Browse the corpus

We acknowledge Dr Mary Thompson and Harshvinder Bhullar from EPIC for providing data from the THIN database. The corresponding author has the right to grant on behalf of all authors and does grant on behalf of all authors, an exclusive license (or non-exclusive for government employees) on a worldwide basis to the BMJ Publishing Group and its licensees to permit this article (if accepted) to be published in Gut editions and any other BMJPGL products to exploit all subsidiary rights, as set out in our license (http://group.bmj.com/products/journals/ instructions-for-authors/licence-forms). SUPPLEMENTARY MATERIAL is linked to the online version of the paper at http://www.nature.com/ajg Guarantor of the article: Gilaad G. Kaplan, MD, MPH. Specific author contributions: Conceiving the study idea, developing the study design, preparing the administrative data, performing the analysis, interpreting results, writing of the manuscript, has full access to all the data in the study and has final responsibility for the decision to submit for publication, and has seen and approved the final version of the manuscript: Gilaad G. Kaplan; preparing the administrative data, analyzing the data, editing the manuscript, and has seen and approved the final version of the manuscript: James Hubbard; contributing to the study design, interpreting results, editing the manuscript, and has seen and approved the final version of the manuscript: Remo Panaccione; contributing to the study design, interpreting results, editing the manuscript, and has seen and approved the final version of the manuscript: Subrata Ghosh; guiding the analysis of the data set, interpreting results, editing of the manuscript, and has seen and approved the final version of the manuscript: Amanda J. Wheeler; contributing to the study design, interpreting results, editing the manuscript, and has seen and approved the final version of the manuscript: Bruce E. Sands; contributing to the study design, interpreting results, editing the manuscript, and has seen and approved the final version of the manuscript: Joshua Korzenik; developing the study design, performing the analysis, interpreting the results, writing of the manuscript, and has seen and approved the final version of the manuscript: Paul J. Villeneuve.

twice daily) and supportive data indicating these agents are highly active against HP (25,26,27). A pilot study in patients previously failing treatment successfully used a similar regimen LEND (levofloxacin, esomeprazole, nitazoxanide, and doxycycline) to achieve a comparable ITT eradication rate of 90% (27/30) (28). Levofloxacin, a bactericidal fluoroquinolone antibiotic, has activity against HP due primarily to the drug's activity on bacterial DNA gyrase (29). Levofloxacin has been advocated for use in second- and third-line “rescue” regimens (4,5,6,7,8,12,30,31). Unfortunately fluoroquinolone resistance, especially in patients who have routinely received a fluoroquinolone for other indications, is of particular concern (4,12,29). A lower dose of levofloxacin was used to aid in the tolerability of this four-drug regimen, however, as levofloxacin is a concentration-dependent agent, higher doses may increase eradication rates. Although the optimal dose of levofloxacin is yet to be determined, our study demonstrated efficacy using the lower dose of 250 mg once daily. Considering the LOAD regimen was well tolerated overall, an increase in levofloxacin dosage may be warranted.

INTRODUCTION Sir Samuel Wilks introduced ulcerative colitis into the medical vernacular in 1875 (1). The predominant air pollutant at that time was sulfurous gases arising from combustion of coal in industrial regions (2). Crohn's disease became readily recognizable in 1932 during the advent of the automobile era (1). Traffic-related pollutants such as nitrogen oxides and particulate matter that arise from the combustion of fossil fuels began to increase in urban societies during the 20th century (2). Early in the 20th century, the incidence of ulcerative colitis was higher than Crohn's disease. However, although the incidence rates of ulcerative colitis have stabilized, the incidence of Crohn's disease began to steadily rise in industrialized nations in Europe (3) and North America (4). In developing nations, inflammatory bowel disease (IBD) rarely occurred; however, as these nations became industrialized, the incidence of ulcerative colitis and then Crohn's disease increased (5,6). Despite numerous studies, few environmental risk factors for IBD have been identified (4); however, they have not completely explained the processes that provoke IBD. Moreover, risk factors associated with industrialization have been incompletely investigated.

lcerative colitis and then Crohn's disease increased (5,6). Despite numerous studies, few environmental risk factors for IBD have been identified (4); however, they have not completely explained the processes that provoke IBD. Moreover, risk factors associated with industrialization have been incompletely investigated. Air pollution directly affects pulmonary diseases including asthma and lung cancer (7). Air pollution exposure has also been shown to impair lung function development in children (8). In addition, air pollution has also been associated with a variety of non-pulmonary diseases including myocardial infarction (9), appendicitis (10), and rheumatoid arthritis (11). Air pollution-mediated inflammation has been implicated as the cause of several adverse health effects (12). Similar pro-inflammatory processes occur in IBD (13). Thus, we studied whether residential exposures to ambient air pollution concentrations were associated with the development of Crohn's disease and ulcerative colitis.

tion-mediated inflammation has been implicated as the cause of several adverse health effects (12). Similar pro-inflammatory processes occur in IBD (13). Thus, we studied whether residential exposures to ambient air pollution concentrations were associated with the development of Crohn's disease and ulcerative colitis. METHODS Patient data source The health improvement network (THIN) database consists of general practice electronic medical records for a sample of patients in the United Kingdom. Although the database is cleared of personal identifiers, it includes electronic information on patients at an individual level that can be tracked through time and includes demographics, physician defined diagnoses, prescription medications, cigarette smoking status, and death certificates. The THIN database also includes postcode linked area based socioeconomic and environmental indicators. Every 3 months, data from practices are exported to the THIN administrators and changes (e.g. new diagnosis) are recorded (14). Clinical diagnoses and prescribing data in the THIN database have been shown to be similar to the general practice research database (15).

rea based socioeconomic and environmental indicators. Every 3 months, data from practices are exported to the THIN administrators and changes (e.g. new diagnosis) are recorded (14). Clinical diagnoses and prescribing data in the THIN database have been shown to be similar to the general practice research database (15). Environmental data source The National Environmental Technology Centre and Department of Environment, Food and Rural Affairs monitored ambient air quality in the United Kingdom using urban and rural fixed site monitors. Air pollutants that were monitored included particulate matter with an aerodynamic diameter of <10 μm (PM10), sulfur dioxide (SO2), and nitrogen dioxide (NO2). Nitrogen oxides were also recorded; however, this pollutant was excluded in this study because nitrogen oxides were highly correlated to NO2. Air pollutants were measured by fixed monitoring sites that include the Automatic Urban Network (103 sites), the Automatic Rural Network (24 sites), and the Automatic London Network (14 sites). Regional mean air pollution concentrations were extrapolated by creating air pollution concentration interpolation maps (16).

Air pollutants were measured by fixed monitoring sites that include the Automatic Urban Network (103 sites), the Automatic Rural Network (24 sites), and the Automatic London Network (14 sites). Regional mean air pollution concentrations were extrapolated by creating air pollution concentration interpolation maps (16). In 2001, the annual means for NO2, SO2, and PM10 were determined for each ward (∼2,000 residents). The wards were then stratified into quintiles based on the distribution of the pollution levels. The range of annual mean air pollutant concentrations for each quintile can be found in Supplementary Appendix online. On the basis of residential postcode, the participants in the THIN database were assigned a pollution score. The first linkage of air pollutant quintiles to postcodes occurred on 1 July 2006 and was updated if the case/control moved residence but remained registered within the practice. In total, 322 practices had air pollution data linked to 2,496,166 active members. Study population A nested case–control study was conducted by identifying incident cases of IBD and matched controls from within the THIN database. Subjects were eligible if they met the following criteria before 1 July 2005: between the ages of 5 and 84 years, enrolled in the THIN database for at least 3 years, and not previously diagnosed with IBD. The cohort members were followed from 1 July 2005 until the earliest of: first code of IBD, migration out of practice, death, or last data collection up to 31 October 2008.

before 1 July 2005: between the ages of 5 and 84 years, enrolled in the THIN database for at least 3 years, and not previously diagnosed with IBD. The cohort members were followed from 1 July 2005 until the earliest of: first code of IBD, migration out of practice, death, or last data collection up to 31 October 2008. During the study period, 2,063 THIN members were recorded with a new code for the IBDs. To minimize misclassification of IBD type, non-IBD patients, and prevalent cases, we excluded (i) 51 cases coded with both Crohn's disease and ulcerative colitis; (ii) 108 cases whose first code was defined as “exacerbation;” (iii) 732 patients coded for indeterminate or unexplained colitis; (iv) three cases without air pollution data; and (v) 211 subjects without a prescription for at least one of corticosteroids, mesalamine, sulfasalazine, azathioprine, 6-mercaptopurine, methotrexate, infliximab, and adalimumab. In total, 961 patients were newly diagnosed with Crohn's disease (n=367) or ulcerative colitis (n=591) between 1 July 2005 and 31 October 2008. The index date for an incident case was the date of a first code with IBD.

mesalamine, sulfasalazine, azathioprine, 6-mercaptopurine, methotrexate, infliximab, and adalimumab. In total, 961 patients were newly diagnosed with Crohn's disease (n=367) or ulcerative colitis (n=591) between 1 July 2005 and 31 October 2008. The index date for an incident case was the date of a first code with IBD. A pool of eligible controls was created by randomly sampling 100,000 members of the THIN database. From the pool of controls, five controls individually matched by age (within 1 year) and sex to each IBD incident cases were randomly selected. Controls were excluded if they were diagnosed with IBD before the index date of their matched case. The selection of controls was performed without replacement. Controls were subjected to the same inclusion/exclusion criteria as the cases. Age- and sex-matched controls were assigned the index date of their case for the purposes of the assignment of exposures. The study protocol was approved by the Conjoint Health Research Ethics Board at the University of Calgary. Data analysis The primary analysis involved stratifying individuals according to their air pollution levels based on quintiles grouped into high exposure (third, fourth, and fifth quintiles) and low exposure (first and second quintiles). Conditional logistic regression analysis was used to determine whether exposures to NO2, PM10, or SO2 were associated with an incident diagnosis of Crohn's disease and/or ulcerative colitis after adjusting for confounders.

high exposure (third, fourth, and fifth quintiles) and low exposure (first and second quintiles). Conditional logistic regression analysis was used to determine whether exposures to NO2, PM10, or SO2 were associated with an incident diagnosis of Crohn's disease and/or ulcerative colitis after adjusting for confounders. Potential confounders that were controlled for included cigarette smoking status, socioeconomic status, prescription aspirin or non-steroidal anti-inflammatory drugs (NSAIDs), and appendectomy (4). Cigarette smoking status was defined as follows: (i) current smoker, coded for smoking within 3 months of the index date; (ii) ex-smoker, coded for smoking before, but not within 3 months of the index date; and (iii) never smoker, coded either as a lifetime non-smoker or not recorded as a smoker. Socioeconomic status was estimated using postcode level indicators generated from an index of deprivation based on the percentage of households: without access to a car, not in owner occupied accommodation, in overcrowded accommodation, and the percentage of the population aged 16–74 years who were unemployed (17). For prescription NSAIDs, the participant was classified as current user, prescription within 1 month of the index date; past user, prescription that ended before the month preceding the index date; or never user.

overcrowded accommodation, and the percentage of the population aged 16–74 years who were unemployed (17). For prescription NSAIDs, the participant was classified as current user, prescription within 1 month of the index date; past user, prescription that ended before the month preceding the index date; or never user. We separately analyzed models that were stratified by age quartiles for Crohn's disease (≤23, 24–43, 44–57, and ≥58 years) and ulcerative colitis (≤35, 36–48, 49–63, and ≥64 years). In a post hoc analysis, younger age groups (≤30, ≤25, and ≤20 years) were also evaluated for ulcerative colitis patients. All estimates were reported as adjusted odds ratios (ORs) with 95% confidence intervals (CIs). Linear trends in the adjusted ORs were evaluated across the quintiles to examine for a dose–response relationship. All analyses were performed using SAS software (SAS, version 9.2, SAS Institute, Cary, NC).

colitis patients. All estimates were reported as adjusted odds ratios (ORs) with 95% confidence intervals (CIs). Linear trends in the adjusted ORs were evaluated across the quintiles to examine for a dose–response relationship. All analyses were performed using SAS software (SAS, version 9.2, SAS Institute, Cary, NC). Sensitivity analysis Additional sensitivity analyses were performed to explore the potential effects of misclassification biases with both exposure and health outcome measures. As air pollution scores were assigned based on residence after 1 July 2006, a sensitivity analysis was performed to restrict the analysis to patients diagnosed after 1 July 2006. To minimize the misclassification of prevalent cases, the analysis was restricted to at least a 5-year washout period without a code for IBD before the index date. A sensitivity analysis was performed to exclude potentially prevalent IBD patients prescribed immunosuppressant within 3 months of the index date. Finally, subjects living in wards classified in the third quintile were excluded allowing for comparisons between the highest (fifth and fourth) and lowest (first and second) exposures. Finally, we restricted our analysis to cases and controls living in urban wards at the index date. Urban areas were defined as settlements, with a population size exceeding 10,000 persons (18).

uintile were excluded allowing for comparisons between the highest (fifth and fourth) and lowest (first and second) exposures. Finally, we restricted our analysis to cases and controls living in urban wards at the index date. Urban areas were defined as settlements, with a population size exceeding 10,000 persons (18). RESULTS We identified 591 individuals newly diagnosed with UC and 367 with Crohn's disease. The demographics at diagnosis of ulcerative colitis and Crohn's disease patients, as well as their matched controls at the index date, are shown in Table 1. Crohn's disease Table 2 presents estimates of the risk of developing Crohn's disease after adjusting for NO2, smoking, socioeconomic status, NSAID use, and prior appendectomy. Compared with lifetime non-smokers, current smokers (OR=1.55; 95% CI=1.14–2.11) and ex-smokers (OR=1.86; 95% CI=1.38–2.52) were at increased risk of being diagnosed with Crohn's disease. Appendectomy (OR=1.94; 95% CI=1.28–2.95) increased the risk of developing Crohn's disease. The adjusted ORs reported in Table 2 did not significantly change when we modeled these variables with SO2 or PM10 rather than NO2.

okers (OR=1.86; 95% CI=1.38–2.52) were at increased risk of being diagnosed with Crohn's disease. Appendectomy (OR=1.94; 95% CI=1.28–2.95) increased the risk of developing Crohn's disease. The adjusted ORs reported in Table 2 did not significantly change when we modeled these variables with SO2 or PM10 rather than NO2. Overall, individuals who lived in regions of higher concentrations of NO2, SO2, and PM10 were not at risk of Crohn's disease ( Table 3). However, individuals ≤23 years were significantly more likely to be diagnosed with Crohn's disease if they lived in regions of higher concentrations of NO2 (OR=2.31; 95% CI=1.25–4.28) after adjusting for confounders. Crohn's patients aged 44–57 years were less likely to live in regions of higher NO2 (OR=0.56; 95% CI=0.33–0.95) and PM10 (OR=0.48; 95% CI=0.29–0.80) ( Table 3). Among patients ≤23 years, the adjusted OR increased linearly across quintile levels for NO2 (P=0.02) and PM10 (P=0.05), but not SO2 (Figure 1). Ulcerative colitis Table 2 shows the risk of developing ulcerative colitis after adjusting for NO2, smoking, socioeconomic status, NSAIDs, and appendectomy. Compared with non-smokers, current smokers were protected from ulcerative colitis (OR=0.59; 95% CI=0.43–0.81), whereas ex-smokers were at increased risk (OR=2.05; 95% CI=1.67–2.53). Patients with an appendectomy were less likely (OR=0.38; 95% CI=0.23–0.64) to be diagnosed with ulcerative colitis. The adjusted ORs reported in Table 2 did not significantly change when we modeled these variables with SO2 or PM10.

5% CI=0.43–0.81), whereas ex-smokers were at increased risk (OR=2.05; 95% CI=1.67–2.53). Patients with an appendectomy were less likely (OR=0.38; 95% CI=0.23–0.64) to be diagnosed with ulcerative colitis. The adjusted ORs reported in Table 2 did not significantly change when we modeled these variables with SO2 or PM10. Overall, NO2, SO2, and PM10 were not associated with the diagnosis of ulcerative colitis at all ages and when the age groups were stratified by quartiles ( Table 4). However, patients diagnosed with ulcerative colitis ≤25 (OR=2.00; 95% CI=1.08–3.72) and ≤20 (OR=2.62; 95% CI=1.15–6.00) years were significantly more likely to live in regions of higher concentrations of SO2 (Figure 2). Reducing the age interval (i.e. ≤25 and ≤20 years) did not influence the risk estimates associated with NO2 or PM10. The ORs did not increase linearly across quintile levels (Figure 3).

nd ≤20 (OR=2.62; 95% CI=1.15–6.00) years were significantly more likely to live in regions of higher concentrations of SO2 (Figure 2). Reducing the age interval (i.e. ≤25 and ≤20 years) did not influence the risk estimates associated with NO2 or PM10. The ORs did not increase linearly across quintile levels (Figure 3). Sensitivity analysis We performed sensitivity analyses on each pollutant among the patients diagnosed with early onset Crohn's disease (≤23 years) and ulcerative colitis (≤25 years). When the washout period was extended from 3 to 5 years, the risk estimates marginally increased for both NO2 in Crohn's disease (OR=2.46; 95% CI=1.29–4.69) and for SO2 in ulcerative colitis (OR=2.12; 95% CI=1.14–3.94); in addition, young incident cases of Crohn's disease were significantly more likely to live in regions of higher PM10. Restricting analysis to IBD cases diagnosed after 1 July 2006 also increased the risk estimates for NO2 in Crohn's disease and for SO2 in ulcerative colitis; in addition, Crohn's patients were more likely to live in wards with higher concentrations of PM10 (2.08; 95% CI=1.04–4.14) exposure. Exclusion of cases/controls living third quintile wards strengthened the association for NO2 in Crohn's disease, whereas the effects of SO2 were no longer significant in ulcerative colitis. Excluding potentially prevalent cases/controls prescribed an immunosuppressant within 3 months of the index date reduced the risk estimates, but the associations were still significant. Restricting the analysis to only cases and controls living in urban areas strengthened the association between NO2 (3.27; 95% CI=1.32–8.06) and Crohn's disease ( Table 5).

/controls prescribed an immunosuppressant within 3 months of the index date reduced the risk estimates, but the associations were still significant. Restricting the analysis to only cases and controls living in urban areas strengthened the association between NO2 (3.27; 95% CI=1.32–8.06) and Crohn's disease ( Table 5). DISCUSSION Overall, air pollution exposure was not associated with the risk of developing IBD. However, children and young adults living in areas with higher levels of SO2 were more likely to develop ulcerative colitis. Crohn's disease was more commonly diagnosed in young patients living in regions of higher concentrations of NO2. The effect of NO2 on early onset Crohn's disease persisted after adjusting for confounders and across sensitivity analyses. The association demonstrated a dose–response effect and was strengthened when the analysis was restricted to urban areas. Finally, middle-aged adults diagnosed with Crohn's disease demonstrated a paradoxical negative association to NO2 and PM10. Collectively, these findings suggest that traffic-related pollutants (i.e. NO2) and industrial-based pollutants (i.e. SO2) may have age-specific effects on the development of Crohn's disease and ulcerative colitis, respectively.

nosed with Crohn's disease demonstrated a paradoxical negative association to NO2 and PM10. Collectively, these findings suggest that traffic-related pollutants (i.e. NO2) and industrial-based pollutants (i.e. SO2) may have age-specific effects on the development of Crohn's disease and ulcerative colitis, respectively. Early onset IBD differs from late-onset IBD by incidence patterns, having a more aggressive prognosis, and a different phenotypic expression (19,20). Likewise, children are more susceptible to the toxic effects of pollutants compared with adults (21). This may be explained by the fact that children and young adults tend to be more active and spend more time outdoors leading to greater exposure to air pollutants (21). Gastrointestinal absorption of pollutants (e.g. lead (22)) has been shown to be greater in children, and children metabolize pollutants less efficiently than adults (21). Furthermore, adverse health effects of pollutants may be greater in children because of dynamic development and growth (21).

to air pollutants (21). Gastrointestinal absorption of pollutants (e.g. lead (22)) has been shown to be greater in children, and children metabolize pollutants less efficiently than adults (21). Furthermore, adverse health effects of pollutants may be greater in children because of dynamic development and growth (21). Adults aged 44–57 years were less likely to develop Crohn's disease in regions of higher concentrations of NO2 and PM10. Differential effects observed across ages may have been due to exposure misclassification. Air pollution exposure was defined by residential measurements. Children and young adults spend more time outside their home and often attend schools nearby, whereas adults tend to work in areas with different air pollution exposures than their homes. Thus, residential air pollution measurements may more accurately reflect personal exposures in the young compared with older adults. Alternatively, environmental risk factors have been shown to have polarizing effects when comparing early onset IBD to late-onset IBD. For example, living with pets before the age of 5 years reduced the risk of developing adult-onset Crohn's disease, whereas living with pets increased the risk of pediatric-onset Crohn's disease (23). Future studies will be needed to determine whether this paradoxical finding was due to methodological considerations or because of the complex interaction between the environment and the age of onset of IBD.

dult-onset Crohn's disease, whereas living with pets increased the risk of pediatric-onset Crohn's disease (23). Future studies will be needed to determine whether this paradoxical finding was due to methodological considerations or because of the complex interaction between the environment and the age of onset of IBD. Air pollution has been associated with other chronic inflammatory diseases that, like IBD, occur in genetically predisposed individuals who are exposed to environmental triggers. Elevated levels of SO2 and NO2 were associated with an 11-fold increased risk for a relapse of multiple sclerosis (24). Moreover, women exposed to traffic pollutants were at increased risk of rheumatoid arthritis (11). IBD patients are at higher risk of developing rheumatoid arthritis and multiple sclerosis (25), which suggest that these diseases may share genetic susceptibility and environmental risk factors. In addition, acute exposure to elevated levels of NO2 and ozone have been shown to increase the risk of appendicitis (26); likewise, appendicitis increased the risk of Crohn's disease in our study, as with others (27). These studies suggest a complex interaction between common environmental exposures such as air pollution and immune-mediated diseases.

evated levels of NO2 and ozone have been shown to increase the risk of appendicitis (26); likewise, appendicitis increased the risk of Crohn's disease in our study, as with others (27). These studies suggest a complex interaction between common environmental exposures such as air pollution and immune-mediated diseases. The mechanisms by which air pollution may influence the development of IBD are speculative. The adverse health effects associated with air pollution are in part due to inflammation. The inhalation of diesel exhaust in healthy volunteers led to a rise in T-helper-1-associated cytokines such as tumor necrosis factor (12). Animals that fed air pollutants experienced oxidative damage of colonic mucosa (28). Thus, exposure of air pollutants, either through inhalation or ingestion, may incite inflammatory pathways that have been postulated to be central in the pathogenesis of IBD (13). Alternatively, diesel exhaust exposure increases susceptibility to bacterial (29) pulmonary infections through impairment of microbial defense (30). If air pollution similarly affects gastrointestinal innate immunity, then pollutants may predispose to IBD through altering interactions between the epithelial barrier and the intestinal microbiome. Future studies will be needed to elucidate the mechanistic relationship between air pollution and IBD.

bial defense (30). If air pollution similarly affects gastrointestinal innate immunity, then pollutants may predispose to IBD through altering interactions between the epithelial barrier and the intestinal microbiome. Future studies will be needed to elucidate the mechanistic relationship between air pollution and IBD. Several limitations should be considered. First, IBD cases may have been misclassified and prevalent cases may have been mislabeled as incident. To minimize misclassification, IBD patients were only included if they also had a prescription for an IBD-related medication (i.e. mesalamine, corticosteroid, immunosuppressant, or biologic). We also performed sensitivity analyses to minimize misclassification of prevalent cases by increasing the washout period from 3 to 5 years and excluding patients prescribed an immunosuppressant within 3 months of the index date. In both sensitivity analyses, the primary results remained unchanged. Second, misclassification of air pollution exposures was possible. Air pollution scores were assigned based on regional estimates (i.e. at the ward level rather than the patient level) derived in 2001 and assigned to patients based on residence after 1 July 2006. Patients diagnosed after 1 July 2005 and who moved residence before 1 July 2006 may have been assigned an incorrect pollution score. However, this misclassification was presumably non-differential, which has the effect of biasing estimates toward the null. In fact, when the analysis was restricted to patients diagnosed after 1 July 2006, the risk estimates were strengthened. Third, the investigators cannot access postcodes in the THIN database and thus, more refined exposure assessments (e.g. using land use regression models) could not be performed. Fourth, other important air pollutants (e.g. ozone) were not available in the THIN database to study. Ozone exposure has been shown to increase the risk of appendicitis (26) and thus, future studies are necessary to evaluate the effects of other pollutants on IBD development. Fifth, the effect of SO2 on ulcerative colitis was not as robust as those observed between NO2 and Crohn's disease. SO2 did not follow a dose–response relationship and when extremes were compared (i.e. fourth and fifth vs. first and second quintiles), the effect of SO2 lost significance. Thus, one should cautiously interpret the effects of SO2 on ulcerative colitis. Sixth, residual confounding may have introduced bias into our risk estimates.

t follow a dose–response relationship and when extremes were compared (i.e. fourth and fifth vs. first and second quintiles), the effect of SO2 lost significance. Thus, one should cautiously interpret the effects of SO2 on ulcerative colitis. Sixth, residual confounding may have introduced bias into our risk estimates. Although important confounders (e.g. smoking) were controlled, others (e.g. occupational exposures) were missing. The lack of information on occupational exposures limited our assessment of adult-onset IBD. For example, adults employed in “driving” occupations, which have greater exposure to traffic-related pollutants such as NO2, have been shown to be at increased risk for developing Crohn's disease (31). Seventh, multiple comparisons were analyzed including three pollutants and several age groups. Furthermore, significant findings were only observed after subgroup analysis. Consequently, significant findings may have occurred by chance and these findings need to be replicated in other IBD populations. Eighth, countries such as China and India have significant air pollution, but have reported lower incidence rates of IBD. However, the incidence of IBD has rapidly increased in the past generation in developing nations as they have become industrialized (5,6) Also, genetic susceptibility to IBD differs across races (32); gene–environment interactions may explain differences between countries. Thus, future studies will be needed to corroborate our findings, to explore late-onset IBD in the context of air pollution exposures at home and at work, and to evaluate gene–pollutant interactions.

susceptibility to IBD differs across races (32); gene–environment interactions may explain differences between countries. Thus, future studies will be needed to corroborate our findings, to explore late-onset IBD in the context of air pollution exposures at home and at work, and to evaluate gene–pollutant interactions. When all age groups were combined, air pollution did not increase the risk of developing IBD. However, exposure to NO2 and SO2 may increase the risk of developing early onset Crohn's disease and ulcerative colitis, respectively. In contrast, exposure of NO2 and PM10 in middle-aged adults was negatively associated with the development of Crohn's disease. If these findings are confirmed, then this work provides novel insight into the complex pathogenesis of IBD. STUDY HIGHLIGHTS We acknowledge Dr Mary Thompson and Harshvinder Bhullar from EPIC for providing data from the THIN database. The corresponding author has the right to grant on behalf of all authors and does grant on behalf of all authors, an exclusive license (or non-exclusive for government employees) on a worldwide basis to the BMJ Publishing Group and its licensees to permit this article (if accepted) to be published in Gut editions and any other BMJPGL products to exploit all subsidiary rights, as set out in our license (http://group.bmj.com/products/journals/ instructions-for-authors/licence-forms). SUPPLEMENTARY MATERIAL is linked to the online version of the paper at http://www.nature.com/ajg Guarantor of the article: Gilaad G. Kaplan, MD, MPH.

ign, interpreting results, editing the manuscript, and has seen and approved the final version of the manuscript: Joshua Korzenik; developing the study design, performing the analysis, interpreting the results, writing of the manuscript, and has seen and approved the final version of the manuscript: Paul J. Villeneuve. Financial support: This work was supported by an operating grant from the Broad Medical Research Program of the Broad Foundation and from the CCFC Chair in IBD Research. G.G.K. is supported through a New Investigator Award from the Canadian Institute of Health Research and a Population Health Investigator Award from the Alberta Heritage Foundation for Medical Research. Potential competing interests: None. Supplementary Material Supplementary Appendix Online Click here for additional data file. Figure 1 Dose–response relationship across quintiles of nitrogen dioxide (NO2), sulfur dioxide (SO2), and particulate matter <10 μm (PM10) exposures for the adjusted odds ratio of developing Crohn's disease ≤23 years. Figure 2 Adjusted risk of developing ulcerative colitis among individuals living in wards with higher concentrations of sulfur dioxide (SO2) stratified by decreasing age at diagnosis. Figure 3 Dose–response relationship across quintiles of nitrogen dioxide (NO2), sulfur dioxide (SO2), and particulate matter <10 μm (PM10) exposures for the adjusted odds ratio of developing ulcerative colitis ≤25 years.

Figure 2 Adjusted risk of developing ulcerative colitis among individuals living in wards with higher concentrations of sulfur dioxide (SO2) stratified by decreasing age at diagnosis. Figure 3 Dose–response relationship across quintiles of nitrogen dioxide (NO2), sulfur dioxide (SO2), and particulate matter <10 μm (PM10) exposures for the adjusted odds ratio of developing ulcerative colitis ≤25 years. Table 1 Characteristics of patients at diagnosis of Crohn's disease or ulcerative colitis, as well as matched controls at the index date Crohn's disease, n=367 Matched controlsa, n=1,833 Ulcerative colitis, n=591 Matched controlsa, n=2,962 Age, median (IQR) 43 (23, 58) 43 (23, 58) 49 (35, 64) 49 (35, 64) Sex, % Male 48.8 48.8 57.5 57.4 Female 51.2 51.2 42.5 42.6 Smoking, % Neverb 49.6 60.5 48.9 56.2 Current 22.6 19.5 9.3 18.3 Ex-smoker 27.8 20.0 41.8 25.5 Socioeconomic statusc, % 1 (Least deprived) 29.4 27.6 25.4 28.7 2 24.8 22.0 24.9 23.0 3 22.1 21.7 23.5 20.7 4 13.6 17.7 17.1 16.9 5 (Most deprived) 10.1 11.0 9.1 10.7 Prescription NSAID, % Never 41.7 44.0 37.5 39.8 Current 9.0 6.8 9.5 9.0 Past 49.3 49.2 53.0 51.2 Appendectomy, % 9.5 5.2 2.7 6.7 Diagnosis IBD medicationsd, % Corticosteroids 38.4 1.4 46.4 1.8 5-ASA 67.6 0.1 81.0 0.2 Azathioprine or 6-mercaptopurine 13.6 0.05 3.9 0.2 Methotrexate 1.1 0.1 0.3 0.3 Infliximab or adalimumab 0 0 0 0 5-ASA, 5-aminosalicylic acid; IBD, inflammatory bowel disease; IQR, interquartile range; NA, not applicable; NSAID, non-steroidal anti-inflammatory drug. a Controls matched on age (within 1 year) and sex.

Table 1 Characteristics of patients at diagnosis of Crohn's disease or ulcerative colitis, as well as matched controls at the index date Crohn's disease, n=367 Matched controlsa, n=1,833 Ulcerative colitis, n=591 Matched controlsa, n=2,962 Age, median (IQR) 43 (23, 58) 43 (23, 58) 49 (35, 64) 49 (35, 64) Sex, % Male 48.8 48.8 57.5 57.4 Female 51.2 51.2 42.5 42.6 Smoking, % Neverb 49.6 60.5 48.9 56.2 Current 22.6 19.5 9.3 18.3 Ex-smoker 27.8 20.0 41.8 25.5 Socioeconomic statusc, % 1 (Least deprived) 29.4 27.6 25.4 28.7 2 24.8 22.0 24.9 23.0 3 22.1 21.7 23.5 20.7 4 13.6 17.7 17.1 16.9 5 (Most deprived) 10.1 11.0 9.1 10.7 Prescription NSAID, % Never 41.7 44.0 37.5 39.8 Current 9.0 6.8 9.5 9.0 Past 49.3 49.2 53.0 51.2 Appendectomy, % 9.5 5.2 2.7 6.7 Diagnosis IBD medicationsd, % Corticosteroids 38.4 1.4 46.4 1.8 5-ASA 67.6 0.1 81.0 0.2 Azathioprine or 6-mercaptopurine 13.6 0.05 3.9 0.2 Methotrexate 1.1 0.1 0.3 0.3 Infliximab or adalimumab 0 0 0 0 5-ASA, 5-aminosalicylic acid; IBD, inflammatory bowel disease; IQR, interquartile range; NA, not applicable; NSAID, non-steroidal anti-inflammatory drug. a Controls matched on age (within 1 year) and sex. b Never smokers were defined as either no record of smoking in the database or coded as a ‘never smoker' before index date. c Socioeconomic status was derived from an index of deprivation by area based on percentage of households: without access to a car, not in owner occupied accommodation, in overcrowded accommodation, and percentage of the economically active population aged 16–74 years who were unemployed.

b Never smokers were defined as either no record of smoking in the database or coded as a ‘never smoker' before index date. c Socioeconomic status was derived from an index of deprivation by area based on percentage of households: without access to a car, not in owner occupied accommodation, in overcrowded accommodation, and percentage of the economically active population aged 16–74 years who were unemployed. d Prescription registered within 3 months of the index date for cases and controls. Patients may be prescribed one or more IBD-related medications. Table 2 The risk of developing Crohn's disease and ulcerative colitis associated with smoking, socioeconomic status, NSAID use, and prior appendectomy Crohn's disease odds ratio (95% CI)a n=367/1,833b Ulcerative colitis odds ratio (95% CI)a n=591/2,962b Smoking Neverc 1.00 1.00 Currentd 1.55 (1.14–2.11) 0.59 (0.43–0.81) Ex-smoker 1.86 (1.38–2.52) 2.05 (1.67–2.53) Socioeconomic status 1 (Least deprived) 1.00 1.00 2 1.01 (0.74–1.38) 1.21 (0.94–1.56) 3 0.89 (0.64–1.23) 1.30 (1.00–1.68) 4 0.68 (0.47–0.98) 1.14 (0.86–1.51) 5 (Most deprived) 0.77 (0.51–1.16) 0.97 (0.68–1.38) NSAID Never 1.00 1.00 Currentd 1.32 (0.83–2.09) 1.05 (0.74–1.49) Past 1.03 (0.80–1.33) 1.06 (0.87–1.29) Appendectomy 1.94 (1.28–2.95) 0.38 (0.23–0.64) CI, confidence interval; NSAID, non-steroidal anti-inflammatory drug. a Matched for age and sex and adjusted for exposure to nitrogen dioxide, smoking, socioeconomic status, prior appendectomy, and NSAID use. b n denotes the number of cases/controls.

Table 2 The risk of developing Crohn's disease and ulcerative colitis associated with smoking, socioeconomic status, NSAID use, and prior appendectomy Crohn's disease odds ratio (95% CI)a n=367/1,833b Ulcerative colitis odds ratio (95% CI)a n=591/2,962b Smoking Neverc 1.00 1.00 Currentd 1.55 (1.14–2.11) 0.59 (0.43–0.81) Ex-smoker 1.86 (1.38–2.52) 2.05 (1.67–2.53) Socioeconomic status 1 (Least deprived) 1.00 1.00 2 1.01 (0.74–1.38) 1.21 (0.94–1.56) 3 0.89 (0.64–1.23) 1.30 (1.00–1.68) 4 0.68 (0.47–0.98) 1.14 (0.86–1.51) 5 (Most deprived) 0.77 (0.51–1.16) 0.97 (0.68–1.38) NSAID Never 1.00 1.00 Currentd 1.32 (0.83–2.09) 1.05 (0.74–1.49) Past 1.03 (0.80–1.33) 1.06 (0.87–1.29) Appendectomy 1.94 (1.28–2.95) 0.38 (0.23–0.64) CI, confidence interval; NSAID, non-steroidal anti-inflammatory drug. a Matched for age and sex and adjusted for exposure to nitrogen dioxide, smoking, socioeconomic status, prior appendectomy, and NSAID use. b n denotes the number of cases/controls. c Never smokers were defined as either no record of smoking in the database or coded as a ‘never smoker' before index date. d Current defined as exposure at index date.

a Matched for age and sex and adjusted for exposure to nitrogen dioxide, smoking, socioeconomic status, prior appendectomy, and NSAID use. b n denotes the number of cases/controls. c Never smokers were defined as either no record of smoking in the database or coded as a ‘never smoker' before index date. d Current defined as exposure at index date. Table 3 The age-stratified adjusted risk of developing Crohn's disease among individuals living in wards with higher concentrations of NO2, SO2, and PM10 Crohn's disease odds ratio (95% confidence interval)a All agesc, n=367/1,833 ≤23 Yearsc, n=93/465 24–43 Yearsc, n=95/474 44–57 Yearsc, n=84/420 ≥58 Yearsc, n=95/474 NO2 1.02 (0.79–1.32) 2.31 (1.25–4.28) 0.68 (0.41–1.13) 0.56 (0.33–0.95) 1.28 (0.78–2.09) SO2 0.95 (0.74–1.21) 1.23 (0.73–2.05) 0.88 (0.55–1.43) 0.67 (0.40–1.11) 1.09 (0.68–1.76) PM10 0.91 (0.71–1.17) 1.73 (0.98–3.03) 0.76 (0.46–1.27) 0.48 (0.29–0.80) 1.10 (0.67–1.82) NO2, nitrogen dioxide; NSAID, non-steroidal anti-inflammatory drug; PM10, particulate matter <10μm; SO2, sulfur dioxide. a Matched for age and sex and adjusted for smoking, socioeconomic status, prior appendectomy, and NSAID use. Air pollution levels were stratified into high exposure (third, fourth, and fifth quintiles) and low exposure (referent). b n denotes the number of cases/controls. c Age cutoffs were defined by the distribution of ages (i.e. quartiles) at diagnosis of Crohn's disease.

a Matched for age and sex and adjusted for smoking, socioeconomic status, prior appendectomy, and NSAID use. Air pollution levels were stratified into high exposure (third, fourth, and fifth quintiles) and low exposure (referent). b n denotes the number of cases/controls. c Age cutoffs were defined by the distribution of ages (i.e. quartiles) at diagnosis of Crohn's disease. Table 4 The age-stratified adjusted risk of developing ulcerative colitis among individuals living in wards with higher concentrations of NO2, SO2, and PM10 Ulcerative colitis odds ratio (95% confidence interval)a All ages, n=591/2,962b ≤35 Years, n=152/763c 36–48 Years, n=140/696c 49–63 Years, n=148/749c ≥64 Years, n=151/754c NO2 1.00 (0.82–1.22) 0.76 (0.51–1.12) 1.14 (0.74–1.77) 1.07 (0.72–1.59) 1.08 (0.73–1.62) SO2 1.16 (0.95–1.41) 1.34 (0.89–2.02) 1.01 (0.67–1.54) 1.07 (0.72–1.57) 1.20 (0.81–1.78) PM10 1.10 (0.90–1.34) 1.11 (0.74–1.67) 1.11 (0.73–1.70) 1.12 (0.75–1.66) 1.08 (0.74–1.58) NO2, nitrogen dioxide; NSAID, non-steroidal anti-inflammatory drug; PM10, particulate matter <10μm; SO2, sulfur dioxide. a Matched for age and sex and adjusted for smoking, socioeconomic status, prior appendectomy, and NSAID use. Air pollution levels were stratified into high exposure (third, fourth, and fifth quintiles) and low exposure (referent). b n denotes the number of cases/controls. c Age cutoffs were defined by the distribution of ages (i.e. quartiles) at diagnosis of ulcerative colitis.

a Matched for age and sex and adjusted for smoking, socioeconomic status, prior appendectomy, and NSAID use. Air pollution levels were stratified into high exposure (third, fourth, and fifth quintiles) and low exposure (referent). b n denotes the number of cases/controls. c Age cutoffs were defined by the distribution of ages (i.e. quartiles) at diagnosis of ulcerative colitis. Table 5 Sensitivity analyses Crohn's disease (≤23 years) odds ratio (95% confidence interval)a Ulcerative colitis (≤25 years) odds ratio (95% confidence interval)a NO2 SO2 PM10 NO2 SO2 PM10 Minimum registry in the THIN database of 5 years before diagnosis 2.46 (1.29–4.69) 1.21 (072–2.05) 1.82 (1.02–3.26) 1.00 (0.58–1.73) 2.12 (1.14–3.94) 0.97 (0.56–1.67) IBD diagnosis after 1 July 2006 2.45 (1.18–5.10) 1.72 (0.91–3.26) 2.08 (1.04–4.14) 1.24 (0.61–2.50) 2.69 (1.20–6.04) 1.13 (0.56–2.28) Fifth/fourth vs. first/second quintileb 2.48 (1.29–4.75) 1.12 (0.64–1.97) 1.71 (0.95–3.08) 1.08 (0.56–1.96) 1.47 (0.75–2.88) 1.06 (0.59–1.88) No immunosuppressant/biologic within 3 months of case/control index datec 2.09 (1.05–4.17) 1.16 (0.65–2.07) 1.57 (0.83–2.97) 0.93 (0.52–1.64) 1.91 (1.02–3.60) 0.93 (0.52–1.66) Restricted to cases and controls living in urban centersd 3.27 (1.32–8.06) 1.20 (0.68–2.12) 1.41 (0.72–2.78) 0.94 (0.48–1.83) 1.97 (0.94–4.10) 0.89 (0.44–1.78) IBD, inflammatory bowel disease; NO2, nitrogen dioxide; NSAID, non-steroidal anti-inflammatory drug; PM10, particulate matter <10 μm; SO2, sulfur dioxide; THIN, the health improvement network.

living in urban centersd 3.27 (1.32–8.06) 1.20 (0.68–2.12) 1.41 (0.72–2.78) 0.94 (0.48–1.83) 1.97 (0.94–4.10) 0.89 (0.44–1.78) IBD, inflammatory bowel disease; NO2, nitrogen dioxide; NSAID, non-steroidal anti-inflammatory drug; PM10, particulate matter <10 μm; SO2, sulfur dioxide; THIN, the health improvement network. a Matched for age and sex and adjusted for smoking, socioeconomic status, prior appendectomy, and NSAID use. Air pollution levels were stratified into high exposure (third, fourth, and fifth quintiles) and low exposure (referent). b Cases and controls living in wards within the third quintile level were excluded. c Immunosuppressant or biologic agent defined as either azathioprine, 6-mercaptopurine, methotrexate, infliximab, or adalimumab. d Urban areas were defined as settlements with a population size exceeding 10,000 persons.

INTRODUCTION Ulcerative colitis (UC) is a chronic inflammatory disease of the colon characterized by bloody diarrhea and abdominal pain. Despite recent advances in the understanding of the genetics, immune and inflammatory mechanisms, as well as potential environmental factors that contribute to the disease, an exact pathogenesis remains elusive. Hence, the treatment is aimed at modifying the pathogenic mechanisms involved, mostly by using anti-inflammatory drugs such as mesalazine, corticosteroids, immunosuppressant agents, or biologics (1). Recently, modulation of the gut flora has been suggested as an approach to manage UC. The role of microbiome in inflammatory bowel disease is clearly supported by many experimental observations. Gut flora can be modified either by antibiotics or by probiotics. Antibiotics are not good candidates for patients with chronic disorders because of antibiotic resistance, potential side effects, and ecological concerns. Probiotics have proven to be effective in the management of pouchitis (3,4), and preliminary data are available for the treatment of UC (5,6), but strong data are still lacking in both UC and Crohn's disease. In particular, there is limited evidence that probiotics, in addition to standard therapy, may provide benefits in terms of reduction of disease activity in patients with mild to moderately active UC because of a lack of well-designed, large, randomized, placebo-controlled trials (7).

ng in both UC and Crohn's disease. In particular, there is limited evidence that probiotics, in addition to standard therapy, may provide benefits in terms of reduction of disease activity in patients with mild to moderately active UC because of a lack of well-designed, large, randomized, placebo-controlled trials (7). The present study has been conducted with VSL#3, a product that has proven to be effective for the treatment and prevention of pouchitis (3). The aim of this investigation was to assess whether, by adding VSL#3 to the current standard treatment of patients with mild-to-moderate UC, it would be possible to decrease the ulcerative colitis disease activity index (UCDAI) score by at least 50% and improve some of the symptoms associated with UC. Positive results would encourage a new approach in managing UC patients to avoid or delay step-up therapies with drugs burdened by potentially serious side effects. METHODS A multicenter, double-blind, randomized, placebo-controlled, parallel study was conducted in a population of UC patients with relapsing disease of mild-to-moderate severity. We defined “relapsing mild-to-moderate UC” as a disease showing symptomatic recurrence after at least 6 months of remission (8), with a new increase in UCDAI (see Table 1) of at least three points (between three and eight) (9). The protocol was approved by the Investigational Review Board of each center. All patients gave written informed consent for their participation.

We defined “relapsing mild-to-moderate UC” as a disease showing symptomatic recurrence after at least 6 months of remission (8), with a new increase in UCDAI (see Table 1) of at least three points (between three and eight) (9). The protocol was approved by the Investigational Review Board of each center. All patients gave written informed consent for their participation. Sample size The sample size was based on a power of 80% and a statistical significance (α) of 95% (P=0.05). This calculation was based on the assumption that a response to treatment at 8 weeks, such as with oral 5-aminosalicylic acid (ASA) preparations, was expected to occur in 71% of patients treated with VSL#3 compared with a 40% expected response for patients treated with placebo. This assumed that the probiotic is as effective as oral 5-aminosalycilic acid. Therefore, 59 patients were required in each group, with an additional 15% for dropouts and 5% for patients failing to undergo final endoscopic assessment; hence a total of 144 patients were planned for the trial. Study procedures The study procedures were conducted for each patient enrolled in the study. At the screening visit, each patient's demographic characteristics, medical history, and current medications were recorded. β-Chorionic gonadotropin hormone was also assessed in women of childbearing age and was collected and analyzed to exclude pregnancy.

Study procedures The study procedures were conducted for each patient enrolled in the study. At the screening visit, each patient's demographic characteristics, medical history, and current medications were recorded. β-Chorionic gonadotropin hormone was also assessed in women of childbearing age and was collected and analyzed to exclude pregnancy. Eligible patients were randomly assigned to receive either VSL#3 or placebo twice daily for 8 weeks. The study product, VSL#3, was provided in plastic sealed individual dose sachets. Placebo was supplied in identical sachets. Patients were asked to take the contents of the sachets in the morning and evening. Individual disease activity quantified by the patient's UCDAI was calculated. The UCDAI was calculated by the investigator, who added the individual scores of the four parameters (bowel frequency, rectal bleeding, endoscopic score, and physician's rating of severity). At each visit, a detailed physical examination and history were performed. All adverse events were documented, classified, and graded. Study participants were supplied with diary cards to assess and record their symptoms (stool frequency, bleeding, and abdominal pain) on a daily basis. Participants' compliance was assessed by the investigators, who counted the unused sachets that the patients were requested to bring back at week 8.

sified, and graded. Study participants were supplied with diary cards to assess and record their symptoms (stool frequency, bleeding, and abdominal pain) on a daily basis. Participants' compliance was assessed by the investigators, who counted the unused sachets that the patients were requested to bring back at week 8. Inclusion criteria Patients had to meet all the inclusion criteria described in Table 2 to be eligible for participation. Moreover, women who had a negative pregnancy test at the screening visit and agreed to use a valid contraceptive method for the duration of the study, as well as patients not requiring hospitalization and patients willing and able to provide written informed consent, were considered eligible for inclusion in the study. Exclusion criteria Patients who met any of the exclusion criteria as described in Table 3 were not enrolled in this study. Significant hepatic, renal, endocrine, respiratory, neurological, or cardiovascular diseases, as determined by the investigator, were also considered as exclusion criteria. Other exclusion criteria that were also taken into consideration included the following: a history of severe adverse reaction or known hypersensitivity to maltose and/or silicon dioxide; patients requiring hospitalization; use of any investigational drug and/or participation in any clinical trial within 3 months before entering this study; inability to give a valid written informed consent or to properly follow the protocol.

Significant hepatic, renal, endocrine, respiratory, neurological, or cardiovascular diseases, as determined by the investigator, were also considered as exclusion criteria. Other exclusion criteria that were also taken into consideration included the following: a history of severe adverse reaction or known hypersensitivity to maltose and/or silicon dioxide; patients requiring hospitalization; use of any investigational drug and/or participation in any clinical trial within 3 months before entering this study; inability to give a valid written informed consent or to properly follow the protocol. Treatment Patients meeting the inclusion criteria were randomly assigned to one of the two groups of treatment and received the product for 8 weeks in addition to their standard pharmaceutical therapy (5-ASA and/or immunosuppressant). VSL#3 consists of sachets, each containing 900 billion viable lyophilized bacteria, comprising four strains of lactobacilli (L. paracasei, L. plantarum, L. acidophilus, and L. delbrueckii subsp bulgaricus), three strains of bifidobacteria (B. longum, B. breve, and B. infantis), and one strain of Streptococcus thermophilus (VSL Pharmaceuticals, MD). The daily dose was two sachets twice a day taken orally (3,600 billion bacteria per day). The patient was asked to mix the contents of the sachets in a glass of cold water or in yogurt. Hot beverages were excluded, as an elevated temperature may inactivate the bacteria. The placebo was in the form of identical sachets that did not contain any lyophilized bacteria.

en orally (3,600 billion bacteria per day). The patient was asked to mix the contents of the sachets in a glass of cold water or in yogurt. Hot beverages were excluded, as an elevated temperature may inactivate the bacteria. The placebo was in the form of identical sachets that did not contain any lyophilized bacteria. Concomitant treatments Patients who were taking maintenance oral 5-ASA and/or azathioprine or 6-mercapropurine continued to do so at stable doses. The 5-ASA doses had to be fixed for 4 weeks and azathioprine or 6-mercaptopurine doses were fixed for at least 3 months before study entry, and had to be maintained at the same dose throughout the study. Any change in dosing of oral 5-ASA or in dosing of oral 6-mercaptopurine and azathioprine drugs throughout the 8-week study period was considered as a protocol violation. Rectally administered medications, steroids, antibiotics, probiotics, and antidiarrheal drugs were not allowed, nor were any fruits, vegetables, milk, or fresh milk by-products. VSL#3 supplementation had to be interrupted for a minimum of 14 days before inclusion in the study. Primary end point The primary end point was the evaluation of the beneficial effects of food supplementation with VSL#3 in relapsing mild-to-moderate UC patients, assessed by a decrease in the UCDAI of 50% or more, from baseline to week 8. Secondary end points Secondary end points were the possible beneficial effects of VSL#3 on the following: activity of relapsing UC; remission, considered as UCDAI ≤2, assessed at week 8;

Primary end point The primary end point was the evaluation of the beneficial effects of food supplementation with VSL#3 in relapsing mild-to-moderate UC patients, assessed by a decrease in the UCDAI of 50% or more, from baseline to week 8. Secondary end points Secondary end points were the possible beneficial effects of VSL#3 on the following: activity of relapsing UC; remission, considered as UCDAI ≤2, assessed at week 8; improvement in endoscopic scores, assessed by the endoscopic subgroup score of the UCDAI at week 8; change in objective symptoms (rectal bleeding and stool frequency); change in subjective symptoms (physician rating of disease activity); lack of beneficial effects, defined by the need for pharmacological treatment or inability to remain on the study regimen until week 8. Randomization Each center enrolled patients according to the randomization list. Patients who fulfilled the eligibility criteria specified above were randomly assigned to receive VSL#3 or placebo in a random order, using only one randomization list. The randomization number was strictly given according to the order of the patient's enrollment, assigning each patient the first available number on the randomization list. The randomization number, or the reason for not enrolling the patient, was reported for each patient in the appropriate forms. Randomization was carried out in a double-blind manner in blocks of four patients using 1:1 allocation to the two groups. Assessment of compliance The investigators assessed compliance by checking the number of unused sachets that the patients brought back at each visit.

Randomization Each center enrolled patients according to the randomization list. Patients who fulfilled the eligibility criteria specified above were randomly assigned to receive VSL#3 or placebo in a random order, using only one randomization list. The randomization number was strictly given according to the order of the patient's enrollment, assigning each patient the first available number on the randomization list. The randomization number, or the reason for not enrolling the patient, was reported for each patient in the appropriate forms. Randomization was carried out in a double-blind manner in blocks of four patients using 1:1 allocation to the two groups. Assessment of compliance The investigators assessed compliance by checking the number of unused sachets that the patients brought back at each visit. Statistical assessment Baseline characteristics of patients were compared using Student's t-test for independent samples or Pearson's χ2-test as appropriate. Values of P≥0.05 were considered statistically significant. Pearson's χ2-test was used to compare the UCDAI score at each visit with the basal visit score after adjustment of data using the last-observation-carried-forward method. Comparison of stool frequency, rectal bleeding, and mucosal appearance at each time between treatment groups and at each visit vs. the basal value was performed using Pearson's χ2-test. The 95% confidence interval (CI) was also assessed.

score after adjustment of data using the last-observation-carried-forward method. Comparison of stool frequency, rectal bleeding, and mucosal appearance at each time between treatment groups and at each visit vs. the basal value was performed using Pearson's χ2-test. The 95% confidence interval (CI) was also assessed. A multivariate analysis was also performed. The general linear model multivariate procedure is based on a general linear model in which factors and covariates are assumed to have linear relationships to the dependent variables. As dependent variables, we chose UCDAI overall response at visit three (increase of 50% or more in the UCDAI score compared with the screening score) and disease extension at visit three (left-sided colitis, distal colitis, pancolitis). Fixed factors categorical predictors were selected as factors in the model (treatment with placebo or VLS#3, and concomitant treatment with or without the combination of 5-ASA and immunosuppressors). The general linear model multivariate procedure assumes that all model factors are fixed, i.e., they are generally thought of as variables, the values of interest of which are all represented in the data file, usually by design.

and concomitant treatment with or without the combination of 5-ASA and immunosuppressors). The general linear model multivariate procedure assumes that all model factors are fixed, i.e., they are generally thought of as variables, the values of interest of which are all represented in the data file, usually by design. The statistical analysis of all the data sets pertaining to efficacy (specifically, primary and secondary end points) and safety (specifically, serious adverse events as defined by federal guidelines) has been independently performed by a biostatistician who is not employed by the corporate entity. The corresponding author had full access to all data and takes full responsibility for the veracity of the data and analysis. RESULTS Participant flow A total of 144 patients (71 in the VSL#3 group and 73 in the placebo group) were enrolled. No patient was withdrawn before treatment assignment (see Figure 1). Baseline data The clinical characteristics of patients in the two groups were comparable ( Table 4). No significant differences were identified in terms of demographic characteristics (mean age, male–female ratio, weight, height, and mean UCDAI). Clinical response The main clinical outcomes of the study according to per-protocol (PP) and intention-to-treat (ITT) methods are shown in Table 5.

Baseline data The clinical characteristics of patients in the two groups were comparable ( Table 4). No significant differences were identified in terms of demographic characteristics (mean age, male–female ratio, weight, height, and mean UCDAI). Clinical response The main clinical outcomes of the study according to per-protocol (PP) and intention-to-treat (ITT) methods are shown in Table 5. Six patients in the VSL#3 group withdrew during the follow-up, two had protocol violations (these patients took beclometasone dipropionate and prednisone), two withdrew their informed consent, and three were lost to follow-up. Among the seven patients in the placebo group who withdrew during the follow-up, five patients experienced a worsening of symptoms, one was lost to follow-up, and one withdrew informed consent (see Table 6). Primary end point Overall, VSL#3 was significantly superior to placebo in reducing the disease activity of mild-to-moderate UC. Significantly more patients in the VSL#3 group experienced an improvement in their UCDAI score of at least 50% at the end of 8 weeks than those who received the placebo (41 (63.1%) vs. 29 (40.8%), respectively; PP P=0.010, CI95% 0.51–0.74; ITT P=0.031, CI95% 0.47–0.69) (see Figure 2).

e activity of mild-to-moderate UC. Significantly more patients in the VSL#3 group experienced an improvement in their UCDAI score of at least 50% at the end of 8 weeks than those who received the placebo (41 (63.1%) vs. 29 (40.8%), respectively; PP P=0.010, CI95% 0.51–0.74; ITT P=0.031, CI95% 0.47–0.69) (see Figure 2). To evaluate a more homogeneous set of patients, we also excluded patients who were under immunosuppressive treatment from the final evaluation. However, no statistical difference was found because VSL#3 was still significantly better in improving UCDAI scores of at least 50% at the end of 8 weeks than placebo (37 (56.1%) vs. 25 (36.2%), respectively; PP P=0.008; ITT P=0.025). Secondary end points Similarly, a significantly higher number of patients in the VSL#3 group had a decrease of three or more points in their UCDAI score from baseline to week 8 than the placebo group (39 (60%) vs. 29 (43.94%), respectively; PP P=0.017, CI95% 0.51–0.74; ITT P=0.046, CI95% 0.47–0.69) (see Figure 2). Regarding the induction of remission, 31 (47.7%) patients in the VLS#3 group and 23 (32.4%) patients in the placebo group experienced remission by the end of 8 weeks; although a Δ value of 15.3% was observed, this difference was not statistically significant (PP P=0.069, CI95% 0.36–0.60; ITT P=0.132, CI95% 0.33–0.56) (see Figure 2). None of the parameters assessed in the multivariate analysis was found to have a significant role in influencing remission.

ssion by the end of 8 weeks; although a Δ value of 15.3% was observed, this difference was not statistically significant (PP P=0.069, CI95% 0.36–0.60; ITT P=0.132, CI95% 0.33–0.56) (see Figure 2). None of the parameters assessed in the multivariate analysis was found to have a significant role in influencing remission. To evaluate a more homogeneous set of patients, we also excluded patients under immunosuppressive treatment from the final evaluation. However, no difference was found because VSL#3 was still better in obtaining remission at the end of 8 weeks than placebo, and the result did not reach statistical significance (28 (42.4%) vs. 20 (29%), respectively; PP P=0.067; ITT P=0.110). It is interesting that Tables 7 and 8 show that none of the patients in the VSL#3 group experienced a worsening of symptoms during the follow-up, whereas several patients in the placebo group showed a worsening of symptoms, and five of them had to be withdrawn from the study. Patients receiving VSL#3 had a significant reduction in rectal bleeding (PP P=0.014, CI95% 0.46–0.70; ITT P=0.036, CI95% 0.41–0.65). On the other hand, we did not find any significant difference in stool frequency (PP P=0.202, CI95% 0.39–0.63; ITT P=0.229, CI95% 0.35–0.57), physician's rating of disease activity (PP P=0.088, CI95% 0.34–0.58; ITT P=0.168, CI95% 0.31–0.53), or mean endoscopy scores (PP P=0.086, CI95% 0.74–0.92; ITT P=0.366, CI95% 0.66–0.86) (see Figure 3).

did not find any significant difference in stool frequency (PP P=0.202, CI95% 0.39–0.63; ITT P=0.229, CI95% 0.35–0.57), physician's rating of disease activity (PP P=0.088, CI95% 0.34–0.58; ITT P=0.168, CI95% 0.31–0.53), or mean endoscopy scores (PP P=0.086, CI95% 0.74–0.92; ITT P=0.366, CI95% 0.66–0.86) (see Figure 3). Safety and tolerability No major adverse event was reported. Eight patients on VSL#3 (11.2%) reported mild side effects (one patient reported dizziness, one reported a flu-like syndrome, and six initially complained of abdominal bloating and discomfort), whereas nine patients on placebo (12.3%) reported mild side effects (one reported a fever, one had cystitis, three had abdominal bloating, and four patients had an unpleasant taste in their mouth). DISCUSSION UC is a chronic inflammatory disease of the colon involving still largely unknown interactions between genetic, environmental, and immunological factors. UC is characterized by flare-ups of inflammation and periods of remission or quiescence that can be achieved or maintained by drugs having, as a common denominator, anti-inflammatory and/or immunosuppressive properties (5-aminosalicylates, 6-mercaptopurine, azathioprine, and anti-TNFα antibodies). If left without any maintenance drug, about 70% of patients will relapse within 12 months (2), and many patients on maintenance drugs will still require step-up therapy.

mon denominator, anti-inflammatory and/or immunosuppressive properties (5-aminosalicylates, 6-mercaptopurine, azathioprine, and anti-TNFα antibodies). If left without any maintenance drug, about 70% of patients will relapse within 12 months (2), and many patients on maintenance drugs will still require step-up therapy. After the initial report by Gionchetti et al. (3) on pouchitis, followed by other confirmatory clinical studies, it is now accepted that VSL#3, a combination of probiotic bacteria, can place this disease in remission or quiescence in a large number of patients with a J-pouch, as recommended in the guidelines of international gastroenterological associations (10,11). We report the results of an Italian multicenter study aimed at evaluating the efficacy of the specific probiotic product, VSL#3, for the treatment of mild-to-moderate UC used in conjunction with standard treatment. Our study is a double-blind randomized placebo-controlled trial on adult patients affected by relapsing mild-to-moderate UC, in which VSL#3 or placebo was added to the standard treatment, and aimed to assess the decrease in UCDAI score of 50% or more. For ethical reasons, the “placebo” group was a group in which the patients continued to take their standard treatment (5-ASA and/or immunosuppressant), with the simple addition of a placebo.

