. PFGE and PT results for Salmonella Enteritidis, Heidelberg, Typhimurium, E. coli O157:H7, Shigella, and Listeria are emailed weekly to the provincial epidemiologists for surveillance purposes. Date information on Salmonella, STEC, Shigella, and Listeria infections reported in 2012 and 2013 were collected from paper CRF and laboratory information systems. Twelve date variables were collected representing the surveillance process from onset of symptoms to case interview (epidemiological surveillance stream) and from onset of symptoms to final laboratory result provided to provincial epidemiologists (laboratory surveillance stream) (Figure 1). The data collected from paper and electronic records were linked using a unique laboratory ID and Personal Health Number. If data from different systems could not be linked due to missing or different unique identifier, only dates where intervals could be measured were included. Time intervals between each date were measured in days. Negative intervals and intervals greater than 150 days were evaluated for accuracy. Obvious date errors were manually corrected (i.e., month and day interchanged) or excluded. Medians, ranges, and interquartile (IQ) intervals were measured for each interval and disease using MS Excel®. Our findings were compared to similar data published from 2005–2015 identified through a literature search using Medline and through a review of the grey literature available on the Internet [3–8].

Time intervals between each date were measured in days. Negative intervals and intervals greater than 150 days were evaluated for accuracy. Obvious date errors were manually corrected (i.e., month and day interchanged) or excluded. Medians, ranges, and interquartile (IQ) intervals were measured for each interval and disease using MS Excel®. Our findings were compared to similar data published from 2005–2015 identified through a literature search using Medline and through a review of the grey literature available on the Internet [3–8]. 3. Results Data on 2615 cases reported in 2012-13 were included: Salmonella (n = 1831), STEC (n = 434), Shigella (n = 323), and Listeria (n = 27). Epidemiological case data was missing for one of the five regional health authorities in 2013.

Our findings were compared to similar data published from 2005–2015 identified through a literature search using Medline and through a review of the grey literature available on the Internet [3–8]. 3. Results Data on 2615 cases reported in 2012-13 were included: Salmonella (n = 1831), STEC (n = 434), Shigella (n = 323), and Listeria (n = 27). Epidemiological case data was missing for one of the five regional health authorities in 2013. In the laboratory surveillance stream, it took a median of 3–7 days from symptom onset to sample collection and a median of 3–5 days from sample collection to sample receipt at the PHL, after initial diagnosis was completed at a local laboratory (Table 1). It took a median of 1–6 days for the PHL to issue results (including confirmatory culture where necessary and speciation or serotyping). There was a median interval of 6–8 days from PHL confirmatory culture to start the PFGE and 8–10 days from start of PFGE run to report PFGE results to provincial epidemiologists. For Salmonella, it took a median of 7 days to send isolates for PT, 5 days to receive the PT results, and another 9 days (IQ range = 2–22 days) to report PT results to provincial epidemiologists. The total median time from onset of symptoms to report PFGE/PT results to provincial epidemiologists was 26–36 days.

ogists. For Salmonella, it took a median of 7 days to send isolates for PT, 5 days to receive the PT results, and another 9 days (IQ range = 2–22 days) to report PT results to provincial epidemiologists. The total median time from onset of symptoms to report PFGE/PT results to provincial epidemiologists was 26–36 days. In the epidemiological surveillance stream, it took a median of 3–5 days from sample collection to the local laboratory reporting the result to the regional health authority (Table 1). The median time to enter the case report into the electronic database (0 days), to make a first attempt to interview (1 day), and to complete the interview (0 days) was negligible. However, the interval range shows that it could take up to 20 days to enter a case into the electronic database, 62 days to make a first attempt to interview a case, and 43 days to complete the interview. These ranges were highest for Salmonella cases. The total median time from onset of symptoms to completed interview was 12–14 days.

erval range shows that it could take up to 20 days to enter a case into the electronic database, 62 days to make a first attempt to interview a case, and 43 days to complete the interview. These ranges were highest for Salmonella cases. The total median time from onset of symptoms to completed interview was 12–14 days. In general, it took more time for cases in BC to collect a sample but less time for samples to arrive and to be confirmed/serotyped at the PHL when compared to findings produced by researchers in other countries (Table 2) [3–8]. The time from sample receipt at the PHL to reporting of PFGE results to provincial epidemiologists was much longer in BC (17–33 days) than it is in some US states (4-5 days) [8]. The time from sample collection to notifying the regional health authority was a little shorter in BC, as was the time to first attempt to interview. Overall, there was a comparable interval from symptom onset to interview between BC and other jurisdictions. 4. Discussion We present the first published Canadian evaluation of the timeliness of enteric disease surveillance. Understanding these time intervals has helped interpret surveillance findings (e.g., know the delay between case reporting by regional health authorities and expected lab results) and has been used during BC outbreak investigations to assess whether and when more cases or information on cases may be expected.

eillance. Understanding these time intervals has helped interpret surveillance findings (e.g., know the delay between case reporting by regional health authorities and expected lab results) and has been used during BC outbreak investigations to assess whether and when more cases or information on cases may be expected. The longest interval was from PFGE run or PT result receipt at PHL to PFGE/PT report to provincial epidemiologists (median 8–10 days). This is in large part due to batching of PFGE and PT results which are sent to provincial epidemiologists only once a week. Individual results may be shared more rapidly during outbreak investigations. The reasons for the long interquartile range (up to 22 days) and overall ranges (up to 57 days) include repeating failed PFGE runs, awaiting NML confirmation or designation and prioritization of other laboratory work. The interval from PHL confirmation to PFGE run is about 1 week because isolates are accumulated and analysed in batches rather than immediately due to resource limitations. The interval from onset to sample collection (3–7 days) is not highly modifiable. The longest epidemiological interval was from sample collection to notification of the regional health authority (3–5 days), which includes the diagnosis at the local laboratory; this was comparable to findings by other authors (Table 2) [4, 6].

erval from onset to sample collection (3–7 days) is not highly modifiable. The longest epidemiological interval was from sample collection to notification of the regional health authority (3–5 days), which includes the diagnosis at the local laboratory; this was comparable to findings by other authors (Table 2) [4, 6]. Once the regional health authority is notified of an enteric disease, reporting and interviews are conducted quickly, on median. In BC, it is recommended that these four diseases be reported provincially and cases be contacted for interview within 24 hours of notification [10]. Our findings indicate that it can take up to 2 months for the first attempt to interview. Delays may be due to prioritization of other disease investigations or other activities, understaffing and waiting for confirmatory test results from the PHL prior to interview. Since the epidemiological data on cases is collected much faster than the isolate subtyping data, it is important that epidemiological information be standardized and compared routinely to enable early identification of outbreaks. In BC, cases are interviewed using standard CRF and are reported regionally and provincially via electronic public health information systems. Statistical algorithms are run weekly to detect aberrations [2]. In addition, many outbreaks are identified via reporting of gastrointestinal syndromes among people who attended common events.

e interviewed using standard CRF and are reported regionally and provincially via electronic public health information systems. Statistical algorithms are run weekly to detect aberrations [2]. In addition, many outbreaks are identified via reporting of gastrointestinal syndromes among people who attended common events. We found six studies which used comparable methods to measure the timeliness of surveillance in North America and Europe in the last decade [3–8]. Most of our findings were comparable to these other jurisdictions. Where they could be measured and compared, the times from symptom onset to completed interview (12–14 days in the epidemiological stream) and from symptom onset to reporting of laboratory results to epidemiologists (28 days for Salmonella PT in the laboratory stream) were very similar. One important difference was noted in the time from receipt of an isolate at the PHL to reporting PFGE results to provincial epidemiologists (medians of 17–33 days in BC). In a subset of US states, it takes a median of 4-5 days from isolate receipt at the state laboratory to PFGE upload into PulseNet where it is accessible by other laboratories for comparison and detection of clusters [8]. US epidemiologists in the same states also have direct real-time access to PulseNet PFGE results via SEDRIC (System for Enteric Disease Response, Investigation and Coordination) in the same 4-5 day interval (personal communications: G Biggerstaff, CDC, May 23, 2016) [11].

s for comparison and detection of clusters [8]. US epidemiologists in the same states also have direct real-time access to PulseNet PFGE results via SEDRIC (System for Enteric Disease Response, Investigation and Coordination) in the same 4-5 day interval (personal communications: G Biggerstaff, CDC, May 23, 2016) [11]. Although many intervals are nonmodifiable, others are. Authors agree that the most effective way to improve the timeliness of surveillance is through increasing electronic access to, or transmission of, data [12–14].

s for comparison and detection of clusters [8]. US epidemiologists in the same states also have direct real-time access to PulseNet PFGE results via SEDRIC (System for Enteric Disease Response, Investigation and Coordination) in the same 4-5 day interval (personal communications: G Biggerstaff, CDC, May 23, 2016) [11]. Although many intervals are nonmodifiable, others are. Authors agree that the most effective way to improve the timeliness of surveillance is through increasing electronic access to, or transmission of, data [12–14]. In order to improve the timeliness of enteric disease surveillance in BC, several improvements were made or planned following this study. For the laboratory surveillance stream, a pilot project will be set up to enable daily automated electronic data transmission of PFGE and PT results from the PHL to the provincial epidemiologists. In order to liberate resources to more rapidly conduct PFGE when it is most useful, the PHL has implemented an algorithm in which only the most common serotypes (e.g., E. coli O157:H7) and clusters of rare serotypes (e.g., S. Montevideo) undergo routine PFGE. Epidemiologists will continue to investigate clusters of rare species and serotypes and the PHL will conduct subtyping of these isolates once a cluster is identified. Timeliness may also improve with the advent of routine whole genome sequencing of enteric pathogens, which started in 2017 in Canada. For the epidemiological stream, public health officials inform the PHL when they identify outbreaks based on epidemiological data to prioritize the testing of outbreak isolates. Finally, all five regional health authorities took measures to decrease the range of time to first attempt to interview cases. These include increased staff awareness, dedicated data entry clerks, active and timely handover of case files, and routine audits.

ological data to prioritize the testing of outbreak isolates. Finally, all five regional health authorities took measures to decrease the range of time to first attempt to interview cases. These include increased staff awareness, dedicated data entry clerks, active and timely handover of case files, and routine audits. Although any improvements in timeliness would be helpful, it is reasonable to expect that, following our improvements, the median time from symptom onset to completed interview should remain stable at 14 days and that the median time from sample receipt at the PHL to reporting of PFGE to provincial epidemiologists should decrease by about a week to 21 days. Ultimately, improving timeliness of surveillance should improve our ability to detect and solve outbreaks. This will be measured using validated metrics such as the proportion of outbreaks solved and the time from first case to initiating an outbreak investigation [2, 15, 16].

hould decrease by about a week to 21 days. Ultimately, improving timeliness of surveillance should improve our ability to detect and solve outbreaks. This will be measured using validated metrics such as the proportion of outbreaks solved and the time from first case to initiating an outbreak investigation [2, 15, 16]. There were some limitations to this work. Due to negative time intervals or intervals greater than 150 days, 349 dates were manually corrected and 335 dates were excluded because they could not be validated. These errors and missing data occurred randomly in all datasets and variables and would not have impacted results substantially. Overall, 29.0% of cases did not have complete data for all time periods of interest. However, since we measured the laboratory and epidemiological streams separately, this had minimal impact on the results. There were several outliers, particularly for date of onset; negative time intervals were corrected or discarded and positive outliers were kept. This may have led to a slight overestimation of intervals. Since dates from local laboratories were not available, we could not measure the time it took for local laboratories to diagnose and report cases. We were not able to identify and exclude outbreak cases; it is possible that such cases and isolates are processed faster [5].

e led to a slight overestimation of intervals. Since dates from local laboratories were not available, we could not measure the time it took for local laboratories to diagnose and report cases. We were not able to identify and exclude outbreak cases; it is possible that such cases and isolates are processed faster [5]. There were several international publications on the timeliness of Salmonella surveillance in the last decade but few on other diseases [3, 4, 6–8]. Some of the time intervals were not directly comparable to other studies as different researchers are interested in and report on different time intervals. This is somewhat surprising given that enteric disease surveillance systems in developed countries are very similar. This may be because different jurisdictions have different diagnostic and reporting processes (e.g., different laboratories conduct different parts of the subtyping; reporting can be paper-based or electronic). Also, authors only measure or report on intervals that they are interested in or able to modify and they use different nomenclature to identify intervals. We agree with Jajosky and Groseclose that a more detailed and standardized approach, including a definition of intervals and use of the same intervals where possible, to evaluate surveillance timeliness would assist in comparability [12]. In addition, we encourage others to measure and report on the smallest intervals possible to facilitate comparison.

se that a more detailed and standardized approach, including a definition of intervals and use of the same intervals where possible, to evaluate surveillance timeliness would assist in comparability [12]. In addition, we encourage others to measure and report on the smallest intervals possible to facilitate comparison. The US FoodCORE initiative is an attempt to set such standards at a national level [17]. Launched in 2011, 11 FoodCORE sites work together to develop improved methods to detect, investigate, and control multistate foodborne disease outbreaks. Their purposeful selection and ongoing measurement of time intervals have improved the timeliness of surveillance in participating sites [8]. Our own evaluation will be repeated after the full implementation of changes, when we also hope to see improved timeliness. 5. Conclusions The measurement of the timeliness of enteric disease surveillance in BC enhanced our understanding of the surveillance process. Our findings were comparable to international benchmarks except for a considerably longer time to reporting of PFGE results. Several changes are underway to improve timeliness, including real-time electronic access to subtyping results. These will lay the foundation for other innovations such as the inclusion of genomic results and their integration with epidemiological findings to further improve our ability to identify and solve outbreaks.

Several changes are underway to improve timeliness, including real-time electronic access to subtyping results. These will lay the foundation for other innovations such as the inclusion of genomic results and their integration with epidemiological findings to further improve our ability to identify and solve outbreaks. Acknowledgments The authors are grateful for the contributions of Greg Tone, Environmental Health Officer, with Northern Health Authority, who passed away during the preparation of this manuscript. They also wish to acknowledge the work of the BC diagnostic laboratories, the Bacteriology team at the BCCDC PHL, and the Enteric Diseases Section at the NML. They also wish to thank Dr. Swastika Narayan for conducting the initial literature review and Dr. Celine Nadon for manuscript review. Conflicts of Interest The authors declare that there are no conflicts of interest regarding the publication of this paper. Figure 1 Enteric disease surveillance process and time intervals measured, British Columbia. PH: public health, PHL: public health laboratory, PFGE: pulsed-field gel electrophoresis, PT: phage type. Table 1 Time intervals between steps in the surveillance of enteric diseases, BC, 2012-13.

Conflicts of Interest The authors declare that there are no conflicts of interest regarding the publication of this paper. Figure 1 Enteric disease surveillance process and time intervals measured, British Columbia. PH: public health, PHL: public health laboratory, PFGE: pulsed-field gel electrophoresis, PT: phage type. Table 1 Time intervals between steps in the surveillance of enteric diseases, BC, 2012-13. Time interval1 Salmonella median [IQ] (range) in days STEC median [IQ] (range) in days Shigella median [IQ] (range) in days Listeria median [IQ] (range) in days Laboratory surveillance 1: symptom onset to sample collection 6 [3–11] (0–163) 4 [3–8] (0–108) 7 [4–13] (0–137) 3 [1–4] (0–13) 2: sample collection to sample receipt at PHL 5 [3–6] (0–36) 4 [2–6] (0–46) 4 [3–5] (0–20) 3 [2–4] (0–7) 3: sample receipt at PHL to PHL confirmation (incl. serotype or species) 3 [2–4] (1–70) 6 [2–21] (1–173) 1 [1–3] (1–30) 6 [3–8] (2–69) 4: PHL confirmation to start of PFGE run 6 [3–8] (0–55) 6 [5–10] (1–83) 8 [5–12] (0–34) 6 [5–8] (0–13) 5: start of PFGE run to report of PFGE result to provincial epidemiologists 10 [8–22] (1–57) 8 [5–10] (1–44) 9 [7–14] (2–49) NA2 6: PHL confirmation to isolate sent for PT 7 [3–8] (0–27) NA NA NA 7: isolate sent for PT to receipt of PT result at PHL 5 [5-6] (2–26) NA NA NA 8: receipt of PT result at PHL to report of PT result to provincial epidemiologists 9 [2–22] (0–38) NA NA NA A1: symptom onset to report of results to provincial epidemiologists 35 [25–48] (15–179) 26 [22–34] (13–99) 36 [27–49] (18–165) NA2

A 7: isolate sent for PT to receipt of PT result at PHL 5 [5-6] (2–26) NA NA NA 8: receipt of PT result at PHL to report of PT result to provincial epidemiologists 9 [2–22] (0–38) NA NA NA A1: symptom onset to report of results to provincial epidemiologists 35 [25–48] (15–179) 26 [22–34] (13–99) 36 [27–49] (18–165) NA2 Epidemiological surveillance 9: sample collection to regional health authority notification 4 [3–6] (0–18) 5 [3–7] (0–30) 4 [3–6] (0–35) 3 [3-4] (1–8) 10: regional health authority notification to case report entry 0 [0-0] (0–20) 0 [0-0] (0–2) 0 [0-0] (0–3) 0 [0-0] (0-0) 11: case report entry to first attempt to contact 1 [0–2] (0–62) 0 [0-1] (0–19) 0 [0-1] (0–31) 1 [1–5] (0–55) 12: first attempt to contact to completed interview 0 [0-1] (0–43) 0 [0-1] (0–38) 0 [0-1] (0–29) 0 [0–3] (0–6) B1: symptom onset to completed interview 14 [9–21] (2–168) 12 [8–18] (3–119) 14 [10–21] (5–148) 13 [8–17] (3–61) 1Numbers refer to intervals in Figure 1. 2Dates for reporting Listeria PFGE to provincial epidemiologists were not available (not in regular weekly report). Table 2 Comparison of time intervals between steps in the surveillance of enteric diseases between BC (2012-13) and other regions1.

Epidemiological surveillance 9: sample collection to regional health authority notification 4 [3–6] (0–18) 5 [3–7] (0–30) 4 [3–6] (0–35) 3 [3-4] (1–8) 10: regional health authority notification to case report entry 0 [0-0] (0–20) 0 [0-0] (0–2) 0 [0-0] (0–3) 0 [0-0] (0-0) 11: case report entry to first attempt to contact 1 [0–2] (0–62) 0 [0-1] (0–19) 0 [0-1] (0–31) 1 [1–5] (0–55) 12: first attempt to contact to completed interview 0 [0-1] (0–43) 0 [0-1] (0–38) 0 [0-1] (0–29) 0 [0–3] (0–6) B1: symptom onset to completed interview 14 [9–21] (2–168) 12 [8–18] (3–119) 14 [10–21] (5–148) 13 [8–17] (3–61) 1Numbers refer to intervals in Figure 1. 2Dates for reporting Listeria PFGE to provincial epidemiologists were not available (not in regular weekly report). Table 2 Comparison of time intervals between steps in the surveillance of enteric diseases between BC (2012-13) and other regions1. Time interval Salmonella median in days STEC median in days Shigella median in days Listeria median in days BC Int'l BC Int'l BC Int'l BC Int'l Laboratory surveillance 1: symptom onset to sample collection 6 4 [3, 4] 4 3 [3] 7 2 [3] 3 NA 2: sample collection to sample receipt at PHL 5 6 [8] 4 5 [8] 4 NA 3 6 [8] 3: sample receipt at PHL to PHL confirmation (incl. serotype or species) 3 5 [4] 4 [8] 6 [7] 6 5 [8] 1 NA 6 NA A2: symptom onset to start of PFGE 21 18 [3] 19 15 [3] 22 21 [3] NA NA A3: sample collection to report of PT results to provincial epidemiologists 28 25 [4] NA NA NA NA NA NA A4: sample receipt at PHL to report of PFGE result to provincial epidemiologists 23 5 [8] 17 4 [8] 33 NA NA 4 [8]

] 4 [8] 6 [7] 6 5 [8] 1 NA 6 NA A2: symptom onset to start of PFGE 21 18 [3] 19 15 [3] 22 21 [3] NA NA A3: sample collection to report of PT results to provincial epidemiologists 28 25 [4] NA NA NA NA NA NA A4: sample receipt at PHL to report of PFGE result to provincial epidemiologists 23 5 [8] 17 4 [8] 33 NA NA 4 [8] Epidemiological surveillance 9: sample collection to local/regional health authority notification 4 6 [4] 6 [6] 5 4 [6] 4 6 [6] 3 5 [6] B1: symptom onset to completed interview 14 14 [3] 12 12 [3] 14 14 [3] 13 NA B2: symptom onset to local/regional health authority notification 11 9 [3] 10 7 [3] 11 [5] 13 8 [3] 7 NA B3: local/regional health authority notification to first attempt to contact 1 2 [8] 1 2 [8] NA NA 42 2 [8] 1Number in brackets refer to the associated references. 2Only 8 cases.

1. Introduction Actinomycosis is a chronic bacterial infection, suppurative and granulomatous in nature, caused by bacteria of the genus Actinomyces [1], a group of Gram-positive anaerobic bacteria that form filamentous microcolonies [2], do not form spores, measure up to 1 μm diameter, and are slow-growing [1]. Actinomycosis is an uncommon condition whose symptomatology imitates some malignant pelvic tumours, tuberculosis, or nocardiosis because it spreads progressively and continuously [3]. This pathology invades tissue layers, causing the formation of abscesses and fistulae. Its diagnosis is difficult, and it results in increased morbimortality. Actinomyces belong to the phylum Actinobacteria and to the order Actinomycetales. Hundreds of Actinomyces species exist, most of which inhabit the soil. Others are associated with plants, which participate in nitrogen fixation, and a few species live in human beings as saprophytic bacteria [2]. It should be highlighted that most Actinomyces spp. are present in microbiota, chiefly inhabiting the oropharynx, gastrointestinal tract, and urogenital tract [3].

bit the soil. Others are associated with plants, which participate in nitrogen fixation, and a few species live in human beings as saprophytic bacteria [2]. It should be highlighted that most Actinomyces spp. are present in microbiota, chiefly inhabiting the oropharynx, gastrointestinal tract, and urogenital tract [3]. Actinomycoses are opportunistic chronic infections [4], as Actinomyces have a low potential for virulence in connection with fimbriae. Therefore, they require normal mucosal barriers to be altered through trauma, surgery, or an infection. In this way, they cross the mucosal membrane or epithelial surface [4–6]. For example, a pulmonary infection can be caused by bronchoaspiration [5, 7], or a pelvic infection can originate from the use of an intrauterine device (IUD), which can injure or perforate the mucosal membrane of the uterus and facilitate infection [3].

way, they cross the mucosal membrane or epithelial surface [4–6]. For example, a pulmonary infection can be caused by bronchoaspiration [5, 7], or a pelvic infection can originate from the use of an intrauterine device (IUD), which can injure or perforate the mucosal membrane of the uterus and facilitate infection [3]. Currently, various clinical characteristics of actinomycosis have been described, and the bacterium has been observed in various anatomical sites (e.g., face, bones and articulations, respiratory tract, urogenital tract, digestive tract, central nervous system, skin, and soft tissue structures). The most frequent clinical form of the disease is cervicofacial actinomycosis, representing approximately 60% of all reported cases, and is associated with odontogenic infection. Other clinical types include thoracic actinomycosis, the third most common type of actinomycosis, which includes pulmonary, bronchial, and laryngeal actinomycosis [3], and abdominal actinomycosis, where the appendix, caecum, and colon are the most common sites of infection. Actinomycosis of the central nervous system is located chiefly in the cerebral abscess. Actinomycosis of the urogenital tract is the second most common clinical form of actinomycosis, and the principal clinical presentation is pelvic actinomycosis [3, 5, 8].

ix, caecum, and colon are the most common sites of infection. Actinomycosis of the central nervous system is located chiefly in the cerebral abscess. Actinomycosis of the urogenital tract is the second most common clinical form of actinomycosis, and the principal clinical presentation is pelvic actinomycosis [3, 5, 8]. Pelvic actinomycosis can affect any age group, with no preference for occupation or season and is secondary to perforation or fistulation [4]. Other possible causes include bacterial vaginosis, which fosters an anaerobic environment and is associated with other microorganisms [51]; the presence of tumours [66]; and the use of IUDs [3–5]. The possibility of a contagion through oral sex has been considered because these bacteria are part of the oral cavity microbiota [72]. One possible route of dissemination is through IUDs, which fosters the growth of microorganisms through wires that are left in the exocervix. In addition, the IUD changes the carbohydrate metabolism in endometrial cells, fostering still more inflammation. Another probable route is the perineum, where the microorganisms could extend from the anus up through the cervicovaginal zone [4]. The most common aetiological agent is Actinomyces israelii [5, 73]. Other reported species include A. naeslundii, A. viscosus, A. odontolyticus, A. pyogenes, A. urogenitalis, and A. turicensis [72, 74, 75].

Pelvic actinomycosis can affect any age group, with no preference for occupation or season and is secondary to perforation or fistulation [4]. Other possible causes include bacterial vaginosis, which fosters an anaerobic environment and is associated with other microorganisms [51]; the presence of tumours [66]; and the use of IUDs [3–5]. The possibility of a contagion through oral sex has been considered because these bacteria are part of the oral cavity microbiota [72]. One possible route of dissemination is through IUDs, which fosters the growth of microorganisms through wires that are left in the exocervix. In addition, the IUD changes the carbohydrate metabolism in endometrial cells, fostering still more inflammation. Another probable route is the perineum, where the microorganisms could extend from the anus up through the cervicovaginal zone [4]. The most common aetiological agent is Actinomyces israelii [5, 73]. Other reported species include A. naeslundii, A. viscosus, A. odontolyticus, A. pyogenes, A. urogenitalis, and A. turicensis [72, 74, 75]. The symptoms of pelvic actinomycosis associated with the use of an IUD can imitate symptoms of gynaecological malignant tumours, uterine myoma, or adenomyosis when presenting as a genital mass without fever [3]. The infection can disseminate to the uterine tubes and can cause salpingitis and the subsequent destruction of the ovarian parenchyma [4]. Organs such as the bladder, ileocaecal (iliac fossa) and rectosigmoid region, colon, urethra, and extension to the skin have been reportedly affected in various published cases.

3]. The infection can disseminate to the uterine tubes and can cause salpingitis and the subsequent destruction of the ovarian parenchyma [4]. Organs such as the bladder, ileocaecal (iliac fossa) and rectosigmoid region, colon, urethra, and extension to the skin have been reportedly affected in various published cases. The diagnosis of pelvic actinomycosis is obtained using various techniques because culturing Actinomyces spp. presents difficulties and also depends on the skill and access to equipment necessary to perform it.First, the signs and symptoms of the patients are considered and can point to a possible abdominal infection, vaginitis, abscess, or possible tumour-forming process. The most common symptoms are weight loss, nonspecific abdominal or pelvic pain, breakthrough bleeding or abundant vaginal flow, and, on rare occasions, fever [3, 4, 51]. Upon medical exploration, the affected zone is palpated to detect hard masses, and a gynaecological exam is performed to check for inflammation of the vaginal mucous membrane, yellowish secretion with a bad smell, or some visible damage to the mucous membrane [4, 51]. In laboratory studies, it is possible to identify leucocytosis, erythropaenia, and high sedimentation rate; high values of C-reactive protein; and tumour marker values within the reference ranges or slightly elevated like Ca 125 (Alpha-fetoprotein), and cancer antigen 15–3 [3, 4, 51].

Upon medical exploration, the affected zone is palpated to detect hard masses, and a gynaecological exam is performed to check for inflammation of the vaginal mucous membrane, yellowish secretion with a bad smell, or some visible damage to the mucous membrane [4, 51]. In laboratory studies, it is possible to identify leucocytosis, erythropaenia, and high sedimentation rate; high values of C-reactive protein; and tumour marker values within the reference ranges or slightly elevated like Ca 125 (Alpha-fetoprotein), and cancer antigen 15–3 [3, 4, 51]. Diagnostic images, such as computed tomography, magnetic resonance, ultrasound, X-rays, and laparoscopy are helpful, as they can be used to observe the affected zone, such as a tumour-forming mass that can induce either actinomycosis or a carcinogenic process [4, 51, 73].

In laboratory studies, it is possible to identify leucocytosis, erythropaenia, and high sedimentation rate; high values of C-reactive protein; and tumour marker values within the reference ranges or slightly elevated like Ca 125 (Alpha-fetoprotein), and cancer antigen 15–3 [3, 4, 51]. Diagnostic images, such as computed tomography, magnetic resonance, ultrasound, X-rays, and laparoscopy are helpful, as they can be used to observe the affected zone, such as a tumour-forming mass that can induce either actinomycosis or a carcinogenic process [4, 51, 73]. In most cases, histological visualisation of biopsy or aspirated samples is employed, where bacilli in the tissue with their typical ramifications, such as in interconnected breasts, are observed. Cervicovaginal cells are collected for Papanicolaou (Pap) staining. In both cases, they are reported as Microorganisms Similar to Actinomyces (MSA) [4, 51]. In many cases, the diagnosis is made a posteriori through a histological examination of samples obtained surgically during laparotomy or laparoscopy, but rarely in a preoperative manner. Histological studies of tissues show inflammatory changes of suppurative and granulomatous nature, connective proliferation, and sulphur granules, which have also been identified in infections caused by Nocardia brasiliensis, Actinomadura madurae, and Staphylococcus aureus. These granules are particles of yellowish colour, which, when viewed by the naked eye, are formed by groups of filamentous Actinomyces surrounded by neutrophils [73].

on, and sulphur granules, which have also been identified in infections caused by Nocardia brasiliensis, Actinomadura madurae, and Staphylococcus aureus. These granules are particles of yellowish colour, which, when viewed by the naked eye, are formed by groups of filamentous Actinomyces surrounded by neutrophils [73]. Two methods exist for completely identifying the causal agent: culture and identification through biochemical tests and identification through sequencing of the 16S rRNA segment, which offers greater precision. Although these methods are very efficient, they are not well reported in the literature due to the conditions under which they must be performed, requiring an anaerobic culture environment and the necessary equipment, which is costly. The usual treatment for actinomycosis consists of high and prolonged doses of penicillin G (20 million units per day) or amoxicillin for 4 to 6 weeks, followed by penicillin V (4 g per day) orally for 6 to 12 months. Clindamycin, tetracycline, and erythromycin are an alternative in cases of allergy to penicillin [4, 5]. In addition to these medicines, it has been observed that Actinomyces is also sensitive to third-generation cephalosporins, ciprofloxacin, trimethoprim-sulfamethoxazole, and rifampicin [4]. However, the elimination of the injured tissue and surgical drainage are necessary measures in some cases [5], and, in these patients, the duration of antimicrobial therapy could be reduced (3 months) [3].

so sensitive to third-generation cephalosporins, ciprofloxacin, trimethoprim-sulfamethoxazole, and rifampicin [4]. However, the elimination of the injured tissue and surgical drainage are necessary measures in some cases [5], and, in these patients, the duration of antimicrobial therapy could be reduced (3 months) [3]. In the review performed by Martínez et al. [74], it is mentioned that there are reports of the presence of Actinomyces in secretions starting from 1877, recorded by Harz. Beedham et al. found that the first reports of intrauterine actinomycosis related to IUDs appeared in the 1920s [76]. Clinical cases of pelvic actinomycosis have been reported in Africa, Oceania, Asia, Europe, and America. However, as pelvic actinomycosis is an uncommon infection, no epidemiological studies have been conducted to determine its prevalence or incidence.

rauterine actinomycosis related to IUDs appeared in the 1920s [76]. Clinical cases of pelvic actinomycosis have been reported in Africa, Oceania, Asia, Europe, and America. However, as pelvic actinomycosis is an uncommon infection, no epidemiological studies have been conducted to determine its prevalence or incidence. 2. Materials and Methods We performed this analysis according to the guidelines of the Meta-Analysis of Observational Studies in Epidemiology Group (MOOSE). A systematic review of worldwide cases of pelvic actinomycosis between the years 1980 and 2014 was performed. Studies that described clinical cases of pelvic actinomycosis with a detailed diagnostic method and cross-sectional studies of cases of actinomycosis available in the PubMed, Scopus, and Google Scholar databases were included using the following keywords: Pelvic actinomycosis, Actinomycosis pélvica, “Actinomycosis prevalence”, “Prevalencia de actinomycosis”, “Actinomyces” AND “female genital tract” and combinations of these terms. The use of quotes was avoided when searching for the terms Pelvic actinomycosis and Actinomycosis pélvica in order to increase the search results.

c actinomycosis, Actinomycosis pélvica, “Actinomycosis prevalence”, “Prevalencia de actinomycosis”, “Actinomyces” AND “female genital tract” and combinations of these terms. The use of quotes was avoided when searching for the terms Pelvic actinomycosis and Actinomycosis pélvica in order to increase the search results. Abstracts of articles identified to be relevant for the objective of this paper were read; studies whose abstract or full text was unavailable were automatically excluded. When an abstract complied with inclusion criteria, the full text was analysed. Case reports that lacked a diagnostic method and a final diagnosis of pelvic actinomycosis were excluded. Studies published in a language that was not English, Spanish, French, or Portuguese were not included. The following information was extracted and analysed from the compiled studies: year, country, type of study, number of cases, prior use of IUD and duration, initial diagnosis, treatment, definitive diagnosis, and method of definitive diagnosis (Figure 1). 3. Results The search yielded a total of 3852 studies; 3693 were excluded from the title, abstract, and language screening; 96 more were excluded for not being available in full text format and for not meeting the selection criteria when reading the full article. A total of 63 studies including 86 case reports of pelvic actinomycosis, along with 8 cross-sectional studies of reports examining populations for cases of Microorganisms Similar to Actinomyces (MSA), were included for this review (Figure 1).

3. Results The search yielded a total of 3852 studies; 3693 were excluded from the title, abstract, and language screening; 96 more were excluded for not being available in full text format and for not meeting the selection criteria when reading the full article. A total of 63 studies including 86 case reports of pelvic actinomycosis, along with 8 cross-sectional studies of reports examining populations for cases of Microorganisms Similar to Actinomyces (MSA), were included for this review (Figure 1). 3.1. Cases of Pelvic Actinomycosis Reported in Africa From the African continent, 3 articles of clinical cases were found, totalling 8 clinical cases. The majority of patients were IUD users; however in most cases the type of IUD used was not disclosed. The pathology that was first diagnosed in these cases was an ovarian tumour. The method of diagnosis that was utilised to definitively diagnose patients with actinomycosis was histopathological reporting (Table 1). The most common treatments were hysterectomy, laparotomy, and antibiotic therapy. No follow-up data was presented. 3.2. Cases of Pelvic Actinomycosis Reported in Oceania From Oceania, 1 article was published that included 3 clinical cases with the following ages: 56, 70, and 37 years; two of them were copper IUD users. In the three cases, malignant lesions were initially diagnosed; the final diagnosis was performed postoperatively. Salpingo-oophorectomy along with antibiotic therapy was used in all the cases; patients fully recovered after treatment. (Table 2).

following ages: 56, 70, and 37 years; two of them were copper IUD users. In the three cases, malignant lesions were initially diagnosed; the final diagnosis was performed postoperatively. Salpingo-oophorectomy along with antibiotic therapy was used in all the cases; patients fully recovered after treatment. (Table 2). 3.3. Cases of Pelvic Actinomycosis Reported in Asia Fourteen articles of 16 clinical cases came from Asia, the age of the patients ranged between 25 and 86 years, and the average age was 45.6 years (SD 15.5). The majority of patients were IUD users, with a usage time of 1 year to more than 20 years; most of the studies did not specify the type of IUD used. However, cases in nonusers were also reported, despite the well-known relationship between IUD use and pelvic actinomycosis. The most common presumptive diagnostic was malignant lesions, while, in other cases, Crohn's disease and acute peritonitis were also suspected. The most utilised diagnostic method was histological reporting after surgical interventions, which were invasive in most cases, such as hysterectomy and salpingo-oophorectomy along with antibiotic therapy. Most of the patients had a full recovery or at least a significant improvement after follow-up; only a case of renal sequelae was reported. (Table 3).

was histological reporting after surgical interventions, which were invasive in most cases, such as hysterectomy and salpingo-oophorectomy along with antibiotic therapy. Most of the patients had a full recovery or at least a significant improvement after follow-up; only a case of renal sequelae was reported. (Table 3). 3.4. Cases of Pelvic Actinomycosis Reported in Europe Twenty clinical case report articles including 39 cases of pelvic actinomycosis originated from Europe, in which ages ranged from 18 to 65 years; average age was 40 years (SD 10.4). The cases principally included female IUD users, with a usage time ranging from 1.5 to 20 years; copper and multiload IUD were the most reported; however in most studies the type of IUD is not specified. The predominant presumptive diagnosis was malignant lesion; other suspected diagnoses included Crohn's disease, acute appendicitis, endometrial infection, pelvic inflammatory disease, and abscesses. Postoperative histopathological reports were the most common definitive diagnostic methods. Other methods of final diagnosis have also been reported, such as the Pap smear, culture, API 20A biochemical assays, and 16S rRNA sequencing techniques. The most common treatments used were damaged tissue excision, laparotomy, and salpingo-oophorectomy together with antibiotic therapy. The majority of the articles do not have follow-up information, nonetheless studies reporting patient follow-up stated that they fully recovered after treatment, and there is one report of death (Table 4).

atments used were damaged tissue excision, laparotomy, and salpingo-oophorectomy together with antibiotic therapy. The majority of the articles do not have follow-up information, nonetheless studies reporting patient follow-up stated that they fully recovered after treatment, and there is one report of death (Table 4). 3.5. Cases of Pelvic Actinomycosis Reported in America With regard to America, 16 articles with reports of 20 clinical cases exist. The ages of patients ranged between 18 and 58 years with an average age of 39.6 years (SD 9.9). All patients were IUD users except one case, and the time of device use ranged from 22 months to 33 years. Reported types of IUD include Dalkon Shield and Lippes loop. Tubo-ovarian and pelvic abscesses along with malignant lesions were the conditions with the greatest diagnostic confusion. Similar to the other summaries, the postsurgical histological reports were the most reported definitive diagnostic methods. Other methods were also utilised, such as culture, the 16S rRNA sequencing technique, haematoxylin-eosin staining microscopy (HE), and the IUD smear. Salpingo-oophorectomy and laparotomy along with prolonged antibiotic therapy were the most used therapeutic measures. After treatment most of the patients had a full or significant recovery. (Table 5).

d, such as culture, the 16S rRNA sequencing technique, haematoxylin-eosin staining microscopy (HE), and the IUD smear. Salpingo-oophorectomy and laparotomy along with prolonged antibiotic therapy were the most used therapeutic measures. After treatment most of the patients had a full or significant recovery. (Table 5). 3.6. Cross-Sectional Studies Eight cross-sectional studies of reports worldwide that examine populations for cases of actinomycosis or MSA were analysed. The prevalence of pelvic actinomycosis was low. Likewise, there is a strong relationship between the use of an IUD and the presence of MSA. In this type of report, the diagnosis methods reviewed were the Pap reports. However, it is important to emphasise that what is reported in these analyses are MSA. Only 3 articles reported actinomycosis as such, and only one report completely identified the causal agent through culture and biochemical assays (Table 6). 4. Discussion According to the analysis of the articles presented, Europe was the continent on which the greatest number of cases of pelvic actinomycosis was reported, followed by Asia and America. However, it is important to emphasise that this summary of information only gives us an approximation of the real epidemiology of this disease, as the cases presented in this article are only those reported. The youngest cases (18 years) are found in the European and American continents, and the oldest case (86 years) is found in the Asian continent.

asise that this summary of information only gives us an approximation of the real epidemiology of this disease, as the cases presented in this article are only those reported. The youngest cases (18 years) are found in the European and American continents, and the oldest case (86 years) is found in the Asian continent. Actinomycosis is an invasive infection that frequently imitates malignant processes in various anatomical zones. Pelvic actinomycosis involves one of the regions that is most often a source of diagnostic confusion. In this review, it is evident that, in many of the cases presented, an erroneous clinical diagnosis was made, confusing pelvic actinomycosis mainly with malignant lesions. Other common suspects were tubo-ovarian and pelvic abscesses and Crohn's disease. As such, as described by Kayikcioglu et al. [35] and Moniruddin et al. [77], pelvic actinomycosis should be considered in the differential diagnosis in any chronic inflammatory lesion of the viscera located in the pelvic zone to prevent a diagnostic error that could lead to unnecessary invasive treatment.

hn's disease. As such, as described by Kayikcioglu et al. [35] and Moniruddin et al. [77], pelvic actinomycosis should be considered in the differential diagnosis in any chronic inflammatory lesion of the viscera located in the pelvic zone to prevent a diagnostic error that could lead to unnecessary invasive treatment. The diagnosis of pelvic actinomycosis is difficult because it does not produce characteristic disease signs or symptoms. According to what was observed in this analysis, the most utilised diagnostic method in all continents was the histopathological report, which is commonly obtained after a surgical intervention due to an initial diagnostic error. This observation was also made by Purola and Paavonen [78] and more recently by Pérez-López et al. [73]. Other highly reported methods are the Pap test, which is generally reliable, but not unequivocal, as Actinomyces could be confused with similar organisms. Cases in which the causal microorganism is completely identified are few, as are cases identified by culture and biochemical assays. Valour et al. [3] mentioned that culturing bacteria in anaerobic medium is the cornerstone for diagnosing actinomycosis. However, this method requires a very controlled and precise technique and specific equipment. Identification by sequencing of the 16S rRNA segment is another technique that offers greater precision. Currently other authors such as Demirezen et al. [79] report the effectiveness of using specific primers to identify the most common Actinomyces species from patients' swabs samples; this technique is more accurate and faster than all the previous ones; the disadvantage is the high cost of reagents and the use of special equipment. However, because of the nature of the pathology, there is no early diagnosis, because, as has already been mentioned, the presence of symptoms occurs in advanced stages of the disease.

is more accurate and faster than all the previous ones; the disadvantage is the high cost of reagents and the use of special equipment. However, because of the nature of the pathology, there is no early diagnosis, because, as has already been mentioned, the presence of symptoms occurs in advanced stages of the disease. According to the observed reports, we conclude that the presentation of symptoms in pelvic actinomycosis is observed in an advanced period of the pathology, which does not include attack to the general state of health or fever, which is oriented towards an infectious pathology. The manifestations found are occupational masses in the pelvic-abdominal cavity that force the surgical procedures to be performed, and the diagnosis is made up to the time of the histopathological study.

does not include attack to the general state of health or fever, which is oriented towards an infectious pathology. The manifestations found are occupational masses in the pelvic-abdominal cavity that force the surgical procedures to be performed, and the diagnosis is made up to the time of the histopathological study. Pelvic actinomycosis is considered to be a rare and unusual disease, although the use of IUDs can promote its appearance. In the articles analysed, the greatest number of patients diagnosed with pelvic actinomycosis on all continents was IUD users, and the periods of use varied widely, from one year to long periods, such as 33 years. However, cases occurred mainly in users that wore IUDs for prolonged periods. Thus, based on experience and observation, it is recommended that IUDs be changed periodically to limit the occurrence of this condition. Some authors, such as Valour et al. [3], recommend changing the IUD every 5 years at a minimum, and others, such as Hernández et al. [4], recommend changes every 3 years. It should be emphasised that those cases in which patients were not IUD users were identified more recently. This observation could suggest that, despite the information that is available regarding the relationship between this condition and IUD use, the aetiology of pelvic actinomycosis could be due to other factors. Like all review studies, the main limitation of the study was the lack of data reported, another limitation was that not all articles were open access, and there were not enough subscriptions to the respective journals to access them.