C, in which VSL#3 or placebo was added to the standard treatment, and aimed to assess the decrease in UCDAI score of 50% or more. For ethical reasons, the “placebo” group was a group in which the patients continued to take their standard treatment (5-ASA and/or immunosuppressant), with the simple addition of a placebo. Overall, VSL#3 was significantly superior to the placebo in reducing the activity of mild-to-moderate UC (primary end point). A significantly higher proportion of patients in the VSL#3 group experienced an improvement in their UCDAI score of at least 50% at week 8 over those who received placebo (63.1% vs. 40.8%, P=0.010). As a secondary end point, 31 individuals (47.7%) in the VLS#3 group and 23 individuals (32.4%) in the placebo group experienced remission by the end of 8 weeks, reaching results that did not show a significant difference (PP P=0.069; ITT P=0.132). We believe that this might represent a type II error and that a larger study might have had enough power to detect a statistically significant difference. None of the patients in the VSL#3 group experienced any worsening of symptoms during follow-up ( Tables 6 and 7), whereas five individuals in the placebo group showed a deterioration in their clinical status and had to be withdrawn from the study. No significant difference in stool frequency, physician rating of disease activity, and mean endoscopy scores was detected between the two groups (P= n. s. (not significant)). However, VSL#3 patients had a significant reduction in rectal bleeding compared with the placebo group (PP P=0.014; ITT P=0.036). Finally, no major adverse event was reported in either group. To confirm the efficacy of VSL#3, we also considered the patients who dropped out because of clinical ineffectiveness. In the “placebo” group, five patients abandoned the study for this specific reason (7%), whereas all VSL#3 patients completed the study.

ration-dependent agent, higher doses may increase eradication rates. Although the optimal dose of levofloxacin is yet to be determined, our study demonstrated efficacy using the lower dose of 250 mg once daily. Considering the LOAD regimen was well tolerated overall, an increase in levofloxacin dosage may be warranted. Doxycycline, a tetracycline analog, has bacteriostatic properties through the inhibition of bacterial protein synthesis (32). The current literature indicates there is little to no resistance to this class (2,9,11), thus supporting the use of tetracyclines as superior alternatives for rescue therapies (33,34). Although tetracycline is traditionally dosed four times a day, doxycycline offers the advantages of once- or twice-daily dosing and thus potentially improving compliance and tolerability (32). In our study, LOAD therapy takes advantage of the prolonged half-life of doxycycline by using a single 100 mg dose at dinner. Dosing doxycycline only once a day may help to increase tolerability while decreasing pill burden, but, as in the case with levofloxacin, it is possible that twice-daily dosing may improve clinical outcomes.

). Finally, no major adverse event was reported in either group. To confirm the efficacy of VSL#3, we also considered the patients who dropped out because of clinical ineffectiveness. In the “placebo” group, five patients abandoned the study for this specific reason (7%), whereas all VSL#3 patients completed the study. VSL#3 has proven to be effective by colonizing the host, changing the epithelial function and the immune response. Experimentally, in murine models of colitis, VSL#3 prevents redistribution and reduced expression of sealing tight-junction proteins (12) and specifically stimulates the expression of genes associated with lipid, xenobiotic, and peroxisome proliferator-activated receptor signaling (13). The roles of probiotics in managing active UC have also been reported in literature. Studies have reported Escherichia coli 1917 Nissle to be as effective as low-dose mesalamine in preventing a relapse of quiescent UC (14,15,16) and treatment with Saccharomyces boulardii for 4 weeks was shown to induce clinical remission in 71% of patients with mild-to-moderate disease; however, very few patients were enrolled to draw any conclusions (17). Moreover, S. boulardii should be managed with caution, especially in immunocompromised patients (e.g., in patients under immunosuppressant treatment) (18).

eeks was shown to induce clinical remission in 71% of patients with mild-to-moderate disease; however, very few patients were enrolled to draw any conclusions (17). Moreover, S. boulardii should be managed with caution, especially in immunocompromised patients (e.g., in patients under immunosuppressant treatment) (18). Other studies have reported the efficacy of VSL#3 in patients affected by UC (19). An open-label study (20) showed that in 5-ASA allergic or nonresponsive UC patients, VSL#3 was able to colonize the intestine and suggested that the product may be useful in maintaining remission (15 out of 20 patients remained in remission during the 1-year study). Thereafter, an open-label study found that 77% of mild-to-moderate UC patients obtained remission with 3,600 billion CFU/day of VSL#3 at 6 weeks (6). An Italian randomized, controlled study found that VSL#3 900 billion CFU/day added to low-dose balsalazide shows better results in treating active UC than balsalazide or mesalazine alone (5). Two studies with VSL#3 in pediatric UC have recently been carried out; the first one is an open-label study showing that 56% of pediatric patients obtained remission, with a combined remission/response rate of 61% (21), and the second one is a double-blind placebo-controlled trial, showing that VSL#3 supplementation was only able to induce remission in 92.8% of UC children compared with 36.4% with steroid alone, and was effective in maintaining remission in 78.6% of patients during a 12-month follow-up compared with 26.7% in the placebo group (22).

one is a double-blind placebo-controlled trial, showing that VSL#3 supplementation was only able to induce remission in 92.8% of UC children compared with 36.4% with steroid alone, and was effective in maintaining remission in 78.6% of patients during a 12-month follow-up compared with 26.7% in the placebo group (22). A recent Indian multicenter placebo-controlled trial investigating VSL#3 in mild-to-moderate UC patients was published (23). Patients were given 3,600 billion CFU/day VSL#3 for 12 weeks. At week 6, the percentage of patients with an improvement in UCDAI >50% was significantly higher in the group given VSL#3 (25, 32.5%) than in the placebo group (7, 10% P=0.001). At week 12, 42.9% of VSL#3 patients achieved remission, compared with 15.7% of placebo patients (P<0.001). Furthermore, significantly more number of patients given VSL#3 (40, 51.9%) achieved a UCDAI decrease of more than three points, compared with those given placebo (13, 18.6% P<0.001).

lacebo group (7, 10% P=0.001). At week 12, 42.9% of VSL#3 patients achieved remission, compared with 15.7% of placebo patients (P<0.001). Furthermore, significantly more number of patients given VSL#3 (40, 51.9%) achieved a UCDAI decrease of more than three points, compared with those given placebo (13, 18.6% P<0.001). Although the design of our study was similar, we recorded a higher placebo response compared with the Sood et al. (23) study (40% in our trial vs. 10% in Indian trial). The high “placebo” response rate of our study (40.8% of placebo patients had a 50% reduced UCDAI) may be easily explained by the continuous standard medical treatment provided to all the patients and allows for the statistically borderline results reached in this study for obtaining remission and mucosal healing. A possible suggestion for future studies, in addition to increasing the number of enrolled patients, may be to extend the study period to 12 weeks, expecting, as the Sood et al. (23) study proved, that a longer treatment with VSL#3 will offer more divergence from the placebo group. As stated by a recent review, another possible explanation for this high “placebo” response is that the country in which the study is conducted significantly influences the placebo response rate (24). In particular, studies carried out exclusively in Europe have a significantly higher placebo remission rate than studies outside Europe, ranging from 20.8% to 33.6% (24). Our placebo results are therefore in line with the literature estimates. This high percentage of placebo response may also account for some results of this study. For example, the failure to improve stool frequency vs. placebo may be very relevant to patients. We found VSL#3 better than placebo when we assessed the objective parameters (UCDAI, rectal bleeding, remission, and mucosal healing). On the contrary, subjective parameters (stool frequency and physician rating of disease activity) do not seem to improve so significantly under VSL#3 treatment. Two reasons may explain these conflicting results. First, the “placebo” response may affect some subjective parameters (e.g., stool frequency). The second is that unchanged stool frequency may be related to overlapping irritable bowel syndrome, as this sometimes affects patients with inflammatory bowel disease (25).

t. Two reasons may explain these conflicting results. First, the “placebo” response may affect some subjective parameters (e.g., stool frequency). The second is that unchanged stool frequency may be related to overlapping irritable bowel syndrome, as this sometimes affects patients with inflammatory bowel disease (25). An important point of discussion to be addressed is the rationale of this study. People may argue that a higher dose of 5-ASA therapy might be just as well tolerated and may be more convenient and less expensive for obtaining remission. This may be a rational and advisable approach. However, we need a new therapeutic approach to relapsing UC, especially when the patient is already under treatment with immunosuppressors. Increased doses of mesalazine formulations may be safe and effective in obtaining remission, but the azo-bonded formulations may be compromised by secretory diarrhea at doses providing >2–2.4 g/day of mesalazine (26). Moreover, biologics are at higher risk of severe side effects and are much more expensive than a high-dose probiotic treatment in obtaining remission in relapsing UC. On the contrary, VSL#3 is classified as a food or food supplement in most countries and is characterized by a very high safety profile that has also been confirmed throughout this study. The safety of VSL#3 has also been proven in pediatric inflammatory bowel disease and intensive care unit patients (21,22,27).

ing UC. On the contrary, VSL#3 is classified as a food or food supplement in most countries and is characterized by a very high safety profile that has also been confirmed throughout this study. The safety of VSL#3 has also been proven in pediatric inflammatory bowel disease and intensive care unit patients (21,22,27). Of course, once remission has been obtained, physicians also need to know how these patients should be managed in the longer term, i.e., with maintenance doses of probiotic. A clinical trial assessing the optimal dose of VSL#3 in maintaining remission of UC is needed. Another criticism may be that the VSL#3 dose used in this study is quite high, compared with other studies reporting an effect on remission of UC or pouchitis (7). This choice was based on the assumption that a high probiotic concentration is needed to treat an extensive and active colonic disease. Of course, the optimal dose to maintain remission may be much lower (e.g., one sachet daily for the maintenance of remission in pouchitis (3), and, as stated, a further trial assessing the optimal dose of VSL#3 in maintaining remission of UC is needed.

tration is needed to treat an extensive and active colonic disease. Of course, the optimal dose to maintain remission may be much lower (e.g., one sachet daily for the maintenance of remission in pouchitis (3), and, as stated, a further trial assessing the optimal dose of VSL#3 in maintaining remission of UC is needed. In this trial, probiotics and 5-ASA seem to have a synergistic activity. It is unclear how the association between probiotic and 5-ASA may take effect. It is possible that VSL#3 may function in synergy with, or perhaps increases, the anti-inflammatory action of 5-ASA compounds. 5-ASA compounds are potent inhibitors of several inflammatory mediators, such as leukotrienes, prostaglandins, and platelet-activating factor, all of which have roles in the pathogenesis of UC (28). In addition, 5-ASA compounds inhibit the production of interleukin-1 and free radicals and have an intrinsic antioxidant activity (29). Probiotics reduce inflammation by a number of mechanisms, including alteration of the mucosal immune system, competitive exclusion of proinflammatory pathogens, production of antimicrobial factors such as bacteriocins and other metabolites (28,30), and support of increased intestinal barrier function (31,32). At present, on the basis of what has recently been published for acetaminophen, we cannot exclude the possibility that gut bacteria may be the principal target for drugs, and that by manipulating the gut flora in the drug treatment, the outcome can be improved (33).

ort of increased intestinal barrier function (31,32). At present, on the basis of what has recently been published for acetaminophen, we cannot exclude the possibility that gut bacteria may be the principal target for drugs, and that by manipulating the gut flora in the drug treatment, the outcome can be improved (33). We do not know whether similar results could have been obtained only by increasing the 5-ASA daily dosage by up to 4 g, provided that the incidence of 5-ASA-related side effects remains unchanged regardless of whether the dose is set at 2 g or 4 g. However, independent of any economic considerations (VSL#3, being a probiotic, is not covered by insurance policies), we believe that the association between 5-ASA and VSL#3 should be preferred, even to a high-dose 5-ASA regimen or to the 5-ASA/immunosuppressant association for the treatment of UC patients with mild-to-moderate UC. Our opinion is based on the fact that, because the mammalian genome does not encode for all functions required for proper immunological responses, it is therefore evident that humans depend on critical interactions with their microbiome for health (34,35).

tion for the treatment of UC patients with mild-to-moderate UC. Our opinion is based on the fact that, because the mammalian genome does not encode for all functions required for proper immunological responses, it is therefore evident that humans depend on critical interactions with their microbiome for health (34,35). In conclusion, our study found that the addition of the high-potency probiotic mixture VSL#3 to the standard UC treatment is able to induce significant symptomatic improvement of relapsing mild-to-moderate UC compared with the placebo group on standard treatment only. This double-blind, placebo-controlled study found that VSL#3 is also able to improve the clinical picture, reduce symptoms, and improve the endoscopic appearance of the colonic mucosa. Therefore, VSL#3 may be considered as a safe and effective option for patients suffering from relapsing mild-to-moderate UC, to avoid or delay the administration of steroids, immunosuppressants, and biologics. STUDY HIGHLIGHTS We are grateful to Mrs Florence Pryen for language assistance in the revision of the paper. We also thank Dr Luciana Mosca for her critical revision of the paper. Guarantor of the article: Antonio Tursi, MD.

In conclusion, our study found that the addition of the high-potency probiotic mixture VSL#3 to the standard UC treatment is able to induce significant symptomatic improvement of relapsing mild-to-moderate UC compared with the placebo group on standard treatment only. This double-blind, placebo-controlled study found that VSL#3 is also able to improve the clinical picture, reduce symptoms, and improve the endoscopic appearance of the colonic mucosa. Therefore, VSL#3 may be considered as a safe and effective option for patients suffering from relapsing mild-to-moderate UC, to avoid or delay the administration of steroids, immunosuppressants, and biologics. STUDY HIGHLIGHTS We are grateful to Mrs Florence Pryen for language assistance in the revision of the paper. We also thank Dr Luciana Mosca for her critical revision of the paper. Guarantor of the article: Antonio Tursi, MD. Specific author contributions: Antonio Tursi conceived the study and wrote the paper. Giovanni Brandimarte, Alfredo Papa, Andrea Giglio, Walter Elisei, Gian Marco Giorgetti, Giacomo Forti, Sergio Morini, Cesare Hassan, Maria Antonietta Pistoia, Maria Ester Modeo, Stefano Rodinò, Teresa D'Amico, Ladislava Sebkova, Natale Saccà, Emilio Di Giulio, Francesco Luzza, Maria Imeneo, Tiziana Larussa, Salvatore Di Rosa, Vito Annese, Silvio Danese, and Antonio Gasbarrini conducted the study and approved the paper before submission. Walter Elisei revised the statistical analysis. Financial support: This trial was sponsored by VSL Pharmaceuticals, Towson, MD. Potential competing interests: None. Figure 1 Patient disposition.

Specific author contributions: Antonio Tursi conceived the study and wrote the paper. Giovanni Brandimarte, Alfredo Papa, Andrea Giglio, Walter Elisei, Gian Marco Giorgetti, Giacomo Forti, Sergio Morini, Cesare Hassan, Maria Antonietta Pistoia, Maria Ester Modeo, Stefano Rodinò, Teresa D'Amico, Ladislava Sebkova, Natale Saccà, Emilio Di Giulio, Francesco Luzza, Maria Imeneo, Tiziana Larussa, Salvatore Di Rosa, Vito Annese, Silvio Danese, and Antonio Gasbarrini conducted the study and approved the paper before submission. Walter Elisei revised the statistical analysis. Financial support: This trial was sponsored by VSL Pharmaceuticals, Towson, MD. Potential competing interests: None. Figure 1 Patient disposition. Figure 2 Percentage of patients with reduction of ulcerative colitis disease activity index (UCDAI) > 50% or of at least three points, and patients in remission at week 8 (on intention-to-treat analysis). n.s., not significant. Figure 3 Percentage of patients with improvement in different subgroups of ulcerative colitis disease activity index (UCDAI; rectal bleeding, stool frequency, physician rating of disease activity, and endoscopic score) at week 8 (on intention-to-treat analysis). n.s., not significant.

Figure 2 Percentage of patients with reduction of ulcerative colitis disease activity index (UCDAI) > 50% or of at least three points, and patients in remission at week 8 (on intention-to-treat analysis). n.s., not significant. Figure 3 Percentage of patients with improvement in different subgroups of ulcerative colitis disease activity index (UCDAI; rectal bleeding, stool frequency, physician rating of disease activity, and endoscopic score) at week 8 (on intention-to-treat analysis). n.s., not significant. Table 1 Ulcerative colitis (UC) disease activity index 1. Stool frequency Normal 0 1–2 Stools/day>normal 1 3–4 Stools/day>normal 2 >4 Stools/day>normal 3 2. Rectal bleeding None 0 Streaks of blood 1 Obvious blood 2 Mostly blood 3 3. Mucosal appearance Normal 0 Mild friability 1 Moderate friability 2 Exudation, spontaneous bleeding 3 4. Physician's rating of disease activity Normal 0 Mild 1 Moderate 2 Severe 3 The index assesses four variables, which include stool frequency, severity of bleeding, colonic mucosal appearance, and the physician's overall assessment of disease activity. Each variable is scored from 0–3 so that the total index score ranges from 0–12; 0–2: remission; 3–6: mild; 7–10: moderate; >10: severe UC.

Table 1 Ulcerative colitis (UC) disease activity index 1. Stool frequency Normal 0 1–2 Stools/day>normal 1 3–4 Stools/day>normal 2 >4 Stools/day>normal 3 2. Rectal bleeding None 0 Streaks of blood 1 Obvious blood 2 Mostly blood 3 3. Mucosal appearance Normal 0 Mild friability 1 Moderate friability 2 Exudation, spontaneous bleeding 3 4. Physician's rating of disease activity Normal 0 Mild 1 Moderate 2 Severe 3 The index assesses four variables, which include stool frequency, severity of bleeding, colonic mucosal appearance, and the physician's overall assessment of disease activity. Each variable is scored from 0–3 so that the total index score ranges from 0–12; 0–2: remission; 3–6: mild; 7–10: moderate; >10: severe UC. Table 2 Inclusion criteria Male and female patients aged more than 18 years; Diagnosis of UC established by previous colonoscopy, with consistent histology and clinical course; UC involving at least the rectosigmoid region; activity confirmed by colonoscopy at the beginning of the study; Mild-to-moderate relapsing UC, defined as a UCDAI score ranging from three to eight; Symptoms (relapsing episodes) for less than 4 weeks before study entry; A minimum endoscopic score of three on the UCDAI at screening (mucosal appearance); Use of oral 5-ASA at least 4 weeks before study entry at a stable dose (mesalazine at least 1.6 g/day or balsalazide at least 4.5 g/day) and/or use of azathioprine (at least 1.5 mg/kg/day) or 6-mercaptopurine (at least 1 mg/kg/day) at least 3 months before study entry at a stable dose. ASA, aminosalicylic acid; UC, ulcerative colitis; UCDAI, ulcerative colitis disease activity index.

dose (mesalazine at least 1.6 g/day or balsalazide at least 4.5 g/day) and/or use of azathioprine (at least 1.5 mg/kg/day) or 6-mercaptopurine (at least 1 mg/kg/day) at least 3 months before study entry at a stable dose. ASA, aminosalicylic acid; UC, ulcerative colitis; UCDAI, ulcerative colitis disease activity index. Table 3 Exclusion Criteria Crohn's disease or pouchitis; A UCDAI score greater than eight (need for emergency surgery or the presence of severe disease); Use of oral steroids within the last 4 weeks before study entry; Use of antibiotics within the last 2 weeks before study entry; Change in dose of oral 5-ASA within the last 4 weeks before study entry and throughout the 8-week study period or a change in dose of oral 6-mercaptopurine and azathioprine drugs within the last 3 months before the study; Use of rectal 5-ASA or steroids within 1 week before entering the study or throughout the 8-week study period; Use of probiotic preparations either prescribed or over-the-counter within 2 weeks before study entry; Use of NSAIDs for 1 week before and throughout the 8-week study period. ASA, aminosalicylic acid; NSAID, non-steroidal anti-inflammatory drug; UCDAI, ulcerative colitis disease activity index.

h sequential therapy is innovative, the regimen is also very complex, requiring mid-therapy transition from dual to triple therapy (18,19,20). Comparatively, concomitant quadruple therapy appears to be an effective, safe, and well-tolerated alternative to triple therapy and is less complex than sequential therapy (20). Our study indicates that a unique four-drug regimen (LOAD) is potentially more efficacious than the standard of care (LAC) for the treatment of HP in a treatment-naive population. Overall, patients receiving LAC had an unsatisfactory cure rate of 73.3% ITT (77.6% per protocol); this outcome is similar to those reported by others (4,8,21,22). On the other hand, the LOAD cure rate of 89.4% ITT (93.6% per protocol) meets the criteria that content experts deem acceptable (22,23). Furthermore, there was a trend toward fewer patients receiving LOAD (9.2%) therapy having HP recurrence 1 year later as compared with those receiving LAC therapy (16.7%). Notably, a majority of those positive for HP 1 year later had traveled out of the United States to areas of known high HP prevalence; in addition, it is possible that noncompliance and antibiotic resistance had roles in disease recurrence. However, these results are in line with a meta-analysis by Niv and Hazazi (24) indicating much higher HP recurrence rates in developing countries.

the 8-week study period; Use of probiotic preparations either prescribed or over-the-counter within 2 weeks before study entry; Use of NSAIDs for 1 week before and throughout the 8-week study period. ASA, aminosalicylic acid; NSAID, non-steroidal anti-inflammatory drug; UCDAI, ulcerative colitis disease activity index. Table 4 Patient demographic and baseline characteristics Characteristic VSL#3 Placebo Gender (male:female) 49:22 (69%) 44:29 (60.3%) Age in years (mean±s.d.) 47.7±14.1 46.4±14.4 Number of previous relapses (mean±s.d.) 2.24±1.05 2.37±1.04 UCDAI at entry (mean±s.d.) 5.52±1.33 5.42±1.43 Disease extent (number of patients) (%) Proctosigmoiditis 36 (50.7%) 38 (52.1%) Left-sided colitis 24 (33.8%) 21 (28.8%) Pancolitis 11 (15.5%) 14 (19.1%) Concomitant medications Mesalamine alone (mean/median±s.d.) 65 (91.55%) (2.08/2.4±0.39) 69 (94.52%) (2.08/2.4±0.40) Balsalazide alone (mean/median±s.d.) 2 (2.82%) (4.5/4.5) 2 (2.74%) (4.5/4.5) Azathioprine alone (mean/median±s.d.) 1 (1.23%) (1.62/1.5±0.25) 0 (0%) Methotrexate alone (mean/median±s.d.) 1 (1.23%) (15 mg i.m./week) 0 (0%) No medications 0 (0%) 0 (0%) Combinations of drugs Mesalamine+azathioprine (mean/median±s.d.) 2 (3.90%) (2.08/2.4±0.39)+(1.62/1.5±0.25) 2 (2.74%) (2.08/2.4±0.40)+ 1.75/1.75±0.25 Balsalazide+azathioprine 0 (0%) 0 (0%) Balsalazide+methotrexate 0 (0%) 0 (0%) Total 71 (100%) 73 (100%) i.m., intra-muscular; UCDAI, ulcerative colitis disease activity index.

inations of drugs Mesalamine+azathioprine (mean/median±s.d.) 2 (3.90%) (2.08/2.4±0.39)+(1.62/1.5±0.25) 2 (2.74%) (2.08/2.4±0.40)+ 1.75/1.75±0.25 Balsalazide+azathioprine 0 (0%) 0 (0%) Balsalazide+methotrexate 0 (0%) 0 (0%) Total 71 (100%) 73 (100%) i.m., intra-muscular; UCDAI, ulcerative colitis disease activity index. Table 5 Clinical outcomes Per-protocol Intention-to-treat VSL#3 Placebo P value VSL#3 Placebo P value ≥50% Improvement in UCDAI (week 8) 41 29 0.010 41 29 0.031 >3 Decrease in UCDAI score (week 8) 39 29 0.017 39 28 0.046 Remission (week 8) 31 23 0.069 31 23 0.132 UCDAI, ulcerative colitis disease activity index. Table 6 Reasons for discontinuation of treatment VSL#3 number of patients (%) Placebo number of patients (%) Lack of efficacy 0 (0.0) 5 (6.8) Clinical episode 0 (0.0) 0 (0.0) Abnormal laboratory result 0 (0.0) 0 (0.0) Death 0 (0.0) 0 (0.0) Protocol violation 2 (1.4) 0 (0.0) Lost to follow-up 3 (4.2) 1 (1.4) Protocol interim criteria not met 0 (0.0) 0 (0.0) Patient's consent withdrawn 2 (2.8) 1 (1.4) Table 7 Overall UCDAI response after 8 weeks (per-protocol analysis) UCDAI after 8 weeks Treatment VLS#3 Placebo Overall n % n % n % None or light (0–2) 31 47.7 23 32.4 54 39.7 Mild (3–5) 27 41.5 34 47.9 61 44.9 Moderate (6–8) 7 10.8 11 15.5 18 13.2 Severe (9–12) 0 0 3 4.2 3 2.2 Overall 65 100 71 100 136 100 UCDAI, ulcerative colitis disease activity index.

weeks (per-protocol analysis) UCDAI after 8 weeks Treatment VLS#3 Placebo Overall n % n % n % None or light (0–2) 31 47.7 23 32.4 54 39.7 Mild (3–5) 27 41.5 34 47.9 61 44.9 Moderate (6–8) 7 10.8 11 15.5 18 13.2 Severe (9–12) 0 0 3 4.2 3 2.2 Overall 65 100 71 100 136 100 UCDAI, ulcerative colitis disease activity index. Table 8 Overall UCDAI response after 8 weeks (on intention-to-treat analysis) UCDAI after 8 weeks Treatment VLS#3 Placebo Overall n % n % n % None or light (0–2) 31 43.7 23 31.5 54 37.5 Mild (3–5) 30 42.3 35 47.9 65 45.1 Moderate (6–8) 10 14.1 12 16.4 22 15.3 Severe (9–12) 0 0 3 4.1 3 2.1 Overall 71 100 73 100 144 100 UCDAI, ulcerative colitis disease activity index.

INTRODUCTION Colonoscopy has been incorporated into colorectal cancer (CRC) screening programs in some countries (1,2,3). However, take-up of screening colonoscopy is limited owing to various factors that include fear of pain and the need for sedation. The alternative platforms for endoscopic imaging of the colon that have recently been presented include diagnosis-only devices such as capsule colonoscopes (4,5) and the Aer-O-Scope (GI View Ltd, Ramat Gan, Israel) (6), and therapeutic colonoscopes with alternative propulsion mechanisms such as the NeoGuide (Neoguide Systems Inc, Los Gatos, CA) (7) and the Invendo SC20 (Invendo Medical GmbH, Kissing, Germany). The Invendo SC20 is a computer-assisted colonoscope, propelled by an “inverted sleeve” mechanism and steered by means of a reusable handheld device (8). The latest version of the Invendo SC20 is CE-marked. We report a prospective trial in volunteers eligible for and willing to undergo screening colonoscopy. METHODS Participants The trial was conducted as a prospective single-arm study in November and December 2009 with the aim of including 60 participants. All the participants were people who were willing to undergo screening colonoscopy, who were at an average risk for CRC, and who were between 50 and 75 years of age. An honorarium of €350 (∼$440) was offered. The study was approved by the Ethical Committee of the Ärztekammer Hamburg (trial no. PV 3314). The inclusion criteria were: Screenee status, i.e., asymptomatic and willing to undergo screening colonoscopy, and at average risk for CRC. Age between 50 and 75 years.

METHODS Participants The trial was conducted as a prospective single-arm study in November and December 2009 with the aim of including 60 participants. All the participants were people who were willing to undergo screening colonoscopy, who were at an average risk for CRC, and who were between 50 and 75 years of age. An honorarium of €350 (∼$440) was offered. The study was approved by the Ethical Committee of the Ärztekammer Hamburg (trial no. PV 3314). The inclusion criteria were: Screenee status, i.e., asymptomatic and willing to undergo screening colonoscopy, and at average risk for CRC. Age between 50 and 75 years. Provision of signed informed consent. The exclusion criteria were: Family or personal history of colorectal neoplasia including familial adenomatous polyposis or hereditary nonpolyposis CRC. Previous colonoscopy, within preceding 10 years. Diagnosis of suspected inflammatory bowel disease, bowel obstruction, or acute diverticulitis, or known severe diverticulosis, or any known large-bowel disease. Clinically significant cardiovascular or pulmonary disease. Gastrointestinal tract-related symptoms, complaints, or diseases suggesting performance of a diagnostic colonoscopy (nonscreening cases). Cancer or other life-threatening disease or significant chronic condition. Blood-clotting disorders and/or anticoagulant therapy (anticoagulant therapy included aspirin within the previous 7 days). Known pregnancy or positive pregnancy-screening test. Previous abdominal surgery, except for uncomplicated cholecystectomy, appendectomy, or minor pelvic surgery (e.g., hernia repair, oöphorectomy).

Cancer or other life-threatening disease or significant chronic condition. Blood-clotting disorders and/or anticoagulant therapy (anticoagulant therapy included aspirin within the previous 7 days). Known pregnancy or positive pregnancy-screening test. Previous abdominal surgery, except for uncomplicated cholecystectomy, appendectomy, or minor pelvic surgery (e.g., hernia repair, oöphorectomy). Morbid obesity (body mass index >40 kg/m2). Clinically significant abnormal screening laboratory findings. Clinically significant abnormal screening electrocardiographic findings. Drug abuse or alcoholism. Inability of the screenee to communicate adequately. Being under custodial care. Participation in a clinical study within the previous 30 days. Colonoscopic examinations Two centers and four investigators (S.G., N.H., D.K.R., and T.R.) were involved in the study. These investigators were experienced in colonoscopy (3,000–30,000 lifetime experience) and two of them had done >100 colonoscopies with the SC20 colonoscope before the start of the study (N.H. and T.R.). The two other investigators received training with one (S.G.) or two (D.K.R.) procedures using the Koken Colonoscopy Training Model and reported proficiency regarding passage after two training procedures in humans (screenees). These two investigators performed nine procedures each with the Invendo SC20 before the start of the study.

nvestigators received training with one (S.G.) or two (D.K.R.) procedures using the Koken Colonoscopy Training Model and reported proficiency regarding passage after two training procedures in humans (screenees). These two investigators performed nine procedures each with the Invendo SC20 before the start of the study. The study was carried out at two sites in Germany, in Frankfurt and Hamburg. Following conventional colon lavage preparation using polyethylene glycol solution (MoviPrep, Norgine GmbH, Marburg, Germany, in split dosage), the colonoscopic examination was begun with the screenee in the left lateral position. An intravenous line was only placed if sedation or antispasmodic agents became necessary during the examination. Participants were reassured that they could receive sedation at any point during the examination if this was requested, and were repeatedly asked by the colonoscopist during the examination whether they were comfortable or wanted sedation. If sedation was requested, propofol was administered by a second endoscopist and the patient was monitored by pulse oximetry, and pulse and blood pressure measurement. Position change and application of abdominal pressure were used at the discretion of the colonoscopist. CO2 was used for insufflation in all cases. Water immersion, administered via a foot pump, was used during insertion at the discretion of the endoscopist.

by pulse oximetry, and pulse and blood pressure measurement. Position change and application of abdominal pressure were used at the discretion of the colonoscopist. CO2 was used for insufflation in all cases. Water immersion, administered via a foot pump, was used during insertion at the discretion of the endoscopist. The instrument was steered to the cecum; this was confirmed by images of the appendiceal orifice from a point proximal to the ileocecal valve and of the ileocecal valve from just distal to the valve. On introduction and especially on withdrawal (this was the main step of the examination for diagnostic activity), pathological findings such as polyps were documented and biopsied, or removed by forceps or snare if the endoscopist deemed that this was indicated and feasible. The quality of bowel preparation was rated by the investigator on a scale of 1 (excellent) to 6 (very poor) for the three segments of the colon (right colon up to mid-transverse, left colon including sigmoid, and rectum). After the examination, participants were interviewed about their general impression of the examination and they rated pain and discomfort on a visual analog scale (1=excellent to 6=very poor, unbearable). If volunteers required sedation, a rating of 6 was automatically assumed for the record immediately after the procedure. Participants were contacted again at 24 h and 7 days.

their general impression of the examination and they rated pain and discomfort on a visual analog scale (1=excellent to 6=very poor, unbearable). If volunteers required sedation, a rating of 6 was automatically assumed for the record immediately after the procedure. Participants were contacted again at 24 h and 7 days. Instrument description: the Invendo SC20 The Invendo SC20, a single-use colonoscope controlled using a handheld unit (Invendo Medical GmbH, Germany) that is not yet commercially available in the United States, has been previously described in detail (8). Briefly, the colonoscope has a working length of 210 cm; the endoscope per se is covered by a 10-mm inner sheath. The sheath is covered by double layers of an “inverted sleeve” that provides the propulsion mechanism. A propulsion (drive) connector allows a mechanical link to the inner layer of the sleeve. Before the examination, the drive connector is locked into the endoscope's external driving unit; the examination is then started (Figure 1). Eight drive wheels in the driving unit grip the inner layer of the inverted sleeve and rotate, causing the inner layer of the inverted sleeve to drive forward. The “inverted sleeve” mechanism causes the colonoscope to “grow,” at a position just 10 cm below the distal end. Similarly, when the colonoscope is being driven backward, the drive wheels rotate in the opposite direction and the endoscope “shrinks.” It can be actively pulled out if needed (e.g., if a sedation emergency occurs); no direct manual maneuvers such as rotation are possible during introduction and withdrawal; this can be compensated in some way by 190° tip rotation in all four directions.

eled out of the United States to areas of known high HP prevalence; in addition, it is possible that noncompliance and antibiotic resistance had roles in disease recurrence. However, these results are in line with a meta-analysis by Niv and Hazazi (24) indicating much higher HP recurrence rates in developing countries. The antibiotics used in this study (levofloxacin, doxycycline, and nitazoxanide) were selected because of ease of dosing (once or twice daily) and supportive data indicating these agents are highly active against HP (25,26,27). A pilot study in patients previously failing treatment successfully used a similar regimen LEND (levofloxacin, esomeprazole, nitazoxanide, and doxycycline) to achieve a comparable ITT eradication rate of 90% (27/30) (28).

eels rotate in the opposite direction and the endoscope “shrinks.” It can be actively pulled out if needed (e.g., if a sedation emergency occurs); no direct manual maneuvers such as rotation are possible during introduction and withdrawal; this can be compensated in some way by 190° tip rotation in all four directions. A handheld control unit is used to activate all the endoscopic and software functions. The endoscope tip can be flexed electrohydraulically 180° in any direction by moving a joystick on the handheld device. Otherwise, the design of the colonoscope is similar to conventional endoscopes, allowing for insufflation, rinsing, and suction. It also has a 3.1-mm working channel. Study parameters The main outcome parameters of the study were: Safety as measured by the frequency and severity of device-related adverse events. Device effectiveness as shown by cecal intubation rate. The secondary outcome parameters were: Utility of the device in the documentation and biopsy of pathological findings. Pathological findings such as polyps, inflammatory changes, and so on. These were to be biopsied and/or polypectomies carried out according to the decision of the investigator. All findings were documented with respect to type, size, and location. The histological findings from polyps and the percentage of polyps that were biopsied or removed were also to be recorded.

changes, and so on. These were to be biopsied and/or polypectomies carried out according to the decision of the investigator. All findings were documented with respect to type, size, and location. The histological findings from polyps and the percentage of polyps that were biopsied or removed were also to be recorded. Sedation was recorded as an observation. Screenees started colonoscopy without sedation and analgesia but had the freedom to choose sedation and/or analgesia at any point during the examination. The need for sedation was recorded as a percentage of all examinations. The following additional parameters were also recorded: Baseline characteristics: age, gender, height, weight, body mass index, and medical history. Vital signs before and after the procedure. Duration of introduction and withdrawal of the colonoscope. The volunteer's perception of the procedure with respect to pain/discomfort, and their general impression, as described above (1=excellent, no pain; 6=unbearable, procedure had to be stopped due to pain, on the visual analog scale). A rating of 6 was automatically given immediately after the procedure in cases where sedation was used. Screenees were asked by examiner and/or study personnel after study completion. Quality of bowel preparation as judged on a scale of 1 (excellent) to 6 (very poor).

The volunteer's perception of the procedure with respect to pain/discomfort, and their general impression, as described above (1=excellent, no pain; 6=unbearable, procedure had to be stopped due to pain, on the visual analog scale). A rating of 6 was automatically given immediately after the procedure in cases where sedation was used. Screenees were asked by examiner and/or study personnel after study completion. Quality of bowel preparation as judged on a scale of 1 (excellent) to 6 (very poor). Statistical analysis The main outcome was evaluated as the proportion of patients with successful examinations, with a 95% confidence interval. The sample size of 60 was chosen in line with the expectation of a ≥90% true success rate. Intubation time and the patients' evaluations of the procedures are presented as means and medians. Further presentation of the results is primarily descriptive; therefore, no power calculation was done. RESULTS Between 25 November 2009 and 19 December 2009, 61 volunteers were included; these were 34 men and 27 women, with a mean age of 57.5 years (range 50–70) and a mean body mass index of 26.3 kg/m2 (19.5–36.8). Of the screenees, 15 (24.6%) had previously undergone minor abdominal surgery. Thirty-four procedures were carried out in Frankfurt and 27 procedures in Hamburg. There were no drop-outs and no nonevaluable participants; the follow-up rate was 100%.

ars (range 50–70) and a mean body mass index of 26.3 kg/m2 (19.5–36.8). Of the screenees, 15 (24.6%) had previously undergone minor abdominal surgery. Thirty-four procedures were carried out in Frankfurt and 27 procedures in Hamburg. There were no drop-outs and no nonevaluable participants; the follow-up rate was 100%. The cecum was reached in 60/61 of participants (98.4, 95% confidence interval 91.2–99.9%). The rate was significantly higher than 90% (exact binomial test; one-sided P value 0.013). In the one case where the cecum was not intubated, the deepest point of advancement was the ascending colon. Water instillation was used in 29 cases. Abdominal compression and/or position change was used in approximately two-thirds of the patients that helped in further advancing the scope; in the case of failed cecal intubation, these measures did not help and a longer scope might have been necessary to reach the cecum. Systematic retroflexion (e.g., in the rectum) was not part of the study protocol; we nevertheless felt that retroflexion, which can happen inadvertently, is safe and can be easily achieved, e.g., in rectum, at flexures, and in the cecum if aimed at (examples of colon views and findings are shown in Figures 2 and 3).

m. Systematic retroflexion (e.g., in the rectum) was not part of the study protocol; we nevertheless felt that retroflexion, which can happen inadvertently, is safe and can be easily achieved, e.g., in rectum, at flexures, and in the cecum if aimed at (examples of colon views and findings are shown in Figures 2 and 3). The median time to reach the cecum was 15 min (range 7–53.5, mean 16.4). The median withdrawal time was also 15 min (range 3.5–51, mean 16.4); for the cases in whom polyp removal was performed it was a mean of 21.2 min (range 6.5–51), and for patients without findings it was a mean of 13.6 min (range 3.5–27.5). There was a nonsignificant trend for shorter times, by 2 to 3 min, with the investigators with more experience with the device (13–14 vs. 16–17 min). The mean value for quality of bowel cleansing in all participants, as subjectively assessed by the examiners, was 1.8 (range 1–5); each individual had a mean value derived from the values in the three colonic segments as described above. In two cases, a second Invendo SC20 had to be used in the same volunteer because of endoscope malfunction. In one case, there was a malfunction of the tip flexion, probably caused by the introduction of a rough-running forceps. In the other case, the working channel had been completely blocked by aspirated bowel content.

In two cases, a second Invendo SC20 had to be used in the same volunteer because of endoscope malfunction. In one case, there was a malfunction of the tip flexion, probably caused by the introduction of a rough-running forceps. In the other case, the working channel had been completely blocked by aspirated bowel content. Sedation was used in three participants (4.9%); the propofol doses used were 120, 130, and 180 mg. The mean ratings from the screenees, immediately after colonoscopy, for overall assessment and pain/discomfort were 1.6 (range 1–3) and 2.3 (range 1–6). Follow-up at 24 h and 7 days was complete for all the study participants. The mean overall ratings at 24 h and at 7 days were 1.4 and 1.3 (range 1–5). The mean pain/discomfort ratings at 24 h and at 7 days were 1.5 and 1.3 (range 1–6). Only three screenees had previous colonoscopy, 12, 17, and 35 years before, with little memory of these procedures. A total of 36 polyps were detected in 23 participants, ranging in size from 2 to 18 mm (mean 4.8 mm). Of these, 32 polyps were removed by either forceps (n=22) or snare (n=10) (Figure 4); one additional polyp, a 12-mm flat lesion, was referred for polypectomy using conventional colonoscopy. Histological investigation showed 11 low-grade adenomas (including the polyp removed conventionally), hyperplastic tissue in 12 cases, and normal colonic mucosa in 10. In two screenees, three small polyps detected on scope introduction could not be found again on withdrawal.

polypectomy using conventional colonoscopy. Histological investigation showed 11 low-grade adenomas (including the polyp removed conventionally), hyperplastic tissue in 12 cases, and normal colonic mucosa in 10. In two screenees, three small polyps detected on scope introduction could not be found again on withdrawal. There were no device-related adverse events. One occurrence of minor bleeding, after snare polypectomy of a 15-mm polyp in the sigmoid colon and insufficient high frequency current delivery during intervention, was observed (considered to be a minor adverse event, and not device related). The investigator decided to check the resection site and treat possible bleeding via conventional colonoscopy; minor postpolypectomy bleeding was successfully treated using hemoclips. There was no other in-procedure adverse event. At follow-up at 24 h and 7 days, no adverse events were detected.

nd not device related). The investigator decided to check the resection site and treat possible bleeding via conventional colonoscopy; minor postpolypectomy bleeding was successfully treated using hemoclips. There was no other in-procedure adverse event. At follow-up at 24 h and 7 days, no adverse events were detected. DISCUSSION We demonstrated that a computer-assisted therapeutic colonoscope controlled by means of a handheld unit could be advanced to the cecum in a high percentage of cases (>98%) with very few of the paid volunteers (<5%) requiring sedation. Furthermore, experienced colonoscopists could successfully and safely use the device in screenees after minimal bench training. Finally, there were no complications in this small number of patients. Thus, the device appears promising as a means of providing unsedated screening colonoscopy. The extreme flexibility of the device suggests a low-risk profile, and the easy operation by joystick suggests that nonspecialists could learn to operate the device, although safety profile and operation with nonspecialists remain to be demonstrated. The device is disposable that confers an advantage in situations where reprocessing methods are suboptimal and/or in countries where there are principal objections to endoscope reprocessing (9,10).

ts could learn to operate the device, although safety profile and operation with nonspecialists remain to be demonstrated. The device is disposable that confers an advantage in situations where reprocessing methods are suboptimal and/or in countries where there are principal objections to endoscope reprocessing (9,10). Compared with the times reported from expert centers in conventional colonoscopy, the Invendo SC20 takes longer to reach the cecum, withdrawal is slower, and interventions such as polypectomy (from our anecdotal impression) take longer to perform. Future versions of the device should improve performance in these respects. Indeed, insertion times decreased from 24 min in the first study of the device (8) to 15 min in this study. In addition, experience appears to improve performance, as the endoscopists with greater experience tended to have shorter insertion times. Experience may also improve the speed for performing polypectomy, as none of the examiners has yet carried out large numbers of polypectomies. The marked flexibility of this device allows for passing of the endoscope keeping the natural anatomical shape of the large bowel without straightening of the endoscope and the colon. In situations where the bowel shape is grossly modified or the colon is rigid (excessive scarring after previous surgery, diverticular disease, and so on) we do not have enough experience with the Invendo SC20. On the other hand, such a situation may also require use of thinner or more flexible endoscopes in the conventional setting. Almost 25% of our screenees had previous pelvic surgery without any difficulties in passage.

ious surgery, diverticular disease, and so on) we do not have enough experience with the Invendo SC20. On the other hand, such a situation may also require use of thinner or more flexible endoscopes in the conventional setting. Almost 25% of our screenees had previous pelvic surgery without any difficulties in passage. The low proportion of screenees in our study that required sedation could reflect that they were paid volunteers, although they were told they could get sedated if required. To prove a reduced sedation rate in comparison with conventional colonoscopy would require a truly comparative, preferentially randomized, and blinded study; the results of this pilot study may be used as hypothesis generating for such a study. Use of water instillation in about half of the cases and CO2 in all cases may also have made colonoscopy easier (11,12,13). In addition, the use of sedation for endoscopic procedures varies widely by country (14,15). In countries such as the United States, where nearly all those undergoing colonoscopy are sedated, the initial evaluation of the Invendo SC20 might best be performed in people who are better candidates for scheduled, unsedated conventional colonoscopy. These individuals are typically older, male, highly educated, have no abdominal pain, and have low anxiety levels (16).

all those undergoing colonoscopy are sedated, the initial evaluation of the Invendo SC20 might best be performed in people who are better candidates for scheduled, unsedated conventional colonoscopy. These individuals are typically older, male, highly educated, have no abdominal pain, and have low anxiety levels (16). The optics of the current device utilize a complementary metal–oxide–semiconductor (CMOS) chip, and although the image resolution allowed recognition of a number of polyps that was in line with expectations, the performance of the device compared with colonoscopes with high resolution or high definition optics (17,18) has not been evaluated. In general, several factors may have contributed to the fact the adenoma rate was slightly lower than known from the German screening colonoscopy registry (∼20–25%) (http://www.zi-berlin.de/cms/fileadmin/images/content/PDFs_alle/Darmkrebsfrueherk_Bericht.pdf). However, it is our impression that there is no principal limitation of the Invendo SC20 with regard to adenoma detection. Furthermore, the low case number and the fact that adenoma detection was not an explicit study aim make conclusions about the diagnostic capability of the new scope impossible. In summary, we demonstrated that a computer-assisted colonoscope controlled with a handheld unit could be advanced to the cecum in a high percentage of cases. Very few patients needed sedation. The device warrants additional investigation as a means of providing screening colonoscopy. STUDY HIGHLIGHTS

The optics of the current device utilize a complementary metal–oxide–semiconductor (CMOS) chip, and although the image resolution allowed recognition of a number of polyps that was in line with expectations, the performance of the device compared with colonoscopes with high resolution or high definition optics (17,18) has not been evaluated. In general, several factors may have contributed to the fact the adenoma rate was slightly lower than known from the German screening colonoscopy registry (∼20–25%) (http://www.zi-berlin.de/cms/fileadmin/images/content/PDFs_alle/Darmkrebsfrueherk_Bericht.pdf). However, it is our impression that there is no principal limitation of the Invendo SC20 with regard to adenoma detection. Furthermore, the low case number and the fact that adenoma detection was not an explicit study aim make conclusions about the diagnostic capability of the new scope impossible. In summary, we demonstrated that a computer-assisted colonoscope controlled with a handheld unit could be advanced to the cecum in a high percentage of cases. Very few patients needed sedation. The device warrants additional investigation as a means of providing screening colonoscopy. STUDY HIGHLIGHTS We are grateful to Dr Marx, Department of Pathology, University Hospital Hamburg-Eppendorf and Drs Falk, Heinrich, Motherby, Lange, Ihling, Technau-Ihling, and Schubert, Pathology Practice Frankfurt, for the histopathological assessment of the study cases. We also thank Ulrich Gauger for statistical support. Guarantor of the article: Thomas Rösch, MD.

We are grateful to Dr Marx, Department of Pathology, University Hospital Hamburg-Eppendorf and Drs Falk, Heinrich, Motherby, Lange, Ihling, Technau-Ihling, and Schubert, Pathology Practice Frankfurt, for the histopathological assessment of the study cases. We also thank Ulrich Gauger for statistical support. Guarantor of the article: Thomas Rösch, MD. Specific author contributions: All four authors performed and documented the examinations; S.G. and N.H. collected and analyzed data, and paper writing was by T.R. and D.K.R. Financial support: Invendo Medical supported study performance by providing instruments and all technical support as well as covering costs for study registration at Ethics Committee as well as honoraria to volunteers and statistician. Data collection after each colonoscopy by company representatives was counterchecked by examiners. All colonoscopy and histopathology reports were reviewed by authors (S.G. and N.H.). Potential competing interests: The investigators hold consultant contracts with Invendo Medical, the manufacturer of the SC20 colonoscope, the device under investigation in this study. Figure 1 The SC20 colonoscope shown in the newest version. (a) The complete device with the instrument, driving unit, and processor; (b) the tip is introduced through the driving motor; (c) tip in full flexion; (d) tip with a biopsy forceps introduced through the working channel. Figure 2 Endoscopic images of the cecal floor. (a) A rather large appendiceal orifice; (b) the cecal valve.

Figure 1 The SC20 colonoscope shown in the newest version. (a) The complete device with the instrument, driving unit, and processor; (b) the tip is introduced through the driving motor; (c) tip in full flexion; (d) tip with a biopsy forceps introduced through the working channel. Figure 2 Endoscopic images of the cecal floor. (a) A rather large appendiceal orifice; (b) the cecal valve. Figure 3 Image quality of the Invendo SC20. (a) Endoscopic image of the colonic mucosa; (b) full lumen view on retroflexion. Figure 4 Polypectomy performed with the Invendo SC20.

INTRODUCTION Helicobacter pylori (HP) is a small, Gram-negative spirochete inhabiting the mucous layer overlying the gastric epithelial cells in humans. It is the most common chronic human bacterial infection and the most common cause of gastritis, with incidence rates as high as 50% worldwide (1). Furthermore, according to the World Health Organization, HP is classified as a type 1 carcinogen and is the primary cause of peptic ulcer disease, gastric carcinoma, and mucosa-associated lymphoid tissue lymphomas (1,2,3,4,5,6,7).

nd the most common cause of gastritis, with incidence rates as high as 50% worldwide (1). Furthermore, according to the World Health Organization, HP is classified as a type 1 carcinogen and is the primary cause of peptic ulcer disease, gastric carcinoma, and mucosa-associated lymphoid tissue lymphomas (1,2,3,4,5,6,7). Current guidelines from the American College of Gastroenterology and the European Helicobacter Study Group recommend either a clarithromycin-based triple therapy (a proton pump inhibitor (PPI) plus amoxicillin and clarithromycin) or a bismuth quadruple therapy (a PPI plus bismuth, metronidazole, and tetracycline) as a standard of care in the treatment of confirmed HP infections (4,7). However, a study by Rokkas et al. (8), based on the Maastricht III guidelines, indicated that treatment with a PPI, amoxicillin, and clarithromycin regimen as first-line therapy will fail in ∼30% of patients on an intention-to-treat (ITT) basis, and treatment with PPI, bismuth, metronidazole, and tetracycline as second-line therapy will fail in another 30%, leaving ∼10% of the total patient population in need of an alternative regimen. Patient tolerability, adherence, and drug resistance rates play a major role in treatment selection and efficacy. Drug resistance has the greatest impact on clarithromycin therapy, with HP resistance rates ranging from 10 to 30% (2,9,10,11), and actual cure rates for PPI, amoxicillin, and clarithromycin triple therapy consistently falling below 80% (5,6,8,10,12). In contrast, global resistance rates to metronidazole have been reported to range from 25 to 66%, but some investigators believe resistance might be overcome with increased doses of metronidazole (2,4,9,10,11,12,13). Unfortunately, this increase in dosage can lead to the potential for intolerable side effects and pill burden, thus affecting treatment tolerability and adherence (4).

reported to range from 25 to 66%, but some investigators believe resistance might be overcome with increased doses of metronidazole (2,4,9,10,11,12,13). Unfortunately, this increase in dosage can lead to the potential for intolerable side effects and pill burden, thus affecting treatment tolerability and adherence (4). In response to increasing rates of resistance with the aforementioned three- and four-drug regimens, newer treatment regimens aimed at eradicating the organism more effectively are increasing in popularity (1,3,5,6,7,8,9,10,12). This study evaluates a novel four-drug regimen, three antibiotics and a PPI, for the eradication of HP infection in treatment-naive patients. METHODS Study patients Consecutive patients with dyspeptic symptoms undergoing upper endoscopy were recruited for study participation. To be included in the study, patients must have had HP-induced gastritis confirmed by endoscopy (four-quadrant stomach biopsy (antro–corpus–incisura); Caris Life Sciences, Irving, TX) and monoclonal stool antigen testing (Premier Platinum HPsA PLUS, Meridian Bioscience, Cincinnati, OH). Exclusion criteria were as follows: partial gastrectomy or gastric malignancy; active bleeding; age <20 years; pregnancy; prior HP treatment or infection; recent history and treatment for Clostridium difficile infection; current use of a PPI, H2 receptor antagonist, antacid, anticoagulant, or misoprostol; recent use of antibiotics (within 6 weeks), or allergy to any medication used in the study.

active bleeding; age <20 years; pregnancy; prior HP treatment or infection; recent history and treatment for Clostridium difficile infection; current use of a PPI, H2 receptor antagonist, antacid, anticoagulant, or misoprostol; recent use of antibiotics (within 6 weeks), or allergy to any medication used in the study. Study design A randomized, open-label trial was conducted to compare the tolerability and efficacy of a novel four-drug regimen in eradicating HP gastritis. The study compared the LOAD regimen (levofloxacin (Levaquin®; Ortho-McNeil-Janssen Pharmaceuticals, Raritan, NJ), omeprazole (Prilosec®; AstraZeneca Pharmaceuticals LP, Wilmington, DE), nitazoxanide (Alinia®; Romark Laboratories, LC, Tampa, FL), and doxycycline (Vibramycin®; Pfizer, New York, NY)) with the current standard of care as represented by the LAC regimen (lansoprazole (Prevacid®; Takeda Pharmaceuticals North America, Deerfield, IL), amoxicillin (Amoxil®; GlaxoSmithKline, Research Triangle Park, NC), and clarithromycin (Biaxin®; Abbott Laboratories, Abbott Park, IL)). Before randomization, patients underwent a washout period of 6 weeks from any antibiotic or PPI use, and provided written informed consent before the initiation of therapy. Eligible patients were then randomized to either LOAD therapy for 7 days (LOAD-7) or 10 days (LOAD-10; levofloxacin 250 mg with breakfast, omeprazole 40 mg before breakfast, nitazoxanide 500 mg twice daily with meals, and doxycycline 100 mg at dinner) or LAC therapy for 10 days (lansoprozole 30 mg, amoxicillin 1 g with breakfast and dinner, clarithromycin 500 mg with breakfast and dinner). Randomization was performed in blocks of three, where the assignment of treatment was drawn in a blinded manner from an envelope. Patients were instructed to return to the clinic at least 4 weeks after the cessation of therapy for repeat stool antigen testing for HP. An ad hoc analysis of those initially meeting the definition of a “cure” (HP-negative by stool antigen at least 4 weeks after the cessation of therapy) was performed. These patients were asked to return to the clinic approximately 1 year later to test for HP recurrence (recrudescence or reinfection).

en testing for HP. An ad hoc analysis of those initially meeting the definition of a “cure” (HP-negative by stool antigen at least 4 weeks after the cessation of therapy) was performed. These patients were asked to return to the clinic approximately 1 year later to test for HP recurrence (recrudescence or reinfection). Outcome assessments The primary efficacy outcome was measured as relative eradication rates between the three regimens, at least 4 weeks after the cessation of therapy, with patient tolerability and compliance addressed as secondary end points. At the cessation of HP therapy, all anti-secretory agents were discontinued and stool antigen testing for HP was performed at least 4 weeks after therapy to evaluate for complete eradication. Usage of a monoclonal antibody stool antigen test to assess HP eradication has been validated in a large meta-analysis and was approved by the United States Food and Drug Administration (14,15). During the course of therapy, all patients were contacted by phone at least four times to assess for side effects and adherence. Helicobacter pylori recrudescence or reinfection was assessed using a 13C–urea breath test (OAPI BreathTeK Otsuka Pharmaceutical, Rockville, MD) and stool antigen testing as previously described. Patients were instructed to discontinue any anti-secretory agent at least 2 weeks before follow-up testing.

s and adherence. Helicobacter pylori recrudescence or reinfection was assessed using a 13C–urea breath test (OAPI BreathTeK Otsuka Pharmaceutical, Rockville, MD) and stool antigen testing as previously described. Patients were instructed to discontinue any anti-secretory agent at least 2 weeks before follow-up testing. The study protocol was approved by the institutional review board at Long Island Jewish Hospital-Forest Hills, and performed in accordance with the Declaration of Helsinki, the International Conference on Harmonisation Good Clinical Practice Guidelines, and applicable local laws and regulations. Signed informed consent was obtained from each patient before study enrollment.

iew board at Long Island Jewish Hospital-Forest Hills, and performed in accordance with the Declaration of Helsinki, the International Conference on Harmonisation Good Clinical Practice Guidelines, and applicable local laws and regulations. Signed informed consent was obtained from each patient before study enrollment. RESULTS Clinical outcomes In total, 653 patients (N=653) were evaluated from a community referral base between February 2008 and January 2010. These patients reported symptoms of dyspepsia (n=439, 67.2%), abdominal pain (n=114, 17.5%), gastroesophageal reflux (n=89, 13.6%), early satiety (n=8, 1.2%), and loss of appetite (n=3, 0.5%). Of these patients, 270 (41.3%) were enrolled with confirmed HP gastritis or non-bleeding peptic ulcers as documented by endoscopy (four-quadrant stomach biopsies) and stool antigen testing. Overall, 278 patients (42.6%) were excluded because of recent or ongoing usage of PPI, H2 receptor antagonist, bismuth, or antacid; 79 (12.1%) patients were excluded because of recent usage of nonsteroidal anti-inflammatory drugs, Cox-II inhibitors, or misoprostol; 24 (3.7%) patients were excluded because of recent antibiotic usage; and 2 (0.3%) declined to participate. Baseline characteristics and patient demographics are further elucidated in Table 1. Approximately two-thirds of those enrolled were of Asian descent; however, there was a broad distribution of patients representing the six distinct regions of the continent.

ecent antibiotic usage; and 2 (0.3%) declined to participate. Baseline characteristics and patient demographics are further elucidated in Table 1. Approximately two-thirds of those enrolled were of Asian descent; however, there was a broad distribution of patients representing the six distinct regions of the continent. No difference in patient adherence among LOAD-10, LOAD-7, and LAC regimens was detected. Overall, 94.4% (85/90), 96.7% (87/90), and 94.4% (85/90) of patients completed the LOAD-10, LOAD-7, and LAC regimens, respectively. In the LOAD-7 and LOAD-10 groups, eight patients discontinued therapy owing to gastrointestinal distress (n=6), dizziness (n=1), or palpitations (n=1). Five patients in the LAC arm discontinued therapy secondary to a minor skin rash (n=2) or gastrointestinal distress (n=3). Follow-up data are unavailable for one patient in the LOAD-10 arm who left the country during treatment; this patient was therefore considered to be a treatment failure. As depicted in Figure 1, ITT analysis in the patient population revealed that eradication rates were 88.9% (80/90) for LOAD-10, 90% (81/90) for LOAD-7, 89.4% (161/180) for combined LOAD regimen patients, and 73.3% (66/90) for LAC. There was no significant difference between LOAD-7 and LOAD-10 in eradication rates or adverse events. When compared with LAC in the ITT population, significant differences were detected between LOAD-7, LOAD-10, and combined LOAD therapy, P=0.013, P=0.006, and P=0.0001, respectively. Per protocol eradication rates comparing LOAD-7 and LOAD-10 patients with LAC-10 patients revealed respective cure rates of 93.6% (161/172) and 77.6% (66/85), P=0.0003 (Figure 2). As depicted in Table 2, there was no difference in reported adverse events between the groups.

P=0.013, P=0.006, and P=0.0001, respectively. Per protocol eradication rates comparing LOAD-7 and LOAD-10 patients with LAC-10 patients revealed respective cure rates of 93.6% (161/172) and 77.6% (66/85), P=0.0003 (Figure 2). As depicted in Table 2, there was no difference in reported adverse events between the groups. Although eradication of HP was of primary concern, recurrence (recrudescence or reinfection) data were collected on 218 “cures” (n=152 LOAD, n=75 LOAD-7, and n=77 LOAD-10; n=66 LAC) at 1 year (range 12 to 14 months) after the completion of therapy using stool antigen and urea breath testing. Nine patients were lost to follow-up or refused retesting. The incidence of recurrence was 11.5% (25/218) for all patients, 9.2% (14/152) for LOAD, and 16.7% (11/66) for LAC, P=0.16. Indistinguishable symptoms of recurrence made delineating between true recrudescence and reinfection difficult. Of note, 80% (20/25) of the patients who were HP positive 1 year later had traveled out of the country to areas of high HP prevalence (Asia and Latin America). Most of these, 15/25 (60%), had traveled to Asia: Bangladesh (n=4), India (n=3), Kazakhstan (n=3), Nepal (n=2), Turkey (n=2), and Mongolia (n=1).

ion difficult. Of note, 80% (20/25) of the patients who were HP positive 1 year later had traveled out of the country to areas of high HP prevalence (Asia and Latin America). Most of these, 15/25 (60%), had traveled to Asia: Bangladesh (n=4), India (n=3), Kazakhstan (n=3), Nepal (n=2), Turkey (n=2), and Mongolia (n=1). DISCUSSION HP resistance to metronidazole and clarithromycin in standard therapy has led to a decline in cure rates and increased the demand for a more efficacious, yet tolerable, therapy. Although there are few regional data regarding HP resistance in our community, previous studies indicate that, in New York and the Northeast region of the United States, antibiotic resistance ranges from 9 to 15% for clarithromycin and 5 to 31% for metronidazole, and is negligible to nonexistent for amoxicillin and tetracycline (2,11). Fluoroquinolone resistance is of concern especially in areas of high fluoroquinolone use; unfortunately, regional data on HP resistance to this class of drugs are not available.

ges from 9 to 15% for clarithromycin and 5 to 31% for metronidazole, and is negligible to nonexistent for amoxicillin and tetracycline (2,11). Fluoroquinolone resistance is of concern especially in areas of high fluoroquinolone use; unfortunately, regional data on HP resistance to this class of drugs are not available. A meta-analyses including over 53,000 patients indicated the cure rate for the recommended first-line therapy is <80%, which is far below the desired 90% cure rate for any HP regimen (16,17,18). Rokkas et al. (8) highlight a greater concern—at least 10% of HP patients will fail standard first- and second-line regimens. In response to rising resistance rates, therapies including concomitant quadruple therapy and sequential therapy have emerged as alternatives to standard therapy (12,18,19,20). Although sequential therapy is innovative, the regimen is also very complex, requiring mid-therapy transition from dual to triple therapy (18,19,20). Comparatively, concomitant quadruple therapy appears to be an effective, safe, and well-tolerated alternative to triple therapy and is less complex than sequential therapy (20).

LOAD therapy takes advantage of the prolonged half-life of doxycycline by using a single 100 mg dose at dinner. Dosing doxycycline only once a day may help to increase tolerability while decreasing pill burden, but, as in the case with levofloxacin, it is possible that twice-daily dosing may improve clinical outcomes. Nitazoxanide is a thiazolide antibiotic indicated for use in adults and children for the treatment of Cryptosporidium and Giardia infections. The bactericidal activity of nitazoxanide is a result of the drug's ability to prevent anaerobic energy metabolism via inhibition of the pyruvate oxidoreductase enzyme (35). Unlike metronidazole, nitazoxanide has been shown to be non-mutagenic for HP (35) and to possess anti-vacuolating toxin activity (36). In vitro studies indicate nitazoxanide is a potent agent against HP and other anaerobes, having activity against metronidazole-resistant strains (27,36). In addition to the in vitro data, others have reported clinical success using nitazoxanide in a two-drug regimen combined with a PPI (27) or sucralfate 37 and a three-drug regimen with a PPI and amoxicillin (38). Clinical studies with nitazoxanide have used total daily doses ranging from 1,000 to 2,000 mg (27,28,37,38) when given in combination with other antibiotics, a dose of 500 mg twice daily has been successful (28,38). This lower dose of nitazoxanide is generally better tolerated and helps in regard to capping the total cost of therapy.

s with nitazoxanide have used total daily doses ranging from 1,000 to 2,000 mg (27,28,37,38) when given in combination with other antibiotics, a dose of 500 mg twice daily has been successful (28,38). This lower dose of nitazoxanide is generally better tolerated and helps in regard to capping the total cost of therapy. Current guidelines recommend a PPI as part of any standard therapy regimen in the treatment of HP (4,7). PPIs enhance therapy by effectively increasing gastric pH, which serves to enhance the activity of the antibiotics and disrupt the acidic environment preferred by HP (39). Omeprazole was selected for use because at the time this study was initiated, it was the only generic option. Comparisons of the available PPIs in the treatment of HP have been reviewed elsewhere and are beyond the scope of this paper (40).

ity of the antibiotics and disrupt the acidic environment preferred by HP (39). Omeprazole was selected for use because at the time this study was initiated, it was the only generic option. Comparisons of the available PPIs in the treatment of HP have been reviewed elsewhere and are beyond the scope of this paper (40). Although the data indicate LOAD therapy is statistically superior to the standard of care, this study has several limitations. Patients were enrolled from a single center, thus narrowing the population size and the applicability to larger populations. Although the population studied was diverse, approximately two-thirds of the patients were of Asian descent. However, the distribution of Asian patients was quite diverse as well, with patient representation from six distinct regions of the continent (South Asia, East Asia, Southeast Asia, the Middle East, Central Asia, and Russia). This unique global distribution of patients is primarily due to our location in New York, which is convenient to many ethnically distinct neighborhoods. Presumably, symptom improvement would follow HP eradication; unfortunately, in this study patients were not assessed for symptom improvement. Secondary to constraints in funding, we were unable to perform HP susceptibility and DNA testing for antibiotic resistance and differentiation of recrudescence and reinfection. Another limitation of the study is the use of a telephone survey to assess for compliance, which is probably not optimal; non-compliance may have been associated with treatment failures and recurrence. On the other hand, this type of less-stringent follow-up may serve as an indicator of expected clinical outcomes in community practice. Finally, patients receiving therapy were not blinded, potentially leaving a margin of error due to surveillance bias.

e may have been associated with treatment failures and recurrence. On the other hand, this type of less-stringent follow-up may serve as an indicator of expected clinical outcomes in community practice. Finally, patients receiving therapy were not blinded, potentially leaving a margin of error due to surveillance bias. In summary, this prospective trial demonstrates that LOAD is a highly active and well-tolerated regimen for the treatment of HP infection in treatment-naive patients. The novel LOAD regimen combines three antibiotics with a PPI, offering the advantages of good tolerability, twice-a-day dosing, low pill burden (five per day), minimal in vitro HP resistance, and superior therapeutic outcomes as compared with a standard triple-therapy regimen. A large, randomized controlled clinical trial is warranted to confirm the therapeutic superiority of this regimen as a first- or second-line treatment option over the standard recommended therapies for HP eradication. STUDY HIGHLIGHTS We thank Matthew Bardin, PharmD, BCPS, for his guidance and technical assistance. Guarantor of the article: P. Patrick Basu, MD. Specific author contributions: Principal investigator, patient recruitment, patient evaluation, data collection, and manuscript preparation: Patrick Basu; data collection, patient evaluation, and manuscript preparation: Krishna Rayapudi; data collection and patient evaluation: Tommy Pacana and NV Krishnaswamy; manuscript preparation, background research, and statistics: Molly Flynn. Financial support: None.

Specific author contributions: Principal investigator, patient recruitment, patient evaluation, data collection, and manuscript preparation: Patrick Basu; data collection, patient evaluation, and manuscript preparation: Krishna Rayapudi; data collection and patient evaluation: Tommy Pacana and NV Krishnaswamy; manuscript preparation, background research, and statistics: Molly Flynn. Financial support: None. Potential competing interests: Patrick Basu has received educational grants from Salix Pharmaceuticals and Romark Labs, and serves on the advisory boards for Roche, Vertex, and Gilead. The other authors declare no conflict of interest. Figure 1 Comparative efficacy, intent-to-treat analysis. *LAC vs. all LOAD, P=0.0001; LAC vs. LOAD-7, P=0.006; LAC vs. LOAD-10, P=0.013. There was no difference between the LOAD regimens as compared with each other. LAC, lansoprozole 30 mg, amoxicillin 1 g with breakfast and dinner, and clarithromycin 500 mg with breakfast and dinner; LOAD, levofloxacin 250 mg with breakfast, omeprazole 40 mg before breakfast, nitazoxanide (Alina®) 500 mg twice daily with meals and doxycycline 100 mg at dinner.

egimens as compared with each other. LAC, lansoprozole 30 mg, amoxicillin 1 g with breakfast and dinner, and clarithromycin 500 mg with breakfast and dinner; LOAD, levofloxacin 250 mg with breakfast, omeprazole 40 mg before breakfast, nitazoxanide (Alina®) 500 mg twice daily with meals and doxycycline 100 mg at dinner. Figure 2 Comparative efficacy, per protocol analysis. *LAC vs. all LOAD, P=0.0003; LAC vs. LOAD-7, P=0.0047; LAC vs. LOAD-10, P=0.0035. There was no difference between the LOAD regimens as compared with each other. LAC, lansoprozole 30 mg, amoxicillin 1 g with breakfast and dinner, and clarithromycin 500 mg with breakfast and dinner; LOAD, levofloxacin 250 mg with breakfast, omeprazole 40 mg before breakfast, nitazoxanide (Alina®) 500 mg twice daily with meals, and doxycycline 100 mg at dinner

egimens as compared with each other. LAC, lansoprozole 30 mg, amoxicillin 1 g with breakfast and dinner, and clarithromycin 500 mg with breakfast and dinner; LOAD, levofloxacin 250 mg with breakfast, omeprazole 40 mg before breakfast, nitazoxanide (Alina®) 500 mg twice daily with meals, and doxycycline 100 mg at dinner Table 1 Patient demographics LOAD-7 (n=90) LOAD-10 (n=90) LAC-10 (n=90) P valuea Mean age (range 22–58 years) 37 (26–58) 36 (22–48) 37 (28–52) ND Male (n=156, 57.8%) 52 (57.8%) 51 (56.7%) 53 (58.9%) ND Female (n=114, 42.2%) 38 (42.2%) 39 (43.3%) 37 (41.1%) ND Race (N=270) Pan-Asian (n=179, 66.3%) 59 61 59 ND African-American (n=36, 13.3%) 12 12 12 ND Caucasian (n=26, 9.6%) 9 8 9 ND Latino (n=29, 10.7%) 10 9 10 ND Peptic ulcer (n=35, 13%) 12 12 11 ND Gastric erosion (n=69, 25.6%) 23 22 23 ND Gastritis Regular (n=77, 28.5%) 28 26 23 ND Nodular (n=29, 10.7%) 5 10 12 ND Without IM (n=66, 24.4%) 22 22 22 ND With IM (n=98, 36.3%) 34 32 33 ND IM, intestinal metaplasia; LAC, lansoprozole 30 mg, amoxicillin 1 g with breakfast and dinner, and clarithromycin 500 mg with breakfast and dinner; LOAD, levofloxacin 250 mg with breakfast, omeprazole 40 mg before breakfast, nitazoxanide (Alina®) 500 mg twice daily with meals, and doxycycline 100 mg at dinner; ND, no difference. a Fisher's exact test for proportions, t-test for means between each group.

Table 1 Patient demographics LOAD-7 (n=90) LOAD-10 (n=90) LAC-10 (n=90) P valuea Mean age (range 22–58 years) 37 (26–58) 36 (22–48) 37 (28–52) ND Male (n=156, 57.8%) 52 (57.8%) 51 (56.7%) 53 (58.9%) ND Female (n=114, 42.2%) 38 (42.2%) 39 (43.3%) 37 (41.1%) ND Race (N=270) Pan-Asian (n=179, 66.3%) 59 61 59 ND African-American (n=36, 13.3%) 12 12 12 ND Caucasian (n=26, 9.6%) 9 8 9 ND Latino (n=29, 10.7%) 10 9 10 ND Peptic ulcer (n=35, 13%) 12 12 11 ND Gastric erosion (n=69, 25.6%) 23 22 23 ND Gastritis Regular (n=77, 28.5%) 28 26 23 ND Nodular (n=29, 10.7%) 5 10 12 ND Without IM (n=66, 24.4%) 22 22 22 ND With IM (n=98, 36.3%) 34 32 33 ND IM, intestinal metaplasia; LAC, lansoprozole 30 mg, amoxicillin 1 g with breakfast and dinner, and clarithromycin 500 mg with breakfast and dinner; LOAD, levofloxacin 250 mg with breakfast, omeprazole 40 mg before breakfast, nitazoxanide (Alina®) 500 mg twice daily with meals, and doxycycline 100 mg at dinner; ND, no difference. a Fisher's exact test for proportions, t-test for means between each group. Table 2 Reported adverse events LOAD-7 and LOAD-10 combined (n=180) LAC-10 (n=90) P valuea Headache 16 (8.9%) 11 (12.2%) ND Bloating 12 (6.7%) 9 (10%) ND Abdominal pain 15 (8.3%) 8 (8.9%) ND Belching 9 (5%) 3 (3.3%) ND Nausea 18 (10%) 9 (10%) ND Palpitations 4 (2.2%) 1 (1.1%) ND Skin rash 2 (1.1%) 2 (2.2%) ND Diarrhea 10 (5.6%) 11 (12.2%) ND Constipation 10 (5.6%) 2 (2.2%) ND Dizziness 6 (3.3%) 2 (2.2%) ND LAC, lansoprozole 30 mg, amoxicillin 1 g with breakfast and dinner, and clarithromycin 500 mg with breakfast and dinner; LOAD, levofloxacin 250 mg with breakfast, omeprazole 40 mg before breakfast, nitazoxanide (Alina®) 500 mg twice daily with meals, and doxycycline 100 mg at dinner; ND, no difference.

.3%) 2 (2.2%) ND LAC, lansoprozole 30 mg, amoxicillin 1 g with breakfast and dinner, and clarithromycin 500 mg with breakfast and dinner; LOAD, levofloxacin 250 mg with breakfast, omeprazole 40 mg before breakfast, nitazoxanide (Alina®) 500 mg twice daily with meals, and doxycycline 100 mg at dinner; ND, no difference. a Fisher's exact test for proportions.

INTRODUCTION Over the past two decades, proton pump inhibitor (PPI) therapy has been established as the standard of care for patients with gastroesophageal reflux disease (GERD) (1,2). Since the introduction of the prototype PPI omeprazole, researchers have hypothesized that elevated intragastric pH caused by blockage of proton pumps may stimulate compensatory mechanisms, leading to an increased capacity to stimulate gastric acid production after therapy is withdrawn (3,4,5). The results of studies evaluating rebound acid hypersecretion (RAHS) following withdrawal of PPI therapy are conflicting. Hunfeld et al. (6) performed the first systematic review of the literature through October 2005 to investigate whether withdrawal of PPI therapy produced RAHS, defined as an increase in basal and/or stimulated gastric acid secretion above pretreatment levels following discontinuation of PPI therapy (7). The authors identified eight studies, six conducted in healthy volunteers and two in patients with acid-related disorders. Of the eight studies—including four double-blind, randomized studies—five did not provide any evidence of RAHS when PPI therapy was withdrawn, whereas three open-label studies suggested that RAHS may occur in Helicobacter pylori-negative patients after withdrawal of 8 weeks of PPI therapy. Based on their overall evaluation, the authors concluded that there was insufficient evidence of a clinically relevant increase in acid production after PPI withdrawal, primarily because of differences in study design, methodology, patient population, and contradictory results.

ter withdrawal of 8 weeks of PPI therapy. Based on their overall evaluation, the authors concluded that there was insufficient evidence of a clinically relevant increase in acid production after PPI withdrawal, primarily because of differences in study design, methodology, patient population, and contradictory results. Since the Hunfeld review was published, data from two randomized, placebo-controlled trials conducted in healthy volunteers (8,9) have shown that withdrawal of short-term PPI therapy (4 or 8 weeks) seemed to induce acid-related symptoms. Based on their findings, the authors speculated that PPIs may aggravate symptoms in patients when therapy is withdrawn and that this symptom rebound may lead to PPI dependence. The potential to worsen symptoms in patients discontinued from PPI therapy is an important topic for clinical practice and warrants further investigation.

eir findings, the authors speculated that PPIs may aggravate symptoms in patients when therapy is withdrawn and that this symptom rebound may lead to PPI dependence. The potential to worsen symptoms in patients discontinued from PPI therapy is an important topic for clinical practice and warrants further investigation. Dexlansoprazole modified release (MR) is a PPI that produces two distinct releases of drug and has been shown to significantly extend the duration of active plasma concentrations and the percentage of time pH >4 beyond that of conventional single-release PPIs (10,11,12,13). Because the degree of RAHS is thought to be correlated with the degree of acid suppression (14), dexlansoprazole MR is a good choice for further investigation. We performed a retrospective analysis of data from four phase III registration trials of dexlansoprazole MR (15,16,17) to determine whether patients with endoscopically confirmed erosive esophagitis (EE) receiving therapy with dexlansoprazole MR or lansoprazole developed RAHS and/or a worsening of symptoms when therapy was discontinued after healing.

lysis of data from four phase III registration trials of dexlansoprazole MR (15,16,17) to determine whether patients with endoscopically confirmed erosive esophagitis (EE) receiving therapy with dexlansoprazole MR or lansoprazole developed RAHS and/or a worsening of symptoms when therapy was discontinued after healing. METHODS Patient population We analyzed data from patients who participated in the pivotal studies for dexlansoprazole MR that evaluated healing of EE (15) and maintenance of EE healing (16,17). All trials were conducted in compliance with institutional review board/ethics committee regulations and within the ethical principles stated in the 1989 Declaration of Helsinki. All patients voluntarily signed informed consent forms in the region in which the patient was participating and completed any Health Insurance Portability and Accountability Act forms (US sites only) before any study-related procedure was initiated.

hin the ethical principles stated in the 1989 Declaration of Helsinki. All patients voluntarily signed informed consent forms in the region in which the patient was participating and completed any Health Insurance Portability and Accountability Act forms (US sites only) before any study-related procedure was initiated. Patients were excluded if they were pregnant or lactating; tested H. pylori positive based on CLOtest (urease test; Kimberly-Clark, Roswell, GA), which is mandatory outside North America and performed in the United States and Canada for patients who tested positive by finger stick or serology at screening; had used prescription or nonprescription PPIs or histamine-2 receptor antagonists within 14 days of screening or throughout the study; had used nonsteroidal anti-inflammatory drugs chronically; had a history of active gastric or duodenal ulcers within 4 weeks of the first dose of study drug; or had acute upper gastrointestinal hemorrhage within 4 weeks of screening endoscopy. Previous use of PPIs (within 90 days of signing the informed consent form) was also recorded by patients enrolled in the trials.

cally; had a history of active gastric or duodenal ulcers within 4 weeks of the first dose of study drug; or had acute upper gastrointestinal hemorrhage within 4 weeks of screening endoscopy. Previous use of PPIs (within 90 days of signing the informed consent form) was also recorded by patients enrolled in the trials. Healing of EE was assessed in two identical double-blind, randomized controlled studies of 4,092 adult patients at 188 US and 118 non-US centers with endoscopically confirmed EE (ClinicalTrials.gov identifiers NCT00251693 and NCT00251719) (15). A total of 896 patients who were healed after 4 or 8 weeks of once-daily therapy with dexlansoprazole MR 60 or 90 mg or lansoprazole 30 mg were then enrolled into one of the two double-blind, placebo-controlled maintenance studies. In the Metz et al. (16) study, patients were randomized to dexlansoprazole MR 30 or 60 mg or placebo for up to 6 months (ClinicalTrials.gov identifier NCT00321737). The Howden et al. (17) study, which originally consisted of two identical study protocols (ClinicalTrial.gov identifiers NCT00255164 and NCT00255151), randomized patients to dexlansoprazole MR 60 or 90 mg or placebo for up to 6 months. Use of approved antacids, up to six tablets in a 24-h period, was permitted during healing and maintenance. Scheduled visits occurred at months 1, 3, and 6 in the maintenance studies. Patients whose healed EE relapsed at month 1 or 3 were discontinued from the studies. The majority of patients who relapsed at month 1 were in the placebo group, leading to a smaller number of placebo-treated patients at later visits.

ntenance. Scheduled visits occurred at months 1, 3, and 6 in the maintenance studies. Patients whose healed EE relapsed at month 1 or 3 were discontinued from the studies. The majority of patients who relapsed at month 1 were in the placebo group, leading to a smaller number of placebo-treated patients at later visits. Among patients who were healed after 4 or 8 weeks of treatment with dexlansoprazole MR or lansoprazole, 287 were subsequently randomized to placebo in the maintenance studies. The occurrence of RAHS was evaluated using available serum gastrin and symptom severity data from these patients. Within the placebo group, 24-h (daytime and nighttime) heartburn symptom severity was analyzed if data were available at baseline (before enrollment in the EE healing studies) and weeks 1, 2, 3, or 4, or months 1 or 2 of the maintenance studies. Heartburn symptom severity was assessed by daily diary. Patients rated heartburn severity using a 5-point scale where none (0) is no heartburn; mild (1) is occasional heartburn that could be ignored and did not influence daily routine or sleep; moderate (2) is heartburn that could not be ignored and/or occasionally influenced daily routine or sleep; severe (3) is heartburn that was present most of the day and/or regularly influenced daily routine or sleep; and very severe (4) is heartburn that was constant and/or markedly influenced daily routine or sleep.

p; moderate (2) is heartburn that could not be ignored and/or occasionally influenced daily routine or sleep; severe (3) is heartburn that was present most of the day and/or regularly influenced daily routine or sleep; and very severe (4) is heartburn that was constant and/or markedly influenced daily routine or sleep. Analyses Evaluation of gastrin levels Fasting serum gastrin levels were measured at baseline before EE healing treatment, at the end of EE healing treatment (day −1), and at the month 1 and month 3 visits in the maintenance studies. For these analyses, we also assessed the change from baseline to each of the post-baseline visits. RAHS was inferred if gastrin levels were increased above pretreatment levels following discontinuation of PPI therapy. An additional analysis was performed to look at patients defined by previous PPI usage to compare data for those who had reported taking a PPI within 90 days of signing an informed consent form with those who had not.

d if gastrin levels were increased above pretreatment levels following discontinuation of PPI therapy. An additional analysis was performed to look at patients defined by previous PPI usage to compare data for those who had reported taking a PPI within 90 days of signing an informed consent form with those who had not. Symptom rebound evaluation For each patient, mean severity of heartburn as assessed by daily diary was calculated and summarized for baseline (the 7-day period before randomization in the EE healing studies), during EE healing treatment, during weeks 1, 2, 3, and 4 of maintenance treatment, during the first month of maintenance treatment, and for the change from baseline to each week and to the first month of maintenance treatment. Patients were considered to have sufficient diary data for the weekly summary of heartburn symptom severity if they had at least 4 days of evaluable diary entries within the given week, and sufficient diary data for the monthly summary if they had at least 15 days of entries. Two-sided Wilcoxon signed-rank tests were performed to test the change from baseline against no change. Symptom rebound was inferred as a worsening of symptoms after withdrawal of PPI therapy compared with baseline symptom assessments.

fficient diary data for the monthly summary if they had at least 15 days of entries. Two-sided Wilcoxon signed-rank tests were performed to test the change from baseline against no change. Symptom rebound was inferred as a worsening of symptoms after withdrawal of PPI therapy compared with baseline symptom assessments. The analysis of 24-h heartburn severity was also performed in subgroups defined by previous PPI usage as reported by patients; healing treatment (dexlansoprazole MR compared with lansoprazole); duration of healing treatment (4 weeks compared with 8 weeks); and baseline body mass index (< or ≥30 kg/m2). In the subgroup of patients who also had heartburn severity results during the second month of maintenance treatment, a similar analysis was performed, which included two-sided Wilcoxon signed-rank tests for the change from baseline to month 2. In addition, we looked at mean heartburn symptom severity separately in the daytime and in the nighttime. EE rebound We also looked at the endoscopic grade of EE from the source documents of patients who relapsed after treatment with placebo in the maintenance studies. We compared these grades with the baseline EE grades of patients to see whether EE worsened after PPI therapy was withdrawn. For this analysis, 124 reports were available.

lso looked at the endoscopic grade of EE from the source documents of patients who relapsed after treatment with placebo in the maintenance studies. We compared these grades with the baseline EE grades of patients to see whether EE worsened after PPI therapy was withdrawn. For this analysis, 124 reports were available. RESULTS Baseline demographics for the cohort of patients (N=287) who were healed after 4–8 weeks of treatment with dexlansoprazole MR or lansoprazole and then randomized to placebo in the maintenance studies are shown in Table 1. Among the 287 patients, 5 were not included in the diary summary because of missing baseline diary. Of the other 282 patients, 60 were not included in the month 1 diary summary, including 48 who prematurely discontinued before day 30 and 12 additional patients who were in the maintenance study for at least 30 days but did not have sufficient diary data during the first 30 days. Of the 222 patients who were included in the month 1 diary summary, 125 patients were in the maintenance study for at least 30 days and the other 97 prematurely discontinued before day 30, but had sufficient diary data. Baseline median of mean heartburn symptom severity of the 222 patients included in the month 1 symptom analysis was 1.36. Of the 124 included in the EE rebound analysis, 31% had grade A esophagitis at baseline, 38% had grade B esophagitis, 27% had grade C esophagitis, and 4% had grade D esophagitis.

ficient diary data. Baseline median of mean heartburn symptom severity of the 222 patients included in the month 1 symptom analysis was 1.36. Of the 124 included in the EE rebound analysis, 31% had grade A esophagitis at baseline, 38% had grade B esophagitis, 27% had grade C esophagitis, and 4% had grade D esophagitis. Gastrin evaluation There was no evidence of RAHS as assessed by serum gastrin levels in this analysis. Among patients randomized to maintenance with placebo with available serum gastrin levels, mean (s.d.) gastrin level at baseline (pretreatment; n=216) was 59.9 (31.4) pg/ml. The changes from baseline to day −1 of maintenance trials (n=207), to maintenance month 1 (n=130), and to maintenance month 3 (n=52) were 85.5 (120.8) pg/ml, 3.5 (42.3) pg/ml, and 3.2 (66.8) pg/ml, respectively. As would be expected, gastrin values did increase from baseline to day −1 of the maintenance studies while patients were receiving active treatment (18,19). In patients randomized to placebo, mean gastrin values at months 1 and 3 were essentially unchanged from baseline. This indicates that gastrin levels normalized within 1 month of discontinuing PPIs and remained flat. Furthermore, data do not suggest any difference in the findings between those who used a PPI within 90 days of informed consent compared with those who did not (Figure 1).

nd 3 were essentially unchanged from baseline. This indicates that gastrin levels normalized within 1 month of discontinuing PPIs and remained flat. Furthermore, data do not suggest any difference in the findings between those who used a PPI within 90 days of informed consent compared with those who did not (Figure 1). Symptom rebound evaluation There was no evidence of symptom rebound in this analysis. The mean 24-h heartburn severity was higher after PPI therapy was withdrawn compared with mean heartburn severity during PPI treatment, as would be expected. However, the mean heartburn severity at maintenance month 1, when gastrin values returned to near-normal levels, was significantly lower than at baseline for all symptom analyses performed, despite the absence of PPI maintenance therapy ( Table 2). When analyzed week by week during the first maintenance month, mean symptom severity scores remained flat—0.71, 0.93, 0.71, and 0.67 during weeks 1, 2, 3, and 4, respectively. Figure 2 illustrates the median change in mean 24-h symptom severity scores from baseline to each week. During the first month of maintenance treatment, the median percentage of days with antacid use among placebo patients was 66%. In contrast, during the last 7 days before randomization in the EE healing study (baseline), the median number of days with antacid use among these patients was 6 days.

baseline to each week. During the first month of maintenance treatment, the median percentage of days with antacid use among placebo patients was 66%. In contrast, during the last 7 days before randomization in the EE healing study (baseline), the median number of days with antacid use among these patients was 6 days. The median of mean heartburn severity was <1 (i.e., less than mild) for all analyses performed 1 month after discontinuing PPI therapy. Data do not suggest any difference in heartburn severity between those who reported using a PPI within 90 days of informed consent compared with those who did not (Figure 3a). Heartburn severity also appeared to be similar in patients healed with dexlansoprazole MR or lansoprazole (Figure 3b), with 4 or 8 weeks of healing treatment (Figure 3c), or in patients with body mass index of < or ≥30 kg/m2 (data not shown). In addition, the changes in daytime heartburn severity and nighttime heartburn severity were consistent with that in 24-h heartburn severity (data not shown). Because most placebo patients failed to maintain healed EE beyond month 1, there are fewer patients with data in the second month or beyond in the maintenance study. Among those with month 2 data, mean heartburn severity at months 1 and 2 was also significantly lower than at baseline (median decrease of 0.54 and 0.58 points; both P<0.001), indicating an ongoing symptom response for up to 2 months after PPI withdrawal.

ith data in the second month or beyond in the maintenance study. Among those with month 2 data, mean heartburn severity at months 1 and 2 was also significantly lower than at baseline (median decrease of 0.54 and 0.58 points; both P<0.001), indicating an ongoing symptom response for up to 2 months after PPI withdrawal. EE rebound evaluation More than half of patients whose EE recurred after maintenance therapy was discontinued relapsed to a less severe grade of EE compared with 18% who relapsed to a more severe grade of EE (Table 3).

ith data in the second month or beyond in the maintenance study. Among those with month 2 data, mean heartburn severity at months 1 and 2 was also significantly lower than at baseline (median decrease of 0.54 and 0.58 points; both P<0.001), indicating an ongoing symptom response for up to 2 months after PPI withdrawal. EE rebound evaluation More than half of patients whose EE recurred after maintenance therapy was discontinued relapsed to a less severe grade of EE compared with 18% who relapsed to a more severe grade of EE (Table 3). DISCUSSION RAHS and possible PPI dependency in patients discontinuing PPI therapy have been postulated to occur because of the hypertrophic effects of acid suppression, induced by increased gastrin levels, on histamine-releasing enterochromaffin-like cells (5,20). There was no evidence of RAHS in our analysis within 3 months of discontinuing 4–8 weeks of PPI therapy in a cohort of H. pylori-negative EE patients. Gastrin levels normalized within 1 month of discontinuing PPI therapy and remained flat at month 3. Furthermore, there was no evidence of symptom rebound within 2 months of PPI withdrawal. Rather, we observed a trend toward less severe, as opposed to more severe, heartburn after healing and PPI withdrawal—the opposite of symptom rebound. This is supported by the lower use of rescue medication during maintenance month 1 compared with baseline. Heartburn symptom severity after PPI withdrawal was similar among patients treated with dexlansoprazole MR or lansoprazole and for those treated for 4 or 8 weeks, suggesting that neither more potent therapy nor longer exposure was associated with rebound in the patients evaluated. These findings are based on data from well-controlled trials in an appropriate patient population and do not support the conjecture that withdrawal of short-term (i.e., <8 weeks) PPI therapy will produce symptomatic rebound that could lead to PPI dependency.

posure was associated with rebound in the patients evaluated. These findings are based on data from well-controlled trials in an appropriate patient population and do not support the conjecture that withdrawal of short-term (i.e., <8 weeks) PPI therapy will produce symptomatic rebound that could lead to PPI dependency. Data from our analysis contradict the findings of two recent trials that evaluated symptom development in healthy volunteers after withdrawal of a short course of PPI therapy (8,9). Reimer et al. (8) randomized subjects to receive either 8 weeks of esomeprazole 40 mg daily followed by 4 weeks of placebo or 12 weeks of placebo only. Symptoms were assessed weekly via questionnaires using a 15-item gastrointestinal symptom rating scale with a 7-point scale (1=absence of bothersome symptoms; 7=very bothersome symptoms). Mean fasting gastrin levels were also measured.

esomeprazole 40 mg daily followed by 4 weeks of placebo or 12 weeks of placebo only. Symptoms were assessed weekly via questionnaires using a 15-item gastrointestinal symptom rating scale with a 7-point scale (1=absence of bothersome symptoms; 7=very bothersome symptoms). Mean fasting gastrin levels were also measured. At least 1 symptom of heartburn, acid regurgitation, or dyspepsia was reported during weeks 9–12 by 44% (26/59) and 15% (9/59) of patients in the PPI and placebo groups, respectively (P<0.001). Mean scores for the subset of questions focusing on heartburn, acid regurgitation, and dyspepsia at baseline were 1.04 and 1.03 for the PPI and placebo groups, respectively, whereas scores for the last 4 weeks combined were 1.35 and 1.12, respectively (P=0.001) (8). Although this difference was statistically significant, scores were low in both and the numerical difference was modest (0.23 points). In addition, it cannot be determined if symptoms of this frequency and very mild severity would cause patients to resume PPI therapy. Plasma gastrin levels were significantly higher in the PPI group than the placebo group during treatment and correlated significantly with symptoms in the PPI group after therapy was discontinued (8). However, gastrin levels were within the normal range (<50 pmol/l) in both groups throughout the study and returned to baseline values by week 12 in the PPI group. There was an imbalance of H. pylori-positive subjects in the placebo group in this trial (13% vs. 2% in the esomeprazole group, P=0.02) (8). Because H. pylori-positive subjects are generally less likely to experience rebound problems (14), and because the numerical differences between groups was quite small, it is possible that H. pylori status had an impact on the findings in this study population. Furthermore, because symptoms were only evaluated weekly, there may have been the possibility of recall bias.

ess likely to experience rebound problems (14), and because the numerical differences between groups was quite small, it is possible that H. pylori status had an impact on the findings in this study population. Furthermore, because symptoms were only evaluated weekly, there may have been the possibility of recall bias. Niklasson et al. (9) enrolled H. pylori-negative subjects in a randomized, double-blind, placebo-controlled trial evaluating dyspeptic symptoms after discontinuation of PPI therapy. A total of 48 female volunteers reported their symptoms daily throughout a 2-week run-in period, during the 4 weeks of treatment with either pantoprazole 40 mg daily or placebo, and for 6 weeks after treatment was discontinued (9). Symptoms were evaluated daily using a modified Swedish version of the Glasgow dyspepsia score questionnaire, which was designed to reflect on symptom severity, nocturnal disturbance, and behavioral response to sustained symptoms, and were scored on a scale of 0 (no symptoms) to 12 (9). Fasting and meal-stimulated gastrin levels were measured the day before treatment was started, on the last day of treatment, and then 6 weeks after treatment was discontinued (9).

tom severity, nocturnal disturbance, and behavioral response to sustained symptoms, and were scored on a scale of 0 (no symptoms) to 12 (9). Fasting and meal-stimulated gastrin levels were measured the day before treatment was started, on the last day of treatment, and then 6 weeks after treatment was discontinued (9). During the first week after treatment was discontinued, 44% (11/25) and 9% (2/23) of subjects in the pantoprazole and placebo groups, respectively, reported dyspeptic symptoms (P=0.009) (9). The difference between the groups was also significant during the second week, but not during the remaining 4 weeks. Mean symptom scores at baseline were 0.54±1.3 and 0.20±0.7 in the pantoprazole and placebo groups, respectively, and 5.7±11.7 and 0.74±2.6 (P<0.01), respectively, 1 week after treatment discontinuation (9). Although symptom scores decreased in the second week post- treatment, they were still significantly different between the two treatment groups (P<0.05), but this significance disappeared in weeks 3–6. Symptoms resolved quickly; the median symptom duration was 4 days in the subgroup of PPI-treated subjects who experienced symptoms (9). Mean fasting and meal-induced serum gastrin levels were significantly higher in the pantoprazole group than in the placebo group during the last week of treatment and were significantly correlated with symptom scores (P<0.01) (9). There was no significant difference between groups at 6 weeks after treatment was discontinued. Because the first measurement of gastrin scores was not until 6 weeks after treatment was discontinued, it is possible that the difference between groups resolved before gastrin levels were assessed post-treatment (9).

(9). There was no significant difference between groups at 6 weeks after treatment was discontinued. Because the first measurement of gastrin scores was not until 6 weeks after treatment was discontinued, it is possible that the difference between groups resolved before gastrin levels were assessed post-treatment (9). Both normal volunteer studies cited in this paper utilized a broader acid peptic symptom complex when defining rebound, including regurgitation and dyspepsia. Our definition of symptom rebound, which was defined as the recurrence of heartburn symptoms of greater severity than reported at baseline, has been utilized in several previously published clinical trials of GERD treatment. Although we cannot exclude that earlier onset of symptom relapse, rather than increased severity, could reflect rebound, timing of relapse rates reported in our study was similar to that reported in other placebo-controlled studies. Overall, it is difficult to say whether the conclusions from either normal volunteer trial (8,9) are clinically relevant and can be extrapolated to patients who are receiving therapy with PPIs to treat acid-related disorders. The fact that healthy volunteers receiving placebo reported symptoms in both trials may be an indication that subjects in these trials were sensitized to report symptoms. It is also possible that the PPI-treated subjects were reporting side effects of therapy rather than symptoms of GERD. The fact that some, but not all, subjects receiving PPIs reported symptoms, and that these symptoms were generally mild and transient, tends to support this hypothesis.

e sensitized to report symptoms. It is also possible that the PPI-treated subjects were reporting side effects of therapy rather than symptoms of GERD. The fact that some, but not all, subjects receiving PPIs reported symptoms, and that these symptoms were generally mild and transient, tends to support this hypothesis. In one double-blind, randomized, placebo-controlled study, Farup et al. (21) evaluated symptoms after PPI withdrawal in a population of GERD patients. In this crossover trial, 62 patients were treated with high doses of lansoprazole (60 mg daily) followed by placebo, or vice versa, to see if PPIs would aggravate symptoms after withdrawal of 5 days of therapy (21). The authors chose this treatment period because a short course of PPI therapy is often used as a confirmatory test in patients suspected of having an acid-related disorder. In this trial, symptoms at 12–14 days before and after the treatment period were generally similar between the two treatment groups. There was no indication of symptom rebound (21). Our analysis does have some limitations. It is difficult to distinguish symptom rebound from symptom recurrence in patients suffering from a chronic, relapsing disorder. We inferred symptom rebound based on a worsening of symptoms when therapy was withdrawn beyond severity recorded at baseline. This definition does test the hypotheses that withdrawal of PPI therapy may somehow aggravate symptoms in patients with an acid-related disorder.

ts suffering from a chronic, relapsing disorder. We inferred symptom rebound based on a worsening of symptoms when therapy was withdrawn beyond severity recorded at baseline. This definition does test the hypotheses that withdrawal of PPI therapy may somehow aggravate symptoms in patients with an acid-related disorder. Furthermore, the EE healing studies preceding the maintenance studies only included active PPI treatment groups (dexlansoprazole MR or lansoprazole) and did not have a placebo group, which could have yielded additional interesting findings. We also used serum gastrin levels as an indirect measure of acid secretion, which can be determined more directly by several methods, including aspiration. Measuring area under the plasma concentration time curve after a meal would also provide a better surrogate than fasting serum gastrin. These more direct assessments of gastric acid levels were not performed in the original studies; as such, they were not available for this secondary analysis.

ds, including aspiration. Measuring area under the plasma concentration time curve after a meal would also provide a better surrogate than fasting serum gastrin. These more direct assessments of gastric acid levels were not performed in the original studies; as such, they were not available for this secondary analysis. Regarding the observation that grades of EE tended to be less severe in the maintenance period than at baseline, we cannot completely exclude a potential effect from regression to the mean. However, any potential effect from regression to the mean would have been indirect because of the previous EE healing studies. The effect could only have contributed to the healing and/or improvement during the EE healing studies, but not the change during the EE maintenance studies. Finally, we only had data available for patients who were treated with PPIs for 4 or 8 weeks before being discontinued from therapy. Although our analysis clearly showed no evidence of acid or symptom rebound that could provoke PPI dependence, we cannot infer the effect of withdrawing patients from long-term PPI therapy based on data from a short-term analysis.

ients who were treated with PPIs for 4 or 8 weeks before being discontinued from therapy. Although our analysis clearly showed no evidence of acid or symptom rebound that could provoke PPI dependence, we cannot infer the effect of withdrawing patients from long-term PPI therapy based on data from a short-term analysis. The recent studies in healthy volunteers (8,9) certainly renewed interest in the study of RAHS and will contribute to the ongoing debate regarding the occurrence and clinical significance of RAHS. Likewise, the data from our analysis in symptomatic EE patients will contribute to this debate. However, pending consistent and compelling data that withdrawal of PPIs produces acid or symptom rebound in patients that results in PPI dependence, we see no reason for clinicians to alter their prescribing habits. We would recommend that physicians continue to follow the evidence-based AGA (American Gastroenterological Association) 2008 guidelines (1), which state that empirical therapy with PPIs is considered appropriate to initiate therapy in patients with uncomplicated heartburn. Maintenance therapy is appropriate in patients with EE. Additionally, the majority of patients with GERD symptoms severe enough to warrant initial PPI therapy will require long-term PPI therapy to achieve adequate symptom control. In these patients, the dose should be titrated to the lowest effective dose that achieves symptom control.

py is appropriate in patients with EE. Additionally, the majority of patients with GERD symptoms severe enough to warrant initial PPI therapy will require long-term PPI therapy to achieve adequate symptom control. In these patients, the dose should be titrated to the lowest effective dose that achieves symptom control. Currently, the majority of total PPI prescriptions are for long-term therapy (variably defined as one prescription repeated over 12 months to continuous therapy that ranges from 4 to >12 months) (22). Although PPIs are often prescribed for long-term usage, patient adherence to PPI regimens appears to be suboptimal. Van Soest et al. (23) evaluated PPI usage among 386,002 patients for a mean of 3.4±2.4 years and found that only 22% of the study cohort continued PPI treatment for at least a year. Adherence was low to moderate in more than half of all patients, reflecting a high prevalence of intermittent therapy among PPI users, which may be problematic for some patients. Therefore, it is also important for clinicians to regularly monitor the status of patients on PPI therapy to ensure that they are receiving optimal treatment.

to moderate in more than half of all patients, reflecting a high prevalence of intermittent therapy among PPI users, which may be problematic for some patients. Therefore, it is also important for clinicians to regularly monitor the status of patients on PPI therapy to ensure that they are receiving optimal treatment. In summary, the findings of our analysis show that discontinuation of 4 or 8 weeks of PPI therapy did not produce any evidence of relapse of heartburn symptom severity to levels worse than pretreatment levels in a patient population that derived benefits from PPI therapy. Furthermore, serum gastrin levels were shown to normalize within a month of discontinuing therapy and symptoms appeared to be less severe as opposed to more severe when PPI therapy was withdrawn. Clearly, there was no indication that discontinuation of a 4- or 8-week course of PPI therapy is causing PPI dependency. However, the current findings do not exclude the possibility of RAHS following discontinuation of long-term PPI therapy. STUDY HIGHLIGHTS Medical writing assistance was provided by Eileen R. Gallagher, Gallagher Medical Communications, LLC, and funded by Takeda Pharmaceuticals North America, Deerfield, IL, USA. Guarantor of the article: David C. Metz, MD.

In summary, the findings of our analysis show that discontinuation of 4 or 8 weeks of PPI therapy did not produce any evidence of relapse of heartburn symptom severity to levels worse than pretreatment levels in a patient population that derived benefits from PPI therapy. Furthermore, serum gastrin levels were shown to normalize within a month of discontinuing therapy and symptoms appeared to be less severe as opposed to more severe when PPI therapy was withdrawn. Clearly, there was no indication that discontinuation of a 4- or 8-week course of PPI therapy is causing PPI dependency. However, the current findings do not exclude the possibility of RAHS following discontinuation of long-term PPI therapy. STUDY HIGHLIGHTS Medical writing assistance was provided by Eileen R. Gallagher, Gallagher Medical Communications, LLC, and funded by Takeda Pharmaceuticals North America, Deerfield, IL, USA. Guarantor of the article: David C. Metz, MD. Specific author contributions: Initiated the study concept and design, analysis, and interpretation of data, critical revision of the paper: David C. Metz; study design, analysis, and interpretation of data; critical revision of the paper: Betsy L. Pilmer; study design, conducted the data analysis, and critical revision of the paper: Cong Han; study design, analysis and interpretation; critical revision of the paper: M. Claudia Perez. All authors have seen and approved the final report. Financial support: This analysis was sponsored by Takeda Global Research & Development Center.

Specific author contributions: Initiated the study concept and design, analysis, and interpretation of data, critical revision of the paper: David C. Metz; study design, analysis, and interpretation of data; critical revision of the paper: Betsy L. Pilmer; study design, conducted the data analysis, and critical revision of the paper: Cong Han; study design, analysis and interpretation; critical revision of the paper: M. Claudia Perez. All authors have seen and approved the final report. Financial support: This analysis was sponsored by Takeda Global Research & Development Center. Potential competing interests: D.C. Metz has served as a consultant and reports receiving honoraria and/or grant support from AstraZeneca, Takeda, Novartis Consumer, and Xenoport. B.L. Pilmer, C. Han, and M.C. Perez are employees of Takeda Global Research & Development Center, Deerfield, IL, USA. Figure 1 Mean serum gastrin levels. EE, erosive esophagitis; PPI, proton pump inhibitor. Error bars represent the s.d. Figure 2 Median change from baseline in mean 24-h heartburn symptom severity in all placebo-treated patients with data for baseline and weeks 1–4 of the maintenance studies. CFB, change from baseline. P<0.001 for comparison of change from baseline with no change for each week; Wilcoxon signed-rank test.

Figure 1 Mean serum gastrin levels. EE, erosive esophagitis; PPI, proton pump inhibitor. Error bars represent the s.d. Figure 2 Median change from baseline in mean 24-h heartburn symptom severity in all placebo-treated patients with data for baseline and weeks 1–4 of the maintenance studies. CFB, change from baseline. P<0.001 for comparison of change from baseline with no change for each week; Wilcoxon signed-rank test. Figure 3 Heartburn severity during follow-up. Mean 24-h heartburn severity (median) in all placebo-treated patients with baseline and month 1 data shown by (a) previous PPI usage, (b) healing therapy, and (c) duration of healing therapy. EE, erosive esophagitis; PPI, proton pump inhibitor. Mean severity is calculated for each patient based on a 5-point scale: 0=none, 1=mild, 2=moderate, 3=severe, and 4=very severe. Table 1 Baseline demographics of placebo cohorta Demographic variable Placebo cohort (N=287) Gender, n (%) Men 142 (49.5) Women 145 (50.5) Ethnicity, n (%) Hispanic/latino 35 (12.2) Not hispanic/latino 252 (87.8) Race, n (%) White 262 (91.3) Black of African Heritage 15 (5.2) Asian 5 (1.7) Other 5 (1.7) Age (years), n (%) <45 98 (34.1) 45 to <65 164 (57.1) ≥65 25 (8.7) Mean±s.d. 48.9±12.90 Body mass index (kg/m2), n (%) <25 48 (16.7) 25 to <30 111 (38.7) ≥30 127 (44.3) Unknown 1 (0.3) Mean±s.d. 30.2±6.09 a Erosive esophagitis patients healed after 4–8 weeks of therapy with dexlansoprazole modified release (MR) or lansoprazole and then randomized to placebo in the maintenance of healing trials.

48.9±12.90 Body mass index (kg/m2), n (%) <25 48 (16.7) 25 to <30 111 (38.7) ≥30 127 (44.3) Unknown 1 (0.3) Mean±s.d. 30.2±6.09 a Erosive esophagitis patients healed after 4–8 weeks of therapy with dexlansoprazole modified release (MR) or lansoprazole and then randomized to placebo in the maintenance of healing trials. Table 2 Change in mean severity of 24-h heartburn Patients Baseline of EE healing trials During EE healing trials Maintenance month 1 CFB to maintenance month 1 n Median (Q1, Q3) n Median (Q1, Q3) n Median (Q1, Q3) n Median (Q1, Q3) All patients with baseline and month 1 data 222 1.36 (0.86, 1.93) 221 0.14 (0.04, 0.50) 222 0.76 (0.37, 1.38) 222 −0.41a (−0.94, −0.02) No PPI used with 90 days of informed consent 127 1.36 (0.86, 2.00) 126 0.15 (0.04, 0.48) 127 0.75 (0.37, 1.29) 127 −0.43a (−0.98, 0.01) PPI used with 90 days of informed consent 95 1.36 (0.79, 1.93) 95 0.13 (0.04, 0.52) 95 0.77 (0.37, 1.45) 95 −0.35a (−0.82, −0.09) Healed on dexlansoprazole MR 151 1.50 (0.86, 2.00) 150 0.15 (0.05, 0.52) 151 0.85 (0.35, 1.45) 151 −0.40a (−0.97, −0.07) Healed on lansoprazole 71 1.21 (0.64, 1.79) 71 0.12 (0.02, 0.38) 71 0.70 (0.38, 1.11) 71 −0.43a (−0.90, 0.00) Healed at week 4 187 1.43 (0.80, 2.00) 186 0.15 (0.04, 0.50) 187 0.82 (0.38, 1.43) 187 −0.35a (−0.93, −0.03) Healed at week 8 34 1.18 (0.86, 1.64) 34 0.09 (0.04, 0.34) 34 0.58 (0.35, 1.03) 34 −0.58a (−1.13, 0.08) EE, erosive esophagitis; CFB, change from baseline; MR, modified release; PPI, proton pump inhibitor. a P≤0.001 from a two-sided Wilcoxon signed-rank test.

Table 2 Change in mean severity of 24-h heartburn Patients Baseline of EE healing trials During EE healing trials Maintenance month 1 CFB to maintenance month 1 n Median (Q1, Q3) n Median (Q1, Q3) n Median (Q1, Q3) n Median (Q1, Q3) All patients with baseline and month 1 data 222 1.36 (0.86, 1.93) 221 0.14 (0.04, 0.50) 222 0.76 (0.37, 1.38) 222 −0.41a (−0.94, −0.02) No PPI used with 90 days of informed consent 127 1.36 (0.86, 2.00) 126 0.15 (0.04, 0.48) 127 0.75 (0.37, 1.29) 127 −0.43a (−0.98, 0.01) PPI used with 90 days of informed consent 95 1.36 (0.79, 1.93) 95 0.13 (0.04, 0.52) 95 0.77 (0.37, 1.45) 95 −0.35a (−0.82, −0.09) Healed on dexlansoprazole MR 151 1.50 (0.86, 2.00) 150 0.15 (0.05, 0.52) 151 0.85 (0.35, 1.45) 151 −0.40a (−0.97, −0.07) Healed on lansoprazole 71 1.21 (0.64, 1.79) 71 0.12 (0.02, 0.38) 71 0.70 (0.38, 1.11) 71 −0.43a (−0.90, 0.00) Healed at week 4 187 1.43 (0.80, 2.00) 186 0.15 (0.04, 0.50) 187 0.82 (0.38, 1.43) 187 −0.35a (−0.93, −0.03) Healed at week 8 34 1.18 (0.86, 1.64) 34 0.09 (0.04, 0.34) 34 0.58 (0.35, 1.03) 34 −0.58a (−1.13, 0.08) EE, erosive esophagitis; CFB, change from baseline; MR, modified release; PPI, proton pump inhibitor. a P≤0.001 from a two-sided Wilcoxon signed-rank test. Mean severity is calculated for each patient based on a 5-point scale: 0=none, 1=mild, 2=moderate, 3=severe, and 4=very severe.

Table 2 Change in mean severity of 24-h heartburn Patients Baseline of EE healing trials During EE healing trials Maintenance month 1 CFB to maintenance month 1 n Median (Q1, Q3) n Median (Q1, Q3) n Median (Q1, Q3) n Median (Q1, Q3) All patients with baseline and month 1 data 222 1.36 (0.86, 1.93) 221 0.14 (0.04, 0.50) 222 0.76 (0.37, 1.38) 222 −0.41a (−0.94, −0.02) No PPI used with 90 days of informed consent 127 1.36 (0.86, 2.00) 126 0.15 (0.04, 0.48) 127 0.75 (0.37, 1.29) 127 −0.43a (−0.98, 0.01) PPI used with 90 days of informed consent 95 1.36 (0.79, 1.93) 95 0.13 (0.04, 0.52) 95 0.77 (0.37, 1.45) 95 −0.35a (−0.82, −0.09) Healed on dexlansoprazole MR 151 1.50 (0.86, 2.00) 150 0.15 (0.05, 0.52) 151 0.85 (0.35, 1.45) 151 −0.40a (−0.97, −0.07) Healed on lansoprazole 71 1.21 (0.64, 1.79) 71 0.12 (0.02, 0.38) 71 0.70 (0.38, 1.11) 71 −0.43a (−0.90, 0.00) Healed at week 4 187 1.43 (0.80, 2.00) 186 0.15 (0.04, 0.50) 187 0.82 (0.38, 1.43) 187 −0.35a (−0.93, −0.03) Healed at week 8 34 1.18 (0.86, 1.64) 34 0.09 (0.04, 0.34) 34 0.58 (0.35, 1.03) 34 −0.58a (−1.13, 0.08) EE, erosive esophagitis; CFB, change from baseline; MR, modified release; PPI, proton pump inhibitor. a P≤0.001 from a two-sided Wilcoxon signed-rank test. Mean severity is calculated for each patient based on a 5-point scale: 0=none, 1=mild, 2=moderate, 3=severe, and 4=very severe. Table 3 Change in Los Angeles grade of erosive esophagitis Start of healing study (baseline) Relapse to grade A Relapse to grade B Relapse to grade C Relapse to grade D Grade, n n n n n Grade A=39 23 14 2 0 Grade B=47 21 20 6 0 Grade C=33 3 17 13 0 Grade D=5 0 0 2 3

INTRODUCTION Chronic idiopathic constipation (CIC) is a common gastrointestinal (GI) condition that affects quality of life and health-care expenditures (1). It has been estimated that 12–19% (2) of the general population experiences CIC. Constipation is more often reported by females and those over 65 years of age (2). The symptoms of CIC are diverse and include infrequent bowel movements (BMs), hard stools, excessive straining, bloating, and abdominal discomfort/pain (3). Though a number of over-the-counter and prescription medications are available, approximately half of patients with CIC are not satisfied with their current treatment (4). As such, there remains an unmet need for medical therapies and approaches addressing novel physiological mechanisms in patients with CIC.