Pelvic actinomycosis is considered to be a rare and unusual disease, although the use of IUDs can promote its appearance. In the articles analysed, the greatest number of patients diagnosed with pelvic actinomycosis on all continents was IUD users, and the periods of use varied widely, from one year to long periods, such as 33 years. However, cases occurred mainly in users that wore IUDs for prolonged periods. Thus, based on experience and observation, it is recommended that IUDs be changed periodically to limit the occurrence of this condition. Some authors, such as Valour et al. [3], recommend changing the IUD every 5 years at a minimum, and others, such as Hernández et al. [4], recommend changes every 3 years. It should be emphasised that those cases in which patients were not IUD users were identified more recently. This observation could suggest that, despite the information that is available regarding the relationship between this condition and IUD use, the aetiology of pelvic actinomycosis could be due to other factors. Like all review studies, the main limitation of the study was the lack of data reported, another limitation was that not all articles were open access, and there were not enough subscriptions to the respective journals to access them. During the first reports of this disease, greater numbers of cases were observed in developed countries, but, presently, reports of cases in developing countries and regions such as the Middle East, Southwest Asia, or Latin America are more common. This change could be because, in the first decades of its observation, this condition and its aetiology were unknown and prevention was difficult. However, with the advance of technology, preventative measures directed at high-risk populations in developed countries began to be applied, as opposed to developing countries, where no such actions were taken.

first decades of its observation, this condition and its aetiology were unknown and prevention was difficult. However, with the advance of technology, preventative measures directed at high-risk populations in developed countries began to be applied, as opposed to developing countries, where no such actions were taken. Conflicts of Interest The authors declare that there are no conflicts of interest regarding the publication of this paper. Figure 1 Search strategy: flow chart of literature research. Table 1 Publications of cases of pelvic actinomycosis in Africa. Definitive diagnosis Presumptive diagnosis Definitive method of diagnosis Previous use of IUD Treatment and resolution Number of cases Age (years) Reference Tumour mass formation caused by MSA∗ Ovarian tumour Histopathological report Yes (15 years) Type: ND∗∗ Laparotomy and hysterectomy, ampicillin No date of resolution 1 56 1989, Ben Nasr et al. Tunisia [9] Pelvic actinomycosis, MSA   Histopathological report No = 1 Yes = 4 Type: ND Total hysterectomy and bilateral oophorectomy Penicillin Resolution different according to patient 5 39.2 (average age) 2008, Chelli et al. Tunisia [10] MSA Ovarian tumour Histopathological report; direct study of right extracted ovary Yes Type: ND Oophorectomy, prolonged antibiotic therapy No date of resolution 2 ND 2010, Abid et al. Tunisia [11] Total cases 8 ∗MSA = microorganisms similar to Actinomyces. ∗∗ND = not disclosed. Table 2 Publications of cases of pelvic actinomycosis in Oceania.

MSA Ovarian tumour Histopathological report; direct study of right extracted ovary Yes Type: ND Oophorectomy, prolonged antibiotic therapy No date of resolution 2 ND 2010, Abid et al. Tunisia [11] Total cases 8 ∗MSA = microorganisms similar to Actinomyces. ∗∗ND = not disclosed. Table 2 Publications of cases of pelvic actinomycosis in Oceania. Definitive diagnosis Presumptive diagnosis Definitive method of diagnosis Previous use of IUD Treatment and resolution Number of cases Age (years) Reference Actinomycosis organisms Actinomyces sp. Ovarian cancer Malignant ovarian lesion Ovarian neoplasm Histopathological report Histopathological report, culture Yes = 1 Type: copper IUD No = 1 Yes = 1 Type: copper IUD Total abdominal elective hysterectomy and bilateral salpingo-oophorectomy, penicillin, and amoxicillin Complete recovery Laparotomy and bilateral salpingo-oophorectomy with left ureterolysis, ceftriaxone, and metronidazole Complete recovery Ceftriaxone and Metronidazole, subsequent laparotomy, left salpingo-oophorectomy, penicillin, and amoxicillin Complete recovery 3 56, 70, and 37 2014, Wan et al. Australia [12] Total cases 3 Table 3 Publications of cases of pelvic actinomycosis in Asia.

Definitive diagnosis Presumptive diagnosis Definitive method of diagnosis Previous use of IUD Treatment and resolution Number of cases Age (years) Reference Actinomycosis organisms Actinomyces sp. Ovarian cancer Malignant ovarian lesion Ovarian neoplasm Histopathological report Histopathological report, culture Yes = 1 Type: copper IUD No = 1 Yes = 1 Type: copper IUD Total abdominal elective hysterectomy and bilateral salpingo-oophorectomy, penicillin, and amoxicillin Complete recovery Laparotomy and bilateral salpingo-oophorectomy with left ureterolysis, ceftriaxone, and metronidazole Complete recovery Ceftriaxone and Metronidazole, subsequent laparotomy, left salpingo-oophorectomy, penicillin, and amoxicillin Complete recovery 3 56, 70, and 37 2014, Wan et al. Australia [12] Total cases 3 Table 3 Publications of cases of pelvic actinomycosis in Asia. Definitive diagnosis Presumptive diagnosis Definitive method of diagnosis Previous use of IUD Treatment and resolution Number of cases Age (years) Reference Actinomycotic abscesses (sulphur granules) Sigmoid colon cancer and tumour in left ovary Histopathological report Yes (1 year) Type: ND∗∗ Segmented resection of the sigmoid colon, elimination of the left distal ureter, the left ovary and Fallopian tube, ampicillin, and amoxicillin Complete recovery 1 36 1995, Kim et al. South Korea [13]

esses (sulphur granules) Sigmoid colon cancer and tumour in left ovary Histopathological report Yes (1 year) Type: ND∗∗ Segmented resection of the sigmoid colon, elimination of the left distal ureter, the left ovary and Fallopian tube, ampicillin, and amoxicillin Complete recovery 1 36 1995, Kim et al. South Korea [13] MSA∗ Crohn's disease or ovarian cancer or pelvic abscess associated with the IUD Histopathological report of the ovary Yes (14 years) Type: ND Laparotomy, total hysterectomy, bilateral salpingo-oophorectomy, and anterior resection No date of resolution 1 45 2009, Lim et al. Korea [14] MSA Tumour in the appendix Histopathological report ND ND 1 50 2010, Lee et al. South Korea [15] Pelvic actinomycosis (A. israelii) ND Sonography-guided transvaginal needle aspiration Yes (4 years) Type: multiload copper IUD Drainage, penicillin, and amoxicillin Complete recovery 1 38 1996, Anteby et al. Israel [16] Actinomyces, pelvic actinomycosis Peritoneal carcinomatosis Schiff and Grocott-Gomori acid tests Yes (10 years) Type: ND Incomplete tumourectomy, ileal resection, partial cystectomy, colostomy and bilateral ureterocutaneostomy, and penicillin Significant improvement 1 43 1999, Maeda et al. Japan [17] Pelvic actinomycosis Pelvic actinomycosis Cervical Papanicolaou Yes (21 years) Type: ND Ampicillin Almost complete recovery 1 51 2007, Nozawa et al. Japan [18] Pelvic actinomycosis Ovarian malignancy Gomori methenamine staining histopathology No Hysterectomy with bilateral salpingo-oophorectomy No date of resolution 1 74 2012, Ikeda and Kato Japan [19]

7.7 MSA Papanicolaou samples Cases without IUD (group 0: 128) 2 to 35 months IUD use (group 1: 121); 36 to 71 months (group 2: 112); more than 72 months (group 3: 107) MSA was found in 2 cases from group 2 and in 7 patients from group 3 MSA was 3.68 times more likely with greater use 1999, Garrido et al. Colombia [68] 1774 January 1996–January 2001 22–51 Actinomycosis Papanicolaou samples Yes = 671 oral contraceptive methods = 343 Other contraceptive methods = 32 No contraceptive use = 728 Actinomycosis in 13 patients with IUDs and in 2 without contraceptive methods 2002, Torres et al. Chile [69] 22   24–58 Genital actinomycosis Biopsy results Yes = 18 (3–19 years)   2003, Madrid et al. Chile [70] 200 ND 25–50 Actinomyces Vaginal secretion culture, Gram stain, Papanicolaou, API 20A biochemical assays Yes = 106 (3–10 years) No = 94 Actinomyces in 14 patients with IUDs, of which 2 had no symptoms of infection; species: Actinomyces israelii, Actinomyces naeslund ii, and Actinomyces odontolyticus 2002, Cano Ramos et al. Mexico [71] ∗MSA = microorganisms similar to Actinomyces. ∗∗ND = not disclosed.

Pelvic actinomycosis Pelvic actinomycosis Cervical Papanicolaou Yes (21 years) Type: ND Ampicillin Almost complete recovery 1 51 2007, Nozawa et al. Japan [18] Pelvic actinomycosis Ovarian malignancy Gomori methenamine staining histopathology No Hysterectomy with bilateral salpingo-oophorectomy No date of resolution 1 74 2012, Ikeda and Kato Japan [19] Puncture pyometra caused by Actinomyces Puncture pyometra Microscopic examination, Gram staining of the uterus and intraperitoneal pus, and culture No Emergency abdominal hysterectomy and bilateral salpingo-oophorectomy, cefmetazole, and meropenem No date of resolution 1 86 2013, Hagiya Japan [20] Inflammation caused by MSA Acute peritonitis due to perforated viscera Histopathological report of the abdominal wall Yes (20 years) Type: ND Laparotomy, resection, and penicillin Complete recovery 1 50 2008, Devendra and Chen Singapore [21] MSA Pelvic actinomycosis Papanicolaou, cervical culture and culture of IUD (without being able to be isolated), and histopathological report Yes (15 years) Type: ND Laparotomy, amoxicillin, and penicillin Significant recovery 1 40 2010, Fu and Tasi Taiwan [22] MSA, Actinomyces spp. Ovarian cancer Histopathological report and culture of purulent material No Laparotomy, hysterectomy, penicillin, and streptomycin Complete recovery 3 25, 31, and 35 2010, Munjal et al. India [23] Endometrial actinomycosis ND Histopathological report of endometrial samples No Augmentin and amoxicillin No date of resolution 1 52 2012, Sharma et al. India [24]

MSA, Actinomyces spp. Ovarian cancer Histopathological report and culture of purulent material No Laparotomy, hysterectomy, penicillin, and streptomycin Complete recovery 3 25, 31, and 35 2010, Munjal et al. India [23] Endometrial actinomycosis ND Histopathological report of endometrial samples No Augmentin and amoxicillin No date of resolution 1 52 2012, Sharma et al. India [24] Ovarian actinomycosis Ovarian cancer Histopathological report No Laparoscopy, hysterectomy with salpingo-oophorectomy, and penicillin Total recovery 1 39 2013, Vijaya et al. India [25] Pelvic actinomycosis, Actinomyces Ovarian cancer Histopathological report No Total hysterectomy with bilateral salpingo-oophorectomy No date of resolution 1 35 2013, Chalageri et al. India [26] Total cases 16 ∗MSA = microorganisms similar to Actinomyces. ∗∗ND = not disclosed. Table 4 Publications of cases of pelvic actinomycosis in Europe. Definitive diagnosis Presumptive diagnosis Definitive method of diagnosis Previous use of IUD Treatment and resolution Number of cases Age (years) Reference Actinomyces Crohn's disease Histopathological report of purulent material Yes (20 months) Type:  ND∗∗ Laparotomy, penicillin, and fusidic acid Complete recovery 1 19 1985, Spickett and Kipping England [27] MSA∗ Ovarian cancer with metastasis Histopathological report Yes (4 years) Type: ND Total abdominal hysterectomy and bilateral salpingo-oophorectomy, and penicillin No date of resolution 1 37 1997, Kirova et al. France [28]

Actinomyces Crohn's disease Histopathological report of purulent material Yes (20 months) Type:  ND∗∗ Laparotomy, penicillin, and fusidic acid Complete recovery 1 19 1985, Spickett and Kipping England [27] MSA∗ Ovarian cancer with metastasis Histopathological report Yes (4 years) Type: ND Total abdominal hysterectomy and bilateral salpingo-oophorectomy, and penicillin No date of resolution 1 37 1997, Kirova et al. France [28] Actinomycosis ND Papanicolaou Yes (15 years) Type: copper IUD Amoxicillin/clavulanic acid and ofloxacin Complete recovery 1 57 2013, Rajaonarison et al. France [29] Actinomyces israelii ND Culture of drained sample Yes (8 years) Type: copper IUD Drainage of abscess (colpoceliotomy), coamoxiclav Complete recovery 1 53 2013, Tholozan et al. France [30] MSA Inflammation caused by Actinomyces ND Histopathological report Yes (19 and 7 years) Type: ND Laparotomy, preoperative biopsy, resection of the tumour, resection of the necrotised tissues and partial cystectomy, hysterectomy, bilateral salpingo-oophorectomy, penicillin, and amoxicillin Complete recovery 2 48 and 52 2000, Pérez García et al. Spain [31] MSA Malignant tumour formation Histopathological report ND Laparotomy, resection of the central part of the epiploon and tumour formation, penicillin, and amoxicillin Significant improvement 1 30 2009, García Martínez et al. Spain [32] MSA ND Histological cervicovaginal observation and histopathological report Yes Type: ND Laparotomy, penicillin, and amoxicillin Complete recovery 2 33 and 35 2003, Bergenhenegouwen et al. Holland [33]

MSA Malignant tumour formation Histopathological report ND Laparotomy, resection of the central part of the epiploon and tumour formation, penicillin, and amoxicillin Significant improvement 1 30 2009, García Martínez et al. Spain [32] MSA ND Histological cervicovaginal observation and histopathological report Yes Type: ND Laparotomy, penicillin, and amoxicillin Complete recovery 2 33 and 35 2003, Bergenhenegouwen et al. Holland [33] Actinomyces ND Aspirate study Yes Type: ND ND 1 45 2005, Lely and Van Es Holland [34] MSA ND Histopathological report Yes (4–9 years) Type: copper IUD Penicillin, bacampicillin Complete recovery 5 32, 35, 44, 44, and 52 2005, Kayikcioglu et al. Turkey [35] Actinomycosis, MSA Tumour formation or abscess in ovary Histopathological report of the ovary Yes = 2 (15 and 6 years) No = 1 Type: ND Sulbactam-ampicillin, penicillin and ceftriaxone, laparotomy, drainage of abscesses salpingo-oophorectomy, and hysterectomy Complete recovery 3 32, 45, and 55 2009, Onal et al. Turkey [36] MSA ND Histopathological report Yes (8 years) Type: multiload copper IUD Extraction of a mass in the internal walls of the abdomen, penicillin Complete recovery 1 48 2010, Carkman et al. Turkey [37] Damage in the organs adjacent to the irregular mass, MSA ND Histopathological report Yes (16 years) Type: ND Laparotomy, total abdominal hysterectomy, bilateral salpingo-oophorectomy, and penicillin Complete recovery 1 48 2000, Yegüez et al. Turkey [38]

MSA ND Histopathological report Yes (8 years) Type: multiload copper IUD Extraction of a mass in the internal walls of the abdomen, penicillin Complete recovery 1 48 2010, Carkman et al. Turkey [37] Damage in the organs adjacent to the irregular mass, MSA ND Histopathological report Yes (16 years) Type: ND Laparotomy, total abdominal hysterectomy, bilateral salpingo-oophorectomy, and penicillin Complete recovery 1 48 2000, Yegüez et al. Turkey [38] Actinomycosis Pelvic inflammatory disease, rectal tumour Histopathological report Yes (12 years) Type: copper IUD Laparotomy, hysterectomy, bilateral salpingo-oophorectomy, appendectomy, lower anterior resection, Hartmann colostomy, penicillin, and amoxicillin Complete recovery 1 44 2012, Yilmaz et al. Turkey [39] Actinomycosis   Necropsy Yes (20 years) Type: ND Death due to sepsis 1 49 2007, Grabiec et al. Poland [40] MSA Acute appendicitis, fistulisation in abdominal wall Histopathological report No Laparotomy, right ileocolic resection with anastomosis of the ileotransverse colon, and amoxicillin No date of resolution 1 46 2008, Pitot et al. Belgium [41] Actinomyces spp. Carcinoma Purulent material culture, histopathological report Yes (3 years) Type: ND Right hemicolectomy, antibiotic therapy Complete recovery 1 35 2009, Čolović et al. Serbia [42] MSA, pseudoactinomycotic radiate granules Endometrial infection, ovarian abscess, and both Histopathological report Yes Type: ND 5 endometrial biopsies and 1 piece of hysterectomy No date of resolution 6 ND 2009, Boyle and McCluggage North Ireland [43]

Actinomyces spp. Carcinoma Purulent material culture, histopathological report Yes (3 years) Type: ND Right hemicolectomy, antibiotic therapy Complete recovery 1 35 2009, Čolović et al. Serbia [42] MSA, pseudoactinomycotic radiate granules Endometrial infection, ovarian abscess, and both Histopathological report Yes Type: ND 5 endometrial biopsies and 1 piece of hysterectomy No date of resolution 6 ND 2009, Boyle and McCluggage North Ireland [43] Actinomyces Muscular neoplasia Postsurgical histopathological report of samples from the abdominal wall abscess Yes Type: ND Laparotomy, adhesiolysis, complete excision of the mass with extensive damage to the anterior abdominal wall, and antibiotic therapy No date of resolution 1 47 2010, Acquaro et al. Italy [44] Anogenital actinomycosis, Actinomyces turicensis Perianal abscesses, pilonidal cyst, and gas gangrene API 20A biochemical assays and 16S rRNA sequencing technique ND ND 7 18, 18, 28, 23, 28, 33, and 65 2010, Chudáčková et al. Czech Republic [45] Actinomycosis Tumoural process in pelvis Histopathological report Yes Type: multiload copper IUD Cystoscopy, penicillin, and Duomox No date of resolution 1 42 2012, Maxová et al. Czech Republic [46] Total cases 39 ∗MSA = microorganisms similar to Actinomyces. ∗∗ND = not disclosed. Table 5 Publications of cases of pelvic actinomycosis in America. Definitive diagnosis Presumptive diagnosis Definitive method of diagnosis Previous use of IUD Treatment and resolution Number of cases Age (years) Reference

Actinomycosis Tumoural process in pelvis Histopathological report Yes Type: multiload copper IUD Cystoscopy, penicillin, and Duomox No date of resolution 1 42 2012, Maxová et al. Czech Republic [46] Total cases 39 ∗MSA = microorganisms similar to Actinomyces. ∗∗ND = not disclosed. Table 5 Publications of cases of pelvic actinomycosis in America. Definitive diagnosis Presumptive diagnosis Definitive method of diagnosis Previous use of IUD Treatment and resolution Number of cases Age (years) Reference Actinomyces israelii Tubo-ovarian abscess Histopathological report and culture Yes (4 years) Type: Dalkon Shield IUD Laparotomy, hysterectomy, bilateral salpingo-oophorectomy, and penicillin Complete resolution 1 29 1980, McLeod et al. United States [47] Actinomycotic tubo-ovarian abscess Tubo-ovarian abscess or malignant tumour Histopathological report Yes Type: ND∗∗ Antibiotic therapy, tumourectomy, and right salpingo-oophorectomy No date of resolution 1 29 1982, Kelly and Aaron United States [48] A. naeslundii Pelvic abscess Microscopic observation of the IUD and culture Yes (10 years) Type: Dalkon Shield IUD Antibiotic therapy hysterectomy, bilateral salpingo-oophorectomy Complete recovery 1 39 1985, Bonnez et al. United States [49]

Actinomycotic tubo-ovarian abscess Tubo-ovarian abscess or malignant tumour Histopathological report Yes Type: ND∗∗ Antibiotic therapy, tumourectomy, and right salpingo-oophorectomy No date of resolution 1 29 1982, Kelly and Aaron United States [48] A. naeslundii Pelvic abscess Microscopic observation of the IUD and culture Yes (10 years) Type: Dalkon Shield IUD Antibiotic therapy hysterectomy, bilateral salpingo-oophorectomy Complete recovery 1 39 1985, Bonnez et al. United States [49] Sulphur granules; Actinomyces israelii; actinomycotic pelvic abscess secondary to IUD involving the bladder, sigmoid colon, left ureter, liver, and superior abdominal wall ND Histopathological report and culture Yes (15 years) Type: Lippes loop IUD Percutaneous drainage and prolonged antibiotic therapy No date of resolution 1 41 1996, Hochsztein et al. United States [50] Actinomycotic granules, tubo-ovarian abscess Abdominal tumour secondary to colon cancer Laparotomy Yes Type: copper IUD Laparotomy, en bloc resection that included compromised abdominal wall, right hemicolectomy, hysterectomy, bilateral salpingo-oophorectomy, partial sigmoidectomy, and penicillin Complete recovery 1 47 1999, Mesa-Castillo et al. Colombia [51] MSA∗ ND Histopathological report Yes (9 and 3 years) Type: ND Laparotomy, hysterectomy, oophorectomy, penicillin, and amoxicillin Significant improvement 2 38 and 23 2006, Urbina et al. Colombia [52]

Actinomycotic granules, tubo-ovarian abscess Abdominal tumour secondary to colon cancer Laparotomy Yes Type: copper IUD Laparotomy, en bloc resection that included compromised abdominal wall, right hemicolectomy, hysterectomy, bilateral salpingo-oophorectomy, partial sigmoidectomy, and penicillin Complete recovery 1 47 1999, Mesa-Castillo et al. Colombia [51] MSA∗ ND Histopathological report Yes (9 and 3 years) Type: ND Laparotomy, hysterectomy, oophorectomy, penicillin, and amoxicillin Significant improvement 2 38 and 23 2006, Urbina et al. Colombia [52] Actinomyces sp. Bilateral cystic teratoma Histopathological report No Laparotomy, bilateral salphingo-oophorectomy, and penicillin Significant improvement 1 47 2001, Burlando et al. Argentina [53] Tubo-ovarian actinomycosis, MSA Tumour formation Histopathological report of the ovary Yes (8 years) Type: ND Oophorectomy, right salpingectomy, and amoxicillin No date of resolution 1 41 2005, Vispo et al. Argentina [54] Actinomycosis, MSA Vesical tumour Histopathological report Yes (33 years) Type: ND Penicillin Progressive improvement 1 58 2003, Alegría et al. Chile [55] Pelvic actinomycosis, sulphur granules Pelvic or neoplastic actinomycosis of the colon or ovary IUD swab Yes (27 years) Type: Lippes loop IUD Penicillin and amoxicillin Significant improvement 1 54 2013, Daniels et al. Chile [56] MSA ND Histopathological report of the right ovary Yes (22 years) Type: ND Laparotomy to drain purulent material, hysterectomy with bilateral salpingo-oophorectomy, and penicillin Significant improvement 1 48 2004, López-Cervantes et al. México [57]

Pelvic actinomycosis, sulphur granules Pelvic or neoplastic actinomycosis of the colon or ovary IUD swab Yes (27 years) Type: Lippes loop IUD Penicillin and amoxicillin Significant improvement 1 54 2013, Daniels et al. Chile [56] MSA ND Histopathological report of the right ovary Yes (22 years) Type: ND Laparotomy to drain purulent material, hysterectomy with bilateral salpingo-oophorectomy, and penicillin Significant improvement 1 48 2004, López-Cervantes et al. México [57] MSA Actinomycotic granuloma Uterine myomatosis Histopathological report Yes (10, 10, and 4 years) Type: ND Laparotomy, hysterectomy, salpingo-oophorectomy, and penicillin Significant improvement 3 36, 37, and 39 2005, Olivera-Reynada et al. Mexico [58] Coinfection by Neisseria gonorrhoeae and Actinomyces naeslundii ND Culture of UID and vaginal exudate Yes (22 months) Type: copper IUD Surgical excision of the appendix, bilateral salpingectomy No date of resolution 1 36 2013, Eiros-Bouza et al. Mexico [59] Urachal actinomycosis, “sulphur granules” Carcinoma Histopathological report ND Partial cystectomy Complete recovery 1 46 2013, Alfonso et al. Mexico [60] Actinomyces spp. Cyst in the left ovary and abscess in the iliac fossa Purulent material, histopathological report Yes (2 years) Type: ND Laparotomy, cefotaxime, metronidazole, and penicillin Gradual recovery 1 18 2004, Mejia et al. Mexico [61] Actinomyces urogenitalis ND Microscopic observation of the IUD, sequencing of the 16S rRNA gene Yes Type: ND Oral amoxicillin Complete recovery 1 38 2006, Elsayed et al. Canada [62]

Actinomyces spp. Cyst in the left ovary and abscess in the iliac fossa Purulent material, histopathological report Yes (2 years) Type: ND Laparotomy, cefotaxime, metronidazole, and penicillin Gradual recovery 1 18 2004, Mejia et al. Mexico [61] Actinomyces urogenitalis ND Microscopic observation of the IUD, sequencing of the 16S rRNA gene Yes Type: ND Oral amoxicillin Complete recovery 1 38 2006, Elsayed et al. Canada [62] Ovarian actinomycosis Left tubo-ovarian abscess Haematoxylin-eosin (HE) staining microscopy No Exploratory laparotomy, unilateral oophorectomy, and penicillin Complete recovery 1 49 2013, Bes et al. Brazil [63] Total cases 20 ∗MSA = microorganisms similar to Actinomyces. ∗∗ND = not disclosed. Table 6 Publications of cross-sectional studies of pelvic actinomycosis. Sample size Period Age Diagnosis Diagnosis method Previous use of IUD Important findings Reference 121,193 March 1977–November 1979 21–51 MSA∗ Papanicolaou Yes = 11,952 (6 months–12 years) 202 cases with MSA∗, 2 patients were not IUD users 1980, Fry et al. South Africa [64] 2290 ND∗∗ 17–76 MSA Papanicolaou Yes (prolonged use) 19 out of 2290 were diagnosed with MSA Statistically significant correlation of the presence of MSA with Trichomonas vaginalis, cocci, lactobacilli, pseudoeosinophils, endocervical cells, and polymorphs 2005, Demirezen et al. Turkey [65]

Sample size Period Age Diagnosis Diagnosis method Previous use of IUD Important findings Reference 121,193 March 1977–November 1979 21–51 MSA∗ Papanicolaou Yes = 11,952 (6 months–12 years) 202 cases with MSA∗, 2 patients were not IUD users 1980, Fry et al. South Africa [64] 2290 ND∗∗ 17–76 MSA Papanicolaou Yes (prolonged use) 19 out of 2290 were diagnosed with MSA Statistically significant correlation of the presence of MSA with Trichomonas vaginalis, cocci, lactobacilli, pseudoeosinophils, endocervical cells, and polymorphs 2005, Demirezen et al. Turkey [65] ND January 1994–January 2010 6–75 Actinomycosis TAC first and later histological report with finding of MSA Yes = 2 23 cases of abdominal pelvic actinomycosis were identified: 18 women: 5 had ovarian and pelvic masses, 2 in the uterus; as an important risk factor, 2 patients used IUDs 2011, Sung et al. Korea [66] 293 March 1978–March 1979 ND MSA Papanicolaou sample observations Yes = 128 plastic IUDs and 167 copper IUDs Oral contraceptives = 300 40 women with IUDs had increased prevalence of MSA, 2 who used copper and none who used oral contraceptives 1980, Duguid et al. England [67] 468 Comparative study with 4 sample groups ND 33 ± 7.7 MSA Papanicolaou samples Cases without IUD (group 0: 128) 2 to 35 months IUD use (group 1: 121); 36 to 71 months (group 2: 112); more than 72 months (group 3: 107) MSA was found in 2 cases from group 2 and in 7 patients from group 3 MSA was 3.68 times more likely with greater use 1999, Garrido et al. Colombia [68]

1. Introduction The genus Mycobacterium causes a wide variety of zoonotic diseases. The best known example is zoonotic tuberculosis due to M. bovis, for which cattle is the main reservoir. M. bovis is part of the “tuberculosis complex,” which also includes the species M. tuberculosis, M. africanum, M. caprae, and M. microti [1]. Within the mycobacterial group are the “nontuberculous mycobacteria (NTM)," which are also associated with infections in humans. The NTM are found in various environmental sources such as soil, water, vegetation, animals, dairy products, and feces and may be transmitted inadvertently by inhalation, ingestion, or skin penetration [2]. The Mexican standard NOM-ZOO-031-1995 regulates the presence of M. bovis in cattle to control and eradicate bovine tuberculosis (bTB); however, no regulation exists for the NTM species. The official diagnosis of bovine tuberculosis due to the presence of M. bovis at the field level is based on the intradermal test using a purified protein derivative (tuberculin) [3]. Although used for several years, this test does not provide good sensitivity and specificity. Approximately 20% of the animals with tuberculosis do not react to the test [4], and the presence of other mycobacterial species, both tuberculosis complex and NTM species, causes interference that leads to false-positive and false-negative diagnoses.

years, this test does not provide good sensitivity and specificity. Approximately 20% of the animals with tuberculosis do not react to the test [4], and the presence of other mycobacterial species, both tuberculosis complex and NTM species, causes interference that leads to false-positive and false-negative diagnoses. Although Mexico has a regulatory standard, bTB prevalence in excess of 2% is reported in some areas [5]. Given the poor specificity and sensitivity of the tuberculin test, the actual presence of M. bovis is likely to be lower and the infection rate of cattle by other mycobacteria is likely to be higher, respectively. Thus, cattle breeders, veterinarians, technicians, and employees working in the livestock industry might be occupationally exposed to infections by M. bovis and NTM. Very little is known about occupational exposure to zoonoses due to NTM species because the identification of these species was a rather difficult task prior to the development of identification techniques based on molecular biology.

livestock industry might be occupationally exposed to infections by M. bovis and NTM. Very little is known about occupational exposure to zoonoses due to NTM species because the identification of these species was a rather difficult task prior to the development of identification techniques based on molecular biology. Currently, the molecular biology techniques most commonly used for the diagnosis of diseases caused by mycobacteria are restriction fragment length polymorphism (RFLP) for the diagnosis of M. tuberculosis [6], spoligotyping for the diagnosis of M. bovis [7], and the detection of a 100-base pair (bp) “specific insertion” located on the 23S rRNA gene characteristic of Gram-positive bacteria with a high guanine-cytosine (HGC) content, which is considered a molecular marker for this group of bacteria [8, 9], followed by sequence analysis of the 16S rRNA gene for the identification of bacteria at the species level [10].

specific insertion” located on the 23S rRNA gene characteristic of Gram-positive bacteria with a high guanine-cytosine (HGC) content, which is considered a molecular marker for this group of bacteria [8, 9], followed by sequence analysis of the 16S rRNA gene for the identification of bacteria at the species level [10]. Among the NTM species identified by the aforementioned techniques are M. balnei, M. marinum, and M. platypoecilus, which have caused superficial and deep skin lesions [11]; M. kansasii from lung lesions [12]; M. simiae from generalized infections [13]; M. scrofulaceum from infections of the skin and internal organs [14]; M. szulgai associated with pulmonary infections, osteomyelitis, tenosynovitis, and lymphadenitis [15]; M. ulcerans associated with subcutaneous granulomas [16]; M. fortuitum and M. chelonae associated with vasculitis, endocarditis, osteomyelitis, mediastinitis, meningitis, keratitis, and hepatitis [17]; M. abscessus, associated with erythematous lesions that progressed to ulcerated nodules [18]; and other species.

[15]; M. ulcerans associated with subcutaneous granulomas [16]; M. fortuitum and M. chelonae associated with vasculitis, endocarditis, osteomyelitis, mediastinitis, meningitis, keratitis, and hepatitis [17]; M. abscessus, associated with erythematous lesions that progressed to ulcerated nodules [18]; and other species. The largest percentage of the state inventory for heads of cattle in the State of Mexico in Mexico is concentrated in the southern region, and one of the main economic activities is cattle ranching [19]. The Mexican regulation for cattle control NOM-ZOO-031-1995 only focuses on the tuberculin test for the diagnosis of M. bovis. Little is known about the presence of NTM in the cattle of the region. Given the possibility of identifying species of actinobacteria by detection of the 100-base pair molecular marker on the 23S rRNA gene and the subsequent sequencing of the 16S rRNA gene, it is possible to identify the aforementioned NTM species. The objective of the present study was to isolate and identify NTM species from cattle of the south region of the State of Mexico. The Mycobacterium species were isolated from samples of nasal exudate and bovine milk and identified by detecting the 100-base pair molecular marker in the 23S rRNA gene with subsequent sequencing of the 16S rRNA gene.

study was to isolate and identify NTM species from cattle of the south region of the State of Mexico. The Mycobacterium species were isolated from samples of nasal exudate and bovine milk and identified by detecting the 100-base pair molecular marker in the 23S rRNA gene with subsequent sequencing of the 16S rRNA gene. 2. Materials and Methods 2.1. Sampling A sampling was performed based on the spatial distribution of herds positive for bovine tuberculosis in the state of Mexico conducted by Zaragoza et al. 2015 [20]. Four herds of cattle were selected in the south region of the State of Mexico, one herd belonging to the Municipality of Temascaltepec and three herds belonging to the municipality of Zacazonapan. A total of 103 samples, 35 milk samples and 68 samples of nasal exudate, were collected. The distribution of the number and type of samples collected in each herd is shown in Table 1. 2.2. Obtaining Samples of Milk and Nasal Exudate The udder and nipples were cleansed with purified water and soap and then dried with paper towels, and nipple asepsis was subsequently performed using swabs soaked in 70% alcohol. Five milliliters of milk was collected directly from the nipple in sterile 20 mL vessels, discarding the initial flow. Nasal exudate was collected directly from the inside of the nasal orifice using a 10 cm long sterile swab, which was then submerged in an isotonic saline solution (0.85%). Samples of milk and nasal exudate were stored at 4°C until processing.

rectly from the nipple in sterile 20 mL vessels, discarding the initial flow. Nasal exudate was collected directly from the inside of the nasal orifice using a 10 cm long sterile swab, which was then submerged in an isotonic saline solution (0.85%). Samples of milk and nasal exudate were stored at 4°C until processing. 2.3. Sample Processing 2.3.1. Isolation of Mycobacteria The milk samples were centrifuged at 2500 revolutions per minute (rpm) for 10 minutes. The pellets from the milk and nasal exudate samples were inoculated into the following culture medium selective for mycobacteria: Stonebrink (BD BBL 220504), Middlebrook (BD BBL 254521), and Middlebrook (BD BBL 254521) supplemented with 6 g of sodium pyruvate per liter (Middlebrook-P). The inoculated media were incubated at 37°C for 8 weeks and were assessed every 3 days. 2.3.2. Classification of Isolated Strains The isolated strains were distributed in groups according to the following characteristics: colony pigmentation, growth time, and colony characteristics (shape, consistency, texture, and pigment production). Isolated strains were stained with Ziehl-Neelsen to confirm the presence of acid-fast bacilli (AFB) [21].

trains The isolated strains were distributed in groups according to the following characteristics: colony pigmentation, growth time, and colony characteristics (shape, consistency, texture, and pigment production). Isolated strains were stained with Ziehl-Neelsen to confirm the presence of acid-fast bacilli (AFB) [21]. 2.4. DNA Extraction Strains with microscopic characteristics similar to mycobacteria (acid-fast positivity) and two representative strains of each group were selected for identification. To obtain biomass, the strains were inoculated into 30 mL of Middlebrook liquid culture medium (BD BBL 254521) in 125 mL flasks and incubated at 37°C for 7 days. The liquid medium was transferred to sterile 15 mL Falcon tubes and centrifuged for 15 minutes at 14,000 rpm. Then, the supernatant was removed and the pellet was transferred to 1.5 mL Eppendorf tubes; the tubes were then centrifuged at 14,000 rpm × 5 minutes, and the supernatant was discarded. DNA extraction was performed on the resulting pellet using the Wizard Genomic DNA Purification kit (Promega A1120). 2.5. Detection of the Molecular Marker in the 23S rRNA Gene The 100 bp molecular marker located on the 23S rRNA gene was amplified according to the methodology described by Roller et al. (1992) using the following primers [8]:  23S InsF, 5′-(AC)A(AGT)GCGTAG(AGCT)CGA(AT)GG-3′, and 23S InsR, 5′-GTG(AT)CGGTTT(AGCT)(GCT)GGTA-3′.

etection of the Molecular Marker in the 23S rRNA Gene The 100 bp molecular marker located on the 23S rRNA gene was amplified according to the methodology described by Roller et al. (1992) using the following primers [8]:  23S InsF, 5′-(AC)A(AGT)GCGTAG(AGCT)CGA(AT)GG-3′, and 23S InsR, 5′-GTG(AT)CGGTTT(AGCT)(GCT)GGTA-3′. The reaction was conducted using a commercial Taq DNA polymerase (Promega M1661). The following thermal cycle conditions were used: a predenaturation step for 5 minutes (94°C); 29 cycles of denaturation for 30 seconds (94°C), hybridization for 45 seconds (46°C), and elongation for 50 seconds (72°C); and, finally, a postelongation cycle of 5 minutes (72°C). The amplified fragments were confirmed on a 2% agarose gel stained with ethidium bromide (SIGMA 46065). 2.6. Amplification of the 16S rRNA Gene Strains that amplified the 100 bp phylogenetic marker were selected for 16S rRNA sequencing analysis. The following primers were used for the amplification:  8f: AGAGTTTGATCMTGGCTCAG and 1492r: TACGGYTACCTTGTTACGACTT. The reaction was conducted using a commercial Taq DNA polymerase (Promega M1661). The following thermal cycle conditions were used: one predenaturation step for 5 minutes (94°C); 34 cycles of denaturation for 30 seconds (94°C), hybridization for 20 seconds (52°C), and elongation for 1 minute 30 seconds (72°C); and, finally, a postelongation cycle of 7 minutes (72°C).

DNA polymerase (Promega M1661). The following thermal cycle conditions were used: one predenaturation step for 5 minutes (94°C); 34 cycles of denaturation for 30 seconds (94°C), hybridization for 20 seconds (52°C), and elongation for 1 minute 30 seconds (72°C); and, finally, a postelongation cycle of 7 minutes (72°C). The amplified fragments were confirmed on a 1% agarose gel stained with ethidium bromide (SIGMA 46065). The products of this amplification were purified using the Amicon Ultra Filter® kit (Millipore UFC901008) and confirmed on a 1% agarose gel to verify their presence and quality. 2.7. Identification of Mycobacterium Species The amplified products of the 16S rRNA gene were sent to the Macrogen Sequencing Service, Maryland, USA. The obtained sequences were analyzed and corrected using the BioEdit program [22]. Consensus sequences were constructed from the forward and reverse fragments, which were compared with sequences deposited previously in GenBank of the National Center for Biotechnology Information (NCBI) using the BLAST program [23] and EzTaxon 2.1 [24].

ences were analyzed and corrected using the BioEdit program [22]. Consensus sequences were constructed from the forward and reverse fragments, which were compared with sequences deposited previously in GenBank of the National Center for Biotechnology Information (NCBI) using the BLAST program [23] and EzTaxon 2.1 [24]. 2.8. Phylogenetic Analysis Sequences of the 16S rRNA gene were obtained for the following mycobacterial species from the American Type Culture Collection (ATCC) and the German Collection of Microorganisms and Cell Cultures (DSM): M. neoaurum ATCC25795, M. parafortuitum DSM43528, M. moriokaense DSM44221T, and M. confluentis DSM44017T. The sequences of the collection strains and those of the strains isolated in the present investigation were aligned with the BioEdit program [22]. Phylogenetic analysis was performed using the maximum parsimony method in MEGA software version 4 [25]. To form the root of the cladogram, the sequence of Pantoea agglomerans DSM 3493 was used. 3. Results The 108 strains isolated from the 103 collected samples were distributed in 13 groups according to their macroscopic and microscopic morphological characteristics (Table 2). Groups 11 and 12, particularly, were composed of acid-fast strains.

2.8. Phylogenetic Analysis Sequences of the 16S rRNA gene were obtained for the following mycobacterial species from the American Type Culture Collection (ATCC) and the German Collection of Microorganisms and Cell Cultures (DSM): M. neoaurum ATCC25795, M. parafortuitum DSM43528, M. moriokaense DSM44221T, and M. confluentis DSM44017T. The sequences of the collection strains and those of the strains isolated in the present investigation were aligned with the BioEdit program [22]. Phylogenetic analysis was performed using the maximum parsimony method in MEGA software version 4 [25]. To form the root of the cladogram, the sequence of Pantoea agglomerans DSM 3493 was used. 3. Results The 108 strains isolated from the 103 collected samples were distributed in 13 groups according to their macroscopic and microscopic morphological characteristics (Table 2). Groups 11 and 12, particularly, were composed of acid-fast strains. For identification at the species level, 39 strains were chosen: 10 of them belonged to group 11 and 7 to group 12. Two strains from each one of the remaining 11 groups were selected to complete the 39 strains. The 100 bp molecular marker was found in the 33% (13/39) of the selected strains. For them, the 16S rRNA gene was amplified for sequencing and identification at the species level. The overall prevalence of NTM on the collected samples was 12.6% (13/103) considering both milk and nasal exudate samples. However, the specific prevalence for nasal exudate samples was 19.1% (13/68).

For identification at the species level, 39 strains were chosen: 10 of them belonged to group 11 and 7 to group 12. Two strains from each one of the remaining 11 groups were selected to complete the 39 strains. The 100 bp molecular marker was found in the 33% (13/39) of the selected strains. For them, the 16S rRNA gene was amplified for sequencing and identification at the species level. The overall prevalence of NTM on the collected samples was 12.6% (13/103) considering both milk and nasal exudate samples. However, the specific prevalence for nasal exudate samples was 19.1% (13/68). According to the sequence comparison, four NTM species of the genus Mycobacterium were identified; 64% (6/13) of the strains had 98% and 99% of similarities with M. neoaurum, while 31% (4/13) had 99% similarity with M. parafortuitum, 15% (2/13) had similarities of 98% and 99% with M. moriokaense, and, finally, 8% (1/13) had 99% similarity with M. confluentis (Table 3). The phylogenetic tree was formed with the genus Mycobacterium and four of its species by which the phylogenetic relationships between the collection strains and the strains isolated in the present investigation were observed (Figure 1).