/pain (3). Though a number of over-the-counter and prescription medications are available, approximately half of patients with CIC are not satisfied with their current treatment (4). As such, there remains an unmet need for medical therapies and approaches addressing novel physiological mechanisms in patients with CIC. The pathophysiology of CIC remains incompletely defined. There is evidence to suggest that modulation of endogenous bile acid (BA) homeostasis might offer a novel therapeutic approach to the treatment of CIC. There are several examples demonstrating that changes in intracolonic BA concentrations alter bowel function: ileal resection results in diarrhea (5), and changes in diurnal rates of BA synthesis and colonic BA concentration have been identified in constipated children (6) and in adults with slow transit constipation and constipation-predominant irritable bowel syndrome (IBS-C) (7,8). Conversely, decreasing free BA concentrations in the colon through the use of the BA sequestrant cholestyramine induces constipation (5), and another BA sequestrant, colesevelam, retards accelerated colonic transit in patients with IBS-D (9).

tipation and constipation-predominant irritable bowel syndrome (IBS-C) (7,8). Conversely, decreasing free BA concentrations in the colon through the use of the BA sequestrant cholestyramine induces constipation (5), and another BA sequestrant, colesevelam, retards accelerated colonic transit in patients with IBS-D (9). Inhibitors of the ileal BA transporter (IBAT; syn ASBT—apical sodium-dependent BA transporter) reduce the active ileal reabsorption of BA, resulting in an increased concentration of BA entering the colon, thereby stimulating colonic motility and secretion (10). BAs induce propulsive contractions in the human colon (11), secretion through activation of adenylate cyclase (12), increased mucosal permeability (13), and inhibition of apical Cl−/OH− exchange (14). These effects on colonic motility and secretion may offer clinical benefits in patients with CIC. A3309 is a selective inhibitor of the IBAT, which has been shown to reduce meat-induced constipation in the dog (15). In humans, A3309 accelerated colonic transit in a multiple ascending dose, phase Ib study in patients with CIC (16). The current phase IIb study in patients with CIC was performed to assess the short-term efficacy and safety of orally administered A3309 at 5, 10, and 15 mg doses, compared with placebo, once daily for 8 weeks.

Inhibitors of the ileal BA transporter (IBAT; syn ASBT—apical sodium-dependent BA transporter) reduce the active ileal reabsorption of BA, resulting in an increased concentration of BA entering the colon, thereby stimulating colonic motility and secretion (10). BAs induce propulsive contractions in the human colon (11), secretion through activation of adenylate cyclase (12), increased mucosal permeability (13), and inhibition of apical Cl−/OH− exchange (14). These effects on colonic motility and secretion may offer clinical benefits in patients with CIC. A3309 is a selective inhibitor of the IBAT, which has been shown to reduce meat-induced constipation in the dog (15). In humans, A3309 accelerated colonic transit in a multiple ascending dose, phase Ib study in patients with CIC (16). The current phase IIb study in patients with CIC was performed to assess the short-term efficacy and safety of orally administered A3309 at 5, 10, and 15 mg doses, compared with placebo, once daily for 8 weeks. METHODS Study design: randomization and protocol A randomized, double-blind, placebo-controlled study was conducted at 45 sites across the United States between November 2009 (first signed informed consent) and August 2010 (last patient visit). The study was conducted in accordance with the Declaration of Helsinki and with Good Clinical Practice guidelines. Written informed consent was obtained from all patients before participation. Central or local Ethical Review Boards approved the study protocol and the informed consent form. The study was registered on ClinicalTrial.Gov [NCT01007123]. Treatment allocation and concealment were conducted by a computerized random-number generator and numbered containers with active and placebo capsules of identical appearance. The randomization list was centrally held by an independent Contract Research Organization, all randomized patients and study center and contract research organization personnel were blinded to study treatment allocation until data were locked and analyzed.

ners with active and placebo capsules of identical appearance. The randomization list was centrally held by an independent Contract Research Organization, all randomized patients and study center and contract research organization personnel were blinded to study treatment allocation until data were locked and analyzed. After informed consent was obtained, patients entered an initial screening period of up to 28 days during which routine blood tests, urinalysis, and pregnancy testing were obtained. Where appropriate, patients were taken off of laxatives and other prohibited medications. Specific washout periods (3- or 10-day periods based on duration of action) were used for prohibited medications. Bulking agents and stool softeners were allowed if the patient was taking a stable dose for at least 30 days before the start of the screening period. The use of rescue medication (bisacodyl suppository) was permitted following consultation with the investigator if the patient had >72 h without a BM. Patients meeting the inclusion and exclusion criteria entered a 14-day pretreatment baseline period. Patients eligible to continue were randomized to receive a 5-, 10-, or 15-mg tablet of A3309 or placebo orally once daily before breakfast for 8 weeks. During the 14 days before randomization and during the treatment period, patients reported daily bowel and abdominal symptoms using an Interactive Voice Response System. Study visits occurred during the screening period, at randomization (day 0), during treatment (day 28), at end of treatment (day 56), and 14 days post-treatment (day 70).

efore randomization and during the treatment period, patients reported daily bowel and abdominal symptoms using an Interactive Voice Response System. Study visits occurred during the screening period, at randomization (day 0), during treatment (day 28), at end of treatment (day 56), and 14 days post-treatment (day 70). Study medication Pharmacokinetic studies show that orally administered A3309 leads to minimal systemic exposure and that systemically available drug is highly protein bound (>99%) (15).

efore randomization and during the treatment period, patients reported daily bowel and abdominal symptoms using an Interactive Voice Response System. Study visits occurred during the screening period, at randomization (day 0), during treatment (day 28), at end of treatment (day 56), and 14 days post-treatment (day 70). Study medication Pharmacokinetic studies show that orally administered A3309 leads to minimal systemic exposure and that systemically available drug is highly protein bound (>99%) (15). Study participants Eligible patients were men or non-pregnant women, 20–80 years of age, who met modified Rome III criteria for functional constipation (17), with a diagnosis of >3 months and with symptom onset at least 6 months before diagnosis. Study participants were required to have a baseline BM frequency of <3 complete spontaneous BMs (CSBMs) per week. In addition, ≥2 of the following characteristics that are included in the Rome III criteria for functional constipation were required during the 14 days before randomization: (1) straining during ≥25% of BMs, (2) sensation of incomplete evacuation in ≥25% of BMs, (3) lumpy or hard stools in ≥25% of BMs, or (4) sensation of blockage/anorectal obstruction in ≥25% of BMs. Eligible patients had a body mass index of >18.0 and <35 kg/m2. Patients were excluded if they had been previously diagnosed with constipation-predominant IBS, had abdominal pain as a predominant symptom, or had drug-induced constipation. Also, patients were excluded if they reported the need for manual maneuvers (digital insertion into the rectum, perianal pressure, or vaginal splinting) or frequent use of enemas to evacuate stool. Other exclusion criteria included known rectal prolapse, dyssynergic defecation (per patient history and clinical examination), previous intestinal resection, neurological cause of constipation, and use of drugs with known effects on GI motility or stool consistency. Patients reporting loose or liquid stool consistency by a Bristol Stool Form Scale score of 6 or 7 (unless within 24 h of having taken a laxative) at least once during the screening period were excluded (18). Patients with persistent constipation symptoms despite a stable dosage of bulking agents for >4 weeks were eligible.

nts reporting loose or liquid stool consistency by a Bristol Stool Form Scale score of 6 or 7 (unless within 24 h of having taken a laxative) at least once during the screening period were excluded (18). Patients with persistent constipation symptoms despite a stable dosage of bulking agents for >4 weeks were eligible. All study participants were required to meet the colorectal cancer screening requirements of the American Gastroenterological Association Guidelines (19). Women of child-bearing age were required to have a negative serum pregnancy test; those who were sexually active were required to use an accepted form of birth control during their participation in the study.

t the colorectal cancer screening requirements of the American Gastroenterological Association Guidelines (19). Women of child-bearing age were required to have a negative serum pregnancy test; those who were sexually active were required to use an accepted form of birth control during their participation in the study. Outcome measures and assessments The primary end point of the study was the change from baseline in frequency of weekly spontaneous bowel movements (SBMs) at week 1. Secondary efficacy end points included weekly assessment of SBMs and CSBMs, daily assessments of other bowel-related complaints (i.e., stool consistency and straining), and abdominal symptoms (i.e., pain, discomfort, and bloating). The time to onset of clinical efficacy was evaluated by “time to first SBM” and “percentage of patients having an SBM within 24 h of first administration of study drug.” Similar assessments for CSBM were also performed. In addition, we chose to look at the stool frequency responder end point included in a recent US Food and Drug Administration guidance document (20); “CSBM stool frequency responder” was defined as a patient who had an increase of ≥1 CSBM per week over baseline for at least 4 of the 8 weeks of the treatment period.

ed. In addition, we chose to look at the stool frequency responder end point included in a recent US Food and Drug Administration guidance document (20); “CSBM stool frequency responder” was defined as a patient who had an increase of ≥1 CSBM per week over baseline for at least 4 of the 8 weeks of the treatment period. Daily symptom assessments Assessments were recorded by phone using Interactive Voice Response System. Each day, patients recorded the time study medication was taken and the number and time of each BM. The characteristics of each BM were assessed using the 7-point Bristol Stool Form Scale (18). The severity of straining was graded using a 5-point ordinal scale (1=not at all, 2=a little bit, 3=more straining than not, 4=a great deal, and 5=an extreme amount). SBM was defined as a BM in which no laxative, enema, or suppository was used in the preceding 24 h, and a CSBM was defined as an SBM, which was associated with a feeling of complete bowel emptying. Each day, patients also recorded the severity of bloating, pain, and discomfort using 5-point scales (1=none, 2=mild, 3=moderate, 4=severe, and 5=very severe) and any use of rescue medication. Surrogate markers of hepatic synthesis and GI loss of BAs Mechanistic effects of A3309 on hepatic BA synthesis rates were evaluated using plasma C4 (7α-hydroxy-4-cholesten-3-one) (21) and on plasma lipids (total cholesterol, low-density lipoprotein (LDL) cholesterol, high-density lipoprotein (HDL) cholesterol, and triglycerides) using a high-performance liquid chromatography method.

stic effects of A3309 on hepatic BA synthesis rates were evaluated using plasma C4 (7α-hydroxy-4-cholesten-3-one) (21) and on plasma lipids (total cholesterol, low-density lipoprotein (LDL) cholesterol, high-density lipoprotein (HDL) cholesterol, and triglycerides) using a high-performance liquid chromatography method. Safety assessment The intensity of adverse events (AEs) was reported by patients as mild, moderate, or severe. Separate from AE intensity, patients were defined as suffering from a serious AE (SAE) if the AE resulted in death, was deemed life threatening, led to long-term disability, or required or prolonged hospitalization for treatment. The site investigator assessed all patient-reported AEs and determined their relationship to study treatment. Safety evaluations included physical examinations, electrocardiogram, vital signs measurements, and standard laboratory tests (clinical chemistry, hematology, fat-soluble vitamins [A, D, E], and coagulation).

. The site investigator assessed all patient-reported AEs and determined their relationship to study treatment. Safety evaluations included physical examinations, electrocardiogram, vital signs measurements, and standard laboratory tests (clinical chemistry, hematology, fat-soluble vitamins [A, D, E], and coagulation). Statistical methods and data analysis Patients were randomized in a 1:1:1:1 ratio to the treatment groups. The sample size for this study was selected to provide 80% power to detect a difference of 2.5 SBMs per week between an active treatment group and the placebo group using a two-sided t-test at α=0.05 significance without adjustment for multiplicity. This calculation was based on the assumption that treatment groups have unequal variances (estimated active s.d.=4.7, estimated placebo s.d.=2.6). Thirty-eight patients per treatment group were required to achieve the desired power under the stated assumptions. To allow for a withdrawal rate of 15%, 45 patients per group (180 patients total) were required. All changes from baseline end points were analyzed with a one-way analysis of covariance, with a fixed effect for treatment group and the baseline value of SBMs used as a covariate. Within the analysis of covariance, each dose of A3309 was compared with placebo in a pairwise manner. The time to the first SBM and the first CSBM were analyzed using a one-way analysis of variance, with a fixed effect for treatment group. The binary response end point “CSBM stool frequency responder” was analyzed with Fisher's exact test comparing each A3309 dose with placebo.

mpared with placebo in a pairwise manner. The time to the first SBM and the first CSBM were analyzed using a one-way analysis of variance, with a fixed effect for treatment group. The binary response end point “CSBM stool frequency responder” was analyzed with Fisher's exact test comparing each A3309 dose with placebo. An observed-cases approach to missing data was applied (i.e., a patient's missing values were not imputed or carried forward). All P values are based on two-sided tests and given that this was a phase IIb dose-range finding study, the reported P values were not adjusted for multiple comparisons. Patients discontinuing treatment were defined as treatment failures for the responder analyses. Efficacy analyses are reported for the ITT (intent to treat) population defined as all randomized patients who received at least one dose of study medication and had any post-baseline patient diary data (BM or GI symptoms) collected (n=190). Evaluations of serum C4 and lipids, which reflect pharmacodynamic effects of A3309, were performed on the completer per protocol population (n=121) defined as all ITT patients who completed the study and were at least 80% compliant with study medication and patient reports during the treatment period.

Efficacy analyses are reported for the ITT (intent to treat) population defined as all randomized patients who received at least one dose of study medication and had any post-baseline patient diary data (BM or GI symptoms) collected (n=190). Evaluations of serum C4 and lipids, which reflect pharmacodynamic effects of A3309, were performed on the completer per protocol population (n=121) defined as all ITT patients who completed the study and were at least 80% compliant with study medication and patient reports during the treatment period. Safety results are reported for the safety population (n=189) defined as all randomized patients who received at least one dose of study treatment and for whom follow-up safety data are available. Assessments of safety were performed evaluating number/proportions of patients experiencing SAEs/AEs and with an analysis of any changes from baseline in laboratory safety tests, electrocardiogram, and vital signs. RESULTS Patient flow and follow-up Of 448 CIC patients screened, 190 were randomized and 161 completed the study (Figure 1). At baseline, the demographic parameters were similar between the treatment groups ( Table 1). Patients had a mean age of 48.1 years (range 20–79), and were predominantly female (90%) and white (78%). The mean number of SBMs and CSBMs per week at screening was 2.82 and 0.40, respectively. Full baseline characteristics of our study population can be found in Supplementary Appendix Table 1 online.

oups ( Table 1). Patients had a mean age of 48.1 years (range 20–79), and were predominantly female (90%) and white (78%). The mean number of SBMs and CSBMs per week at screening was 2.82 and 0.40, respectively. Full baseline characteristics of our study population can be found in Supplementary Appendix Table 1 online. Bowel function Stool frequency At week 1 of the double-blind treatment phase (primary efficacy end point), patients treated with 10 or 15 mg of A3309 reported a significantly (P<0.002 and P<0.001, respectively) greater change from baseline in the number of SBMs compared with placebo; Figure 2 shows a clear dose–response with mean (95% confidence interval) increases of 2.5 (1.5–3.5, NS), 4.0 (2.9–5.0, P<0.002), and 5.4 (4.4–6.4; P<0.001) SBMs for the A3309 dose levels of 5, 10, and 15 mg, compared with an increase of 1.7 (0.7–2.8) SBMs for placebo. The beneficial effect of A3309 on weekly SBM (Figure 3a) and CSBM (Figure 3b) frequency was maintained over placebo for the 8-week treatment period.

5, NS), 4.0 (2.9–5.0, P<0.002), and 5.4 (4.4–6.4; P<0.001) SBMs for the A3309 dose levels of 5, 10, and 15 mg, compared with an increase of 1.7 (0.7–2.8) SBMs for placebo. The beneficial effect of A3309 on weekly SBM (Figure 3a) and CSBM (Figure 3b) frequency was maintained over placebo for the 8-week treatment period. A3309 at doses of 10 and 15 mg significantly reduced the time to first SBM relative to placebo (P=0.033 and 0.039, respectively); an SBM occurred (mean) 12 and 7 h after first dose of A3309 10 and 15 mg, respectively, compared with (mean) 27 h for placebo. A significantly higher proportion of patients reported an SBM within 24 h of taking their first dose of A3309 at a dose of 10 (74%) or 15 mg (75%) compared with placebo (45%) (P=0.012 for both). When considering time to first CSBM or the proportion who reported a CSBM within 24 h of dosing, significant differences were again observed with the two higher doses of A3309 compared with placebo (P≤0.02 for all comparisons). CSBM stool frequency responder rates were also significantly improved with all three A3309 dose levels compared with placebo; 58% (25/43), 64% (27/42), 75% (30/40) for the 5, 10, and 15 mg groups compared with placebo (33%, 14/42) (P=0.030, 0.008, <0.001, respectively). Stool consistency and straining Patients treated with 10 and 15 mg doses of A3309 reported significant improvements in stool consistency and significant decreases in straining at all treatment weeks (P<0.001 for all comparisons) (Figure 4a and b).

CSBM stool frequency responder rates were also significantly improved with all three A3309 dose levels compared with placebo; 58% (25/43), 64% (27/42), 75% (30/40) for the 5, 10, and 15 mg groups compared with placebo (33%, 14/42) (P=0.030, 0.008, <0.001, respectively). Stool consistency and straining Patients treated with 10 and 15 mg doses of A3309 reported significant improvements in stool consistency and significant decreases in straining at all treatment weeks (P<0.001 for all comparisons) (Figure 4a and b). Abdominal symptoms Bloating, pain, and discomfort The mean change from baseline in overall bloating across the 8-week treatment period improved in patients receiving the 15-mg dose of A3309 (P=0.048) (Figure 4c). There were no differences in effects on abdominal pain or discomfort between the A3309 groups and placebo. Rescue medication use There were no significant differences in the use of rescue medication between active treatment groups and placebo at any of the study intervals.

Abdominal symptoms Bloating, pain, and discomfort The mean change from baseline in overall bloating across the 8-week treatment period improved in patients receiving the 15-mg dose of A3309 (P=0.048) (Figure 4c). There were no differences in effects on abdominal pain or discomfort between the A3309 groups and placebo. Rescue medication use There were no significant differences in the use of rescue medication between active treatment groups and placebo at any of the study intervals. Surrogate markers of hepatic BA synthesis and GI loss Plasma C4 increased in all of the active A3309 dose levels with mean (±s.e.m.) changes from baseline being 1.2 (±4.2) for placebo, 14.9 (±4.0) for the 5-mg group, 21.1 (±4.4) for the 10-mg group, and 12.9 (±5.1) for the 15-mg group (P=0.020, 0.002, and 0.082, respectively). Baseline total cholesterol levels were >200 mg/dl in 43% of the patients; the mean (±s.e.m.) baseline total plasma cholesterol level was 195 (±5.2) mg/dl (ITT population). Significant decreases in total cholesterol were observed for the 10- and 15-mg dose levels (P=0.014 and P=0.027, respectively). Compared with baseline, LDL cholesterol and the LDL/HDL ratio each decreased while on treatment with A3309 (Table 2). Neither A3309 nor placebo caused significant changes in plasma HDL cholesterol or triglycerides.

in total cholesterol were observed for the 10- and 15-mg dose levels (P=0.014 and P=0.027, respectively). Compared with baseline, LDL cholesterol and the LDL/HDL ratio each decreased while on treatment with A3309 (Table 2). Neither A3309 nor placebo caused significant changes in plasma HDL cholesterol or triglycerides. Safety assessment AEs were reported in 102 of the 189 patients (54%); 14 patients (7%) had events of severe intensity. The main difference between the groups was that five patients in the 15-mg A3309 group had abdominal cramping/pain and/or diarrhea of severe intensity. Treatment-emergent SEs occurred in 44, 46, 62, and 65% in the placebo, 5, 10, 15 mg A3309 groups, respectively. The most common AEs were of GI origin including abdominal pain (0, 10, 11, and 27%) and diarrhea (2, 8, 6, and 13%). In total, 29 patients (15%) discontinued study medication (placebo: 12.8%, 5 mg A3309: 12.5%, 10 mg A3309: 12.8%, and 15 mg A3309: 22.9%). Table 3 outlines the treatment-emergent AEs and discontinuations during the study. The rate of AEs and discontinuations was greater in the 15-mg A3309 group compared with the 5- or 10-mg A3309 or placebo groups. Discontinuations due to GI AEs are outlined in Table 3. Three SAEs were reported, one in the placebo group (shoulder pain), one in the 5-mg A3309 group (colonoscopy verified bleeding colonic diverticulum 2 weeks after the end of treatment), and one in the 10-mg A3309 group (breast carcinoma diagnosed on day 5 of treatment). None of the SAEs were considered related to the study drug. No deaths occurred during the study.

roup (shoulder pain), one in the 5-mg A3309 group (colonoscopy verified bleeding colonic diverticulum 2 weeks after the end of treatment), and one in the 10-mg A3309 group (breast carcinoma diagnosed on day 5 of treatment). None of the SAEs were considered related to the study drug. No deaths occurred during the study. There were no differences in the incidence of clinically significant changes in laboratory parameters, vital signs, or other safety parameters.

roup (shoulder pain), one in the 5-mg A3309 group (colonoscopy verified bleeding colonic diverticulum 2 weeks after the end of treatment), and one in the 10-mg A3309 group (breast carcinoma diagnosed on day 5 of treatment). None of the SAEs were considered related to the study drug. No deaths occurred during the study. There were no differences in the incidence of clinically significant changes in laboratory parameters, vital signs, or other safety parameters. DISCUSSION This randomized, placebo-controlled trial evaluated the efficacy and safety of three doses of the novel IBAT inhibitor, A3309 in patients with CIC. Compared with placebo, A3309 significantly increased SBMs during week 1 (the primary end point) and each subsequent week of the 8-week treatment period. A3309 also significantly increased weekly CSBMs, an end point that includes the patient's qualitative assessment of completeness of a bowel movement and arguably provides an even more rigorous assessment of stool frequency than SBM. A3309 led to significant improvements in other constipation-associated symptoms (stool consistency, straining, and bloating). There was also benefit for the rigorous CSBM stool frequency responder definition (an increase of ≥1 CSBM per week over baseline for at least 4 of the 8 weeks of the treatment period), with all doses of A3309 compared with placebo. The effects of A3309 on the frequency and consistency of BMs occurred within the first week of drug administration and were maintained over the full 8 weeks of treatment. Roughly three quarters of the patients in the 10- and 15-mg A3309 groups reported passing an SBM within 24 h of initial drug dosing. Importantly, there was no diminution of the clinical benefits of A3309 over time, suggesting that there was no development of tolerance.

ere maintained over the full 8 weeks of treatment. Roughly three quarters of the patients in the 10- and 15-mg A3309 groups reported passing an SBM within 24 h of initial drug dosing. Importantly, there was no diminution of the clinical benefits of A3309 over time, suggesting that there was no development of tolerance. It has long been known that BAs are physiological promoters of colonic secretion and motility (10). Reduced synthesis and/or colonic concentrations of specific BAs have been associated with the development of constipation (6,7,8). Therefore, recent work has investigated the potential role of raising colonic BA concentration (by direct supplementation or through prevention of ileal reabsorption) as a novel target for new drug development in patients with CIC (8,16). BAs are mainly absorbed in the ileum by a specific BA transporter or IBAT. The IBAT is an integral brush border membrane glycoprotein that co-transports sodium and BAs. Available evidence suggests that the IBAT is a major regulator of the BA pool in animals and humans. BA reabsorption from the intestine is very efficient; 95% of the secreted BAs are reabsorbed and delivered back to the liver completing the process of enterohepatic circulation (22).

n that co-transports sodium and BAs. Available evidence suggests that the IBAT is a major regulator of the BA pool in animals and humans. BA reabsorption from the intestine is very efficient; 95% of the secreted BAs are reabsorbed and delivered back to the liver completing the process of enterohepatic circulation (22). IBAT inhibitors prevent the reabsorption of BAs from the ileum and consequently, their return to the liver. The liver compensates for the decrease in BAs by upregulating cholesterol 7α-hydroxylase, the rate-limiting enzyme for BA synthesis. This results in increased levels of 7α-hydroxy-4-cholesten-3-one (C4), an intermediate in BA synthesis. In addition, lower hepatic cholesterol levels and increased number of LDL receptors in the liver occur leading to reduced plasma LDL cholesterol (23,24). These concepts were confirmed in our study where A3309 but not placebo resulted in increased C4 values and decreased plasma total and LDL cholesterol levels, consistent with increased BA synthesis due to inhibition of IBAT. A3309's cholesterol-lowering effect may be an incremental attribute of this treatment that may prove useful in patients with CIC.

our study where A3309 but not placebo resulted in increased C4 values and decreased plasma total and LDL cholesterol levels, consistent with increased BA synthesis due to inhibition of IBAT. A3309's cholesterol-lowering effect may be an incremental attribute of this treatment that may prove useful in patients with CIC. The improvements to constipation symptoms observed with A3309 are likely the consequence of increased delivery of BAs to the colon leading to increased fluid secretion and motility and resultant acceleration of transit. This suggestion is supported by a recent single center, randomized, double-blind dose escalation study in 30 patients with CIC, which found that A3309 significantly accelerated colon transit measured by radio-opaque markers and increased stool frequency compared with placebo (16). Contrary to the clear benefit of A3309 for constipation-associated symptoms, there was no statistically significant difference in abdominal pain or discomfort ratings between the A3309 and placebo groups. The interpretation of these findings may be confounded by low baseline scores for abdominal pain and discomfort (floor effect) likely resulting from the exclusion of patients with a prior history of IBS or abdominal pain or discomfort as a predominant complaint. Further evaluation of A3309 in patients with IBS-C should be considered to more definitively evaluate the impact of this drug on abdominal pain and abdominal discomfort.

(floor effect) likely resulting from the exclusion of patients with a prior history of IBS or abdominal pain or discomfort as a predominant complaint. Further evaluation of A3309 in patients with IBS-C should be considered to more definitively evaluate the impact of this drug on abdominal pain and abdominal discomfort. A3309 was generally well tolerated and no SAEs related to the study drug were reported. There was a dose-related trend in the number of AEs and in the percentage of patients discontinuing treatment. Because A3309 has a very low systemic exposure (resulting in estimated plasma concentrations in the picomolar range), it is reasonable to speculate that AEs, which were largely GI in origin, were the consequence of the increased concentration of colonic BAs (11,25). Based upon data from this phase IIb dose-range study, it appears that the 10-mg dose of A3309 provided the best balance of efficacy and safety in patients with CIC. Patients on the 15-mg dose level had more AEs of abdominal pain and diarrhea. It is conceivable that this may reflect a direct effect of the higher dose of A3309 on motility, rather than a secondary effect on transit arising from increased colonic secretion. Concentrations of >5 m of BAs infused directly into the human rectum and sigmoid stimulate colonic phasic contractions (11); in addition, Bampton et al. (26) observed HAPCs in the colon with administration of 1 m chenodexoycholate into the rectum of healthy volunteers. Such concentrations of 1–5 m BAs are seldom achieved in the absence of ileal resection (27).

ctly into the human rectum and sigmoid stimulate colonic phasic contractions (11); in addition, Bampton et al. (26) observed HAPCs in the colon with administration of 1 m chenodexoycholate into the rectum of healthy volunteers. Such concentrations of 1–5 m BAs are seldom achieved in the absence of ileal resection (27). Despite the chronic nature of CIC, many patients use medical therapy intermittently; the speed of onset of A3309 is also potentially beneficial as it significantly shortened the time to first SBM and CSBM. Treatment with A3309 may therefore be suited for individualized treatment, either on a fixed schedule or “on demand.” Further studies specifically designed to address these potential uses for A3309 in CIC patients are eagerly awaited. Forty-three percent of CIC patients enrolled in this study had baseline total plasma cholesterol values above the desirable cutoff level of 200 mg/dl. A3309 treatment induced a decrease in LDL cholesterol and in the important predictor of cardiovascular disease—the LDL/HDL ratio—(14% decreases in the 10-mg dose group for both variables). The beneficial effects on lipid profile are a unique feature of A3309 among current medications and those in the pipeline for treatment of CIC. This improvement in serum lipids might provide incremental benefits to a subset of patients with CIC.

LDL/HDL ratio—(14% decreases in the 10-mg dose group for both variables). The beneficial effects on lipid profile are a unique feature of A3309 among current medications and those in the pipeline for treatment of CIC. This improvement in serum lipids might provide incremental benefits to a subset of patients with CIC. When considering our results, several additional issues merit discussion. As is common in drug trials of this size, a large number of investigative sites were included and patient-reported outcomes were captured using an Interactive Voice Response System or with electronic handheld devices. Studies that rely upon electronic data capture might bias the randomized population toward a younger, more affluent demographic. What effect, if any, this could have on our results is unknown. Patients with CIC entered into this study were required to be passing fewer than 3 CSBMs per week. This may limit the generalizability of our data to the entire population of CIC patients. Acknowledging this point, the decision for this entry requirement was largely predicated upon current regulatory recommendations, which support stool frequency as the primary outcome for trials evaluating therapies for CIC. Further, though weekly CSBM rate was used to determine study eligibility, weekly SBM rate after the first week of therapy served as the primary outcome measure. This primary outcome was chosen after consultation with the US Food and Drug Administration. In addition to meeting the primary outcome, the two higher doses of A3309 provided robust and significant benefits to the rate of weekly SBMs and CSBMs for the entire 8-week treatment period. One further point, which merits consideration, is that our study population likely included a subset of patients with dyssynergic defecation. The study protocol called for the exclusion of patients with a history of dyssynergic defecation or in whom the history and physical examination was felt to indicate the presence of this type of constipation. We acknowledge that this is an imperfect way of identifying patients with dyssynergic defecation and that the inclusion of such patients could have affected the study results. However, we would point out that nearly all other recently performed trials evaluating novel drug therapies for chronic constipation have handled the issue of dyssynergic defecation in a similar manner (28,29).

ts with dyssynergic defecation and that the inclusion of such patients could have affected the study results. However, we would point out that nearly all other recently performed trials evaluating novel drug therapies for chronic constipation have handled the issue of dyssynergic defecation in a similar manner (28,29). In addition, randomization would be expected to level the effect of the inclusion of such patients on results yielded by the study groups. We would also argue that the practice used in this protocol closely mimics clinical practice where diagnostic testing for dyssynergic defecation is typically reserved for patients who have failed multiple laxative therapies as recommended by the American Gastroenterological Association (30) and the Rome Foundation (31). In conclusion, the results of this placebo-controlled, dose-range, phase IIb study in CIC patients demonstrated that the IBAT inhibitor A3309 significantly improved BM frequency, stool consistency, and constipation symptoms. The 10-mg dose of A3309 provided the best balance of efficacy and safety. Further large clinical trials evaluating the efficacy of A3309 in patients with CIC and IBS-C are warranted. STUDY HIGHLIGHTS We thank the ACCESS Study team investigators and nurses for their participation in this study. SUPPLEMENTARY MATERIAL is linked to the online version of the paper at http://www.nature.com/ajg Guarantor of the article: William D. Chey, MD, AGAF, FACG, FACP.

In conclusion, the results of this placebo-controlled, dose-range, phase IIb study in CIC patients demonstrated that the IBAT inhibitor A3309 significantly improved BM frequency, stool consistency, and constipation symptoms. The 10-mg dose of A3309 provided the best balance of efficacy and safety. Further large clinical trials evaluating the efficacy of A3309 in patients with CIC and IBS-C are warranted. STUDY HIGHLIGHTS We thank the ACCESS Study team investigators and nurses for their participation in this study. SUPPLEMENTARY MATERIAL is linked to the online version of the paper at http://www.nature.com/ajg Guarantor of the article: William D. Chey, MD, AGAF, FACG, FACP. Specific author contributions: Full access to all the data in the study and took responsibility for the integrity of the data and the accuracy of the data analyses and wrote the first draft of the manuscript: William D. Chey and Hans Graffner. No professional writing assistance was utilized in the preparation of this manuscript. All authors contributed to study concept and design and data analyses, drafted and reviewed the manuscript revisions, and approved the final draft for submission. The statistical analysis of the entire data sets has been independently analyzed by a biostatistician not employed by Albireo and with no links to the institution performing the study. Financial support: Albireo funded this study and provided A3309.

Specific author contributions: Full access to all the data in the study and took responsibility for the integrity of the data and the accuracy of the data analyses and wrote the first draft of the manuscript: William D. Chey and Hans Graffner. No professional writing assistance was utilized in the preparation of this manuscript. All authors contributed to study concept and design and data analyses, drafted and reviewed the manuscript revisions, and approved the final draft for submission. The statistical analysis of the entire data sets has been independently analyzed by a biostatistician not employed by Albireo and with no links to the institution performing the study. Financial support: Albireo funded this study and provided A3309. Potential competing interests: Rikner and Graffner are employees of Albireo. Chey is a consultant for Albireo. Camilleri receives a research grant from Albireo, and serves as a consultant for Albireo with payment of consultant fee to Mayo Clinic. Chang serves as a consultant for Albireo. Supplementary Material Supplementary Appendix Table 1 Click here for additional data file. Figure 1 Patient flow diagram. Completer PP (per protocol) population=all intention to treat (ITT) patients who completed the study and were at least 80% compliant with study medication and patient reports during the treatment period. AE, adverse event. Figure 2 Change from baseline in weekly spontaneous bowel movement (SBM) frequency for treatment week 1. ITT, intention to treat.

Figure 1 Patient flow diagram. Completer PP (per protocol) population=all intention to treat (ITT) patients who completed the study and were at least 80% compliant with study medication and patient reports during the treatment period. AE, adverse event. Figure 2 Change from baseline in weekly spontaneous bowel movement (SBM) frequency for treatment week 1. ITT, intention to treat. Figure 3 Effect of A3309 and placebo on weekly stool frequency. Mean weekly SBM (a) and CSBM (b) frequency. Significance values shown are for the overall change from baseline between A3309 and placebo. ***P<0.001, *P<0.05. CSBM, complete spontaneous bowel movement; ITT, intention to treat; SBM, spontaneous bowel movement. Figure 4 Effect of A3309 and placebo on weekly stool consistency, straining, and bloating. Changes in stool consistency (a), straining (b), and bloating (c) over the treatment period. Significance values shown are for the overall change from baseline between A3309 and placebo. ***P<0.001, *P<0.05. BSFS, Bristol Stool Form Scale (7-graded scale, see text for explanations); ITT, intention to treat.

d bloating. Changes in stool consistency (a), straining (b), and bloating (c) over the treatment period. Significance values shown are for the overall change from baseline between A3309 and placebo. ***P<0.001, *P<0.05. BSFS, Bristol Stool Form Scale (7-graded scale, see text for explanations); ITT, intention to treat. Table 1 Patient demographics—ITT population Placebo A3309 Overall 5 mg 10 mg 15 mg N 47 48 47 48 190 Age, years Mean 49.9 48.0 48.4 46.2 48.1 Min–max 21–72 20–79 22–66 20–72 20–79 Age group <60 years 33 (70.2) 40 (83.3) 39 (83.0) 40 (83.3) 152 (80.0) ≥60 years 14 (29.8) 8 (16.7) 8 (17.0) 8 (16.7) 38 (20.0) Gender Male 8 (17.0) 5 (10.4) 2 (4.3) 5 (10.4) 20 (10.5) Female 39 (83.0) 43 (89.6) 45 (95.7) 43 (89.6) 170 (89.5) Race, n (%) White 35 (74.5) 41 (85.4) 36 (76.6) 36 (75.0) 148 (77.9) Black/African American 12 (25.5) 7 (14.6) 10 (21.3) 12 (25.0) 41 (21.6) BMI, kg/m2 Mean 26.87 26.09 27.09 27.71 26.94 s.d. 4.330 4.327 3.728 3.841 4.076 BMI, body mass index; ITT, intention to treat; min–max, minimum and maximum age. Numbers in parentheses are percentages. Table 2 Mean change from baseline for plasma LDL and LDL/HDL ratio Placebo A3309 5 mg 10 mg 15 mg N 32 37 29 22 LDL Mean change 2.6 −2.4 −15.6*** −11.1** s.d. 17.30 15.61 18.23 22.90 LDL/HDL Mean change −0.024 −0.111 −0.287** −0.221* s.d. 0.333 0.405 0.318 0.393 HDL, high-density lipoprotein; ITT, intention to treat; LDL, low-density lipoprotein.

eline for plasma LDL and LDL/HDL ratio Placebo A3309 5 mg 10 mg 15 mg N 32 37 29 22 LDL Mean change 2.6 −2.4 −15.6*** −11.1** s.d. 17.30 15.61 18.23 22.90 LDL/HDL Mean change −0.024 −0.111 −0.287** −0.221* s.d. 0.333 0.405 0.318 0.393 HDL, high-density lipoprotein; ITT, intention to treat; LDL, low-density lipoprotein. Completer PP (per protocol) population (all ITT patients who completed the study and were at least 80% compliant with study medication and patient reports during the treatment period). ***P<0.001, **P=0.01,*P<0.05. Table 3 Discontinuations and related treatment-emergent gastrointestinal adverse events (TEAEs) Placebo (N=46) A3309 5 mg (N=48) A3309 10 mg (N=47) A3309 15 mg (N=48) Any TEAE, n (%) 20 (44) 22 (46) 29 (62) 31 (65) Treatment discontinuations 6 (12.8) 6 (12.5) 6 (12.8) 11 (22.9) Gastrointestinal disorders Total 5 (10.9) 12 (25.0) 11 (23.4) 19 (39.6) Mild 2 (4.3) 6 (12.5) 5 (10.6) 6 (12.5) Moderate 2 (4.3) 5 (10.4) 4 (8.5) 8 (16.7) Severe 1 (2.2) 1 (2.1) 2 (4.3) 5 (10.4) Abdominal pain 0 5 (10.4) 5 (10.6)a 13 (27.1)aaaa Diarrhea 1 (2.2) 4 (8.3) 3 (6.4)a 6 (12.5)aaaa Abdominal distension 1 (2.2) 4 (8.3) 2 (4.3)a 3 (6.3)aa Flatulence 3 (6.5) 4 (8.3) 3 (6.4)a 2 (4.2) Nausea 2 (4.3) 2 (4.2) 2 (4.3) 3 (6.3) Defaecation urgency 0 0 0 2 (4.2) Abdominal pain upper 1 (2.2) 1 (2.1) 0 1 (2.1) a Drug withdrawn. Numbers in parentheses are percentages.

INTRODUCTION Hepatocellular carcinoma (HCC) is the fifth most common malignant neoplasm in the world (1). Only 20% of HCC patients are candidates for resection (2). Furthermore, recurrence is frequent even after apparently curative resection. Liver transplantation is restricted by organ donor shortage. Thus, various nonsurgical therapies have been introduced (3,4,5). Among these, image-guided percutaneous ablation is considered best for early-stage HCC. Ethanol injection was formerly the standard procedure among the various percutaneous ablation techniques. Randomized controlled trials, however, have demonstrated that radiofrequency ablation (RFA) has a more reliable local antitumor effect, leading to a lower local tumor progression risk and higher survival rates (6,7,8,9). RFA has largely replaced ethanol injection (10). Several reports on 5-year outcome of RFA exist (11,12,13,14,15,16,17); however, no study has covered 10-year outcome. We report on a 10-year consecutive case series at a tertiary referral center. We analyzed antitumor effect, patient survival, local tumor progression, and distant recurrence rates, variables relevant to these outcomes, and complications. To our knowledge, this study documents the largest number of RFA treatments performed at a single institution.

cutive case series at a tertiary referral center. We analyzed antitumor effect, patient survival, local tumor progression, and distant recurrence rates, variables relevant to these outcomes, and complications. To our knowledge, this study documents the largest number of RFA treatments performed at a single institution. METHODS RFA indications RFA was the treatment of choice in HCC patients satisfying the following criteria: (i) ineligible for surgical resection/liver transplantation or patient refusal for surgery; (ii) no extrahepatic metastasis/vascular invasion; and (iii) no other malignancies that may determine the patient's prognosis. Exclusion criteria were as follows: (i) tumor not visualized by ultrasonography/not accessible percutaneously; (ii) total bilirubin level ≥3.0 mg/dl; (iii) platelet count <50 × 109/l or prothrombin activity <50% (iv) refractory ascites; (v) enterobiliary reflux; and (vi) adhesion between the tumor and the gastrointestinal tract. In general, we performed RFA on Child-Pugh class A or B patients, a single tumor ≤5 cm in diameter, or three or fewer tumors ≤3 cm in diameter. In cases beyond these conditions, we performed RFA on patients who were likely to benefit from this procedure for possible cure or prolongation of life. No patients were excluded solely on account of tumor location (18). Informed consent was obtained from each patient. This study was conducted according to the ethical guidelines of the 1975 Declaration of Helsinki and approved by the institutional review board (Registration ID: P98C05-11Y).

ure or prolongation of life. No patients were excluded solely on account of tumor location (18). Informed consent was obtained from each patient. This study was conducted according to the ethical guidelines of the 1975 Declaration of Helsinki and approved by the institutional review board (Registration ID: P98C05-11Y). Patients In this cohort study, we analyzed a prospectively collected computerized database. Between February 1999 and December 2009, 2,825 HCC patients were admitted once or more to the Department of Gastroenterology, the University of Tokyo (Figure 1). At initial hospitalization, 1,485 had primary HCC and the remaining 1,340 had recurrent HCC. In the recurrent HCC patients, primary HCC had previously been treated by therapies other than RFA. Of the 1,485 primary HCC patients, 1,294 (87.1%) underwent percutaneous ablation as the initial treatment, including RFA. The remaining 191 patients underwent other therapies: hepatic resection, 29 patients with good liver function and who consented to an operation; transarterial chemoembolization, 149 with multinodular or large tumors that could not be treated by ablation therapies; systemic chemotherapy, three with extrahepatic metastasis; and only supportive care, 10 with decompensated cirrhosis or poor general condition.

h good liver function and who consented to an operation; transarterial chemoembolization, 149 with multinodular or large tumors that could not be treated by ablation therapies; systemic chemotherapy, three with extrahepatic metastasis; and only supportive care, 10 with decompensated cirrhosis or poor general condition. Of the 1,294 patients treated by percutaneous ablation, 1,170 underwent RFA and the other 124 underwent ethanol injection. The choice of therapy was made as follows: between April 1999 and January 2001, 232 patients with three or fewer tumors, each ≤3 cm in diameter, and Child-Pugh class A or B liver function were entered into a randomized controlled trial to compare RFA with ethanol injection (6). Patients outside these inclusion criteria were mostly treated by RFA. After this trial, RFA was generally the treatment of choice, and ethanol injection was administered only to those considered unsuitable for RFA; ethanol injection was administered to those with either enterobiliary reflux or adhesion of the tumor to the gastrointestinal tract. HCC was diagnosed based on typical imaging findings; that is, early-phase enhancement and late-phase contrast washout on dynamic computed tomography (CT) (19). HCC diagnosis was also confirmed by biopsy in 1,078 (92.1%) of the 1,170 patients with RFA-treated primary HCC. A total of 998 (85.3%) were diagnosed as having liver cirrhosis.

d based on typical imaging findings; that is, early-phase enhancement and late-phase contrast washout on dynamic computed tomography (CT) (19). HCC diagnosis was also confirmed by biopsy in 1,078 (92.1%) of the 1,170 patients with RFA-treated primary HCC. A total of 998 (85.3%) were diagnosed as having liver cirrhosis. In general, transarterial chemoembolization was combined with RFA in patients with either ≥4 tumors or those with even one tumor >3.0 cm in diameter, although indication criteria of this combination had changed over time. The combination of transarterial chemoembolization with RFA was performed in 324 primary HCC patients. Treatment methods RFA was performed on an inpatient basis. Preoperative planning including evaluation of all imaging studies, and careful ultrasound examination was performed to identify the tumors and determine the access routes. The procedure was performed according to an institutional protocol and in the presence of three physicians. One physician inserted the electrode under ultrasound guidance while another assisted the procedure; at least one had 8-year or longer experience of percutaneous ablation therapies. The remaining physician was responsible for the ultrasound machine and data recording. Video recording was performed occasionally to improve and standardize the procedure.

ode under ultrasound guidance while another assisted the procedure; at least one had 8-year or longer experience of percutaneous ablation therapies. The remaining physician was responsible for the ultrasound machine and data recording. Video recording was performed occasionally to improve and standardize the procedure. The precise techniques of RFA are described elsewhere (6). Briefly, all RFA procedures were performed percutaneously under ultrasound guidance (Power Vision 8000, Aplio XV or Aplio XG; Toshiba, Tokyo, Japan). We used artificial pleural effusion (20) or artificial ascites (21) for tumors, which were in the hepatic dome or adjacent to the gastrointestinal tract. After administration of sedatives and local anesthesia, a 17-gauge cooled-tip electrode (Cool-Tip; RF Ablation System, Covidien, Boulder, Colombia, CO) was inserted. Radiofrequency energy was delivered for 6–12 min for each application. For large tumors, the electrode was repeatedly inserted into different sites, such that the entire tumor could be enveloped by assumed necrotic volumes. Following the procedure, the patient remained in bed until the next morning.

, CO) was inserted. Radiofrequency energy was delivered for 6–12 min for each application. For large tumors, the electrode was repeatedly inserted into different sites, such that the entire tumor could be enveloped by assumed necrotic volumes. Following the procedure, the patient remained in bed until the next morning. A CT scan with a 5-mm section thickness was performed 1–3 days after RFA to evaluate technique effectiveness (22). Complete ablation was defined as hypoattenuation of the entire tumor. We intended to ablate not only the tumor but also some of the liver parenchyma surrounding it. When we suspected that unablated tumor portions remained, the procedure was repeated. We did not predefine the procedure number in a treatment: treatment was generally continued until CT imaging demonstrated necrosis of the entire tumor. Follow-up To detect recurrence at an early stage, serum α-fetoprotein (AFP), lectin-reactive AFP (AFP-L3), and des-γ-carboxy-prothrombin (DCP) levels were measured monthly, and CT and ultrasonography were performed every 4 months. Local tumor progression was defined as the appearance of viable cancer tissue touching the initially treated tumor (22) and distant recurrence as the emergence of one or several tumor(s) separate from the primary site. Chest CT or bone scintigraphy was performed if extrahepatic recurrence was suspected. RFA was used for recurrence if the patient still met the indication criteria. If multiple recurrences were not treatable with RFA, transarterial chemoembolization was generally performed.

veral tumor(s) separate from the primary site. Chest CT or bone scintigraphy was performed if extrahepatic recurrence was suspected. RFA was used for recurrence if the patient still met the indication criteria. If multiple recurrences were not treatable with RFA, transarterial chemoembolization was generally performed. Statistical analysis This is a report of a consecutive case series: all RFA treatments performed on primary HCC patients at the Department of Gastroenterology, University of Tokyo between February 1999 and December 2009 were included and none was excluded. Data are presented as mean±s.d. for quantitative variables, and as absolute frequencies for qualitative variables. A “procedure” was defined as a single intervention episode comprising one or more ablation performed on one or more tumors and a “treatment” as the completed effort to ablate one or more tumors. A treatment comprised one or more procedures (22). “Technique effectiveness” rate was defined as the percentage of successfully eradicated macroscopic tumors, as evidenced by CT scan 1–3 days after the last procedure (22).

A “procedure” was defined as a single intervention episode comprising one or more ablation performed on one or more tumors and a “treatment” as the completed effort to ablate one or more tumors. A treatment comprised one or more procedures (22). “Technique effectiveness” rate was defined as the percentage of successfully eradicated macroscopic tumors, as evidenced by CT scan 1–3 days after the last procedure (22). Overall survival was calculated in the 1,170 primary HCC patients. Survival curves were generated by the Kaplan–Meier method. In addition to overall survival, some subgroup analyses were performed with clinical characteristics including tumor size, tumor number, and liver function. Recurrence was evaluated in 1,138 of the 1,170 primary HCC patients; the remaining 32 patients were excluded from the recurrence analysis because some small tumors had been left untreated by RFA on account of detection failure by ultrasonography. Recurrence rates were calculated by the Gaynor's method (23). All time estimates were made from the date of the first RFA. The follow-up was finalized at either death or the last visit to the outpatient clinic before 31 December 2009. Transplanted patients were censored from this study at the date of transplantation.

. Recurrence rates were calculated by the Gaynor's method (23). All time estimates were made from the date of the first RFA. The follow-up was finalized at either death or the last visit to the outpatient clinic before 31 December 2009. Transplanted patients were censored from this study at the date of transplantation. The prognostic relevance of 19 baseline variables ( Table 1), the combination of transcatheter arterial chemoembolization (TACE) with RFA, HCC recurrence, and the number of RFA sessions to survival was analyzed by univariate and multivariate Cox proportional hazards regression models. The prognostic relevance of 19 baseline variables ( Table 1), the combination of TACE with RFA, and the number of RFA sessions to local tumor progression and distant recurrence was also analyzed by univariate and multivariate models. All variables with a P value <0.05 by univariate comparison were subjected to multivariate analysis. Some continuous variables in which log-linearity could not be assumed were transformed into categorical variables. In multivariate analysis, we evaluated two models that contained either Child-Pugh class or its components to avoid multicollinearity. A stepwise variable selection was performed with Akaike Information Criteria in multivariate analysis. The results of multivariate analyses were presented as a hazard ratio with corresponding 95% confidence interval (CI), with P values from the Wald test. All significance tests were two-tailed, and differences with a P value <0.05 were considered statistically significant.

Information Criteria in multivariate analysis. The results of multivariate analyses were presented as a hazard ratio with corresponding 95% confidence interval (CI), with P values from the Wald test. All significance tests were two-tailed, and differences with a P value <0.05 were considered statistically significant. Complications were defined according to the guidelines of the Society of Interventional Radiology (24). RESULTS Antitumor effect We performed a total of 2,982 RFA treatments for the 1,170 primary HCC patients, comprising 4,514 procedures. Thus, procedure number per treatment was 1.52±0.78. Many patients undergoing RFA for treatment of primary HCC received iterative RFA treatments for recurrence. A total of 485 patients underwent RFA treatment once, 247 twice, 177 thrice, 94 four times, 70 five times, 35 six times, 23 seven times, 14 eight times, 7 nine times, 7 ten times, 6 eleven times, 2 twelve times, 2 thirteen times, and 1 fourteen times.

ent of primary HCC received iterative RFA treatments for recurrence. A total of 485 patients underwent RFA treatment once, 247 twice, 177 thrice, 94 four times, 70 five times, 35 six times, 23 seven times, 14 eight times, 7 nine times, 7 ten times, 6 eleven times, 2 twelve times, 2 thirteen times, and 1 fourteen times. Technique effectiveness rate was 99.4% (2,964/2,982 treatments). It was similar between the initial RFA treatments and the other RFA treatments for recurrence (P=0.98). Complete ablation of the tumor was achieved in 1,163 (99.4%) of the 1,170 initial treatments and in 1,801 (99.4%) of the 1,812 other RFA treatments. However, technique effectiveness rate significantly differed with tumor size (P=0.023). No apparent viable portions remained in the treated tumor in 1,642 (99.6%) of 1,648 treatments for tumors ≤2.0 cm in diameter, in 923 (99.2%) of 930 treatments for tumors 2.1–3.0 cm, in 356 (98.9%) of 360 treatments for tumors 3.1–5.0 cm, and in 43 (97.7%) of 44 treatments for tumors >5.0 cm. Final CT imaging demonstrated residual cancer tissue in the remaining 18 treatments. We decided against performing additional procedures, because liver failure rather than HCC seemed to determine the prognosis in 10 treatments, and because additional RFA would have caused complications on account of poor visualization or inaccessibility in the other eight treatments.

sue in the remaining 18 treatments. We decided against performing additional procedures, because liver failure rather than HCC seemed to determine the prognosis in 10 treatments, and because additional RFA would have caused complications on account of poor visualization or inaccessibility in the other eight treatments. Survival The 19 baseline clinical characteristics of the 1,170 patients who underwent RFA for treatment of primary HCC are shown in Table 1. A total of 269 patients (23.0%) were >75 years old. In all, 422 patients had tumors ≤2.0 cm in diameter, 467 had tumors 2.1–3.0 cm, 246 had tumors 3.1–5.0 cm, and 35 had tumors >5.0 cm; 685 patients had 1 tumor, 395 had 2 or 3 tumors, and 90 had ≥4 tumors. As of December 2009 (with a median follow-up of 38.2 months), 692 patients (59.1%) remained alive, 39 (3.3%) were lost to follow-up, and 439 (37.5%) had died. Of the 1,170 patients, two were transplanted. The number of 5-, 7-, and 10-year survivors was 325, 131, and 16, respectively. The cause of death was HCC in 245 patients (55.8%), liver failure in 89 (20.3%), upper gastrointestinal bleeding in 11 (2.5%), complications related to the procedure in 3 (0.7%), liver-unrelated diseases in 81 (18.5%), and undetermined in 10 (2.3%).

r of 5-, 7-, and 10-year survivors was 325, 131, and 16, respectively. The cause of death was HCC in 245 patients (55.8%), liver failure in 89 (20.3%), upper gastrointestinal bleeding in 11 (2.5%), complications related to the procedure in 3 (0.7%), liver-unrelated diseases in 81 (18.5%), and undetermined in 10 (2.3%). The 1-, 3-, 5-, 7-, and 10-year survival rates of all 1,170 primary HCC patients were 96.6% (95% CI: 95.5–97.7%), 80.5% (95% CI: 78.0–83.1%), 60.2% (95% CI: 56.7–63.9%), 45.1% (95% CI: 40.9–49.6%), and 27.3% (95% CI: 21.5–34.7%), respectively (Figure 2; Table 2). Survival rates differed significantly with tumor size (P<0.0001), tumor number (P=0.0003), and Child-Pugh class (P<0.0001). In the Child-Pugh class A or B patients with a single tumor ≤5 cm in diameter, or three or fewer tumors ≤3 cm in diameter, the 5-year survival rate was 63.8% (95% CI: 59.9–67.9%), while in those outside these criteria, it was 46.4% (95% CI: 39.4–54.8%). Univariate analysis showed 19 of the 22 variables relevant to survival. In multivariate analysis that contained Child-Pugh class but not its components, a model that contained age, antibody to hepatitis C virus (anti-HCV), Child-Pugh class, tumor size, tumor number, serum DCP level, and serum AFP-L3 level was selected ( Table 3). The other model that contained the components of Child-Pugh class is shown in Supplementary Table online.

-Pugh class but not its components, a model that contained age, antibody to hepatitis C virus (anti-HCV), Child-Pugh class, tumor size, tumor number, serum DCP level, and serum AFP-L3 level was selected ( Table 3). The other model that contained the components of Child-Pugh class is shown in Supplementary Table online. Recurrence Recurrence developed in 741 patients. Local tumor progression alone was found in 25, local tumor progression with distant recurrence was found in 9, and distant recurrence alone was found in the other 707 patients. Of these 707 patients, 13 had the first recurrence in extrahepatic sites: 7 had lymph node metastasis, 3 had peritoneal seeding, 1 had lung metastasis, 1 had bone metastasis, and the remainder had both peritoneal seeding and lung metastasis. No recurrence developed in the remaining 397 patients. Of the 741 patients, the first recurrence was treated by iterative RFA in 659 (88.9%), transarterial chemoembolization in 69 (9.3%), systemic chemotherapy in 4 (0.5%), surgical resection in 3 (0.4%), radiation therapy in 2 (0.3%), and supportive care in 4 (0.5%).

Recurrence Recurrence developed in 741 patients. Local tumor progression alone was found in 25, local tumor progression with distant recurrence was found in 9, and distant recurrence alone was found in the other 707 patients. Of these 707 patients, 13 had the first recurrence in extrahepatic sites: 7 had lymph node metastasis, 3 had peritoneal seeding, 1 had lung metastasis, 1 had bone metastasis, and the remainder had both peritoneal seeding and lung metastasis. No recurrence developed in the remaining 397 patients. Of the 741 patients, the first recurrence was treated by iterative RFA in 659 (88.9%), transarterial chemoembolization in 69 (9.3%), systemic chemotherapy in 4 (0.5%), surgical resection in 3 (0.4%), radiation therapy in 2 (0.3%), and supportive care in 4 (0.5%). The 1-, 3-, 5-, 7-, and 10-year rates of local tumor progression with or without distant recurrence were 1.4% (95% CI: 0.7–2.1%), 3.2% (95% CI: 2.1–4.3%), 3.2% (95% CI: 2.1–4.3%), 3.2% (95% CI: 2.1–4.3%), and 3.2% (95% CI: 2.1–4.3%), respectively (Figure 3). Univariate analysis demonstrated that prothrombin time and serum AFP, DCP, and AFP-L3 levels were correlated to local tumor progression, whereas multivariate analysis showed that serum DCP level alone was significantly correlated. Tumor size was not correlated to local tumor progression.

4.3%), respectively (Figure 3). Univariate analysis demonstrated that prothrombin time and serum AFP, DCP, and AFP-L3 levels were correlated to local tumor progression, whereas multivariate analysis showed that serum DCP level alone was significantly correlated. Tumor size was not correlated to local tumor progression. The 1-, 3-, 5-, 7-, and 10-year rates of distant recurrence without local tumor progression were 25.6% (95% CI: 23.0–28.2%), 63.3% (95% CI: 60.2–66.4%), 74.8% (95% CI: 71.8–77.8%), 78.1% (95% CI: 75.1–81.2%), and 80.8% (95% CI: 77.4–84.3%), respectively. Univariate analysis demonstrated 14 variables relevant to distant recurrence, whereas multivariate analysis showed that anti-HCV, Child-Pugh class, platelet count, tumor size, tumor number, serum AFP level, and serum DCP level were significantly related to distant recurrence ( Table 3). Complications A total of 67 complications were encountered ( Table 4). The incidence rates of complications per treatment and per procedure were 2.2% (67/2,982) and 1.5% (67/4,514), respectively. One patient died of hepatic failure on account of massive hepatic infarction 7 days after the last RFA procedure. He had undergone 12 RFA treatments in 8 years. The treatment mortality rate was 0.03%.

ncidence rates of complications per treatment and per procedure were 2.2% (67/2,982) and 1.5% (67/4,514), respectively. One patient died of hepatic failure on account of massive hepatic infarction 7 days after the last RFA procedure. He had undergone 12 RFA treatments in 8 years. The treatment mortality rate was 0.03%. DISCUSSION This study describes our 10-year clinical experience with RFA at a high-volume center. We performed the 2,982 RFA treatments on a total of the 1,170 primary HCC patients, showing that RFA has a high antitumor effect. Tumors were judged to have been completely ablated by final CT imaging in 99.4% of the treatments. Complete response was achieved not only in the first RFA but also in iterative RFA for recurrence. Although complete response rate differed with tumor size, there was not a sharp drop-off in effectiveness. The complete response rate may be higher in this study than others probably because we generally repeated the procedure until CT imaging demonstrated complete tumor necrosis, whereas many other studies limited the procedure number of RFA to 2–3 (11,13,15). Complete ablation of tumors has been reported to be related to improved survival (25). There were the 18 treatments in which we did not perform additional RFA for residual cancer tissue. In those treatments, usefulness of RFA had been unclear at the initial session because of liver dysfunction or tumor burden.

,13,15). Complete ablation of tumors has been reported to be related to improved survival (25). There were the 18 treatments in which we did not perform additional RFA for residual cancer tissue. In those treatments, usefulness of RFA had been unclear at the initial session because of liver dysfunction or tumor burden. This study shows that RFA could achieve long-term survival for as long as 10 years. Sixteen patients treated by RFA survived for >10 years. The variables relevant to survival were similar to those found in previous studies on ethanol injection (26,27), RFA, hepatic resection (28), and transarterial chemoembolization (29). Both liver function and tumor-related factors were associated with survival. In addition, age and anti-HCV were relevant to survival in this study. Age was among the prognostic factors, probably because 23.0% of the patients were >75 years old, which resulted in a higher percentage (18.5%) of liver-unrelated deaths in this study compared with others. Anti-HCV was among the prognostic factors, probably because anti-HCV-positive patients developed distant recurrence more frequently.

prognostic factors, probably because 23.0% of the patients were >75 years old, which resulted in a higher percentage (18.5%) of liver-unrelated deaths in this study compared with others. Anti-HCV was among the prognostic factors, probably because anti-HCV-positive patients developed distant recurrence more frequently. HCC frequently recurred after RFA; most recurrences were, however, not local tumor progression but distant recurrence. Frequent recurrence is not specific to RFA. After hepatic resection, the tumor recurrence rate exceeds 70% at 5 years (30,31). In this study, periodic follow-up detected most recurrences at limited stage. RFA was performed again for first recurrence in almost 90% of cases, although multimodal treatments were used in a long-term follow-up. On the other hand, repeat resection rate for first recurrence has been reported to range from 10.4 to 30.6% (31,32). Because RFA is less invasive than hepatic resection, iterative RFA can be performed for recurrence more easily.

in almost 90% of cases, although multimodal treatments were used in a long-term follow-up. On the other hand, repeat resection rate for first recurrence has been reported to range from 10.4 to 30.6% (31,32). Because RFA is less invasive than hepatic resection, iterative RFA can be performed for recurrence more easily. Local tumor progression was found less frequently in this study than in other studies, having been reported to be around 10% at 3 years following RFA (13,14). Furthermore, different from the findings in previous reports (33,34), tumor size was not related to local tumor progression in this study. These differences are probably because we repeated RFA until we considered we had ablated not only the tumor itself but also some of the liver tissue surrounding it. Furthermore, to avoid local tumor progression, we were more cautious in the treatment of larger tumors when deciding whether sufficient ablation had been performed. Only serum DCP level was significantly related to local tumor progression in this study. Elevated serum DCP level may be related to the malignant potential of HCC such as the development of portal venous invasion (35).

tious in the treatment of larger tumors when deciding whether sufficient ablation had been performed. Only serum DCP level was significantly related to local tumor progression in this study. Elevated serum DCP level may be related to the malignant potential of HCC such as the development of portal venous invasion (35). The frequency of distant recurrence in this study was similar to that reported in other studies (13). Among the variables significantly related to distant recurrence, tumor size, tumor number, serum AFP level, and serum DCP level were probably related to micrometastasis, which had not been detected by imaging modalities before the treatment, while anti-HCV, Child-Pugh class, and platelet count were related to metachronous multicentric carcinogenesis, which developed based on underlying chronic liver disease. From the viewpoint of survival and distant recurrence, patients with 2.1–5.0 cm tumors had significantly worse outcomes than those with ≤2.0 cm tumors while those with tumors >5.0 cm did not have worse rates than those with tumors ≤2 cm. This is probably because the number of patients with tumors >5.0 cm (n=35) were not large enough for the difference to be statistically significant. Another possibility is selection bias. It is possible that patient with tumors >5.0 cm who underwent RFA had more favorable conditions for survival and distant recurrence except tumor size than those with 2.1–5.0 cm tumors.

tumors >5.0 cm (n=35) were not large enough for the difference to be statistically significant. Another possibility is selection bias. It is possible that patient with tumors >5.0 cm who underwent RFA had more favorable conditions for survival and distant recurrence except tumor size than those with 2.1–5.0 cm tumors. In this study, 324 of the 1,170 patients were treated with combination of TACE and RFA at the initial treatment. Thus, we evaluated the combination as a possible variable that influences survival or recurrence. Univariate analysis demonstrated that the combined therapy was significantly correlated to overall survival, whereas multivariate analysis did not show the relationship. TACE was generally combined with RFA in patients with either ≥4 tumors or those with even one tumor >3.0 cm in diameter. This is why the correlation was significant in univariate analysis, while it was not in multivariable model in which the effect of other risk factors, such as tumor number and tumor size were adjusted. The combination of TACE and RFA was not significantly related to either local tumor progression or distant recurrence. RFA was a safe procedure. Although many patients treated by RFA in this study were at high risk for surgical treatment because of advanced cirrhosis or other comorbidities, complications occurred in only 2.2% of the treatments. Other investigators have also reported low complication rates of 0–6.1% (11,13,14,15,16). For hepatic resection, morbidity rates of 38–47% have been reported even in recent studies (36,37,38).

surgical treatment because of advanced cirrhosis or other comorbidities, complications occurred in only 2.2% of the treatments. Other investigators have also reported low complication rates of 0–6.1% (11,13,14,15,16). For hepatic resection, morbidity rates of 38–47% have been reported even in recent studies (36,37,38). To date, percutaneous ethanol injection has been considered the standard in ablation (5). However, randomized controlled trials have demonstrated the superiority of RFA (6,7,8,9), with RFA now largely replacing ethanol injection. We have also shifted from ethanol injection to RFA (10). At our department, RFA is currently the first option and ethanol injection is performed only on patients on whom RFA cannot be performed safely because of either enterobiliary reflux, adhesion between the tumor and the gastrointestinal tract, or other reasons. Surgical resection has been considered the treatment of choice for HCC. Our first option for resectable HCC was also surgery. However, most patients who came to our department visited us because they did not want surgical resection. Thus, many patients in this study underwent RFA not because of unresectable tumor but because of refusal of surgery. Those who preferred surgery would have directly gone to the surgical department that has extensive experience in hepatic resection (38).

epartment visited us because they did not want surgical resection. Thus, many patients in this study underwent RFA not because of unresectable tumor but because of refusal of surgery. Those who preferred surgery would have directly gone to the surgical department that has extensive experience in hepatic resection (38). It is not easy to compare outcomes between RFA and surgical resection; the indications are different between the two treatments. Furthermore, indications for each treatment are different from institution to institution. Thus, a case adjudged to be treatable by RFA or surgical resection at an institution may not be given the same treatment at another. The best known indication criteria for surgical resection may be those proposed in the Barcelona Clinic Liver Cancer (BCLC) protocol (5), which states that surgical resection should be restricted to patients with performance status 0, Child-Pugh class A, single HCC, normal portal pressure, and normal serum bilirubin level. In patients satisfying those criteria, the 5-year survival rate is expected to be >70% (30). In this study, 237 (20.3%) of 1,170 patients satisfied those criteria and were thus considered good candidates for surgical resection; their 5-year survival rate was 75.9%, which appears satisfactory when compared with outcomes following surgical resection. Furthermore, in all 1,170 primary HCC patients treated by RFA, 5- and 10-year survival rates were 60.2% and 27.3%, respectively. In patients treated by surgical resection, 5- and 10-year survival rates were 34.4–70.0% and 10.5–52.0%, respectively (32,39,40,41,42,43,44,45). Although this is an observational study with no control, survivals following RFA appear comparable to those reported following surgical resection.

.2% and 27.3%, respectively. In patients treated by surgical resection, 5- and 10-year survival rates were 34.4–70.0% and 10.5–52.0%, respectively (32,39,40,41,42,43,44,45). Although this is an observational study with no control, survivals following RFA appear comparable to those reported following surgical resection. Two recent randomized controlled trials showed no significant difference in survival between RFA and surgical resection (46,47). Several nonrandomized controlled trials reported that RFA had similar overall survival rates to resection (48,49,50), while others found resection to be associated with higher survival rates (51,52,53). Further studies are necessary to resolve comparison of RFA with resection. We have made strenuous efforts to standardize the RFA procedure. Although many physicians have participated in RFA at our institution, the procedure was invariably performed according to the institutional protocol and in the presence of experienced physicians. Video recording was also used to monitor the procedure. Additionally, preoperative planning and postoperative evaluation of technique effectiveness were also carried out by at least three physicians. We also believe that not only proficient practice of RFA but also detailed preoperative planning, cautious postoperative evaluation of therapeutic effect, and careful follow-up are vital to achieve satisfactory outcomes.

operative evaluation of technique effectiveness were also carried out by at least three physicians. We also believe that not only proficient practice of RFA but also detailed preoperative planning, cautious postoperative evaluation of therapeutic effect, and careful follow-up are vital to achieve satisfactory outcomes. Source population in this study may represent selection bias, as we performed RFA on most patients who were hospitalized at our department; however, many patients with unfavorable tumor conditions for RFA might not have been referred to us. Therefore, caution is required when extrapolating our findings to the general population of HCC patients. A second limitation is that study population cannot be clearly defined. This study was based on daily clinical practice over a 10-year period. Indication criteria of RFA have changed over time, mainly because another percutaneous ablation, that is, ethanol injection has also been performed. Furthermore, various treatments besides percutaneous ablation were available for HCC, such as surgical resection and transarterial chemoembolization, with frequently overlapping indications.

of RFA have changed over time, mainly because another percutaneous ablation, that is, ethanol injection has also been performed. Furthermore, various treatments besides percutaneous ablation were available for HCC, such as surgical resection and transarterial chemoembolization, with frequently overlapping indications. One further limitation is the fact that this was a single-center study; these results might not be reproducible consistently in other settings. To extrapolate the findings in this study to patients at other institutions, careful consideration should be given to differences in the indications, methods, expertise, performance of available ultrasound and CT equipment, and others. Treatment outcome may be influenced by the physicians' expertise and the institution's volume of care. We started ethanol injection in 1985 and microwave ablation in 1995, that is, before the introduction of RFA. Recently, we have performed over 900 RFA treatments per year, which may represent a far greater number of treatments than those in most other institutions. We would not recommend any change in daily clinical practice solely on the strength of our study findings. In conclusion, our 10-year clinical experience shows that RFA could be locally curative, resulting in survival for as long as 10 years, and was a safe procedure. RFA might be a first-line treatment for selected patients with early-stage HCC. STUDY HIGHLIGHTS SUPPLEMENTARY MATERIAL is linked to the online version of the paper at http://www.nature.com/ajg Guarantor of the article: Shuichiro Shiina, MD, PhD.

In conclusion, our 10-year clinical experience shows that RFA could be locally curative, resulting in survival for as long as 10 years, and was a safe procedure. RFA might be a first-line treatment for selected patients with early-stage HCC. STUDY HIGHLIGHTS SUPPLEMENTARY MATERIAL is linked to the online version of the paper at http://www.nature.com/ajg Guarantor of the article: Shuichiro Shiina, MD, PhD. Specific author contributions: Study concept and design, analysis and interpretation of data, and drafting of the manuscript: Shuichiro Shiina; analysis and interpretation of data and statistical analysis: Ryosuke Tateishi; study execution and data acquisition: Toru Arano, Koji Uchino, Kenichiro Enooku, Hayato Nakagawa, Yoshinari Asaoka, Takahisa Sato, Ryota Masuzaki, Yuji Kondo, and Tadashi Goto; revised the article critically for important intellectual content: Haruhiko Yoshida; Masao Omata, and Kazuhiko Koike. All authors have read and approved the submitted manuscript. Financial support: This study was partly supported by Health Sciences Research Grants of The Ministry of Health, Labor and Welfare of Japan (Research on Hepatitis). Potential competing interests: None. Supplementary Material Supplementary Information Click here for additional data file. Figure 1 Flow of patients in this study. HCC, hepatocellular carcinoma. Figure 2 Overall survival in 1,170 primary hepatocellular carcinoma patients who underwent radiofrequency ablation. Figure 3 Local tumor progression and distant recurrence in patients who underwent radiofrequency ablation.

(7)* 5.3% (14) 0.8% (3)* 4.7% (9) 1.1% (10)*** 5.1% (23) Absolute risk reduction (95% CI) 4.1 (1.2, 7.0%) 3.9% (0.8, 7.1%) 4.0% (1.9, 6.1%) CI, confidence interval; GI, gastrointestinal; REDUCE, Registration Endoscopic Studies to Determine Ulcer Formation of HZT-501 Compared with Ibuprofen: Efficacy and Safety Studies. *P<0.05; **P<0.001; ***P<0.0001.

Supplementary Material Supplementary Information Click here for additional data file. Figure 1 Flow of patients in this study. HCC, hepatocellular carcinoma. Figure 2 Overall survival in 1,170 primary hepatocellular carcinoma patients who underwent radiofrequency ablation. Figure 3 Local tumor progression and distant recurrence in patients who underwent radiofrequency ablation. Table 1 Baseline characteristics of the 1,170 patients undergoing radiofrequency ablation for primary hepatocellular carcinoma Variable Age (years) 68.3±8.6 Males, n (%) 751 (64.1) Viral infection HBs-Ag-positive, n (%) 127 (10.9) Anti-HCV-positive, n (%) 870 (74.4) Both positive, n (%) 13 (1.1) Both negative, n (%) 159 (13.6) Alcohol consumption >80 g/d 170 (14.5) Ascites, n (%) 117 (10.0) Encephalopathy, n (%) 24 (2.1) Albumin (g/dl) 3.65±0.47 Total bilirubin (mg/dl) 0.95±0.49 Prothrombin time (%) 79.6±14.1 Platelet count (× 104/mm3) 11.9±5.6 AST (IU/l) 61.5±35.9 ALT (IU/l) 57.3±40.8 Child-Pugh class, n (%) A 868 (74.2) B 291 (24.9) C 11 (0.9) Tumor size (cm) 2.54±1.04 Tumor number 1.8±1.2 Serum AFP (ng/dl), n (%) ≤100 928 (793) 101–400 146 (12.5) >400 96 (8.2) Serum DCP (mAU/ml), n (%)a ≤100 964 (83.1) 101–400 126 (10.9) >400 70 (6.0) Serum AFP-L3 (%), n (%) ≤15 1,015 (86.8) 15.1–40 74 (6.3) >40 81 (6.9) AFP, α-fetoprotein; AFP-L3, lectin-reactive α-fetoprotein; ALT, alanine aminotransferase; AST, asparatate aminotransferase; DCP, des-γ-carboxy-prothrombin; HCV, hepatitis C virus. Data are expressed as mean±s.d. a Serum DCP level could not be measured in 10 patients because they were being administered warfarin.

Table 1 Baseline characteristics of the 1,170 patients undergoing radiofrequency ablation for primary hepatocellular carcinoma Variable Age (years) 68.3±8.6 Males, n (%) 751 (64.1) Viral infection HBs-Ag-positive, n (%) 127 (10.9) Anti-HCV-positive, n (%) 870 (74.4) Both positive, n (%) 13 (1.1) Both negative, n (%) 159 (13.6) Alcohol consumption >80 g/d 170 (14.5) Ascites, n (%) 117 (10.0) Encephalopathy, n (%) 24 (2.1) Albumin (g/dl) 3.65±0.47 Total bilirubin (mg/dl) 0.95±0.49 Prothrombin time (%) 79.6±14.1 Platelet count (× 104/mm3) 11.9±5.6 AST (IU/l) 61.5±35.9 ALT (IU/l) 57.3±40.8 Child-Pugh class, n (%) A 868 (74.2) B 291 (24.9) C 11 (0.9) Tumor size (cm) 2.54±1.04 Tumor number 1.8±1.2 Serum AFP (ng/dl), n (%) ≤100 928 (793) 101–400 146 (12.5) >400 96 (8.2) Serum DCP (mAU/ml), n (%)a ≤100 964 (83.1) 101–400 126 (10.9) >400 70 (6.0) Serum AFP-L3 (%), n (%) ≤15 1,015 (86.8) 15.1–40 74 (6.3) >40 81 (6.9) AFP, α-fetoprotein; AFP-L3, lectin-reactive α-fetoprotein; ALT, alanine aminotransferase; AST, asparatate aminotransferase; DCP, des-γ-carboxy-prothrombin; HCV, hepatitis C virus. Data are expressed as mean±s.d. a Serum DCP level could not be measured in 10 patients because they were being administered warfarin. Table 2 Survival of patients undergoing radiofrequency ablation, based on tumor number, tumor size, and Child-Pugh class Grading n Survival (%) Median (years) P value 1-Year 3-Year 5-Year 7-Year 10-Year Overall survival 1,170 96.6 80.5 60.2 45.1 27.3 6.4 — Tumor number Solitary 685 97.2 82.6 64.6 50.5 32.0 7.0 0.0003 2–3 395 95.7 77.9 54.4 39.4 19.9 5.6 ≥4 90 96.5 76.4 53.6 30.1 17.6 5.3 Tumor size ≤3 cm 889 97.2 83.8 65.1 47.3 30.7 6.7 <0.0001 >3 cm 281 94.8 71.0 46.5 38.0 18.6 4.6 Child-Pugh class A 868 98.0 86.0 65.9 50.2 30.1 7.0 <0.0001 B 291 93.2 66.4 46.5 32.4 20.6 4.6 C 11 81.8 58.4 23.4 23.4 — 3.1 Combination of tumor number, tumor size, and Child-Pugh class Solitary, ≤3 cm 534 97.6 84.7 68.0 51.4 34.3 7.1 — Solitary, ≤3 cm, Child-Pugh A 401 98.7 90.1 74.0 57.4 41.3 8.2 — 1–3 Tumors, ≤3 cm 822 97.1 83.7 65.2 48.8 32.5 6.9 — Solitary, ≤5 cm, or 1–3 tumors, ≤3 cm 947 97.2 82.8 63.8 48.8 30.6 6.9 — Child-Pugh A/B Satisfied the indication criteria of surgical resection proposed in the BCLC protocola 237 98.6 90.5 75.9 61.1 38.1 8.7 — BCLC, Barcelona Clinic Liver Cancer; HCC, hepatocellular carcinoma.

, ≤3 cm 822 97.1 83.7 65.2 48.8 32.5 6.9 — Solitary, ≤5 cm, or 1–3 tumors, ≤3 cm 947 97.2 82.8 63.8 48.8 30.6 6.9 — Child-Pugh A/B Satisfied the indication criteria of surgical resection proposed in the BCLC protocola 237 98.6 90.5 75.9 61.1 38.1 8.7 — BCLC, Barcelona Clinic Liver Cancer; HCC, hepatocellular carcinoma. a Child-Pugh class A with a normal level of bilirubin, no significant portal hypertension, and a single HCC.

, ≤3 cm 822 97.1 83.7 65.2 48.8 32.5 6.9 — Solitary, ≤5 cm, or 1–3 tumors, ≤3 cm 947 97.2 82.8 63.8 48.8 30.6 6.9 — Child-Pugh A/B Satisfied the indication criteria of surgical resection proposed in the BCLC protocola 237 98.6 90.5 75.9 61.1 38.1 8.7 — BCLC, Barcelona Clinic Liver Cancer; HCC, hepatocellular carcinoma. a Child-Pugh class A with a normal level of bilirubin, no significant portal hypertension, and a single HCC. Table 3 Multivariate analysis of variables relevant to survival, local tumor progression, and distant recurrence Variable Multivariate analysis Hazard ratio (95% CI) P value Survival Age (per year) 1.03 (1.02–1.04) <0.0001 Anti-HCV-positive 1.34 (1.03–1.76) 0.03 Child-Pugh class A 1 B or C 2.08 (1.69–2.56) <0.0001 Tumor size (cm) ≤2.0 1 2.1–3.0 1.40 (1.10–1.80) 0.007 3.1–5.0 1.80 (1.37–2.38) <0.0001 >5.0 1.50 (0.90–2.49) 0.12 Tumor number Solitary 1 2–3 1.28 (1.04–1.59) 0.02 ≥4 1.58 (1.13–2.21) 0.008 Serum DCP (mAU/ml) ≤100 1 101–400 1.22 (0.88–1.69) 0.24 >400 1.66 (1.14–2.42) 0.008 Serum AFP-L3 (%) ≤15 1 >15 1.45 (1.11–1.91) 0.008 Local tumor progression Serum DCP (mAU/ml) ≤100 1 101–400 2.51 (1.02–6.20) 0.05 >400 6.52 (2.63–16.1) <0.0001 Distant recurrence Anti-HCV-positive 1.44 (1.19–1.75) 0.0002 Child-Pugh class A 1 B or C 1.23 (1.03–1.45) 0.02 Platelet count (/l) >1011 1 ≤1011 1.36 (1.12–1.64) 0.002 Tumor size (cm) ≤2.0 1 2.1–3.0 1.30 (1.10–1.55) 0.003 3.1–5.0 1.29 (1.05–1.60) 0.02 >5.0 1.25 (0.75–2.08) 0.4 Tumor number Solitary 1 2–3 1.36 (1.16–1.59) 0.0002 ≥4 2.02 (1.53–2.66) <0.0001 Serum AFP (ng/dl) ≤100 1 101–400 1.15 (0.92–1.44) 0.22 >400 1.36 (1.03–1.81) 0.03 Serum DCP (mAU/ml) ≤100 1 101–400 1.19 (0.92–1.54) 0.19 >400 1.72 (1.22–2.42) 0.002 AFP, α-fetoprotein; CI, confidence interval; DCP, des-γ-carboxy-prothrombin; HCV, hepatitis C virus.

3 1.36 (1.16–1.59) 0.0002 ≥4 2.02 (1.53–2.66) <0.0001 Serum AFP (ng/dl) ≤100 1 101–400 1.15 (0.92–1.44) 0.22 >400 1.36 (1.03–1.81) 0.03 Serum DCP (mAU/ml) ≤100 1 101–400 1.19 (0.92–1.54) 0.19 >400 1.72 (1.22–2.42) 0.002 AFP, α-fetoprotein; CI, confidence interval; DCP, des-γ-carboxy-prothrombin; HCV, hepatitis C virus. Table 4 Complications in 2,982 treatments of radiofrequency ablation for hepatocellular carcinoma Complication No. of complications Neoplastic seeding 24 Liver abscess 6 Hemoperitoneum 12 Hemothorax 5 Symptomatic pleural effusion 1 Massive hepatic infarction 6 Gastrointestinal perforation or penetration 5 Hemobilia 2 Skin burn 1 Pneumothorax 3 Gallbladder injury 1 Cerebral infarction 1

INTRODUCTION Inflammatory bowel disease (IBD) is a chronic, inflammatory disorder of the gastrointestinal tract. As such, standard therapies for IBD have focused on nonspecific inhibition of inflammation with sulfasalazine, mesalazine, steroids, the thiopurines azathioprine (AZA) and 6-mercaptopurine (6-MP), and methotrexate (MTX) (1). While these agents can be moderately effective in maintaining corticosteroid-induced remission, relapse can be common (2,3). In addition, thiopurine therapy of IBD carries an increased risk of lymphoproliferative disorders (4). On the basis of the unmet need for patients intolerant of or unresponsive to standard therapy, several antagonists of the proinflammatory cytokine tumor necrosis factor (TNF) have been developed, including adalimumab, etanercept, infliximab, and certolizumab pegol. Although TNF antagonist therapy is generally well tolerated by patients with IBD, a unique spectrum of safety issues related to blocking TNF, including life-threatening and opportunistic infection, malignancy, and mortality, must be considered. While researchers have attempted to obtain consensus on the relationship between TNF antagonist therapy and these relatively rare events, analyses have generally been limited by comparatively small study populations and short periods of patient follow-up.

nistic infection, malignancy, and mortality, must be considered. While researchers have attempted to obtain consensus on the relationship between TNF antagonist therapy and these relatively rare events, analyses have generally been limited by comparatively small study populations and short periods of patient follow-up. The TNF antagonist infliximab has been used to treat patients with moderately-to-severely active IBD for more than a decade. As infliximab was approved for the treatment of moderate-to-severe Crohn's disease (CD), the sponsor has conducted several large, randomized, controlled phase 3 clinical trials of infliximab maintenance therapy in IBD (hereafter referred to as “pivotal phase 3 trials”), including the ACCENT I (5,6), ACCENT II (7), and SONIC (8) trials in CD and the ACT 1 and ACT 2 trials (9) that formed the basis for approval of infliximab in ulcerative colitis (UC). Although safety findings related to each of these studies have been reported in separate publications, a pooled analysis of key safety outcomes has been recently conducted for all sponsor-initiated infliximab trials in IBD, with emphasis on the five pivotal phase 3 trials that contribute nearly 90% of the data for patients in the pooled IBD safety analyses. Given the continued need for safety data related to TNF antagonism, findings of these analyses are detailed herein.

cently conducted for all sponsor-initiated infliximab trials in IBD, with emphasis on the five pivotal phase 3 trials that contribute nearly 90% of the data for patients in the pooled IBD safety analyses. Given the continued need for safety data related to TNF antagonism, findings of these analyses are detailed herein. METHODS The 10 (7 CD, 3 UC) clinical trials included in these pooled analyses represent the totality of the sponsor's clinical safety database for infliximab in the treatment of adult patients with IBD. The pooled studies comprise five smaller trials (4 CD, 1 UC) evaluating a total of 244 patients (10,11,12,13,14,15), 22 of whom contributed data to more than one treatment group per study, and five pivotal phase 3 trials evaluating infliximab maintenance therapy in 1,644 patients with CD (5,6,7,8) and 741 patients with UC (9), 184 of whom contributed data to more than one treatment group per study (Tables 1, 2, 3, 4, 5).

2,13,14,15), 22 of whom contributed data to more than one treatment group per study, and five pivotal phase 3 trials evaluating infliximab maintenance therapy in 1,644 patients with CD (5,6,7,8) and 741 patients with UC (9), 184 of whom contributed data to more than one treatment group per study (Tables 1, 2, 3, 4, 5). The five pivotal phase 3 trials, i.e., ACCENT I, ACCENT II, SONIC, ACT 1, and ACT 2, contributed approximately 89% (2,119/2,385) of all patients with data in the overall pooled analyses and were generally consistent in terms of study designs, in that they were all randomized, multicenter, double-blind trials that included a control group (Table 1). Note that for four of the five pivotal phase 3 trials (ACCENT I, ACCENT II, ACT 1, and ACT 2), treatment with the immunomodulators AZA, 6-MP, and MTX or with corticosteroids was allowed during study participation, but such use was not a randomized study treatment. Each of these study protocols stipulated that patients would continue a stable regimen of baseline immunomodulator therapy throughout study participation. Conversely, the SONIC trial enrolled exclusively immunomodulator-naïve patients, and these patients were randomized to receive AZA, 5 mg/kg of infliximab, or AZA plus 5 mg/kg of infliximab (8).

s stipulated that patients would continue a stable regimen of baseline immunomodulator therapy throughout study participation. Conversely, the SONIC trial enrolled exclusively immunomodulator-naïve patients, and these patients were randomized to receive AZA, 5 mg/kg of infliximab, or AZA plus 5 mg/kg of infliximab (8). All study protocols were approved by the institutional review board at each participating site, and all patients provided written informed consent before beginning study participation. Janssen Biotech, Inc. (Horsham, PA) provided infliximab, active comparator (AZA in the SONIC trial), and placebo (as appropriate) for intravenous infusion. To evaluate the occurrence of uncommon events, safety data from the seven studies in CD (with the majority of the data coming from the pivotal ACCENT I, ACCENT II, and SONIC studies) were pooled and are reported as ‘CD studies'. When pooled with data from the three ‘UC studies' (with the majority of the data deriving from the pivotal ACT 1 and ACT 2 studies), they are reported as ‘IBD studies'. Safety data from the pivotal phase 3 trials in IBD, i.e., ACCENT I, ACCENT II, SONIC, ACT 1, and ACT 2, were also separately pooled across the three CD, two UC, and all five pivotal phase 3 IBD studies. Summaries of key design features of these studies are provided in Table 1.

dies), they are reported as ‘IBD studies'. Safety data from the pivotal phase 3 trials in IBD, i.e., ACCENT I, ACCENT II, SONIC, ACT 1, and ACT 2, were also separately pooled across the three CD, two UC, and all five pivotal phase 3 IBD studies. Summaries of key design features of these studies are provided in Table 1. The incidences of adverse events per 100 patient-years (pt-yrs) of follow-up were calculated for infections, malignancies (including both solid tumors and hematological malignancies and excluding nonmelanoma skin cancers) and deaths by treatment group (infliximab vs. placebo) and also by immunomodulator use (treatment vs. no treatment) for infections and malignancies as the quotient of the total number of events and pt-yrs of follow-up multiplied by 100; exact 95% confidence intervals (CIs) were also calculated. For malignancies (excluding nonmelanoma skin cancer), standardized incidence ratios (SIRs) were also calculated as the quotient of the observed and expected numbers of patients with malignancy; 95% CIs were determined using exact methodology. The expected numbers of malignancy were derived using data adjusted for age, sex, and race from the general US population in the Surveillance, Epidemiology, and End Results (SEER) database (16).

e quotient of the observed and expected numbers of patients with malignancy; 95% CIs were determined using exact methodology. The expected numbers of malignancy were derived using data adjusted for age, sex, and race from the general US population in the Surveillance, Epidemiology, and End Results (SEER) database (16). Fisher's exact test was used to compare the proportions of patients who experienced an adverse event of interest (e.g., infection, malignancy, or death) between treatment groups. Because a large number of safety parameters were evaluated, the Fisher's exact test is employed not for hypothesis testing but rather as an aid in signal detection to highlight differences requiring closer examination. All patients in ACCENT I and ACCENT II received infliximab 5 mg/kg at week 0 and were therefore counted in the infliximab column in the calculation of pt-yrs of follow-up. Pt-yrs of follow-up for placebo were determined for the 161 placebo plus AZA-treated CD patients in the SONIC study plus additional placebo-treated patients from other CD studies (T08, T11, T16, and T20; see Table 1) as applicable to the subpopulation being assessed. Note that infliximab use, both in combination with immunomodulators and alone, and immunomodulator use, both in combination with infliximab and alone, are pooled in these analyses such that infliximab use refers to any use of infliximab and immunomodulator use refers to any use of immunomodulators. Also note that the placebo group includes only patients who never received infliximab.

s and alone, and immunomodulator use, both in combination with infliximab and alone, are pooled in these analyses such that infliximab use refers to any use of infliximab and immunomodulator use refers to any use of immunomodulators. Also note that the placebo group includes only patients who never received infliximab. As noted above, infliximab treatment was generally randomized and blinded, while immunomodulator treatment, with the exception of the SONIC trial, reflects immunomodulator use at baseline, i.e., immunomodulator use was not randomized or blinded and assumes that such use continued during the study. Protocols for these four of five pivotal phase 3 trials mandated that patients receiving a stable immunomodulator regimen at baseline would continue such use throughout study participation.

tor use at baseline, i.e., immunomodulator use was not randomized or blinded and assumes that such use continued during the study. Protocols for these four of five pivotal phase 3 trials mandated that patients receiving a stable immunomodulator regimen at baseline would continue such use throughout study participation. RESULTS Analysis groups and extent of exposure Across the 10 sponsor-initiated infliximab trials in IBD (five smaller studies and five pivotal phase 3 studies conducted following infliximab's initial approval), safety data for 2,385 patients were available for pooled safety analyses (see Tables 3, 4, 5). Note that some patients presented in Tables 1 and 2 contributed data to more than one treatment group in the analyses presented in Tables 3, 4, 5, e.g., 22 of the 244 patients in the five smaller studies and 33 of the 2,086 patients in the five pivotal trials. Data from the five pivotal trials were pooled for additional analyses, both across all five studies (n=2,119) and across the three CD studies (n=1,389) or 2 UC (n=730) studies due to similarities in study design and homogeneity of the patient populations. In the 5 pivotal IBD trials, 406 and 1,713 patients were treated with placebo and infliximab, respectively (Table 3). Among the 2,117 patients with documentation of immunomodulator use at baseline (yes/no), 947 patients did and 1,170 patients did not receive the immunomodulators AZA, 6-MP, or MTX (Tables 3 and 5).

ons. In the 5 pivotal IBD trials, 406 and 1,713 patients were treated with placebo and infliximab, respectively (Table 3). Among the 2,117 patients with documentation of immunomodulator use at baseline (yes/no), 947 patients did and 1,170 patients did not receive the immunomodulators AZA, 6-MP, or MTX (Tables 3 and 5). The extent of exposure to individual study agents for each of the pivotal phase 3 trials is summarized in Table 2, which includes infliximab infusions received during the main studies, as well as the blinded study extensions of SONIC and ACT 2. Excluding patients who initially received infliximab but who were later randomized to placebo maintenance treatment, CD patients received an average of 6.1–7.5 infliximab infusions and UC patients received an average of 6.3–6.5 infliximab infusions. Note that study agent administration in the SONIC and ACT 2 main studies ended with the week-22 infusion and that safety data from week 30 through week 54 were collected as part of a blinded study extension, during which patients continued to receive blinded study agent.

ived an average of 6.3–6.5 infliximab infusions. Note that study agent administration in the SONIC and ACT 2 main studies ended with the week-22 infusion and that safety data from week 30 through week 54 were collected as part of a blinded study extension, during which patients continued to receive blinded study agent. Infections The incidences of infections and serious infections were determined across the five pivotal phase 3 IBD trials. A larger proportion of infliximab- than placebo-treated UC patients (50.1% vs. 36.3% P<0.001) had at least one infection. Among CD patients, however, the proportions of patients who experienced at least one infection were similar between the infliximab- and placebo-treated patients (49.1% vs. 45.3% P=0.402). The proportions of patients who experienced at least one serious infection were also similar between placebo- and infliximab-treated patients (Table 3). When expressed on the basis of length of patient follow-up, the incidences (95% CIs) per 100 pt-yrs of infections were 132.81 (112.00, 156.36) in placebo- vs. 119.98 (113.67, 126.56) in infliximab-treated CD patients; 106.98 (93.43, 121.94) in placebo- vs. 105.41 (98.54, 112.63) in infliximab-treated UC patients; and 115.79 (104.26, 128.25) in placebo- vs. 113.80 (109.12, 118.62) in infliximab-treated IBD patients. Similar patterns of overlapping 95% CIs between the placebo and infliximab groups were also observed for serious infections (Table 3).

121.94) in placebo- vs. 105.41 (98.54, 112.63) in infliximab-treated UC patients; and 115.79 (104.26, 128.25) in placebo- vs. 113.80 (109.12, 118.62) in infliximab-treated IBD patients. Similar patterns of overlapping 95% CIs between the placebo and infliximab groups were also observed for serious infections (Table 3). When assessed by the patient's baseline immunomodulator treatment (yes/no), the 95% CIs surrounding the incidences of infections and serious infections overlapped between patients treated with immunomodulators and those not treated with immunomodulators in all patient populations (CD, UC, all IBD), with one exception. In patients with UC, but not CD, immunomodulator treatment (vs. no treatment) yielded a higher incidence (95% CI) of infections (120.07 (110.66, 130.08)/100 pt-yrs vs. 92.47 (84.54, 100.94)/100 pt-yrs) (Table 3). A summary of the incidences of serious infections (per 100 pt-yrs of follow-up) by system–organ class and preferred term is also provided in Table 3. The most common serious infections were those considered resistance mechanism disorders (3.15 and 3.32/100 pt-yrs in placebo- and infliximab-treated IBD patients, respectively), gastrointestinal disorders (0.63 and 1.07/100 pt-yrs), and respiratory system disorders (0.94 and 0.97/100 pt-yrs). Within these three system–organ classes, the most common serious infections were abscess (1.57 and 1.89/100 pt-yrs in placebo- and infliximab-treated IBD patients, respectively), gastroenteritis (0.63 and 0.26/100 pt-yrs), and pneumonia (0.31 and 0.66/100 pt-yrs) (Table 3).

system disorders (0.94 and 0.97/100 pt-yrs). Within these three system–organ classes, the most common serious infections were abscess (1.57 and 1.89/100 pt-yrs in placebo- and infliximab-treated IBD patients, respectively), gastroenteritis (0.63 and 0.26/100 pt-yrs), and pneumonia (0.31 and 0.66/100 pt-yrs) (Table 3). Malignancy Across the 10 IBD trials and excluding nonmelanoma skin cancers, 13 patients (two placebo-treated, 11 infliximab-treated) had a malignancy during study participation. Of the malignancies, 11 were non-lymphoma, while two patients had lymphoma. Both patients with lymphoma were infliximab-treated CD patients. One patient who received infliximab 5 mg/kg at week 0 followed by placebo maintenance and AZA had natural killer cell lymphoma diagnosed after study participation ended. A second patient with a history of AZA use received a single infusion of infliximab 10 mg/kg and was diagnosed with intravascular B-cell lymphoma 9.5 months after the single infliximab infusion. The non-lymphoma malignancies included breast (n=2), colon (n=2), prostate (n=2), bladder (n=1), lung (n=1), renal (n=1), skin (n=1), and rectal (n=1) cancers (9 infliximab-treated, 2 placebo).

n of infliximab 10 mg/kg and was diagnosed with intravascular B-cell lymphoma 9.5 months after the single infliximab infusion. The non-lymphoma malignancies included breast (n=2), colon (n=2), prostate (n=2), bladder (n=1), lung (n=1), renal (n=1), skin (n=1), and rectal (n=1) cancers (9 infliximab-treated, 2 placebo). When expressed on the basis of incidence (95% CI) per 100 pt-yrs of follow-up, overlapping 95% CIs indicated that the incidences of malignancies were similar in the placebo- and infliximab-treated patients with CD (1.61 (0.19, 5.82) vs. 0.49 (0.18, 1.06), respectively) and with UC (0.00 (0.00, 1.43) vs. 0.60 (0.20, 1.40), respectively). Findings observed within the lymphoma and non-lymphoma malignancy subcategories were similar (Table 4). The incidences of malignancies during only the controlled portions of the 10 IBD trials were also determined. Four patients, two placebo-treated and two infliximab-treated, had a malignancy during the controlled study phases. No cases of lymphoma were documented during the controlled portions of the 10 IBD studies. When comparing the proportions of patients who had malignancy diagnosed during the studies (both overall and during the controlled portions), results of Fisher's exact testing indicated no significant difference between infliximab- and placebo-treated patients (Table 4).

g the controlled portions of the 10 IBD studies. When comparing the proportions of patients who had malignancy diagnosed during the studies (both overall and during the controlled portions), results of Fisher's exact testing indicated no significant difference between infliximab- and placebo-treated patients (Table 4). Further, when expressed as incidence (95% CI) per 100 pt-yrs of follow-up, the incidences of malignancy appeared similar in placebo- vs. infliximab-treated patients with CD (1.65 (0.20, 5.97) vs. 0.00 (0.00, 1.00), respectively) and UC (0.00 (0.00, 2.18) vs. 0.60 (0.07, 2.17), respectively). Consistent findings were observed within the lymphoma and non-lymphoma malignancy subcategories (Table 4). The incidence of malignancy was also assessed by immunomodulator use in the controlled portions of the five pivotal phase 3 IBD trials. When comparing the proportions of patients who had malignancy diagnosed during the studies (both overall and during the controlled portions), results of Fisher's exact testing indicated no significant difference between infliximab- and placebo-treated patients or between patients with and without immunomodulator use (Table 4).

paring the proportions of patients who had malignancy diagnosed during the studies (both overall and during the controlled portions), results of Fisher's exact testing indicated no significant difference between infliximab- and placebo-treated patients or between patients with and without immunomodulator use (Table 4). Further, in all three of the patient populations (CD, UC, and all IBD), the 95% CIs surrounding the incidences of malignancy overlapped when compared between patients who were treated vs. those who were not treated with immunomodulators, with one exception. Among placebo-treated patients with CD, but not UC, those with immunomodulator use demonstrated a higher incidence (95% CI) of malignancy vs. no immunomodulator treatment (1.84 (0.22, 6.66)/100 pt-yrs vs. 0.00 (0.00, 0.00)/100 pt-yrs). In a separate analysis comparing the observed incidences of malignancy with rates expected in the general US population based on the SEER database, all 95% CIs surrounding the malignancy SIRs contained 1 in all analysis groups (i.e., placebo and infliximab, with and without baseline immunomodulator use). This indicates that the incidences of malignancy observed in the controlled portions of the pivotal phase 3 trials are not significantly different from the expected rates in the general US population. Similar findings were observed within the lymphoma and non-lymphoma malignancy subcategories (Table 4).

modulator use). This indicates that the incidences of malignancy observed in the controlled portions of the pivotal phase 3 trials are not significantly different from the expected rates in the general US population. Similar findings were observed within the lymphoma and non-lymphoma malignancy subcategories (Table 4). In addition to the malignancies discussed above, among the 2,385 patients with IBD included in these analyses (465 placebo, 1,920 infliximab), five patients (3 infliximab-treated, 2 placebo-treated) had basal cell carcinoma and two patients (both infliximab-treated) had malignant skin neoplasm. Mortality Five patients died during the 10 IBD trials. The deaths included a patient (63-year-old female with baseline immunomodulator use) with CD in SONIC who received AZA monotherapy and died of sepsis following a colectomy; three infliximab-treated patients with CD in ACCENT I who died of septic shock (35-year-old female with no baseline immunomodulator use), sepsis (57-year-old female with no baseline immunomodulator use), and myocardial infarction (37-year-old male with baseline immunomodulator use); and one infliximab-treated patient with UC in ACT 2 (56-year-old male with baseline immunomodulator use) who died following diagnosis of pulmonary histoplasmosis during the open-label, long-term, follow-up period.

line immunomodulator use), and myocardial infarction (37-year-old male with baseline immunomodulator use); and one infliximab-treated patient with UC in ACT 2 (56-year-old male with baseline immunomodulator use) who died following diagnosis of pulmonary histoplasmosis during the open-label, long-term, follow-up period. Given the longer follow-up periods for infliximab-treated patients, when summarized as incidence (95% CI) per 100 pt-yrs of follow-up, overlapping CIs indicated no increase in mortality with infliximab vs. placebo treatment among patients with CD (0.24 (0.05, 0.71) vs. 0.80 (0.02, 4.48), respectively), UC (0.12 (0.00, 0.67) vs. 0.00 (0.00, 1.43), respectively), or IBD (0.19 (0.05, 0.50) vs. 0.30 (0.01, 1.67), respectively). The same was true for infection-related deaths, which accounted for four of the five deaths. Overlapping 95% CIs were also observed in a separate analysis of mortality by immunomodulator use in the five pivotal phase 3 IBD trials (Table 5).

respectively), or IBD (0.19 (0.05, 0.50) vs. 0.30 (0.01, 1.67), respectively). The same was true for infection-related deaths, which accounted for four of the five deaths. Overlapping 95% CIs were also observed in a separate analysis of mortality by immunomodulator use in the five pivotal phase 3 IBD trials (Table 5). DISCUSSION Since receiving marketing authorization for the treatment of CD more than a decade ago, infliximab has gained wide acceptance as a highly effective treatment option for IBD. As a result, longer-term safety data are becoming available. No difference in the incidence of neoplasia between adult patients with CD who were (n=404) and were not (n=404) treated with infliximab was reported by Biancone et al (17). An additional 4 years of follow-up of 591 of the patients in this matched-pair study yielded consistent findings, with 3.9% (12/304) of infliximab-treated patients and 4.2% (12/287) of patients not treated with infliximab diagnosed with neoplasia (P=0.95) (18). Fidder et al. (19), who retrospectively examined medical records of 734 infliximab-treated IBD patients and 666 controls, also observed no difference between the two groups in mortality, malignancy including lymphoma, or infection rate. In a retrospective safety analysis of 799 German IBD patients treated with thiopurines and/or TNF antagonists between 2002 and 2010, an elevated risk of malignancy (4.2% vs. 1.5%, P=0.024, odds ratio=2.86), but not infection (14.4% vs. 15.5%, P=0.69), was observed in patients treated with only thiopurines relative to patients treated with TNF antagonists with or without thiopurines (20). Results of several large randomized phase 3 clinical trials of infliximab maintenance therapy in IBD, including the ACCENT I (5,6), ACCENT II 7 and SONIC (8) trials in CD and the ACT 1 and ACT 2 trials (9) in UC also contribute to the growing body of infliximab safety data.

antagonists with or without thiopurines (20). Results of several large randomized phase 3 clinical trials of infliximab maintenance therapy in IBD, including the ACCENT I (5,6), ACCENT II 7 and SONIC (8) trials in CD and the ACT 1 and ACT 2 trials (9) in UC also contribute to the growing body of infliximab safety data. Findings presented herein from a pooled analysis of key safety outcomes, derived from 10 sponsor-conducted IBD studies with large cohorts of IBD patients treated by referral centers in daily practice, are generally consistent with conclusions drawn by Biancone et al. (17,18) and Fidder et al. (19). Specifically, results of our pooled analyses of infliximab safety data in the treatment of IBD indicate no increase in infections or serious infections with infliximab vs. placebo treatment in patients with CD or UC. Independent of infliximab use, immunomodulator treatment did not appear to increase the incidence of infections or serious infections in patients with CD but did yield a higher incidence of infection vs. no immunomodulator treatment in patients with UC. One explanation for the lack of an increase in the incidence of serious infections with either infliximab or immunomodulators in CD is that many of the complications of CD are inherently infectious in nature and may be decreased by effective CD therapy. Results derived from the TREAT Registry of CD patients, which included assessment of the role of corticosteroids in infectious complications, have shown that infliximab allows for steroid tapering and discontinuation in CD (21). Several of the protocols for the pivotal phase 3 trials included in this report mandated steroid tapering during the early phase of the trial. The lower steroid consumption in the reported CD trials may also contribute to the relatively low incidence of serious infections, regardless of immunomodulator use, as such an effect could offset any small increase in serious infections that might be associated with the immunosuppressive nature of these drugs. These findings are also consistent with additional data from the TREAT Registry that indicated that the risk for serious infection in infliximab-treated CD patients is similar to that for patients receiving conventional immunomodulators (21). Despite our observation of no increase in serious infections, all patients should be screened for pre-existing infections before the start of any immunosuppressive therapy.

that the risk for serious infection in infliximab-treated CD patients is similar to that for patients receiving conventional immunomodulators (21). Despite our observation of no increase in serious infections, all patients should be screened for pre-existing infections before the start of any immunosuppressive therapy. Overall, 13 patients (2 placebo-treated, 11 infliximab-treated) had a malignancy (excluding nonmelanoma skin cancer) during the 10 IBD trials, equating to incidences of 0.60 and 0.53/100 pt-yrs, among placebo- and infliximab-treated IBD patients, respectively. Two of the malignancies in infliximab-treated CD patients were lymphomas; the 11 other malignancies across both cohorts were non-lymphoma. During the controlled portions of the 10 IBD trials, four patients (2 placebo, 2 infliximab) had a malignancy, all non-lymphoma, equating to incidences of 0.77 and 0.32/100 pt-yrs, respectively, of placebo- and infliximab-treated IBD patients. The incidence of malignancy was not higher with infliximab treatment, nor was it significantly impacted by immunomodulator use. In a separate analysis comparing the observed incidences of malignancy with rates expected in the general US population based on the SEER database, the incidence of malignancy was not significantly different.

malignancy was not higher with infliximab treatment, nor was it significantly impacted by immunomodulator use. In a separate analysis comparing the observed incidences of malignancy with rates expected in the general US population based on the SEER database, the incidence of malignancy was not significantly different. A higher incidence of malignancy was observed in placebo-treated patients receiving vs. those not receiving immunomodulators. These findings may support others suggesting that the thiopurines AZA and 6-MP are associated with a moderately increased risk of malignancy, particularly lymphoproliferative disease. Specifically, in a French cohort of nearly 20,000 patients with IBD (60% with CD, 40% with UC or unclassified IBD) followed for an average of 35 months, the risk of lymphoproliferative disorder was 5 times higher in patients exposed to thiopurines than in those never exposed to these drugs. Older age, male sex, and longer duration of IBD were also associated with increased risk of lymphoproliferative disorder in the French cohort (22). Also, as noted above, results of a retrospective safety analysis of 799 German IBD patients treated with thiopurines and/or TNF antagonists between 2002 and 2010 indicated an elevated risk of malignancy (4.2% vs. 1.5%, P=0.024, odds ratio=2.86) in patients treated with only thiopurines relative to patients treated with TNF antagonists with or without thiopurines (20).

afety analysis of 799 German IBD patients treated with thiopurines and/or TNF antagonists between 2002 and 2010 indicated an elevated risk of malignancy (4.2% vs. 1.5%, P=0.024, odds ratio=2.86) in patients treated with only thiopurines relative to patients treated with TNF antagonists with or without thiopurines (20). We also determined malignancy SIRs within the groups of patients with and without baseline immunomodulator use. Results of these pooled analyses indicated no significant difference from the expected rates in the general US population. Conversely, recently reported results of a meta-analysis of 26 studies of infliximab, adalimumab, and certolizumab, including almost 9,000 CD patients, indicated that use of anti-TNF agents with immunomodulators is associated with an increased, albeit small (6.1/10,000 pt-yrs), risk of non-Hodgkin's lymphoma relative to the expected rate of non-Hodgkin's lymphoma derived from the SEER database, i.e., 1.9/10,000 pt-yrs (SIR (95% CI)=3.23 (1.5, 6.9)), and also relative to the expected rate for CD patients treated with only immunomodulators, i.e., 4.0/10,000 pt-yrs (SIR (95% CI)=1.70 (0.5, 7.1)) (23). As noted above, no cases of lymphoma were reported during the controlled portions of the trials comprising our pooled analyses. In addition to the malignancies discussed above, among the 2,385 patients with IBD included in these analyses (465 placebo, 1,920 infliximab), five patients (3 infliximab-treated, 2 placebo-treated) had basal cell carcinoma and two patients (both infliximab-treated) had nonmelanoma malignant skin neoplasm. It is still not certain whether or not infliximab use increases the risk of malignancy, but it is possible that the impact of infliximab is no worse than that of conventional immunomodulators and that, by effectively controlling inflammation, infliximab may contribute to a lower malignancy risk in IBD. The latter possibility requires further confirmation with longer-term data.

ases the risk of malignancy, but it is possible that the impact of infliximab is no worse than that of conventional immunomodulators and that, by effectively controlling inflammation, infliximab may contribute to a lower malignancy risk in IBD. The latter possibility requires further confirmation with longer-term data. Consistent with published reports citing no evidence of increased mortality in CD patients treated with anti-TNF agents (19,20,24), we observed no difference in mortality between placebo- and infliximab-treated patients with either CD or UC. The same was true for infection-related deaths, which accounted for four of the five deaths. Immunomodulator treatment was also unassociated with increases in mortality in these IBD patients.

agents (19,20,24), we observed no difference in mortality between placebo- and infliximab-treated patients with either CD or UC. The same was true for infection-related deaths, which accounted for four of the five deaths. Immunomodulator treatment was also unassociated with increases in mortality in these IBD patients. One point to note in the interpretation of these data is that, with the exception of the SONIC trial (8), immunomodulator treatment was not randomized, blinded, or controlled and reflects patient treatment at the time of study entry (as study protocols stipulated that any baseline immunomodulator treatment regimen would remain stable throughout study participation). It therefore remains possible that any higher event incidence in the immunomodulator-treated patients reflects their having more severe IBD rather than immunomodulator use itself. For this reason, any comparison between event rates between infliximab vs. immunomodulators must be made with caution. Still, the data of Fidder et al. (19) showed no difference in the rates of infection, malignancy including lymphoma, and mortality between IBD patients treated with infliximab and IBD patients treated with conventional therapies. It should also be noted, however, that the overall pooled results do not differ from those of the SONIC trial, in which both treatment with infliximab and treatment with immunomodulators were randomized in a controlled trial. It is also important to note that the relatively short period of follow-up, along with the relative lack of power inherent in these clinical trial data for determining treatment group differences in rare safety events, limit our ability to draw definitive conclusions from these analyses. The powering, however, is fairly good for detecting a doubling or tripling of the malignancy rate, both of which are clinically important to exclude. The 2,061 pt-yrs of follow-up from the infliximab-treated IBD cohort in this paper would yield 71% or 99% power to detect a doubling or tripling in malignancy incidence, respectively. Representing data derived from rigorous clinical trials, therefore, the current data are somewhat reassuring and have merit when assessing the overall safety of anti-TNF agents.

fliximab-treated IBD cohort in this paper would yield 71% or 99% power to detect a doubling or tripling in malignancy incidence, respectively. Representing data derived from rigorous clinical trials, therefore, the current data are somewhat reassuring and have merit when assessing the overall safety of anti-TNF agents. When taken together, results of these pooled analyses indicate no increase in serious infection, mortality, or malignancy, including lymphoma in association with infliximab treatment of IBD. In addition, the safety of infliximab in these analyses appears comparable to that of conventional immunomodulators. STUDY HIGHLIGHTS We thank James P. Barrett and Mary Ann Thomas of Janssen Biotech, Inc., for their work on the study protocols, Kathryn Mingione of Janssen for her work pertaining to data interpretation, and Michelle Perate and Mary H. Whitman of Janssen Biotech, Inc., for their writing support. Guarantor of the article: Gary R. Lichtenstein, MD. Specific author contributions: Design and conduct of clinical trials contributing data to these pooled analyses: Gary R. Lichtenstein, Paul Rutgeerts, William J. Sandborn, Bruce E. Sands, and Robert H. Diamond; data collection and analysis: Robert H. Diamond, Linda Tang, Jennifer Montello, and Freddy Cornillie; and all authors provided critical content review and final approval of this manuscript. Financial support: Funding for the clinical trials contributing data to these pooled analyses was provided by Janssen Research & Development, a Johnson & Johnson (J&J) pharmaceutical company.