According to the sequence comparison, four NTM species of the genus Mycobacterium were identified; 64% (6/13) of the strains had 98% and 99% of similarities with M. neoaurum, while 31% (4/13) had 99% similarity with M. parafortuitum, 15% (2/13) had similarities of 98% and 99% with M. moriokaense, and, finally, 8% (1/13) had 99% similarity with M. confluentis (Table 3). The phylogenetic tree was formed with the genus Mycobacterium and four of its species by which the phylogenetic relationships between the collection strains and the strains isolated in the present investigation were observed (Figure 1). 4. Discussion The NTM species were isolated from samples of nasal exudate only, which eliminated the samples from one of the local farms of this study (Table 1). We found that the specific prevalence was 19.1% in herds of the south region of the State of Mexico. Similar studies in the United States, South Africa, Tanzania, and Brazil reported NTM prevalence values of 3.4%, 24.5%, 7%, and 7.8%, respectively; therefore, the prevalence value found in this study lies within the range reported previously [26–29]. In this study, 13 of the 39 analyzed strains were identified as the NTM species M. neoaurum, M. moriokaense, M. confluentis, and M. parafortuitum.

rted NTM prevalence values of 3.4%, 24.5%, 7%, and 7.8%, respectively; therefore, the prevalence value found in this study lies within the range reported previously [26–29]. In this study, 13 of the 39 analyzed strains were identified as the NTM species M. neoaurum, M. moriokaense, M. confluentis, and M. parafortuitum. M. neoaurum, a member of the Mycobacterium parafortuitum complex, is responsible for a broad spectrum of illnesses, most of them device related infections such as Hickman catheters, BROVIAC catheters, PICC lines [30–33], arteriovenous fistula that included a polytetrafluoroethylene graft [34], pace makers [35], and prosthetic valve endocarditis [36]. Immunocompromised patients holding these devices are the principal hosts, for example, patients suffering from cancer [32] and diabetics with renal failure [31, 33, 34] and heart problems [35]. M. neoaurum has also been isolated from patients with urinary infections [37], meningoencephalitis and alterations in the central nervous system [38], bacteremia and endocarditis [39], and pulmonary infection [40, 41]. Although it has been mainly isolated from clinical cases, there are also reports about its isolation from milk and cattle [28, 42, 43].

from patients with urinary infections [37], meningoencephalitis and alterations in the central nervous system [38], bacteremia and endocarditis [39], and pulmonary infection [40, 41]. Although it has been mainly isolated from clinical cases, there are also reports about its isolation from milk and cattle [28, 42, 43]. M. moriokaense was isolated from sputum sample [44]. Although it is considered nonpathogenic for humans, it has been associated with pulmonary diseases [45]. M. confluentis was isolated from sputum samples as well [46], and, along with M. parafortuitum, both are considered nonpathogenic species. M. confluentis, M. moriokaense, and M. neoaurum have been isolated from different bovine and wildlife tissues with tuberculous lesions, whereas M. parafortuitum has only been isolated from bovine milk [26, 28, 47–49]. However, in our work, M. parafortuitum was only isolated from nasal exudate samples.

thogenic species. M. confluentis, M. moriokaense, and M. neoaurum have been isolated from different bovine and wildlife tissues with tuberculous lesions, whereas M. parafortuitum has only been isolated from bovine milk [26, 28, 47–49]. However, in our work, M. parafortuitum was only isolated from nasal exudate samples. The nutritional requirements of mycobacteria differ among various species, which was the reason for using different culture media. Notably, seven of the 13 strains identified in this study were isolated in Stonebrink medium, including M. neoaurum, M. parafortuitum, and M. moriokaense. This result is consistent with that described by Sepúlveda et al. [50] who indicated that Stonebrink medium is suitable for the recovery of different species of the genus Mycobacterium. García-Martos and García-Agudo [51] reported that Middlebrook medium is optimal for the isolation of actinomycetes, which is in accord with the present investigation considering that two species, M. neoaurum and M. parafortuitum, were isolated in this medium. Notably, M. confluentis was isolated only in Middlebrook medium supplemented with sodium pyruvate; thus, the strategy of using different culture media was appropriate because it allowed the isolation of different species of the genus Mycobacterium.

pecies, M. neoaurum and M. parafortuitum, were isolated in this medium. Notably, M. confluentis was isolated only in Middlebrook medium supplemented with sodium pyruvate; thus, the strategy of using different culture media was appropriate because it allowed the isolation of different species of the genus Mycobacterium. The detection of the molecular marker present in the 23S rRNA gene of Gram-positive bacteria with HGC content allowed discrimination between strains of eubacteria and mycobacteria. The sequencing analysis of the 16S rRNA gene made the identification at the species level possible; therefore, the combination of these methodologies is appropriate for the identification of NTM species. 5. Conclusions Using the methodology described in this study, four NTM species were isolated and identified: M. confluentis, M. moriokaense, M. neoaurum, and M. parafortuitum. These species were isolated for the first time from nasal exudates of bovines from the south region of the State of Mexico. Three of the identified species (M. neoaurum, M. moriokaense, and M. confluentis) are of public health and veterinary importance.

M. confluentis, M. moriokaense, M. neoaurum, and M. parafortuitum. These species were isolated for the first time from nasal exudates of bovines from the south region of the State of Mexico. Three of the identified species (M. neoaurum, M. moriokaense, and M. confluentis) are of public health and veterinary importance. Acknowledgments The authors would like to acknowledge the financial assistance from the Secretary of Research and Advanced Studies of Universidad Autónoma del Estado de México (UAEMex) through the following research grants: (i) “Implementation of Geographic Information Systems and Techniques of Molecular Biology, as tools in the detection and identification of Mycobacterium spp.,” SIEA-UAEM 3486/2013CHT, and (ii) the network “Microbiología y química en las Ciencias de la Salud,” 039/2014RIF. Disclosure This work is derived from the thesis for the degree of Doctorate in Health Sciences (Universidad Autónoma del Estado de México), registered in the PNPC-CONACYT. Conflicts of Interest All authors declare that they do not have any conflicts of interest. Figure 1 Phylogenetic tree constructed by comparing the 16S rRNA gene sequences from the isolated and reference strains. Table 1 Samples collected in cattle herds in the south region of the State of Mexico.

Disclosure This work is derived from the thesis for the degree of Doctorate in Health Sciences (Universidad Autónoma del Estado de México), registered in the PNPC-CONACYT. Conflicts of Interest All authors declare that they do not have any conflicts of interest. Figure 1 Phylogenetic tree constructed by comparing the 16S rRNA gene sequences from the isolated and reference strains. Table 1 Samples collected in cattle herds in the south region of the State of Mexico. Characteristic Herd 1 Herd 2 Herd 3 Herd 4 Total Municipality Temascaltepec Zacazonapan Zacazonapan Zacazonapan Breed F1 Swiss-Cebu Holstein Friesian Holstein Friesian Holstein Friesian Geographic location La-19°03′13.7′′Lo-100°13′36.7′′ La-19°03′39.5′′Lo-100°16′30.9′′ La-19°04′0.4′′Lo-100°15′11.5′′ La-19°03′41′′Lo-100°16′06′′ History of bTB Prevalence 0.2%∗ Prevalence 0.2%∗ Prevalence 0.2%∗ Prevalence 0.2%∗ Samples obtained Milk 15 20 0 0 35 Nasal exudate 0 18 23 27 68 103 La: latitude; Lo: longitude; bTB: bovine tuberculosis. ∗Information obtained from the Committee on the Promotion and Protection of Livestock of the State of Mexico. Table 2 Isolated strains are grouped according to their morphological characteristics and the presence of the molecular marker (100 bp) on the 23S rRNA gene.

La: latitude; Lo: longitude; bTB: bovine tuberculosis. ∗Information obtained from the Committee on the Promotion and Protection of Livestock of the State of Mexico. Table 2 Isolated strains are grouped according to their morphological characteristics and the presence of the molecular marker (100 bp) on the 23S rRNA gene. Group Number of Strains Morphological characteristics Molecular marker (bp) Macroscopic Microscopic 23S rRNA Pigmentation Appearance Ziehl-Neelsen 1 29 Yellow Creamy − 250 2 10 White Creamy − 250 3 14 White Dry − 250 4 2 Salmon Creamy − 250 5 2 Salmon Dry − 250 6 2 White, dark Creamy − 250 7 2 White Creamy − 250 8 3 White, dark Dry − 250 9 4 White Dry − 250 10 10 White Creamy, dry − 250 11 10 White Dry + 350 and 250 12 7 Yellow Creamy + 350 13 13 White Dry − 250 −: absence of acid-fast bacilli; +: presence of acid-fast bacilli; Bp: base pairs. Table 3 Comparison of 16S rRNA gene sequences of strains isolated from cattle with those documented in GenBank, using BLAST and EzTaxon.

Group Number of Strains Morphological characteristics Molecular marker (bp) Macroscopic Microscopic 23S rRNA Pigmentation Appearance Ziehl-Neelsen 1 29 Yellow Creamy − 250 2 10 White Creamy − 250 3 14 White Dry − 250 4 2 Salmon Creamy − 250 5 2 Salmon Dry − 250 6 2 White, dark Creamy − 250 7 2 White Creamy − 250 8 3 White, dark Dry − 250 9 4 White Dry − 250 10 10 White Creamy, dry − 250 11 10 White Dry + 350 and 250 12 7 Yellow Creamy + 350 13 13 White Dry − 250 −: absence of acid-fast bacilli; +: presence of acid-fast bacilli; Bp: base pairs. Table 3 Comparison of 16S rRNA gene sequences of strains isolated from cattle with those documented in GenBank, using BLAST and EzTaxon. Strain Origin of the herd Culture medium Amplified fragment size (bp) Similarity (Blast) % Similarity (EzTaxon) % 1-AZ 2 Middlebrook 1428 M. neoaurum 98 M. neoaurum 98.3 2-AZ 2 Stonebrink 1408 M. neoaurum 99 M. neoaurum 99.1 3-AZ 2 Stonebrink 1428 M. neoaurum 98 M. neoaurum 98.2 5-AZ 3 Stonebrink 1416 M. neoaurum 99 M. neoaurum 99.2 8-AZ 4 Middlebrook 1415 M. neoaurum 99 M. neoaurum 99.0 12-AZ 2 Middlebrook 1415 M. neoaurum 99 M. neoaurum 99.4 4-AZ 4 Middlebrook-P 1420 M. parafortuitum 99 M. parafortuitum 98.2 9-AZ 3 Middlebrook 1415 M. parafortuitum 99 M. parafortuitum 98.9 10-AZ 4 Stonebrink 1411 M. parafortuitum 99 M. parafortuitum 98.4 11-AZ 3 Stonebrink 1414 M. parafortuitum 99 M. parafortuitum 98.2 6-AZ 2 Stonebrink 1455 M. moriokaense 99 M. moriokaense 98.2 13-AZ 4 Stonebrink 1417 M. moriokaense 98 M. moriokaense 98.2 7-AZ 2 Middlebrook-P 1420 M. confluentis 99 M. confluentis 99.1 2: Zacazonapan Holstein-F; 3: Zacazonapan Holstein-F; 4: Zacazonapan Holstein-F.

1. Introduction Tuberculosis (TB) remains a global public health problem. According to World Health Organization (WHO), an estimated number of 10.4 million new cases occurred in the world in 2015 [1]. The African Region recorded the highest incidence rate, almost twice that of the world [1]. In Benin in West Africa, 4,092 cases were detected in 2015 [2]. Despite use of standardized treatment regimens and a well-established National TB Program (NTP) in the country, the treatment success rate as well as the number of previously treated cases (failure, relapse, and default) has remained stable over years [2]. In contrast to new cases, previously treated cases are much more likely to harbour multidrug-resistant (MDR) strains, defined as resistance to rifampicin (R) and isoniazid (H), and their characteristics may differ from those of new cases [3, 4]. Molecular tools are useful for better understanding of TB transmission dynamics in a given area. Nevertheless, molecular studies on TB are scarce in high-incidence, low-income countries [5]. In Benin, the only molecular epidemiologic study available to our knowledge recruited only TB new cases in one city [6, 7]. The scarcity of these studies in TB endemic countries is partly due to lack of resources and relative complexity of some molecular techniques. Among them, spoligotyping has the advantage of being relatively simple, inexpensive, and generally sufficient as a first approach of molecular epidemiology of TB [8].

ty [6, 7]. The scarcity of these studies in TB endemic countries is partly due to lack of resources and relative complexity of some molecular techniques. Among them, spoligotyping has the advantage of being relatively simple, inexpensive, and generally sufficient as a first approach of molecular epidemiology of TB [8]. In this study, we aimed to evaluate a nationwide genotypic diversity of Mycobacterium tuberculosis complex strains in previously treated pulmonary TB patients in Benin, using spoligotyping technique. 2. Materials and Methods 2.1. Setting Benin is a country with a size of 114,763 square kilometres and an estimated population of 11 million. It has 70 TB facilities spread all over the country and a well-established National TB Program. Every year, about 4,000 TB cases including new and previously treated cases are detected in the country [2].

2.1. Setting Benin is a country with a size of 114,763 square kilometres and an estimated population of 11 million. It has 70 TB facilities spread all over the country and a well-established National TB Program. Every year, about 4,000 TB cases including new and previously treated cases are detected in the country [2]. 2.2. Specimens A total of 100 isolates obtained from 100 sputum samples collected from smear-positive previously treated pulmonary TB patients all over the country were sent to the National Reference Laboratory (NRL) in Cotonou for processing. Previously treated TB patients were from relapse (n = 71), failure (n = 24), and default (n = 5) cases. Two sputum samples were collected (spot and early morning) from each patient, stored at 4°C, and sent in a cool box to the NRL within a week. Upon arrival at the NRL, the two samples were processed for culture but only one strain per patient was used for drug susceptibility testing (DST) and DNA fingerprinting. Samples were systematically collected between January and December 2014, and, for each of them, demographic data was retrieved, while after obtaining consent from each patient, HIV screening was performed on blood using rapid immunochromatography-based tests: Alere Determine HIV-1/2® (Alere Medical, Japan) was used for HIV screening, while samples that were reactive were confirmed by ImmunoComb HIV 1&2 BiSpot® (Orgenics, France).

was retrieved, while after obtaining consent from each patient, HIV screening was performed on blood using rapid immunochromatography-based tests: Alere Determine HIV-1/2® (Alere Medical, Japan) was used for HIV screening, while samples that were reactive were confirmed by ImmunoComb HIV 1&2 BiSpot® (Orgenics, France). 2.3. Culture and DST Samples were decontaminated using the Petroff method and cultured on Löwenstein-Jensen (LJ) media [9]. The M. tuberculosis isolates (one per patient) were tested for susceptibility against rifampicin (R), isoniazid (H), streptomycin (S), and ethambutol (E) using the proportion method on LJ medium at the following concentrations: 40 μg/mL, 0.2 μg/mL, 4 μg/mL, and 2 μg/mL, respectively [9]. Internal quality control was routinely performed, while annual external quality assurance was carried out by the WHO Supranational Reference Laboratory at the Institute of Tropical Medicine, Antwerp, Belgium. In case of resistance to R, DST for second-line drugs was performed using the proportion method on LJ medium at the following concentrations: kanamycin (30 μg/mL), capreomycin (40 μg/mL), amikacin (40 μg/mL), and ofloxacin (2 μg/mL) [10]. All strains were stored upon routine processing at −80°C and subcultured on LJ for spoligotyping.

tance to R, DST for second-line drugs was performed using the proportion method on LJ medium at the following concentrations: kanamycin (30 μg/mL), capreomycin (40 μg/mL), amikacin (40 μg/mL), and ofloxacin (2 μg/mL) [10]. All strains were stored upon routine processing at −80°C and subcultured on LJ for spoligotyping. 2.4. Spoligotyping DNA was extracted by making a suspension of bacteria with a loop of colonies into 300 μL of molecular grade water followed by heating at 100°C for 20 minutes. Spoligotyping was performed as previously described [11]. Mycobacterium tuberculosis H37Rv was used as a positive control, while molecular grade water served as negative control. Spoligotype patterns obtained were then translated into binary code with 1 and 0 for presence and absence of “spacer” and then entered on an Excel file. From these codes, lineages and families of strains were determined using TB lineage database http://tbinsight.cs.rpi.edu/run_tb_lineage.html [12] and the SPOTCLUST database http://tbinsight.cs.rpi.edu/run_spotclust.html [13], respectively. Spoligotype data were compared to the SITVIT WEB database (http://www.pasteur-guadeloupe.fr:8081/SITVIT_ONLINE/) [14] to determine the Spoligotype International Type (SIT) if already described. 2.5. Data Analysis Data were analyzed using EpiData 3.1. Chi-square test and Fisher's exact test were used to compare proportions. p value < 0.05 was considered significant.

2.4. Spoligotyping DNA was extracted by making a suspension of bacteria with a loop of colonies into 300 μL of molecular grade water followed by heating at 100°C for 20 minutes. Spoligotyping was performed as previously described [11]. Mycobacterium tuberculosis H37Rv was used as a positive control, while molecular grade water served as negative control. Spoligotype patterns obtained were then translated into binary code with 1 and 0 for presence and absence of “spacer” and then entered on an Excel file. From these codes, lineages and families of strains were determined using TB lineage database http://tbinsight.cs.rpi.edu/run_tb_lineage.html [12] and the SPOTCLUST database http://tbinsight.cs.rpi.edu/run_spotclust.html [13], respectively. Spoligotype data were compared to the SITVIT WEB database (http://www.pasteur-guadeloupe.fr:8081/SITVIT_ONLINE/) [14] to determine the Spoligotype International Type (SIT) if already described. 2.5. Data Analysis Data were analyzed using EpiData 3.1. Chi-square test and Fisher's exact test were used to compare proportions. p value < 0.05 was considered significant. 3. Results In total, 100 viable strains (one single strain per patient) were used for spoligotyping. They were 71, 24, and 5 isolates from relapse, failure, and default patients, respectively. In total, 74 (74.0%) isolates were from male patients, while 26 (26.0%) were from females. HIV positivity rate was 15.2%, all of whom were infected with HIV1.

strains (one single strain per patient) were used for spoligotyping. They were 71, 24, and 5 isolates from relapse, failure, and default patients, respectively. In total, 74 (74.0%) isolates were from male patients, while 26 (26.0%) were from females. HIV positivity rate was 15.2%, all of whom were infected with HIV1. Resistance pattern to first-line drugs by type of previously treated cases is presented in Table 1. Resistance rate to any first-line drug was 40.0%, while 12.0% of strains were multidrug-resistant (MDR). In addition, two other strains were resistant to R but not to H; one was monoresistant to R and another one was resistant to both R and S. Thus, resistance rate to R was 14.0%. MDR rates were 20.8% and 9.9% for failure and relapse cases, respectively, while none was found among defaulters. Second-line DST results were available for nine MDR strains, of which six (66.7%) were susceptible, two (22.2%) showed resistance to ofloxacin, and one (11.1%) showed resistance to kanamycin, while none was resistant to both fluoroquinolones and injectable drugs. Thus, no strain was extensively drug-resistant (XDR) (Table 2).

line DST results were available for nine MDR strains, of which six (66.7%) were susceptible, two (22.2%) showed resistance to ofloxacin, and one (11.1%) showed resistance to kanamycin, while none was resistant to both fluoroquinolones and injectable drugs. Thus, no strain was extensively drug-resistant (XDR) (Table 2). A total of 40 distinct spoligotypes were found to be corresponding to a genotypic diversity of 40.0%. Of these, 21 (52.5%) corresponded to spoligotypes already identified in the SITVIT database and had shared-type (ST) denominations (SIT), while 19 (47.5%) were newly found spoligotypes. ST61, ST53, and ST1 were the most predominant spoligotypes with prevalence rates of 33.0%, 13.0%, and 8.0%, respectively. In this study, 31 single spoligotypes and nine clusters were observed with 2 to 33 strains per cluster, giving a rate of 69.0% (Table 3). Most prevalent families were LAM 10, T1, and M. africanum West-African 1 with prevalence rates of 46.0%, 17.0%, and 12.0%, respectively. For lineages, the more prevalent lineages were Euro-American (Lineage 4), M. africanum West-African 1 (Lineage 5), and East-Asian (Lineage 2), with prevalence rates of 74.0%, 12.0%, and 8.0%, respectively. Interestingly, one strain was identified as M. bovis, representing 1.0% of the total strains tested (Table 4).

ly. For lineages, the more prevalent lineages were Euro-American (Lineage 4), M. africanum West-African 1 (Lineage 5), and East-Asian (Lineage 2), with prevalence rates of 74.0%, 12.0%, and 8.0%, respectively. Interestingly, one strain was identified as M. bovis, representing 1.0% of the total strains tested (Table 4). By comparing characteristics of patients within lineages, we found no association between sex, HIV status, types of treatment, and lineages; however, drug resistance particularly resistance to S was associated with lineages distribution. Strains belonging to Lineage 2 were more likely to be resistant to S than the other strains (p = 0.001) (Table 5). 4. Discussion There are still several gaps in understanding TB dynamics in Africa. For example, the reason why M. africanum is mainly restricted to the Western and Central parts of the continent remains unclear [5, 7]. Studies using molecular tools may be useful in this respect. Unfortunately, the few molecular studies available either were limited to a city or a region or only focused on new TB cases and if previously treated cases were included, the number was usually low [15, 16].

ral parts of the continent remains unclear [5, 7]. Studies using molecular tools may be useful in this respect. Unfortunately, the few molecular studies available either were limited to a city or a region or only focused on new TB cases and if previously treated cases were included, the number was usually low [15, 16]. In this study, we carried out a nationwide molecular study on previously treated pulmonary TB cases detected in Benin over a period of one year. In total, 40 different spoligotypes were found, corresponding to a genotypic diversity of 40.0%. This percentage was higher than the 19.1% found by Ouassa et al. in previously treated cases in Côte d'Ivoire but was quite similar to 35.1% obtained on the genetic diversity in a mixed population of new and previously treated cases in Rwanda [15, 16]. A genotypic diversity of 49.0% was reported in 2005 among new cases in Cotonou, the biggest city in Benin, suggesting that genetic diversities were similar among new and previously treated cases [6]. However, the previous study among new cases was carried out 10 years ago and distribution of spoligotypes in new cases might have changed over time. In addition, the national figure might be different from what was obtained in Cotonou.

ng that genetic diversities were similar among new and previously treated cases [6]. However, the previous study among new cases was carried out 10 years ago and distribution of spoligotypes in new cases might have changed over time. In addition, the national figure might be different from what was obtained in Cotonou. This study showed that the most frequent spoligotype was ST61 (33%) belonging to the Latino-American and Mediterranean (LAM) family. This finding was similar to what was previously reported in the same country in 2005, indicating that ST61 was the most prevalent spoligotype in new cases [6]. This same genotype was previously described to be prevalent in countries within the West-African coast [17]. At a lineage level, Lineage 4 was the most prevalent lineage (74.0%). High prevalence of Lineage 4 was also found in both new and previously treated cases at a similar rate in Ethiopia (72.4%) and in Guinea (78.8%) [18, 19]. In comparison with other lineages, Lineage 4 appears to have certain characteristics that promote its rapid expansion. For M. bovis, the prevalence rate (1.0%) is similar to those found elsewhere in a mixed population of new and previously treated cases in Ethiopia (1.2%), Nigeria (1.0%), and Mali (0.8%) [20–22]. These low proportions could be explained by the fact that M. bovis is usually involved in extrapulmonary TB in humans, whereas most of these studies, including the present one, were on pulmonary TB [20–22].

mixed population of new and previously treated cases in Ethiopia (1.2%), Nigeria (1.0%), and Mali (0.8%) [20–22]. These low proportions could be explained by the fact that M. bovis is usually involved in extrapulmonary TB in humans, whereas most of these studies, including the present one, were on pulmonary TB [20–22]. A significant association was found between Lineage 2 (Beijing strains) and resistance to streptomycin (p = 0.001). This same association was observed in a study of new cases in 2005 in Cotonou, where an outbreak characterised by Information Geographical System was identified [23]. The positive correlation between Lineage 2 and resistance to streptomycin suggests a clonal distribution of Beijing strains in Benin. This study showed that, in Benin, molecular signatures of strains causing TB retreatment cases are similar to those causing new cases. The reasons why these strains did not respond to first-line TB treatment are likely to be related to human and environmental factors rather than the intrinsic molecular characteristics of strains. Therefore, for TB control, national TB programs in Benin as well as in neighbouring countries should make efforts to reduce the impact of these factors in order to decrease the number of TB retreatment cases.

o be related to human and environmental factors rather than the intrinsic molecular characteristics of strains. Therefore, for TB control, national TB programs in Benin as well as in neighbouring countries should make efforts to reduce the impact of these factors in order to decrease the number of TB retreatment cases. In this study, the resistance rate to R was 14.0%. This rate was slightly higher than what was observed by Affolabi et al. in Benin in 2013 (10%) but was comparable to that of the national drug resistance survey in 2010 [24, 25]. Furthermore, this rate was similar to what was reported by Homolka et al. among previously treated cases in Sierra Leone (14.4%) but was lower than 43.7% reported by Dia et al. in Senegal [26, 27]. Among MDR cases, 33.3% were pre-XDR but no XDR strain was found, contrary to findings from Burkina Faso, Ethiopia, and many other countries in Sub-Saharan Africa [28–30]. This absence of XDR strains in this study could be explained by the rigorous management of the MDR-TB program in Benin with strict application of directly observed therapy during the whole nine-month treatment course. However, the threat of XDR-TB is always present with the emergence of pre-XDR cases and a need for more vigilance cannot be overemphasized.

tudy could be explained by the rigorous management of the MDR-TB program in Benin with strict application of directly observed therapy during the whole nine-month treatment course. However, the threat of XDR-TB is always present with the emergence of pre-XDR cases and a need for more vigilance cannot be overemphasized. In conclusion, this first insight into the genetic diversity of TB in previously treated cases in Benin showed a genetic diversity of 40.0%, with most strains belonging to Lineage 4, similar to previous data in new TB cases. Occurrence of retreatment cases is more likely to be related to human and environmental factors rather than the intrinsic molecular characteristics of strains. Abbreviations DST:Drug susceptibility test E:Ethambutol H:Isoniazid LJ:Löwenstein-Jensen LAM:Latino-American and Mediterranean MDR:Multidrug-resistant NRL:National Reference Laboratory R:Rifampicin SIT:Spoligotype International Type S:Streptomycin ST:Shared-type TB:Tuberculosis WHO:World Health Organization XDR:Extensively drug-resistant. Conflicts of Interest The authors declare that there are no conflicts of interest regarding the publication of this paper. Table 1 Resistance pattern of strains to first-line drugs.

NRL:National Reference Laboratory R:Rifampicin SIT:Spoligotype International Type S:Streptomycin ST:Shared-type TB:Tuberculosis WHO:World Health Organization XDR:Extensively drug-resistant. Conflicts of Interest The authors declare that there are no conflicts of interest regarding the publication of this paper. Table 1 Resistance pattern of strains to first-line drugs. Type of resistance Failure (n = 24) n (%) Relapse (n = 71) n (%) Default (n = 5) n (%) Total (n = 100) n (%) Susceptible to all drugs 12 (50.0) 45 (63.4) 3 (60.0) 60 (60.0) Monoresistance H 1 (4.2) 1 (1.4) 0 (0.0) 2 (2.0) S 3 (12.5) 10 (14.1) 1 (20.0) 14 (14.0) R 0 (0.0) 0 (0.0) 1 (20.0) 1 (1.0) E 1 (4.2) 1 (1.4) 0 (0.0) 2 (2.0) Total 5 (20.8) 12 (16.9) 2 (40.0) 19 (19.0) Multidrug resistance HR 0 (0.0) 2 (2.8) 0 (0.0) 2 (2.0) HRE 1 (4.2) 1 (1.4) 0 (0.0) 2 (2.0) HRS 1 (4.2) 1 (1.4) 0 (0.0) 2 (2.0) HRES 3 (12.5) 3 (4.2) 0 (0.0) 6 (6.0) Total 5 (20.8) 7 (9.9) 0 (0.0) 12 (12.0) Other patterns HS 1 (4.2) 1 (1.4) 0 (0.0) 2 (2.0) HSE 0 (0.0) 2 (2.8) 0 (0.0) 2 (2.0) RS 0 (0.0) 1 (1.4) 0 (0.0) 1 (1.0) ES 1 (4.2) 3 (4.2) 0 (0.0) 4 (4.0) Total 2 (8.3) 7 (9.9) 0 (0.0) 9 (9.0) H: isoniazid; E: ethambutol; S: streptomycin; R: rifampicin. Table 2 Resistance patterns to second-line drugs on MDR strains. Type of resistance MDR strains n = 9 n (%) Susceptible to all second-line drugs 6 (66.7) Monoresistance Ofloxacin 2 (22.2) Kanamycin 1 (11.1) Capreomycin 0 Amikacin 0 Total 3 (33.3) XDR 0 XDR: extensively drug-resistant. Table 3 Strains per family. Family Spoligotype ST Strains n (%) Family33 761777767775771 U 1 (1.0%) 777777777763771 54 1 (1.0%)

Type of resistance MDR strains n = 9 n (%) Susceptible to all second-line drugs 6 (66.7) Monoresistance Ofloxacin 2 (22.2) Kanamycin 1 (11.1) Capreomycin 0 Amikacin 0 Total 3 (33.3) XDR 0 XDR: extensively drug-resistant. Table 3 Strains per family. Family Spoligotype ST Strains n (%) Family33 761777767775771 U 1 (1.0%) 777777777763771 54 1 (1.0%) Family34 777777770000000 46 1 (1.0%) Beijing 000000000003771 1 8 (8.0%) CAS 703777740001171 1199 1 (1.0%) LAM1 677777607760771 20 1 (1.0%) LAM9 377777607760771 177 1 (1.0%) LAM10 777777743760771 61 33 (33.0%) 767740741760751 U 1 (1.0%) 777770343760771 U 1 (1.0%) 777677743760771 U 3 (3.0%) 777770343740771 U 2 (2.0%) 777777743460771 772 3 (3.0%) 777777742760771 U 1 (1.0%) 777777743760731 403 1 (1.0%) 777777743740771 U 1 (1.0%) T1 777777777760771 53 13 (13.0%) 777777777760731 51 1 (1.0%) 737777777760731 848 1 (1.0%) 777777757760771 44 1 (1.0%) 737777777760531 U 1 (1.0%) T2 777417707700000 U 1 (1.0%) T4 777740017760771 159 1 (1.0%) Haarlem1 (H1) 777777770020731 316 1 (1.0%) Haarlem2 (H2) 000000000020731 U 1 (1.0%) Haarlem3 (H3) 777777777720731 49 2 (2.0%) 777777777720771 50 2 (2.0%) Family36 000000007760771 4 1 (1.0%) M. africanum West-African 1 774077607777071 331 3 (3.0%) 674077717777071 U 1 (1.0%) 774077400603031 U 1 (1.0%) 770002607777071 U 1 (1.0%) 574077607777071 319 1 (1.0%) 774077600000071 U 1 (1.0%) 374077607777031 U 1 (1.0%) 574017607777071 U 1 (1.0%) 774040077777071 U 1 (1.0%) 774077777777071 438 1 (1.0%) M. africanum West-African 2 700000377777671 U 1 (1.0%) M. bovis 000040000200000 U 1 (1.0%) ST: shared-type; U: unknown. Table 4 Strains per lineage.

M. africanum West-African 1 774077607777071 331 3 (3.0%) 674077717777071 U 1 (1.0%) 774077400603031 U 1 (1.0%) 770002607777071 U 1 (1.0%) 574077607777071 319 1 (1.0%) 774077600000071 U 1 (1.0%) 374077607777031 U 1 (1.0%) 574017607777071 U 1 (1.0%) 774040077777071 U 1 (1.0%) 774077777777071 438 1 (1.0%) M. africanum West-African 2 700000377777671 U 1 (1.0%) M. bovis 000040000200000 U 1 (1.0%) ST: shared-type; U: unknown. Table 4 Strains per lineage. Lineage Denomination n % 1 Indo-Oceanic 3 3.0 2 East-Asian 8 8.0 3 East-African-Indian 1 1.0 4 Euro-American 74 74.0 5 M. africanum West-African 1 12 12.0 6 M. africanum West-African 2 1 1.0 M. bovis M. bovis 1 1.0 Total 100 100.0 Table 5 Association between lineages and resistance. Lineage Resistance to S n (%) No resistance to S n (%) 1 1 (33.3) 2 (66.7) 2 7 (87.5) 1 (12.5) 3 1 (100.0) 0 (0.0) 4 17 (23.0) 57 (77.0) 5 4 (40.0) 6 (60.0) 6 0 (0.0) 1 (100.0) M. bovis 1 (100.0) 0 (0.0) Total 31 (31.6) 67 (68.4) S: streptomycin.

1. Introduction Fungal infections of the feet including tinea pedis and tinea unguium are very common in the general population [1]. Tinea pedis, generally known as athlete's foot, is divided into three clinical forms such as interdigital, plantar (moccasin foot), and vesiculobullous [2]. Interdigital is the most common clinical manifestation characterized by maceration and fissuring of the skin mainly in the space between the toes. Plantar athlete's foot presents with hyperkeratosic and squamous plaques which cover the soles, heels, and sides of the foot. In inflammatory condition vesicles, pustules and sometimes bullae are present on the sole of the foot [3]. Tinea unguium is classified into four clinical types depending on the mode of penetration of the fungus in the nail plate: distal lateral subungual onychomycosis (DLSO); proximal subungual onychomycosis (PSO); white superficial onychomycosis (WSO); and total dystrophic onychomycosis (TDO) [4]. Because of the prolonged period of treatment and the recurrence of infections, foot mycoses are still considered as a major public health problem affecting quality of life [5]. These fungal infections depends on many factors especially lifestyle and environmental and climatic conditions and can be influenced by individual factors such as age and host defenses [6]. Foot mycoses are mainly caused by dermatophytes, sometimes yeasts, and uncommonly by non-dermatophyte molds (NDMs).

y of life [5]. These fungal infections depends on many factors especially lifestyle and environmental and climatic conditions and can be influenced by individual factors such as age and host defenses [6]. Foot mycoses are mainly caused by dermatophytes, sometimes yeasts, and uncommonly by non-dermatophyte molds (NDMs). Many epidemiological studies have investigated the variability of the frequency of tinea pedis and tinea unguium in different geographical regions [7–11]. In fact, the practice of epidemiological studies at regular intervals is necessary for monitoring the evolution of foot mycoses over time. To our knowledge, there are few recent studies regarding clinical and mycological features of foot mycoses in Tunisia. The aim of our study was to determine the frequency of foot mycoses, their clinical patterns, predisposing factors, and etiological agents in Tunisian patients.

the evolution of foot mycoses over time. To our knowledge, there are few recent studies regarding clinical and mycological features of foot mycoses in Tunisia. The aim of our study was to determine the frequency of foot mycoses, their clinical patterns, predisposing factors, and etiological agents in Tunisian patients. 2. Patients and Methods It was a prospective study that was carried during one year from March 2013 and included all patients referred to the Mycology Unit in the Department of Dermatology and Venereology of La Rabta Universal Hospital in Tunis (Tunisia). Three hundred ninety-two patients were examined to establish the presence of clinical signs of tinea unguium and/or tinea pedis. The questionnaire allowed documentation of potential predisposing factors for foot mycoses, age, sex, diabetes, vascular disease, immunosuppressive drug treatment, psoriasis, fungal infection of the skin, dermatological pathology, associated fingernails onychomycosis, family history of foot mycoses, ritual religious washing, physical activities, used shoes, occlusive shoes, using of publics showers, swimming pools, smoking, walking barefoot, thermal station, pedicure, and the application of henna. Also, the type of tinea pedis (interdigital, hyperkeratosis, and dyshidrosis) and the type of tinea unguium (DLSO, PSO, WSO, or TDO) were documented.

cal activities, used shoes, occlusive shoes, using of publics showers, swimming pools, smoking, walking barefoot, thermal station, pedicure, and the application of henna. Also, the type of tinea pedis (interdigital, hyperkeratosis, and dyshidrosis) and the type of tinea unguium (DLSO, PSO, WSO, or TDO) were documented. Clinical specimens of skin scrapings and nail clippings were collected in sterile Petri dishes for direct examination and culture. All specimens were submitted to a microscopic examination in Chlorazol noir (Sigma-Aldrich, Germany) solution and inoculated into Sabouraud chloramphenicol dextrose agar with and without cycloheximide (Biorad, France) all in duplicate. The culture was incubated at 27°C and examined after 48–72 h for yeast detection and every four days for at least four weeks for fungal detection. The identification of filamentous fungi was based on macroscopic and microscopic examination in the Lactophenol Cotton Blue (Sigma-Aldrich, Germany). The identification of yeast was carried out using the API ID 32C system (Biomerieux, France). The mycological examination was considered to be positive if direct microscopic examination and/or culture were positive. In the absence of any dermatophyte or yeast growth, a mold was only considered to be the causal agent of onychomycosis when the culture was repeated on two separate occasions.

(Biomerieux, France). The mycological examination was considered to be positive if direct microscopic examination and/or culture were positive. In the absence of any dermatophyte or yeast growth, a mold was only considered to be the causal agent of onychomycosis when the culture was repeated on two separate occasions. Statistical analysis was performed with SPSS software (Statistical Package for Social Scientists version 20.0, SPSS, Inc., Armonk, NY). The Chi-square (χ2) was used to calculate significant differences in characteristics between patients. Differences with p < 0.05 were considered statistically significant. 3. Results A total of 392 patients from various regions in Tunisia were included in this study, 125 males (31.88%) and 267 females (68.11%), with an age range between 3 and 85 years and an average age of 44.7 years. The diagnosis of foot mycosis was confirmed through a mycological diagnosis in 346 (88.26%) cases; the frequency was higher in females (67.05%) compared with males (32.94%) but this prevalence according to the sex was not statistically significant (p = 0.217) (Table 1). As shown in Figure 1, the frequency of foot mycoses according to the age groups revealed that the patients most commonly infected were between 41 and 50 years (23.1%) followed by those between 51 and 60 years (21.9%) but the differences were not statistically significant (p = 0.0658; p = 0.71, resp.). However, the prevalence was less frequent in children less than 10 years old (0.8%) and this prevalence was significant (p = 0.0126).

only infected were between 41 and 50 years (23.1%) followed by those between 51 and 60 years (21.9%) but the differences were not statistically significant (p = 0.0658; p = 0.71, resp.). However, the prevalence was less frequent in children less than 10 years old (0.8%) and this prevalence was significant (p = 0.0126). Related to the site of infection, we noted that tinea unguium was confirmed in 268 subjects (77.4%) and tinea pedis was confirmed in 78 cases (22.5%). The subtype most frequently observed in tinea pedis was plantar keratoderma in 70 cases (89.7%), followed by interdigital 23 cases (29.4%). 57 cases (73.07%) of the subjects whom presented with tinea pedis have toenail onychomycosis (Table 2). Clinical patterns of foot mycoses are cited in Table 3; DLSO represent the most common clinical form of tinea unguium (64.3%) followed by TDO (15.6%) and SWO (12.9%). The big toenail was the most infected in 114 cases (35.07%), bilateral nail infection was observed in 55 cases (16.9%), and multiple toes were affected in 103 cases (31.6%). For tinea pedis plantar hyperkeratosis form was observed in 44 cases (62.8%) and plantar dyshidrosis in 26 cases (37.1%). A total of 485 samples were collected from all patients; the direct microscopic examination was positive in 385 specimens (79.3%) showing filaments in 371 cases (96.3%), yeast and pseudomycelium in two cases (0.5%), and both filaments with spores in 12 cases (3.1%).

(62.8%) and plantar dyshidrosis in 26 cases (37.1%). A total of 485 samples were collected from all patients; the direct microscopic examination was positive in 385 specimens (79.3%) showing filaments in 371 cases (96.3%), yeast and pseudomycelium in two cases (0.5%), and both filaments with spores in 12 cases (3.1%). We have obtained 299 positive cultures (61.6%), including dermatophytes in 211 cases (70.5%), yeasts in 53 (17.7%), NDMs in 24 cases (8.02%), and mixed culture (dermatophytes + yeast) in 11 cases (3.6%). The most frequently isolated dermatophyte was Trichophyton rubrum (98.1%), followed by T. violaceum, T. tonsurans, T. verrucosum, and T. interdigitale with 0.47% for each species. While Candida parapsilosis was the most isolated yeast (60.3%), also Trichosporon spp. were isolated (3.7%). The remaining were due to NDMs like Fusarium (29.1%), Penicillium (25%), Aspergillus (20.8%), Scopulariopsis (16.6%), and Scytalidium (8.3%). In mixed culture, C. parapsilosis was most frequently detected with T. rubrum (72.7%) (Table 4). Since our survey was conducted during one year, we have seen a lower frequency of patients in the winter (16.5%) and most cases were observed in spring (33.4%) and the summer (21.1%) (Figure 2).

We have obtained 299 positive cultures (61.6%), including dermatophytes in 211 cases (70.5%), yeasts in 53 (17.7%), NDMs in 24 cases (8.02%), and mixed culture (dermatophytes + yeast) in 11 cases (3.6%). The most frequently isolated dermatophyte was Trichophyton rubrum (98.1%), followed by T. violaceum, T. tonsurans, T. verrucosum, and T. interdigitale with 0.47% for each species. While Candida parapsilosis was the most isolated yeast (60.3%), also Trichosporon spp. were isolated (3.7%). The remaining were due to NDMs like Fusarium (29.1%), Penicillium (25%), Aspergillus (20.8%), Scopulariopsis (16.6%), and Scytalidium (8.3%). In mixed culture, C. parapsilosis was most frequently detected with T. rubrum (72.7%) (Table 4). Since our survey was conducted during one year, we have seen a lower frequency of patients in the winter (16.5%) and most cases were observed in spring (33.4%) and the summer (21.1%) (Figure 2). Considering the possible risk factors, we noted that the high prevalence was observed in patients who practice ritual ablutions (56.6%) followed by communal shower (50.5%) and family history of foot mycoses (28.6%) but there was no statistically significant association between these factors and foot infection (p = 0.41, 0.631, and 0.246, resp.). However, we noted a significant association between foot mycoses and nail trauma (26.5%; p = 0.019), wearing used shoes (26.3%; p = 0.001), antifungal drugs (25.7%; p = 0.013), physical activities (14.7%; p = 0.049), occlusive shoes (13.2%; p = 0.008), swimming pools (8.09%; p = 0.045), attending thermal station (8.3%; p = 0.021), pedicure (14.1%; p = 0.006), associated fingernails onychomycosis (7.5%; p = 0.010), and those taking immunosuppressive drugs (5.4%; p = 0.018).

(25.7%; p = 0.013), physical activities (14.7%; p = 0.049), occlusive shoes (13.2%; p = 0.008), swimming pools (8.09%; p = 0.045), attending thermal station (8.3%; p = 0.021), pedicure (14.1%; p = 0.006), associated fingernails onychomycosis (7.5%; p = 0.010), and those taking immunosuppressive drugs (5.4%; p = 0.018). In our study, we did not find a significant association between foot mycoses and diabetes, vascular disease, psoriasis, fungal infection of the skin, dermatological pathology, smoking, obesity, walking barefoot, and the application of henna (p > 0.05) (Table 5). 4. Discussion Foot mycosis is the most common superficial infection and represents a major public health problem over the world. Many epidemiological studies have reported the high frequency of this fungal infection, but the prevalence varies with many factors like geographic and demographic parameters and the number of selected population. To our knowledge, the latest epidemiological studies about foot mycoses in Tunisia were established by El Fekih et al. [12] in Tunisia between January and April 2009 and by Dhib et al. [13] during 22 years from 1986 to 2007 in the center of Tunisia (Sousse).

demographic parameters and the number of selected population. To our knowledge, the latest epidemiological studies about foot mycoses in Tunisia were established by El Fekih et al. [12] in Tunisia between January and April 2009 and by Dhib et al. [13] during 22 years from 1986 to 2007 in the center of Tunisia (Sousse). In the present study, the prevalence of tinea unguium and tinea pedis in the population studied were 77.4% and 22.5%, respectively; females were more commonly affected than males which agree with some reports [14–16]. But there was no significant relationship in the occurrence of foot mycoses with respect to the sex and these results are in accordance with Dhib et al. [13]. This may be caused by aesthetics reasons such as repeated aggressive pedicure and manicure, frequent housework, and using detergents that cause nail trauma and generally females consulted more frequently for onychomycosis. However, several studies concluded that males are more infected than females due to the fact that males are more exposed to nail trauma and using occlusive footwear [1, 17].

e and manicure, frequent housework, and using detergents that cause nail trauma and generally females consulted more frequently for onychomycosis. However, several studies concluded that males are more infected than females due to the fact that males are more exposed to nail trauma and using occlusive footwear [1, 17]. The frequency of tinea pedis and tinea unguium increased gradually with age; a maximum prevalence was seen in adults aged between 31 and 60 years. These results were confirmed by many studies [10, 12, 18], and this increase may be explained by many conditions such as full-time work activities, frequent nail trauma, reduced nail growth, and inadequate foot care [19]. However, the frequency is less prevalent in the elderly aged between 71 and 80 years and >80 years; this is in agreement with a study reported in Rio Grande do Sul, Brazil [20]. This decreasing frequency can be due to the negligence of old people who do not give importance to nail infections. Our results also showed that children are rarely infected with foot mycoses; this frequency is in accordance with results observed in school children in Spain [21] and in Turkey [22]. Mycoses infections in children can be due to several factors including the difference in the nail plate, the rapid nail growth, and less exposure to fungal infection than adults.

rarely infected with foot mycoses; this frequency is in accordance with results observed in school children in Spain [21] and in Turkey [22]. Mycoses infections in children can be due to several factors including the difference in the nail plate, the rapid nail growth, and less exposure to fungal infection than adults. Tinea pedis is known as the significant reservoir of other dermatophytes in the body and can be a cause of tinea unguium [23]. In the current study, tinea pedis was associated with tinea unguium in more than half of cases (73.07%); this rate was higher than reported in USA [24], in Tokyo [17], and even in another study in Tunisia [12]. This association confirmed the hypothesis that the toenails were infected by toe-webs. Various clinical patterns of onychomycosis have been reported in the literature. In this work, DLSO was the most frequent clinical form as well as in other studies carried out in Turkey and in Tunisia [25–27]. The most affected toes were the big ones; this observation was expected because of the slow growth of the nail which facilitates the invasion of the pathogenic fungal.

rted in the literature. In this work, DLSO was the most frequent clinical form as well as in other studies carried out in Turkey and in Tunisia [25–27]. The most affected toes were the big ones; this observation was expected because of the slow growth of the nail which facilitates the invasion of the pathogenic fungal. In investigating the causative agents of tinea pedis and tinea unguium, we found that the most common isolated pathogens were dermatophytes [28, 29]; among them, T. rubrum was the most common causative agent. These results are similar with other studies [13, 30–32] and are interpreted that T. rubrum is a virulent anthropophilic dermatophyte producing arthrospores which have the capacity to persist on the floor surface and on shoes. The second agent responsible for foot mycoses is yeasts, with a high frequency of C. parapsilosis. This agrees with the study of El Fekih et al. [12] and can be explained by the fact that C. parapsilosis represents a saprophyte yeast of the human skin. In our results, the anthropophilic T. violaceum was isolated from one patient with tinea pedis who had no history of tinea capitis, whereas this species has been classified as the second and the third etiological agent of onychomycosis in few cases as related agent to tinea capitis [26, 33].