Specific author contributions: Design and conduct of clinical trials contributing data to these pooled analyses: Gary R. Lichtenstein, Paul Rutgeerts, William J. Sandborn, Bruce E. Sands, and Robert H. Diamond; data collection and analysis: Robert H. Diamond, Linda Tang, Jennifer Montello, and Freddy Cornillie; and all authors provided critical content review and final approval of this manuscript. Financial support: Funding for the clinical trials contributing data to these pooled analyses was provided by Janssen Research & Development, a Johnson & Johnson (J&J) pharmaceutical company. Potential competing interests: Dr Lichtenstein has received research grants and/or has served as a consultant for Abbott Corporation, Alaven, Bristol-Myers Squibb, Janssen, Elan, Ferring, Meda Pharmaceuticals, Merck/Schering-Plough, Millennium Pharmaceuticals, Pfizer Pharmaceuticals, Proctor and Gamble, Prometheus Laboratories, Salix Pharmaceuticals, Santarus, Shire Pharmaceuticals, UCB, Warner Chilcotte, and Wyeth. Professor Rutgeerts has received research support from Janssen, Merck/Schering-Plough, UCB, and Abbott Laboratories, as well as consulting and/or speaking honoraria from Janssen, Merck/Schering-Plough, UCB, Abbott Laboratories, Elan-Biogen, NovImmune, Italfarmako, Bristol-Myers Squibb, Millennium Pharmaceuticals, Tillots, Glaxo SmithKline, and ChemoCentryx. Dr Sandborn has received has received consulting fees and research support from Abbott Laboratories, Janssen, and UCB Pharma, as well as consulting fees from Merck/Schering-Plough. Dr Sands has received consulting fees from Abbott Immunology, Axcan Pharma, Avaxia Biologics, Bristol-Myers Squibb, Elan Pharmaceuticals, Emmi Solutions, Janssen, Glaxo Wellcome SmithKline, Millennium Pharmaceuticals/Takeda, Novartis Pharmaceuticals, Pfizer, Prometheus Laboratories, and owns common stock in Avaxia Biologics, a company that is not publicly traded. Professor Colombel has received consulting fees and has participated in continuing medical education events supported by unrestricted educational grants from Janssen and Merck/Schering-Plough. Dr Diamond and Linda Tang are employed by Janssen Research & Development, a J&J pharmaceutical company. Jennifer Montello is an employee of J&J. Freddy Cornillie is an employee of Janssen Biologics BV, a J&J pharmaceutical company.

nts supported by unrestricted educational grants from Janssen and Merck/Schering-Plough. Dr Diamond and Linda Tang are employed by Janssen Research & Development, a J&J pharmaceutical company. Jennifer Montello is an employee of J&J. Freddy Cornillie is an employee of Janssen Biologics BV, a J&J pharmaceutical company. Table 1 Key features of 10 sponsor-initiated studies of infliximab in IBD Study (reference) Pt. population Study design Treatment regimens (no. of pts.

nts supported by unrestricted educational grants from Janssen and Merck/Schering-Plough. Dr Diamond and Linda Tang are employed by Janssen Research & Development, a J&J pharmaceutical company. Jennifer Montello is an employee of J&J. Freddy Cornillie is an employee of Janssen Biologics BV, a J&J pharmaceutical company. Table 1 Key features of 10 sponsor-initiated studies of infliximab in IBD Study (reference) Pt. population Study design Treatment regimens (no. of pts. evaluated) AE reporting period (wks) Smaller trials of infliximab in IBD C0168T08 (15) Severe CD (CDAI >150) refractory to corticosteroid therapy Phase 1, SC, OL, single dose Grp 1: Infliximab 10 mg/kg (n=8)Grp 2: Infliximab 20 mg/kg (n=2) 8 C0168T11 (10) Moderate-to-severea CD Phase 2, MC, OL, single dose, sequential dose-escalating trial Grp 1: Infliximab 1 mg/kg (n=5) 12 Grp 2: Infliximab 5 mg/kg (n=5) Grp 3: Infliximab 10 mg/kg (n=5) Grp 4: Infliximab 20 mg/kg (n=6) C0168T16 (10,12,14) Moderate-to-severea CD Phase 2/3, MC, DB, PC, with initial dose-ranging treatment phase followed by repeated-treatment phase plus OL treatment for safety assessments Initial dose-ranging phase (single dose)Grp 1: Placebo (n=25)Grp 2: Infliximab 5 mg/kg (n=27)Grp 3: Infliximab 10 mg/kg (n=28)Grp 4: Infliximab 20 mg/kg (n=28) 16 (n=35) and 48 (n=73) Open-label phase: Infliximab 10 mg/kg (n=48) Repeated-treatment phase (4 DB infusions)Grp 1: Infliximab 10 mg/kg q8wks (n=37)Grp 2: Placebo q8wks (n=36) C0168T20 (11) Fistulizing CD MC, DB, PC, randomized phase 3 trial Grp 1: Infliximab 10 mg/kg at wks 0, 2, 6 (n=32) 52 Grp 2: Infliximab 5 mg/kg at wks 0, 2, 6 (n=31) Grp 3: Placebo at wks 0, 2, 6 (n=31) C0168T12 (13) Active UC (modified Truelove and Witts score >10) MC, DB, PC, randomized phase 2 trial Single dose of:Grp 1: Placebo (n=3)Grp 2: Infliximab 5 mg/kg (n=3)Grp 3: Infliximab 10 mg/kg (n=3)Grp 4: Infliximab 20 mg/kg (n=2) 12 Pivotal phase 3 trials of infliximab in IBD ACCENT I (5,6) Moderate-to-severea CD MC, DB, PC, phase 3 randomized trialAZA, 6-MP, MTX, corticosteroids allowed but not randomized treatments All pts.: Infliximab 5 mg/kg at wk 0 (n=573)Grp 1: Placebo at wks 2, 6, and q8wks through wk 46 (n=188)Grp 2: Infliximab 5 mg/kg at wks 2, 6, and q8wks through wk 46 (n=192)Grp 3: Infliximab 5 mg/kg at wks 2 and 6, then 10 mg/kg q8wks through wk 46 (n=193) 54 ACCENT II (7) Fistulizing CD MC, DB, PC, phase 3 randomized trial All pts.: Infliximab 5 mg/kg at wks 0, 2, 6 (n=306) 54 AZA, 6-MP, MTX, corticosteroids allowed but not randomized treatments Grp 1: Placebo at wk 14 and q8wks through wk 46 (crossover to 5 mg/kg possible; n=143 for placebo maintenance) Grp 2: Inflixi

54 ACCENT II (7) Fistulizing CD MC, DB, PC, phase 3 randomized trial All pts.: Infliximab 5 mg/kg at wks 0, 2, 6 (n=306) 54 AZA, 6-MP, MTX, corticosteroids allowed but not randomized treatments Grp 1: Placebo at wk 14 and q8wks through wk 46 (crossover to 5 mg/kg possible; n=143 for placebo maintenance) Grp 2: Inflixi mab 5 mg/kg at wk 14 and q8wks through wk 46 (crossover to 10 mg/kg possible; n=139 for infliximab maintenance) SONIC (8) Moderate-to-severeb CD MC, DB, ACC, phase 3 randomized trial Grp 1: AZA 2.5 mg/kg capsules/placebo infusions (n=161) 54c Naïve to IMs and biologics; patients randomized to IM treatment Grp 2: Placebo capsules/infliximab 5 mg/kg infusions (n=163)Grp 3: AZA 2.5 mg/kg capsules/infliximab 5 mg/kg infusions (n=179)Capsules (daily)/infusions (wks 0, 2, 6, q8wks through wk 22) ACT 1 (9) UC (364) in pts.

kg capsules/placebo infusions (n=161) 54c Naïve to IMs and biologics; patients randomized to IM treatment Grp 2: Placebo capsules/infliximab 5 mg/kg infusions (n=163)Grp 3: AZA 2.5 mg/kg capsules/infliximab 5 mg/kg infusions (n=179)Capsules (daily)/infusions (wks 0, 2, 6, q8wks through wk 22) ACT 1 (9) UC (364) in pts. with Mayo score of 6–12 pts., Mayo endoscopic subscore of ≥2, and an inadequate response to or tolerance of oral corticosteroids, 6-MP, and/or AZA MC, DB, PC, phase 3 randomized trialAZA, 6-MP, corticosteroids allowed but not randomized treatments Grp 1: Placebo at wks 0, 2, 6, and q8wks through wk 46 (n=121)Grp 2: Infliximab 5 mg/kg at wks 0, 2, 6, and q8wks through wk 46 (n=121)Grp 3: Infliximab 10 mg/kg at wks 0, 2, 6, and q8wks through wk 46 (n=122) 54 ACT 2 (9) UC (364) in pts. with Mayo score of 6–12 pts., Mayo endoscopic subscore of ≥2, and an inadequate response to or tolerance of 5-ASAs, oral corticosteroids, 6-MP, and/or AZA MC, DB, PC, phase 3 randomized trialAZA, 6-MP, corticosteroids allowed but not randomized treatments Grp 1: Placebo at wks 0, 2, 6, and q8wks through wk 22 (n=123)Grp 2: Infliximab 5 mg/kg at wks 0, 2, 6, and q8wks through wk 22 (n=121)Grp 3: Infliximab 10 mg/kg at wks 0, 2, 6, and q8wks through wk 22 (n=120) 54c ACC, active-comparator-controlled; AE, adverse event; 5-ASAs, 5-aminosalicylates; AZA, azathioprine; CD, Crohn's disease; CDAI, Crohn's disease activity index; DB, double-blind; IBD, inflammatory bowel disease; IM, immunomodulators; MC, multicenter; 6-MP, 6-mercaptopurine; MTX, methotrexate; OL, open label; PC, placebocontrolled; pts., patients; q8wks, every 8 weeks; SC, single center; UC, ulcerative colitis; wks, weeks.

Crohn's disease; CDAI, Crohn's disease activity index; DB, double-blind; IBD, inflammatory bowel disease; IM, immunomodulators; MC, multicenter; 6-MP, 6-mercaptopurine; MTX, methotrexate; OL, open label; PC, placebocontrolled; pts., patients; q8wks, every 8 weeks; SC, single center; UC, ulcerative colitis; wks, weeks. a Baseline CDAI score between 220 and 400, inclusive. b Baseline CDAI score between 220 and 450, inclusive. c Dosing in the SONIC and ACT 2 main studies ended with the week-22 infusion. Safety data from week 30 through week 54 were collected as part of study extensions. Table 2 Extent of exposure to infliximab in the pivotal phase 3 IBD trials through week 46 Crohn's disease Ulcerative colitis ACCENT I ACCENT II SONIC ACT 1 ACT 2 PBOa INF 5 mg/kg INF 10 mg/kg PBOb INF 5 mg/kg AZA+PBO INF 5 mg/kg+PBOc INF 5 mg/kg+AZAc PBO INF 5 mg/kg INF 10 mg/kg PBOc INF 5 mg/kgc INF 10 mg/kgc Pts. treated 188 192 193 143 139 161 163 179 121 121 122 123 121 120 Average no. of infliximab infusions 2.2 6.7 6.8 4.3 7.5 0.0 6.1 6.1 0.0 6.5 6.3 0.0 6.4 6.4 AZA, azathioprine; IBD, inflammatory bowel disease; INF, infliximab; pts., patients; PBO, placebo infusions, except for the INF+PBO group in SONIC, in which case: PBO, placebo capsules. a In ACCENT 1, placebo patients received 5 mg/kg infliximab at week 0, and some of them also received episodic infusions of infliximab 5 mg/kg. b In ACCENT II, placebo patients received 5 mg/kg infliximab at weeks 0, 2, and 6 before randomization at week 14, and some of them also crossed over to receive infusions of infliximab 5 mg/kg.

a In ACCENT 1, placebo patients received 5 mg/kg infliximab at week 0, and some of them also received episodic infusions of infliximab 5 mg/kg. b In ACCENT II, placebo patients received 5 mg/kg infliximab at weeks 0, 2, and 6 before randomization at week 14, and some of them also crossed over to receive infusions of infliximab 5 mg/kg. c Including infliximab infusions received during the main study and the blinded study extension. Table 3 Summary of infections and serious infections through week 54 of the pivotal phase 3 infliximab IBD studies by treatment and immunomodulator use Crohn's diseasea Ulcerative colitisa All inflammatory bowel diseasea Placebob Infliximab Placebob Infliximab Placebob Infliximab Pts. treated 161 1,228 245 485 406 1,713 Total/median pt-yrs of follow-up 108/0.7 1,127/1.0 209/0.6 831/1.0 318/0.6 1,958/1.0 No. (%) of pts. with infection 73 (45.3%) 603 (49.1%) 89 (36.3%) 243 (50.1%) 162 (39.9%) 846 (49.4%) P-valuec 0.402 <0.001 <0.001 Total incidence 144 1,352 224 876 368 2,228 Incidence per 100 pt-yrs 132.81 119.98 106.98 105.41 115.79 113.8 95% CId (112.00, 156.36) (113.67, 126.56) (93.43, 121.94) (98.54, 112.63) (104.26, 128.25) (109.12, 118.62) No. (%) of pts.

03 (49.1%) 89 (36.3%) 243 (50.1%) 162 (39.9%) 846 (49.4%) P-valuec 0.402 <0.001 <0.001 Total incidence 144 1,352 224 876 368 2,228 Incidence per 100 pt-yrs 132.81 119.98 106.98 105.41 115.79 113.8 95% CId (112.00, 156.36) (113.67, 126.56) (93.43, 121.94) (98.54, 112.63) (104.26, 128.25) (109.12, 118.62) No. (%) of pts. with serious infection 9 (5.6%) 55 (4.5%) 6 (2.4%) 26 (5.4%) 15 (3.7%) 81 (4.7%) P-value 0.547 0.085 0.427 Total incidence 9 86 6 42 15 128 Incidence per 100 pt-yrs 8.3 7.63 2.87 5.05 4.72 6.54 95% CI (3.80, 15.76) (6.10, 9.43) (1.05, 6.24) (3.64, 6.83) (2.64, 7.78) (5.45, 7.77) System–organ class/common preferred terms (>0.20/per 100 pt-yrs per group) Resistance mechanism disorder 5.53 4.44 1.91 1.8 3.15 3.32 Abscess 2.77 3.02 0.96 0.36 1.57 1.89 Fever 0 0.27 0 0.24 0 0.26 Infection 0 0.18 0.48 0.84 0.31 0.46 Sepsis 0.92 0.27 0 0.12 0.31 0.2 Cellulitis 0 0.27 0 0 0 0.15 Herpes zoster 0 0.27 0 0 0 0.15 Bacterial infection 1.84 0.09 0.48 0 0.94 0.05 Gastrointestinal system disorder 1.84 1.24 0 0.84 0.63 1.07 Gastroenteritis 1.84 0.18 0 0.36 0.63 0.26 Abdominal pain 0 0.27 0 0 0 0.15 Respiratory system disorder 0.92 0.44 0.96 1.68 0.94 0.97 Pneumonia 0.92 0.44 0 0.96 0.31 0.66 Sinusitis 0 0 0.48 0.12 0.31 0.05 Upper respiratory infection 0 0 0.48 0 0.31 0 Skin and appendages disorder 0 0.53 0 0 0 0.31 Urinary system disorder 0 0.18 0 0.24 0 0.2 Body as a whole–general disorder 0 0.18 0 0 0 0.1 Cardiovascular disorder 0 0.18 0 0 0 0.1 Liver and biliary system disorder 0 0.09 0 0.12 0 0.1 Musculoskeletal system disorder 0 0.18 0 0 0 0.1 Reproductive disorder 0 0.09 0 0.12 0 0.1 Nervous system disorder 0 0.09 0 0 0 0.05 Ear and hearing disorder 0 0 0 0.12 0 0.05 Myo-, endo-, pericardial, coronary and valve disorder 0 0 0 0.12 0 0.05 No immunomodulatore Immuno modulatorf No immunomodulatore Immunomodulatorf No immunomodulatore Immunomodulatorf Pts. treated 776 613 394 334 1,170 947 Total/median pt-yrs of follow-up 715/1.0 520/1.0 541/0.8 500/1.0 1,256/1.0 1,020/1.0 Number (%) of pts. with infection 397 (51.2%) 279 (45.5%) 169 (42.9%) 163 (48.8%) 566 (48.4%) 442 (46.7%) P-value 0.04 0.117 0.457 Total incidence 918 578 500 600 1,418 1,178 Incidence per 100 pt-yrs 128.42 111.07 92.47 120.07 112.93 115.48 95% CI (120.24, 137.00) (102.20, 120.51) (84.54, 100.94) (110.66, 130.08) (107.13, 118.97) (108.98, 122.27) Number (%) of pts.

79 (45.5%) 169 (42.9%) 163 (48.8%) 566 (48.4%) 442 (46.7%) P-value 0.04 0.117 0.457 Total incidence 918 578 500 600 1,418 1,178 Incidence per 100 pt-yrs 128.42 111.07 92.47 120.07 112.93 115.48 95% CI (120.24, 137.00) (102.20, 120.51) (84.54, 100.94) (110.66, 130.08) (107.13, 118.97) (108.98, 122.27) Number (%) of pts. with serious infections 36 (4.6%) 28 (4.6%) 14 (3.6%) 18 (5.4%) 50 (4.3%) 46 (4.9%) P-value 1 0.277 0.53 Total incidence 63 32 18 30 81 62 Incidence per 100 pt-yrs 8.81 6.15 3.33 6 6.45 6.08 95% CI (6.77, 11.28) (4.21, 8.68) (1.97, 5.26) (4.05, 8.57) (5.12, 8.02) (4.66, 7.79) AZA, azathioprine; CI, confidence interval; IBD, inflammatory bowel disease; 6-MP, 6-mercaptopurine; MTX, methotrexate; pts., patients; pt-yrs, patient-years. a Includes 3 Crohn's disease and 2 ulcerative colitis, and thus a total of 5, pivotal phase 3 IBD studies. b With or without concomitant conventional therapy. c P-values comparing treatment or immunomodulator use subgroups were calculated with the use of Fisher's exact test. d 95% CIs based on an exact method. e No receipt of AZA, 6-MP, or MTX at baseline. f Receipt of AZA, 6-MP, or MTX at baseline.

a Includes 3 Crohn's disease and 2 ulcerative colitis, and thus a total of 5, pivotal phase 3 IBD studies. b With or without concomitant conventional therapy. c P-values comparing treatment or immunomodulator use subgroups were calculated with the use of Fisher's exact test. d 95% CIs based on an exact method. e No receipt of AZA, 6-MP, or MTX at baseline. f Receipt of AZA, 6-MP, or MTX at baseline. Table 4 Summary of malignancies (excluding nonmelanoma skin cancers) by treatment both overall and during the main portions of all infliximab IBD studiesa and by immunomodulator use during the controlled portions of the pivotal phase 3 IBD trialsb Crohn's disease Ulcerative colitis All inflammatory bowel disease Placeboc Infliximab Placeboc Infliximab Placeboc Infliximab Overall among all infliximab IBD studiesa Pts. treated 217 1,427 248 493 465 1,920 Total/median pt-yrs of follow-up 124/0.5 1,229/1.0 210/0.6 832/1.0 334/0.6 2,061/1.0 All malignancies No. (%) of pts. with malignancy 2 (0.9%) 6 (0.4%) 0 (0.0%) 5 (1.0%) 2 (0.4%) 11 (0.6%) P-valued 0.286 0.175 1 Incidence per 100 pt-yrs 1.61 0.49 0 0.6 0.6 0.53 95% CIe (0.19, 5.82) (0.18, 1.06) (0.00, 1.43) (0.20, 1.40) (0.07, 2.16) (0.27, 0.95) Lymphoma No. (%) of pts. with malignancy 0 (0.0%) 2 (0.1%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 2 (0.1%) P-value 1 0 1 Incidence per 100 pt-yrs 0 0.16 0 0 0 0.1 95% CI (0.00, 2.41) (0.02, 0.59) (0.00, 1.43) (0.00, 0.36) (0.00, 0.90) (0.01, 0.35) Non-lymphoma malignancies No. (%) of pts. with malignancy 2 (0.9%) 4 (0.3%) 0 (0.0%) 5 (1.0%) 2 (0.4%) 9 (0.5%) P-value 0.182 0.175 1 Incidence per 100 pt-yrs 1.61 0.33 0 0.6 0.6 0.44 95% CI (0.19, 5.82) (0.09, 0.83) (0.00, 1.43) (0.20, 1.40) (0.07, 2.16) (0.20, 0.83) Controlled portions of all infliximab IBD studiesa Pts. treated 217 488 245 483 462 971 Total/median pt-yrs of follow-up 121/0.5 298/0.6 137/0.6 333/0.6 258/0.6 631/0.6 All malignancies No. (%) of pts. with malignancy 2 (0.9%) 0 (0.0%) 0 (0.0%) 2 (0.4%) 2 (0.4%) 2 (0.2%) P-value 0.094 0.553 0.598 Incidence per 100 pt-yrs 1.65 0 0 0.6 0.77 0.32 95% CI (0.20, 5.97) (0.00, 1.00) (0.00, 2.18) (0.07, 2.17) (0.09, 2.80) (0.04, 1.15) Lymphoma No. (%) of pts. with malignancy 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) Incidence per 100 pt-yrs 0 0 0 0 0 0 95% CI (0.00, 2.48) (0.00, 1.00) (0.00, 2.18) (0.00, 0.90) (0.00, 1.16) (0.00, 0.47) Non-lymphoma malignancies No. (%) of pts.

00, 1.00) (0.00, 2.18) (0.07, 2.17) (0.09, 2.80) (0.04, 1.15) Lymphoma No. (%) of pts. with malignancy 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) Incidence per 100 pt-yrs 0 0 0 0 0 0 95% CI (0.00, 2.48) (0.00, 1.00) (0.00, 2.18) (0.00, 0.90) (0.00, 1.16) (0.00, 0.47) Non-lymphoma malignancies No. (%) of pts. with malignancy 2 (0.9%) 0 (0.0%) 0 (0.0%) 2 (0.4%) 2 (0.4%) 2 (0.2%) P-value 0.094 0.553 0.598 Incidence per 100 pt-yrs 1.65 0 0 0.6 0.77 0.32 95% CI (0.20, 5.97) (0.00, 1.00) (0.00, 2.18) (0.07, 2.17) (0.09, 2.80) (0.04, 1.15) No immunomodulatorf Immuno modulatorg No immunomodulatorf Immuno modulatorg No immunomodulatorf Immunomodulatorg Controlled portions of 5 pivotal IBD studiesb Pts. treated 166 337 394 334 560 671 All malignancies Total/median pt-yrs of follow-up 129/1.0 250/0.9 250/0.6 220/0.6 378/0.6 470/0.7 No. (%) of pts. with malignancy 0 (0.0%) 2 (0.6%) 1 (0.3%) 1 (0.3%) 1 (0.2%) 3 (0.5%) P-value 1 1 0.631 Incidence per 100 pt-yrs 0 0.8 0.4 0.45 0.26 0.64 95% CI (0.00, 2.33) (0.10, 2.89) (0.01, 2.23) (0.01, 2.53) (0.01, 1.47) (0.13, 1.87) Expected no. of pts.h 0.43 0.71 1.22 0.88 1.65 1.6 SIRi 0 2.8 0.82 1.13 0.61 1.88 SIR 95% CI (0.00, 6.92) (0.34, 10.11) (0.02, 4.58) (0.03, 6.30) (0.02, 3.38) (0.39, 5.48) No immunomodulatorf Immuno modulatorg No immunomodulatorf Immuno modulatorg No immunomodulatorf Immunomodulatorg Placebo Pts. treated 0 161 137 108 137 269 Total/median pt-yrs of follow-up 0/0.0 108/0.7 75/0.6 62/0.6 75/0.6 170/0.6 No. of pts. (%) with malignancy 0 (0.0%) 2 (1.2%) 0 (0.0%) 0 (0.0%) 0 (0. 0%) 2 (0.7%) P-value 0 0 0.552 Incidence per 100 pt-yrs 0 1.84 0 0 0 1.17 95% CI (0.00, 0.00) (0.22, 6.66) (0.00, 3.97) (0.00, 4.85) (0.00, 3.97) (0.14, 4.24) Expected no. of pts. 0 0.35 0.35 0.23 0.36 0.59 SIR 0 5.7 0 0 0 3.41 SIR 95% CI (0.00, 0.00) (0.69, 20.59) (0.00, 8.59) (0.00, 12.76) (0.00, 8.24) (0.41, 12.33) Infliximab Pts. treated 166 176 257 226 423 402 Total/median pt-yrs of follow-up 129/1.0 142/1.0 174/0.6 158/0.6 303/0.6 300/0.7 No. of pts. (%) with malignancy 0 (0.0%) 0 (0.0%) 1 (0.4%) 1 (0.4%) 1 (0.2%) 1 (0.3%) P-value (infliximab vs. placebo) 0 0.228 1 1 1 0.568 P-value (no immunomodulator vs. immunomodulator) 0 1 1 Incidence per 100 pt-yrs 0 0 0.57 0.63 0.33 0.33 95% CI (0.00, 2.33) (0.00, 2.12) (0.01, 3.20) (0.02, 3.52) (0.01, 1.84) (0.01, 1.86) Expected no. of pts. 0.43 0.36 0.82 0.65 1.29 1.01 SIR 0 0 1.22 1.54 0.78 0.99 SIR 95% CI (0.00, 6.92) (0.00, 8.24) (0.03, 6.82) (0.04, 8.57) (0.02, 4.33) (0.02, 5.50) Lymphoma No.

0 1 1 Incidence per 100 pt-yrs 0 0 0.57 0.63 0.33 0.33 95% CI (0.00, 2.33) (0.00, 2.12) (0.01, 3.20) (0.02, 3.52) (0.01, 1.84) (0.01, 1.86) Expected no. of pts. 0.43 0.36 0.82 0.65 1.29 1.01 SIR 0 0 1.22 1.54 0.78 0.99 SIR 95% CI (0.00, 6.92) (0.00, 8.24) (0.03, 6.82) (0.04, 8.57) (0.02, 4.33) (0.02, 5.50) Lymphoma No. (%) of pts. with malignancy 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) Incidence per 100 pt-yrs 0 0 0 0 0 0 95% CI (0.00, 2.33) (0.00, 2.12) (0.00, 1.72) (0.00, 1.89) (0.00, 0.99) (0.00, 1.00) Expected no. of pts. 0.02 0.02 0.04 0.03 0.05 0.05 SIR 0 0 0 0 0 0 SIR 95% CI (0.00, 157.21) (0.00,163.60) (0.00, 83.69) (0.00, 101.69) (0.00, 54.61) (0.00, 62.71) Non-lymphoma malignancies No. (%) of pts. with malignancy 0 (0.0%) 0 (0.0%) 1 (0.4%) 1 (0.4%) 1 (0.2%) 1 (0.3%) P-value 0 1 1 Incidence per 100 pt-yrs 0 0 0.57 0.63 0.33 0.33 95% CI (0.00, 2.33) (0.00, 2.12) (0.01, 3.20) (0.02, 3.52) (0.01, 1.84) (0.01, 1.86) Expected no. of pts. 0.41 0.35 0.82 0.62 1.23 0.97 SIR 0 0 1.22 1.61 0.81 1.03 SIR 95% CI (0.00, 7.24) (0.00, 8.67) (0.03, 6.82) (0.04, 8.98) (0.02, 4.53) (0.03, 5.77) AZA, azathioprine; CI, confidence interval; IBD, inflammatory bowel disease; 6-MP, 6-mercaptopurine; MTX, methotrexate; pts., patients; pt-yrs, patient-years; SEER, Surveillance, Epidemiology, and End Results; SIR, standardized incidence ratio. a Includes 7 Crohn's disease and 3 ulcerative colitis, and thus a total of 10, IBD studies. b Includes 3 Crohn's disease and 2 ulcerative colitis, and thus a total of 5 pivotal phase 3 IBD studies. c With or without concomitant conventional therapy.

(%) of pts. with malignancy 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) Incidence per 100 pt-yrs 0 0 0 0 0 0 95% CI (0.00, 2.33) (0.00, 2.12) (0.00, 1.72) (0.00, 1.89) (0.00, 0.99) (0.00, 1.00) Expected no. of pts. 0.02 0.02 0.04 0.03 0.05 0.05 SIR 0 0 0 0 0 0 SIR 95% CI (0.00, 157.21) (0.00,163.60) (0.00, 83.69) (0.00, 101.69) (0.00, 54.61) (0.00, 62.71) Non-lymphoma malignancies No. (%) of pts. with malignancy 0 (0.0%) 0 (0.0%) 1 (0.4%) 1 (0.4%) 1 (0.2%) 1 (0.3%) P-value 0 1 1 Incidence per 100 pt-yrs 0 0 0.57 0.63 0.33 0.33 95% CI (0.00, 2.33) (0.00, 2.12) (0.01, 3.20) (0.02, 3.52) (0.01, 1.84) (0.01, 1.86) Expected no. of pts. 0.41 0.35 0.82 0.62 1.23 0.97 SIR 0 0 1.22 1.61 0.81 1.03 SIR 95% CI (0.00, 7.24) (0.00, 8.67) (0.03, 6.82) (0.04, 8.98) (0.02, 4.53) (0.03, 5.77) AZA, azathioprine; CI, confidence interval; IBD, inflammatory bowel disease; 6-MP, 6-mercaptopurine; MTX, methotrexate; pts., patients; pt-yrs, patient-years; SEER, Surveillance, Epidemiology, and End Results; SIR, standardized incidence ratio. a Includes 7 Crohn's disease and 3 ulcerative colitis, and thus a total of 10, IBD studies. b Includes 3 Crohn's disease and 2 ulcerative colitis, and thus a total of 5 pivotal phase 3 IBD studies. c With or without concomitant conventional therapy. d P-values comparing treatment or immunomodulator use subgroups were calculated with the use of Fisher's exact test. Note that P-values cannot be computed when a group has no patients or when neither group has such an event. e 95% CIs based on an exact method. f No receipt of AZA, 6-MP, or MTX at baseline.

c With or without concomitant conventional therapy. d P-values comparing treatment or immunomodulator use subgroups were calculated with the use of Fisher's exact test. Note that P-values cannot be computed when a group has no patients or when neither group has such an event. e 95% CIs based on an exact method. f No receipt of AZA, 6-MP, or MTX at baseline. g Receipt of AZA, 6-MP, or MTX at baseline. h Based on the SEER database (2002) adjusted for age, sex, and race. i Calculated as the quotient of the observed and expected numbers of patients with malignancy.

d P-values comparing treatment or immunomodulator use subgroups were calculated with the use of Fisher's exact test. Note that P-values cannot be computed when a group has no patients or when neither group has such an event. e 95% CIs based on an exact method. f No receipt of AZA, 6-MP, or MTX at baseline. g Receipt of AZA, 6-MP, or MTX at baseline. h Based on the SEER database (2002) adjusted for age, sex, and race. i Calculated as the quotient of the observed and expected numbers of patients with malignancy. Table 5 Summary of mortality through week 54 by treatment (all IBD studies) and immunomodulator use status (pivotal phase 3 IBD trials) Crohn's diseasea Ulcerative colitisa All inflammatory bowel diseasea Placebob Infliximab Placebob Infliximab Placebob Infliximab Pts. treated 217 1,427 248 493 465 1,920 Total/median pt-yrs of follow-up 124/0.5 1,230/1.0 210/0.6 833/1.0 334/0.6 2,063/1.0 Deaths No. (%) of pts. 1 (0.5%) 3 (0.2%) 0 (0.0%) 1 (0.2%) 1 (0.2%) 4 (0.2%) P-valuec 0.433 1 1 Incidence/100 pt-yrs 0.8 0.24 0 0.12 0.3 0.19 95% CId (0.02, 4.48) (0.05, 0.71) (0.00, 1.43) (0.00, 0.67) (0.01, 1.67) (0.05, 0.50) Infection-related deaths No. (%) of pts. 1 (0.5%) 2 (0.1%) 0 (0.0%) 1 (0.2%) 1 (0.2%) 3 (0.2%) P-value 0.346 1 0.58 Incidence/100 pt-yrs 0.8 0.16 0 0.12 0.3 0.15 95% CI (0.02, 4.48) (0.02, 0.59) (0.00, 1.43) (0.00, 0.67) (0.01, 1.67) (0.03, 0.43) No immunomodulatore Immunomodulatorf No immunomodulatore Immunomodulatorf No immunomodulatore Immunomodulatorf Pts. treated 776 613 394 334 1,170 947 Total/median pt-yrs of follow-up 715/1.0 520/1.0 541/0.8 500/1.0 1,256/1.0 1,020/1.0 Deaths No. (%) of pts. 2 (0.3%) 2 (0.3%) 0 (0.0%) 1 (0.3%) 2 (0.2%) 3 (0.3%) P-value 1 0.459 0.662 Incidence/100 pt-yrs 0.28 0.38 0 0.2 0.16 0.29 95% CI (0.03, 1.01) (0.05, 1.39) (0.00, 0.55) (0.01, 1.11) (0.02, 0.58) (0.06, 0.86) AZA, azathioprine; CI, confidence interval; IBD, inflammatory bowel disease; 6-MP, 6-mercaptopurine; MTX, methotrexate; pts., patients; pt-yrs, patient-years.

(0.3%) P-value 1 0.459 0.662 Incidence/100 pt-yrs 0.28 0.38 0 0.2 0.16 0.29 95% CI (0.03, 1.01) (0.05, 1.39) (0.00, 0.55) (0.01, 1.11) (0.02, 0.58) (0.06, 0.86) AZA, azathioprine; CI, confidence interval; IBD, inflammatory bowel disease; 6-MP, 6-mercaptopurine; MTX, methotrexate; pts., patients; pt-yrs, patient-years. a Includes 7 Crohn's disease and 3 ulcerative colitis, or a total of 10, IBD studies when summarized by treatment and 3 Crohn's disease and 2 ulcerative colitis, or a total of 5, pivotal phase 3 IBD studies when summarized by baseline immunomodulator use. b With or without concomitant conventional therapy. c P-values comparing treatment or immunomodulator use subgroups were calculated with the use of Fisher's exact test. d 95% CIs based on an exact method. e No receipt of AZA, 6-MP, or MTX at baseline. f Receipt of AZA, 6-MP, or MTX at baseline.

INTRODUCTION Non-steroidal anti-inflammatory drugs (NSAIDs) are widely used as analgesics (1,2). Ibuprofen is the most commonly used non-aspirin NSAID, with about one quarter of the US population aged 17 years or older reporting ibuprofen intake in the past month (2). The main factor that limits NSAID use is the development of upper gastrointestinal (GI) adverse effects including ulcers, complications such as bleeding, and dyspepsia. Strategies recommended (3,4,5,6) to decrease GI injury in NSAID users include co-therapy with misoprostol or proton-pump inhibitors and/or use of COX-2 selective inhibitors. Histamine2-receptor antagonists (H2RAs) have not been recommended for preventive therapy in NSAID users because, when given in standard doses, they significantly decrease duodenal but not gastric ulcers (3,7). However, a Cochrane systematic review of placebo-controlled randomized trials ≥3 months (7) identified one study showing a significant benefit of double-dose H2RAs in reducing both gastric ulcers (relative risk (RR)=0.42, 95% confidence interval, 0.18–0.97) and duodenal ulcers (RR=0.19, 0.04–0.85) (8). The RRs for proton-pump inhibitor vs. placebo co-therapy in the meta-analysis were 0.40 (0.32–0.51) for gastric ulcers and 0.19 (0.09–0.37) for duodenal ulcers (7).

uble-dose H2RAs in reducing both gastric ulcers (relative risk (RR)=0.42, 95% confidence interval, 0.18–0.97) and duodenal ulcers (RR=0.19, 0.04–0.85) (8). The RRs for proton-pump inhibitor vs. placebo co-therapy in the meta-analysis were 0.40 (0.32–0.51) for gastric ulcers and 0.19 (0.09–0.37) for duodenal ulcers (7). Although many national and international groups recommend protective therapy in NSAID users at increased risk of GI events (3,4,5,6), most such patients are not prescribed protective therapy (8,9). In addition, patients may not take their protective co-therapy along with their NSAID, especially if they are not experiencing any symptoms, and decreased adherence is associated with a significantly increased risk of developing upper GI ulcers or bleeding (10,11,12,13). For example, van Soest et al. (13) reported that the risk of an upper GI clinical event (bleeding, perforation, or symptomatic ulcer) in NSAID users at increased GI risk rose 16% for every 10% decrease in the proportion of time a proton-pump inhibitor or H2RA was prescribed. Mechanisms that ensure adherence to protective therapy, such as the combination of an NSAID and a protective agent in one pill, theoretically should decrease the GI risk associated with the NSAID therapy.

d GI risk rose 16% for every 10% decrease in the proportion of time a proton-pump inhibitor or H2RA was prescribed. Mechanisms that ensure adherence to protective therapy, such as the combination of an NSAID and a protective agent in one pill, theoretically should decrease the GI risk associated with the NSAID therapy. We performed the REDUCE trials (Registration Endoscopic Studies to Determine Ulcer Formation of HZT-501 Compared with Ibuprofen: Efficacy and Safety Studies) to assess whether double-dose famotidine given in a single-tablet combination with ibuprofen (HZT-501 (ibuprofen 800 mg plus famotidine 26.6 mg); Horizon Pharma, Northbrook, IL) significantly reduces the proportion of patients who develop gastric ulcers (REDUCE-1) or upper GI (gastric or duodenal) ulcers (REDUCE-2) during 24 weeks of treatment as compared with ibuprofen in adult NSAID users.

ion with ibuprofen (HZT-501 (ibuprofen 800 mg plus famotidine 26.6 mg); Horizon Pharma, Northbrook, IL) significantly reduces the proportion of patients who develop gastric ulcers (REDUCE-1) or upper GI (gastric or duodenal) ulcers (REDUCE-2) during 24 weeks of treatment as compared with ibuprofen in adult NSAID users. METHODS Patients Male and female patients aged 40–80 years expected to require daily NSAID therapy for at least 6 months for pain and/or inflammatory conditions were eligible. Exclusion criteria included history of erosive esophagitis; history of GI complications (bleeding, perforated ulcer, gastric outlet obstruction due to an ulcer); history of NSAID-associated asthma exacerbations, acute renal failure, interstitial nephritis, or hepatitis; history of GI malignancy; history of myocardial infarction, unstable cardiac arrhythmias, or stroke within 6 months of study entry; coronary artery bypass graft surgery within 14 days of study entry; uncontrolled congestive heart failure or hypertension at entry; acid-suppressive therapy or misoprostol within 14 days before study entry or investigational drug or NSAIDs (including aspirin >325 mg daily) within 30 days before study entry; ulcer or >5 erosions on screening upper GI tract endoscopy; or one of the following abnormalities on baseline laboratory testing: creatinine clearance <45 ml/min; aminotransferase >2.5 times upper limit of normal; fasting blood sugar >200 mg/dl; serum pregnancy test positive; serologic tests positive for human immunodeficiency virus, hepatitis B, or hepatitis C; or stool antigen for H. pylori positive.

bnormalities on baseline laboratory testing: creatinine clearance <45 ml/min; aminotransferase >2.5 times upper limit of normal; fasting blood sugar >200 mg/dl; serum pregnancy test positive; serologic tests positive for human immunodeficiency virus, hepatitis B, or hepatitis C; or stool antigen for H. pylori positive. Study design Patients were randomly assigned, using a computer-generated randomization schedule, from a central location utilizing an interactive voice response system with blinded medication kit number allocation in a 2:1 ratio to identical-appearing tablets of HZT-501 (800 mg ibuprofen and 26.6 mg famotidine) or ibuprofen (800 mg) thrice daily for 24 weeks. Patients, care providers, and all study personnel were blinded to the treatment. Patients were stratified for two risk factors for ulcer development: concomitant use of low-dose aspirin (≤325 mg daily) and/or anticoagulant medication and history of gastric or duodenal ulcer. In addition to the screening upper endoscopy at baseline, patients had endoscopy at weeks 8, 16, and 24 (or earlier if premature study termination) of study therapy. The following medications were proscribed during the study: medications that may reduce ulcers (e.g., misoprostol, proton-pump inhibitors, and non-study H2RAs), non-study NSAIDs other than low-dose aspirin taken for cardiovascular prophylaxis. In addition, antacids could not be taken for >3 days in any 2-week period; patients requiring further antacid therapy were to be discontinued from the trial. The study medication was dispensed in an 8-week supply at 0, 8, and 16 weeks. Compliance was determined by pill count of returned bottles of study medication. Serum chemistries, complete blood count, and prothrombin time were performed at screening, week 8, week 16, and the final study visit (week 24 or earlier if early termination). Urinalysis was done at baseline and final visit.

and 16 weeks. Compliance was determined by pill count of returned bottles of study medication. Serum chemistries, complete blood count, and prothrombin time were performed at screening, week 8, week 16, and the final study visit (week 24 or earlier if early termination). Urinalysis was done at baseline and final visit. End points and analysis The primary end point for REDUCE-1 was gastric ulcers identified at endoscopy during the 24-week study period, with three secondary end points: upper GI ulcers (gastric and duodenal), duodenal ulcers, and GI complications (bleeding, ulcer perforation, and gastric outlet obstruction due to ulcer). The primary end point for REDUCE-2 was upper GI (gastric or duodenal) ulcers identified at endoscopy during the 24-week study period, with three secondary end points: gastric ulcers, duodenal ulcers, and GI complications. An endoscopic diagnosis of ulcer required unequivocal depth and diameter of ≥3 mm. The predefined primary population for analysis included all patients who were randomized, received a dose of study medication, and had at least one study-mandated follow-up endoscopy.

astric ulcers, duodenal ulcers, and GI complications. An endoscopic diagnosis of ulcer required unequivocal depth and diameter of ≥3 mm. The predefined primary population for analysis included all patients who were randomized, received a dose of study medication, and had at least one study-mandated follow-up endoscopy. A sample size for REDUCE-1 of 875 was calculated based on a 90% power to detect a difference of 6% vs. 14% in the incidence of gastric ulcers with a two-sided α of 0.05 and assuming a 15% drop-out rate. A sample size for REDUCE-2 of 600 was calculated based on a 90% power to detect a difference of 6% vs. 16% in the incidence of upper GI ulcers with a two-sided α of 0.05 and assuming a 15% drop-out rate. Statistical comparisons were predefined to be done with a fixed testing sequence (hierarchical) in the following order of primary followed by secondary end points: REDUCE-1: gastric ulcers, upper GI ulcers, duodenal ulcers, GI complications; REDUCE-2: upper GI ulcers, gastric ulcers, duodenal ulcers, GI complications. With this approach, the first null hypothesis that is accepted (i.e., P≥0.05) will cause immediate acceptance of all subsequent null hypotheses in the sequence (subsequent comparisons will be considered not significantly different and no statistical comparison will be performed). This approach also requires no α adjustment for multiple comparisons (14).

hypothesis that is accepted (i.e., P≥0.05) will cause immediate acceptance of all subsequent null hypotheses in the sequence (subsequent comparisons will be considered not significantly different and no statistical comparison will be performed). This approach also requires no α adjustment for multiple comparisons (14). The two different primary end points in the individual studies were chosen to address both the more clinically relevant end point of upper GI ulcers (gastric and/or duodenal) and the traditional US FDA (Food and Drug Administration) end point of gastric ulcers. In clinical practice, physicians and patients are concerned about preventing ulcers, whether they are gastric or duodenal, and many clinical trials use this primary end point. However, the FDA generally has approved antisecretory medications for prevention of gastric ulcers (e.g., lansoprazole and esomeprazole). These two trials were designed to be used for registration, and, in pretrial meetings, the FDA agreed to the use of the different end points in the two trials.

his primary end point. However, the FDA generally has approved antisecretory medications for prevention of gastric ulcers (e.g., lansoprazole and esomeprazole). These two trials were designed to be used for registration, and, in pretrial meetings, the FDA agreed to the use of the different end points in the two trials. Comparison between treatment arms of the crude proportions of patients with ulcers at 24 weeks with a Cochran–Mantel–Haenszel test stratified by the two randomization risk factor strata (use of low-dose aspirin and other anticoagulants, prior ulcer history) was specified as the primary statistical analysis at the time of sample size calculation and study initiation. Before study termination and unblinding, the primary analysis was changed to a comparison between treatment arms of the life table estimates of the proportion of patients with ulcers at 24 weeks employing a modified χ2 using the sum of squares from the life table in the error term, with the comparison of crude proportions using the Cochran–Mantel–Haenszel test maintained as a secondary analysis. Numbers-needed-to-treat (NNTs) and absolute risk reductions were calculated using crude proportions.

ers at 24 weeks employing a modified χ2 using the sum of squares from the life table in the error term, with the comparison of crude proportions using the Cochran–Mantel–Haenszel test maintained as a secondary analysis. Numbers-needed-to-treat (NNTs) and absolute risk reductions were calculated using crude proportions. A pooled analysis of the patients in REDUCE-1 and -2 was prespecified, with the primary end point being upper GI ulcers, and the secondary end points being gastric ulcers and duodenal ulcers. The proportion of patients developing ulcers was also prespecified to be assessed in the following subgroups: use of low-dose aspirin and/or anticoagulants, prior ulcer history, age (≥65 vs. <65 years), gender, and race. In addition, post hoc subgroup analysis included use of low-dose aspirin alone, and presence or absence of erosions at screening endoscopy. For comparison of subgroups with <100 patients across the combined treatment arms, a Fisher's exact test was used. The independent variables prior ulcer history, gender, age, low-dose aspirin use, baseline erosions, and therapy also were included in a proportional hazards model to determine the effect on the dependent variable of development of upper GI ulcer. Treatment-by-subgroup interaction was also assessed among these subgroups in the model. A post hoc fixed effect meta-analysis of the results of the two studies for end points of upper GI, gastric, and duodenal ulcers was also performed (Review Manager 5.1, Cochrane Collaboration, Copenhagen, Denmark).

f upper GI ulcer. Treatment-by-subgroup interaction was also assessed among these subgroups in the model. A post hoc fixed effect meta-analysis of the results of the two studies for end points of upper GI, gastric, and duodenal ulcers was also performed (Review Manager 5.1, Cochrane Collaboration, Copenhagen, Denmark). Statistical comparison between the study groups in the population of all patients randomized was also prespecified for common adverse events (occurring in >5% of patients) with a Cochran–Mantel–Haenszel test. In addition, we prespecified comparison of the proportion of patients who reported any symptom consistent with dyspepsia (e.g., dyspepsia, upper abdominal pain or discomfort, epigastric pain or discomfort, stomach pain or discomfort; with and without nausea). All study patients provided written informed consent and the study was approved by institutional review boards for all participating centers. RESULTS The progress of the studies from enrollment through analysis is shown in the CONSORT diagrams in Figure 1. The REDUCE-1 trial took place from March 2007 through August 2008 at 68 centers in the United States and REDUCE-2 took place from March 2007 through September 2008 at 68 centers in the United States. The trials continued until their planned completion based on the prespecified sample sizes. Selected baseline characteristics of the primary population analyzed in the two study groups for each trial are shown in Table 1. Reason for NSAID use across both trials included pain (37.9%), osteoarthritis (50.6%), rheumatoid arthritis (3.5%), and other (7.8%).

anned completion based on the prespecified sample sizes. Selected baseline characteristics of the primary population analyzed in the two study groups for each trial are shown in Table 1. Reason for NSAID use across both trials included pain (37.9%), osteoarthritis (50.6%), rheumatoid arthritis (3.5%), and other (7.8%). The initial results after data lock and unblinding in REDUCE-1 for the primary end point of proportion of patients with gastric ulcers at 24 weeks was 12.9% for HZT-501 and 25.3% for ibuprofen (P=0.0009; NNT=11). The initial results for the primary end point in REDUCE-2 showed the proportions of patients with upper GI ulcers of 13.8% for HZT-501 and 22.6% for ibuprofen (P=0.0304; NNT=11).

primary end point of proportion of patients with gastric ulcers at 24 weeks was 12.9% for HZT-501 and 25.3% for ibuprofen (P=0.0009; NNT=11). The initial results for the primary end point in REDUCE-2 showed the proportions of patients with upper GI ulcers of 13.8% for HZT-501 and 22.6% for ibuprofen (P=0.0304; NNT=11). After unblinding and performance of the statistical analyses, a review of endoscopy reports revealed that some patients with esophageal ulcers had been incorrectly assigned as having gastric ulcers. An adjudication committee of two independent gastroenterologists then reviewed all positive endoscopy reports to determine how many ulcers had been misclassified. In the primary population, five patients in REDUCE-1 (one HZT-501 (no ulcer), four ibuprofen (four esophageal ulcers)) and seven patients in REDUCE-2 (three HZT-501 (two no ulcer, one esophageal ulcer), four ibuprofen (one no ulcer, three esophageal ulcers)) had esophageal ulcers misclassified as gastric ulcers or had no gastric ulcer present; duodenal ulcers were not affected. The results for the analyses were then re-run with these patients no longer listed as having gastric ulcers. These post-adjudication results are presented in Table 2. In REDUCE-1, the primary end point of proportion of patients with gastric ulcers at 24 weeks was 12.7% for HZT-501 and 22.9% for ibuprofen (P=0.0044; NNT=12). The results for the primary end point in REDUCE-2 showed the proportions of patients with upper GI ulcers of 13.0% for HZT-501 and 20.5% for ibuprofen (P=0.0587).

n REDUCE-1, the primary end point of proportion of patients with gastric ulcers at 24 weeks was 12.7% for HZT-501 and 22.9% for ibuprofen (P=0.0044; NNT=12). The results for the primary end point in REDUCE-2 showed the proportions of patients with upper GI ulcers of 13.0% for HZT-501 and 20.5% for ibuprofen (P=0.0587). Assessment of the secondary ulcer end points in REDUCE-1 revealed that significantly fewer patients receiving HZT-501 developed upper GI ulcers in both the initial (NNT=9) and post-adjudication analysis (NNT=10), and significantly fewer patients developed duodenal ulcers (NNT=25). Assessment of secondary end points in REDUCE-2 showed that the 6.7% reduction in life table estimate of gastric ulcers was not significant in the initial comparison (P=0.0795) and the 5.3% reduction in the post-adjudication analysis was not formally tested based on the predefined hierarchical testing sequence for the analysis. The combined trial cumulative incidence for crude rate of ulcers by visit is shown in Figure 2. Results of the pooled analysis of the two studies are also shown in Table 2. When a fixed effect meta-analysis of the two studies was performed, the differences and 95% confidence intervals were identical to those shown in Table 2 for the pooled analyses of upper GI, gastric, and duodenal ulcers without evidence of heterogeneity (I2=0 for upper GI, gastric, and duodenal ulcer analyses). In addition, when the Cochran–Mantel–Haenszel analysis was re-run with study (REDUCE-1 or -2) included in the model, the P values for comparisons of all three end points remained <0.0001.

INTRODUCTION Non-steroidal anti-inflammatory drugs (NSAIDs) are widely prescribed for the treatment of pain and inflammation in patients with various musculoskeletal conditions. It is well known that NSAIDs induce gastrointestinal (GI) adverse effects, including serious complications such as upper GI bleeding, perforation, obstruction, and death (1,2). Patients with risk factors are more likely to develop serious complications. Risk factors for GI complications are well known and depend on patient characteristics, medical history, and the type of NSAID prescribed (1,2). To reduce the risk of these adverse events, different scientific societies and regulatory authorities have developed guidelines and recommendations to indicate that patients with risk factors should receive preventive treatment including, among others, the co-prescription of gastroprotective agents (GPAs), namely a standard dose of proton pump inhibitors (PPIs), misoprostol, or high-dose famotidine (3,4).

upper GI, gastric, and duodenal ulcers without evidence of heterogeneity (I2=0 for upper GI, gastric, and duodenal ulcer analyses). In addition, when the Cochran–Mantel–Haenszel analysis was re-run with study (REDUCE-1 or -2) included in the model, the P values for comparisons of all three end points remained <0.0001. The combined trial data of the primary population was used in a proportional hazards analysis to examine multiple risk factors in relation to development of upper GI ulcers. Treatment with HZT-501 (vs. ibuprofen) was associated with a significantly lower risk ratio (RR) for upper GI ulcer formation in the initial (RR=0.44, 0.33–0.58) and post-adjudication analyses (RR=0.46, 0.34–0.61). Patients ≥65 years had a significantly higher risk ratio of upper GI ulcer formation (initial: 1.47, 1.06–2.05; post-adjudication: 1.54, 1.10–2.17) than those <65 years. RRs of upper GI ulcer formation for low-dose aspirin use were 1.40 (0.97–2.02) for initial and 1.46 (1.004–2.11) for post-adjudication analyses and for prior ulcer history they were 1.55 (0.95–2.53) for initial and 1.65 (1.01–2.69) for post-adjudication analyses. Baseline erosions did not show a significant association with upper GI ulcer formation (initial: 1.20, 0.86–1.67; post-adjudication: 1.15, 0.82–1.62). The effect of treatment was also examined within subgroups for the combined studies. The forest plot of these is shown in Figure 3. No significant treatment-by-subgroup interactions were identified for the subgroups assessed (all P values >0.10).

The combined trial data of the primary population was used in a proportional hazards analysis to examine multiple risk factors in relation to development of upper GI ulcers. Treatment with HZT-501 (vs. ibuprofen) was associated with a significantly lower risk ratio (RR) for upper GI ulcer formation in the initial (RR=0.44, 0.33–0.58) and post-adjudication analyses (RR=0.46, 0.34–0.61). Patients ≥65 years had a significantly higher risk ratio of upper GI ulcer formation (initial: 1.47, 1.06–2.05; post-adjudication: 1.54, 1.10–2.17) than those <65 years. RRs of upper GI ulcer formation for low-dose aspirin use were 1.40 (0.97–2.02) for initial and 1.46 (1.004–2.11) for post-adjudication analyses and for prior ulcer history they were 1.55 (0.95–2.53) for initial and 1.65 (1.01–2.69) for post-adjudication analyses. Baseline erosions did not show a significant association with upper GI ulcer formation (initial: 1.20, 0.86–1.67; post-adjudication: 1.15, 0.82–1.62). The effect of treatment was also examined within subgroups for the combined studies. The forest plot of these is shown in Figure 3. No significant treatment-by-subgroup interactions were identified for the subgroups assessed (all P values >0.10). Only two adverse event terms were reported in either study at a rate of ≥5%: dyspepsia and nausea. The rates for HZT-501 vs. ibuprofen in the pooled results were 4.7% vs. 8.0% for dyspepsia and 5.8% vs. 4.7% for nausea. However, the proportion of patients with one or more predefined symptoms consistent with dyspepsia was not significantly different between HZT-501 and ibuprofen (12.3% vs. 14.9%).

spepsia and nausea. The rates for HZT-501 vs. ibuprofen in the pooled results were 4.7% vs. 8.0% for dyspepsia and 5.8% vs. 4.7% for nausea. However, the proportion of patients with one or more predefined symptoms consistent with dyspepsia was not significantly different between HZT-501 and ibuprofen (12.3% vs. 14.9%). GI complications were reported by investigators for three patients (0.6%) given HZT-501 and zero given ibuprofen in REDUCE-1. These were GI bleeding episodes in which hemoglobin dropped 1.6, 2.1, and 3.1 g/dl without clinical evidence of overt GI bleeding, transfusions, or hospitalizations. Gastric erosions were noted at endoscopy in all three patients. No GI complications were reported in REDUCE-2. One death occurred in the two studies: a 48-year-old female in the ibuprofen group died of cardiorespiratory arrest and multiorgan system failure attributed to acetaminophen toxicity. DISCUSSION The goal of the REDUCE trials was to confirm the results of a single smaller study by Taha et al. (15), and demonstrate a significant decrease in NSAID-associated gastric and duodenal ulcers with double-dose famotidine, using a combination tablet of ibuprofen and famotidine.

GI complications were reported by investigators for three patients (0.6%) given HZT-501 and zero given ibuprofen in REDUCE-1. These were GI bleeding episodes in which hemoglobin dropped 1.6, 2.1, and 3.1 g/dl without clinical evidence of overt GI bleeding, transfusions, or hospitalizations. Gastric erosions were noted at endoscopy in all three patients. No GI complications were reported in REDUCE-2. One death occurred in the two studies: a 48-year-old female in the ibuprofen group died of cardiorespiratory arrest and multiorgan system failure attributed to acetaminophen toxicity. DISCUSSION The goal of the REDUCE trials was to confirm the results of a single smaller study by Taha et al. (15), and demonstrate a significant decrease in NSAID-associated gastric and duodenal ulcers with double-dose famotidine, using a combination tablet of ibuprofen and famotidine. Two studies of identical design were performed, as is generally required when phase 3 trials are used to apply for regulatory approval. The trials differed only in their sample size and their primary vs. secondary end point (gastric ulcers followed by upper GI ulcers in REDUCE-1 and upper GI ulcers followed by gastric ulcers in REDUCE-2). After unblinding and initial analyses were performed, it was discovered on adjudication of endoscopy reports that 12 patients listed as having gastric ulcers actually had esophageal ulcers or no ulcers. Therefore, analyses were re-run using the post-adjudication data, with these 12 patients no longer listed as having gastric ulcers.

ng and initial analyses were performed, it was discovered on adjudication of endoscopy reports that 12 patients listed as having gastric ulcers actually had esophageal ulcers or no ulcers. Therefore, analyses were re-run using the post-adjudication data, with these 12 patients no longer listed as having gastric ulcers. The larger REDUCE-1 trial demonstrated a significant decrease in the primary end point of gastric ulcers (P=0.0044) as well as in duodenal ulcers. The smaller REDUCE-2 trial did not show a significant difference in the post-adjudication primary analysis of life table estimates of upper GI ulcers, with a P value of 0.0587; the secondary statistical comparison of crude proportions of upper GI ulcers revealed a P value of 0.0070. Pooled analysis of the trials showed a significant decrease in upper GI ulcers with an NNT of 11, as well as significant decreases in both components, gastric ulcers and duodenal ulcers. Multivariable analysis revealed a significant RR reduction in ulcers of ∼55% with HZT-501 after adjustment for risk factors that may influence the development of ulcers, such as age, ulcer history, aspirin use, and baseline erosions. Furthermore, the treatment effect of HZT-501 was not significantly different in subgroups of patients with or without these characteristics. Of interest, there was no decrement in treatment effect noted among patients taking low-dose aspirin.

nt of ulcers, such as age, ulcer history, aspirin use, and baseline erosions. Furthermore, the treatment effect of HZT-501 was not significantly different in subgroups of patients with or without these characteristics. Of interest, there was no decrement in treatment effect noted among patients taking low-dose aspirin. Ibuprofen was chosen as the NSAID studied because it is the most widely used non-aspirin NSAID in the United States (2). Although famotidine is generally given twice daily, ibuprofen at full doses for arthritis is typically given thrice daily. The double-dose famotidine was therefore given as three daily doses of 26.6 mg in the combination tablet. A preliminary study suggested that 26.6 mg thrice daily produces gastric acid suppression similar to 40 mg twice daily after 1 day (16). Also, oral administration of HZT-501 is bioequivalent to concurrent oral administration of equivalent separate doses of ibuprofen and famotidine (17). Previous trials have suggested that H2RAs may decrease upper GI symptoms in patients taking NSAIDs (18,19). We did not find a significant decrease in the proportion of patients spontaneously reporting a composite of symptoms consistent with dyspepsia. Future studies should formally assess NSAID users with a patient-reported outcome instrument validated for NSAID-associated dyspepsia.

mptoms in patients taking NSAIDs (18,19). We did not find a significant decrease in the proportion of patients spontaneously reporting a composite of symptoms consistent with dyspepsia. Future studies should formally assess NSAID users with a patient-reported outcome instrument validated for NSAID-associated dyspepsia. Upper GI complications such as bleeding are more important clinically than ulcers identified at a scheduled endoscopy, although results of endoscopic ulcer trials in NSAID users generally correlate with results in outcome trials assessing upper GI complications (7,20). Nevertheless, our study was insufficient in size and duration to reasonably assess complications, and the results do not demonstrate a benefit in complications. No patients had overt bleeding, obstruction, or perforation. Two patients in the HZT-501 arms (and none in the ibuprofen arms) had asymptomatic hemoglobin decreases >2.0 g/dl, the level typically used in outcome trials as a clinically relevant hemoglobin drop (21,22), and a third had a 1.6 g/dl decrease. Chan et al. (22) recently showed that such hemoglobin drops are much more common than overt complications in patients taking an NSAID plus antisecretory therapy, and that a majority are probably of small intestinal origin. Our patients with hemoglobin drops had gastric erosions noted, although no examination was performed to assess potential small intestinal injury.

lobin drops are much more common than overt complications in patients taking an NSAID plus antisecretory therapy, and that a majority are probably of small intestinal origin. Our patients with hemoglobin drops had gastric erosions noted, although no examination was performed to assess potential small intestinal injury. Very few patients were on antiplatelet agents other than low-dose aspirin so our study is unable to assess any potential interactions of famotidine with clopidogrel. Famotidine, like ranitidine and unlike cimetidine, shows little or no interaction with the hepatic cytochrome P450 system (23). In summary, the combined results of two studies indicate that double-dose famotidine plus ibuprofen, given as a combination tablet, decreases endoscopic upper GI ulcers as compared with ibuprofen alone. Use of a combination tablet theoretically should improve adherence to antisecretory therapy as compared with use of separate individual NSAID and antisecretory agents, but future studies will be necessary to prospectively study adherence and the effect on endoscopic or clinical outcomes. STUDY HIGHLIGHTS We would like to thank the patients, investigators, and study personnel who made these trials possible. We also thank Dr Weinblatt for his review of the manuscript. Guarantor of the article: Loren Laine, MD.

y authorities have developed guidelines and recommendations to indicate that patients with risk factors should receive preventive treatment including, among others, the co-prescription of gastroprotective agents (GPAs), namely a standard dose of proton pump inhibitors (PPIs), misoprostol, or high-dose famotidine (3,4). Two major challenges to reduce serious GI complications in at-risk patients are the low prescription rates of preventive therapy and poor patient adherence to prescribed GPAs. Several studies have reported a lack of correspondence between patterns of NSAID and GPA prescription (5,6,7,8), as well as reduced levels of patient adherence to prescribed GPAs, with reported rates of non-adherence ranging from 9 to 71% (9,10,11). Adherence to GPAs below the optimum level (which is defined as taking GPAs for ≥80% of the prescribed days) has been associated with a 2.5- to 4-fold increase in the risk of upper GI bleeding in patients receiving NSAIDs (9,10,11).

In summary, the combined results of two studies indicate that double-dose famotidine plus ibuprofen, given as a combination tablet, decreases endoscopic upper GI ulcers as compared with ibuprofen alone. Use of a combination tablet theoretically should improve adherence to antisecretory therapy as compared with use of separate individual NSAID and antisecretory agents, but future studies will be necessary to prospectively study adherence and the effect on endoscopic or clinical outcomes. STUDY HIGHLIGHTS We would like to thank the patients, investigators, and study personnel who made these trials possible. We also thank Dr Weinblatt for his review of the manuscript. Guarantor of the article: Loren Laine, MD. Specific author contributions: Planning and conducting study, analysis/interpretation of data, drafting and revision of the manuscript, approved the final draft submitted: Loren Laine; planning and conducting study, collecting data, interpretation of data, critical review of the manuscript, approved the final draft submitted: Alan J. Kivitz, Alfonso E. Bello, and Michael H. Schiff; analysis/interpretation of data, critical review of the manuscript, approved the final draft submitted: Amy Y. Grahn; analysis/interpretation of data; drafting and critical review of the manuscript, approved the final draft submitted: Ali S. Taha.

raft submitted: Alan J. Kivitz, Alfonso E. Bello, and Michael H. Schiff; analysis/interpretation of data, critical review of the manuscript, approved the final draft submitted: Amy Y. Grahn; analysis/interpretation of data; drafting and critical review of the manuscript, approved the final draft submitted: Ali S. Taha. Financial support: This study received funding and drugs from Horizon Pharma. The study was designed in collaboration between the sponsors and authors; the sponsor was responsible for collecting data; statistical analyses were performed by the sponsor under the direction of the sponsor and the authors; interpretation of the data and writing of the manuscript were performed by the authors. Potential competing interests: Loren Laine is a consultant for Horizon Pharma, AstraZeneca, Eisai, and Logical Therapeutics; is on the Data Safety Monitoring Board for Merck, Bayer, and BMS. Alfonso E. Bello is a consultant for Horizon Pharma and Pfizer; is a speaker for Abbott and Amgen. Amy Y. Grahn is an employee of Horizon Pharma. Michael H. Schiff is a consultant for Horizon Pharma. Ali S. Taha is a consultant for Horizon Pharma and Astellas Pharma. Figure 1 Trial flow and patient disposition.

Potential competing interests: Loren Laine is a consultant for Horizon Pharma, AstraZeneca, Eisai, and Logical Therapeutics; is on the Data Safety Monitoring Board for Merck, Bayer, and BMS. Alfonso E. Bello is a consultant for Horizon Pharma and Pfizer; is a speaker for Abbott and Amgen. Amy Y. Grahn is an employee of Horizon Pharma. Michael H. Schiff is a consultant for Horizon Pharma. Ali S. Taha is a consultant for Horizon Pharma and Astellas Pharma. Figure 1 Trial flow and patient disposition. Figure 2 The cumulative incidence (crude rate (%)) of ulcers at 8, 16, and 24 weeks in patients taking HZT-501 or ibuprofen (IBU) for the initial and post-adjudication analysis of the pooled data for the primary population of REDUCE-1 and REDUCE-2 trials. GI, gastrointestinal; REDUCE, Registration Endoscopic Studies to Determine Ulcer Formation of HZT-501 Compared with Ibuprofen: Efficacy and Safety Studies. Figure 3 Forest plots of the relative risks (95% confidence interval (CI)) for upper gastrointestinal ulcers for HZT-501 vs. ibuprofen (IBU) in subgroup analyses of the pooled data for the primary population of REDUCE-1 and REDUCE-2 trials. The sample sizes for each of the subgroups are shown at the right side. REDUCE, Registration Endoscopic Studies to Determine Ulcer Formation of HZT-501 Compared with Ibuprofen: Efficacy and Safety Studies.

buprofen (IBU) in subgroup analyses of the pooled data for the primary population of REDUCE-1 and REDUCE-2 trials. The sample sizes for each of the subgroups are shown at the right side. REDUCE, Registration Endoscopic Studies to Determine Ulcer Formation of HZT-501 Compared with Ibuprofen: Efficacy and Safety Studies. Table 1 Baseline characteristics of patients randomly assigned to HZT-501 or ibuprofen Characteristic Group REDUCE-1 trial REDUCE-2 trial Pooled data HZT-501 (N=550) Ibuprofen (N=262) HZT-501 (N=380) Ibuprofen (N=190) HZT-501 (N=930) Ibuprofen (N=452) Age (years) Median 55.0 55.0 54.0 54.0 55.0 55.0 Range 40–80 40–78 39–79 40–78 39–80 40–78 Age class % (n) <65 years 81.5% (448) 82.1% (215) 82.6% (314) 82.1% (156) 81.6% (762) 82.1% (371) ≥65 years 18.5% (102) 17.9% (47) 17.4% (66) 17.9% (34) 18.1% (168) 17.9% (81) Gender % (n) Male 32.0% (176) 31.3% (82) 34.2% (130) 28.4% (54) 32.9% (306) 30.1% (136) Female 68.0% (374) 68.7% (180) 65.8% (250) 71.6% (136) 67.1% (624) 69.9% (316) Race % (n) White 77.1% (424) 77.5% (203) 81.6% (310) 84.7% (161) 78.9% (734) 80.5% (364) Black 19.5% (107) 20.2% (53) 15.0% (57) 10.5% (20) 17.6% (164) 16.2% (73) Other 3.5% (19) 2.3% (6) 3.4% (13) 4.7% (9) 3.4% (32) 3.3% (15) Potential risk factors % (n) Prior ulcer history 7.6% (42) 5.7% (15) 4.7% (18) 5.8% (11) 6.5% (60) 5.8% (26) Use of low-dose aspirin and/or anticoagulants 17.1% (94) 13.4% (35) 15.5% (59) 13.2% (25) 16.5% (153) 13.3% (60) Use of low-dose aspirin 16.7% (92) 12.6% (33) 15.0% (57) 13.2% (25) 16.0% (149) 12.8% (58) Use of corticosteroids 0 0 0.3% (1) 0 0.11% (1) 0 Erosions at baseline 23.8% (131) 17.2% (45) 21.3% (81) 19.5% (37) 22.8% (212) 18.1% (82) REDUCE, Registration Endoscopic Studies to Determine Ulcer Formation of HZT-501 Compared with Ibuprofen: Efficacy and Safety Studies.

3) 15.0% (57) 13.2% (25) 16.0% (149) 12.8% (58) Use of corticosteroids 0 0 0.3% (1) 0 0.11% (1) 0 Erosions at baseline 23.8% (131) 17.2% (45) 21.3% (81) 19.5% (37) 22.8% (212) 18.1% (82) REDUCE, Registration Endoscopic Studies to Determine Ulcer Formation of HZT-501 Compared with Ibuprofen: Efficacy and Safety Studies. Table 2 The proportions of patients (life table and crude rate) in the primary population developing an upper GI, gastric, or duodenal ulcer in the post-adjudication analysis (excluding esophageal or no ulcers previously labeled as gastric ulcers) over 24 weeks of treatment REDUCE-1 trial REDUCE-2 trial Pooled results HZT-501 N =550 Ibuprofen N =262 HZT-501 N =380 Ibuprofen N =190 HZT-501 N =930 Ibuprofen N =452 Upper gastrointestinal (gastric or duodenal) ulcers Life table % 14.5%* 26.9% 13.0% 20.5% 13.9%** 24.3% Crude rate % (n) 11.1% (61)*** 21.8% (57) 9.7% (37)* 17.9% (34) 10.5% (98)*** 20.1% (91) Absolute risk reduction (95% CI) 10.7% (5.0, 16.3%) 8.2% (1.9, 14.4%) 9.6% (5.4, 13.8%) Gastric ulcers Life table % 12.7%* 22.9% 12.2% 17.5% 12.5%** 20.7% Crude rate % (n) 9.8% (54)** 18.3% (48) 8.9% (34)* 15.8% (30) 9.5% (88)*** 17.3% (78) Absolute risk reduction (95% CI) 8.5% (3.2, 13.8%) 6.8% (0.9, 12.8%) 7.8% (3.8, 11.8%) Duodenal ulcers Life table % 2.1%* 7.1% 0.9% 6.6% 1.6%* 6.9% Crude rate % (n) 1.3% (7)* 5.3% (14) 0.8% (3)* 4.7% (9) 1.1% (10)*** 5.1% (23) Absolute risk reduction (95% CI) 4.1 (1.2, 7.0%) 3.9% (0.8, 7.1%) 4.0% (1.9, 6.1%) CI, confidence interval; GI, gastrointestinal; REDUCE, Registration Endoscopic Studies to Determine Ulcer Formation of HZT-501 Compared with Ibuprofen: Efficacy and Safety Studies.

o prescribed GPAs, with reported rates of non-adherence ranging from 9 to 71% (9,10,11). Adherence to GPAs below the optimum level (which is defined as taking GPAs for ≥80% of the prescribed days) has been associated with a 2.5- to 4-fold increase in the risk of upper GI bleeding in patients receiving NSAIDs (9,10,11). Although these studies have consistently demonstrated reduced levels of adherence to GPA therapy among NSAID users, several issues remain unresolved. One such issue is that these studies have basically evaluated the adherence to PPI therapy of NSAID users with varying GI risk levels; however, the actual pattern of NSAID prescription in this population and whether patients exhibit differences in adherence to either NSAIDs or GPAs is unclear. Another important concern is the reasons for low adherence. Many patients take multiple drugs, which may be a factor in cases of poor adherence (12), and patients may elect to take these drugs only if they have symptoms. For example, patients may take NSAIDs if they have musculoskeletal pain or a PPI if they have dyspepsia, although the occurrence of these respective symptoms may not be simultaneous. It must be noted that up to 60% of patients developing upper GI complications have no previous abdominal symptoms (13). Based on these considerations, the primary objectives of this study were to determine the levels of adherence to prescribed GPAs and NSAIDs in at-risk patients. Secondary objectives were to describe the type of prescription, and to investigate factors associated with adherence.

have no previous abdominal symptoms (13). Based on these considerations, the primary objectives of this study were to determine the levels of adherence to prescribed GPAs and NSAIDs in at-risk patients. Secondary objectives were to describe the type of prescription, and to investigate factors associated with adherence. METHODS Settings This was a multicenter, observational, longitudinal study with prospective data collection. The study was conducted between 15 May 2008 and 16 January 2009. There were a total of 296 doctors involved in recruiting patients in 158 different outpatient clinics (mostly rheumatology, traumatology/orthopedic or internal medicine) distributed throughout Spain. Patients Inclusion criteria were (i) patients attending outpatient clinics with a musculoskeletal condition and an indication for NSAID prescription; (ii) age ≥18 years; (iii) presence of at least one GI risk factor of those described below, and (iv) receipt of prescriptions for both an NSAID and a GPA for a minimum of 15 days. The only exclusion criterion was treatment with a GPA for reasons other than the prevention of NSAID-related complications (e.g., gastroesophageal reflux disease). All included patients signed an informed consent form agreeing to participate in the study.

scriptions for both an NSAID and a GPA for a minimum of 15 days. The only exclusion criterion was treatment with a GPA for reasons other than the prevention of NSAID-related complications (e.g., gastroesophageal reflux disease). All included patients signed an informed consent form agreeing to participate in the study. GI risk factors (3,4,5,6,7,8) for this study were (i) age ≥60 years; (ii) a history of peptic ulcer, ulcer complications, or dyspepsia (a marker for increased risk of peptic ulcer, especially in populations with high H. pylori infection rates (14,15,16,17)); (iii) the use of aspirin, corticosteroids, or anticoagulants in addition to a prescribed NSAID; (iv) the use of a high-dose NSAID or the use of two NSAIDs. High-dose NSAID, which has been previously defined elsewhere (18,19), included treatment with any NSAID at the maximum dose recommended for the symptomatic treatment of arthritis pain (e.g., diclofenac ≥150 mg/day, aceclofenac ≥100 mg/day, meloxicam ≥15 mg/day, naproxen ≥1,000 mg/day, piroxicam ≥20 mg/day, and ibuprofen >1,800 mg/day). The doses of PPI for gastroprotection were as follows: omeprazole 20 mg/day, lansoprazole 30 mg/day, pantoprazole 20 mg/day, and esomeprazole 20 mg/day. Among the H2 receptor antagonists, the doses were 40 mg/12 h for famotidine. The appropriate doses for misoprostol were 200 μg/6–8 h.

ay, and ibuprofen >1,800 mg/day). The doses of PPI for gastroprotection were as follows: omeprazole 20 mg/day, lansoprazole 30 mg/day, pantoprazole 20 mg/day, and esomeprazole 20 mg/day. Among the H2 receptor antagonists, the doses were 40 mg/12 h for famotidine. The appropriate doses for misoprostol were 200 μg/6–8 h. Questionnaires and follow-up Investigators enrolled consecutive patients (with the above-mentioned inclusion criteria and no exclusion criteria) who agreed to participate in the study for at least 1 month. Investigators collected data in a closed and pre-printed questionnaire that included data concerning demographics (age and sex), GI risk factors, and current medication for pre-existing conditions, as well as doses, duration of use, time of use, and reason for prescription of NSAID plus GPA. Each questionnaire was anonymized, and patients were only identified by a number. Each questionnaire contained a telephone number provided by the patient where they could be reached for follow-up. Once completed, each questionnaire was faxed to the coordinating center and the principal investigator (AL) evaluated the consistency and completeness of the data provided and requested additional information or clarification, if needed. To be contacted for follow-up, patients signed an informed consent form. They were also informed that they would receive one or two telephone calls from independent researchers who would ask questions concerning their disease and the medication they take within an investigational project.

Questionnaires and follow-up Investigators enrolled consecutive patients (with the above-mentioned inclusion criteria and no exclusion criteria) who agreed to participate in the study for at least 1 month. Investigators collected data in a closed and pre-printed questionnaire that included data concerning demographics (age and sex), GI risk factors, and current medication for pre-existing conditions, as well as doses, duration of use, time of use, and reason for prescription of NSAID plus GPA. Each questionnaire was anonymized, and patients were only identified by a number. Each questionnaire contained a telephone number provided by the patient where they could be reached for follow-up. Once completed, each questionnaire was faxed to the coordinating center and the principal investigator (AL) evaluated the consistency and completeness of the data provided and requested additional information or clarification, if needed. To be contacted for follow-up, patients signed an informed consent form. They were also informed that they would receive one or two telephone calls from independent researchers who would ask questions concerning their disease and the medication they take within an investigational project. Patients were followed up with telephone calls at a maximum of two different times. The first contact was an early call within 15–18 days after the medical visit. If the prescription of the NSAID plus GPA was for 30−60 days or longer, then the patients received a second call within a window of 60±7 days. Two independent and trained investigators (MPT and PR) carried out the calls and completed a structured questionnaire that was originally validated in a small group of patients to assess the feasibility of the questions. The questions focused on adherence to NSAID plus GPA therapy and evaluated levels of adherence and reasons for not taking the pills. In general, the call lasted ∼10 min and patients were asked to provide the number of prescriptions obtained and the number of pills that remained in the package or to be refilled at the end of the interview. The study flow is summarized in Figure 1.

nd evaluated levels of adherence and reasons for not taking the pills. In general, the call lasted ∼10 min and patients were asked to provide the number of prescriptions obtained and the number of pills that remained in the package or to be refilled at the end of the interview. The study flow is summarized in Figure 1. Statistical analysis Descriptive analysis of the patients included demographic and clinical characteristics, pharmacological treatments, and frequencies of the main variables of the study (rates of adherence, factors associated with adherence and type of prescription). Quantitative variables were analyzed using measurements of central tendency (mean and median) and dispersion (95% confidence intervals (CIs), standard deviation, quartiles, and ranges). Qualitative variables were defined according to their absolute and relative frequencies. Student's t-test was used to analyze quantitative variables. Categorical variables were analyzed using the χ2 test or the Fisher's exact test. Tests were two-tailed with a significance level of 5%. Multivariate analyses were used to determine risk factors of poor adherence to either NSAID or GPA therapy. Optimal adherence was defined as taking GPAs for ≥80% of the prescribed days. Models of logistic regression were constructed based on variables of interest (age, gender, ulcer history, concomitant medications (including aspirin, corticosteroids, and anticoagulants) history of dyspepsia, dose of NSAIDs, use of two NSAIDs, duration of treatment, dose timing, number of pills and reasons for not taking medication) to provide adjusted odds ratios for each factor. A backward selection method was used and those variables with a significance level of >0.2 were excluded from the model.

s) history of dyspepsia, dose of NSAIDs, use of two NSAIDs, duration of treatment, dose timing, number of pills and reasons for not taking medication) to provide adjusted odds ratios for each factor. A backward selection method was used and those variables with a significance level of >0.2 were excluded from the model. Data were analyzed with SAS 8.2 statistical software (SAS Institute, Cary, NC). A sample size of 1,200 patients would provide an error <3% for 50% levels of adherence. An initial analysis of the data on GPA adherence showed an unexplainably high proportion of patients who did not provide a reason for not starting PPI therapy. This led us to do a manual post hoc review of the original data collected during the telephone call. We found that this proportion was lower, since many of those patients had already taken the medication (which was prescribed for 2 weeks) at the time of the call, and were incorrectly introduced into the database as patients who were not taking the medication rather than patients who had finished the prescribed treatment. This made us revise the whole database and the questionnaire to reconfirm that it was the only problem with data entry and the interpretation of questions.

ll, and were incorrectly introduced into the database as patients who were not taking the medication rather than patients who had finished the prescribed treatment. This made us revise the whole database and the questionnaire to reconfirm that it was the only problem with data entry and the interpretation of questions. Ethical considerations This study complied with all ethical considerations involving human subjects, as adopted by the 18th World Medical Assembly, Helsinki, Finland. All recorded information was obtained following the standard clinical guidelines, and patients were not subjected to any therapeutic or diagnostic experimentation. The study followed standard security and confidentiality measures, complying fully with Spanish legislation regarding data protection (Ley Orgánica de 15/99). The Regional Ethics Committee for Clinical Research, Hospital San Carlos (Madrid) approved this study. The patients' names remained confidential; identification numbers were used instead. RESULTS Demographics A total of 296 specialists participated in the study and 1,232 patients agreed to participate, of whom 192 were excluded due to incomplete data in the original questionnaire completed by the investigator (n=9), duration of treatment <15 days (n=34), or telephone interview not carried out within the pre-specified time window (n=149). Therefore, 1,040 patients were included in the final analysis. Table 1 presents patients' clinical characteristics.

complete data in the original questionnaire completed by the investigator (n=9), duration of treatment <15 days (n=34), or telephone interview not carried out within the pre-specified time window (n=149). Therefore, 1,040 patients were included in the final analysis. Table 1 presents patients' clinical characteristics. Type of prescription The most common type of anti-inflammatory drug prescribed was traditional NSAIDs (862/1,040, 82.9%); COX-2 selective inhibitors represented 13.1% (136) of all prescriptions. Among traditional NSAIDs, in 682 (79.1%) cases the prescription was below the recommended doses. The standard dose was prescribed in 157 (18.2%) cases. In 568 (65.9%) cases, duration of treatment was short term (<30 days). Among prescriptions for COX-2 selective inhibitors, 31.6% were below recommended doses, 66.2% were for the standard dose, and 61.0% were short-term prescriptions. The majority of prescriptions for traditional NSAIDs (81.7%) were either b.i.d. or t.i.d.; by contrast, 89.7% of COX-2 selective inhibitor prescriptions were to be taken once daily.

COX-2 selective inhibitors, 31.6% were below recommended doses, 66.2% were for the standard dose, and 61.0% were short-term prescriptions. The majority of prescriptions for traditional NSAIDs (81.7%) were either b.i.d. or t.i.d.; by contrast, 89.7% of COX-2 selective inhibitor prescriptions were to be taken once daily. Among patients who were prescribed a GPA, 1,028 (99.4%) were also prescribed a PPI, while the remaining patients were prescribed H2 receptor antagonists (0.39%) or misoprostol (0.19%). The doses of PPI prescribed were the approved dose for the indication of GI prevention of NSAID damage in 70.6% of cases, while 28% of patients were prescribed higher doses. PPI prescription was once daily in 95.6% of cases. Prescribed PPI treatment was short term (<30 days) in 63.1% of cases; 512 (49.2%) patients were additionally taking other types of medication (different from NSAIDs or GPAs) for different reasons. Adherence to treatment The telephone interview was conducted with the patient in 92.4% of cases. In the remaining 7.6% of cases, a family member of the patient was also involved in the interview.

Among patients who were prescribed a GPA, 1,028 (99.4%) were also prescribed a PPI, while the remaining patients were prescribed H2 receptor antagonists (0.39%) or misoprostol (0.19%). The doses of PPI prescribed were the approved dose for the indication of GI prevention of NSAID damage in 70.6% of cases, while 28% of patients were prescribed higher doses. PPI prescription was once daily in 95.6% of cases. Prescribed PPI treatment was short term (<30 days) in 63.1% of cases; 512 (49.2%) patients were additionally taking other types of medication (different from NSAIDs or GPAs) for different reasons. Adherence to treatment The telephone interview was conducted with the patient in 92.4% of cases. In the remaining 7.6% of cases, a family member of the patient was also involved in the interview. NSAIDs In 92.5% (962/1,040) of cases, the patient reported starting NSAID treatment. Among 77 patients (1 case with data missing) who did not start the prescribed NSAID, the main reasons for not initiating treatment were quite diverse: not properly understanding the doctor's instructions was the most frequently cited reason (24 [35.8%]), followed by infrequent/low-intensity pain (16 [23.9%]), fear of adverse events (11 [16.4%]) and taking medications other than those prescribed for pain (13 [19.4%]).

s for not initiating treatment were quite diverse: not properly understanding the doctor's instructions was the most frequently cited reason (24 [35.8%]), followed by infrequent/low-intensity pain (16 [23.9%]), fear of adverse events (11 [16.4%]) and taking medications other than those prescribed for pain (13 [19.4%]). Of the patients who did initiate therapy, 233 (24.2%) failed to take the prescribed NSAID at some point for a mean of 12.2±19.1 days. Table 2 summarizes the reasons given by patients for not taking the prescribed NSAID at some point during treatment which were infrequent/low-intensity pain (29.4%), development of adverse events (15.7%), that the prescribed NSAID was ineffective (9.3%), forgetfulness (12.5%), and not getting a second prescription (31.4%). The majority of patients (79.7% 95% CI: 76.9−82.2%) exhibited optimal adherence, taking the prescribed NSAID for 80% or more of the days prescribed. GPAs In 85.9% (893/1,040) of cases, the patient reported starting GPA therapy. Reasons for not initiating GPA treatment (146 patients, 1 case with data missing) were infrequent/low-intensity pain (64 [43.8%]), fear of adverse events (24 [16.4%]), taking different analgesics (19 [13.0%]), not understanding the doctor's instructions (15 [10.3%]), and taking too many/unnecessary pills (5 [3.4%]). Fifteen (10.3%) did not provide an answer, and three patients (2.1%) gave other reasons.

infrequent/low-intensity pain (64 [43.8%]), fear of adverse events (24 [16.4%]), taking different analgesics (19 [13.0%]), not understanding the doctor's instructions (15 [10.3%]), and taking too many/unnecessary pills (5 [3.4%]). Fifteen (10.3%) did not provide an answer, and three patients (2.1%) gave other reasons. Of the patients who initiated GPA therapy, 48 (5.4%) failed to take the drug at some point for a mean of 8.5±25.9 days. Table 2 summarizes the reasons given by patients for not taking the prescribed GPA at some point during treatment which were forgetfulness (47.4%) and the absence of either rheumatic or abdominal symptoms (39.5%). Overall, 84.1% took the drug for 80% or more of the days prescribed. Of patients who reported initiating NSAID therapy, 9.3% did not take concomitant GPA therapy at any point. In 11% of cases (95% CI: 9.0−13.2%), GPA therapy either was not initiated, or was taken <80% of the time. Only eight patients had short-term prescription of GPA (<30 days) together with longer (>30 days) prescription of NSAIDs.

o reported initiating NSAID therapy, 9.3% did not take concomitant GPA therapy at any point. In 11% of cases (95% CI: 9.0−13.2%), GPA therapy either was not initiated, or was taken <80% of the time. Only eight patients had short-term prescription of GPA (<30 days) together with longer (>30 days) prescription of NSAIDs. To assess the concordance between patients' reported behavior and actual behavior, participants were asked to count the pills remaining from the last prescription at the end of the interview. Concerning GPA prescriptions, in 338/1,040 (32.5%) cases the pill count was not performed because the patient had not started therapy, or was unable to perform the count for other reasons. Among patients who counted the pills, there was agreement in 94.7% of cases between the patients' self-report and the actual count. Regarding NSAID prescriptions, 55% of patients did not perform the pill count; among those who did, agreement was present in 91% of cases. For both drug types, the highest disagreement occurred in patients who reported adherence between 20 and 80% (GPAs: 27/38, 71% NSAIDs: 8/30, 26.7%).

elf-report and the actual count. Regarding NSAID prescriptions, 55% of patients did not perform the pill count; among those who did, agreement was present in 91% of cases. For both drug types, the highest disagreement occurred in patients who reported adherence between 20 and 80% (GPAs: 27/38, 71% NSAIDs: 8/30, 26.7%). Adverse events The frequency of adverse events was higher in patients who reported not optimal adherence to either GPA or NSAID prescriptions; 22.1% (35/158) of patients with low adherence (<80%) to GPA had an adverse event, compared with 1.9% (16/838) of patients who were optimally adherent (P<0.0001). Similarly, 17.0% (32/188) of patients who were not optimally adherent to NSAIDs had an adverse event, compared with 1.6% (12/737) of optimally adherent patients (P<0.0001). Adverse events were GI (30 dyspepsia, 3 diarrhea, and 1 bleeding event) in most cases (34/49). The remaining events were non-GI, including cardiovascular/renal (hypertension, edema (n=4), allergic reactions (n=2), headache (n=1), and unspecified (n=13)).

red with 1.6% (12/737) of optimally adherent patients (P<0.0001). Adverse events were GI (30 dyspepsia, 3 diarrhea, and 1 bleeding event) in most cases (34/49). The remaining events were non-GI, including cardiovascular/renal (hypertension, edema (n=4), allergic reactions (n=2), headache (n=1), and unspecified (n=13)). Multivariate analysis for adherence Univariate analysis revealed that of all clinical variables considered, concomitant use of a non-aspirin antiplatelet drug, overall use of any antiplatelet drug, high-dose GPA (PPI), short-term GPA treatment, and the presence of adverse events were associated with poor patient adherence to GPA prescription, while history of uncomplicated peptic ulcer disease, use of a non-aspirin antiplatelet drug, dosing regimen for NSAID treatment, short-term NSAID treatment, and the presence of adverse events were associated with poor patient adherence to NSAID prescription. Logistic regression analysis demonstrated that of all variables considered in the models, short-term prescription of GPA therapy and the presence of adverse events were independent determinants for poor adherence to the prescribed GPA. The presence of uncomplicated ulcer history, short-term NSAID prescription, frequent NSAID dosing, and the presence of adverse events were associated with poor adherence to the prescribed NSAID ( Tables 3 and 4). The major determinants of poor adherence to either NSAID or GPA prescription were the development of adverse events, which in most of cases were GI adverse events (dyspepsia being the most common).

dosing, and the presence of adverse events were associated with poor adherence to the prescribed NSAID ( Tables 3 and 4). The major determinants of poor adherence to either NSAID or GPA prescription were the development of adverse events, which in most of cases were GI adverse events (dyspepsia being the most common). We conducted additional analysis with (i) age as a risk factor when >65 years, instead of 60 years; (ii) the presence of very high GI risk, as defined by the guidelines of the American College of Gastroenterology (3), and (iii) excluding history of dyspepsia as a risk factor. Neither the presence/absence of very high GI risk, nor the presence/absence of dyspepsia showed statistically significant differences in either NSAID or GPA adherence; however, patients with very high GI risk showed a trend for greater adherence to GPAs and less adherence to NSAIDs. When compared with patients <65 years old, those >65 years of age were associated with a trend (P=0.07) toward higher levels of adherence to GPA, but not NSAID therapy (similarly to the results obtained by using 60 years of age as the cutoff point). When this variable (≥65 vs. <65 years of age) was included in the logistic regression model, it was not independently associated with poor adherence to GPA (data not shown). In addition to short-term prescription of GPA therapy and the presence of adverse events, being <65 years was independently associated with poor GPA adherence (odds ratio: 2.2; 95% CI: 1.3−3.9) if the analysis was restricted to patients who reported initiating the prescribed NSAID therapy.

ence to GPA (data not shown). In addition to short-term prescription of GPA therapy and the presence of adverse events, being <65 years was independently associated with poor GPA adherence (odds ratio: 2.2; 95% CI: 1.3−3.9) if the analysis was restricted to patients who reported initiating the prescribed NSAID therapy. DISCUSSION This study focuses on patients' adherence to both NSAIDs and GPAs. Previous studies showed low rates of GPA prescription to at-risk patients receiving NSAIDs, and low levels of adherence to GPA prescriptions among those who did receive co-therapy (9), which was in turn associated with the increased risk of GI complications (10,11). However, these studies did not investigate the reasons for poor adherence or the determinants of poor adherence with regard to GPAs or NSAIDs. We believe that these aspects are of paramount importance because they are probably linked. Here, we have investigated these features by examining adherence to both GPA and NSAID therapies using a different approach, which is based on the direct questioning of patients concerning their reasons for not taking the prescribed medication. Consequently, we were able to discriminate and report on two aspects of the same spectrum: (i) failure to initiate the prescribed treatment and (ii) lack of adherence to the prescribed drugs.

pproach, which is based on the direct questioning of patients concerning their reasons for not taking the prescribed medication. Consequently, we were able to discriminate and report on two aspects of the same spectrum: (i) failure to initiate the prescribed treatment and (ii) lack of adherence to the prescribed drugs. Rates of adherence to both therapies were high; however, contrary to what may be expected the proportion of patients who did not initiate the prescribed GPA therapy was higher than the proportion that did not initiate NSAID therapy. This pattern may be due to the fact that patients who start NSAID therapy because they seek rheumatic pain relief do not necessarily experience GI symptoms, and some patients may not be aware of the increased GI risk associated with NSAID use. Interestingly, among patients who did not initiate NSAID therapy, failure to properly understand their doctor's instructions was most often cited as the primary reason, suggesting that this aspect should be taken into consideration during the prescription process. On the contrary, among those who did not initiate GPA therapy, most patients did not do so because they had no (or mild) GI symptoms. Among patients who actually started therapy, a high proportion reported optimal drug adherence (defined as taking the prescribed drug >80% of the days prescribed) for both NSAIDs and GPAs; this proportion was actually higher than reported in other studies (9,10,11), but is in agreement with the increasing trend of concomitantly prescribing a GPA to NSAID users (20). Our different methodological approach may explain the findings of higher adherence rates.

of the days prescribed) for both NSAIDs and GPAs; this proportion was actually higher than reported in other studies (9,10,11), but is in agreement with the increasing trend of concomitantly prescribing a GPA to NSAID users (20). Our different methodological approach may explain the findings of higher adherence rates. Patients may falsely report high compliance levels because they have a false perception of compliance. We have tried to evaluate this possibility by asking patients to report on the number of NSAID and GPA pills remaining at the time of the follow-up interview. We could not obtain that information from all patients; however, among those who could actually count the pills, we found a high degree of agreement between the reported adherence and the number of pills taken from the prescribed boxes for both GPAs and NSAIDs.

PA pills remaining at the time of the follow-up interview. We could not obtain that information from all patients; however, among those who could actually count the pills, we found a high degree of agreement between the reported adherence and the number of pills taken from the prescribed boxes for both GPAs and NSAIDs. Among patients reporting poor adherence to medication, there was a higher level of adherence to GPA than to NSAID therapy; the reasons for poor adherence were different for the two drug types. Although the main reasons given for stopping NSAID therapy were infrequent/low-intensity rheumatic pain or developing adverse events, the main reason for stopping GPA therapy was forgetfulness, followed by the absence of either rheumatic or abdominal symptoms. If we consider that some of the reasons given (e.g., not getting a second prescription) may also reflect infrequent/low-intensity rheumatic pain, then this was the most often-cited reason for non-adherence to prescribed NSAIDs. These findings reveal the primary underlying reasons driving drug use behavior in clinical practice, demonstrating that a substantial number of patients with chronic musculoskeletal conditions take their NSAID prescription irregularly depending on the level of pain. These results also document that the development of adverse events (especially GI-related adverse events) is another major factor affecting drug use. On the contrary, the main reason for not taking the GPA is probably linked to lack of GI symptoms in most cases.

prescription irregularly depending on the level of pain. These results also document that the development of adverse events (especially GI-related adverse events) is another major factor affecting drug use. On the contrary, the main reason for not taking the GPA is probably linked to lack of GI symptoms in most cases. The development of adverse events is a well-known characteristic of NSAID therapy. The design and size of this study did not allow us to detect GI complications or determine whether poor adherence was associated with this serious adverse event. However, we were able to evaluate other patient-reported minor adverse events, which often are not recorded in databases, but are suspected to be the main reasons for stopping NSAID use (2). It should be noted that dyspepsia was the most commonly reported adverse event, and that patients who were non-adherent to GPA therapy had a significantly higher risk of this type of adverse event.

se events, which often are not recorded in databases, but are suspected to be the main reasons for stopping NSAID use (2). It should be noted that dyspepsia was the most commonly reported adverse event, and that patients who were non-adherent to GPA therapy had a significantly higher risk of this type of adverse event. This study also investigated clinical determinants of poor adherence to either NSAID or GPA prescriptions. History of peptic ulcer disease and the presence of adverse events were predictors of poor adherence to NSAID prescriptions and seem related to the well-known GI risk associated with NSAIDs. Concomitant use of a non-aspirin antiplatelet agent was also associated with poor adherence to GPA prescriptions, and may be related to the current warning from regulatory agencies to take PPIs together with clopidogrel (21), although eventually this did not emerge as an independent factor. Frequent dosing (more than once daily) was a predictor of poor adherence to NSAID but not GPA prescriptions, which may be due to the fact that PPIs are taken once daily, while NSAIDs are taken several times per day. Short-term treatment (<30 days) was a predictor of poor adherence for both therapies. The reason for this is unclear, because short-term NSAID treatment was the most frequent prescription type in our study. It is possible that patients who received longer periods of therapy suffered from more severe musculoskeletal diseases and pain, which would increase adherence during the relatively short period of observation (15 and 60 days); however, this characteristic was not recorded in our study.

prescription type in our study. It is possible that patients who received longer periods of therapy suffered from more severe musculoskeletal diseases and pain, which would increase adherence during the relatively short period of observation (15 and 60 days); however, this characteristic was not recorded in our study. This study has also evaluated the prescription characteristics of both NSAIDs and GPAs in patients who are at risk for GI complications. NSAID prescriptions were usually short-term and at lower than-recommended doses, in clear contrast with the type of treatment and dosing prescribed for similar indications in randomized controlled trials (22,23). However, the doses of GPA prescribed were either standard or high. This prescription pattern may be guided by the perceived GI risk with NSAID treatment in an attempt to minimize adverse events, given that dose and duration are two factors linked to increased risk of upper GI complications (2).

ntrolled trials (22,23). However, the doses of GPA prescribed were either standard or high. This prescription pattern may be guided by the perceived GI risk with NSAID treatment in an attempt to minimize adverse events, given that dose and duration are two factors linked to increased risk of upper GI complications (2). Our study has several strengths and limitations. This study evaluated a real clinical sample, and direct contact with patients allowed us to take a different approach to evaluate the patient-reported reasons for non-adherence, a factor that studies based on the data extracted from database platforms cannot report. This study is limited by the lack of a direct, objective measure of prescription use. Instead, we had to rely on the patients' self-reporting, which may introduce recall bias. We have tried to limit the impact of recall bias by having patients report the number of pills remaining in their prescriptions. Among patients who were able to provide this information, we found a high level of agreement, which supports the validity of our study. In any case, it must also be recognized that having issued a prescription does not mean that patients will take the medication, an aspect that cannot be controlled in database studies. Another limitation is that our study reports mostly on short-term NSAID and GPA therapy, which we found to be the most frequent type of prescription in clinical practice. We did not evaluate the long-term use of these drugs, which might have provided different results. The fact that most patients received short-term prescriptions justified the early telephone call to interview patients about adherence, because a later call might have had a negative impact on the accuracy of our data. It is possible that the study design induced a selection bias for patients who were prescribed short-term and not long-term treatment.

ved short-term prescriptions justified the early telephone call to interview patients about adherence, because a later call might have had a negative impact on the accuracy of our data. It is possible that the study design induced a selection bias for patients who were prescribed short-term and not long-term treatment. In summary, this study investigated the type of prescription, rate of adherence, and reasons for non-adherence to NSAID and GPA therapy in patients at increased risk of developing GI-related adverse events. NSAIDs and GPAs were prescribed short term in most cases. More subjects initiated NSAID than GPA therapy. We report high levels of adherence to both NSAID and GPA therapies, which supports recent data suggesting an important time trend decrease in the rate of upper GI complications in our country (24). Still, there were more side effects among patients with non-optimal adherence to GPA. Not understanding the doctor's instructions regarding drug use, infrequent/mild-intensity pain, and forgetfulness were the most frequently cited reasons for non-adherence. Adverse events and short-term treatment were the main clinical predictors of poor adherence for both NSAIDs and GPAs. History of peptic ulcer and frequent dosing were additional factors for poor NSAID adherence. We believe that these findings are relevant to attempts to improve adherence to both GPA and NSAID prescriptions among at-risk patients. STUDY HIGHLIGHTS We are indebted to all participating investigators and patients involved in the study. Guarantor of the article: Angel Lanas, MD, DSc.

In summary, this study investigated the type of prescription, rate of adherence, and reasons for non-adherence to NSAID and GPA therapy in patients at increased risk of developing GI-related adverse events. NSAIDs and GPAs were prescribed short term in most cases. More subjects initiated NSAID than GPA therapy. We report high levels of adherence to both NSAID and GPA therapies, which supports recent data suggesting an important time trend decrease in the rate of upper GI complications in our country (24). Still, there were more side effects among patients with non-optimal adherence to GPA. Not understanding the doctor's instructions regarding drug use, infrequent/mild-intensity pain, and forgetfulness were the most frequently cited reasons for non-adherence. Adverse events and short-term treatment were the main clinical predictors of poor adherence for both NSAIDs and GPAs. History of peptic ulcer and frequent dosing were additional factors for poor NSAID adherence. We believe that these findings are relevant to attempts to improve adherence to both GPA and NSAID prescriptions among at-risk patients. STUDY HIGHLIGHTS We are indebted to all participating investigators and patients involved in the study. Guarantor of the article: Angel Lanas, MD, DSc. Specific author contributions: Designed the study: Angel Lanas and Javier Zapardiel; drafted the first version of the manuscript: Angel Lanas; made the phone calls: Mónica Polo-Tomás and Pilar Roncalés; conducted the statistical analysis: Miguel Angel Gonzalez; all authors interpreted the results and contributed to the subsequent drafting and versions of the manuscript.

ngel Lanas and Javier Zapardiel; drafted the first version of the manuscript: Angel Lanas; made the phone calls: Mónica Polo-Tomás and Pilar Roncalés; conducted the statistical analysis: Miguel Angel Gonzalez; all authors interpreted the results and contributed to the subsequent drafting and versions of the manuscript. Financial support: This study was supported by an unrestricted grant from AstraZeneca Spain, which had no role in the study design; the collection, analysis, or interpretation of the data; or in the writing of the report. Potential competing interests: Angel Lanas is an advisor to AstraZeneca, Pfizer, and Nicox. Miguel A. Gonzalez works for Quintiles, an external company engaged in statistical support for companies including AstraZeneca. Javier Zapardiel was an employee of AstraZeneca. Figure 1 Study flow. Investigators collected consecutive patients who met inclusion and exclusion criteria and who agreed to participate in the study. After data collection, the anonymized information was sent to the coordinating center. Patients were followed up with telephone calls at two different times and the follow-up information was added to the database. GI, gastrointestinal.

who met inclusion and exclusion criteria and who agreed to participate in the study. After data collection, the anonymized information was sent to the coordinating center. Patients were followed up with telephone calls at two different times and the follow-up information was added to the database. GI, gastrointestinal. Table 1 Demographics of patients and gastrointestinal risk factors Variable Mean (s.d.) 95% CI Age (years) 57.1+16.0 56.1, 58.1 Gender Female: 722 (69.4%) 66.5%, 72.2% Male: 318 (30.6%) 27.8%, 33.5% Main risk factors N (%) 1,040 (100%) 95% CI Age ≥60 years 522 (50.2%) 47.1%, 53.3% History of complicated peptic ulcer 34 (3.3%) 2.3%, 4.5% History of uncomplicated peptic ulcer 101 (9.7%) 8.0%, 11.7% History of dyspepsia 304 (29.2%) 26.5%, 32.1% Co-therapy with anticoagulants 47 (4.5%) 3.3%, 6.0% Co-therapy with aspirin 82 (7.9%) 6.3%, 9.7% Co-therapy with non-aspirin antiplatelet agents 32 (3.1%) 2.1%, 4.3% Treatment with two NSAIDs 77 (7.4%) 5.9%, 9.2% High NSAID dose 12 (1.2%) 0.6%, 2.1% Other 68 (6.5%) 5.1%, 8.2% CI, confidence interval; GI, gastrointestinal; NSAIDs, non-steroidal anti-inflammatory drugs. The guidelines of the American College of Gastroenterology consider age to be a risk factor when patients are >65 years of age. The corresponding figure is 366/1,040 (35.2%). Also, the proportion of patients with very high GI risk (patients with a history of complicated ulcer or >2 risk factors)=81/1,040 (7.8%). The total number of patients with one or more GI risk factors considering the age cutoff at 65 is 77.3%.

atients are >65 years of age. The corresponding figure is 366/1,040 (35.2%). Also, the proportion of patients with very high GI risk (patients with a history of complicated ulcer or >2 risk factors)=81/1,040 (7.8%). The total number of patients with one or more GI risk factors considering the age cutoff at 65 is 77.3%. Table 2 Main reasons reported by patients for not taking the prescribed NSAID or GPA at some point (any day) during treatment among those who started therapy Reason for lack of compliance Number (%) 95% CI NSAIDs N=233 (100%) Adverse event 32 (15.7%) 11.0%, 21.4% Not having enough rheumatic symptoms 60 (29.4%) 23.2%, 36.2% Taking a different drug, or the prescribed drug was not effective 19 (9.3%) 5.7%, 14.1% Forgetfulness 26 (12.2%) 8.5%, 18.1% Not getting a second prescription 64 (31.4%) 25.1%, 38.2% Other 3 (1.4%) 0.3%, 4.2% Missing 29 — GPAs N=48 (100%) Adverse event 3 (7.9%) 1.7%, 21.4% Infrequent/mild-intensity pain 15 (39.5%) 24.0%, 56.6% Taking a different drug, or the prescribed drug was not effective 1 (2.6%) 0.1%, 13.8% Forgetfulness 18 (47.4%) 31.0%, 64.2% Other 1 (2.1%) 0.1%, 13.8% Missing 10 — CI, confidence interval; GPA, gastroprotective agent; NSAIDs, non-steroidal anti-inflammatory drugs.

Infrequent/mild-intensity pain 15 (39.5%) 24.0%, 56.6% Taking a different drug, or the prescribed drug was not effective 1 (2.6%) 0.1%, 13.8% Forgetfulness 18 (47.4%) 31.0%, 64.2% Other 1 (2.1%) 0.1%, 13.8% Missing 10 — CI, confidence interval; GPA, gastroprotective agent; NSAIDs, non-steroidal anti-inflammatory drugs. Table 3 Factors associated with poor adherence to NSAID treatment Variablea   N (%) Crude odds ratio (95% CI) Adjusted odds ratio (95% CI)a History of uncomplicated peptic ulcer No 160 (85.1%) Yes 28 (14.9%) 1.8 (1.1, 2.9) 2.3 (1.4, 3.9) Non-aspirin antiplatelet treatment Yes 1 (0.5%) No 187 (99.5%) 6.8 (0.9, 50.7) — Antiplatelet treatment Yes 14 (7.5%) No 174 (92.5%) 1.6 (0.9, 2.8) — Number of pills/day Once daily 39 (20.7%) More than once daily 149 (79.3%) 1.6 (1.1, 2.3) 1.6 (1.1, 2.5) Length of prescription >4 weeks 32 (17.8%) ≤ 4 weeks 148 (82.2%) 2.4 (1.6, 3.6) 2.7 (1.7, 4.2) Adverse events No 156 (83.0%) Yes 32 (17.0%) 12.4 (6.2, 24.6) 14.9 (7.1, 31.2) CI, confidence interval; NSAIDs, non-steroidal anti-inflammatory drugs; OR, odds ratio. a Adjusted ORs in the final model. ORs of variables that were not statistically significant in the logistic regression model are not reported.

Table 3 Factors associated with poor adherence to NSAID treatment Variablea   N (%) Crude odds ratio (95% CI) Adjusted odds ratio (95% CI)a History of uncomplicated peptic ulcer No 160 (85.1%) Yes 28 (14.9%) 1.8 (1.1, 2.9) 2.3 (1.4, 3.9) Non-aspirin antiplatelet treatment Yes 1 (0.5%) No 187 (99.5%) 6.8 (0.9, 50.7) — Antiplatelet treatment Yes 14 (7.5%) No 174 (92.5%) 1.6 (0.9, 2.8) — Number of pills/day Once daily 39 (20.7%) More than once daily 149 (79.3%) 1.6 (1.1, 2.3) 1.6 (1.1, 2.5) Length of prescription >4 weeks 32 (17.8%) ≤ 4 weeks 148 (82.2%) 2.4 (1.6, 3.6) 2.7 (1.7, 4.2) Adverse events No 156 (83.0%) Yes 32 (17.0%) 12.4 (6.2, 24.6) 14.9 (7.1, 31.2) CI, confidence interval; NSAIDs, non-steroidal anti-inflammatory drugs; OR, odds ratio. a Adjusted ORs in the final model. ORs of variables that were not statistically significant in the logistic regression model are not reported. Table 4 Factors associated with poor adherence to GPA treatment Variablea Crude % N (%) Crude odds ratio (95% CI) Adjusted odds ratio (95% CI)a Age (years) ≥ 60 72 (45.6%)   — <60 86 (54.4%) 1.3 (0.9, 1.8) History of complicated peptic ulcer Yes 2 (1.3%)   — No 156 (98.7%) Non-aspirin antiplatelet treatment Yes 1 (0.6%) 2.9 (0.7, 12.2) — No 157 (99.4%) Antiplatelet treatment Yes 9 (5.7%) 5.8 (0.8, 43.0) — No 149 (94.3%) Dose of gastroprotectant Low 2 (1.3%) 2.1 (1.1, 4.3) — Standard 93 (59.6%) 0.9 (0.2, 4.5) High 61 (39.1%) 1.9 (0.4, 8.5) Length of prescription >4 weeks 34 (21.8%) ≤ 4 weeks 122 (78.2%) 2.3 (1.5, 3.5) 2.4 (1.6, 3.7) Adverse events No 123 (77.9%) Yes 35 (22.1%) 14.6 (7.9, 27.2) 15.5 (8.2, 29.3) CI, confidence interval; GPA, gastroprotective agent; OR, odds ratio.

.1, 4.3) — Standard 93 (59.6%) 0.9 (0.2, 4.5) High 61 (39.1%) 1.9 (0.4, 8.5) Length of prescription >4 weeks 34 (21.8%) ≤ 4 weeks 122 (78.2%) 2.3 (1.5, 3.5) 2.4 (1.6, 3.7) Adverse events No 123 (77.9%) Yes 35 (22.1%) 14.6 (7.9, 27.2) 15.5 (8.2, 29.3) CI, confidence interval; GPA, gastroprotective agent; OR, odds ratio. a Adjusted ORs in the final model. ORs of variables that were not statistically significant in the logistic regression model are not reported.

INTRODUCTION Irritable bowel syndrome (IBS) is a common gastrointestinal disorder characterized by frequent and intermittent episodes of abdominal pain and abdominal discomfort that are associated with altered bowel habits (1,2). The symptoms of IBS not only adversely affect a patient's health-related quality of life (3) but also place a significant financial burden on society due to reduced work productivity and increased use of healthcare-related resources (4,5). IBS with constipation (IBS-C) affects approximately one-third of IBS patients (3), occurs more commonly in women than men (6), and frequently includes additional symptoms, such as abdominal bloating, hard stools, straining, and sensation of incomplete evacuation (7,8). Traditional therapies for IBS-C, generally directed towards the patient's predominant symptoms (9), are frequently associated with patient dissatisfaction (10). More recent therapies, including tegaserod and lubiprostone, have been shown to improve global symptoms of IBS-C (9). Tegaserod, a 5-HT4 partial agonist approved by the Food and Drug Administration (FDA) for the short-term treatment of women with IBS-C, was removed from the market in 2007 due to increased cardiovascular events in patients receiving the medication. Lubiprostone, a chloride channel activator that was approved by the FDA for the treatment of women with IBS-C in 2008, has shown efficacy using a global end point (global symptom relief) (11). Given the limited treatments currently available for patients with IBS-C, additional therapeutic options would be of value.

on. Lubiprostone, a chloride channel activator that was approved by the FDA for the treatment of women with IBS-C in 2008, has shown efficacy using a global end point (global symptom relief) (11). Given the limited treatments currently available for patients with IBS-C, additional therapeutic options would be of value. Linaclotide, a minimally absorbed 14-amino-acid peptide structurally related to the endogenous guanylin peptide family of hormones that regulate fluid and electrolyte homeostasis in the intestine, binds to and activates GCC (guanylate cyclase-C) on the luminal surface of the intestinal epithelium. Activation of GCC results in the generation of cyclic guanosine monophosphate (cGMP), which is increased in both the intracellular and extracellular compartments. The increase in cGMP within intestinal epithelial cells triggers a signal transduction cascade activating the cystic fibrosis transmembrane conductance regulator (12). This activation causes secretion of chloride and bicarbonate into the intestinal lumen; sodium ions and water follow, resulting in increased luminal fluid secretion and a reflex acceleration of intestinal transit. Extracellular cGMP, actively transported out of intestinal epithelial cells, is believed to reduce visceral hyperalgesia by modulating the activity of afferent pain fibers (13). In animal models, linaclotide treatment accelerated gastrointestinal transit and reduced visceral nociception (14); in human phase 2 clinical studies, it accelerated colonic transit (15) and improved abdominal pain and constipation associated with IBS-C (16). Likewise, in two large phase 3 trials in patients with chronic constipation, linaclotide significantly improved bowel and abdominal symptoms over 12 weeks (17).

ciception (14); in human phase 2 clinical studies, it accelerated colonic transit (15) and improved abdominal pain and constipation associated with IBS-C (16). Likewise, in two large phase 3 trials in patients with chronic constipation, linaclotide significantly improved bowel and abdominal symptoms over 12 weeks (17). The objective of this phase 3 clinical trial was to assess the efficacy and safety of linaclotide administered once daily as an oral capsule at a dose of 290 μg vs. placebo to patients with IBS-C. A 4-week randomized withdrawal (RW) period was included in this trial to assess the effect of discontinuing treatment with linaclotide. METHODS Trial design This multicenter, randomized, double-blind, placebo-controlled, parallel-group trial was conducted at 118 outpatient clinical research centers (111 in the United States, 7 in Canada) from 14 July 2009 (first patient enrolled) to 12 July 2010 (last patient completed). The protocol and all trial procedures were approved by an Institutional Review Board, and the trial was designed, conducted, and reported in accordance with the principles of Good Clinical Practice guidelines. All patients gave written informed consent before their participation in the trial.

(last patient completed). The protocol and all trial procedures were approved by an Institutional Review Board, and the trial was designed, conducted, and reported in accordance with the principles of Good Clinical Practice guidelines. All patients gave written informed consent before their participation in the trial. After a screening period of up to 21 days followed by a pretreatment baseline period of 14–21 days, eligible patients were randomly assigned with the use of an interactive voice-response system (IVRS) to receive once daily an oral capsule of either linaclotide 290 μg or placebo, in a 1:1 ratio. Patients who completed all 12 weeks of the double-blind treatment period were eligible to enter the double-blind 4-week RW period in which patients initially randomized to linaclotide were re-randomized (1:1) to linaclotide 290 μg or placebo, and patients previously randomized to placebo were assigned to receive linaclotide 290 μg once a day. Randomization assignments were generated in blocks of four and stratified according to trial center. All sponsor staff involved in the trial, trial center personnel, and patients were blinded to the allocation of trial treatment. Trial visits were conducted at screening, at the start of the pretreatment baseline period, at randomization (day 1), throughout the treatment period (weeks 2, 4, 8, and 12), and at the beginning and end of the RW period (weeks 13 and 16 (end of trial)). Patients made daily calls to the IVRS to report their symptoms throughout the trial.

onducted at screening, at the start of the pretreatment baseline period, at randomization (day 1), throughout the treatment period (weeks 2, 4, 8, and 12), and at the beginning and end of the RW period (weeks 13 and 16 (end of trial)). Patients made daily calls to the IVRS to report their symptoms throughout the trial. Trial patients Female and male patients were eligible to participate if they were at least 18 years of age, and met modified Rome II criteria for IBS (18). In the 12 months before the screening visit, eligible patients were to have for at least 12 weeks, which need not be consecutive, abdominal pain, or abdominal discomfort that had ≥2 of these three features: (i) relieved with defecation, (ii) onset associated with a change in frequency of stool, and (iii) onset associated with a change in form (appearance) of stool, before starting chronic treatment with tegaserod or lubiprostone (if patients had taken these medications); and <3 spontaneous bowel movements (SBMs) per week (SBM=a bowel movement (BM) occurring in the absence of any laxative, suppository, or enema use during the preceding 24 h), and had at least one additional bowel symptom (straining, lumpy or hard stools, and sensation of incomplete evacuation during >25% of BMs), before starting chronic treatment with tegaserod, lubiprostone, polyethylene glycol 3350, or any laxative (if patients had taken these medications). In addition, patients had to report an average score ≥3.0 for daily abdominal pain at its worst (11-point NRS (numerical rating scale)) as well as an average of <3 complete SBMs (CSBMs) per week (CSBM=an SBM associated with a sense of complete evacuation, as reported by the patient) and ≤5 SBMs per week during the 14 days immediately before randomization (i.e., the baseline period).

ly abdominal pain at its worst (11-point NRS (numerical rating scale)) as well as an average of <3 complete SBMs (CSBMs) per week (CSBM=an SBM associated with a sense of complete evacuation, as reported by the patient) and ≤5 SBMs per week during the 14 days immediately before randomization (i.e., the baseline period). Patients were excluded if they reported loose (mushy) or watery stools for >25% of their BMs during the 12 weeks before screening or, during the baseline period, a BSFS (Bristol Stool Form Scale) (19) score of 7 (watery, no solid pieces) for any SBM, or a BSFS score of 6 (fluffy pieces with ragged edges, a mushy stool) for >1 SBM. Other key exclusion criteria included history of cathartic colon, laxative or enema abuse, ischemic colitis, or pelvic floor dysfunction (unless successful treatment had been documented by a normal balloon expulsion test); bariatric surgery for treatment of obesity or surgery to remove a segment of the gastrointestinal tract at any time before the screening visit, surgery of the abdomen, pelvis, or retroperitoneal structures during the 6 months before the screening visit, appendectomy or cholecystectomy during the 60 days before the screening visit, or other major surgery during the 30 days before the screening visit; history of diverticulitis or any chronic condition that could be associated with abdominal pain or discomfort and could confound the assessments in the trial (e.g., inflammatory bowel disease, chronic pancreatitis, polycystic kidney disease, ovarian cysts, endometriosis, lactose intolerance); family history of a familial form of colorectal cancer. In general, patients were excluded if they were taking drugs that could cause constipation (e.g., narcotics); however, patients taking certain drugs for IBS that might be constipating (e.g., tricyclic antidepressants) were eligible provided that they were on a stable dose for at least 30 days before the screening visit and there was no plan to change the dose after the screening visit. Colonoscopy requirements were based on the American Gastroenterological Association guidelines (20). Women of childbearing potential were required to use contraceptives and have a negative serum pregnancy test. Patients were asked to refrain from making any major lifestyle changes (e.g., starting a new diet or changing their exercise pattern) during the trial.