In investigating the causative agents of tinea pedis and tinea unguium, we found that the most common isolated pathogens were dermatophytes [28, 29]; among them, T. rubrum was the most common causative agent. These results are similar with other studies [13, 30–32] and are interpreted that T. rubrum is a virulent anthropophilic dermatophyte producing arthrospores which have the capacity to persist on the floor surface and on shoes. The second agent responsible for foot mycoses is yeasts, with a high frequency of C. parapsilosis. This agrees with the study of El Fekih et al. [12] and can be explained by the fact that C. parapsilosis represents a saprophyte yeast of the human skin. In our results, the anthropophilic T. violaceum was isolated from one patient with tinea pedis who had no history of tinea capitis, whereas this species has been classified as the second and the third etiological agent of onychomycosis in few cases as related agent to tinea capitis [26, 33]. Molds are cosmopolitan filamentous fungi; most of them are saprophyte and can be contaminants; however they become opportunistic under unfavorable conditions. In addition to the causative dermatophytes and yeasts, NDMs have been described as etiological agents of foot mycoses. Traditionally, these molds have been considered as secondary pathogens of nails which affect a keratin already degraded and their frequency rates between 1.45 and 17.6% [34]. The primary molds that cause onychomycosis are species belonging to the genus Scopulariopsis, Aspergillus, and Fusarium [35]. In our survey, we found a low incidence of NDMs and the most prevalent species were Fusarium sp. (29.1%). This result agreed with the study reported in America showing that Fusarium spp. seem to be the most frequent agents [36].

omycosis are species belonging to the genus Scopulariopsis, Aspergillus, and Fusarium [35]. In our survey, we found a low incidence of NDMs and the most prevalent species were Fusarium sp. (29.1%). This result agreed with the study reported in America showing that Fusarium spp. seem to be the most frequent agents [36]. In contrast, a study reported in Italy showed that Scopulariopsis brevicaulis was the dominant causative mold [37], although a recent paper from Morocco reports the increasing frequency of Aspergillus onychomycosis [38]. In the present work, two cases of onychomycosis were due to Scytalidium species. These molds caused lesions similar to those engendered by dermatophytes, called pseudodermatophytes, and widespread in the environment especially in tropical areas [39]. Walking barefoot represents the main risk factor of infections caused by these species because of their telluric reservoir. The presence of NDMs in foot mycoses may be related to many factors that predispose the development of nail infections, such as direct contact with the soil by walking barefoot, wearing open shoes sandals, practicing sports, and trauma of the nail. Actually onychomycosis caused by NDMs is more prevalent in tropical and subtropical regions with a hot and humid climate making them endemic areas [34, 37]. The contaminated culture can be detected. This may be related to the nonpathogenic nature of the fungi that infects the nail or the presence of a saprophyte mold that inhibits the growth of the pathogenic fungi [11].

Actually onychomycosis caused by NDMs is more prevalent in tropical and subtropical regions with a hot and humid climate making them endemic areas [34, 37]. The contaminated culture can be detected. This may be related to the nonpathogenic nature of the fungi that infects the nail or the presence of a saprophyte mold that inhibits the growth of the pathogenic fungi [11]. In the current study, we note that more cases of fungal infections of the foot were observed in the spring and the summer. The higher frequency during certain seasons can be related to the wearing of occlusive shoes in warm climates, causing heat of the feet which causes maceration and hyperhidrosis, considered as a risk factor of developing foot mycoses. In another way, patients become more aware that the duration of antifungal therapy requires a long period of time and prepare to wear summer shoes because they are more interested in beautiful feet. Considering the risk factors of foot mycoses, we found a significant association with patients who practice physical activities [32], wear used shoes, or have frequent nail trauma [19] and with patients who receive immunosuppressive therapy [40].

In the current study, we note that more cases of fungal infections of the foot were observed in the spring and the summer. The higher frequency during certain seasons can be related to the wearing of occlusive shoes in warm climates, causing heat of the feet which causes maceration and hyperhidrosis, considered as a risk factor of developing foot mycoses. In another way, patients become more aware that the duration of antifungal therapy requires a long period of time and prepare to wear summer shoes because they are more interested in beautiful feet. Considering the risk factors of foot mycoses, we found a significant association with patients who practice physical activities [32], wear used shoes, or have frequent nail trauma [19] and with patients who receive immunosuppressive therapy [40]. Many other possible risk factors can be related to fungal infections of the foot but our study did not show a significant association; it was found to be most common in persons with family history of fungal infection of the foot [12], having peripheral vascular diseases [41]. Although foot mycoses can be linked to many chronic diseases like diabetes [42], HIV infection [43], and psoriasis [44], this can be explained that people with chronic infections are now more attentive to their health due to awareness raising sessions.

fection of the foot [12], having peripheral vascular diseases [41]. Although foot mycoses can be linked to many chronic diseases like diabetes [42], HIV infection [43], and psoriasis [44], this can be explained that people with chronic infections are now more attentive to their health due to awareness raising sessions. Interestingly, we found a high prevalence of subjects who practice ritual washing. Firstly, this can be explained by the religious custom of ablutions five times every day which can cause maceration of the feet which represents a risk factor of fungal penetration through the stratum corneum of the skin. It also can be related to the spread of fungal species in areas used for washing and in prayer carpets of the mosques; this has been confirmed in other studies [12, 45–47]. Moreover, in the current study, 50.5% of patients with foot mycoses attend communal showers and bathing. This high frequency may be the consequence of the culture and the tradition of Tunisian population to frequent hammams, which are humid and warm locations that are a source of fungal contagion; this also has been found in Algerian population [10, 48]. We observed that some patients have onyxis of both fingers and toes. The association of fingernails onychomycosis can be a risk factor for developing a foot infection that is reported in previous surveys [13, 19]. This association can be related to autoinfection that represents an important source of transmission to another location of the same body [49].

nyxis of both fingers and toes. The association of fingernails onychomycosis can be a risk factor for developing a foot infection that is reported in previous surveys [13, 19]. This association can be related to autoinfection that represents an important source of transmission to another location of the same body [49]. In the present work, 25.7% subjects with foot mycosis have taken antifungal therapy. This finding may be related to the recurrent infection [50] that can be due to various causes which include lack of diagnosis, misidentification of the causative pathogen, and inappropriate choice of antifungal treatment. On the other hand, it can be related to resistant fungal species or the presence of dormant arthroconidia in the nail bed as a reservoir for recurrent infection [51]. 5. Conclusion The epidemiological profile of fungal foot infections seems to be related to age, life style, and the presence of comorbidities. Our study shows that the prevalence of these infections is common in the general population of Tunisia, and the frequency is higher than reported in Maghreb, African, and European countries. Our data can be useful to eradicate these infections and provide further measures regarding the personal hygiene and education about prophylaxis in order to reduce the risk factors of tinea pedis and tinea unguium. Acknowledgments This work was supported by funds from the Ministry of Higher Education and Scientific Research of Tunisia (LR16ES05). The authors gratefully acknowledge the laboratory staff at La Rabta Hospital for the skilled technical assistance.

5. Conclusion The epidemiological profile of fungal foot infections seems to be related to age, life style, and the presence of comorbidities. Our study shows that the prevalence of these infections is common in the general population of Tunisia, and the frequency is higher than reported in Maghreb, African, and European countries. Our data can be useful to eradicate these infections and provide further measures regarding the personal hygiene and education about prophylaxis in order to reduce the risk factors of tinea pedis and tinea unguium. Acknowledgments This work was supported by funds from the Ministry of Higher Education and Scientific Research of Tunisia (LR16ES05). The authors gratefully acknowledge the laboratory staff at La Rabta Hospital for the skilled technical assistance. Disclosure The authors alone are responsible for the content and the writing of the paper. Conflicts of Interest The authors report no conflicts of interest. Figure 1 Frequency of foot mycoses according to age groups. Percentage of patients with foot mycoses according to different age groups. p value is mentioned under each histogram. Figure 2 Seasonal evolution of fungal infections of the feet. Percentage of patients consulted during the four seasons. Table 1 Distribution of foot mycosis according to sex. Males Females p value Number % Number % Positive 114 91.20 232 86.89 0.217 Negative 11 8.80 35 13.10 Total 125 31.88 267 68.11 Table 2 Distribution of foot mycoses according to anatomic sites.

Figure 2 Seasonal evolution of fungal infections of the feet. Percentage of patients consulted during the four seasons. Table 1 Distribution of foot mycosis according to sex. Males Females p value Number % Number % Positive 114 91.20 232 86.89 0.217 Negative 11 8.80 35 13.10 Total 125 31.88 267 68.11 Table 2 Distribution of foot mycoses according to anatomic sites. Nature of lesion Site of infection Number of patients Direct examination Culture (+) (−) (+) (−) Tinea unguium Nails 268 253 15 201 67 Tinea pedis Interdigital 3 2 1 3 0 Plantar 14 13 1 6 8 Interdigital and plantar 4 4 0 3 1 Interdigital and nails 5 5 0 5 0 Plantar and nails 41 41 0 33 8 Interdigital, plantar, and nails 11 10 1 7 4 Total   346 328 18 258 88 (+): positive; (−): negative. Table 3 Clinical patterns of foot mycoses. Clinical patterns Number of cases (%) Tinea unguium 325 (100) DLSO 209 (64.3) PSO 23 (7.07) SWO 42 (12.9) TDO 51 (15.6) Tinea pedis 78 (100) Plantar hyperkeratosis 44 (62.8) Plantar dyshidrosis 26 (37.1) Interdigital 23 (29.4) DLSO, distal lateral subungual onychomycosis; PSO, proximal subungual onychomycosis; SWO, superficial white onychomycosis; TDO, total dystrophic onychomycosis. Table 4 Etiological agents responsible for foot mycoses. Causative agents Tinea unguium Tinea pedis Total (%) Dermatophytes 165 46 211 (70.5) T. rubrum 163 44 207 (98.1) T. violaceum — 1 1 (0.47) T. tonsurans 1 — 1 (0.47) T. verrucosum — 1 1 (0.47) T. interdigitale 1 — 1 (0.47) Yeasts 53 — 53 (17.7)

Tinea pedis 78 (100) Plantar hyperkeratosis 44 (62.8) Plantar dyshidrosis 26 (37.1) Interdigital 23 (29.4) DLSO, distal lateral subungual onychomycosis; PSO, proximal subungual onychomycosis; SWO, superficial white onychomycosis; TDO, total dystrophic onychomycosis. Table 4 Etiological agents responsible for foot mycoses. Causative agents Tinea unguium Tinea pedis Total (%) Dermatophytes 165 46 211 (70.5) T. rubrum 163 44 207 (98.1) T. violaceum — 1 1 (0.47) T. tonsurans 1 — 1 (0.47) T. verrucosum — 1 1 (0.47) T. interdigitale 1 — 1 (0.47) Yeasts 53 — 53 (17.7) C. parapsilosis 32 — 32 (60.3) C. tropicalis 1 — 1 (1.8) C. metapsilosis 2 — 2 (3.7) C. famata 2 — 2 (3.7) C. lusitaniae 1 — 1 (1.8) C. pelliculosa 1 — 1 (1.8) C. sake 2 — 2 (3.7) C. guilliermondii 5 — 5 (9.4) Trichosporon asahii 1 — 1 (1.8) Trichosporon mucoides 1 — 1 (1.8) Rhodotorula 1 — 1 (1.8) Other Candida sp. 4 — 4 (7.5) NDMs 24 — 24 (8.02) Aspergillus sp. 5 — 5 (20.8) Fusarium sp. 7 — 7 (29.1) Scopulariopsis brevicaulis 4 — 4 (16.6) Penicillium sp. 6 — 6 (25) Scytalidium 2 — 2 (8.3) Mixed culture 10 1 11 (3.6) T. rubrum + C. parapsilosis 7 1 8 (72.7) T. rubrum + C. metapsilosis 1 — 1 (9.09) T. rubrum + Trichosporon sp. 1 — 1 (9.09) T. rubrum + Rhodotorula 1 — 1 (9.09) C., Candida; T., Trichophyton; NDMs, non-dermatophyte molds. Table 5 Possible risk factors associated with foot mycosis based on the questionnaire. Risk factors Patients with foot mycosis p value Number % Chronic diseases

T. rubrum + C. parapsilosis 7 1 8 (72.7) T. rubrum + C. metapsilosis 1 — 1 (9.09) T. rubrum + Trichosporon sp. 1 — 1 (9.09) T. rubrum + Rhodotorula 1 — 1 (9.09) C., Candida; T., Trichophyton; NDMs, non-dermatophyte molds. Table 5 Possible risk factors associated with foot mycosis based on the questionnaire. Risk factors Patients with foot mycosis p value Number % Chronic diseases Diabetes history Present 41 11.8 0.815 Absent 305 88.1 Peripheral vascular disease Present 76 21.9 0.293 Absent 270 78.03 Immunosuppressive drugs Present 19 5.4 0.018 Absent 327 94.5 Skin disorders Psoriasis Present 6 1.7 0,368 Absent 340 98.2 Fungal infection of the skin Present 7 2.02 0,330 Absent 339 97.9 Dermatological pathology Present 11 3.1 0.220 Absent 335 96.8 Associated fingernails onychomycosis Present 26 7.5 0.010 Absent 320 92.4 Life style

Diabetes history Present 41 11.8 0.815 Absent 305 88.1 Peripheral vascular disease Present 76 21.9 0.293 Absent 270 78.03 Immunosuppressive drugs Present 19 5.4 0.018 Absent 327 94.5 Skin disorders Psoriasis Present 6 1.7 0,368 Absent 340 98.2 Fungal infection of the skin Present 7 2.02 0,330 Absent 339 97.9 Dermatological pathology Present 11 3.1 0.220 Absent 335 96.8 Associated fingernails onychomycosis Present 26 7.5 0.010 Absent 320 92.4 Life style Family history of foot mycosis Present 99 28.6 0.244 Absent 247 71.3 Ritual washing Present 196 56.6 0.410 Absent 150 43.3 Physical activities Present 51 14.7 0.049 Absent 295 85.2 Wearing used shoes Present 91 26.3 0.001 Absent 255 73.6 Occlusive shoes Present 46 13.2 0,008 Absent 300 86.8 Nail trauma Present 92 26.5 0.019 Absent 254 73.4 Swimming pools Present 28 8.09 0,045 Absent 318 91.9 Communal shower Present 175 50.5 0.631 Absent 171 49.4 Smoking Present 13 3.7 0,181 Absent 333 96.2 Obesity Present 8 2.3 0.297 Absent 338 97.6 Walking barefoot Present 34 9.82 0,524 Absent 312 90.1 Thermal station Present 29 8.3 0.021 Absent 317 91.6 Pedicure Present 49 14.1 0,006 Absent 297 85.9 Application of henna Present 6 1.73 0.832 Absent 340 98.2 Antifungal therapy Present 89 25.7 0.013 Absent 257 74.2

1. Introduction Buruli ulcer (BU) is a neglected emerging disease that has recently been reported in some countries as the second most frequent mycobacterial disease in humans after tuberculosis [1–3]. BU continues to be one of the most debilitating cutaneous diseases causing significant morbidity. The disease is characterized by severe subcutaneous necrotic lesions that lead to chronic opened sores and ulcerations, ultimately affecting the bone in extreme cases [4]. Mycolactone, a secreted exotoxin, is the only virulence factor identified to date for Mycobacterium ulcerans (MU) [5]. During the last two decades, there has been a reemergence of BU across diverse regions of the world [3, 6, 7]. Its prevalence has increased and currently is seen in over 33 countries worldwide [8]. Although the distribution of BU is global and affects people of all ages, the burden of this disease is most severe in West and Central Africa, as well as some parts of Australia [1, 9, 10]. More than 30,000 cases of BU have been reported in Africa over the last decade and the West African region accounts for more than 67% of the reported cases [6].

is global and affects people of all ages, the burden of this disease is most severe in West and Central Africa, as well as some parts of Australia [1, 9, 10]. More than 30,000 cases of BU have been reported in Africa over the last decade and the West African region accounts for more than 67% of the reported cases [6]. The environmental pathogen Mycobacterium ulcerans (MU) is the etiologic agent of Buruli ulcer [11]. Merritt et al. [1] provided a series of hierarchical criteria analogous to Koch's postulates and/or the Bradford Hill guidelines emphasizing epidemiological/ecological association and the use of logical inference for establishing cause and effect in biological disease transmission. They further discussed the application of this process to indictment of insect vectors for transmission of MU. However, the mode of transmission of MU from the risk environments to humans remains unknown and its reservoirs in the environment are still being uncovered [1]. The direct transmission of MU from human-to-human is extremely rare and cases usually occur in proximity to slow moving or stagnant bodies of water and among rural and economically deprived populations [12–18]. Recent studies in Australia have demonstrated that mosquitoes may be potential reservoirs or vectors of BU [7, 19–23]. Similarly, a recent study conducted in an endemic area of Cameroon revealed the presence of MU molecular markers in hematophagous families of insects like Culicidae (mosquito's family), Ceratopogonidae, and Psychodidae [24]. However, a similar study in an endemic area of Benin did not detect MU molecular markers in mosquito species [25]. An experimental laboratory study conducted by Wallace et al. [26] also failed to confirm the implication of mosquitoes as biological vectors in the transmission of BU. These recent studies highlight controversial concerns whether mosquitoes actually play a role in the transmission dynamics of BU. Mosquitoes are the most important group of insects involved in the spread of human and animal diseases [27]. One hypothesis is that they could transmit MU to humans. However, there is no scientific or historic precedent for mosquitoes transmitting a bacterium to host in any disease system, either directly or mechanically [1]. In the vector ecology, they may serve as biological vectors and hosts for pathogen replication, or mechanical vectors carrying organisms from hosts to hosts without serving as a site of replication [26, 27].

dent for mosquitoes transmitting a bacterium to host in any disease system, either directly or mechanically [1]. In the vector ecology, they may serve as biological vectors and hosts for pathogen replication, or mechanical vectors carrying organisms from hosts to hosts without serving as a site of replication [26, 27]. This last hypothesis has recently been reinforced by Wallace et al. [28] who reported a biologically plausible mechanical transmission mode of BU via natural or anthropogenic skin punctures (trauma). These authors further highlighted that a significant low quantity of MU delivered beneath the skin surface of animal (BALB/C mice) by a minor injury created by mosquitoes might cause BU in return [28]. Previously in 1974, Meyers et al. [29] reported that skin trauma could be an important mode of transmitting MU infections or of introducing MU into the dermis of subcutaneous tissue from superficially contaminated skin. However, Williamson et al. [30] recently established that abrasions (trauma) of the skin in Guinea pig models and subsequent application of MU are not sufficient enough to cause an ulcer. Mosquitoes contamination or colonization by MU remains an event which has only been reported in Australia and which could vary according to mosquito species. As BU infections occur in humid areas of Africa where high densities of mosquito species are recorded, there is a need to further investigate whether they could be involved in the transmission cycle of BU in African settings and more specifically in Benin.

n Australia and which could vary according to mosquito species. As BU infections occur in humid areas of Africa where high densities of mosquito species are recorded, there is a need to further investigate whether they could be involved in the transmission cycle of BU in African settings and more specifically in Benin. In this study, we tested the hypothesis of the implication of mosquito species in the transmission of MU in an endemic area of Sedje-Denou in the Southern Benin. We further evaluated whether mosquitoes could pick MU bacteria from water breeding sites during larval developmental stages leading to colonization and whether colonization continues into the adult stage where they become infective to humans (vertical transmission of MU by mosquitoes). Based on these assumptions, we screened wild mosquitoes populations collected from three endemic villages found in Sedje-Denou for molecular targets of MU. Coupled to this field based activity, we also investigated the potential for vertical transmission of MU within mosquitoes populations using the laboratory strain Anopheles kisumu. 2. Methods 2.1. Ethical Considerations This research which was mainly laboratory based received administrative clearance from the International Institute of Tropical Agriculture (IITA). In addition, the community consent was obtained prior to mosquitoes sampling in the three villages of Sedje-Denou.

In this study, we tested the hypothesis of the implication of mosquito species in the transmission of MU in an endemic area of Sedje-Denou in the Southern Benin. We further evaluated whether mosquitoes could pick MU bacteria from water breeding sites during larval developmental stages leading to colonization and whether colonization continues into the adult stage where they become infective to humans (vertical transmission of MU by mosquitoes). Based on these assumptions, we screened wild mosquitoes populations collected from three endemic villages found in Sedje-Denou for molecular targets of MU. Coupled to this field based activity, we also investigated the potential for vertical transmission of MU within mosquitoes populations using the laboratory strain Anopheles kisumu. 2. Methods 2.1. Ethical Considerations This research which was mainly laboratory based received administrative clearance from the International Institute of Tropical Agriculture (IITA). In addition, the community consent was obtained prior to mosquitoes sampling in the three villages of Sedje-Denou. 2.2. Study Area This study was carried out in three endemic communities (Agbahounsou, Agodenou, and Agongbo) of Sedje-Denou (6°32′N and 2°13′E) in the Southern Benin (Figure 1). One nonendemic village, Tanongou (10°48′N and 1°26′E), in the Northern Benin was selected as a negative control village for data comparison. Sedje-Denou (also named Sedje) is located in the Commune of Ze which is the second most endemic locality in Benin with a reported prevalence of 450 cases of BU per 100,000 inhabitants [14]. The presence of rivers and wetlands make this locality an appropriate environment for BU. According to Wagner et al. [31], drainage basins as well as forest land cover with variable wetness patterns are prolific for the growth of MU and associated with higher BU disease prevalence rates. These patterns could also influence the distribution and abundance of vectors, or mediating vector-human interactions. The climate at Sedje is a subequatorial type with two discontinuous dry and wet seasons. The annual average rainfall measures 1,000 mm with an annual average temperature of 24°C and a mean altitude of 20 m. The population of 5,496 inhabitants are distributed into six different villages. Sedje is a rural area where agricultural works being the predominant occupation could contribute to increased exposure to MU due to the close spatial proximity with the risk environments [31].

mperature of 24°C and a mean altitude of 20 m. The population of 5,496 inhabitants are distributed into six different villages. Sedje is a rural area where agricultural works being the predominant occupation could contribute to increased exposure to MU due to the close spatial proximity with the risk environments [31]. Tanongou is also a rural locality under the Department of Atakora in Northern Benin (Figure 1). This village is administratively subdivided into two close villages named Tanongou 1 and Tanongou 2. BU epidemiological data in Benin show that this locality is a nonendemic area for the disease. The climate is a wet Sudanese type with one long dry season (November to May) and a short rainy season (June to October). This region is dominated by hills of up to 800 m of altitude and several small water bodies, which makes the region colder and relatively wet. Annual rainfall ranges from 1200 to 1300 mm per year, the vegetation is partially of wet savanna type, and the temperature in this part of the country ranges between 23 and 31°C. Agriculture is the prominent activity of the region.

titude and several small water bodies, which makes the region colder and relatively wet. Annual rainfall ranges from 1200 to 1300 mm per year, the vegetation is partially of wet savanna type, and the temperature in this part of the country ranges between 23 and 31°C. Agriculture is the prominent activity of the region. 2.3. Sampling of Mosquito Species in BU Endemic and Nonendemic Areas 2.3.1. Sampling of Adult Mosquitoes Field surveys for mosquitoes collections were conducted during rainy seasons at the 3 villages of Sedje-Denou from 2014 to 2016. Similarly, mosquitoes samples were collected at Tanongou (Tanongou 1 + Tanongou 2) during rainy seasons as well. Adult mosquitoes were caught indoors using insecticide spraying technique which is one of the effective methods for collecting indoors resting mosquitoes [32]. Mosquitoes were harvested about twenty minutes after house spraying. They were safely transferred into Petri dishes labeled with room/house references and were taken to the laboratory. In the laboratory, each mosquitoes sample was morphologically identified using Edward identification keys [33]. Mosquitoes were identified to genus and to species. No molecular test was performed for mosquito identification. For each identified species (Anopheles gambiae s.l., Culex quinquefasciatus, Mansonia africana, and Aedes aegypti), pools of 10 mosquitoes each were prepared and kept at −20°C in Eppendorf tubes filled with silica gel. Mosquitoes from Tanongou 1 and Tanongou 2 were pooled and considered as from a single control village of Tanongou.

tified species (Anopheles gambiae s.l., Culex quinquefasciatus, Mansonia africana, and Aedes aegypti), pools of 10 mosquitoes each were prepared and kept at −20°C in Eppendorf tubes filled with silica gel. Mosquitoes from Tanongou 1 and Tanongou 2 were pooled and considered as from a single control village of Tanongou. 2.3.2. Sampling of Mosquito Larvae Mosquito larvae were collected from temporal, semipermanent, and permanent breeding areas using the WHO protocol [34]. Collected larvae were transported to the laboratory where they were morphologically identified and pooled as were the adults. Larvae pools were prepared and stored at −20°C in 70% alcohol.

quito Larvae Mosquito larvae were collected from temporal, semipermanent, and permanent breeding areas using the WHO protocol [34]. Collected larvae were transported to the laboratory where they were morphologically identified and pooled as were the adults. Larvae pools were prepared and stored at −20°C in 70% alcohol. 2.4. Molecular Identification of MU in Mosquitoes Samples 2.4.1. Extraction of Genomic DNA Genomic DNA was extracted from a total of 721 pools of mosquitoes samples (adult and larvae) using the phenol/chloroform extraction method described by Sambrook and Russell [35]. Several types of controls were put in place to guide against false positive and negative results. To reduce cross-contaminations, extractions were conducted in batches of 10 pools and the 10 pools completely processed (extraction and PCRs) before moving back to a new set of extractions. Negative controls (nuclease-free water, Sigma-Aldrich) were added at a frequency of 10% (1 control per batch of extraction) to monitor potential cross-contaminations. Pooled mosquitoes samples were ground using an electric grinder in sterile 100 µl 1x PBS and the homogenates were suspended in 300 µl preheated lysis buffer made of 5 M NaCl, 0.5 M EDTA, 1 M Tris-HCl (pH 8.0), 10% SDS, and proteinase K (Qiagen, Hilden). The mixture was heated at 60°C for one hour and DNA extracted with phenol/chloroform/isoamyl acid in the ratio 25 : 24 : 1. This was briefly mixed by a pulse vortex and centrifuged for 2 min at 13,000g. The DNA was precipitated by adding 2 volumes of pure ethanol and the mixture was incubated for 2 hours and centrifuged 10 min at 13,000g. The DNA was washed by 70% cold ethanol, dried 20 min at room temperature, and eluted in 50 µl of nuclease-free water (Sigma-Aldrich).

lse vortex and centrifuged for 2 min at 13,000g. The DNA was precipitated by adding 2 volumes of pure ethanol and the mixture was incubated for 2 hours and centrifuged 10 min at 13,000g. The DNA was washed by 70% cold ethanol, dried 20 min at room temperature, and eluted in 50 µl of nuclease-free water (Sigma-Aldrich). 2.4.2. Detection of MU DNA in Mosquitoes Samples Using TaqMan qPCRs The TaqMan IS2404 qPCR analysis described by Fyfe et al. [36] was performed on extracted mosquitoes DNA samples to detect Mycobacterium DNA in these samples. A total of 404 pools of DNA samples from adult mosquitoes and a total of 317 pools of DNA samples from mosquitoes larvae from the 4 villages (3 endemic villages and one control village) were subjected to PCR analysis for detecting the presence of MU in these wild mosquitoes populations. Briefly, 2.5 µl of the DNA extract was amplified in 12.5 µl PCR mixture using the SensiMix buffer system (BioLine). Each reaction mixture contained 7.5 µl SensiMix (2x SensiMix II probe, No-Rox Mix, BioLine), 0.9 µM IS2404 primer pair, 0.25 µM IS2404 probe, a reference Rox dye (Rox Passive Reference Dye, Bio-Rad), and sterile nuclease-free water (Sigma-Aldrich). One positive control (MU Agy99 DNA) as well as a no-template negative control (nuclease-free water, Sigma-Aldrich) was used to guide this experiment against false positive and negative results. The amplification process was performed in the Mx3500P automate (MxPro Agilent Technologies, Stratagene Mx3500P) under the following cycling conditions: 50°C for 2 min, 95°C for 10 min, 40 cycles of 95°C for 15 s, and 60°C for 1 min. Negative samples to IS2404 were diluted 1/10 and resubmitted to molecular analyses for the detection of PCR inhibitors. In addition to the screening of the IS2404 target, other quantitative real time PCR IS2606/KR multiplex assays were performed on IS2404-positive samples to screen the presence of Mycobacterium conservative insertion sequence 2606 (IS2606) and the Ketoreductase B (KR-B) domain of the mycolactone polyketide synthase gene of MU plasmid (pMUM001) [36]. QPCR mixtures here contained 1 µl of DNA template, 0.9 µM of each primer, 0.25 µM of each probe, 12.5 µl of the SensiMix buffer system (2x SensiMix II probe, No-Rox Mix, BioLine), and nuclease-free water (Sigma-Aldrich) in a total volume of 25 µl. Amplification and detection conditions were performed as described above.

PCR mixtures here contained 1 µl of DNA template, 0.9 µM of each primer, 0.25 µM of each probe, 12.5 µl of the SensiMix buffer system (2x SensiMix II probe, No-Rox Mix, BioLine), and nuclease-free water (Sigma-Aldrich) in a total volume of 25 µl. Amplification and detection conditions were performed as described above. 2.5. Investigations on the Capability of Mosquitoes to Pick and Host MU Bacteria from Larval to Adult Stages (Vertical Transmission of MU in Mosquitoes) This experiment was carried out in the insectary of the AgroEcoHealth Platform of the International Institute of Tropical Agriculture (IITA-Benin). The laboratory strain Anopheles gambiae kisumu and MU strain isolates were used in this experiment. 2.5.1. The Mosquitoes Strain Anopheles kisumu Anopheles kisumu is a reference laboratory strain originating from the Kisumu region in Western Kenya. This strain is commonly used in standardization experiments and is well maintained in most malaria entomology research laboratories. 2.5.2. The Bacterial Strain Mycobacterium ulcerans Agy99 Mycobacterium ulcerans Agy99 (MU Agy99) is a well-characterized Ghanaian human isolate obtained from the Department of Bacteriology at the Noguchi Memorial Institute for Medical Research (NMIMR, Ghana). Agy99 is a reference MU strain with a sequenced genome [37].

Anopheles kisumu is a reference laboratory strain originating from the Kisumu region in Western Kenya. This strain is commonly used in standardization experiments and is well maintained in most malaria entomology research laboratories. 2.5.2. The Bacterial Strain Mycobacterium ulcerans Agy99 Mycobacterium ulcerans Agy99 (MU Agy99) is a well-characterized Ghanaian human isolate obtained from the Department of Bacteriology at the Noguchi Memorial Institute for Medical Research (NMIMR, Ghana). Agy99 is a reference MU strain with a sequenced genome [37]. 2.5.3. Experimental Infection of Mosquitoes Larvae with Mycobacterium ulcerans and Monitoring of Infected Larvae Mosquitoes larvae were infected by ingestion of MU contaminated Tetramin® Baby Fish Food (Charterhouse Aquatics, London, UK). The infection protocol was adapted from Wallace et al. [26]. Prior to infection, the preserved stock of MU strain was diluted in 1X PBS and vortexed 5 min.

rans and Monitoring of Infected Larvae Mosquitoes larvae were infected by ingestion of MU contaminated Tetramin® Baby Fish Food (Charterhouse Aquatics, London, UK). The infection protocol was adapted from Wallace et al. [26]. Prior to infection, the preserved stock of MU strain was diluted in 1X PBS and vortexed 5 min. (1) Experimental Infection of Mosquitoes Larvae with MU. Six groups (4 tests and 2 controls) of 100 eggs of An. kisumu each were distributed for rearing into labeled plastic bowls containing 250 ml sterile water. Prior to introducing eggs into bowls, the breeding/rearing water in test groups received 80 mg of Tetramin Baby Fish Food (Charterhouse Aquatics, London, UK) contaminated with 100 µl of MU (2.0 105 CFU/ml). The control groups (2 bowls) were prepared in the same way as the test bowls without introducing MU contaminated Tetramin Baby Fish Food (Charterhouse Aquatics, London, UK). The mixture (eggs-food-MU) was kept in the insectary at 27°C, 75% RH, and 12 : 12 LD for eggs hatching. The first instars larvae progeny (L1) obtained was kept in the contaminated breeding water for ingestion of the bacteria (MU) for 24 hours after which the breeding water was completely replaced with a new MU free breeding water (water + food only). The L1 larvae were fed with Tetramin and bred till obtaining the second, third, and fourth instars larvae, as well as the pupae and adult mosquitoes. To avoid cross-contaminations during the experiments, all materials and consumables such as rearing bowls, rearing water, and larvae food used for mosquitoes breeding were replaced on daily basis. Rearing waters as well as Tetramin Baby Fish Food were initially tested (qPCR analysis) and confirmed free of MU prior to be used in the experiments. Breeding bowls remained covered throughout larval rearing. The entire experiment was repeated thrice to ascertain the accuracy of the data.

were replaced on daily basis. Rearing waters as well as Tetramin Baby Fish Food were initially tested (qPCR analysis) and confirmed free of MU prior to be used in the experiments. Breeding bowls remained covered throughout larval rearing. The entire experiment was repeated thrice to ascertain the accuracy of the data. (2) Monitoring of Infected Mosquitoes. Pools of 10 individuals per developmental stage (egg, L1, L2, L3, L4, pupae, and adult) were prepared from test and control bowls. These pools of individuals were kept in labeled Eppendorf tubes with 70% ethanol and stored at −20°C for molecular screening of MU. In addition, we also harvested from breeding water the cuticles from the different larval molting phases and preserved them for similar molecular analysis. Finally, the third group of stored samples was constituted of small volumes (1 ml) of breeding water collected during the entire larval developmental stages. Collected breeding waters were spun at 14,000 rpm for 5 minutes; then, the condensate was vortexed vigorously and 250 µl was used for DNA extraction. The rationale of preserving cuticles and breeding water is to be certain after analysis that the bacterium was effectively ingested by the larvae and is inside the larvae system and not on its skin (due to cuticle colonization). For example, the presence of the bacterium DNA in larvae and its total absence in the water and the cuticle at a given developmental stage will imply that the bacterium was not on the larva skin (colonization of the skin) but is within/inside the larvae. For this infection monitoring experiment, preserved pools of larvae/adults were screened for 2 MU markers (IS2404 and KR-B which is more specific to MU). A standard curve of the qPCR values (Cts) and the bacterial loads was plotted and this curve was used to determine the bacterial infection rate and to monitor the presence of the bacteria at all larval developmental stages and also at the mosquitoes emergence (the adult stage).

4 and KR-B which is more specific to MU). A standard curve of the qPCR values (Cts) and the bacterial loads was plotted and this curve was used to determine the bacterial infection rate and to monitor the presence of the bacteria at all larval developmental stages and also at the mosquitoes emergence (the adult stage). 2.6. Statistical Analysis Statistical analysis of generated data was performed using SPSS 17.0 software (SPSS Inc., Chicago IL, USA). Chi-square test was used to set the difference in proportions (mosquitoes distribution and distribution of MU targets between localities and eggs hatching rates). Nonparametric ANOVA test (Kruskal-Wallis) was used to set the difference in means (bacterial loads and “Ct” values according to mosquitoes developmental stage), whereas the Pearson logistic regression test was used to establish the correlation between MU bacterial loads and the corresponding “Ct” values (Table 5). A pool of mosquitoes (adults or larvae) was defined infected with MU if found positive for the three targets (IS2404, IS2606, and KR-B) for field screened samples and two targets (IS2404 and KR) for laboratory infected samples. Two standard curves were plotted from serial dilutions of MU strain (Agy99) and the Ct values for IS2404 and KR-B genes. Based on these standard curves, the cycle threshold (Ct) cut-off was set at less than 35 cycles for IS2404 and less than 37 cycles for KR-B. Threshold for statistical significance was set at p < 5%.

Two standard curves were plotted from serial dilutions of MU strain (Agy99) and the Ct values for IS2404 and KR-B genes. Based on these standard curves, the cycle threshold (Ct) cut-off was set at less than 35 cycles for IS2404 and less than 37 cycles for KR-B. Threshold for statistical significance was set at p < 5%. 3. Results 3.1. Distribution of Mosquito' Species Collected in Studied Localities A total of 4,043 adult mosquitoes were collected during surveyed periods in the three targeted BU endemic villages (Agongbo, Agodenou, and Agbahounsou) and the single BU nonendemic village (Tanongou). 404 pools of 10 adults were generated from sampled mosquitoes which were identified to genus and to species. Pools were grouped by identified species of mosquitoes in each village. Four mosquito species were found in surveyed localities, namely, Mansonia africana (34.63%), Culex quinquefasciatus (32.95%), Anopheles gambiae s.l. (30.05%), and Aedes aegypti (2.37%) (Table 1). In addition to sampled adult mosquitoes, 3,175 mosquitoes larvae were collected from mosquitoes breeding sites found in the endemic villages and the control site. These larvae were used to generate 317 pools of 10 larvae. Larvae identified in the endemic sites included 60.6% of Culex quinquefasciatus, 27.86% of Anopheles gambiae s.l., and 11.54% of the pupal stage of an unknown species (unknown sp.). In the nonendemic control site only two larvae of two genera were detected, Anopheles gambiae s.l. (67.02%) and Culex quinquefasciatus (32.98%) (Table 1).

d in the endemic sites included 60.6% of Culex quinquefasciatus, 27.86% of Anopheles gambiae s.l., and 11.54% of the pupal stage of an unknown species (unknown sp.). In the nonendemic control site only two larvae of two genera were detected, Anopheles gambiae s.l. (67.02%) and Culex quinquefasciatus (32.98%) (Table 1). 3.2. Screening of IS2404, IS2606, and KR-B Targets in Wild Populations of Mosquitoes from Endemic and Nonendemic Localities 3.2.1. Screening of IS2404 Out of 301 pools of adult mosquitoes (3,010 mosquitoes) from endemic villages subjected to real time quantitative PCR analysis, 26 pools (8.63%) were found positive to IS2404 target (Table 2). At Agbahounsou, 8 pools (12.12%) of mosquitoes were found positive to IS2404, 12 pools (6.82%) at Agongbo, and 6 pools (10.17%) at Agodenou for this same molecular marker. Unexpectedly, we recorded an identical trend of positive number of pools (10/103, 9.7%) in samples from the nonendemic control site (Table 2). Out of 223 pools of collected mosquitoes larvae (2,235 mosquitoes larvae) from endemic villages subjected to qPCR analysis, 39 pools (17.49%) were found positive to IS2404 target with 10 pools (13.51%) at Agbahounsou, 24 pools (32.88%) at Agongbo, and 5 pools (6.58%) at Agodenou. At Tanongou the control site, 11 pools (11.70%) out of 94 tested from 940 mosquitoes larvae were found positive to IS2404 target (Table 3). No statistical difference was found in the distribution of this target between the test and control localities in both adult and larval mosquitoes (p > 0.05).

(6.58%) at Agodenou. At Tanongou the control site, 11 pools (11.70%) out of 94 tested from 940 mosquitoes larvae were found positive to IS2404 target (Table 3). No statistical difference was found in the distribution of this target between the test and control localities in both adult and larval mosquitoes (p > 0.05). 3.2.2. Screening of IS2606 Out of 26 pools of adult mosquitoes tested positive to IS2404 target in the three endemic villages (Agongbo, Agodenou, and Agbahounsou), none was found to be positive for the IS2606 target. The same finding was observed after real time quantitative PCR analysis of the 10 pools of mosquitoes tested positive to IS2404 in the control site. No sample was found positive to IS2606 in the nonendemic site (Table 2). However, the IS2606 target was detected in 3/39 (7.7%) pools of larvae which were positive to IS2404 target in the endemic sites (Table 3). None of the IS2404 positive mosquitoes larvae (positive pools) from one endemic site (Agodenou) or the control site were positive for the IS2606 target (Table 3). 3.2.3. Screening of KR-B Only one pool (3.84%) out of 26 pools of adult mosquitoes tested positive to IS2404 target was found positive for the KR-B target in samples from the endemic villages. This single KR-B positive pool of mosquitoes belonged to the genus Anopheles caught at Agbahounsou. However, it is worth indicating that this unique KR-B (MU plasmid marker) positive pool was not found positive to the IS2606. None of the 10 pools of IS2404 positive mosquitoes from the control site tested positive for the KR-B target (Table 2).

pool of mosquitoes belonged to the genus Anopheles caught at Agbahounsou. However, it is worth indicating that this unique KR-B (MU plasmid marker) positive pool was not found positive to the IS2606. None of the 10 pools of IS2404 positive mosquitoes from the control site tested positive for the KR-B target (Table 2). In addition, none of the mosquitoes larvae that tested positive to IS2404 target was found to be positive for the KR-B in both the endemic and the nonendemic areas (Table 3). 3.2.4. Summary of Results from the Screening of the 3 Targets Related to the Presence of MU in Analyzed Wild Populations of Mosquitoes None of the adult and larvae pools was found to contain the three MU targets (IS2404, IS2606, and KR-B). This demonstrated the absence of MU in the wild mosquitoes populations in the endemic region surveyed. Although the IS2404 target was detected in mosquitoes caught in the nonendemic village, these samples also lacked the three targets related to the presence of MU and most likely represent the presence of other environmental mycobacterial species (Tables 2 and 3).

d mosquitoes populations in the endemic region surveyed. Although the IS2404 target was detected in mosquitoes caught in the nonendemic village, these samples also lacked the three targets related to the presence of MU and most likely represent the presence of other environmental mycobacterial species (Tables 2 and 3). 3.3. Analysis of the Low Capability of Mosquitoes to Pick and Host MU from Larval to Adult Stages Following the inoculation of Anopheles kisumu eggs in simulated laboratory breeding experiment (bowls containing water, larvae food) fed with MU, we recorded an average hatching rate of 94.010 ± 1.289% in the 4 bowls which served as “test bowls” (water + food + MU + eggs of An. kisumu) and an average hatching rate of 93.87 ± 0.546% in the 2 bowls serving as “control bowls” (water + food + eggs of An. kisumu). Overall, the bacterial load decreased throughout the experiment from the young (1st instars larvae) to the old (pupae and adult stages) developmental stages of An. kisumu (Figure 2). No significance difference was observed in the decrease of the bacterial loads throughout the mosquitoes developmental stages in mosquitoes samples (p = 0.220), cuticles (p = 0.199), and breeding waters (p = 0.092).

(1st instars larvae) to the old (pupae and adult stages) developmental stages of An. kisumu (Figure 2). No significance difference was observed in the decrease of the bacterial loads throughout the mosquitoes developmental stages in mosquitoes samples (p = 0.220), cuticles (p = 0.199), and breeding waters (p = 0.092). 3.3.1. Distribution of MU in First Instars Larvae (L1) of Anopheles kisumu Randomly selected L1 larvae from the 4 “test” bowls (1 pool of 10 L1 larvae per bowl, making a total of 4 pools for the 4 bowls) showed after qPCR analysis that all 4 pools of L1 mosquitoes larvae were infected/colonized by MU. Real time PCR analysis targeting the KR-B domain of MU revealed a mean Ct value of 31.592 ± 3.151 cycles which corresponds to a mean bacterial load of E + 2.779 ± E + 0.817 CFU/ml in L1 larvae. The analysis of cuticles (1 pool of 10 cuticles from each bowl, making a total of 4 pools of cuticles from test bowls) released from the metamorphosis of L1 larvae revealed the presence of MU in all the 4 pools from “test” bowls (100% infection rate with MU, 4/4 pools). The mean Ct value of 36.516 ± 2.096 cycles corresponding to the mean bacterial load of E + 1.503 ± E + 0.523 CFU/ml was from L1 released cuticles. When the breeding water was analyzed during L1 larval development, the mean planktonic bacterial load in the water was E + 3.034 ± E + 1.024 CFU/ml, corresponding to a mean Ct value of 30.610 ±2.801 cycles. As observed with larvae and cuticles, MU was also detected in all breeding waters (4/4) during the L1 developmental stage of An. kisumu. MU was not found in L1 larvae, cuticles, or breeding water collected from the 2 bowls constituting the “control group” (Table 4).

esponding to a mean Ct value of 30.610 ±2.801 cycles. As observed with larvae and cuticles, MU was also detected in all breeding waters (4/4) during the L1 developmental stage of An. kisumu. MU was not found in L1 larvae, cuticles, or breeding water collected from the 2 bowls constituting the “control group” (Table 4). 3.3.2. Distribution of MU in Second Instars Larvae (L2) of Anopheles kisumu Randomly selected L2 larvae pools from the “test” bowls and controls were collected as was the case for the L1 larval stages. All 4 test pools of L2 mosquitoes larvae were infected or colonized by MU. Real time PCR of the MU KR-B domain of MU yielded a mean Ct value of 33.063 ± 2.984 cycles equivalents to a mean bacterial load of E + 2.399 ± E + 0.773 CFU/ml. Cuticles released from the metamorphosis of L2 larvae had MU in 3/4 (75% infection/colonization rate) pools from “test” bowls. The mean Ct value of 36.823 ± 1.652 cycles equivalent to a mean bacterial load of E + 1.424 ± E + 0.428 CFU/ml was recorded. When the breeding water was analyzed during L2 larval development, the estimated mean planktonic bacterial load found in the water was E + 2.705 ± E + 0.680 CFU/ml. Bacteria was found in all 4 tests during the L2 developmental stage of An. kisumu. Traces of bacteria were not found in L2 larvae, cuticles, or breeding water in the “control group” (Table 4).

during L2 larval development, the estimated mean planktonic bacterial load found in the water was E + 2.705 ± E + 0.680 CFU/ml. Bacteria was found in all 4 tests during the L2 developmental stage of An. kisumu. Traces of bacteria were not found in L2 larvae, cuticles, or breeding water in the “control group” (Table 4). 3.3.3. Distribution of MU in Third Instars Larvae (L3) of Anopheles kisumu Randomly selected L3 larvae pools as previously described for L1 and L2 showed that all 4 pools of L3 mosquitoes larvae were infected or colonized by MU and had mean Ct values to the KR-B region of 34.33 ± 3.349 cycles equivalent to a mean bacterial load of E + 2.070 ± E + 0.031 CFU/ml in L3. The analysis of cuticles only showed the presence of MU in 1/4 (25%) pools from “test” bowls. The breeding water during L3 larval development had an estimated mean planktonic bacterial load of E + 2.277 ± E + 0.023 CFU/ml. MU was not detected in any of the control group samples (Table 4).

.070 ± E + 0.031 CFU/ml in L3. The analysis of cuticles only showed the presence of MU in 1/4 (25%) pools from “test” bowls. The breeding water during L3 larval development had an estimated mean planktonic bacterial load of E + 2.277 ± E + 0.023 CFU/ml. MU was not detected in any of the control group samples (Table 4). 3.3.4. Distribution of MU in Fourth Instars Larvae (L4) of Anopheles kisumu Randomly selected L4 larvae from the 4 “test” bowls showed that only 3/4 (75%) pools of L4 mosquitoes larvae were infected or colonized by MU. The mean Ct value of 35.03 ± 1.177 cycles equivalent to a mean bacterial load of E + 1.88 ± E + 0.441 CFU/ml was recorded in L4. MU was not detected in samples of cuticles released from the metamorphosis of L4 larvae. Breeding water samples had estimated mean planktonic bacterial loads of E + 1.652 ± E + 0.019 CFU/ml. Three out of 4 (75%) breeding water samples were contaminated with the bacteria during the L4 developmental stage. MU was not detected in any of the samples from the L4 control group samples (Table 4). 3.3.5. Distribution of MU in Pupae Stages of Anopheles kisumu MU was not detected from the randomly selected pupae from the 4 “test” bowls. In addition, MU was not detected in the cuticles released from the emergence of adult mosquitoes from pupae. Only one out of 4 (25%) breeding water samples was contaminated with the bacteria MU during pupae developmental stage, with a KR-B Ct value of 35.47 cycles. As above, MU was not detected in samples constituting the control group (Table 4).

cted in the cuticles released from the emergence of adult mosquitoes from pupae. Only one out of 4 (25%) breeding water samples was contaminated with the bacteria MU during pupae developmental stage, with a KR-B Ct value of 35.47 cycles. As above, MU was not detected in samples constituting the control group (Table 4). 3.3.6. Distribution of MU in Adult Stages of Anopheles kisumu Overall, MU was not detected in any of adult stage samples or their controls (Table 4).

cted in the cuticles released from the emergence of adult mosquitoes from pupae. Only one out of 4 (25%) breeding water samples was contaminated with the bacteria MU during pupae developmental stage, with a KR-B Ct value of 35.47 cycles. As above, MU was not detected in samples constituting the control group (Table 4). 3.3.6. Distribution of MU in Adult Stages of Anopheles kisumu Overall, MU was not detected in any of adult stage samples or their controls (Table 4). 4. Discussion 4.1. Wild Populations of Mosquitoes Are Unlikely to Be MU Reservoirs in Sedje-Denou According to WHO, a reservoir is any person, animal, arthropod, plant, soil, or substance, or a combination of these, in which an infectious agent lives and multiplies and where it reproduces itself in such a manner that it can be transmitted to a susceptible host [38]. Difficulties to cultivate Mycobacterium ulcerans (MU) from contaminated environmental samples remains the main challenge in the identification of reproductive reservoir(s) for this Mycobacterium as well as the understanding of its transmission mode(s) from the MU contaminated environment to humans. Most environmental samples that have been identified with MU have been classified as “potential reservoirs” [1, 39]. Aquatic water bugs have been shown to replicate MU in their salivary glands [40] and MU has been successfully recovered by culture from theses insects [11], and thus, the “reservoir” capacity of other “suspected organisms” remains unclear. The aquatic environment has been identified as the most predominant source of MU contamination [12–18, 31, 41–45]. This research was conducted in the wet agroecosystem of Sedje-Denou region and more specifically in three endemic villages which served as test sites for this study. From the three thousand and ten adult mosquitoes subjected to real time PCR, twenty-six pools (8.64%) were positive to the insertion sequence IS2404, which is not specific enough to infer the presence of MU. We recorded the presence of this insertion (IS2404) in mosquitoes samples collected from nonendemic location (Tanongou in the Northern Benin). These results further highlight the nonspecificity of this marker for MU detection from environmental samples [4, 36, 45]. The use of two additional targets (IS2606 and KR) to increase the specificity of MU detection in our study showed that none of the mosquitoes tested to be simultaneously positive for all three targets. These results certainly confirm the low capability of wild mosquitoes populations to carry MU as previously described by others in this same Southern region of Benin [25].

increase the specificity of MU detection in our study showed that none of the mosquitoes tested to be simultaneously positive for all three targets. These results certainly confirm the low capability of wild mosquitoes populations to carry MU as previously described by others in this same Southern region of Benin [25]. However, our data seems to contradict works conducted in Australia which detected MU in mosquitoes samples [7, 19–23, 26]. Johnson et al. [22] described the contamination of mosquito species by MU as a consequence of resting and feeding or breeding in storm water drains, whereas Wallace et al. [26], in an experimental study, suggested an unlikely role for mosquitoes as BU biological vectors. In their study using mice and both natural and anthropogenic forms of inoculation, they emphasized that reducing exposure to insect bites and destroying mosquitoes breeding sites around households would break the chain of BU transmission [28]. These series of studies on the role of mosquitoes in the transmission of MU show the need of further investigations whether mosquitoes can act as both reservoir and vector of MU. In this current study, none of the 2,235 mosquitoes larvae collected from both endemic and nonendemic areas for BU were found to be positive for MU, suggesting that mosquitoes larvae in the wild were unlikely to be reservoirs for MU. Although our results generated from wild mosquitoes populations are in favor of previous studies conducted in Benin which revealed the inability of mosquitoes to be involved in MU transmission [25], a laboratory designed experimental model was designed to better understand the poor implication of mosquitoes in increased number of BU cases in West and Central Africa [1, 6].

ulations are in favor of previous studies conducted in Benin which revealed the inability of mosquitoes to be involved in MU transmission [25], a laboratory designed experimental model was designed to better understand the poor implication of mosquitoes in increased number of BU cases in West and Central Africa [1, 6]. 4.2. Inability of An. kisumu Larvae to Pick Up MU from Their Environment and Remain Colonized through the Larval Developmental Stages to the Adult Stage Mosquitoes (Culicidae) development, as characteristic of all holometabolous insects, proceeds through embryonic, larval, pupal, and adult stages that reflect considerable morphological and physiological differences [34]. These stages exhibit distinct niches; larvae and pupae are aquatic while adults are free-flying and terrestrial. In mosquitoes vectors, vertical transmission has been demonstrated for certain pathogens which include yellow fever virus, dengue virus, St. Louis encephalitis virus, Japanese encephalitis virus, and West Nile virus (WNV) [46]. Vertical transmission involves the transmission of pathogens from female mosquitoes to their offspring. The laboratory experimental model showed that mosquitoes larvae readily ingested MU and that MU colonized the larval stages through the pupal stage. However, at the pupae series of high energy demanding [47], metabolism taking place in the mosquitoes certainly affects MU development leading to the clearing of MU colonization by the end of pupation and at the adult stage (Figure 2). Our research demonstrated the total absence of MU at both pupae and adult stages as reported by Wallace et al. [28] and, thus, highlights the inability of these biting dipterans to act as a good vector/host of MU in an endemic environment. Results from this laboratory based experiment are consistent with those obtained from the analysis of thousands of wild populations of mosquitoes collected in the endemic locations which did not show any MU colonization through molecular testing. Data published by Wallace et al. [26] suggested that MU is unlikely to persist in the mosquito's body system, a behavior which stands as a natural protective mechanism of mosquitoes to bacterial infections. According to Hoxmeier et al. [48], the contamination of Anopheles gambiae mosquitoes with MU resulted in disruptions to phospholipid metabolic pathways in the mosquitoes, especially the use of glycolipid molecules.

system, a behavior which stands as a natural protective mechanism of mosquitoes to bacterial infections. According to Hoxmeier et al. [48], the contamination of Anopheles gambiae mosquitoes with MU resulted in disruptions to phospholipid metabolic pathways in the mosquitoes, especially the use of glycolipid molecules. Moreover, glycolipids are actively involved in signaling and are mediators in cellular and immune processes [49]. The disruption of synthesis of this molecule probably has a negative impact on the various interactions between MU cells and Anopheles and the poor capability of mosquitoes to serve as biological vectors for MU [45]. Instead of acting as biological vectors for MU as described in this study, mosquitoes might act as mechanical vectors as recently described in an experimental study with Aedes notoscriptus and BALB/C mice [28]. However, mechanical transmission of MU seems to happen only after skin trauma either by an insect bite or by any other environmental stress (e.g., a thorn, penetrating wood splinters, and scorpion stings) [29]. The traumatized skin should initially be colonized by MU, a phenomenon that could naturally happen during repetitive contacts with the risk environments such as water bodies or contaminated biofilms [1, 17, 28]. Furthermore, in behavioral study with Aedes aegypti, Sanders et al. [50] suggested that if a biofilm of MU was on a person, the bacteria may be attracting mosquitoes which in return would lead to a puncture insertion of MU as recently reported by Wallace et al. [28]. Although mechanical transmission of MU stands as a common mechanism that could correlate transmission studies from both Africa and Australia, Williamson et al. [30] recently established that abrasions (trauma) of the skin in Guinea pig models and subsequent application of MU are not sufficient enough to cause an ulcer. Further laboratory and epidemiological studies are therefore required to understand the extent of the mechanical transmission of MU and how frequent animals including humans can carry and remain colonized with MU on their skin to facilitate such transmission mode. MU could be traced from the risk environments to humans or animals directly after they had contact with colonized environments.

understand the extent of the mechanical transmission of MU and how frequent animals including humans can carry and remain colonized with MU on their skin to facilitate such transmission mode. MU could be traced from the risk environments to humans or animals directly after they had contact with colonized environments. In such hypothetical situations and for preventive measures, individuals from endemic areas should remain aware and avoid frequent contacts with mosquito's bites by sleeping under mosquitoes bed nets, wearing protective clothing while farming or using clean water for bathing and cleaning [1, 7, 15, 17, 19, 28].

understand the extent of the mechanical transmission of MU and how frequent animals including humans can carry and remain colonized with MU on their skin to facilitate such transmission mode. MU could be traced from the risk environments to humans or animals directly after they had contact with colonized environments. In such hypothetical situations and for preventive measures, individuals from endemic areas should remain aware and avoid frequent contacts with mosquito's bites by sleeping under mosquitoes bed nets, wearing protective clothing while farming or using clean water for bathing and cleaning [1, 7, 15, 17, 19, 28]. Mosquitoes larvae breeding in an MU contaminated water body are capable of ingesting this bacterium as shown by Hoxmeier et al. [48] and Wallace et al. [26] in Aedes aegypti, Aedes albopictus, Culex restuans, and Ochlerotatus triseriatus larvae. Although several experimental studies have established the potential of predaceous aquatic insects to temporally maintain MU during their developmental stages in water [37, 40], our findings in addition to confirming these previous results also show that MU colonization of mosquitoes larvae is very temporal as larvae system is capable of clearing the bacterial load during pupae and adult developmental stages. The vertical transmission of MU therefore seems not to be effective in mosquitoes populations as documented with several viruses. The noncontamination/colonization of field-caught mosquito species by MU as found in this study might suggest that mosquitoes are unable to move MU from one source to another in endemic areas in Benin.

l transmission of MU therefore seems not to be effective in mosquitoes populations as documented with several viruses. The noncontamination/colonization of field-caught mosquito species by MU as found in this study might suggest that mosquitoes are unable to move MU from one source to another in endemic areas in Benin. 5. Conclusion This study revealed the absence of MU in hematophagous mosquitoes trapped in households in BU endemic locations in the Sedje-Denou division in Benin. Using an experimental model, we also showed the inability of laboratory infected or colonized An. kisumu larvae to transfer the bacteria to their pupae and the emerging adults. This low ability of mosquitoes to vertically transmit MU pathogens to their offspring coupled with the absence of MU in field-caught mosquitoes further highlights the low probability of these biting insects as biological vectors for MU in endemic villages in Benin. Mosquitoes may therefore not be involved in the dissemination of this pathogen from the risk environments to humans in investigated areas. However, further studies should be performed to evaluate their mechanical implication, before completely excluding whether they are involved or not in the transmission cycle of this emerging disease.

refore not be involved in the dissemination of this pathogen from the risk environments to humans in investigated areas. However, further studies should be performed to evaluate their mechanical implication, before completely excluding whether they are involved or not in the transmission cycle of this emerging disease. Acknowledgments This work was supported in part by a research grant from the International Society for Infectious Diseases (ISID, United States) attributed to Francis Zeukeng and the AgroEcoHealth Platform of the International Institute of Tropical Agriculture (IITA-Benin). The authors express their gratitude to household heads and community health workers of the Sedje-Denou and Tanongou localities for their assistance during sample collections. They also extend their gratitude to Professor Dorothy Yeboah-Manu and Dr. Anthony Ablordey for providing the bacterial strain used in this study, their assistance, and the supervision of Francis Zeukeng internships in the Bacteriology Department of Noguchi Memorial Institute for Medical Research (NMIMR) in Accra, Ghana. Conflicts of Interest The authors declare that they have no conflicts of interest.

Acknowledgments This work was supported in part by a research grant from the International Society for Infectious Diseases (ISID, United States) attributed to Francis Zeukeng and the AgroEcoHealth Platform of the International Institute of Tropical Agriculture (IITA-Benin). The authors express their gratitude to household heads and community health workers of the Sedje-Denou and Tanongou localities for their assistance during sample collections. They also extend their gratitude to Professor Dorothy Yeboah-Manu and Dr. Anthony Ablordey for providing the bacterial strain used in this study, their assistance, and the supervision of Francis Zeukeng internships in the Bacteriology Department of Noguchi Memorial Institute for Medical Research (NMIMR) in Accra, Ghana. Conflicts of Interest The authors declare that they have no conflicts of interest. Authors' Contributions Rousseau Djouaka, Francis Zeukeng, Jude Daiga Bigoga, Wilfred Fon Mbacham, Solange E. Kakou-Ngazoa, and Anthony Ablordey conceived and designed the experiments. Rousseau Djouaka, Francis Zeukeng, Genevieve Tchigossou, Sylla Aboubacar, Romaric Akoton, David N'golo Coulibaly, Sodjinin Jean-Eudes Tchebe, Clavella Nantcho Nguepdjo, Romaric Akoton, Innocent Djegbe, and Solange E. Kakou-Ngazoa participated in collection and analyses of the data. Rousseau Djouaka, Francis Zeukeng, Razack Adeoti, Sodjinin Jean-Eudes Tchebe, and Innocent Djegbe performed the experimental study on the vertical transmission of M. ulcerans in mosquitoes. Jude Daiga Bigoga, Wilfred Fon Mbacham, Solange E. Kakou-Ngazoa, Manuele Tamo, and Anthony Ablordey supervised the experiments. Rousseau Djouaka and Francis Zeukeng performed the statistical analysis of the data. Rousseau Djouaka and Francis Zeukeng drafted the manuscript. Anthony Ablordey, Solange E. Kakou-Ngazoa, and Manuele Tamo made critical revisions of the manuscript. All authors read and approved the final manuscript.

ed the experiments. Rousseau Djouaka and Francis Zeukeng performed the statistical analysis of the data. Rousseau Djouaka and Francis Zeukeng drafted the manuscript. Anthony Ablordey, Solange E. Kakou-Ngazoa, and Manuele Tamo made critical revisions of the manuscript. All authors read and approved the final manuscript. Figure 1 Location map of study sites in the Buruli ulcer endemic area in Southern Benin and the nonendemic area in Northern Benin. Figure 2 Distribution of average bacterial load during mosquito developmental stages. L1, L2, L3, and L4 correspond to first, second, third, and fourth instars larvae, respectively. Values are given with error bars at 5%. Table 1 Distribution of field-caught mosquito species in study sites. Mosquito species Developmental stages Study areas Total Agbahounsou Agongbo Agodenou Tanongou Anopheles gambiae s.l. Adult 134 162 119 800 1215 Larvae 210 110 303 630 1253 Mansonia africana Adult 190 870 200 140 1400 Culex quinquefasciatus Adult 320 690 232 90 1332 Larvae 450 550 354 310 1664 Aedes aegypti Adult 20 46 25 5 96 Unknown sp. Pupae 80 65 113 0 258 Total 1404 2493 1346 1975 7218 Table 2 Distribution of MU targets in field-caught adult mosquitoes. Study sites Pools of 10 adult mosquitoes analyzed IS2404-qPCR KR-qPCR IS2606-qPCR MU distribution Positive P (%) Positive P (%) Positive P (%) BU endemic villages Agbahounsou 66 8 12.12 1 12.5 0 — Absent Agongbo 176 12 6.82 0 — 0 — Absent Agodenou 59 6 10.17 0 — 0 — Absent Total   301 26 8.63 1 12.5 0 — Absent

Table 2 Distribution of MU targets in field-caught adult mosquitoes. Study sites Pools of 10 adult mosquitoes analyzed IS2404-qPCR KR-qPCR IS2606-qPCR MU distribution Positive P (%) Positive P (%) Positive P (%) BU endemic villages Agbahounsou 66 8 12.12 1 12.5 0 — Absent Agongbo 176 12 6.82 0 — 0 — Absent Agodenou 59 6 10.17 0 — 0 — Absent Total   301 26 8.63 1 12.5 0 — Absent BU nonendemic village Tanongou 103 10 9.7 0 — 0 — Absent P: percentage of targets distribution. No statistical difference was found in the distribution of IS2404 target between the endemic and nonendemic localities (p = 0.601). Table 3 Distribution of MU targets in field collected mosquitoes larvae. Study sites Pools of 10 mosquito larvae analyzed IS2404-qPCR KR-qPCR IS2606-qPCR MU distribution Positive P (%) Positive P (%) Positive P (%) BU endemic villages Agbahounsou 74 10 13.51 0 — 2 16.67 Absent Agongbo 73 24 32.88 0 — 1 5.26 Absent Agodenou 76 5 6.58 0 — 0 — Absent Total   223 39 17.49 0 — 0 — Absent BU nonendemic village Tanongou 94 11 11.70 0 — 0 — Absent P: percentage of targets distribution. No statistical difference was found in the distribution of IS2404 target between the endemic and nonendemic localities (p = 0.347). Table 4 MU distribution among mosquitoes developmental stages, cuticles, and breeding waters.

BU nonendemic village Tanongou 94 11 11.70 0 — 0 — Absent P: percentage of targets distribution. No statistical difference was found in the distribution of IS2404 target between the endemic and nonendemic localities (p = 0.347). Table 4 MU distribution among mosquitoes developmental stages, cuticles, and breeding waters. Nature of the samples Mosquitoes developmental stages Distribution of MU targets Pool positive/pool tested Presence of MU Mean Cts (IS2404) Mean Cts (KR-B) Mosquito∗ Eggs 19 ± 1.79 21 ± 2.22 4/4 Yes L1 27.67 ± 2.66 31.59 ± 3.15 4/4 Yes L2 29.92 ± 2.58 33.06 ± 2.98 4/4 Yes L3 31.36 ± 2.98 34.33 ± 3.34 4/4 Yes L4 31.38 ± 2.20 35.03 ± 1.17 3/4 Yes Pupae NoCt NoCt 0/4 No Adults 37.89 NoCt 0/4 No Mosquito cuticles∗ Eggs NA NA NA NA L1 30.72 ± 1.78 36.51 ± 2.09 4/4 Yes L2 34.25 ± 2.83 36.82 ± 1.65 3/4 Yes L3 34.13 39.53 1/4 Yes L4 NoCt NoCt 0/4 No Pupae 38 NoCt 0/4 No Adults NA NA NA NA Mosquito breeding waters∗ Eggs 18.43 ± 2.03 21.49 ± 1.63 4/4 Yes L1 23.04 ± 3.19 30.61 ± 2.80 4/4 Yes L2 22.71 ± 2.59 31.88 ± 2.60 4/4 Yes L3 28.4 ± 2.86 33.53 ± 3.00 4/4 Yes L4 32.00 ± 2.64 35.94 ± 1.04 3/4 Yes Pupae 33.65 35.47 1/4 Yes Adults NA NA NA NA L1, 2, 3, and 4 correspond to first, second, third, and fourth instars larvae, respectively. Yes or no corresponds to the presence or the absence of the bacteria in analyzed samples. NA stands for not applicable. ∗The bacterial loads did not vary significantly among the developmental stages (p < 0.05). Table 5 Characteristics of the standard curves linking “Ct” values of MU targets and corresponding bacterial loads (MU Agy99 serial dilutions).

Mosquito breeding waters∗ Eggs 18.43 ± 2.03 21.49 ± 1.63 4/4 Yes L1 23.04 ± 3.19 30.61 ± 2.80 4/4 Yes L2 22.71 ± 2.59 31.88 ± 2.60 4/4 Yes L3 28.4 ± 2.86 33.53 ± 3.00 4/4 Yes L4 32.00 ± 2.64 35.94 ± 1.04 3/4 Yes Pupae 33.65 35.47 1/4 Yes Adults NA NA NA NA L1, 2, 3, and 4 correspond to first, second, third, and fourth instars larvae, respectively. Yes or no corresponds to the presence or the absence of the bacteria in analyzed samples. NA stands for not applicable. ∗The bacterial loads did not vary significantly among the developmental stages (p < 0.05). Table 5 Characteristics of the standard curves linking “Ct” values of MU targets and corresponding bacterial loads (MU Agy99 serial dilutions). MU targets Regression coefficient (R2) Regression equation (95% CI) p  value IS2404 0.9955 Y = 9.6569 − 0.2396X 0.000008 KR-B 0.9968 Y = 10.9682 − 0.2592X 0.000004 Independent variable, Ct values (cycle threshold); dependent variable, log10 MU (CFU/ml).

1. Background Dengue is a febrile disease that is transmitted by Aedes mosquitoes. It is estimated that approximately 3900 million people in 128 countries are at risk of infection by dengue viruses [1]. This virus infects approximately 390 million people per year, 96 million of whom present disease [2]. The causative agent of dengue is dengue virus (DENV), which belongs to the family Flaviviridae. The clinical phenotype of dengue varies from a self-limiting febrile illness to severe, occasionally life-threatening disease. Typically, symptomatic disease follows three phases: a febrile phase lasting 3 to 7 days; a critical phase characterized by defervescence, during which complications appear in a small proportion of patients; and a spontaneous recovery phase. Complications primarily affect the vascular system and include an unusual plasma leakage syndrome that may result in dengue hemorrhagic fever (DHF) and dengue shock syndrome (DSS), which are classified as severe dengue according to the WHO guidelines for dengue diagnosis, treatment, and prevention [3, 4]. In 2015-2016, Zika virus caused large outbreaks in Pacific areas and in Central and South America [5]. Increasing data have shown that Zika virus infection is associated with adverse pregnancy outcomes such as fetal death, premature birth, and microcephalus [6]. Thus, DENV infection which belongs to the same family of Flaviviridae as Zika virus has become a serious concern in pregnant women.

ntral and South America [5]. Increasing data have shown that Zika virus infection is associated with adverse pregnancy outcomes such as fetal death, premature birth, and microcephalus [6]. Thus, DENV infection which belongs to the same family of Flaviviridae as Zika virus has become a serious concern in pregnant women. 2. Dengue in Pregnant Woman With an estimated yearly case number of 390 million from more than 100 countries and regions in Asia, Oceania, America, and Africa, dengue fever (DF) is now one of the most important vector-borne diseases worldwide [2]. As reported in recent studies, a high rate of dengue infection has been observed in pregnant women. An investigation of 358 pregnant women in Malaysia, a dengue endemic country, showed that a seropositivity for dengue infection of approximately 35.8% by ELISA [7]. Additionally, in another dengue endemic country, Thailand, the dengue seroprevalence can reach 90.3% [8]. It must be noted, however, that these data were collected during the Chikungunya disease outbreak [8], raising the possibility of false positivity. In the African country, the Democratic Republic of Sao Tome and Principe, 39.74% of pregnant women were found to be positive for DENV antibodies [9]. These findings indicate that pregnant women have high a risk of dengue virus infection.

the Chikungunya disease outbreak [8], raising the possibility of false positivity. In the African country, the Democratic Republic of Sao Tome and Principe, 39.74% of pregnant women were found to be positive for DENV antibodies [9]. These findings indicate that pregnant women have high a risk of dengue virus infection. However, the prognosis for the pregnant women infected with dengue remains unknown. There might be two explanations. One explanation is that dengue infection is aggravated in pregnant women. Adam et al. reported that, among 30 pregnant women infected by dengue virus, approximately 38.3% of them developed dengue hemorrhagic fever or dengue shock syndrome, and the mortality could reach 21.7% [10]. Machado et al. also reported that 46.5% of pregnant women infected by dengue virus developed severe dengue, as compared with 22.5% of nonpregnant women [11]. These reports indicated that pregnant women infected with dengue virus had a greater tendency to experience illness progression to severe dengue and higher mortality, as compared with infected nonpregnant women. The other explanation is that adverse pregnancy outcomes emerge after infection. A recent meta-analysis indicated that common adverse outcomes of pregnant women with dengue infection were stillbirth (crude relative risk: 6.7), miscarriage (odds ratio: 3.51), preterm birth (odds ratio: 1.71), and low birth weight (odds ratio: 1.71) [12]. A miscarriage was recently reported to occur in a pregnant woman infected by dengue virus, and DENV RNA was detected in fetal material [13]. Thus, dengue infection in pregnant women should undoubtedly be taken seriously.

(odds ratio: 3.51), preterm birth (odds ratio: 1.71), and low birth weight (odds ratio: 1.71) [12]. A miscarriage was recently reported to occur in a pregnant woman infected by dengue virus, and DENV RNA was detected in fetal material [13]. Thus, dengue infection in pregnant women should undoubtedly be taken seriously. 3. Initial Immune Response to Dengue in Pregnant Women Dengue fever is widely accepted as an immunopathological disease. Thus, the pathogenesis of dengue can be explained by various mechanisms related to immune factors as follows: (1) enhancement of viral infection through cross-reactive antibodies; (2) activation of cross-reactive memory T cells; (3) a cytokine or inflammatory storm; and (4) complement activation. However, a very large part of dengue pathogenesis remains elusive despite these arduous studies. Since dengue is an immune-related disease, whether infection becomes limited or progressive depends on the competition between defensive and destructive immune responses. The initial immune response to viral infection is mediated by interferons (IFNs) [14]. IFNs have displayed increasing diversity and activity to “interfere” with viral replication. IFNs are divided into three types: type I mainly represented by IFN-α [15] and IFN-β [16], type II represented by INF-γ [17], and type III including the recently discovered IFN-lambda (IFN-λ) family [18]. Each IFN family member mediates important antiviral activity via engagement with its specific IFN receptor.

cation. IFNs are divided into three types: type I mainly represented by IFN-α [15] and IFN-β [16], type II represented by INF-γ [17], and type III including the recently discovered IFN-lambda (IFN-λ) family [18]. Each IFN family member mediates important antiviral activity via engagement with its specific IFN receptor. 3.1. IFN Level in Dengue Infection Few studies have evaluated serum IFNs level in pregnant women during dengue infection. However, a larger number of studies have explored serum IFN levels in dengue patients. For type I IFNs, the serum level of IFN-β was significantly higher in primary DHF patients than in patients with dengue fever [19], suggesting that high levels of IFN-β might accompany a worsened progression of the disease. However, the other type 1 IFN, IFN-α, displayed an opposite phenomenon. The serum IFN-α level was significantly higher in DF patients than in DHF patients, irrespective of the DENV serotypes [20], and the dynamic change of serum level significantly reduced in 3–5 days after fever onset (DENV-1: reduced from 94.42 pg/mL to 36.12 pg/mL; DENV-2: reduced from 53.39 pg/mL to 38.25 pg/mL). Also the serum level of IFN-α was significantly higher in dengue patients than in healthy individuals [21]. These findings suggest that the early robust production of IFN-α may be correlated with a better clinical condition with respect to dengue infection and disease progression. Thus, the roles of type I IFNs in dengue infection are contradictory and complicated.

higher in dengue patients than in healthy individuals [21]. These findings suggest that the early robust production of IFN-α may be correlated with a better clinical condition with respect to dengue infection and disease progression. Thus, the roles of type I IFNs in dengue infection are contradictory and complicated. For type II IFNs, the serum level of IFN-γ was higher in dengue patients than in healthy individuals [22]. The same finding was also reported by Feitosa et al. [23]. The serum level of IFN-γ was gradually decreased in 3–5 days, 6-7 days, 8–10 days, and 14–17 days since illness onset in both mild and severe dengue patients [24], and it was higher in severe dengue patients in 8–10 days after illness onset compared with the mild dengue patients, but no difference was found in the other times between these two groups [24]. This observation was confirmed by a cohort study reported by Cui et al. [25]. However, the higher level of IFN-γ was found in patients within 96 hours from fever onset. According to these findings, the high concentration of IFN-γ in dengue patient serum may indicate a high risk of disease progression to severe dengue. For type III IFNs, there are few clinical studies reporting mild or severe dengue. Nevertheless, in vitro studies found that dengue virus infection could induce the production of IFN-λ [26]. An elevated serum IFN-λ level was also observed in patients with dengue fever, compared to healthy blood donors, and IFN-λ could inhibit dengue virus replication in vitro [27].

es reporting mild or severe dengue. Nevertheless, in vitro studies found that dengue virus infection could induce the production of IFN-λ [26]. An elevated serum IFN-λ level was also observed in patients with dengue fever, compared to healthy blood donors, and IFN-λ could inhibit dengue virus replication in vitro [27]. In general, a high serum level of IFNs in dengue patients accompanies a high risk of disease progression to severe dengue. This raises the question of whether lower serum IFN level indicates a lower risk of disease progression to severe dengue. The answer to this question is “no.” It has been reported that the mean level of serum IFN-γ in DF cases was higher than that in DHF patients, which implied that low serum IFN-γ level might be associated with severe diseases [28]. A Recent research has shown that the serum IFN-γ concentration in dengue patients is negatively correlated with the dengue virus load, indicating that a lower the serum level of IFN-γ is correlated with a higher dengue viral load in dengue patients [29]. Although dengue virus is regarded as an immunopathological disease, it is also dire [30]. Additionally, a high viral load is one important cause of severe dengue because the virus can destroy host cells and lead to vascular leakage, which accelerates disease progression in a short period of time [31, 32]. Thus, one strategy to reduce the possibility of disease progression to severe dengue is to decrease the dengue viral load. Therefore, a high level of type II IFNs is needed for the treatment of dengue patients in clinics. It has been also reported that, during acute infection, higher serum levels of IFN-β [6.69 (0.15~34.9) pg/mL] were found in DHF patients with primary infection than that [2.35 (0~10.1) pg/mL] in the DHF patients with secondary infection [19]. To this point, the serum IFN-β level of dengue patients was not associated with dengue progression.

ted that, during acute infection, higher serum levels of IFN-β [6.69 (0.15~34.9) pg/mL] were found in DHF patients with primary infection than that [2.35 (0~10.1) pg/mL] in the DHF patients with secondary infection [19]. To this point, the serum IFN-β level of dengue patients was not associated with dengue progression. 3.2. IFN Level in Pregnant Women IFNs are expressed by the trophectoderm of primates, rodents, and ungulates during peri-implantation [33]. Conceptus IFNs are pregnancy signals for maternal recognition in ruminants (cattle, sheep, and goats); they act on the endometrium to indirectly maintain progesterone synthesis [34]. Additionally, studies investigating IFN signaling in the uterus of livestock species suggest that it modulates maternal immune tolerance to the implanting conceptus, changes in the endometrial architecture for uterine receptivity, and vascular remodeling for maternal-fetal nutrient and waste exchange [35, 36]. Although these findings for IFNs have rarely been reported in humans, they are still useful references for pregnant women. Although few studies have described the outcomes of pregnancy with dengue and the relationship between IFNs and pregnancy after DENV infection, such findings have been extensively reported for Zika virus, which belongs to the same family of Flaviviruses as DENV. In a mouse model, type I IFNs could control vaginal Zika virus replication [37]. In vitro, Zika virus infection has been shown to cause the constitutive release of IFN-λ, the antiviral type III IFN, by human placental trophoblast cells [38]. These findings raise the question of whether a higher level of IFNs is beneficial for pregnant women with dengue infection. The answer to this question is also “no.” In an in vitro embryo culture study, IFN-β showed no macroscopic teratogenic effect on embryonic development, but it caused growth retardation in embryos [39].

findings raise the question of whether a higher level of IFNs is beneficial for pregnant women with dengue infection. The answer to this question is also “no.” In an in vitro embryo culture study, IFN-β showed no macroscopic teratogenic effect on embryonic development, but it caused growth retardation in embryos [39]. Li et al. also showed that IFN-γ could induce pregnancy failure by moderating natural killer (NK) cells [40], and it could dramatically increase decidual apoptosis [41]. In a case control study, women with extremely low birth weight infants with accepted induced spontaneous conception or assisted reproductive technology showed increased production of IFN-γ [42]. Furthermore, IFN use in pregnant women is not advised, especially given its known antiproliferative effects [43]. A cohort study has indicated that IFN therapy during the first trimester of pregnancy can lead to a high risk of fetal loss and low birth weight [44]. Animal trials have demonstrated that IFN-γ administration can result in pregnancy failure [45, 46]. Thus, in pregnant women, IFNs should be maintained at a certain level for a stable pregnancy.

ated that IFN therapy during the first trimester of pregnancy can lead to a high risk of fetal loss and low birth weight [44]. Animal trials have demonstrated that IFN-γ administration can result in pregnancy failure [45, 46]. Thus, in pregnant women, IFNs should be maintained at a certain level for a stable pregnancy. 3.3. The Mechanism by Which IFNs Induce Adverse Pregnancy Outcomes Most adverse pregnancy outcomes are associated with pathogen infection, although the precise mechanism leading to adverse outcomes via infection is unknown. One established factor is the specific immune response to pathogen infection in pregnant women. Whether IFNs, particularly IFN-γ, play a positive or negative role during pregnancy has raised numerous controversies. For other IFN family members, few studies have focused on their influence on pregnant women with dengue virus. Our knowledge is derived mostly from animal or in vitro experimental studies, because data for human pregnancies are scarce. In vitro, IFN-γ is cytotoxic to human trophoblast cells [47] and inhibits their proliferation [48]. An in vitro experiment showed that the levels of IFN-γ secreted by decidual NK cells were closely correlated with trophoblasts apoptosis in response to Toxoplasma gondii infection [49]. Comba et al. reported that the levels of IFN-γ in blood and endometrial tissue were significantly higher in patients with recurrent pregnancy loss than in women with normal fertility [50].

secreted by decidual NK cells were closely correlated with trophoblasts apoptosis in response to Toxoplasma gondii infection [49]. Comba et al. reported that the levels of IFN-γ in blood and endometrial tissue were significantly higher in patients with recurrent pregnancy loss than in women with normal fertility [50]. IFN-γ can also lead to adverse pregnancy outcomes through the following signaling pathways. In a mouse model to induce fetal resorption through the injection of α-galactosylceramide (αGC), decidual invariant NK T cells (iNKT) were activated followed by the upregulation of IFN-γ levels. This adoptive change resulted in pregnancy loss in IFN-γ−/− mice [51]. One signal transduction pathway leads to fetal mortality is class I phosphoinositide 3-kinase (PI3K), which converts phosphatidylinositol-(4,5)-phosphate to phosphatidylinositol-(3,4,5)-phosphate (PIP3). Acting as a second messenger, PIP3 recruits proteins to the plasma membrane, where they activate signaling pathways that promote cell proliferation, survival, and differentiation [52]. Based on structural similarity, PI3K can be divided into two classes: class IA and class IB. Class IA PI3K forms heterodimers of p85 regulatory subunits and one of the three isoforms of the catalytic p110 subunit (p110-α, p110-β, and p110-δ). PI3K p110-δ is a key mediator of NK cell maturation and function. An absence of p110-δ signaling leads to reduced cytokine release, aberrant maturation, and incorrect trafficking to peripheral organs, including the uterus during pregnancy [53, 54]. When PI3K and p110-δ were deactivated, the uterus showed a decreased level of IFN-γ and elevated IL-6, resulting in fetal death or growth retardation [55]. Simultaneously, IFN-γ can exert feedback control on Ly-49 receptors to regulate NK cell effector functions during pregnancy failure [40]. The other pathway is the Notch signaling pathway, which exerts its effects throughout pregnancy, playing an important role in placental angiogenesis and trophoblast function [56]. During peptidoglycan and polyinosinic:cytidylic acid-induced preterm labor, Notch signaling is activated, resulting in the upregulation of proinflammatory responses with upregulated levels of IFN-γ, TNF-α, and IL-6, and its inhibition improves the in utero survival of live fetuses [57]. In general, the mechanism by which IFNs induce adverse pregnancy outcomes is complicated and requires further study.

gnaling is activated, resulting in the upregulation of proinflammatory responses with upregulated levels of IFN-γ, TNF-α, and IL-6, and its inhibition improves the in utero survival of live fetuses [57]. In general, the mechanism by which IFNs induce adverse pregnancy outcomes is complicated and requires further study. 4. Conclusion Dengue is a serious public health problem and also a threat to the pregnant women who have a higher risk of progressing to severe dengue or experiencing adverse pregnancy outcomes after infection. The initial immune response to dengue virus infection, especially with respect to IFN production, was reviewed in this group of patients. Dengue is much more prevalent in pregnant women in comparison to the other groups. The severity of dengue is correlated with the high level of some IFNs in patient' serum, while the serum IFN level must be maintained at a sufficiently high level to maintain the pregnancy and inhibit dengue virus replication. However, many of the above-described findings represent only clinical phenomena or were derived from in vivo/in vitro animal experiments. The effects of IFNs on human pregnancies are more difficult to study. Hence, further analyses are needed to reach an unambiguous conclusion regarding the roles of IFNs during dengue virus infection, especially in pregnant women.

represent only clinical phenomena or were derived from in vivo/in vitro animal experiments. The effects of IFNs on human pregnancies are more difficult to study. Hence, further analyses are needed to reach an unambiguous conclusion regarding the roles of IFNs during dengue virus infection, especially in pregnant women. Acknowledgments This study was partly supported by the grant from the National Natural Science Foundation of China (nos. 81271866 and 81572012), the Guangdong Province Universities and Colleges Pearl River Scholar Funded Scheme (2014), the Guangdong Provincial Natural Science Foundation Key Project (2016A030311025), and Guangzhou Health and Medical Collaborative Innovation Major Special Project (201604020011) to Hong-Juan Peng. Abbreviations DENV:Dengue virus DHF:Dengue hemorrhagic fever DSS:Dengue shock syndrome DF:Dengue fever IFNs:Interferons NK:Nature killer αGC: α-Galactosylceramide iNKT:Invariant natural killer T cells PI3K:Phosphoinositide 3-kinases PIP3:Phosphatidylinositol-(4,5)-phosphate to phosphatidylinositol-(3,4,5)-phosphate. Disclosure The funding agencies had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Conflicts of Interest The authors declare that they have no conflicts of interest. Authors' Contributions All authors were involved in the study design, including setting up the keywords search and project protocol. Hao Zhang and Zhiyi He collected the data information. Hao Zhang drafted manuscript. Wenting Zeng and Hong-Juan Peng were responsible for the supervision of the project and revision of the manuscript.

1. Introduction Antibiotic resistance poses a growing threat to public health all over the globe. Undoubtedly inappropriate prescription is a major issue, and that in agriculture and husbandry deserves more focus. Moreover, surging level of antimicrobial agents employed by veterinary medicine makes animals today potential reservoirs of resistant bacteria [1]. Categorized as opportunistic pathogen, nonetheless, Pseudomonas aeruginosa infection could be lethal in the immunocompromised population [2]. Yet the pivotal problem lies in the intrinsic antibiotic resistance of P. aeruginosa that counteracts the effects of medical treatments [3]. To current knowledge, P. aeruginosa is more likely to be an environmental species and passed on by animal reservoirs, the snakes, for instance [3]. In fact, the latest report from Taiwanese Centers of Disease Control illustrates that incidence of snakebites reaches more than 1,000 cases annually [4] and implicates underestimated role of P. aeruginosa of snake origin.

y to be an environmental species and passed on by animal reservoirs, the snakes, for instance [3]. In fact, the latest report from Taiwanese Centers of Disease Control illustrates that incidence of snakebites reaches more than 1,000 cases annually [4] and implicates underestimated role of P. aeruginosa of snake origin. However, little is detailed about the determination regarding susceptibility of antibiotics that are widely administered within human population such as piperacillin/tazobactam and meropenem [5]. Since disease-free snakes might be carriers as well, captive ones of commercial origin and from research institution were also enrolled in our study. In brief, sensitivity testing of P. aeruginosa from snakes' oral cavities and subsequent molecular typing of resistance genes will be outlined in the following paragraph, with further implication of clinical translation and public health responses.

l origin and from research institution were also enrolled in our study. In brief, sensitivity testing of P. aeruginosa from snakes' oral cavities and subsequent molecular typing of resistance genes will be outlined in the following paragraph, with further implication of clinical translation and public health responses. 2. Materials and Methods 2.1. Pseudomonas aeruginosa Strains The study was implemented based on 58 P. aeruginosa isolates from disease-free captivated snakes, wild snakes at Endemic Species Research Institute (ESRI), and clinical cases at National Chung Hsing University Veterinary Medicine Teaching Hospital, Taichung, Taiwan. All samples were collected through sterilized swabbing and stored in the transport media on the way to microbiology laboratory. Following aerobic culture routine, the isolates were plated onto cetrimide agar and incubated at 35°C in 5% CO2 and 95% air for 48 hours. Based on the acquired colonies, Pseudomonas aeruginosa was further determined by colony morphology, Gram stain, oxidase test, and API ID 32 GN strips (bioMérieux, Marcy l'Etoile, France).

wing aerobic culture routine, the isolates were plated onto cetrimide agar and incubated at 35°C in 5% CO2 and 95% air for 48 hours. Based on the acquired colonies, Pseudomonas aeruginosa was further determined by colony morphology, Gram stain, oxidase test, and API ID 32 GN strips (bioMérieux, Marcy l'Etoile, France). 2.2. Antimicrobial Susceptibility Testing The minimum inhibitory concentrations (MICs) were identified by Clinical Laboratory Standards Institute (CLSI) [6]. Targeting P. aeruginosa, cation-adjusted Müller-Hinton broth was engaged to subsequent serial twofold dilutions of representative antimicrobial agents. P. aeruginosa suspension was then adjusted to the turbidity of 0.5 McFarland standard of P. aeruginosa ATCC27853, and 100 μL of the bacteria with inoculum density at around 105 CFU/mL was further transferred onto 96-well plates. After incubation at 37°C for 24 hours, MICs of respective antibiotics were then categorized into susceptible, intermediate, and resistant based on the interpretation table within CLSI guidance. The antimicrobials were employed in the study: penicillins (piperacillin/tazobactam), cephems (cefotaxime), aminoglycosides (gentamicin and amikacin), lipopeptides (colistin), monobactams (aztreonam), and carbapenems (meropenem).

tible, intermediate, and resistant based on the interpretation table within CLSI guidance. The antimicrobials were employed in the study: penicillins (piperacillin/tazobactam), cephems (cefotaxime), aminoglycosides (gentamicin and amikacin), lipopeptides (colistin), monobactams (aztreonam), and carbapenems (meropenem). 2.3. Detection of Resistance Genes in P. aeruginosa Isolates The Nucleic Acid Automatic Extraction System (Taco, Taichung, Taiwan) was adopted for DNA isolation through the manufacturer's protocol. The extracts served as PCR templates, where primers and the amplification outcomes of blaTEM, blaOXA-group I/III, and blaAmpC specific PCR are outlined in Table 1. In brief, 2 μL of DNA template was mixed with 1 μL upstream primer, 1 μL downstream primer, 5 μL of Fast‐Run™ Taq Master Mix (Protech Technology Enterprise Co., Ltd., Taipei, Taiwan), and 16 μL distilled water to reach 25 μL in total. Each premixed set was then anchored within GenePro Thermal Cycler (Bioer Technology, Hangzhou, China). Subsequent to the initial denaturation session of 5 minutes at 95°C, whole PCR was subjected to 30 cycles of 30 seconds at 94°C, half an hour at 55–65°C, 45–75 seconds at 72°C, and final 7-minute-long extension at 72°C. 2.4. Statistical Analysis All data were processed through basic statistic tools of Microsoft Excel (Microsoft Corporation, Redmond, USA).

2.3. Detection of Resistance Genes in P. aeruginosa Isolates The Nucleic Acid Automatic Extraction System (Taco, Taichung, Taiwan) was adopted for DNA isolation through the manufacturer's protocol. The extracts served as PCR templates, where primers and the amplification outcomes of blaTEM, blaOXA-group I/III, and blaAmpC specific PCR are outlined in Table 1. In brief, 2 μL of DNA template was mixed with 1 μL upstream primer, 1 μL downstream primer, 5 μL of Fast‐Run™ Taq Master Mix (Protech Technology Enterprise Co., Ltd., Taipei, Taiwan), and 16 μL distilled water to reach 25 μL in total. Each premixed set was then anchored within GenePro Thermal Cycler (Bioer Technology, Hangzhou, China). Subsequent to the initial denaturation session of 5 minutes at 95°C, whole PCR was subjected to 30 cycles of 30 seconds at 94°C, half an hour at 55–65°C, 45–75 seconds at 72°C, and final 7-minute-long extension at 72°C. 2.4. Statistical Analysis All data were processed through basic statistic tools of Microsoft Excel (Microsoft Corporation, Redmond, USA). 3. Results 3.1. Antimicrobial Resistance MIC distribution of indicative antimicrobial agents is summarized in Table 2, which illustrates further susceptibility translation according to CLSI standards for broth microdilution method and preexisting references of cefotaxime [6]. Of note, only one out of 58 isolates (2%) demonstrated resistance to meropenem otherwise sensitive (MIC ≤ 2 mg/L). Second to meropenem, approximately 80–90% of P. aeruginosa isolates also exhibited susceptibility to amikacin, piperacillin/tazobactam, and aztreonam. On the opposite, 52 isolates (90%) were found to be nonsusceptible to cefotaxime, which was followed by colistin at 55% and gentamicin at 33% resistance rate, respectively.

ond to meropenem, approximately 80–90% of P. aeruginosa isolates also exhibited susceptibility to amikacin, piperacillin/tazobactam, and aztreonam. On the opposite, 52 isolates (90%) were found to be nonsusceptible to cefotaxime, which was followed by colistin at 55% and gentamicin at 33% resistance rate, respectively. 3.2. Prevalence of β-Lactamase Genes Meanwhile, β-lactamase gene within the context was identified through PCR and subsequent molecular typing as well. Intriguingly, blaTEM was discovered among all isolates whereas blaAmPC was detected in 55 out of 58 isolates (94.8%). As for Ambler class D β-lactamase family, frequency of OXA-1 was over three times higher than that of OXA-10 (89.7% over 27.6%). Remarkably, presence of OXA-10 displayed high correlation (93.8%) with coexistence of other three resistance genes. Distribution pattern of β-lactamase genes is abstracted in Table 3. 4. Discussion In our study, antibiotic-resistant P. aeruginosa strains from the snakes' oral cavity were identified for the very first time. Of these, resistance rate even reached approximately 50% with certain antibiotics. High prevalence of β-lactamases genes was disclosed through our work as well.

3.2. Prevalence of β-Lactamase Genes Meanwhile, β-lactamase gene within the context was identified through PCR and subsequent molecular typing as well. Intriguingly, blaTEM was discovered among all isolates whereas blaAmPC was detected in 55 out of 58 isolates (94.8%). As for Ambler class D β-lactamase family, frequency of OXA-1 was over three times higher than that of OXA-10 (89.7% over 27.6%). Remarkably, presence of OXA-10 displayed high correlation (93.8%) with coexistence of other three resistance genes. Distribution pattern of β-lactamase genes is abstracted in Table 3. 4. Discussion In our study, antibiotic-resistant P. aeruginosa strains from the snakes' oral cavity were identified for the very first time. Of these, resistance rate even reached approximately 50% with certain antibiotics. High prevalence of β-lactamases genes was disclosed through our work as well. In terms of other natural reservoirs, P. aeruginosa could only be discovered in a limited number of animals like canine, feline, swan, and reptile [5]. Most studies of animal P. aeruginosa strains from Pan-Pacific regions reported relatively subtle resistance [7]. Reportedly, merely 17.8% of P. aeruginosa from dogs and cats demonstrated resistance of cefotaxime whereas over half of reptile strains recruited in our study were insensitive to the chemotherapeutic agents [8]. As for carbapenems, it is rather striking to detect 2% resistance in reptile strains as their canine and feline parallels were all sensitive according to current literature [8]. Although polymyxin B resistance was once reported in canine P. aeruginosa [9], there exists no direct evidence to implicate colistin (also known as polymyxin E) resistance that was expressed by almost 50% of snake isolates in our study. Combined, to our knowledge, P. aeruginosa from snakes' oral cavity are highly antibiotic-resistant.

in B resistance was once reported in canine P. aeruginosa [9], there exists no direct evidence to implicate colistin (also known as polymyxin E) resistance that was expressed by almost 50% of snake isolates in our study. Combined, to our knowledge, P. aeruginosa from snakes' oral cavity are highly antibiotic-resistant. Indeed P. aeruginosa possesses diverse intrinsic resistance mechanisms such as multidrug efflux pumps and antibiotics inactivating enzymes and naturally tolerates a number of antimicrobial agents [10]. Also, the species is likely to develop acquired or adaptive resistance under selective pressure. Although there are certain antipseudomonal antibiotics available, emergence of multidrug resistant (MDR) strains poses a considerable threat to the public health. Since colistin resistance was accidentally discovered through our work, the problem is far more complicated than initially expected. In Taiwan, MDR P. aeruginosa took up 18.6% among all hospital-acquired infections in a retrospective review [11], of which the level was similar to that worldwide. A parallel Taiwanese study also illustrated insufficient colistin susceptibility of P. aeruginosa, at a mere 76.7% [12]. Mirrored to our study, colistin resistance of snake P. aeruginosa was still 31.7% higher than that of human strains. On the other hand, MDR reptile strains stood at approximately one-tenth in our study, which was similar to 10–20% among human P. aeruginosa infections in Taiwan [13].

P. aeruginosa, at a mere 76.7% [12]. Mirrored to our study, colistin resistance of snake P. aeruginosa was still 31.7% higher than that of human strains. On the other hand, MDR reptile strains stood at approximately one-tenth in our study, which was similar to 10–20% among human P. aeruginosa infections in Taiwan [13]. With regard to antibiotic resistance among P. aeruginosa in other Asian-Pacific countries, surveillance report in 2011 outlined the highest carbapenem resistance in Philippines (50%) that was followed by India (32%) and Thailand (30%) [14]. Intriguingly, cross-national colistin susceptibility stood at 98% on average at that time [14]. Applied to clinical aspects, secondary infection of snakebites often arises from mixed pathogen like other animal bites. In a 10-year-long Taiwanese cross-sectional study, P. aeruginosa was detected in 23.8% of patients with subsequent infections [15]. In fact, among all Gram-negative aerobes, P. aeruginosa was the second most prevalent that followed M. morganii and Enterococcus species [15]. At the time, there are no guidelines or consensus of antibiotics selection for snakebites [16]. Nonetheless, healthcare professionals should be alerted as MDR and colistin-resistant P. aeruginosa strains were recognized in our study. That is, in the context of snakebites in Taiwan, the victim is actually running a high risk of treatment failure.

uidelines or consensus of antibiotics selection for snakebites [16]. Nonetheless, healthcare professionals should be alerted as MDR and colistin-resistant P. aeruginosa strains were recognized in our study. That is, in the context of snakebites in Taiwan, the victim is actually running a high risk of treatment failure. On the other hand, adaptive colistin resistance of P. aeruginosa could also be induced by subinhibitory doses [17]. Surprisingly, the systemic review on pharmacokinetics and pharmacodynamics of colistin is inadequate, even after over-50-year-long clinical adoption [18]. It is foreseeable that colistin administration at a sublethal concentration might be insensible to clinicians and leads to adaptive resistance subsequently. Furthermore, stimulation from colistin at a sublethal level is likely to trigger resistance to aminoglycosides and quinolones [19].

ar-long clinical adoption [18]. It is foreseeable that colistin administration at a sublethal concentration might be insensible to clinicians and leads to adaptive resistance subsequently. Furthermore, stimulation from colistin at a sublethal level is likely to trigger resistance to aminoglycosides and quinolones [19]. Generated to P. aeruginosa of reptile origin, appalling colistin resistance at 55% probably results from exposure to unspecified antibacterial peptides [20]. Concerning serine-β-lactamases, 100% prevalence of the resistance gene blaTEM among P. aeruginosa within natural environments was once reported by Igbinosa et al. before [21], whereas strains extracted from human infection expressed blaTEM at a far below ratio [22]. Nonetheless one-tenth of isolates were still susceptible to cefotaxime and the discrepancy between genotypes and phenotypes probably indicated conditional expression of Ambler class A β-lactamases genes. In addition, our work highlighted high frequency of blaAmpC. Since several antibiotics could upregulate AmpC production [23], none of them were employed in our work in order to avoid batch errors. In other words, route of exposure prior to the recruitment might be the untold story. As all three Ambler classes of β-lactamases contributed to cefotaxime resistance, that raised a question about the hierarchy of single resistance gene. In contrast, it could also be possible for a universal cue to orchestrate expression of those genes, simultaneously.

prior to the recruitment might be the untold story. As all three Ambler classes of β-lactamases contributed to cefotaxime resistance, that raised a question about the hierarchy of single resistance gene. In contrast, it could also be possible for a universal cue to orchestrate expression of those genes, simultaneously. Undoubtedly, certain limitations of our work exist and it takes further studies to discover the missing puzzle pieces. Since meropenem-resistant P. aeruginosa was detected, extensive work on genes encoding for carbapenemases such as OXA-48 and metallo-β-lactamases is required [24, 25]. Similarly, polymyxin resistance genes, PA4773-PA4775, for example, require thorough investigation [26]. However, impact of resistance genes might be overestimated as they are not constitutive to gene expression. Meanwhile, susceptibility profiles of quinolones should be established in the near future as MDR and colistin resistance among P. aeruginosa from snakes' oral cavity was noted.

equire thorough investigation [26]. However, impact of resistance genes might be overestimated as they are not constitutive to gene expression. Meanwhile, susceptibility profiles of quinolones should be established in the near future as MDR and colistin resistance among P. aeruginosa from snakes' oral cavity was noted. Of note, oral flora suggests the gut microbiota of the preys, to some extent, as they might defecate even when being consumed [27]. Indeed, any creatures that are exposed to animal performance enhancers (APEs) and in contact with snakes are skeptical culprits. Also, environmental dissemination and unreported preexposures should be taken into account. The authorities should be cautious of disorganized regulation and reporting system of off-label antibiotic administration, of which biochemical characteristics of APEs and other administration details remain unclear as only gross doses in the unit of kilograms and animal species are reported [28, 29]. Otherwise the rooting issue of disseminating antibiotics at a sublethal dose will continue to threaten public health. 5. Conclusion MDR and colistin-resistant P. aeruginosa were identified from the snakes' oral cavity in Taiwan for the first time, which indicated high risk of medical treatment failure in the context of snakebites and the crucial role of antibiotic stewardship to prevent adaptive resistance. On the other hand, wide prevalence of serine-β-lactamases was also disclosed through our work. Nonetheless the correlation between genotypes and phenotypes still requires further investigation.

treatment failure in the context of snakebites and the crucial role of antibiotic stewardship to prevent adaptive resistance. On the other hand, wide prevalence of serine-β-lactamases was also disclosed through our work. Nonetheless the correlation between genotypes and phenotypes still requires further investigation. Acknowledgments The authors would like to express their sincere appreciation to Dr. Robert Chan and Ms. Chin Chang-Chien for their support and assistance. This work was supported by grants from the Taichung Veterans General Hospital (103DHA0500493) Taichung, Taiwan. Additional Points Highlights. (1) Pseudomonas aeruginosa are not uncommon in snakes' oral cavity. (2) Multidrug resistant and colistin-resistant P. aeruginosa from snakes' oral cavity were discovered. (3) The lowest resistance rates were in aztreonam, piperacillin/tazobactam, and amikacin. (4) 100% of the isolates harbor resistance gene blaTEM and high frequency of blaAmpC among P. aeruginosa isolates. Conflicts of Interest The authors declare that they have no conflicts of interest.

Highlights. (1) Pseudomonas aeruginosa are not uncommon in snakes' oral cavity. (2) Multidrug resistant and colistin-resistant P. aeruginosa from snakes' oral cavity were discovered. (3) The lowest resistance rates were in aztreonam, piperacillin/tazobactam, and amikacin. (4) 100% of the isolates harbor resistance gene blaTEM and high frequency of blaAmpC among P. aeruginosa isolates. Conflicts of Interest The authors declare that they have no conflicts of interest. Authors' Contributions Po-Yu Liu wrote the manuscript and collected samples. Yan-Chiao Mao revised the manuscript. Ling-Ling Weng conducted the data analysis. Shu-Ying Tseng contributed to the microbiological analysis. Chou-Chen Huang and Ching-Chang Cheng collected samples of snakes from different vet hospitals. Kwong-Chung Tung revised the manuscript. All authors contributed to data analysis, drafting, and critically revising the paper, read and approved the final manuscript, and agreed to be accountable for all aspects of the work. Table 1 Primers sequences used in the amplification and antimicrobial resistance genes in P. aeruginosa.

Authors' Contributions Po-Yu Liu wrote the manuscript and collected samples. Yan-Chiao Mao revised the manuscript. Ling-Ling Weng conducted the data analysis. Shu-Ying Tseng contributed to the microbiological analysis. Chou-Chen Huang and Ching-Chang Cheng collected samples of snakes from different vet hospitals. Kwong-Chung Tung revised the manuscript. All authors contributed to data analysis, drafting, and critically revising the paper, read and approved the final manuscript, and agreed to be accountable for all aspects of the work. Table 1 Primers sequences used in the amplification and antimicrobial resistance genes in P. aeruginosa. Class of β-lactamases Primer name Sequence (5′ to 3′) Product size (bp) Target Class A TEM-A GAGTATTCAACATTTCCGTGTC 851 blaTEM TEM-B TAATCAGTGAGGCACCTATCTC Class D OXA-10F TCTTTCGAGTACGGCATTAGC 760 OXA group I OXA-10B CCAATGATGCCCTCACTTTCC OXA-1A AGCCGTTAAAATTAAGCCC 911 OXA group III OXA-1B CTTGATTGAAGGGTTGGGCG Class C AmpC-PA1 ATGCAGCCAACGACAAAGG 1243 blaAmpC AmpC-PA2 CGCCCTCGCGAGCGCGCTTC Table 2 Distribution of MICs and antimicrobial sensitivity test of gentamicin, amikacin, cefotaxime, piperacillin/tazobactam, colistin, meropenem, and aztreonam for P. aeruginosa.

TTAAAATTAAGCCC 911 OXA group III OXA-1B CTTGATTGAAGGGTTGGGCG Class C AmpC-PA1 ATGCAGCCAACGACAAAGG 1243 blaAmpC AmpC-PA2 CGCCCTCGCGAGCGCGCTTC Table 2 Distribution of MICs and antimicrobial sensitivity test of gentamicin, amikacin, cefotaxime, piperacillin/tazobactam, colistin, meropenem, and aztreonam for P. aeruginosa. Agent Number of isolates with indicated MIC values (mg/L) Percentage of indicated susceptibility 0.125 0.25 0.5 1 2 4 8 16 32 64 128 ≥256 S (%) I (%) R (%) Gentamicin 0 0 0 0 5 34 11 2 0 0 1 5 39 (67%) 11 (19%) 8 (14%) Amikacin 0 0 0 0 0 28 19 6 0 0 0 5 53 (91%) 0 5 (9%) Cefotaxime 0 0 0 0 0 2 4 18 17 9 6 2 6 (10%) 18 (31%) 34 (59%) Pip + taz 0 0 0 0 0 14 28 9 5 2 0 0 51 (88%) 7 (12%) 0 Colistin 0 0 0 0 26 25 2 0 2 0 0 3 26 (45%) 25 (43%) 7 (12%) Meropenem 6 18 8 17 8 1 0 0 0 0 0 0 57 (98%) 1 (2%) 0 Aztreonam 0 0 0 0 4 19 26 6 2 0 0 1 49 (85%) 6 (10%) 3 (5%) Pip + taz: piperacillin/tazobactam; S: susceptible; I: intermediate; R: resistant. Table 3 Prevalence of Ambler classes A, C, and D β-lactamases among fifty-eight clinical isolates of P. aeruginosa. Class Type of β-lactamases Number (%) of isolates Class A blaTEM 58 (100%) Class D OXA group I (OXA-1) 52 (89.7%) OXA group III (OXA-10) 16 (27.6%) Class C blaAmpC 55 (94.8%) Combined blaTEM, OXA-1, OXA-10, blaAmpC 15 (25.9%) blaTEM, OXA-1, blaAmpC 34 (58.6%) blaTEM, blaAmpC 5 (8.6%) blaTEM, OXA-10, blaAmpC 1 (1.7%) OXA-1, blaAmpC 3 (5.2%)

1. Introduction In Canada, an estimated 75,500 people were living with HIV by the end of 2014 [1]. Undiagnosed HIV represents a serious public health challenge, with about 21% of infected individuals unaware of their status [1]. Early recognition and initiation of antiretroviral therapy has been shown to reduce transmission and morbidity, and ultimately improve quality of life [2–4]. Accordingly, Health Canada recommends consideration of routine HIV screening during medical care [5]. Individuals who are newly diagnosed with HIV in Manitoba are immediately linked to the Manitoba HIV Program, which provides specialized and long-term care to all individuals living with HIV in the province. In Manitoba, there were 79,924 tests performed, of which 135 (0.17%) were positive; 68,026 unique individuals were tested, of which 134 (0.20%) were positive. There were 78 cases of previously undiagnosed HIV identified in 2015, representing a crude rate of 8.0 cases per 100,000 population. A large percentage (30%) of these cases presented as late diagnoses, with CD4 counts below 200 cells/ml, signifying advanced disease associated with high morbidity and healthcare-related costs [6]. Furthermore, late diagnosis contributes to transmission of undiagnosed HIV to sexual partners. The high proportion of late diagnoses highlights the need for earlier diagnosis through routine screening [7].

s below 200 cells/ml, signifying advanced disease associated with high morbidity and healthcare-related costs [6]. Furthermore, late diagnosis contributes to transmission of undiagnosed HIV to sexual partners. The high proportion of late diagnoses highlights the need for earlier diagnosis through routine screening [7]. Point-of-care (POC) HIV testing has been implemented as a method for overcoming some of the patient and provider barriers to routine screening. POC HIV testing has been shown to be an acceptable method for increasing HIV testing uptake, by providing immediate results, decreasing turnaround time, and guiding urgent decision-making [8]. The INSTI HIV-1/HIV-2 Antibody Test has been validated and licensed for use by Health Canada since 2005 [9]. In 2010, the British Columbia Centre for Disease Control (BCCDC) introduced a centralized POC HIV testing program within the province to expand the availability of HIV testing. BCCDC guidelines suggest POC HIV testing in clinical scenarios where there is an urgent need to determine HIV status, including in acutely ill individuals or in settings in which the prevalence of HIV is expected to be high [10]. Several studies examining the use of POC HIV testing in emergency departments of tertiary care centres have found a prevalence of undiagnosed HIV between 0% and 1.4% [8, 11–13]. Evidence suggests that the prevalence of undiagnosed HIV may be comparable on hospital inpatient wards [14].

HIV is expected to be high [10]. Several studies examining the use of POC HIV testing in emergency departments of tertiary care centres have found a prevalence of undiagnosed HIV between 0% and 1.4% [8, 11–13]. Evidence suggests that the prevalence of undiagnosed HIV may be comparable on hospital inpatient wards [14]. The aim of this study was to evaluate the prevalence of previous outpatient HIV testing, the prevalence of undiagnosed HIV, and the attitudes towards routine POC HIV testing in patients admitted to the adult medicine inpatient unit, thereby informing the expansion of HIV testing in the province. 2. Methods 2.1. Study Design A prospective cross-sectional study was conducted on patients admitted to two adult medicine inpatient units at Health Sciences Centre (HSC) in Winnipeg, Manitoba, during the months of July, August, and October, 2016. Data was not collected in September due to the unavailability of study staff during this month. HSC is an 850-bed tertiary care centre located in inner-city Winnipeg that provides services to rural Manitoba, Nunavut, and parts of Northwestern Ontario. All patients admitted directly to two adult medicine inpatient teaching units at HSC were approached for participation in the study. Exclusion criteria included known HIV-positive status, disturbed cognition, inability to provide consent, and if the patient declined to be tested for HIV. Ethics approval was obtained from the Health Research Ethics Board of the University of Manitoba.

ching units at HSC were approached for participation in the study. Exclusion criteria included known HIV-positive status, disturbed cognition, inability to provide consent, and if the patient declined to be tested for HIV. Ethics approval was obtained from the Health Research Ethics Board of the University of Manitoba. 2.2. Data Collection Study staff was trained on the use of POC HIV test (INSTI HIV-1/HIV-2 Antibody Test, bioLytical Labs, Canada) and standardized pre- and posttest counselling prior to the study. Pretest HIV counselling included information about modes of transmission, risks of infection, diagnosis, the Manitoba HIV Program, and an explanation of the POC HIV test. Posttest counselling was tailored to the POC HIV test result and provided information about the window period and recommendations for future testing. After admission to the inpatient unit, patients were approached by the study staff for participation and consent. The time elapsed between admission and test administration ranged from 0 to 3 days. After obtaining written informed consent, pretest counselling was provided. A brief questionnaire was administered, which collected information on demographics, reason for admission to hospital, previous HIV testing, and risk factors for HIV. The POC HIV test was then performed at the bedside, and results were recorded and shared with the patient immediately. A venous blood sample was drawn from patients who had either a reactive or an indeterminate POC HIV test and sent to Cadham Provincial Laboratory (Winnipeg, Manitoba) for confirmatory testing, consisting of an immunoassay targeted to detect HIV-1/-2 antibodies and p24 antigen. In the event of a positive confirmatory test, patients were to be linked to the Manitoba HIV Program for further interdisciplinary management of newly diagnosed HIV. Posttest counselling was administered, which included information regarding the importance of retesting and avoidance of high-risk behaviours. Afterwards, patients were provided a posttest questionnaire to evaluate their degree of satisfaction with the POC HIV test.

interdisciplinary management of newly diagnosed HIV. Posttest counselling was administered, which included information regarding the importance of retesting and avoidance of high-risk behaviours. Afterwards, patients were provided a posttest questionnaire to evaluate their degree of satisfaction with the POC HIV test. 2.3. Data Analysis Descriptive statistics were used to represent data. Data were summarized using percentages and medians with interquartile ranges. 3. Results 3.1. Patient Recruitment A total of 379 patients were admitted to the adult medicine inpatient teaching units during the study period. Of these patients, 308 were approached for participation in the study, of which 144 ultimately participated in the study and had a POC HIV test (Figure 1). A total of 71 patients were never approached for consent to participate in the study because they were either receiving medical care, were having investigations, or were unwilling to be approached by study staff. A total of 164 patients were excluded due to reasons listed in Table 1. Another two patients were excluded following the administration of POC HIV testing because they disclosed their previously known HIV positive status after testing was performed. The remaining 142 patients were included in analysis.

by study staff. A total of 164 patients were excluded due to reasons listed in Table 1. Another two patients were excluded following the administration of POC HIV testing because they disclosed their previously known HIV positive status after testing was performed. The remaining 142 patients were included in analysis. 3.2. Baseline Characteristics Baseline characteristics of the study population are displayed in Table 2. The median age of study participants was 58 (42–68) years, with 68 (48%) males, 73 (51%) females, and 1 (1%) who identified as transgender. The majority of participants self-identified as Caucasian (57%) or Indigenous, including First Nations, Inuit, and Métis (36%), and most (89%) participants were born in Canada. At the time of study enrolment, 98 (69%) listed their primary location of residence to be within Winnipeg, 19 (13%) lived outside Winnipeg in another community within Manitoba, 18 (13%) lived within a First Nation Reserve, and 7 (4%) lived outside Manitoba. A total of 55 (39%) participants listed the reason for admission as primarily related to an infectious process. A total of 111 participants reported having a regular primary care provider, of whom 39 (35%) reported ever having been tested for HIV previously. Of the 106 participants who had seen their primary care provider within the last year, 7 (7%) reported being tested for HIV during that time.

lated to an infectious process. A total of 111 participants reported having a regular primary care provider, of whom 39 (35%) reported ever having been tested for HIV previously. Of the 106 participants who had seen their primary care provider within the last year, 7 (7%) reported being tested for HIV during that time. 3.3. POC HIV Testing Results and Patient Acceptance A total of 138 (97%) participants had a negative POC HIV test result, 1 (1%) had a reactive result, and 3 (2%) had an indeterminate result. All participants who tested reactive or indeterminate were subsequently tested negative with routine serologic testing. The results of the posttest questionnaire are shown in Table 3. Of the participants who underwent POC HIV testing, 131 (92%) reported satisfaction with the testing experience and 123 (87%) would choose to have the POC method of testing again if repeat HIV testing was required.

d negative with routine serologic testing. The results of the posttest questionnaire are shown in Table 3. Of the participants who underwent POC HIV testing, 131 (92%) reported satisfaction with the testing experience and 123 (87%) would choose to have the POC method of testing again if repeat HIV testing was required. 4. Discussion Health Canada recommends that consideration of HIV testing be made a component of routine medical care [5]. In 2014, the Joint United Nations Programme on HIV/AIDS (UNAIDS) launched a strategy to improve the uptake, quality, and outcome of antiretroviral therapy for HIV worldwide, called the “90-90-90” campaign. This campaign aims to have 90% who have HIV diagnosed, 90% of those diagnosed to receive sustained antiretroviral therapy, and 90% of those on therapy to achieve viral suppression [15]. In our study, only 7% of individuals who had been seen by their family physician in the previous year had HIV testing performed during that year. This low rate of testing is consistent with provincial data suggesting that approximately 6% of adults aged 15 years and above receive HIV testing. The fact that only 35% indicated to ever having been tested further illustrates that HIV testing is suboptimal in Manitoba, Canada. The low rate of routine HIV testing reflects a high rate of missed opportunities: instances where HIV-infected individuals present to medical care with possible HIV-related illnesses, but who are not tested for HIV. The prevalence of missed opportunities is generally high across various populations [14, 16–19]. Inpatient units provide a unique opportunity for HIV recognition and linkage to long-term follow-up care. A study by Rucker et al. conducted at three hospitals in Chicago, Illinois, found that inpatient areas had the highest seroprevalence of undiagnosed HIV (0.6%) compared to emergency room (0.4%) and outpatient care areas (0.1%), highlighting the need to focus on inpatient HIV screening [14].

to long-term follow-up care. A study by Rucker et al. conducted at three hospitals in Chicago, Illinois, found that inpatient areas had the highest seroprevalence of undiagnosed HIV (0.6%) compared to emergency room (0.4%) and outpatient care areas (0.1%), highlighting the need to focus on inpatient HIV screening [14]. In our study, participant satisfaction with POC HIV testing was high, with 93% of respondents indicating that they were satisfied with the overall testing experience. This is consistent with previous studies showing similar degrees of overall satisfaction with POC testing [8, 20]. A clear majority (87%) of participants indicated that they would choose to have the POC HIV test performed in lieu of the standard HIV test. This falls in line with previous studies that have demonstrated 40–90% of patients prefer rapid POC testing as opposed to standard testing [21–24].

ction with POC testing [8, 20]. A clear majority (87%) of participants indicated that they would choose to have the POC HIV test performed in lieu of the standard HIV test. This falls in line with previous studies that have demonstrated 40–90% of patients prefer rapid POC testing as opposed to standard testing [21–24]. The present study did not detect any undiagnosed HIV on the inpatient units during the study period; however, there continues to be a role for routine inpatient HIV screening. Among several urban tertiary care centres in the United States, the prevalence of undiagnosed HIV on inpatient hospital units is highly variable. This variability appears to be reflective of the study sample size: in the study by Padrnos and colleagues, 283 patients were screened with a reported prevalence of undiagnosed HIV of 0%; Osorio and colleagues screened 1537 patients and reported a prevalence of 0.4%; and finally, the group of Rucker and colleagues screened 7546 patients with a prevalence of 0.6% [14, 25, 26]. Additionally, the failure to identify HIV in an inpatient hospital setting—where a considerable population of patients receive the majority of their medical care—would constitute a critical missed opportunity. This may contribute to the finding that patients diagnosed with HIV while admitted to an inpatient hospital unit tend to have more advanced disease compared to patients diagnosed with HIV in an outpatient setting [27]. Considering that there are few existing studies specifically addressing inpatient HIV prevalence, the finding of no cases of undiagnosed HIV in this present study of relatively small sample size should not dismiss the importance of inpatient screening. Further study to specifically address the prevalence of HIV on an inpatient unit is required.

few existing studies specifically addressing inpatient HIV prevalence, the finding of no cases of undiagnosed HIV in this present study of relatively small sample size should not dismiss the importance of inpatient screening. Further study to specifically address the prevalence of HIV on an inpatient unit is required. Overall, POC HIV testing is relatively simple to administer on the medicine inpatient unit and requires minimal prerequisite training. The method of serum acquisition by fine-needle finger poke provides patients and their care providers with a rapid and accurate result through comparatively minimally invasive means. Comparing the cost of one POC test ($15.57 CAD) and the cost of running one ELISA for HIV-1/2 and p24 antigen ($10.72 CAD), the cost-effectiveness of implementing routine POC testing would not be derived from testing alone, but rather, by reducing impending healthcare costs on the burden of illness in patients living with undiagnosed HIV. In practice, POC HIV testing may be particularly useful in settings where knowing HIV status changes the selection of empiric antibiotic coverage, in very short inpatient admissions or in patients with decreased access to services in rural or remote regions.

he burden of illness in patients living with undiagnosed HIV. In practice, POC HIV testing may be particularly useful in settings where knowing HIV status changes the selection of empiric antibiotic coverage, in very short inpatient admissions or in patients with decreased access to services in rural or remote regions. There are several limitations in this study. First, the opt-in nature of the study design resulted in a significant number of patients who declined POC HIV testing, which may have contributed to selection bias towards patients who were at low risk of having HIV. Previous studies also show low rates of HIV testing uptake in patients offered HIV POC testing by an opt-in approach [28–30]. Reasons for patient refusal of HIV testing include recent testing outside of the study, lack of perceived risk, desire to focus on the primary reason for the visit, fear of psychosocial consequences of diagnosis, and concerns about confidentiality following a diagnosis [20, 31]. It has been shown that patients who decline HIV testing may be at higher risk than those who accept testing [32, 33]. Information that may have provided further insight into reasons for testing refusal, such as demographics, HIV risk factors, and previous HIV testing, was not collected from patients who refused to participate in the study. Without this information, there can only be speculation as to how the collected data may have been biased towards having tested patients at lower risk of having HIV in our study sample. Second, three participants had either indeterminate or falsely positive POC HIV testing results. The sensitivity and specificity of the bioLytical INSTI HIV-1/HIV-2 antibody tests have been reported to be 99–100% and 99%, respectively, based on previously published validation studies [34, 35]. As the objective of this study was not to validate the intrinsic accuracy of this test, there can be no conclusions made from the results about this aspect of the test. However, it is well known that the positive predictive value of the test is dependent upon the pretest probability and population prevalence of HIV. It is, therefore, possible that the aforementioned selection bias for low-risk individuals influences the positive predictive value of the POC HIV test in this study.

est. However, it is well known that the positive predictive value of the test is dependent upon the pretest probability and population prevalence of HIV. It is, therefore, possible that the aforementioned selection bias for low-risk individuals influences the positive predictive value of the POC HIV test in this study. 5. Conclusions POC HIV testing is well-accepted among patients who receive it and can be used to be used in conjunction with other strategies to overcome the barriers to HIV testing. This study did not identify cases of undiagnosed HIV. However, the sample size was small and may have been biased towards testing lower-risk patients given the opt-in nature of performing POC testing. The proportion of results that were falsely positive or indeterminate may have been anomalously high when compared to previously reported data regarding the sensitivity and specificity of the POC test. This may relate to the overall low prevalence of HIV on the inpatient units we studied; however, unfortunately, this can only be speculated as we did not collect data on the group that refused to participate. Nevertheless, there was a favourable attitude towards the POC HIV test among patients who agreed to be tested, and the test itself was simple and unobtrusive to perform on the medicine unit. Further study utilizing an opt-out approach to testing may more accurately characterize the prevalence of undiagnosed HIV on the adult internal medicine inpatient unit. It has been shown that POC HIV testing would be a feasible method of approaching a screening initiative of this kind and may also capture important data missing from this present study.

approach to testing may more accurately characterize the prevalence of undiagnosed HIV on the adult internal medicine inpatient unit. It has been shown that POC HIV testing would be a feasible method of approaching a screening initiative of this kind and may also capture important data missing from this present study. Acknowledgments HIV point-of-care testing kits were supplied by Gilead Sciences. Ethical Approval Ethics approval was obtained prior to the commencement of this study. The University of Manitoba Human Research Ethics Board protocol number for this study is 2015:414. Conflicts of Interest The authors declare that there are no conflicts of interest regarding the publication of this paper. Figure 1 Study design and patient recruitment. Table 1 Reasons for patient exclusion from study participation (n = 164). Reason Prevalence Refused to participate in study (too tired, too ill, etc.) 58 (35%) Did not want to know HIV status 7 (4.3%) Perceived to lack risk 6 (3.6%) Palliative 3 (1.8%) No reason given 37 (23%) Unable to consent 51 (31%) Already known to have HIV 2 (1.2%) Table 2 Self-reported patient baseline demographics (n = 142). n (%) Gender Male 68 (48%) Female 73 (51%) Transgender 1 (1%) Primary location of residence Winnipeg 98 (68%) Rural Manitoba, nonreserve 18 (12%) First Nation Reserve 19 (15%) Outside Manitoba 7 (4%) Ethnicity Caucasian 77 (54%) First Nation, Métis, or Inuit 52 (37%) Asian 4 (3%) African Canadian 2 (1%) Latin American 1 (1%) Arab, West Asian, South Asian and others 6 (4%) Country of birth

Primary location of residence Winnipeg 98 (68%) Rural Manitoba, nonreserve 18 (12%) First Nation Reserve 19 (15%) Outside Manitoba 7 (4%) Ethnicity Caucasian 77 (54%) First Nation, Métis, or Inuit 52 (37%) Asian 4 (3%) African Canadian 2 (1%) Latin American 1 (1%) Arab, West Asian, South Asian and others 6 (4%) Country of birth Canada 126 (89%) Outside of Canada 16 (11%) Europe 6 (4%) South America 4 (3%) West Pacific 4 (3%) Africa 1 (1%) Unknown 1 (1%) Table 3 Results of post-POC testing questionnaire regarding attitudes towards POC HIV test. Aspect of POC HIV test on questionnaire Strongly agree or agree, n (%) Neutral, disagree, or strongly disagree, n (%) No response, n (%) Was satisfied with the POC HIV testing experience 131 (92%) 9 (6%) 2 (1%) Would choose to have a POC HIV test again 123 (87%) 18 (13%) 1 (1%) Felt anxious during the POC HIV test 35 (25%) 107 (75%) 0 (0%)

1. Introduction Mycobacterium tuberculosis (M.tb), the causative agent of tuberculosis (TB), is a facultative intracellular bacterium that can survive and replicate within host macrophages [1, 2]. By avoiding critical components of macrophage-killing repertoire such as phagosome-lysosome fusion, phagosome acidification, activity of lysosomal enzymes or reactive oxygen, and nitrogen intermediates, M.tb evades killing and eradication [3]. In addition to phagocytic activity and ability to present antigens to T-cells, macrophages are key cells that regulate the antimycobacterial immune response via secreted cytokines. The functional capacity of macrophages in fighting infection depends on the degree of their activation. Inactive macrophages have limited ability to inhibit the growth of ingested mycobacteria, thereby serving as a safe life niche. After activation by interferon-gamma (IFN-γ) that is secreted by T-cells, macrophages acquire enhanced bactericidal strength enabling them to kill mycobacteria growing intracellularly [4]. The IFN-γ-driven antimicrobial properties of phagocytes are augmented by IL-18 and IL-1β, two proinflammatory cytokines processed by caspase-1 that are recruited to the inflammasomes, multiprotein platforms composed inter alia of intracellular sensors for pathogen- or host-derived molecules. IL-18, belonging to the IL-1 family, is produced by a wide range of immune and nonimmune cells [5–7]. The IL-18 precursor (pro-IL-18) is converted by caspase-1 into an active molecule, which forms a signaling complex with IL-18R [8, 9]. The receptor is composed of two chains: alpha (IL-18Ra) and beta (IL-18Rb). IL-18Rb is a signal transduction chain, essential for the formation of a high affinity complex and cell activation. The primary role of IL-18 is to induce IFN-γ production in cooperation with IL-12 or IL-15, although immunological effects exerted by IL-18 are dependent on the cytokine microenvironment.

beta (IL-18Rb). IL-18Rb is a signal transduction chain, essential for the formation of a high affinity complex and cell activation. The primary role of IL-18 is to induce IFN-γ production in cooperation with IL-12 or IL-15, although immunological effects exerted by IL-18 are dependent on the cytokine microenvironment. IL-18 is able to polarize T lymphocyte response towards Th1, induce T-cell proliferation, activate NK cells, enhance CD8(+) T cytolytic activity, and augment, apart from IFN-γ, the production of varied cytokines including tumor necrosis factor-α (TNF-α), interleukin- (IL-) 4, IL-5, IL-13, IL-17, and granulocyte-macrophage colony stimulating factor (GM-CSF) [8, 10, 11]. Thus, the multifaceted activity of IL-18 seems to play a prominent role in host defense against both extracellular and intracellular pathogens, including M.tb. However, an excessive IL-18 response might contribute to the induction of pathomechanisms leading to the damage of cells and tissues [12, 13]. Therefore, the proinflammatory activity of IL-18 is balanced by a constitutively secreted IL-18 binding protein (IL-18BP), whose binding to IL-18 decreases the production of IFN-γ and other cytokines, thereby reducing the risk of immunopathology [14]. The other inflammasome-dependent cytokine, IL-1β, which is mainly produced by monocytes and macrophages, plays an important role in inflammation and host immune response by affecting the function of various cells, either alone or in combination with other cytokines [15–17]. The activity of IL-1β is tightly regulated at the levels of its transcription and release. The production of IL-1β is regulated by several proteins including pyrin, PI-9 (the caspase-1 inhibitor proteinase inhibitor 9), and some CARD-containing proteins, which interfere with the recruitment of caspase-1 or directly neutralize its activity [18]. The effects of IL-1β are exerted via binding specific cell surface receptors—IL-1RI and IL-1RII [19]. As in the mature IL-18 form, active IL-1β is created after the proteolytic cleavage of its precursor by inflammasome-dependent caspase-1. Mature IL-1β plays important homeostatic functions in organisms and is implicated in the initiation of antimicrobial immunity via the induction of TNF-α and IL-6 release and polarization of Th17 response, which improve protective mucosal host defense by the secretion of IL-17 and IL-22 [20, 21].

me-dependent caspase-1. Mature IL-1β plays important homeostatic functions in organisms and is implicated in the initiation of antimicrobial immunity via the induction of TNF-α and IL-6 release and polarization of Th17 response, which improve protective mucosal host defense by the secretion of IL-17 and IL-22 [20, 21]. The proinflammatory role of IL-1β in the resistance against M.tb has been confirmed by the observation that IL-1β or IL-1R knockout mice were found to be more susceptible to TB showing high mortality and increased bacterial burden in the lungs [22]. Additionally, double-deficient IL-1α/β mice had significantly larger granulomas, and their alveolar macrophages produced less nitric oxide than the cells from wild-type animals [23].

β or IL-1R knockout mice were found to be more susceptible to TB showing high mortality and increased bacterial burden in the lungs [22]. Additionally, double-deficient IL-1α/β mice had significantly larger granulomas, and their alveolar macrophages produced less nitric oxide than the cells from wild-type animals [23]. 2. Inflammasomes—Mediators of Inflammation Inflammation is an evolutionarily conserved protective response to noxious stimuli mounted by the innate immune system of the host. Immune deficiencies leading to insufficient development of inflammation processes may result in severe and recurrent infections, although overly intense activation of the inflammation cascade may be a cause of chronic systemic inflammatory disorders [24, 25]. The development of innate immunity starts from the recognition of conservative antigenic structures called DAMPs (danger-associated molecular patterns) and PAMPs (pathogen-associated molecular patterns) by pattern recognition receptors (PRRs) presented on the surface of first-line defense immune cells—macrophages and neutrophils. Activation of these receptors triggers a cascade of signals that results in the induction of multiple proinflammatory cytokines. The final step of the activation is the production of oxygen and nitrogen radicals, essential elements of the intracellular killing system. The secretion of these radicals is under strict control of a variety of monocyte/macrophage-derived cytokines such as IL-1β and IL-18. The key role in this process is played by structures called inflammasomes, multimeric protein complexes that control many aspects of innate and adaptive immunity. Through their cooperation with PRRs, inflammasomes activate host defense pathways resulting in clearance of various viral and bacterial infections, including those caused by mycobacteria. They function as an activating scaffold for inflammatory caspases that play an essential role in the maturation and secretion of proinflammatory cytokines as well as in pyroptosis, an inflammatory death of infected cells [26, 27]. Caspases are produced as inactive proenzymes that dimerize and undergo cleavage to form active molecules. Assembly into dimers, facilitated by various adaptor proteins binding to specific regions of their precursor forms—procaspases, is achieved through inflammasome formation [28]. Activated inflammatory caspases, typically caspase-1, lead to the generation of active IL-1β, IL-18, and IL-33 from their proprotein precursors.

ssembly into dimers, facilitated by various adaptor proteins binding to specific regions of their precursor forms—procaspases, is achieved through inflammasome formation [28]. Activated inflammatory caspases, typically caspase-1, lead to the generation of active IL-1β, IL-18, and IL-33 from their proprotein precursors. The mature cytokines are engaged in the recruitment of immune cells to the sites of infection and enhancement of the host's defensive responses against invading pathogens [26]. The inflammasomes are activated by multiple recognition receptors, which determine their structure and function. The canonical inflammasome sensors are nucleotide-binding domain–like (NLR) proteins and absent in melanoma 2–like (ALR) proteins and PYRIN. All of them have the ability to assemble inflammasomes and activate the inflammatory caspase-1.

ed by multiple recognition receptors, which determine their structure and function. The canonical inflammasome sensors are nucleotide-binding domain–like (NLR) proteins and absent in melanoma 2–like (ALR) proteins and PYRIN. All of them have the ability to assemble inflammasomes and activate the inflammatory caspase-1. The NLR family contains the NLRPs (or NALPs) and the IPAF (ICE-protease-activating factor) subfamilies [29, 30]. Each NLR molecule (NLRP1, NLRP3, NLRP6, NLRP7, NLRP12, or NAIP/NLRC4) recognizes specific ligands that activate the assembly of the inflammasome. NLR proteins consist of the conserved nucleotide-binding and oligomerization domain (NACHT or NOD), an N-terminal caspase recruitment domain (CARD) or pyrin domain (PYD) or baculovirus inhibitor repeat- (BIR-) like domain, and C-terminal leucine rich repeats (LRRs) [26, 31–35]. LRRs are responsible for the recognition of PAMPs, while the NACHT domain activates proinflammatory cytokine pathways via ATP-dependent oligomerization [26, 29]. The NLRP1 inflammasome has a CARD that activates caspase-1 [36, 37], and therefore the recruitment of ASC is not required to interact directly with procaspase-1. However, it has been shown that the participation of ASC in the process enhanced the activation of the enzyme. In contrast, NLRP3 contains no typical CARD domain that contributes to the activation of caspase-1 through the interaction of the PYD domain of NLRP3 with ASC [25]. Compared with NLRP1 and NLRP3, the IPAF protein does not contain a PYD but instead has a CARD that interacts directly with procaspase-1 without the need for ASC [38].

, NLRP3 contains no typical CARD domain that contributes to the activation of caspase-1 through the interaction of the PYD domain of NLRP3 with ASC [25]. Compared with NLRP1 and NLRP3, the IPAF protein does not contain a PYD but instead has a CARD that interacts directly with procaspase-1 without the need for ASC [38]. The members of the ALR group (known as the PYHIN family) are characterized by the presence of the pyrin domain (PYD) and one or two hematopoietic IFN-inducible nuclear antigens with 200 amino acid repeat (HIN-200) domains [26]. The PYD recruits proteins for the formation of inflammasomes, while the HIN domain recognizes and binds to DNA that can be found in the cytosol [26]. The best-known ALRs, absent in melanoma 2 (AIM2) and IFN-γ inducible protein 16 (IFI16), function as intracellular immune sensors that detect microbial DNA. The PYHIN proteins differ in their localization in the cell compartments; AIM2 can be found in the cytosol, whereas IFI16 is usually localized in the nucleus [39].

The best-known ALRs, absent in melanoma 2 (AIM2) and IFN-γ inducible protein 16 (IFI16), function as intracellular immune sensors that detect microbial DNA. The PYHIN proteins differ in their localization in the cell compartments; AIM2 can be found in the cytosol, whereas IFI16 is usually localized in the nucleus [39]. PYRIN, another canonical inflammasome-activating protein, is composed of an N-terminal PYD followed by two central B-box zinc finger and coiled-coil domains and in humans, a C-terminal B30.2/rfp/SPRY domain [40]. PYRIN associates through a PYD-PYD interaction with ASC protein, leading to its oligomerization that results in caspase-1 activation and interleukin-1β processing [40]. The activation of the PYRIN inflammasome is induced by the inactivation of RhoA GTPase by bacterial toxins [26, 41]. The process of activation has been detected in both mice and humans, suggesting that the B30.2/rfp/SPRY domain is not necessary for its initiation. 3. Inflammasomes in Mycobacterium tuberculosis Infection The inflammasomes have been found to play important roles in host immunity against mycobacteria since it has been found that mice deficient in IL-18, IL-1β, or IL-1 receptor type I (IL-1R1) are more susceptible to M.tb infection [42–46]. Two inflammasomes, containing NLRP3 and AIM2 molecules as sensor proteins, were found to play a crucial role in M.tb-induced immunity (Figure 1) [20, 47, 48].

gainst mycobacteria since it has been found that mice deficient in IL-18, IL-1β, or IL-1 receptor type I (IL-1R1) are more susceptible to M.tb infection [42–46]. Two inflammasomes, containing NLRP3 and AIM2 molecules as sensor proteins, were found to play a crucial role in M.tb-induced immunity (Figure 1) [20, 47, 48]. The NLRP3-containing inflammasome can be activated by a wide group of stimuli including whole mycobacterial cells, as well as viruses, fungi, environmental chemical irritants, and host-derived molecules such as extracellular ATP, fibrillar amyloid-β peptide, and hyaluronan [22, 49–53]. The NLRP3 inflammasome-activated responses result in the release of significant amounts of caspase-1, which leads to maturation and secretion of IL-1β and IL-18 and activation of pyroptosis [26]. The process of NLRP3 activation is triggered by at least two signals: (1) a priming signal eliciting the expression of NLRP3, pro-IL-1β, and pro-IL-18 genes after TLR stimulation and (2) an activation signal leading to the autocatalytic activation of procaspase-1 and proteolytic cleavage of pro-IL-1β and pro-IL-18. In most cell types, NLRP3 priming is a prerequisite for deubiquitination and assembly of the NLRP3 inflammasome. Relocalization of NLRP3 to the mitochondria is followed by the secretion of mitochondrial factors into the cytosol, potassium efflux through membrane ion channels, and release of cathepsin resulting in destabilization of lysosomal membranes. Apoptosis-associated speck-like protein (ASC) plays an important role in the formation of an effective inflammasome. ASC recruits procaspase-1 through its C-terminal caspase recruitment domain (CARD) and interacts with NLRP3 via its pyrin domain (PYD), serving as a bridge between these two molecules. The autocatalysis of procaspase-1 results in its cleavage and transformation into active caspase-1, which in turn cleaves the precursors of two proinflammatory cytokines, IL-1β and IL-18, leading to their secretion into the cytoplasm or induction [24, 25, 48, 54, 55]. However, the mechanism of triggering the NLRP3 inflammasome complex activation cascade is still a subject of debate, and at least three models for the process have been proposed. The first suggestion is that the activation mechanism is associated with an efflux of potassium ions out of the cell and a reduction in their intracellular concentration.

ggering the NLRP3 inflammasome complex activation cascade is still a subject of debate, and at least three models for the process have been proposed. The first suggestion is that the activation mechanism is associated with an efflux of potassium ions out of the cell and a reduction in their intracellular concentration. Such a model of activation occurs in monocytes/macrophages after stimulation with numerous stimuli including ATP, nigericin, bacterial cells, or their components [56, 57]. Recently, NEK7 protein, a member of the family of NIMA-related kinases (NEK proteins), has been identified as an NLRP3-binding protein that acts downstream of potassium efflux to regulate NLRP3 assembly and activation [58]. He et al. demonstrated that in the absence of NEK7, caspase-1 activation and IL-1β release were abrogated in response to signals that activate NLRP3 [58]. According to the second suggested mechanism, inflammasome activation is a result of lysosomal membrane damage and release of the phagosome content into cytosol [22, 59]. The third and most accepted model assumes that the induction of the NLRP3 inflammasome complex is caused by mitochondrial reactive oxygen species (ROS) [60–63]. The common final step in all of these models is the release of cathepsins into the cytosol leading to the lysosomal destabilization and conversion of procaspase-1 into a biologically active caspase-1 form. It should also be mentioned that formation of the NLRP3 inflammasome and cytokine release occur independently of transcriptional upregulation [64]. Juliana et al. showed that TLR4 signaling through MyD88 nontranscriptionally primed the NLRP3 inflammasome by its deubiquitination. The mechanism was dependent on mitochondria-derived reactive oxygen species and was involved in the secretion of cytokines, such as IL-18, and other inflammatory mediators such as high-mobility group protein 1 (HMGB1) [64, 65].

ng through MyD88 nontranscriptionally primed the NLRP3 inflammasome by its deubiquitination. The mechanism was dependent on mitochondria-derived reactive oxygen species and was involved in the secretion of cytokines, such as IL-18, and other inflammatory mediators such as high-mobility group protein 1 (HMGB1) [64, 65]. The AIM2 (absent in melanoma 2) receptor, possessing a C-terminal HIN-200 domain and an N-terminal pyrin domain (PYD), triggers AIM2 inflammasome activation, inflammatory cell death (pyroptosis), and release of IL-1β and IL-18 in response to cytosolic double-stranded (ds) DNA [66, 67]. Studies of gene-targeted AIM2-deficient mice have shown that AIM2 inflammasomes play a role in host defense against viruses and intracellular bacterial pathogens such as listeriae and mycobacteria [68–70]. AIM2 inflammasomes can be activated by DNA sequences having at least 80 base pairs in length in a sequence-independent manner [71, 72]. The HIN-200 and PYD domains take part in forming a complex, which is maintained in an inactive state during homeostasis [71, 73]. Binding of dsDNA to HIN-200 facilitates oligomerization of AIM2, and the resulting conformational change exposes the N-terminal PYD to allow the recruitment of the adaptor protein ASC. The CARD of ASC binds the CARD of procaspase-1, that forms an active AIM2 platform. Upon autoactivation, caspase-1 directs maturation and secretion of proinflammatory cytokines [48, 55, 66, 68, 74].

, and the resulting conformational change exposes the N-terminal PYD to allow the recruitment of the adaptor protein ASC. The CARD of ASC binds the CARD of procaspase-1, that forms an active AIM2 platform. Upon autoactivation, caspase-1 directs maturation and secretion of proinflammatory cytokines [48, 55, 66, 68, 74]. The latest data suggest that NLRP3- or ASC-deficient animals are characterized by impaired inflammasome formation and increased susceptibility to TB [20, 54, 68, 75, 76]. However, NLRP3−/− and ASC−/− mice produced IL-18 and IL-1β levels comparable to those of wild-type mice, which suggests the involvement of inflammasome-independent pathways in the secretion of these cytokines [21, 42, 47]. Many reports have demonstrated that a wide range of microorganisms are able to inhibit inflammasome activation and function. Viruses and many bacterial pathogens develop several mechanisms of repression of inflammasome folding; however, not all mechanisms are clearly understood. Yersinia enterocolitica produce YopE and YopT proteins that supress caspase-1 maturation, whereas YopK protein of Y. pseudotuberculosis binds to the type III secretion system, thereby preventing the recognition of the pathogen by host cell inflammasome. Pseudomonas aeruginosa mediates suppression of NLRC4-inflammasome by secreting ExoU and ExoS effectors, whose mechanism of action still needs elucidation. Virulent M.tb can inhibit the formation of AIM2 and NLRP3 inflammasomes both directly and indirectly, but the factors responsible for the inhibition have not been recognized thus far. One of the likely mechanisms is the activity of Zn-metalloprotease called ZMP1, which inhibits the activation of NLRP3 inflammasome and, as a consequence, leads to the reduction of caspase-1 activity [77–79]. Master et al. showed that infection of mice macrophages with zmp1-deleted M.tb induces activation of the inflammasome, resulting in enhanced maturation of phagosomes, increased IL-1β secretion, and better M.tb clearance in lungs [79]. It is probable that M.tb is able to restrain the activation of other inflammasome types, but evidence is needed to confirm this hypothesis. In addition to the induction of inflammasome activation via PRRs, M.tb antigens can modulate other innate immunity-associated functions.

n, and better M.tb clearance in lungs [79]. It is probable that M.tb is able to restrain the activation of other inflammasome types, but evidence is needed to confirm this hypothesis. In addition to the induction of inflammasome activation via PRRs, M.tb antigens can modulate other innate immunity-associated functions. One recently identified protein, tyrosine phosphatase (Ptp) A, enters the nucleus of the host cells and regulates the transcription of many host genes involved in the mechanisms of innate immunity, cell proliferation, and migration [80]. The enzyme is also able to dephosphorylate certain host proteins (p-JNK, p-p38, and p-VPS33B), leading to inhibition of phagosome-lysosome fusion and blocking the acidification of phagosomes. Both activities are crucial for M.tb virulence in vivo through the promotion of M.tb's intracellular survival in macrophages [80]. M.tb often escapes from the phagosome within a few days of the invasion of the host organism and creates difficulties in assessing the potential role of inflammasomes during the initial stages of mycobacterial infection. Moreover, the evaluation of IL-1β and IL-18 produced as a result of inflammasome activation is inadequate in revealing the significance of formed multiprotein platforms in the course of developing infection. The initiation of phagocytosis causes a decrease in the levels of potassium ions in macrophages, which have been found to be one of the crucial inflammasome activators during infections with M.tb and nontuberculous mycobacteria [81]. Other regulators such as thioredoxin-interacting proteins, activated by the increase in reactive oxygen species in cytosol, are thought to have minor effect on the formation of inflammasomes in M.tb infection [47]. The signaling cascade can also be activated by the mycobacterial type VII secretion system (ESX-1), which is responsible for translocation of extracellular DNA (eDNA) in cytosol and the production of IFN-β. Many studies have demonstrated that, at the molecular level, IFN-β regulates the AIM2 inflammasome activity [82, 83]. Some ESX-1-deficient M. smegmatis mutants have been shown to possess limited capacity for AIM2 inflammasome activation. However, in contrast to nontuberculous mycobacteria (NTM), M.tb mutants lacking ESX-1 system failed to inhibit AIM2 formation, while the wild-type strain inhibited the inflammasome activation [47, 84].

ome ESX-1-deficient M. smegmatis mutants have been shown to possess limited capacity for AIM2 inflammasome activation. However, in contrast to nontuberculous mycobacteria (NTM), M.tb mutants lacking ESX-1 system failed to inhibit AIM2 formation, while the wild-type strain inhibited the inflammasome activation [47, 84]. The suggested mechanism of inhibition involves the IFN-β-mediated induction of IL-10, which in turn suppresses IL-1β production [85, 86]. However, further investigation is needed to elucidate the molecular mechanism of M.tb-driven AIM2 inhibition and its consequences for bacterial virulence. M. bovis BCG vaccine strain, which does not possess the ESX-1 system, poorly activates multiple NLR and inflammasome complex components including caspase-1 [87]. The bacilli repress the expression of thioredoxin-interacting protein (TXNIP), an antioxidant inhibitor recruiting caspase-1 to the NLRP3 inflammasome. The inhibition of TXNIP by BCG limits NLRP3 activation and restrains pyroptosis following mycobacterial infection. Proinflammatory responses to BCG bacilli was found to be driven primarily through Toll-like receptors (TLRs), since BCG does not activate expression of genes downstream of TLR/MyD88- and NOD-2-driven NF-κβ and AP-1 pathways. However, BCG is still able to induce moderate IL-1β secretion as measured by transcription of inflammasome network genes [87, 88]. Understanding BCG-induced pathways of inflammasome activation can be helpful in improving the existing vaccine or developing new antituberculous vaccines. The recombinant BCG ΔureC::hly vaccine candidate (VPM1002) has been shown to induce improved protection against TB over the parental BCG strain [4]. Saiga et al. demonstrated that VPM1002 activated the AIM2 inflammasome and caspase-1 through the ability of listeriolysin to perforate phagosome membranes, which is encoded by the hly gene integrated into BCG genome [4]. The perforation facilitates the release of mycobacterial DNA into the cytosol, in a way that is similar to the ESX-1 system of M.tb. Mice vaccinated with VPM1002 showed increased production of IL-1β and IL-18 as well as induction of the stimulator of IFN genes (STING)-dependent autophagy, which promotes delivery of BCG antigens to MHC molecules and improves their presentation to T-cells [4].

sol, in a way that is similar to the ESX-1 system of M.tb. Mice vaccinated with VPM1002 showed increased production of IL-1β and IL-18 as well as induction of the stimulator of IFN genes (STING)-dependent autophagy, which promotes delivery of BCG antigens to MHC molecules and improves their presentation to T-cells [4]. Apart from direct induction of proinflammatory cytokine secretion, the activated caspase-1 triggers the pyroptotic death of infected cells. The cytosolic protein Gasdermin D (GSDMS) is a key mediator of this process. The cleavage of GSDMD by activated caspase-1 results in the release of its N-terminal fragment (GSDMD-NT), which forms pores in the plasma membrane of the infected cell leading to the elimination of the pathogen [26, 89–91]. The pores disrupt cell membrane integrity allowing water influx, cell swelling, and osmotic lysis together with an efflux of small molecules, including proinflammatory cytokines. GSDMD-NT is able to kill both cell-free and intracellular microorganisms and can be thought as a new antibacterial agent. However, it is still not known whether GSDMD-NT is able to permeabilize the membrane of the phagosomes and kill the bacteria hidden within these organelles. So far, there is no evidence of such a function. It is probable that the inhibition of bacterial growth is mediated by other caspase components. Using single-cell analysis, Thurston et al. demonstrated that the replication of cytosolic Salmonella typhimurium was inhibited independently or prior to the onset of cell death, suggesting that caspase-1 and caspase-11 might have additional functions in the elimination of cytosolic bacteria [92].

r caspase components. Using single-cell analysis, Thurston et al. demonstrated that the replication of cytosolic Salmonella typhimurium was inhibited independently or prior to the onset of cell death, suggesting that caspase-1 and caspase-11 might have additional functions in the elimination of cytosolic bacteria [92]. 4. Therapeutics Targeting Inflammasome Pathways Biologic agents interfering with inflammasome activation may provide new means of therapeutical interventions for many diseases. These agents may target either upstream processes of inflammasome regulation or downstream IL-1 signaling [41]. Inappropriate activity of inflammasomes has been found to be involved in the pathogenesis of certain autoinflammatory skin disorders such as cryopyrin-associated periodic syndrome (CAPS) or familial Mediterranean fever (FMF) as well as a number of chronic inflammatory diseases such as multiple sclerosis, gouty arthritis (gout), atherosclerosis, type 2 diabetes, and obesity [29, 93, 94]. Moreover, mechanisms controlling the NLRP3 inflammasome arrangement have also been implicated in the development of lung, kidney, and liver diseases [95–97]. Colchicine, a drug used for treatment of gout, has been shown to inhibit macrophage NLRP3 inflammasome assembly and activation in vitro and in vivo [98]. Colchicine blocks monosodium urate crystal-induced NLRP3 inflammasome-driven caspase-1 activation and IL-1β processing and release, suppresses the expression of genes involved in cell regulation, and inhibits IL-1-induced L-selectin expression on neutrophils [99]. Other therapeutics that target inflammasome-driven end products include VX-765 (inhibitor of caspase-1 activation), Anakinra (recombinant form of IL-1 receptor antagonist), Canakinumab (monoclonal antibody against IL-1β), Rilonacept (IL-1 inhibitor), IL-18 binding protein, and anti-IL-18 receptors antibodies [8, 41, 100]. A number of new molecules have been identified as inhibitors of IL-1β processing (glyburide, parthenolide, CRID3, auranofin, isoliquiritigenin, β-hydroxybutyrate, and MCC950); however, confirming their clinical utility will require additional time and research [24].

in, and anti-IL-18 receptors antibodies [8, 41, 100]. A number of new molecules have been identified as inhibitors of IL-1β processing (glyburide, parthenolide, CRID3, auranofin, isoliquiritigenin, β-hydroxybutyrate, and MCC950); however, confirming their clinical utility will require additional time and research [24]. 5. Conclusion Inflammasomes have been implicated as specialized signaling platforms critical for the regulation of both innate immunity and inflammation. M.tb has been shown to modulate the host innate immune response by delaying cell death systems of the host, thereby facilitating its own proliferation. Understanding the molecular mechanisms of inflammasome activation during intracellular pathogen infections such as with M.tb, and the evasive mechanisms employed by this evading pathogen, may lead to development of more potent therapies to combat the proliferation of M.tb. Acknowledgments This work was supported by the National Science Centre Grant nos. 2015/19/N/NZ6/01385 and 2016/21/B/NZ7/01771. Conflicts of Interest The authors declare that there are no conflicts of interest regarding the publication of this article. Figure 1 AIM2 and NLRP3 inflammasome activation pathways induced by Mycobacterium tuberculosis.

1. Introduction Recently, infections with drug-resistant bacterial strains have increased which poses a challenge for anti-infection treatments in the clinical context and has become a serious problem in the field of public health. The incidence of bloodstream infections is also increasing and has become a major cause of the occurrence of infectious diseases and deaths worldwide [1]. Bloodstream infections with multidrug-resistant Enterobacteriaceae account for 71.5% of all bloodstream infections with multidrug-resistant bacterial strains [2]. Carbapenem antibiotics are typically one of the most effective drug classes for the treatment of Enterobacteriaceae infections. However, carbapenem-resistant Enterobacteriaceae (CRE) has been identified recently within all Enterobacteriaceae spp. The CHINET monitoring results show that Klebsiella pneumoniae accounts for the majority of cases of CRE infections among the Enterobacteriaceae spp., with drug resistance rates to imipenem and meropenem > 10% [3]. Studies have shown that the mortality rate is significantly higher for CRE infections than for carbapenem-sensitive Enterobacteriaceae infections [4]. This study was a retrospective analysis of medical information on bloodstream infections with Enterobacteriaceae (mainly Klebsiella pneumoniae, Escherichia coli, Enterobacter cloacae, and Enterobacter aerogenes) with intent to identify the risk factors and prognosis for CRE bloodstream infections.

1. Introduction Recently, infections with drug-resistant bacterial strains have increased which poses a challenge for anti-infection treatments in the clinical context and has become a serious problem in the field of public health. The incidence of bloodstream infections is also increasing and has become a major cause of the occurrence of infectious diseases and deaths worldwide [1]. Bloodstream infections with multidrug-resistant Enterobacteriaceae account for 71.5% of all bloodstream infections with multidrug-resistant bacterial strains [2]. Carbapenem antibiotics are typically one of the most effective drug classes for the treatment of Enterobacteriaceae infections. However, carbapenem-resistant Enterobacteriaceae (CRE) has been identified recently within all Enterobacteriaceae spp. The CHINET monitoring results show that Klebsiella pneumoniae accounts for the majority of cases of CRE infections among the Enterobacteriaceae spp., with drug resistance rates to imipenem and meropenem > 10% [3]. Studies have shown that the mortality rate is significantly higher for CRE infections than for carbapenem-sensitive Enterobacteriaceae infections [4]. This study was a retrospective analysis of medical information on bloodstream infections with Enterobacteriaceae (mainly Klebsiella pneumoniae, Escherichia coli, Enterobacter cloacae, and Enterobacter aerogenes) with intent to identify the risk factors and prognosis for CRE bloodstream infections. 2. Materials and Methods 2.1. Study Subjects A total of 148 cases of bloodstream infections with Enterobacteriaceae that occurred during the period between January 2013 and October 2015 in Fuxing Hospital of Capital Medical University, China, were selected. Clinical information from the patients, including demographic data, underlying diseases, pathogen drug resistance and drug sensitivity analyses, history of medical institution admission, level of care (ward versus ICU), history of invasive surgery, and prognosis, was recorded and analyzed. The characteristics of each case were collected from the electronic medical record database, and the files of all patients were complete.

tance and drug sensitivity analyses, history of medical institution admission, level of care (ward versus ICU), history of invasive surgery, and prognosis, was recorded and analyzed. The characteristics of each case were collected from the electronic medical record database, and the files of all patients were complete. Inclusion criteria: On the day of collection of a positive blood culture specimen, the patient had a temperature ≥ 38°C or < 36°C in combination with chills, and the multiple blood culture results showed the same type of bacteria. Exclusion criteria: (1) blood culture test results showing two or more types of bacteria and (2) only one blood culture specimen was positive, whereas multiple subsequent blood culture results were negative or identified other types of pathogens. This study conformed to the standards of medical ethics and was approved by the Ethics Committee of Fuxing Hospital of Capital Medical University.

Inclusion criteria: On the day of collection of a positive blood culture specimen, the patient had a temperature ≥ 38°C or < 36°C in combination with chills, and the multiple blood culture results showed the same type of bacteria. Exclusion criteria: (1) blood culture test results showing two or more types of bacteria and (2) only one blood culture specimen was positive, whereas multiple subsequent blood culture results were negative or identified other types of pathogens. This study conformed to the standards of medical ethics and was approved by the Ethics Committee of Fuxing Hospital of Capital Medical University. 2.2. Relevant Definitions The microbiology laboratory in our hospital currently uses an FX200 automatic blood culture instrument (BACTE) and a WalkAway 40 automatic microbiological drug sensitivity analysis instrument (USA). The drug sensitivity results were determined according to the Clinical and Laboratory Standards Institute (CLSI, USA) standards [5]. The definition of CRE was resistance to imipenem, meropenem, or ertapenem [3]. Extensively drug-resistant (XDR) bacteria were defined based on nonsusceptibility to all antimicrobial agents except colistin and tigecycline [6, 7]. Long-term bedbound patients were defined as patients who were bedbound for more than 14 days and could not recover [7].

sistance to imipenem, meropenem, or ertapenem [3]. Extensively drug-resistant (XDR) bacteria were defined based on nonsusceptibility to all antimicrobial agents except colistin and tigecycline [6, 7]. Long-term bedbound patients were defined as patients who were bedbound for more than 14 days and could not recover [7]. 2.3. Statistical Methods The data processing and analysis were performed using SPSS 21.0 statistical software. P < 0.05 indicated that the difference was significant. Measurement data with a normal distribution are expressed using the mean ± standard deviation. Measurement data with a nonnormal distribution are expressed using the median. Significance testing for differences between groups was performed using the t-test (normal distribution) or rank sum test (nonnormal distribution). Significance testing for differences in count data between groups was performed using the χ2 test. The multivariate analysis was performed using logistic regression.

edian. Significance testing for differences between groups was performed using the t-test (normal distribution) or rank sum test (nonnormal distribution). Significance testing for differences in count data between groups was performed using the χ2 test. The multivariate analysis was performed using logistic regression. 3. Results 3.1. Demographic Data and Clinical Characteristics This study involved a total of 148 cases of bloodstream infections with Enterobacteriaceae. The mean age of the patients was 69.1 (±19.2) years. Male patients accounted for 50.7% of the cases, and 20.3% of the patients were admitted to the ICU. Long-term bedbound patients accounted for 54.1% of the cases, and 62.2% of the patients had a medical institution admission history within 3 months. The bloodstream infections mainly originated from the urinary tract (28.4%) or respiratory system (25.0%) or were catheter related (22.3%). Patients who had prior use of broad-spectrum antibiotics within 30 days accounted for 43.2% of the cases. The total average length of hospital stay was 21.7 (±12.5) days. Patients with CRE infections accounted for 17.6% of the cases, and patients with XDR Enterobacteriaceae infections accounted for 4.0% of the cases.

. Patients who had prior use of broad-spectrum antibiotics within 30 days accounted for 43.2% of the cases. The total average length of hospital stay was 21.7 (±12.5) days. Patients with CRE infections accounted for 17.6% of the cases, and patients with XDR Enterobacteriaceae infections accounted for 4.0% of the cases. The results of the univariate and multivariate analyses of risk factors related to CRE bloodstream infections are shown in Tables 1 and 2, respectively (the relevant factors listed in the tables were all conditions before the presence of the bloodstream infection). The logistic regression analysis revealed that indwelling urethral catheterization, admission to the ICU, CRE infection of the respiratory system, and prior use of broad-spectrum antibiotics within 30 days were independent risk factors for CRE bloodstream infections.

s before the presence of the bloodstream infection). The logistic regression analysis revealed that indwelling urethral catheterization, admission to the ICU, CRE infection of the respiratory system, and prior use of broad-spectrum antibiotics within 30 days were independent risk factors for CRE bloodstream infections. 3.2. Analysis of Risk Factors for Mortality in Bloodstream Infections with Enterobacteriaceae According to the comparison of the 30-day prognosis between groups, the 30-day mortality rate for bloodstream infections with Enterobacteriaceae was 25.7%. Univariate analysis indicated that community acquired, admission to the ICU, long-term bedbound, anemia, hypoalbuminemia, lung infection, abdominal infection, liver dysfunction, mechanic ventilation, central venous catheterization, indwelling urethral catheterization, sources of bloodstream infection from catheter related and urinary tract, antibiotics within 30 days, combined fungal infection, prior use of glucocorticoids within 30 days, and CRE infection were significantly associated with 30-day risk of mortality (Tables 3). A logistic regression analysis revealed that lung infection, abdominal infection, central venous catheterization, and prior use of glucocorticoids within 30 days were independently associated with mortality (Tables 4).

in 30 days, and CRE infection were significantly associated with 30-day risk of mortality (Tables 3). A logistic regression analysis revealed that lung infection, abdominal infection, central venous catheterization, and prior use of glucocorticoids within 30 days were independently associated with mortality (Tables 4). 4. Discussion Currently, the incidence of carbapenemase-producing Enterobacteriaceae is increasing. Data have shown that many countries have interregional spread or an endemic situation [8]. The results of this study showed that the 30-day mortality rate for CRE bloodstream infections reached 65.4%, which was significantly higher than the rate for non-CRE bloodstream infections (17.2%) or for CRE-infected patients in literature reports (26%–44%) [4]. To date, many studies have evaluated risk factors related to bloodstream infections with multidrug-resistant Enterobacteriaceae. For example, factors including admission to the ICU, length of hospital stay, history of use of broad-spectrum antibiotics, such as quinolones and cephalosporins, history of resistant strain colonization, indwelling urethral catheterization, and central venous catheterization have been considered independent risk factors for bloodstream infections with multidrug-resistant Enterobacteriaceae [4, 9]. The analytical results in this study showed that indwelling urethral catheterization, admission to the ICU, prior use of broad-spectrum antibiotics within 30 days, and CRE infection of the respiratory system were independent risk factors for CRE bloodstream infections. The first three factors are similar to previous reports. However, what is different from previous reports is that we identified CRE infection of the respiratory system as an independent risk factor for CRE bloodstream infections. Respiratory tract infection was the most common source of bloodstream infection in our study, likely because of a decreased capacity for bacterial clearance resulting from the altered lung tissue in severe chronic obstructive lung disease.

respiratory system as an independent risk factor for CRE bloodstream infections. Respiratory tract infection was the most common source of bloodstream infection in our study, likely because of a decreased capacity for bacterial clearance resulting from the altered lung tissue in severe chronic obstructive lung disease. The results of this study on risk factors for death due to bloodstream infections with Enterobacteriaceae showed that lung infections, abdominal infections, central venous catheterization, and hormone use within 30 days were all independent risk factors for death due to bloodstream infections with Enterobacteriaceae. The study results showed that the mortality rate was significantly higher in the CRE group than in the non-CRE group. The high mortality associated with CRE bloodstream infections was partially attributed to host conditions, such as endogenous infections and invasive surgery. Additionally, the mortality rate was associated with a lack of timely and accurate antibiotic treatment for infection with drug-resistant bacterial strains or even no antibiotic treatment [10]. XDR strains accounted for 23% of the 26 cases of CRE infections in this study, which was similar to the 28% rate of XDR bacterial strains isolated from CRE infections reported in the literature [8]. This result indicated that approximately one-fourth of CRE infections might be XDR, which is associated with a high mortality rate.

s accounted for 23% of the 26 cases of CRE infections in this study, which was similar to the 28% rate of XDR bacterial strains isolated from CRE infections reported in the literature [8]. This result indicated that approximately one-fourth of CRE infections might be XDR, which is associated with a high mortality rate. During the analysis of mortality-related risk factors in patients with bloodstream Enterobacteriaceae infections in this study, CRE infection was a relevant but not an independent risk factor, which might be associated with the low overall number of cases. In summary, CRE bloodstream infections have a high mortality rate, and treatment has poor clinical efficacy. The use of broad-spectrum antibiotics, indwelling urethral catheterization, admission to the ICU, and infection from the respiratory system were independent risk factors for CRE bloodstream infections. Lung infections, abdominal infections, hormone use, and central venous catheterization were independent risk factors for death in patients with Enterobacteriaceae bloodstream infections. Therefore, the standard use of antibiotics, minimization of invasive operations, use of strict aseptic manipulation, and strengthening of hand-hygiene awareness of healthcare workers may be effective measures for the prevention of CRE bloodstream infections. Conflicts of Interest The authors declare that there are no conflicts of interest. Table 1 Univariate analysis of the clinical characteristics of bloodstream infections caused by CRE and non-CRE.

In summary, CRE bloodstream infections have a high mortality rate, and treatment has poor clinical efficacy. The use of broad-spectrum antibiotics, indwelling urethral catheterization, admission to the ICU, and infection from the respiratory system were independent risk factors for CRE bloodstream infections. Lung infections, abdominal infections, hormone use, and central venous catheterization were independent risk factors for death in patients with Enterobacteriaceae bloodstream infections. Therefore, the standard use of antibiotics, minimization of invasive operations, use of strict aseptic manipulation, and strengthening of hand-hygiene awareness of healthcare workers may be effective measures for the prevention of CRE bloodstream infections. Conflicts of Interest The authors declare that there are no conflicts of interest. Table 1 Univariate analysis of the clinical characteristics of bloodstream infections caused by CRE and non-CRE. Clinical characteristics Bloodstream infections with Enterobacteriaceae, n (%) P value CRE group (n = 26) Non-CRE group (n = 122) Gender (male) 16 (61.5) 59 (48.4) 0.222 Age > 60 years 22 (84.6) 83 (68.0) 0.091 Community acquired 0 30 (24.6) 0.005∗ Admission to the ICU 20 (76.9) 10 (8.2) <0.001∗ Long-term bedbound 19 (73.1) 61 (50) 0.032∗ Medical institution admission history within 3 months 17 (65.4) 75 (61.5) 0.709 Underlying diseases Cardiovascular disease 19 (73.1) 54 (44.3) 0.008∗ Cerebrovascular disease 21 (80.8) 57 (46.7) 0.002∗ Chronic respiratory disease 6 (23.1) 9 (7.4) 0.040∗ Chronic kidney disease 13 (50.0) 33 (27.0) 0.022∗ Diabetes mellitus 9 (34.6) 44 (36.1) 0.889 Malignant tumor 4 (15.4) 44 (36.1) 0.041∗ Anemia 26 (100.0) 87 (71.3) 0.002∗ Hypoalbuminemia 26 (100.0) 98 (80.3) 0.029∗ Lung infection 22 (84.6) 61 (50.0) 0.001∗ Urinary tract infection 12 (46.2) 62 (50.8) 0.666 Abdominal infection 5 (19.2) 14 (11.5) 0.453 Tracheotomy 12 (46.2) 21 (17.2) 0.001∗ Liver dysfunction 13 (50.0) 24 (19.7) 0.001 Invasive surgery before infection Mechanical ventilation 18 (69.2) 12 (9.8) <0.001∗ Central venous catheterization 19 (73.1) 28 (23.0) <0.001∗ Indwelling urethral catheterization 25 (96.2) 45 (36.9) <0.001∗ Sources of bloodstream infection Catheter related 10 (38.5) 23 (18.9) 0.029∗ Respiratory system 11 (42.3) 26 (21.3) 0.025∗ Urinary tract infection 0 (0) 42 (34.4) <0.001∗ Others 5 (19.2) 31 (25.4) 0.505 Prior use of broad-spectrum antibiotics within 30 days 25 (96.2) 39 (32.0) <0.001∗ Combined fungal infection 10 (38.5) 10 (8.2) <0.001∗ Prior use of glucocorticoids within 30 days 8 (30.8) 11 (9.0) 0.007∗ Total length of hospital stay 25.4 ± 11.1 21.0 ± 12.7 0.102 30-Day mortality rate 17 (65.4) 21 (17.2) <0.001∗ 48-Hour mortality rate 6 (23.1) 3 (2.5) <0.001∗ ICU: intensive care unit; ∗P < 0.05.

01∗ Combined fungal infection 10 (38.5) 10 (8.2) <0.001∗ Prior use of glucocorticoids within 30 days 8 (30.8) 11 (9.0) 0.007∗ Total length of hospital stay 25.4 ± 11.1 21.0 ± 12.7 0.102 30-Day mortality rate 17 (65.4) 21 (17.2) <0.001∗ 48-Hour mortality rate 6 (23.1) 3 (2.5) <0.001∗ ICU: intensive care unit; ∗P < 0.05. Table 2 Analysis of risk factors for CRE bloodstream infections (logistic regression analysis). Risk factor OR value 95% CI P value Indwelling urethral catheterization 11.40 1.05 to 124.18 0.046 Admission to the ICU 9.42 2.14 to 41.50 0.003 Respiratory system source 8.95 1.73 to 46.38 0.009 Prior use of broad-spectrum antibiotics within 30 days 11.25 1.19 to 106.79 0.035 OR: odds ratio; 95% CI: 95% confidence interval; respiratory system source: bloodstream infection with Enterobacteriaceae from the respiratory system. Table 3 Univariate analysis for predictors of mortality caused by bloodstream infections with Enterobacteriaceae.

Risk factor OR value 95% CI P value Indwelling urethral catheterization 11.40 1.05 to 124.18 0.046 Admission to the ICU 9.42 2.14 to 41.50 0.003 Respiratory system source 8.95 1.73 to 46.38 0.009 Prior use of broad-spectrum antibiotics within 30 days 11.25 1.19 to 106.79 0.035 OR: odds ratio; 95% CI: 95% confidence interval; respiratory system source: bloodstream infection with Enterobacteriaceae from the respiratory system. Table 3 Univariate analysis for predictors of mortality caused by bloodstream infections with Enterobacteriaceae. Clinical characteristics Bloodstream infections with Enterobacteriaceae, n (%) P value Survival group (n = 110) Mortality group (n = 38) Gender (male) 55 (50.0) 20 (52.6) 0.780 Age 67.5 ± 20.0 73.9 ± 15.9 0.075 Community acquired 28 (25.5) 2 (5.3) 0.008∗ Admission to the ICU 10 (9.1) 20 (52.6) <0.001∗ Long-term bedbound 50 (45.5) 30 (78.9) <0.001∗ Medical institution admission history within 3 months 69 (62.7) 23 (60.5) 0.809 Underlying diseases Cardiovascular disease 51 (46.4) 22 (57.9) 0.220 Cerebrovascular disease 54 (49.1) 24 (63.2) 0.134 Chronic respiratory disease 12 (10.9) 3 (7.9) 0.827 Chronic kidney disease 30 (27.3) 16 (42.1) 0.089 Diabetes mellitus 38 (34.5) 15 (39.5) 0.585 Malignant tumor 37 (33.6) 11 (28.9) 0.594 Anemia 77 (70.0) 36 (94.7) 0.002∗ Hypoalbuminemia 87 (79.1) 37 (97.4) 0.008∗ Lung infection 53 (48.2) 30 (78.9) 0.001∗ Urinary tract infection 60 (54.5) 14 (36.8) 0.060 Abdominal infection 9 (8.2) 10 (26.3) 0.009∗ Tracheotomy 22 (20.0) 11 (28.9) 0.253 Liver dysfunction 20 (18.2) 17 (44.7) 0.001∗ Invasive surgery before infection Mechanic ventilation 10 (9.1) 20 (52.6) <0.001∗ Central venous catheterization 20 (18.2) 27 (71.1) <0.001∗ Indwelling urethral catheterization 41 (37.3) 29 (76.3) <0.001∗ Sources of bloodstream infection Catheter related 19 (17.3) 14 (36.8) 0.012∗ Respiratory system 25 (22.7) 12 (31.6) 0.277 Urinary tract 39 (35.5) 3 (7.9) 0.001∗ Others 27 (24.5) 9 (23.7) 0.915 Prior use of broad-spectrum antibiotics within 30 days 36 (32.7) 28 (73.7) <0.001∗ Combined fungal infection 9 (8.2) 11 (28.9) <0.001∗ Prior use of glucocorticoids within 30 days 4 (3.6) 15 (39.5) 0.007∗ CRE infection 9 (8.2) 17 (44.7) <0.001∗ Total length of hospital stay 21.4 ± 13.2 22.7 ± 10.4 0.528 CRE: carbapenem-resistant Enterobacteriaceae; ∗P < 0.05.

within 30 days 36 (32.7) 28 (73.7) <0.001∗ Combined fungal infection 9 (8.2) 11 (28.9) <0.001∗ Prior use of glucocorticoids within 30 days 4 (3.6) 15 (39.5) 0.007∗ CRE infection 9 (8.2) 17 (44.7) <0.001∗ Total length of hospital stay 21.4 ± 13.2 22.7 ± 10.4 0.528 CRE: carbapenem-resistant Enterobacteriaceae; ∗P < 0.05. Table 4 Logistic regression analysis of risk factors for mortality in patients with Enterobacteriaceae bloodstream infections. Risk factor OR value 95% CI P value Lung infection 6.62 1.86 to 23.61 0.004 Abdominal infection 11.15 2.54 to 48.92 0.001 Central venous catheterization 5.25 1.84 to 14.94 0.002 Prior use of glucocorticoids within 30 days 22.40 4.07 to 123.38 <0.001 OR: odds ratio; 95% CI: 95% confidence interval.

1. Background Leishmania parasite cultivation in routine diagnostic and research laboratories is facing a major problem of microbial contamination despite strict adherence on aseptic microbiological practices. In general, the common antibiotics like penicillin and streptomycin are frequently used to prevent the bacterial contamination, but in particular these drugs have also limited a spectrum of inhibition to rule out the possibility of contamination. In contrary, fungal contamination during the parasite cultivation is another challenging arm in most of the laboratories. Yeast and yeast-like organisms are always reported as a major source of contamination during parasite culture. In fact, the antifungal compounds are rarely used to prevent fungal contamination in parasite culture. Recently, the report of cell banks showed that 39% of specimens are contaminated over a 2-year period and fungi were identified in 8% of these [1]. Polyene macrolide drugs (amphotericin and nystatin) have been used to reduce fungal contamination in cell line cultures [2, 3], and they target ergosterol in the fungal cell membrane. Lipid composition in Leishmania parasite cell membrane is similar that in fungi and yeast [4]. These drugs have also been shown to interact with Toll-like receptors and CD14 to induce signal transduction and release of inflammatory cytokines [5]. Hence, these drugs significantly affect parasite's cell physiology that can disrupt parasite cell cycle. Knowing the fact of antifungal antibiotics that are also potential inhibitors to the parasite growth itself, the exploration of antifungal compound which is not inhibiting the parasite growth could prevent the fungal contamination.

gs significantly affect parasite's cell physiology that can disrupt parasite cell cycle. Knowing the fact of antifungal antibiotics that are also potential inhibitors to the parasite growth itself, the exploration of antifungal compound which is not inhibiting the parasite growth could prevent the fungal contamination. Caspofungin, an echinocandin drug, is used clinically for the invasive fungal disease. It kills the fungi by inhibiting 1-3-β-D-glucan synthesis in fungal cell wall [6]. It has been reported that caspofungin can effectively inhibit several contaminating clinical isolates of Apergillus spp., Penicillium, etc. [7]. The mammalian cell lines have found high resistance towards caspofungin [8]. The cells of Leishmania parasite are more similar to the mammalian cells. Therefore, the current study has an aim to determine the effect of caspofungin on Leishmania parasite as a proof of principle for their application in the culture of Leishmania parasite in order to avoid fungal contamination.

rds caspofungin [8]. The cells of Leishmania parasite are more similar to the mammalian cells. Therefore, the current study has an aim to determine the effect of caspofungin on Leishmania parasite as a proof of principle for their application in the culture of Leishmania parasite in order to avoid fungal contamination. 2. Materials and Methods 2.1. Parasite Promastigote Cultivation The clinical isolates of L. donovani were isolated from the bone marrow of visceral leishmaniasis patients using the Tobie's agar with Locke's overlay. Later on, the culture was transferred to M199 with 20% fetal calf serum and cryopreserved at −80°C with 10% glycerol. The parasite was taken out of cryobank and M199 (Sigma-Aldrich, Cat. No. 2520), supplemented with 20% heat-inactivated fetal calf serum (Invitrogen), and was used to maintain promastigote parasite. 5 × 105 parasite per/ml was inoculated in final 5 ml medium and incubated at 26°C [9]. The growth curve of the parasite was maintained for a week to determine the parasite growth rate. The Leishmania species isolates were determined by using the PCR assays targeting L. donovani(donovani) specific ribosomal DNA and followed by the PCR-RFLP heat shock protein 70. 2.2. Fungi Cultivation

2. Materials and Methods 2.1. Parasite Promastigote Cultivation The clinical isolates of L. donovani were isolated from the bone marrow of visceral leishmaniasis patients using the Tobie's agar with Locke's overlay. Later on, the culture was transferred to M199 with 20% fetal calf serum and cryopreserved at −80°C with 10% glycerol. The parasite was taken out of cryobank and M199 (Sigma-Aldrich, Cat. No. 2520), supplemented with 20% heat-inactivated fetal calf serum (Invitrogen), and was used to maintain promastigote parasite. 5 × 105 parasite per/ml was inoculated in final 5 ml medium and incubated at 26°C [9]. The growth curve of the parasite was maintained for a week to determine the parasite growth rate. The Leishmania species isolates were determined by using the PCR assays targeting L. donovani(donovani) specific ribosomal DNA and followed by the PCR-RFLP heat shock protein 70. 2.2. Fungi Cultivation Candida spp. was isolated from the clinical specimen at BPKIHS by Sabouraud dextrose agar (SDA). Candida species was confirmed by using germ tube test, sugar fermentation, growth at 45°C on SDA broth, and assimilation tests with xylose and α-methyl-D-glucoside. It was transferred to M199 with 20% heat-inactivated fetal calf serum (invitrogen) with 5 × 105 parasite per/ml and then incubated at 26°C [10]. The growth rate of Candida was determined by maintaining the growth curve for a week.

growth at 45°C on SDA broth, and assimilation tests with xylose and α-methyl-D-glucoside. It was transferred to M199 with 20% heat-inactivated fetal calf serum (invitrogen) with 5 × 105 parasite per/ml and then incubated at 26°C [10]. The growth rate of Candida was determined by maintaining the growth curve for a week. 2.3. Preparation of Drug 50 mg of caspofungin (Merck, USA) was dissolved in 5 ml sterile distilled water (10 mg/ml) and mixed thoroughly for 5 minutes. The drug solution was sterilized through a filter of pore size 0.2 µm (PVDF), and drug solution was preserved at −80°C. 2.4. Drug Inhibition Measurement 2.4.1. Leishmania Parasite Log-phase promastigote parasites from third day of subculture were centrifuged at 3200 rpm for 10 minutes; parasites were counted and diluted to make 106 parasite/ml. 100 µl of parasites was plated out in duplicate on sterile 96-well plates (Corning Costar, USA, Cat. No. CLS3596). Each well was topped with 100 µl of seven different concentrations of caspofungin ranging from 512 µg/ml to 8 µg/ml, including one negative control. The 96-well plates were covered with a sealing tape and incubated at 26°C for 4 days. Finally, viable parasites after exposure to caspofungin were determined by the trypan blue assay.

s topped with 100 µl of seven different concentrations of caspofungin ranging from 512 µg/ml to 8 µg/ml, including one negative control. The 96-well plates were covered with a sealing tape and incubated at 26°C for 4 days. Finally, viable parasites after exposure to caspofungin were determined by the trypan blue assay. 2.4.2. Candida spp Day 2 subculture of Candida was centrifuged at 3200 rpm for 10 minutes and resuspended to final cell density 106 cells/ml. 100 µl of Candida was transferred to duplicate wells of 96-well plates. Each well was topped with 100 µl of seven different concentrations of caspofungin ranging from 8 µg/ml to 0.00051 µg/ml. The 96-well plates were covered with a sealing tape and incubated at 26°C for 3 days. Then, viable fungal cells after exposure to caspofungin were determined by the trypan blue assay. 2.5. Trypan Blue Viable Count Assay The viable cells were counted in each well using the trypan blue (0.2%) dye exclusion method, where viable parasite became colourless and dead cells appear blue [11]. 2.6. Calculation of IC50 Viable count of each well was entered in Microsoft Excel 2007, and log transformation was made for each drug concentration. These values were transferred in a GraphPad Prism version 5, and IC50 values were analysed for each strain using sigmoidal dose-response model (nonlinear regression) for caspofungin assay [12].

2.6. Calculation of IC50 Viable count of each well was entered in Microsoft Excel 2007, and log transformation was made for each drug concentration. These values were transferred in a GraphPad Prism version 5, and IC50 values were analysed for each strain using sigmoidal dose-response model (nonlinear regression) for caspofungin assay [12]. 3. Statistical Analysis All data were subjected to statistical analysis using Graphpad Prism Version 5. P values were calculated by Student's t-test or analysis of variance depending on the data. P values of less than 0.05 were considered as significant. 4. Results The two Leishmania strains BPK206/0 cl10 and BPK632/0 were tested to determine IC50 against caspofungin, and detailed results were presented in Table 1. The IC50 values of Leishmania strains ranged from 23.02 to 155.80 µg/ml (mean = 90.25 ± 39.01), as shown in Table 1. The parasite growth curve was also maintained with different concentrations of caspofungin (10 µg/ml, 25 µg/ml, and 50 µg/ml) along with control. Caspofungin concentration more than 25 µg/ml significantly inhibits the parasite growth than others, as shown in Figure 1. IC50 values of Candida spp. were ranged from 0.001 to 0.12 µg/ml (Table 1). The IC50 of Candida is significantly lower than Leishmania strains (P value = 0.02).

4. Results The two Leishmania strains BPK206/0 cl10 and BPK632/0 were tested to determine IC50 against caspofungin, and detailed results were presented in Table 1. The IC50 values of Leishmania strains ranged from 23.02 to 155.80 µg/ml (mean = 90.25 ± 39.01), as shown in Table 1. The parasite growth curve was also maintained with different concentrations of caspofungin (10 µg/ml, 25 µg/ml, and 50 µg/ml) along with control. Caspofungin concentration more than 25 µg/ml significantly inhibits the parasite growth than others, as shown in Figure 1. IC50 values of Candida spp. were ranged from 0.001 to 0.12 µg/ml (Table 1). The IC50 of Candida is significantly lower than Leishmania strains (P value = 0.02). 5. Discussion In vitro cultivation of slow-growing pathogenic protozoan parasites is a difficult process as compared to bacterial and fungal culture. In fact, the longer generation time and requirement of complex nutrients are the main reasons for slow growth in parasite culture. Hence, it requires to maintain an extensive aseptic procedure during the primary isolation and culture maintenance in order to avoid contamination. However, abundant numbers of viable organisms are essentially required not only for confirmatory diagnosis but also for several downstream applications and drug susceptibility assay. Among others, L. donovani was most commonly cultivated in modified Tobie's blood agar at different tropics, but it was most commonly observed with yeast-like contamination during the isolation and maintenance of the parasite. The contamination in parasite culture could be due to the humid environmental condition to support rapid growth of yeast in the different endemic zones. Moreover, the lack of well-equipped laboratory facility to process the clinical specimen for in vitro culture would be another factor for the contamination. Therefore, there is requirement of certain antifungal agents which have no inhibitory effect to the L. donovani culture and also protect from the fungal contamination. That is why we undertook to assess the inhibitory effect of antifungal drugs in L. donovani during the parasite cultivation.

ctor for the contamination. Therefore, there is requirement of certain antifungal agents which have no inhibitory effect to the L. donovani culture and also protect from the fungal contamination. That is why we undertook to assess the inhibitory effect of antifungal drugs in L. donovani during the parasite cultivation. Our experimental results showed that L. donovani had significantly higher IC50 to caspofungin than Candida spp. which indicates that the Leishmania parasite had higher tolerance properties to caspofungin. Both Leishmania and Candida cells are eukaryotic in nature, but the Candida cell contains the cell wall but Leishmania lacks the cell wall. In Candida, the enzyme 1-3-β-D-glucan synthase in cell membrane has essential roles in cell wall formation. Since the caspofungin has a potential power to inhibit the function of 1-3-β-D-glucan synthase, it inhibits the growth of Candida. Therefore, the least growth inhibition was found in L. donovani isolates due to the lack of 1-3-β-D-glucan synthase. In addition, the growth curve of Leishmania culture isolates was also assessed in the presence of three different caspofungin concentrations (10 µg/ml, 25 µg/ml, and 50 µg/ml). It showed that 50 µg/ml caspofungin could inhibit a few strains of Leishmania, and no inhibition was determined in less than 50 µg/ml of caspofungin. In contrast, Candida growth was inhibited with lesser concentration of caspofungin, and no growth was observed in culture incorporated with 10 µg/ml caspofungin. Moreover, Leishmania parasite growth can adapt in five times more concentration of caspofungin than the Candida spp. We found that it could be advantageous to Leishmania culture procedure in order to prevent fungal contamination since Leishmania parasite growth can tolerate 25 µg/ml concentration of caspofungin. Hence, 25 µg/ml caspofungin concentration could be useful to overcome the fungal contamination in Leishmania culture although further study with large sample is required.

mania culture procedure in order to prevent fungal contamination since Leishmania parasite growth can tolerate 25 µg/ml concentration of caspofungin. Hence, 25 µg/ml caspofungin concentration could be useful to overcome the fungal contamination in Leishmania culture although further study with large sample is required. Leishmania parasites have digenetic life stage, and they are transforming from nonmotile amastigote to motile promastigote stage during in vitro isolation of the parasite from clinical specimens [9]. Caspofungin must not inhibit the parasite promastigote transformation from the beginning of isolation attempts in order to use against the fungal contamination since the Leishmania promastigote transformation might be inhibited by a few antibiotics [13]. The efficacy on the development of Leishmania promastigote in the presence of caspofungin was also examined. But, there was no significant reduction in the development of promastigote in parasite culture from clinical specimen (data not shown). Furthermore, Leishmania cell morphology in viable cultures was examined by phase-contrast microscopy, and promastigote cell morphology was not significantly different with/without caspofungin. This evidence indicates that the Leishmania parasite promastigote development was not negatively affected by the caspofungin.

. Furthermore, Leishmania cell morphology in viable cultures was examined by phase-contrast microscopy, and promastigote cell morphology was not significantly different with/without caspofungin. This evidence indicates that the Leishmania parasite promastigote development was not negatively affected by the caspofungin. However, there could be a question of efficacy of the drug towards other fungal contaminations rather than Candida. Undoubtedly, it could also inhibit the growth of all fungi since the key enzyme 1-3-β-D-glucan synthase is essential for the cell wall formation in all fungus and yeast. The fungal growth inhibition of Apergillus spp. and Penicillium spp. [7] was already reported in other than Candida spp. In conclusion, in this study, we show that caspofungin could be the potential agent for the selective growth of L. donovani culture isolation since it has no inhibitory effect towards the concentration in which the fungal cell viability is terminated. Acknowledgments The authors would like to acknowledge Mr. Ganesh Prasad Sah and Mrs. Iccha Ghale for their technical assistance at Kala-Azar Lab, B. P. Koirala Institute of Health Sciences. Conflicts of Interest The authors declare that they have no conflicts of interest. Figure 1 Growth curve of Leishmania parasite with different concentrations of caspofungin. Error bar indicates the 95% confidence interval of the mean based on triplicate culture of different strains. (a) BPK206/0 cl10 growth curve. (b) BPK632/0 growth curve. Table 1 IC50 of strains tested with caspofungin.

Conflicts of Interest The authors declare that they have no conflicts of interest. Figure 1 Growth curve of Leishmania parasite with different concentrations of caspofungin. Error bar indicates the 95% confidence interval of the mean based on triplicate culture of different strains. (a) BPK206/0 cl10 growth curve. (b) BPK632/0 growth curve. Table 1 IC50 of strains tested with caspofungin. Strain IC50 µg/ml (mean ± SD) Range (µg/ml) BPK206/0 cl10 61.70 ± 35.97 23.02 to 121.90 BPK632/0 118.8 ± 3.25 46.87 to 155.80 Candida spp. 0.05 ± 0.07 0.001 to 0.12

1. Introduction Acute gastrointestinal illness (AGI) is an important public health issue, with substantial economic and human health impact [1–4]. Numerous countries in Europe, North and South America, Asia, and Australia and New Zealand have conducted population-based studies to estimate the incidence and burden of AGI in the community [5–21]. In Canada, the National Studies on Acute Gastrointestinal Illness (NSAGI) initiative was initiated in 1999 and developed population estimates in select regions of Canada [22–24]. The NSAGI studies have been used to inform Canadian burden of enteric illness studies [25–27]. In 2014-15, a population-based telephone survey, the Foodbook study, was conducted nationally to describe Canadians' exposure to foods, animals, and water that may serve as sources for enteric illness pathogens and included questions related to AGI symptoms and care seeking behaviours [28]. The objective of this paper is to describe the prevalence, distribution, and symptoms of AGI and the health care seeking behaviours of individuals with AGI, across Canada for 2014-15 based on the Foodbook study.

r enteric illness pathogens and included questions related to AGI symptoms and care seeking behaviours [28]. The objective of this paper is to describe the prevalence, distribution, and symptoms of AGI and the health care seeking behaviours of individuals with AGI, across Canada for 2014-15 based on the Foodbook study. 2. Materials and Methods 2.1. Study Design and Data Collection The Foodbook study was conducted in Canada's 10 provinces and 3 territories over a one-year period (April 2014-April 2015) using a population-based telephone survey and included questions on AGI symptoms and related health care seeking behaviours. Households were randomly selected from a sampling frame of telephone numbers that consisted of land lines (70% listed and 10% random digit-dialing) and cell phones (20%). The survey was designed with a target total sample size of 11,016 surveys collected evenly over a 12-month period across four age groups (0–9, 10–19, 20–64, and 65+ years) and all 13 Canadian provinces and territories. In order to improve completion rates for younger age groups, when households contained children less than 18 years old, 50% of surveys were conducted with the child who would have the next birthday and 50% were conducted with the adult who would have the next birthday. If there were no children in the household, the survey was conducted with the adult who would have the next birthday. Complete details on participant selection and questionnaire administration can be found in the Foodbook Report [28]. The surveys were conducted by an independent research company contracted by the Public Health Agency of Canada (the Agency). Individuals were excluded if they could not speak the supported languages (English, French, Inuktitut, and on-demand verbal translation for other languages), if they did not have a listed land line or cellular telephone number or travelled outside their province or territory of residence during the seven days prior to interview. The Foodbook study was reviewed and approved by Health Canada and the Agency's Research Ethics Board (REB 2013-0025) as well as the Newfoundland and Labrador Health Research Ethics Authority to meet a unique provincial legal requirement (HREB 13.238).

or territory of residence during the seven days prior to interview. The Foodbook study was reviewed and approved by Health Canada and the Agency's Research Ethics Board (REB 2013-0025) as well as the Newfoundland and Labrador Health Research Ethics Authority to meet a unique provincial legal requirement (HREB 13.238). Weighted selection of survey participants to reflect the Canadian population was assigned using the following method. The forward sortation area (FSA, or first three digits of the postal code) collected for each respondent was converted to the most likely census metropolitan area (CMA). Using 2011 Census data, the CMA indicator along with age group, household type, province or territory, the number of people in the household, the number of land lines and cell phones in the household, and gender were used to calculate the individual-level survey expansion weight. To create the final weighting variable, a poststratification step used iterative proportional ranking with available control tables. The population reference year was 2011, representing a population of 33,400,000 [29].

nes in the household, and gender were used to calculate the individual-level survey expansion weight. To create the final weighting variable, a poststratification step used iterative proportional ranking with available control tables. The population reference year was 2011, representing a population of 33,400,000 [29]. The survey questions relating to AGI were developed to be consistent with the NSAGI studies previously conducted in Canada [22–24]. Respondents were asked if they had experienced any vomiting or diarrhea in the 28 days prior to the interview. The module included questions about symptoms, their frequency and duration, existing chronic conditions and medication use, respiratory symptoms, and care seeking and stool submission behaviours. The survey was pilot-tested over a two-week period before full survey implementation. 2.2. Case Definitions Illnesses reported to have started prior to the 28 days of the interview were excluded (3.6% weighted). Respondents who identified more than one episode of AGI during the 28 days prior to the interview were asked to respond only for their most recent episode. Respondents who did not report symptoms of AGI, as well as those identified as having self-reported that their diarrhea or vomiting was due to pregnancy, medical treatment (e.g., chemotherapy), or medical conditions (e.g., Crohn's disease, colitis, irritable bowel syndrome, and alcoholism), were included in the noncase category.

s who did not report symptoms of AGI, as well as those identified as having self-reported that their diarrhea or vomiting was due to pregnancy, medical treatment (e.g., chemotherapy), or medical conditions (e.g., Crohn's disease, colitis, irritable bowel syndrome, and alcoholism), were included in the noncase category. Two case definitions of AGI were assessed: (a) a person reporting three or more loose stools in 24 hours or any vomiting in the past 28 days, according to the international AGI definition [24] and (b) a person reporting three or more loose stools in 24 hours or any vomiting in the past 28 days without concurrent respiratory symptoms (cough or sore throat) [25]. Removing cases with respiratory symptoms creates a more specific definition attempting to exclude respiratory infections that may cause gastrointestinal symptoms such as vomiting or diarrhea [30]. 2.3. Missing Data and Resolution Starting in November 2014 to the end of the survey time period (April 2015), due to an error in the Computer Assisted Telephone Interviewing (CATI) survey tool, 225 respondents who indicated AGI symptoms were not asked all the relevant survey questions. These participants responded “yes” to having any symptoms of vomiting or diarrhea but “no” to vomiting symptoms specifically; therefore, it is assumed that they only had diarrhea symptoms. Questions missed pertained to the subsection specific to diarrhea and included duration of diarrhea symptoms, number of stools, and related chronic conditions or medication use.

ny symptoms of vomiting or diarrhea but “no” to vomiting symptoms specifically; therefore, it is assumed that they only had diarrhea symptoms. Questions missed pertained to the subsection specific to diarrhea and included duration of diarrhea symptoms, number of stools, and related chronic conditions or medication use. An adjustment for missing data was made by weighting completed interviews for respondents that reported symptoms of diarrhea only (April 2014 to October 2014; weighted n = 988,114.51) to account for the missing data (November 2014 to April 2015; weighted n = 855,500.56) by province (and respiratory symptoms for the more specific case definition), therefore increasing the weight assigned to the completed interviews (Appendix A). An assumption was made that there would be similar chronic disease and duration of symptoms between the diarrhea-only cases from April to October and November to April. 3. Analysis Data analysis was performed in Stata 13.0 (StataCorp., Texas Station, TX) using the survey weight, and only weighted results are reported. Categorical variables were described using weighted percentages and the relative 95% confidence interval (CI). Individuals responding “don't know” or “refused” to any question were excluded from the analysis of that question. Mean and median values were used to describe continuous variables.

ghted results are reported. Categorical variables were described using weighted percentages and the relative 95% confidence interval (CI). Individuals responding “don't know” or “refused” to any question were excluded from the analysis of that question. Mean and median values were used to describe continuous variables. For incidence rate calculations, respondents identifying multiple episodes were counted as a single episode. The primary outcome measure of monthly prevalence was defined as the number of respondents reporting AGI in the previous 28 days divided by the total number of respondents. Annual incidence rate and incidence proportion calculations were performed using formulas found in Appendix B [31]. The null hypothesis of no overall association between the prevalence of disease and province was tested using the Wald χ2 test, with a p value cutoff of 0.05. The difference between the proportion of cases (i.e., the prevalence of AGI) in a specific province and the proportion of cases in all other provinces combined was tested using the χ2 test.

erall association between the prevalence of disease and province was tested using the Wald χ2 test, with a p value cutoff of 0.05. The difference between the proportion of cases (i.e., the prevalence of AGI) in a specific province and the proportion of cases in all other provinces combined was tested using the χ2 test. 4. Results The survey response rate was 19.9%, and a total of 10,798 residents responded to the survey (Table 1). There were 975 respondents (weighted n = 2,803,946) that indicated symptoms of diarrhea or vomiting in the past 28 days, reflecting a monthly prevalence of 8.5%. Of these respondents, 12% (weighted) reported that their diarrhea or vomiting in the past 4 weeks was caused by a medical condition, medication, or pregnancy and were counted in the noncase group. Using the international AGI case definition of three or more loose stools in 24 hours and any vomiting [24], the overall monthly prevalence was 5.7% (95% CI 4.6–7.2, weighted n = 1,887,588) corresponding to an annual incidence rate of 0.77 episodes/person-year (95% CI 0.61–0.97) and 26.3 million episodes of AGI per year in Canada (Table 2). After removal of cases with concurrent respiratory symptoms (25%), the monthly prevalence was 4.3% (95% CI 3.1–5.8, weighted n = 1,407,698), with an annual incidence rate of 0.57 episodes/person-year (95% CI 0.41–0.78). This reflects 19.4 million episodes of AGI per year in Canada.

of AGI per year in Canada (Table 2). After removal of cases with concurrent respiratory symptoms (25%), the monthly prevalence was 4.3% (95% CI 3.1–5.8, weighted n = 1,407,698), with an annual incidence rate of 0.57 episodes/person-year (95% CI 0.41–0.78). This reflects 19.4 million episodes of AGI per year in Canada. Estimates of monthly prevalence and incidence of AGI nationally and by province for both AGI case definitions are presented in Table 2. There were some regional differences identified: the monthly prevalence for the province of Quebec was significantly lower compared to the rest of the provinces/territories combined (p < 0.01) when assessing the international AGI case definition, and the monthly prevalence for the province of Saskatchewan was significantly lower compared to the rest of the provinces/territories combined (p=0.02) when assessing the more specific case definition of three or more loose stools in 24 hours and any vomiting without concurrent respiratory symptoms. A higher monthly prevalence was identified in Manitoba, Newfoundland, and the Territories; however, these differences were not statistically significant. The Territories and New Brunswick showed the greatest difference between prevalence of AGI when comparing AGI case definitions with and without concurrent respiratory symptoms (Figure 1).

alence was identified in Manitoba, Newfoundland, and the Territories; however, these differences were not statistically significant. The Territories and New Brunswick showed the greatest difference between prevalence of AGI when comparing AGI case definitions with and without concurrent respiratory symptoms (Figure 1). When considering predisposing factors, 5.6% of AGI cases took prescription antibiotics in the previous 28 days and 5.0% of AGI cases with no respiratory symptoms in the previous 28 days. There was no clear seasonal pattern: lower monthly prevalence in February, June, and October and higher monthly prevalence in December and April (Figure 2). When assessing the most specific AGI case definition, 61.1% of respondents reported experiencing diarrhea symptoms only, while 24.3% reported both vomiting and diarrhea and 14.6% reported vomiting only (Table 3). Of the cases who experienced diarrhea, 9.9% reported bloody diarrhea (95% CI 3.4–25.8). Duration of symptoms was longest for those who experienced both vomiting and diarrhea compared to those experiencing only one symptom (Table 4). Cases of the more specific AGI definition reported a mean of 4.43 episodes of diarrhea and 3.97 episodes of vomiting in a 24-hour period.

ed bloody diarrhea (95% CI 3.4–25.8). Duration of symptoms was longest for those who experienced both vomiting and diarrhea compared to those experiencing only one symptom (Table 4). Cases of the more specific AGI definition reported a mean of 4.43 episodes of diarrhea and 3.97 episodes of vomiting in a 24-hour period. When using the more specific AGI case definition to assess care seeking behaviour, overall, 8.8% (95% CI 4.9–15.1) of cases visited a physician (Table 5). Of these, 17.1% (95% CI 7.5–34.5) were requested to submit a stool sample by a physician, and 49.0% (95% CI 17.6–81.2) of these submitted a stool sample. Hospitalizations were reported by 0.68% of cases with a mean hospital stay of 3.38 days (median 2). Of all cases, 0.11% reported taking antibiotics to treat their illness. 5. Discussion This is the first nationwide survey conducted in Canada to describe the magnitude and distribution of AGI in the general population. Based on the more specific definition of AGI, excluding cases with respiratory symptoms, it is estimated that there are 0.57 (95% CI 0.41–0.78) self-reported AGI episodes per person-year or almost 19.5 million episodes of AGI in Canada each year. This estimate is lower than the rate of 0.63 (95% CI 0.57–0.69) episodes per person-year that was estimated based on the combined previous NSAGI studies and used in the Canadian estimates of foodborne illnesses [25]; however, the 95% confidence intervals of the current and previous estimates overlap indicating a lack of statistical difference.

an the rate of 0.63 (95% CI 0.57–0.69) episodes per person-year that was estimated based on the combined previous NSAGI studies and used in the Canadian estimates of foodborne illnesses [25]; however, the 95% confidence intervals of the current and previous estimates overlap indicating a lack of statistical difference. This lower estimated incidence in the current study year (2014/2015) compared to the previous NSAGI rates (2002–2006) may be related to the different approach in survey design, specifically the use of a weighted sampling technique. Other differences include the exclusion of respondents who travelled outside their province or territory of residence during the seven days prior to the interview in the current study that may have been experiencing symptoms. As well, a true lower incidence may be explained by epidemiological trends including variability in norovirus trends from year to year, the impact of rotavirus vaccine on illness associated with rotavirus [32, 33], or possibly other public health interventions.

he current study that may have been experiencing symptoms. As well, a true lower incidence may be explained by epidemiological trends including variability in norovirus trends from year to year, the impact of rotavirus vaccine on illness associated with rotavirus [32, 33], or possibly other public health interventions. The annual rate of AGI when using the specific definition and excluding respiratory symptoms is comparable to estimates from the United States (US) (individual population studies 0.49, 0.54, and 0.73 and overall 0.60 episodes per person-year) [34] and lower than Italy (0.76) [14]. When comparing the international AGI case definition, the Canadian annual incidence rate (0.77 episodes per person-year) is lower than Germany [12], Denmark [35], Italy [14], Chile [6], Australia, and the US [24], which ranged from 0.83 to 1.4 episodes per person-year, but is higher than Ireland (0.64) and Malta (0.37) [24]. The proportion of respondents excluded due to chronic conditions, medication use, or pregnancy as the cause of their symptoms in the present study (12%) was lower than that in the previous NSAGI studies (16–19%) [22, 23, 36].

0.83 to 1.4 episodes per person-year, but is higher than Ireland (0.64) and Malta (0.37) [24]. The proportion of respondents excluded due to chronic conditions, medication use, or pregnancy as the cause of their symptoms in the present study (12%) was lower than that in the previous NSAGI studies (16–19%) [22, 23, 36]. Comparison of provincial/territorial results for Ontario, Quebec, and Nunavut using the international AGI definition showed that the estimates were lower than previous provincial/territorial illness estimates using the same definition [36–38]. The large variation in incidence between provinces/territories, though not statistically significant for many of them, does speak to the apparent regional differences of AGI incidence in Canada and the importance of capturing national information that reflects all provinces and territories. Furthermore, having provincial and territorial specific estimates enables individual jurisdictions to assess their AGI burden more specifically. This could be used to generate regional estimates of enteric illness and specific transmission routes (e.g., foodborne illness estimates for a specific province or territory) that could be used to inform public health activities (e.g., education and prevention campaigns).

s to assess their AGI burden more specifically. This could be used to generate regional estimates of enteric illness and specific transmission routes (e.g., foodborne illness estimates for a specific province or territory) that could be used to inform public health activities (e.g., education and prevention campaigns). The proportion of cases with respiratory symptoms (25%) is at the low end of the range reported by other countries reporting between 19% and 48% of cases experiencing concurrent respiratory symptoms [13, 14, 24, 35]. Using the more specific case definition creates a more conservative estimate, attempting to account for cases whose AGI symptoms may be caused by respiratory infections [14, 30, 34]. Diarrhea only was the predominant symptom profile of cases (55.9% and 61.1%, resp., for the two case definitions); this result falls between other studies, reporting a higher proportion in Germany (78%) and Denmark (64%) and lower proportions in Sweden, Italy, and Chile (30–40%) [6, 12, 14, 35, 39]. The proportion of diarrheal cases with bloody diarrhea (9.9%) was higher than that in other countries (3-4%) [5, 12, 13, 35]. This may be due to the small number of cases reporting bloody diarrhea and a large assigned weight due to study design.

ower proportions in Sweden, Italy, and Chile (30–40%) [6, 12, 14, 35, 39]. The proportion of diarrheal cases with bloody diarrhea (9.9%) was higher than that in other countries (3-4%) [5, 12, 13, 35]. This may be due to the small number of cases reporting bloody diarrhea and a large assigned weight due to study design. The proportion of persons with AGI varied somewhat by season, which is similar to higher rates of AGI in winter months as what has been reported by previous Canadian studies [22, 40] as well as internationally in the US, Denmark, Italy, Sweden, and Germany [5, 12, 14, 35, 39]. This pattern is likely driven by viruses circulating in the winter months, particularly norovirus which is the most common cause of AGI in Canada [41]. The higher monthly prevalence observed in April may be an artefact due to the lower number of survey respondents in April (4% of the survey) compared to other months (approximately 8% each).

driven by viruses circulating in the winter months, particularly norovirus which is the most common cause of AGI in Canada [41]. The higher monthly prevalence observed in April may be an artefact due to the lower number of survey respondents in April (4% of the survey) compared to other months (approximately 8% each). The percentage of cases who reported seeing a physician was low with only 8.8% seeking care, and 17.1% of these were requested to submit a stool sample. These values are weighted and are lower than those in the previous NSAGI studies (11–23% and 26–54%) [22, 23, 36]. The previous NSAGI studies were not age adjusted; therefore, the more frequent care seeking among the elderly may contribute to the higher overall results in the previous studies. The exclusion of individuals who travelled in the past seven days in the Foodbook study may also influence the lower results. Recent travel is associated with seeking medical care and having a sample requested [42, 43]; thus, these individuals who travelled may have been more likely to seek care and be requested to submit a stool sample. Lower care seeking rates would influence pathogen-specific estimates as it indicates greater underdiagnosis of cases. This should be considered among future burden of illness activities and how survey respondent weighting may influence this phenomenon.

more likely to seek care and be requested to submit a stool sample. Lower care seeking rates would influence pathogen-specific estimates as it indicates greater underdiagnosis of cases. This should be considered among future burden of illness activities and how survey respondent weighting may influence this phenomenon. Possible limitations of this study include the retrospective study design as it may be subject to recall bias. Retrospective studies in the UK (IID2) gave higher estimates of disease burden than prospective studies [19]. However, retrospective studies with longer recall periods gave lower estimated rates than studies with shorter recall periods [44]. Extrapolation from a reported seven-day prevalence was almost twice the rate of illness estimated when extrapolating from the month recall period [6, 35, 45]. The study response rate of 20% is lower than that in previous NSAGI studies [22, 40] and may be a source of bias if those who did not respond had different symptom profiles compared to those who participated in the study. Furthermore, misclassification of cases may have occurred due to excluding cases with chronic conditions or respiratory symptoms that might have been true infectious AGI cases.

2, 40] and may be a source of bias if those who did not respond had different symptom profiles compared to those who participated in the study. Furthermore, misclassification of cases may have occurred due to excluding cases with chronic conditions or respiratory symptoms that might have been true infectious AGI cases. The missing data for diarrheal cases from November to April due to the survey interview error were adjusted for based on known diarrheal cases captured from April to October; this however may not have accurately reflected the symptoms and behaviours of the missing cases and may have impacted the results. An assumption was made that there would be similar patterns (e.g., chronic disease and medical causes of symptoms, duration of symptoms, and care seeking behaviours) between the diarrhea-only cases from April to October and November to April. However, there may have been seasonal differences due to different pathogens circulating (e.g., norovirus in the winter or bacterial pathogens in the summer) or behavioural patterns (e.g., international winter travel or domestic summer recreational water exposure). From previous NSAGI studies, the monthly prevalence of AGI fluctuates seasonally with peaks seen in winter/early spring and again in summer [22–24]. Symptom-specific monthly variations were observed in Ontario where diarrhea only was the predominant symptom for most months, followed by both vomiting and diarrhea combined; however, the statistical significance of these variations was not reported, and the general relationship between symptom profiles does not vary much across seasons [23]. The adjustment was made based on province/territory only as there were insufficient data to allow for age-gender-province/territory-based adjustments. Differences due to age and gender would be inherently incorporated into the province-based adjustment. However, age- and gender-specific results could not be described due to this adjustment approach.

n province/territory only as there were insufficient data to allow for age-gender-province/territory-based adjustments. Differences due to age and gender would be inherently incorporated into the province-based adjustment. However, age- and gender-specific results could not be described due to this adjustment approach. Generating an estimate of the total amount of AGI in Canada provides the foundation for pathogen- and transmission route-specific burden of illness estimates. The lower incidence of AGI reported here will inform future activities to refine estimates of food and waterborne illness in Canada. Additionally, provincial- and territorial-specific estimates will enable individual jurisdictions to assess their AGI burden and generate region-specific public health plans that could include, for example, focused education campaigns, public health policies, or resource allocations toward prevention of AGI strategies. Acknowledgments The authors thank the Centre for Foodborne, Environmental and Zoonotic Infectious Diseases (Public Health Agency of Canada), Outbreak Management Division and Enteric Surveillance and Population Studies Division; the Foodbook team; Jennifer Cutler and Matt Hurst of the Foodbook team for their advice on adjustment for missing data; the staff at R.A. Malatest & Associates Ltd. for their expert interviewing; and the survey respondents for their participation. The authors thank the Public Health Agency of Canada for their funding support.

dbook team; Jennifer Cutler and Matt Hurst of the Foodbook team for their advice on adjustment for missing data; the staff at R.A. Malatest & Associates Ltd. for their expert interviewing; and the survey respondents for their participation. The authors thank the Public Health Agency of Canada for their funding support. Appendices A. Adjustment for Missing Data A.1. Missing Data Starting in November 2014 to the end of the survey time period (April 2015), due to an error in the CATI survey tool, 225 respondents (representing 30.5% of the respondents with any symptoms in the past 28 days) who indicated AGI symptoms were not asked all the relevant survey questions. These participants responded “yes” to having any symptoms of vomiting or diarrhea but “no” to vomiting symptoms specifically; therefore, it is assumed that they only had diarrhea symptoms. Questions missed pertained to the subsection specific to diarrhea and included duration of diarrhea symptoms, number of stools, and related chronic conditions or medication use. Information on respiratory symptoms was captured for all cases.

s specifically; therefore, it is assumed that they only had diarrhea symptoms. Questions missed pertained to the subsection specific to diarrhea and included duration of diarrhea symptoms, number of stools, and related chronic conditions or medication use. Information on respiratory symptoms was captured for all cases. A.2. Approach to Adjustment The responses from diarrhea-only cases from April to October (i.e., respondents where all questions were correctly asked) were adjusted to account for the diarrhea-only cases from November to April that had missing variables, so they could not be assessed if they met the case definition. Complete responses were given additional weight to account for those that were incomplete. The weighted value of the respondents with incomplete interviews (n = 855,500.56) was assigned to the weighted responses for the respondents with complete interviews (n = 988,114.51).

be assessed if they met the case definition. Complete responses were given additional weight to account for those that were incomplete. The weighted value of the respondents with incomplete interviews (n = 855,500.56) was assigned to the weighted responses for the respondents with complete interviews (n = 988,114.51). To estimate the more specific case definition of AGI without respiratory symptoms, a different adjustment was made as information on the respiratory symptoms was available for all respondents. The weighted value of the respondents with AGI and no respiratory symptoms but with incomplete interviews (n = 661,095.32) was assigned to the weighted responses for the respondents with AGI and no respiratory symptoms and complete interviews (n = 702,464.68). As this more specific definition necessitated cases not have concurrent respiratory symptoms and that information was available from all respondents, noncases could be identified directly, and thus, their true weight was incorporated into the noncase definition without need for adjustment.

interviews (n = 702,464.68). As this more specific definition necessitated cases not have concurrent respiratory symptoms and that information was available from all respondents, noncases could be identified directly, and thus, their true weight was incorporated into the noncase definition without need for adjustment. Using this approach, province/territory-specific multipliers were developed for cases of AGI (regardless of respiratory symptoms) and for cases of AGI without respiratory symptoms (Table 6). Essentially, the weight of completed interviews from each province/territory was given additional weight to account for those with missing responses in that province. This approach assumed that the distribution of chronic disease, medical condition, or medication use as the cause of AGI symptoms would be the same over time and that the number of stools for cases occurring in April to October would have the same distribution as that for cases occurring in November to April. Similarly, care seeking behaviours, duration of illness, hospitalization, etc. would also have the same distribution. A.3. Adjustment 1 Example for Cases of AGI (Regardless of Respiratory Symptoms) (A.1) C=A+BA, where A = weighted value of complete interviews for diarrhea only; B = weighted value of incomplete interviews for diarrhea only; C = adjustment multiplier (province specific) for complete interviews for diarrhea only; BC multiplier: (172,666.37 + 85,951.20)/172,666.37 = 1.50. Therefore, the individual complete diarrhea-only responses for BC are weighted an additional 1.5 times.

(A.1) C=A+BA, where A = weighted value of complete interviews for diarrhea only; B = weighted value of incomplete interviews for diarrhea only; C = adjustment multiplier (province specific) for complete interviews for diarrhea only; BC multiplier: (172,666.37 + 85,951.20)/172,666.37 = 1.50. Therefore, the individual complete diarrhea-only responses for BC are weighted an additional 1.5 times. A.4. Adjustment 2 Example for Cases of AGI without Respiratory Symptoms (A.2) Z=X+YX, where X = weighted value of complete interviews for diarrhea only and no respiratory symptoms; Y = weighted value of incomplete interviews for diarrhea only and no respiratory symptoms; Z = adjustment multiplier (province/territory specific) for complete interviews for diarrhea only and no respiratory symptoms; BC multiplier: (159,460.83 + 44,916.74)/159,460.83 = 1.28. Therefore, the individual complete diarrhea-only and no respiratory symptom responses for BC are weighted an additional 1.28 times.

(A.2) Z=X+YX, where X = weighted value of complete interviews for diarrhea only and no respiratory symptoms; Y = weighted value of incomplete interviews for diarrhea only and no respiratory symptoms; Z = adjustment multiplier (province/territory specific) for complete interviews for diarrhea only and no respiratory symptoms; BC multiplier: (159,460.83 + 44,916.74)/159,460.83 = 1.28. Therefore, the individual complete diarrhea-only and no respiratory symptom responses for BC are weighted an additional 1.28 times. A.5. Adjustment for Seasonality Comparison To report on seasonality by symptom, a different approach was needed to account for the missing data. The proportion of cases that experienced diarrhea only from April to October was calculated (60.04% and 46.72% for the international AGI case definition and the more specific case definition where cases with concurrent respiratory symptoms were removed, resp.). This proportion was then multiplied by the weighted total of respondents reporting diarrhea only that had missing information for each month to generate the estimate of the monthly number of cases that would have experienced diarrhea only. This value was combined with the reported values for the other symptom profiles (i.e., vomiting only and both diarrhea and vomiting) to estimate the seasonality of each symptom profile by month. (A.3) H=E×FG, where E = weighted value for month with incomplete interviews for diarrhea only; F = weighted value of completed interviews for diarrhea only that met case definition; G = weighted value of completed interviews for diarrhea only; H = estimated weighted value of respondents from month with incomplete data that met case definition.

eighted value for month with incomplete interviews for diarrhea only; F = weighted value of completed interviews for diarrhea only that met case definition; G = weighted value of completed interviews for diarrhea only; H = estimated weighted value of respondents from month with incomplete data that met case definition. A.6. Example For January and the international AGI case definition, there were 47 respondents with a weighted value of 98,976, indicating that they experienced diarrhea symptoms, but with missing data, the weighted value of these respondents was multiplied by the proportion of completed weighted interviews that were diarrhea only (60.04% or 593,306.07/988,144.51) to estimate that 59,430 weighted respondents would have experienced diarrhea only in January. These values were then used to estimate the monthly prevalence by the symptom profile reported in Figure 2. The difference of cases (98,967–59,430) would have been considered noncases that did not meet the case definition of having three or more loose stools in 24 hours or that their symptoms were related to a chronic condition or medication use.

A.6. Example For January and the international AGI case definition, there were 47 respondents with a weighted value of 98,976, indicating that they experienced diarrhea symptoms, but with missing data, the weighted value of these respondents was multiplied by the proportion of completed weighted interviews that were diarrhea only (60.04% or 593,306.07/988,144.51) to estimate that 59,430 weighted respondents would have experienced diarrhea only in January. These values were then used to estimate the monthly prevalence by the symptom profile reported in Figure 2. The difference of cases (98,967–59,430) would have been considered noncases that did not meet the case definition of having three or more loose stools in 24 hours or that their symptoms were related to a chronic condition or medication use. A.7. Limitations As the survey was a weighted study design, it required a weighted analysis, and thus, it was necessary to devise a weighted adjustment to address the CATI survey error. Ideally, we would have liked to perform this adjustment based on province/territory + age + gender weights so that we could comment on the differences among these demographics. However, due to data limitations in the province/territory + age + gender combinations where certain combinations were missing, this was not possible. Of the interviews with missing data from November 2014 to April 2015, about 8% (19/225) of AGI cases and 12% (18/144) of AGI cases without respiratory symptoms did not have completed interview data, affecting 11 and 14 of the province/territory + age + gender combinations, respectively. Therefore, province/territory + age + gender combinations were not used in the adjustment, and results on demographics could not be reported. The differences between age + gender combinations are inherently captured in the province-specific adjustment but are not able to be described explicitly. Available age and gender data were explored, and AGI results indicated little demographic difference to previous NSAGI studies (e.g., higher rates in children and lower rates in 65 years+ age group). Table 7 illustrates the demographic distribution of AGI cases with known responses and without adjustment for the missing 225 respondents for comparison with the adjusted values reported in the manuscript. This information however underestimates the true burden of AGI and thus cannot be used as the result for this research and is included only to demonstrate the general demographic conclusions that age and gender estimates do not differ greatly from what has been seen in previous NSAGI studies.