American Gastroenterological Association guidelines (20). Women of childbearing potential were required to use contraceptives and have a negative serum pregnancy test. Patients were asked to refrain from making any major lifestyle changes (e.g., starting a new diet or changing their exercise pattern) during the trial. Rescue medication (bisacodyl 5 mg tablet or 10 mg suppository) was allowed for severe constipation (i.e., 72 h after the patient's previous BM or when symptoms became intolerable). Use of rescue medication was not allowed on the day before, the day of, and the calendar day after the randomization visit. Patients on a stable, continuous regimen of fiber, bulk laxatives, stool softeners, or probiotics during the 30 days before the screening visit were allowed to continue, provided they maintained a stable dosage throughout the trial.

owed on the day before, the day of, and the calendar day after the randomization visit. Patients on a stable, continuous regimen of fiber, bulk laxatives, stool softeners, or probiotics during the 30 days before the screening visit were allowed to continue, provided they maintained a stable dosage throughout the trial. Efficacy assessments and end points Daily reports by patients to IVRS included symptom ratings of worst abdominal pain, abdominal discomfort, abdominal cramping, abdominal fullness, and abdominal bloating (all abdominal symptoms were measured using an 11-point NRS), as well as the number of BMs and whether rescue medication was used. Each BM was assessed for sensation of complete bowel emptying (yes/no), stool consistency (7-point BSFS with 1=“separate hard lumps like nuts” to 7=“watery, no solid pieces”), and severity of straining (5-point ordinal scale). Weekly IVRS assessments included IBS severity and constipation severity (both using a 5-point ordinal scale), degree of IBS relief (7-point balanced scale), and adequate relief of IBS-C symptoms (yes/no). Assessment of satisfaction with the trial-medication's ability to relieve IBS symptoms (5-point ordinal scale) was captured at all study visits following randomization.

stipation severity (both using a 5-point ordinal scale), degree of IBS relief (7-point balanced scale), and adequate relief of IBS-C symptoms (yes/no). Assessment of satisfaction with the trial-medication's ability to relieve IBS symptoms (5-point ordinal scale) was captured at all study visits following randomization. Primary end points There were four prespecified primary end points in the trial, which were all responder end points. One of the four primary end points was based on the FDA recommendations for IBS-C trial design and end points in the recently finalized guidance for IBS clinical trials (May 2012) (21); a responder for this end point (to be referred to hereafter as “FDA end point”) was defined as a patient who met both of the following criteria in the same week for at least 6 of the 12 weeks of the treatment period: (i) an improvement of ≥30% from baseline in the average of the daily worst abdominal pain scores (to be referred to hereafter as “abdominal pain”) and (ii) an increase of ≥1 CSBM from baseline. This combined end point was added after the initiation of the trial, but before completion of enrollment and database lock, with a protocol amendment (no unblinding had occurred). The other three primary end points also required patients to meet weekly responder definitions, but for at least 9 of the 12 weeks of the treatment period. These weekly responder definitions were (i) an improvement of ≥30% in abdominal pain, (ii) ≥3 CSBMs and an increase of ≥1 CSBM from baseline, and (iii) a combined end point that defined a responder as a patient who met criteria for both i and ii in the same week.

for at least 9 of the 12 weeks of the treatment period. These weekly responder definitions were (i) an improvement of ≥30% in abdominal pain, (ii) ≥3 CSBMs and an increase of ≥1 CSBM from baseline, and (iii) a combined end point that defined a responder as a patient who met criteria for both i and ii in the same week. Secondary end points The secondary end points included 12-week change from baseline in abdominal pain, abdominal discomfort, abdominal bloating, stool frequency (CSBM and SBM weekly rates), stool consistency (BSFS), and severity of straining; secondary responder end points included abdominal pain and CSBM responders (using the individual components of the FDA end point). A number of other additional end points were also assessed, including 12-week change from baseline in abdominal fullness and abdominal cramping, IBS symptom severity, constipation severity, adequate relief of IBS-C symptoms, degree of relief of IBS symptoms, and treatment satisfaction. Safety assessments At each scheduled study visit, all patients were asked an open-ended question regarding adverse events (AEs). Patients reported AEs by recalling instances since the prior visit. The site investigator assessed all patient-reported AEs and judged each event for severity and relationship to the blinded trial medication. Other safety evaluations included physical examinations, electrocardiogram recordings, vital sign measurements, and standard clinical laboratory tests.

nces since the prior visit. The site investigator assessed all patient-reported AEs and judged each event for severity and relationship to the blinded trial medication. Other safety evaluations included physical examinations, electrocardiogram recordings, vital sign measurements, and standard clinical laboratory tests. Pharmacokinetic assessments During the treatment period, a subset of patients had blood samples taken at the randomization and week 4 visits to determine if linaclotide or its active metabolite, MM 419447, could be detected at quantifiable levels in the plasma. Statistical methods and data analysis The overall family-wise type I error rate for testing the primary and secondary efficacy end points was controlled at the 0.05 significance level using a five-step serial gate-keeping, multiple-comparison procedure. Based on this multiple-comparison procedure and the results of a previous phase 2b study (16), a sample size of 400 patients per treatment arm was selected to provide >85% overall power to simultaneously detect a difference between the placebo and linaclotide groups for the primary end points.

ple-comparison procedure. Based on this multiple-comparison procedure and the results of a previous phase 2b study (16), a sample size of 400 patients per treatment arm was selected to provide >85% overall power to simultaneously detect a difference between the placebo and linaclotide groups for the primary end points. Responder end points were analyzed using a Cochran–Mantel–Haenszel (CMH) test controlling for geographic region. Continuous change-from-baseline end points were analyzed using an ANCOVA (analysis of covariance) model with fixed-effect terms for treatment group and geographic region and the corresponding baseline value as a covariate. Least-squares means (i.e., means adjusted for the other effects) from the ANCOVA model based on patients' overall average scores (except for SBMs and CSBMs, for which the overall weekly rates were calculated) are presented. Geographic region was used as a factor in the analyses rather than individual trial centers due to the potential for very small numbers of patients at some trial centers.

ANCOVA model based on patients' overall average scores (except for SBMs and CSBMs, for which the overall weekly rates were calculated) are presented. Geographic region was used as a factor in the analyses rather than individual trial centers due to the potential for very small numbers of patients at some trial centers. If a patient dropped out of the trial or otherwise did not report efficacy data for a particular treatment-period week (patients were required to complete at least four IVRS calls during a treatment week), the patient was not considered a responder for that week. An observed-cases approach to missing data was applied to the change-from-baseline secondary end points, such that if a patient dropped out of the trial or otherwise did not report data, the average of the non-missing data over the 12 weeks of the treatment period was the patient's value. Patients were assumed to have not had BMs nor taken rescue medication if the corresponding daily question was not answered. For the analysis of adequate relief, degree of relief of IBS symptoms, and treatment satisfaction, a last observation carried forward method was used. All P values were based on two-sided tests. All randomized patients who took at least one dose of trial medication were included in safety analyses (safety population). Efficacy analyses were based on the ITT (intent-to-treat) population, which included all patients in the safety population who had at least one post-randomization entry of the primary efficacy assessment (i.e., IVRS assessment of abdominal pain or CSBMs).

ased paracellular permeability in patients with celiac disease due to gliadin-induced innate and adaptive immune responses (11,26,27,28) and subsequent tight junction disassembly (28,29,30). In addition, genetic defects have been identified in the cytoskeletal proteins involved in tight junction functioning (31,32,33). Larazotide acetate (formerly referred to as AT-1001) is a first-in-class, tight-junction regulator peptide that in vitro prevents the opening of intestinal epithelial tight junctions induced by multiple stimuli, including cytokines, bacterial antigens, and gluten peptides (34,35). Its immunological activity is limited to the luminal surface of the small intestine (36,37). In an inpatient, randomized, double-blind, placebo-controlled, single-dose study 37), patients who received larazotide acetate had a significant reduction in gastrointestinal symptoms, particularly diarrhea, after a 2.5-g gluten challenge compared with those who received placebo. In addition, patients in the placebo group had a 70% increase in the urinary lactulose-to-mannitol (LAMA) ratio, a measure of intestinal permeability, whereas those receiving larazotide acetate had no change. Larazotide acetate was not detected in the plasma after supra-therapeutic doses, and no significant systemic toxicities were observed.

medication were included in safety analyses (safety population). Efficacy analyses were based on the ITT (intent-to-treat) population, which included all patients in the safety population who had at least one post-randomization entry of the primary efficacy assessment (i.e., IVRS assessment of abdominal pain or CSBMs). RESULTS Patient disposition, demographics, and baseline characteristics Of the 2,424 patients who were screened for participation in this trial, 803 (33%) were randomized to treatment (Figure 1). Two patients were randomized at more than one trial center but only data from the trial center in which they were first randomized were included in statistical analyses. Of the 802 patients who received double-blind trial medication (safety population), 800 patients had at least one post-randomization entry of the primary efficacy assessment (ITT population). The demographics of the ITT population are shown in Table 1. Following completion of the treatment period, a total of 647 (81%) ITT patients entered the RW period of the trial, of which 645 received at least one dose of trial medication and were included in the RW population. Mean compliance with the trial-medication dosing (assessed by counting pills returned at trial visits) up to trial discontinuation/completion of the 12-week treatment period was 95 and 94% for the placebo and linaclotide groups, respectively. Compliance with the daily IVRS call-in (patients who completed ≥80% of scheduled calls) during the treatment period was 73 and 71% for placebo- and linaclotide-treated patients, respectively. During the pretreatment baseline period, 88% of patients experienced abdominal pain every day and 76% of patients had no CSBMs.

Compliance with the daily IVRS call-in (patients who completed ≥80% of scheduled calls) during the treatment period was 73 and 71% for placebo- and linaclotide-treated patients, respectively. During the pretreatment baseline period, 88% of patients experienced abdominal pain every day and 76% of patients had no CSBMs. Efficacy results For all primary and secondary efficacy end points, the linaclotide 290-μg group demonstrated statistically significant improvement compared with the placebo group, controlling for multiplicity.

Compliance with the daily IVRS call-in (patients who completed ≥80% of scheduled calls) during the treatment period was 73 and 71% for placebo- and linaclotide-treated patients, respectively. During the pretreatment baseline period, 88% of patients experienced abdominal pain every day and 76% of patients had no CSBMs. Efficacy results For all primary and secondary efficacy end points, the linaclotide 290-μg group demonstrated statistically significant improvement compared with the placebo group, controlling for multiplicity. For the individual components of the FDA end point, a significantly greater percentage of linaclotide-treated patients, compared with placebo-treated patients, reported a reduction of ≥30% in abdominal pain for at least 6 out of the 12 weeks of the treatment period (50.1 vs. 37.5%, P=0.0003 (Figure 2)) or an increase of ≥1 CSBM from baseline for at least 6 out of the 12 weeks of the treatment period (48.6 vs. 29.6%, P<0.0001 (Figure 2)). A total of 136 of 405 patients (33.6%) receiving linaclotide compared with 83 of 395 patients (21.0%) receiving placebo (odds ratio: 1.9, 95% confidence interval: 1.4, 2.7; P<0.0001) met the FDA end point ( Table 2; Figure 2). A significantly greater percentage of linaclotide-treated patients than placebo-treated patients also met the responder requirements for the other three primary end points, which required improvement in abdominal pain (i.e., a reduction of ≥30% in abdominal pain), CSBM rate (i.e., ≥3 CSBMs and an increase of ≥1 CSBM), or both for at least 9 of the 12 weeks of the treatment period ( Table 2). The NNT (number needed to treat) for the primary end points ranged from 7.6 to 14.3.

y end points, which required improvement in abdominal pain (i.e., a reduction of ≥30% in abdominal pain), CSBM rate (i.e., ≥3 CSBMs and an increase of ≥1 CSBM), or both for at least 9 of the 12 weeks of the treatment period ( Table 2). The NNT (number needed to treat) for the primary end points ranged from 7.6 to 14.3. Linaclotide-treated patients also experienced statistically significantly greater improvements compared with placebo-treated patients for the secondary and additional end points ( Table 3). During the first week of treatment and for each subsequent week of treatment, linaclotide-treated patients reported greater improvements in worst abdominal pain and CSBM frequency compared with placebo-treated patients (P<0.001; Figure 3). At week 12, the mean decrease from baseline in worst abdominal pain was 2.4% for linaclotide vs. 1.5% for placebo (P<0.0001), and the mean increase from baseline in the weekly CSBM rate was 2.4 and 0.9 for linaclotide and placebo, respectively (P<0.0001). At the end of the Treatment Period (week 12), 52% of linaclotide-treated patients were either “very satisfied” or “quite satisfied” with treatment compared with 23% of placebo-treated patients (P<0.0001).

ncrease from baseline in the weekly CSBM rate was 2.4 and 0.9 for linaclotide and placebo, respectively (P<0.0001). At the end of the Treatment Period (week 12), 52% of linaclotide-treated patients were either “very satisfied” or “quite satisfied” with treatment compared with 23% of placebo-treated patients (P<0.0001). During the 4-week RW Period, patients who were re-randomized from linaclotide to placebo showed an increase in worst abdominal pain and a decrease in CSBMs to levels similar to those observed in the placebo group during the Treatment Period. The patients who continued to take linaclotide showed sustained improvement in worst abdominal pain and CSBMs similar to that previously observed during the Treatment Period. These improvements were statistically significant compared to patients re-randomized to placebo for weeks 13–16 for CSBMs (P<0.001) and weeks 14–16 for worst abdominal pain (P<0.05). Patients who switched from placebo to linaclotide showed levels of improvement similar to those experienced by linaclotide-treated patients during the Treatment Period (Figure 3).

ificant compared to patients re-randomized to placebo for weeks 13–16 for CSBMs (P<0.001) and weeks 14–16 for worst abdominal pain (P<0.05). Patients who switched from placebo to linaclotide showed levels of improvement similar to those experienced by linaclotide-treated patients during the Treatment Period (Figure 3). Safety A total of 228 of 406 linaclotide-treated patients (56.2%) reported at least one treatment-emergent AE (TEAE) compared with 210 of 396 placebo-treated patients (53.0%) in the 12-week treatment period ( Table 4). Most TEAEs were mild or moderate in severity (93.8%, linaclotide; 98.1%, placebo). The incidences of diarrhea (P<0.0001), flatulence (P=0.0084), and abdominal pain (P=0.0462) TEAEs were significantly greater in the linaclotide-treated patients compared with placebo-treated patients. The most common TEAE in the 12-week treatment period was diarrhea, experienced by 19.5% of linaclotide-treated patients compared with 3.5% of placebo-treated patients. The occurrences of diarrhea were reported to be mild or moderate in 71 of 79 linaclotide-treated patients (89.9%) and 13 of 14 placebo-treated patients (92.9%) who experienced diarrhea. There were no SAEs of diarrhea reported during the trial. None of the patients who reported diarrhea experienced clinically significant sequelae (e.g., orthostatic hypotension or dehydration). More than half of linaclotide-treated patients who experienced diarrhea had onset within the first 2 weeks of treatment. Diarrhea was the most common AE resulting in treatment discontinuation in linaclotide-treated patients (5.7 vs. 0.3% in placebo-treated patients); overall, AEs resulted in the premature discontinuation of 32 patients (7.9%) and 11 patients (2.8%) taking linaclotide and placebo, respectively, in the treatment period.

t. Diarrhea was the most common AE resulting in treatment discontinuation in linaclotide-treated patients (5.7 vs. 0.3% in placebo-treated patients); overall, AEs resulted in the premature discontinuation of 32 patients (7.9%) and 11 patients (2.8%) taking linaclotide and placebo, respectively, in the treatment period. Rates of serious AEs (SAEs) did not differ between linaclotide and placebo groups (two patients in each group (0.5%)). In the linaclotide group, the SAEs consisted of one patient who experienced asthma and a second patient who experienced pericardial effusion and pericarditis leading to withdrawal from the trial. In the placebo group, the SAEs consisted of one patient who experienced chronic cholecystitis and a second patient who experienced duodenitis, gastroenteritis, hiatal hernia, esophagitis, renal cyst, and urinary tract infection. There were no deaths during the treatment period; one screened patient died as a result of cardiorespiratory arrest and ventricular fibrillation due to a possible drug overdose, but this patient died before randomization and did not receive trial medication.

hernia, esophagitis, renal cyst, and urinary tract infection. There were no deaths during the treatment period; one screened patient died as a result of cardiorespiratory arrest and ventricular fibrillation due to a possible drug overdose, but this patient died before randomization and did not receive trial medication. There were no clinically significant differences between the linaclotide and placebo groups in the incidence of abnormal laboratory parameters, vital signs, or electrocardiogram parameters. Serum bicarbonate levels were below the lower limit of normal at the end of treatment in seven patients receiving linaclotide compared with one patient receiving placebo. None of these patients reported diarrhea as an AE or other AEs that were considered to be related to low bicarbonate levels. In the subset of patients who were assessed for linaclotide exposure, no quantifiable plasma levels of linaclotide were detected following trial-medication dosing at the randomization and week 4 trial visits. All patients tested (72 placebo and 64 linaclotide) had levels lower than the limit of quantification for linaclotide (<0.2 ng/ml) and its primary metabolite, MM-419447 (<2.0 ng/ml).

o quantifiable plasma levels of linaclotide were detected following trial-medication dosing at the randomization and week 4 trial visits. All patients tested (72 placebo and 64 linaclotide) had levels lower than the limit of quantification for linaclotide (<0.2 ng/ml) and its primary metabolite, MM-419447 (<2.0 ng/ml). During the RW period, TEAEs occurred in 22.2% of linaclotide–linaclotide patients, 22.1% of linaclotide–placebo patients, and 30.6% of placebo–linaclotide patients. With the exceptions of diarrhea and abdominal pain, the incidence of TEAEs was similar across the three treatment sequences. The incidence of diarrhea was 1.9, 0.6, and 11.7%, in linaclotide–linaclotide, linaclotide–placebo, and placebo–linaclotide patients, respectively. The incidence of abdominal pain was 1.3% in the linaclotide–linaclotide patients and 2.4% in the placebo–linaclotide patients; there were no TEAEs of abdominal pain in the linaclotide–placebo patients. There was no evidence of “rebound” (i.e., worsening in IBS-C symptoms compared with the baseline period in the linaclotide–placebo patients). No SAEs were reported during the RW period.

tide patients and 2.4% in the placebo–linaclotide patients; there were no TEAEs of abdominal pain in the linaclotide–placebo patients. There was no evidence of “rebound” (i.e., worsening in IBS-C symptoms compared with the baseline period in the linaclotide–placebo patients). No SAEs were reported during the RW period. DISCUSSION In this large phase 3 clinical trial, a greater percentage of IBS-C patients who were treated with linaclotide achieved statistically significant improvement in the key symptoms of IBS-C, including abdominal pain and constipation, compared with placebo. Four primary outcomes measures were assessed, including the FDA end point. This end point required that patients experience a benefit of at least 30% when compared with baseline in abdominal pain and an increase of ≥1 CSBM from baseline in the same week for at least 6 out of the 12 weeks of the treatment period. In spite of the rigor of this end point, 33.6% of linaclotide-treated patients were responders compared with 21.0% of placebo-treated patients (P<0.0001). Furthermore, statistically significant differences in responder rates were also demonstrated for the three other primary end points, which required (i) a decrease in abdominal pain of ≥30%, (ii) both an absolute value of ≥3 CSBMs and an increase of ≥1 CSBM from baseline, and (iii) both abdominal pain and CSBM criteria for at least 9 of the 12 weeks of the treatment period.

esponder rates were also demonstrated for the three other primary end points, which required (i) a decrease in abdominal pain of ≥30%, (ii) both an absolute value of ≥3 CSBMs and an increase of ≥1 CSBM from baseline, and (iii) both abdominal pain and CSBM criteria for at least 9 of the 12 weeks of the treatment period. Although IBS is a disorder with multiple symptoms, abdominal pain is one of the cardinal manifestations and strongly correlates with IBS severity (22) and utilization of healthcare resources (23). Also, an improvement in abdominal pain of ≥30% has been shown to be clinically important in IBS patients (23), and in patients reporting pain relief in general (24). In this trial, more than half of linaclotide-treated patients reported an improvement in abdominal pain of ≥30% for at least 6 out of 12 weeks compared with 37.5% of placebo-treated patients, for an NNT of 7.9. Improvement in abdominal pain began within the first week of therapy, and once reaching maximum at 6–8 weeks, was sustained throughout the remainder of the treatment period. By the last week of the treatment period (week 12), linaclotide-treated patients reported a mean improvement of 43.2% in abdominal pain compared with 27.5% for placebo-treated patients. During the 4-week RW period, patients re-randomized to remain on linaclotide had continued relief of abdominal pain, showing durability of response, while those re-randomized from linaclotide to placebo showed a gradual worsening of abdominal pain symptoms to the level experienced by patients receiving placebo during the treatment period, but without signs of a “rebound” or worsening of symptoms relative to baseline.

abdominal pain, showing durability of response, while those re-randomized from linaclotide to placebo showed a gradual worsening of abdominal pain symptoms to the level experienced by patients receiving placebo during the treatment period, but without signs of a “rebound” or worsening of symptoms relative to baseline. In addition to abdominal pain, linaclotide improved several other important abdominal symptoms that are frequently reported by IBS-C patients, including abdominal bloating and abdominal discomfort, beginning during the first week of treatment and continuing throughout the 12-week treatment period. Linaclotide also improved bowel function, including SBM and CSBM frequency, straining, stool consistency, and constipation severity. However, in contrast to the gradual improvement in abdominal symptoms, improvement in bowel function occurred more rapidly. Most linaclotide-treated patients experienced an SBM within 24 h of the first dose of linaclotide (67.4 vs. 43.8% for placebo, P<0.0001); maximal improvement in bowel function usually occurred within the first week. Thus, improvement with linaclotide in abdominal (sensory) symptoms such as abdominal pain may be attributable to more than improvement in bowel function alone. Preclinical data suggest that cGMP, which is released intra- and extracellularly following GCC activation by linaclotide, can reduce the firing of pain-sensing visceral afferent fibers (13). Further studies are under way that may provide a better understanding of the mechanisms by which linaclotide exerts its beneficial effects directly on abdominal sensory symptoms.

ased intra- and extracellularly following GCC activation by linaclotide, can reduce the firing of pain-sensing visceral afferent fibers (13). Further studies are under way that may provide a better understanding of the mechanisms by which linaclotide exerts its beneficial effects directly on abdominal sensory symptoms. Diarrhea was the most common TEAE in linaclotide-treated patients and appears to be an extension of linaclotide's pharmacological effects. Although diarrhea was reported in 19.5% of linaclotide-treated patients, only 2% reported that they had severe diarrhea and only 5.7% discontinued the drug due to diarrhea. The incidence of SAEs was similar between linaclotide- and placebo-treated patients (n=2 patients in each group); diarrhea was not reported as an SAE. In conclusion, linaclotide significantly improved abdominal and bowel symptoms in this phase 3 trial (12-week treatment period+4-week RW period). STUDY HIGHLIGHTS Annie Neild, PhD, of Ironwood Pharmaceuticals, provided editorial assistance. Portions of this manuscript were presented as an oral presentation at Digestive Disease Week, May 2011, Chicago, IL. Guarantor of the article: Jeffrey M. Johnston, MD.

In conclusion, linaclotide significantly improved abdominal and bowel symptoms in this phase 3 trial (12-week treatment period+4-week RW period). STUDY HIGHLIGHTS Annie Neild, PhD, of Ironwood Pharmaceuticals, provided editorial assistance. Portions of this manuscript were presented as an oral presentation at Digestive Disease Week, May 2011, Chicago, IL. Guarantor of the article: Jeffrey M. Johnston, MD. Specific author contributions: Wrote the initial draft of the manuscript, assisted in the interpretation of data, and provided critical revision: Satish Rao and Anthony J. Lembo; designed the trial: Jeffrey M. Johnston, Bernard J. Lavins, Harvey A. Schneier, and Steven J. Shiff; assisted in the interpretation of data and critical revision of the manuscript for important intellectual content: Jeffrey M. Johnston, Bernard J. Lavins, Caroline B. Kurtz, Mark G. Currie, Harvey A. Schneier, and Steven J. Shiff; provided statistical design, analyses, and interpretation: James E. MacDougall, Xinwei D. Jia, Kelvin Shi, and James Z. Shao; coordinated acquisition of data and trial supervision: Paul Eng and Susan M. Fox. Financial support: This trial was funded by Forest Research Institute and Ironwood Pharmaceuticals, Inc.

participating institutions and all patients provided written informed consent (ClinicalTrials. gov Identifier: NCT00362856). The study was conducted at 10 clinical sites in the United States between August 2006 and March 2007. Patients were recruited by the investigators directly and through peer-to-peer communication. Patients Patients aged 18 to 65 years were eligible for the study if they had a body mass index between 18.5 and 38 kg/m2 and a diagnosis of celiac disease confirmed via biopsys ≥6 months before study entry. Patients must have been following a gluten-free diet for ≥6 months before study entry and have been in remission, as measured by antibodies to tTG (≤10 ELISA (enzyme-linked immunosorbent assay) Units (EU) for IgA and IgG, with 10 EU as the cutoff level for a positive test result). Women must have been post-menopausal or surgically sterile or must have had a negative result on the serum β-human chorionic gonadotropin pregnancy test and agreed to use acceptable methods of contraception.

Specific author contributions: Wrote the initial draft of the manuscript, assisted in the interpretation of data, and provided critical revision: Satish Rao and Anthony J. Lembo; designed the trial: Jeffrey M. Johnston, Bernard J. Lavins, Harvey A. Schneier, and Steven J. Shiff; assisted in the interpretation of data and critical revision of the manuscript for important intellectual content: Jeffrey M. Johnston, Bernard J. Lavins, Caroline B. Kurtz, Mark G. Currie, Harvey A. Schneier, and Steven J. Shiff; provided statistical design, analyses, and interpretation: James E. MacDougall, Xinwei D. Jia, Kelvin Shi, and James Z. Shao; coordinated acquisition of data and trial supervision: Paul Eng and Susan M. Fox. Financial support: This trial was funded by Forest Research Institute and Ironwood Pharmaceuticals, Inc. Potential competing interests: Jeffrey M. Johnston, Caroline B. Kurtz, James E. MacDougall, James Z. Shao, Bernard J. Lavins, and Mark G. Currie are employees of Ironwood Pharmaceuticals and own stock/stock options in Ironwood Pharmaceuticals. Harvey A. Schneier, Steven J. Shiff, Paul Eng, Susan M. Fox, Xinwei D. Jia, and Kelvin Shi are employees of Forest Laboratories and own stock/stock options in Forest Laboratories. Anthony J. Lembo and Satish Rao are paid consultants to Ironwood Pharmaceuticals and Forest Research Institute. Figure 1 Patient flow through the study.

Potential competing interests: Jeffrey M. Johnston, Caroline B. Kurtz, James E. MacDougall, James Z. Shao, Bernard J. Lavins, and Mark G. Currie are employees of Ironwood Pharmaceuticals and own stock/stock options in Ironwood Pharmaceuticals. Harvey A. Schneier, Steven J. Shiff, Paul Eng, Susan M. Fox, Xinwei D. Jia, and Kelvin Shi are employees of Forest Laboratories and own stock/stock options in Forest Laboratories. Anthony J. Lembo and Satish Rao are paid consultants to Ironwood Pharmaceuticals and Forest Research Institute. Figure 1 Patient flow through the study. Figure 2 FDA end point and components. FDA end point: ≥30% abdominal pain reduction and increase ≥1 CSBM from baseline in the same week for ≥6/12 weeks. ****P value <0.0001, ***<0.001 for linaclotide vs. placebo (Cochran–Mantel–Haenszel (CMH) test). P values met the criterion for statistical significance based on the multiple-comparison procedure. CSBM, complete spontaneous bowel movement; FDA, Food and Drug Administration; Lin, linaclotide; NNT, number needed to treat.

***P value <0.0001, ***<0.001 for linaclotide vs. placebo (Cochran–Mantel–Haenszel (CMH) test). P values met the criterion for statistical significance based on the multiple-comparison procedure. CSBM, complete spontaneous bowel movement; FDA, Food and Drug Administration; Lin, linaclotide; NNT, number needed to treat. Figure 3 Weekly results for complete spontaneous bowel movement (CSBM) frequency (a, *P<0.0001 for linaclotide patients compared with placebo patients for each of the 12 Treatment-Period weeks, **P<0.001 for linaclotide–linaclotide patients compared with linaclotide–placebo patients for RW Period weeks 13–16); reduction in worst abdominal pain (b, *P<0.001 for linaclotide patients compared with placebo patients for each of the 12 Treatment Period weeks, **P<0.05 for linaclotide–linaclotide patients compared with linaclotide–placebo patients for RW Period weeks 14–16); and percent reduction in worst abdominal pain (c,*P<0.001 for linaclotide patients compared with placebo patients for each of the 12 Treatment Period weeks, **P<0.05 for linaclotide–linaclotide patients compared with linaclotide–placebo patients for RW Period week 14). All P values were derived from an analysis of covariance model.

eduction in worst abdominal pain (c,*P<0.001 for linaclotide patients compared with placebo patients for each of the 12 Treatment Period weeks, **P<0.05 for linaclotide–linaclotide patients compared with linaclotide–placebo patients for RW Period week 14). All P values were derived from an analysis of covariance model. Table 1 Summary of patient demographic and baseline characteristics (ITT population) Placebo, N=395 Linaclotide 290 μg, N=405 Demographic data Age (years), mean (range) 43.7 (18–84) 43.3 (19–81) ≥65 years, n (%) 26 (6.6) 19 (4.7) Sex, n (%) Female 357 (90.4) 367 (90.6) Male 38 (9.6) 38 (9.4) Race, n (%) White 301 (76.2) 314 (77.5) Black 75 (19.0) 78 (19.3) Other 19 (4.8) 13 (3.2) BMI, mean (s.d.) 27.6 (6.2) 28.3 (6.4) Abdominal symptoms, mean (s.d.) Abdominal paina 5.6 (1.7) 5.7 (1.7) Abdominal discomforta 6.0 (1.7) 6.2 (1.6) Abdominal bloatinga 6.5 (1.9) 6.7 (1.8) Abdominal fullnessa 6.5 (1.8) 6.8 (1.7) Abdominal crampinga 5.4 (1.9) 5.4 (1.9) Bowel symptoms, mean (s.d.) CSBMs/week 0.2 (0.5) 0.2 (0.5) SBMs/week 1.9 (1.4) 1.9 (1.4) Stool consistencyb 2.4 (1.0) 2.3 (1.0) Strainingc 3.4 (0.8) 3.6 (0.8) Constipation severityd 3.7 (0.6) 3.8 (0.6) IBS severityd 3.7 (0.6) 3.7 (0.6) BMI, body mass index; CSBM, complete SBM; IBS, irritable bowel syndrome; ITT, intent-to-treat; SBM, spontaneous bowel movement. a Assessed using an 11-point Numerical Rating Scale: 0=none; 10=very severe.

Table 1 Summary of patient demographic and baseline characteristics (ITT population) Placebo, N=395 Linaclotide 290 μg, N=405 Demographic data Age (years), mean (range) 43.7 (18–84) 43.3 (19–81) ≥65 years, n (%) 26 (6.6) 19 (4.7) Sex, n (%) Female 357 (90.4) 367 (90.6) Male 38 (9.6) 38 (9.4) Race, n (%) White 301 (76.2) 314 (77.5) Black 75 (19.0) 78 (19.3) Other 19 (4.8) 13 (3.2) BMI, mean (s.d.) 27.6 (6.2) 28.3 (6.4) Abdominal symptoms, mean (s.d.) Abdominal paina 5.6 (1.7) 5.7 (1.7) Abdominal discomforta 6.0 (1.7) 6.2 (1.6) Abdominal bloatinga 6.5 (1.9) 6.7 (1.8) Abdominal fullnessa 6.5 (1.8) 6.8 (1.7) Abdominal crampinga 5.4 (1.9) 5.4 (1.9) Bowel symptoms, mean (s.d.) CSBMs/week 0.2 (0.5) 0.2 (0.5) SBMs/week 1.9 (1.4) 1.9 (1.4) Stool consistencyb 2.4 (1.0) 2.3 (1.0) Strainingc 3.4 (0.8) 3.6 (0.8) Constipation severityd 3.7 (0.6) 3.8 (0.6) IBS severityd 3.7 (0.6) 3.7 (0.6) BMI, body mass index; CSBM, complete SBM; IBS, irritable bowel syndrome; ITT, intent-to-treat; SBM, spontaneous bowel movement. a Assessed using an 11-point Numerical Rating Scale: 0=none; 10=very severe. b Assessed using the BSFS: 1=separate hard lumps, like nuts (hard to pass); 2=sausage-shaped, but lumpy; 3=like a sausage but with cracks on its surface; 4=like a sausage or snake, smooth and soft; 5=soft blobs with clear cut edges (passed easily); 6=fluffy pieces with ragged edges, a mushy stool; 7=watery, no solid pieces (entirely liquid). c Assessed using a 5-point ordinal scale: 1=not at all; 2=a little bit; 3=a moderate amount; 4=a great deal; 5=an extreme amount.

b Assessed using the BSFS: 1=separate hard lumps, like nuts (hard to pass); 2=sausage-shaped, but lumpy; 3=like a sausage but with cracks on its surface; 4=like a sausage or snake, smooth and soft; 5=soft blobs with clear cut edges (passed easily); 6=fluffy pieces with ragged edges, a mushy stool; 7=watery, no solid pieces (entirely liquid). c Assessed using a 5-point ordinal scale: 1=not at all; 2=a little bit; 3=a moderate amount; 4=a great deal; 5=an extreme amount. d Assessed using a 5-point ordinal scale: 1=none; 2=mild; 3=moderate; 4=severe; 5=very severe. All demographic characteristics were similar between treatment groups. For baseline clinical characteristics, significant differences were observed for abdominal fullness (P=0.011), stool consistency (P=0.046), and straining (P=0.020).

d Assessed using a 5-point ordinal scale: 1=none; 2=mild; 3=moderate; 4=severe; 5=very severe. All demographic characteristics were similar between treatment groups. For baseline clinical characteristics, significant differences were observed for abdominal fullness (P=0.011), stool consistency (P=0.046), and straining (P=0.020). Table 2 Primary efficacy parameter results (ITT population) Primary efficacy parameters Placebo responder (N=395), n (%) Linaclotide responder (N=405), n (%) Difference Odds ratio (95% CI) P valuea NNT (95% CI) FDA end point (each week, ≥30% decrease in worst abdominal pain+an increase ≥1 CSBM from baseline for at least 6/12 weeks) 83 (21.0) 136 (33.6) 12.6 1.9 (1.4, 2.7) <0.0001 8.0 (5.4, 15.5) ≥30% Decrease in worst abdominal pain (each week, ≥30% decrease in abdominal pain from baseline for at least 9/12 weeks) 107 (27.1) 139 (34.3) 7.2 1.4 (1.0, 1.9) 0.0262 13.8 (7.4, 116.1) ≥3 CSBMs and an increase of ≥1 CSBM (each week, ≥3 CSBM+an increase ≥1 CSBM from baseline for at least 9/12 weeks) 25 (6.3) 79 (19.5) 13.2 3.7 (2.3, 5.9) <0.0001 7.6 (5.6, 11.6) Combined responder (each week ≥30% decrease in worst abdominal pain+≥3 CSBM+an increase ≥1 CSBM from baseline for at least 9/12 weeks) 20 (5.1) 49 (12.1) 7.0 2.6 (1.5, 4.5) 0.0004 14.2 (9.2, 31.3) CI, confidence interval; CSBM, complete spontaneous bowel movement; FDA, Food and Drug Administration; ITT, intent-to-treat; NNT, number needed to treat. a P values were based on a comparison of linaclotide vs. the placebo group using the Cochran–Mantel–Haenszel test.

Table 2 Primary efficacy parameter results (ITT population) Primary efficacy parameters Placebo responder (N=395), n (%) Linaclotide responder (N=405), n (%) Difference Odds ratio (95% CI) P valuea NNT (95% CI) FDA end point (each week, ≥30% decrease in worst abdominal pain+an increase ≥1 CSBM from baseline for at least 6/12 weeks) 83 (21.0) 136 (33.6) 12.6 1.9 (1.4, 2.7) <0.0001 8.0 (5.4, 15.5) ≥30% Decrease in worst abdominal pain (each week, ≥30% decrease in abdominal pain from baseline for at least 9/12 weeks) 107 (27.1) 139 (34.3) 7.2 1.4 (1.0, 1.9) 0.0262 13.8 (7.4, 116.1) ≥3 CSBMs and an increase of ≥1 CSBM (each week, ≥3 CSBM+an increase ≥1 CSBM from baseline for at least 9/12 weeks) 25 (6.3) 79 (19.5) 13.2 3.7 (2.3, 5.9) <0.0001 7.6 (5.6, 11.6) Combined responder (each week ≥30% decrease in worst abdominal pain+≥3 CSBM+an increase ≥1 CSBM from baseline for at least 9/12 weeks) 20 (5.1) 49 (12.1) 7.0 2.6 (1.5, 4.5) 0.0004 14.2 (9.2, 31.3) CI, confidence interval; CSBM, complete spontaneous bowel movement; FDA, Food and Drug Administration; ITT, intent-to-treat; NNT, number needed to treat. a P values were based on a comparison of linaclotide vs. the placebo group using the Cochran–Mantel–Haenszel test. Table 3 Other efficacy parameter results (ITT population) Placebo, N=395 Linaclotide 290 μg, N=405 Difference P value NNT (95% CI) Worst abdominal pain Mean (11-point NRS scale) 4.4 3.7 aChange from baseline, meanb,c −1.1 −1.9 −0.7 <0.0001 a% of patients with ≥30% decrease in worst abdominal pain for at least 6/12 weeksd 37.5 50.1 12.7 0.0003 7.9 (5.1, 17.1) Abdominal discomfort Mean (11-point NRS scale) 4.7 4.1 aChange from baseline, meanb,c −1.2 −2.0 −0.7 <0.0001 % of patients with ≥30% decrease in abdominal discomfort for at least 6/12 weeksd 37.0 48.1 11.2 0.0013 8.9 (5.6, 22.8) Abdominal bloating Mean (11-point NRS scale) 5.3 4.6 aChange from baseline, meanb,c −1.1 −1.9 −0.8 <0.0001 % of patients with ≥30% decrease in abdominal bloating for at least 6/12 weeksd 29.9 43.5 13.6 <0.0001 7.4 (5.0, 14.3) Abdominal fullness Mean (11-point NRS scale) 5.3 4.6 Change from baseline, meanb,c −1.1 −2.0 −0.9 <0.0001 % of patients with ≥30% decrease in abdominal fullness for at least 6/12 weeksd 32.9 44.0 11.0 0.0012 9.1 (5.6, 23.0) Abdominal cramping Mean (11-point NRS scale) 4.1 3.5 Change from baseline, meanb,c −1.1 −1.7 −0.6 <0.0001 % of patients with ≥30% decrease in abdominal cramping for at least 6/12 weeksd 39.5 49.9 10.4 0.0029 9.6 (5.8, 28.3) CSBMs Mean CSBMs/week 1.0 2.6 aChange from baseline, meanb,c 0.7 2.3 1.6 <0.0001 CSBM ≤24 h first dose (%)c 13.2 32.3 19.2 <0.0001 5.2 (4.0, 7.4) a% of patients w/ CSBM rate increase ≥1 per week for at least 6/12 weeksd 29.6 48.6 19.0 <0.0001 5.3 (3.9, 8.1) SBMs Mean SBMs/week 3.2 6.0 aChange from baselineb,c 1.1 3.9 2.8 <0.0001 SBM ≤24 h after first dose (%)c 43.8 67.4 23.6 <0.0001 4.2 (3.3, 5.9) % of patients w/ SBM rate increase ≥2 per week from baseline for at least 6/12 weeksd 29.4 57.5 28.2 <0.0001 3.6 (2.9, 4.6) Stool consistency Mean BSFS score (1–7) 3.1 4.5 aChange from baseline, meanb,c 0.7 2.1 1.4 <0.0001 Mean weekly % of SBMs without hard or lumpy stools (BSFS ≥3) 60.7 79.4 18.7 <0.0001 Straining Mean straining score (1–5) 2.8 2.2 aChange from baseline, meanb,c −0.7 −1.3 −0.7 <0.0001 Mean weekly % of SBMs without significant straining (i.e., score ≤3) 71.7 85.3 13.6 <0.0001 Constipation severity Mean constipati

<0.0001 Mean weekly % of SBMs without hard or lumpy stools (BSFS ≥3) 60.7 79.4 18.7 <0.0001 Straining Mean straining score (1–5) 2.8 2.2 aChange from baseline, meanb,c −0.7 −1.3 −0.7 <0.0001 Mean weekly % of SBMs without significant straining (i.e., score ≤3) 71.7 85.3 13.6 <0.0001 Constipation severity Mean constipati on severity score (1–5) 3.1 2.6 Change from baseline, meanb,c −0.6 −1.2 −0.6 <0.0001 % of patients with decrease of ≥1 for at least 6/12 weeksd 42.5 59.5 17.0 <0.0001 5.9 (4.2, 9.9) IBS severity Mean IBS severity score (1–5) 3.1 2.7 Change from baseline, meanb,c −0.5 −1.0 −0.5 <0.0001 % of patients with decrease of ≥1 for at least 6/12 weeksd 37.5 56.3 18.8 <0.0001 5.3 (3.9, 8.3) Adequate relief % of patients reporting adequate relief of IBS symptoms for at least 75% of the weeks (i.e., 9/12 weeks)d 21.3 36.8 15.5 <0.0001 6.4 (4.6, 10.7) % of patients reporting adequate relief of IBS symptoms for at least 50% of the weeks (i.e., 6/12 weeks)d 34.2 48.9 14.7 <0.0001 6.8 (4.7, 12.6) Degree of reliefe % of patients reporting “Somewhat Relieved,” “Considerably Relieved,” or “Completely Relieved” for 100% of the weekly scores or “Considerably Relieved” or “Completely Relieved” for at least 50% of the weekly scoresd 24.3 41.2 16.9 <0.0001 5.9 (4.3, 9.5) BSFS, Bristol Stool Forms Scale; CI, confidence interval; CSBM, complete SBM; IBS, irritable bowel syndrome; ITT, intent-to-treat; NNT, number needed to treat; NRS, numerical rating scale; SBM, spontaneous bowel movement. a Secondary end point.

n the placebo group had a 70% increase in the urinary lactulose-to-mannitol (LAMA) ratio, a measure of intestinal permeability, whereas those receiving larazotide acetate had no change. Larazotide acetate was not detected in the plasma after supra-therapeutic doses, and no significant systemic toxicities were observed. This dose-ranging, exploratory study was designed to evaluate the effect of multiple doses of larazotide acetate in patients with celiac disease who were given a 2-week gluten challenge in an outpatient setting. In addition, we sought to evaluate the tolerability of multiple doses of larazotide acetate, to explore elements of the design of studies involving controlled gluten challenges, and to gain experience with experimental outcome measures for drug development in celiac disease. METHODS The study protocol was approved by the institutional review boards of the participating institutions and all patients provided written informed consent (ClinicalTrials. gov Identifier: NCT00362856). The study was conducted at 10 clinical sites in the United States between August 2006 and March 2007. Patients were recruited by the investigators directly and through peer-to-peer communication.

on severity score (1–5) 3.1 2.6 Change from baseline, meanb,c −0.6 −1.2 −0.6 <0.0001 % of patients with decrease of ≥1 for at least 6/12 weeksd 42.5 59.5 17.0 <0.0001 5.9 (4.2, 9.9) IBS severity Mean IBS severity score (1–5) 3.1 2.7 Change from baseline, meanb,c −0.5 −1.0 −0.5 <0.0001 % of patients with decrease of ≥1 for at least 6/12 weeksd 37.5 56.3 18.8 <0.0001 5.3 (3.9, 8.3) Adequate relief % of patients reporting adequate relief of IBS symptoms for at least 75% of the weeks (i.e., 9/12 weeks)d 21.3 36.8 15.5 <0.0001 6.4 (4.6, 10.7) % of patients reporting adequate relief of IBS symptoms for at least 50% of the weeks (i.e., 6/12 weeks)d 34.2 48.9 14.7 <0.0001 6.8 (4.7, 12.6) Degree of reliefe % of patients reporting “Somewhat Relieved,” “Considerably Relieved,” or “Completely Relieved” for 100% of the weekly scores or “Considerably Relieved” or “Completely Relieved” for at least 50% of the weekly scoresd 24.3 41.2 16.9 <0.0001 5.9 (4.3, 9.5) BSFS, Bristol Stool Forms Scale; CI, confidence interval; CSBM, complete SBM; IBS, irritable bowel syndrome; ITT, intent-to-treat; NNT, number needed to treat; NRS, numerical rating scale; SBM, spontaneous bowel movement. a Secondary end point. b Changes from baseline are the least-squares means from the analysis of covariance (ANCOVA) model. c P values were based on a comparison of linaclotide vs. the placebo group using the ANCOVA model. d P values were based on a comparison of linaclotide vs. the placebo group using the Cochran–Mantel–Haenszel test.

a Secondary end point. b Changes from baseline are the least-squares means from the analysis of covariance (ANCOVA) model. c P values were based on a comparison of linaclotide vs. the placebo group using the ANCOVA model. d P values were based on a comparison of linaclotide vs. the placebo group using the Cochran–Mantel–Haenszel test. e Degree of Relief scale: 1=completely relieved; 2=considerably relieved; 3=somewhat relieved; 4=unchanged; 5=somewhat worse; 6=considerably worse; 7=as bad as I can imagine. Table 4 Treatment-emergent adverse events (safety population) Adverse event (preferred term) Placebo (N=396), n (%) Linaclotide 290 μg (N=406), n (%) P value Patients with at least 1 TEAE 210 (53.0) 228 (56.2) 0.3949 Diarrhea 14 (3.5) 79 (19.5) <0.0001 Abdominal pain 10 (2.5) 22 (5.4) 0.0462 Flatulence 6 (1.5) 20 (4.9) 0.0084 Headache 14 (3.5) 20 (4.9) 0.3825 Abdominal distension 3 (0.8) 9 (2.2) 0.1434 Treatment-emergent adverse events (TEAEs) reported in ≥2% of linaclotide-treated patients and at an incidence greater than reported in placebo-treated patients during the treatment period. P value was based on a Fisher's exact test comparing linaclotide and placebo.

INTRODUCTION Celiac disease is one of the most common autoimmune disorders, affecting 1% of individuals in many regions (1,2,3,4). Subjects with celiac disease frequently present with intestinal symptoms such as diarrhea, abdominal pain and bloating, and may also experience extraintestinal signs (5,6,7,8,9). Severe complications of celiac disease include gastrointestinal carcinoma and T-cell lymphoma, which may develop owing to chronic inflammation and sustained activation of intestinal lymphocytes and T cells (10). Disease activity is triggered and sustained by the entry of gluten peptides into the lamina propria of the intestine after crossing the epithelial barrier. Gluten, an amorphous mixture of proteins found in the endosperm of cereals like wheat, rye, and barley, is a major component of the human diet. In the lamina propria, tissue transglutaminase (tTG) modifies gluten peptides and potentiates their immunogenicity. These events subsequently trigger both T-helper type 1–mediated immune responses (11).

re of proteins found in the endosperm of cereals like wheat, rye, and barley, is a major component of the human diet. In the lamina propria, tissue transglutaminase (tTG) modifies gluten peptides and potentiates their immunogenicity. These events subsequently trigger both T-helper type 1–mediated immune responses (11). Currently, the only management option for patients with celiac disease is strict adherence to a gluten-free diet. Adherence to this highly restrictive diet is difficult due to the pervasiveness of gluten in foods. Patients maintaining a gluten-free diet may still be inadvertently exposed to up to 2 g per day of gluten (12,13,14,15). Exposure even to small amounts (i.e., 50 mg per day) can trigger signs and symptoms of celiac disease (4,14,16). Consequently, even after long-term maintenance of a gluten-free diet, many patients still have symptoms and/or mucosal damage (2,17,18,19). Therefore, a gluten-free diet alone may be insufficient to fully control the disease in some patients, and safe and effective pharmacological therapy are needed (20,21).

(4,14,16). Consequently, even after long-term maintenance of a gluten-free diet, many patients still have symptoms and/or mucosal damage (2,17,18,19). Therefore, a gluten-free diet alone may be insufficient to fully control the disease in some patients, and safe and effective pharmacological therapy are needed (20,21). Immune responses in patients with celiac disease are initiated when immunogenic, incompletely digested gluten peptides gain entry into the lamina propria of the small intestine by transcellular transport (22,23,24,25) and/or through the paracellular space between epithelial cells (23,24). In transcellular transport, partially degraded gliadin moves through epithelial cells in an immunoglobulin-mediated process, making them available for antigen presentation (24). Paracellular transport of gliadin peptides occurs in the setting of increased paracellular permeability in patients with celiac disease due to gliadin-induced innate and adaptive immune responses (11,26,27,28) and subsequent tight junction disassembly (28,29,30). In addition, genetic defects have been identified in the cytoskeletal proteins involved in tight junction functioning (31,32,33).

ked immunosorbent assay) Units (EU) for IgA and IgG, with 10 EU as the cutoff level for a positive test result). Women must have been post-menopausal or surgically sterile or must have had a negative result on the serum β-human chorionic gonadotropin pregnancy test and agreed to use acceptable methods of contraception. Patients were excluded if they had any food intolerances or allergies other than to gluten that would have interfered with the conduct of the study, had any chronic active gastrointestinal disease other than celiac disease, had diabetes mellitus, were receiving any medications that could have interfered with LAMA testing or other study measures, used nicotine-containing products for 6 months before study entry, had clinically significant abnormal laboratory test results at the time of screening, had significant comorbidities (including positive HIV, hepatitis B surface antigen, or hepatitis C test results), were pregnant or breast-feeding mothers, had donated blood within 56 days of randomization or plasma within 7 days, had abused alcohol or drugs within 2 years of randomization, had a positive urine drug test result at screening, or had participated in any clinical trial with an active drug within 30 days of randomization. Alcohol consumption of ≥3 fluid ounces within 48 h of producing urine samples for LAMA ratio testing was not permitted.

sed alcohol or drugs within 2 years of randomization, had a positive urine drug test result at screening, or had participated in any clinical trial with an active drug within 30 days of randomization. Alcohol consumption of ≥3 fluid ounces within 48 h of producing urine samples for LAMA ratio testing was not permitted. Study design and dosing In this prospective, multicenter, double-blind trial, 86 patients were randomly assigned to one of seven treatment groups (Figure 1a). Patients in four groups received a gluten challenge along with doses of 0.25, 1, 4, or 8 mg larazotide acetate three times daily. The other three groups were a safety control arm with the highest dose of larazotide acetate (8 mg) and no gluten challenge, a gluten-free control arm of patients in remission (placebo drug/placebo gluten), and a gluten challenge control arm (placebo drug/gluten challenge). The study treatment regimen started with breakfast on day 1 and continued through day 14. Patients ingested study medication capsules three times daily 15 min before meals. Larazotide acetate capsules contained enteric-coated multi-particulate beads. Placebo drug was provided in similarly colored capsules, with beads that were composed like those of the active drug except for the absence of larazotide acetate. Neither larazotide acetate nor placebo drug capsules contained gluten.

e meals. Larazotide acetate capsules contained enteric-coated multi-particulate beads. Placebo drug was provided in similarly colored capsules, with beads that were composed like those of the active drug except for the absence of larazotide acetate. Neither larazotide acetate nor placebo drug capsules contained gluten. The gluten challenge consisted of two capsules of gluten 400 mg (Amgluten 160 powder, Tate and Lyle of Decatur, Illinois, consisting of 45% gliadin, 45% glutenins, and 10% globulins). Matching placebo gluten capsules consisted of 100% cornstarch. Gluten challenge or placebo was ingested three times daily during meals. Thus, the gluten challenge consisted of a total of 2.4 g per day. Patients were asked to eat three meals per day and remain on a gluten-free diet during the study. Study conduct and assessments This study was performed according to good clinical practice guidelines established by the International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use. There were five clinic visits: screening and days 0, 7, 14, and 21.

The gluten challenge consisted of two capsules of gluten 400 mg (Amgluten 160 powder, Tate and Lyle of Decatur, Illinois, consisting of 45% gliadin, 45% glutenins, and 10% globulins). Matching placebo gluten capsules consisted of 100% cornstarch. Gluten challenge or placebo was ingested three times daily during meals. Thus, the gluten challenge consisted of a total of 2.4 g per day. Patients were asked to eat three meals per day and remain on a gluten-free diet during the study. Study conduct and assessments This study was performed according to good clinical practice guidelines established by the International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use. There were five clinic visits: screening and days 0, 7, 14, and 21. An experimental biomarker, the urinary LAMA fractional excretion ratio, was used to quantify changes in intestinal permeability. In patients with celiac disease, mucosal injury leads to a simultaneous reduction in the transmembrane absorption of monosaccharides (e.g., mannitol) and an increase in the paracellular absorption of disaccharides (e.g., lactulose), resulting in an increase in the LAMA ratio. A probe solution containing 7.5 g lactulose and 2 g mannitol in approximately 100 ml of water was administered orally in the clinic on day 0 and in the evening before the study visits on days 7, 14, and 21. On the day-0 collection in the clinic, urine was collected for 6 h. Patients fasted for at least 4 h before drinking the sugar solution and did not eat or drink (except water) until the end of the 6-h collection. For the overnight collections before days 7, 14, and 21, patients were asked to have a normal dinner around 1800hrs, not to eat or drink afterward (except water until 2200 hrs), void completely at 2200 hrs, drink the sugar solution, fast overnight (drinking water was permitted), and collect their overnight and morning urine. Urine samples were frozen at the clinic and stored for analysis of lactulose and mannitol using standardized methods (Dionex MA-1 ion exchange column with pulsed amperometric detection on a Dionex Ion Chromatograph 3000, Thermo Scientific, Sunnyvale, CA) (38).

permitted), and collect their overnight and morning urine. Urine samples were frozen at the clinic and stored for analysis of lactulose and mannitol using standardized methods (Dionex MA-1 ion exchange column with pulsed amperometric detection on a Dionex Ion Chromatograph 3000, Thermo Scientific, Sunnyvale, CA) (38). Serum was analyzed at a central laboratory for antibodies to tTG at screening and day 21 (ACM Medical Laboratory, Rochester, New York). Patients completed the Gastrointestinal Symptom Rating Scale (GSRS) (39,40,41) and the Psychological General Well-Being Index (42,43) on days 0, 7, 14, and 21. The GSRS is a widely used questionnaire completed by the patient to assess gastrointestinal symptom severity. The instrument consists of 15 questions that are grouped into five domains (diarrhea, abdominal pain, indigestion, constipation, and reflux). In a post hoc analysis, the three domains that are most relevant to celiac disease (diarrhea, abdominal pain, and indigestion) were evaluated separately, hereafter referred to as the Celiac Disease GSRS (CeD-GSRS). Both the GSRS (42,44,45) and the Psychological General Well-Being Index (42,46,47) have been previously used to evaluate patients with celiac disease. However, the CeD-GSRS has not been previously used.

ominal pain, and indigestion) were evaluated separately, hereafter referred to as the Celiac Disease GSRS (CeD-GSRS). Both the GSRS (42,44,45) and the Psychological General Well-Being Index (42,46,47) have been previously used to evaluate patients with celiac disease. However, the CeD-GSRS has not been previously used. Tolerability was evaluated at each visit by adverse event surveillance, measurement of vital signs, and clinical laboratory analysis for blood chemistry, hematology, and urinalysis. Twelve-lead electrocardiograms were performed at screening and on days 0 and 21. Plasma levels of larazotide acetate and metabolites were measured on days 0, 7, and 14 using validated high-performance liquid chromatography with tandem mass spectrometry, with a lower limit of quantification of 0.5 ng/ml. In order to differentiate adverse events that were related to the gluten challenge from those that were related to study medication, the following events were identified a priori as the signs and symptoms of gluten toxicity: abdominal discomfort, dyspepsia, nausea, diarrhea, vomiting, flatulence, constipation, and rash/dermatitis herpetiformis. These events were evaluated as part of the efficacy assessment and were not included in the assessment of safety.

lowing events were identified a priori as the signs and symptoms of gluten toxicity: abdominal discomfort, dyspepsia, nausea, diarrhea, vomiting, flatulence, constipation, and rash/dermatitis herpetiformis. These events were evaluated as part of the efficacy assessment and were not included in the assessment of safety. Outcome measures and statistical analysis The primary outcome measure was the change in LAMA ratios from day 0 to day 14, which was calculated by dividing the LAMA ratio at day 14 by the LAMA ratio at day 0. This calculation results in a fold-ratio that indicates the intestinal permeability, with a ratio of 1 indicating no change in permeability, a ratio >1 indicating an increase in permeability, and a ratio <1 indicating a reduction. LAMA values were log-transformed for the analysis and sample-size calculations that made it possible to use standard sample-size calculation methods for determining the differences between the treatment groups. Because of the log transformation, LAMA values were summarized using the geometric mean. All analyses presented in this report were conducted using the intent-to-treat population. All analyses were pre-specified unless otherwise indicated. Safety analyses were descriptive in nature.

Outcome measures and statistical analysis The primary outcome measure was the change in LAMA ratios from day 0 to day 14, which was calculated by dividing the LAMA ratio at day 14 by the LAMA ratio at day 0. This calculation results in a fold-ratio that indicates the intestinal permeability, with a ratio of 1 indicating no change in permeability, a ratio >1 indicating an increase in permeability, and a ratio <1 indicating a reduction. LAMA values were log-transformed for the analysis and sample-size calculations that made it possible to use standard sample-size calculation methods for determining the differences between the treatment groups. Because of the log transformation, LAMA values were summarized using the geometric mean. All analyses presented in this report were conducted using the intent-to-treat population. All analyses were pre-specified unless otherwise indicated. Safety analyses were descriptive in nature. In the sample-size calculation, it was estimated that eight evaluable patients per group would have provided 80% power to demonstrate a statistically significant difference (α=0.05) between the gluten-challenge control group and the larazotide acetate groups that received the gluten challenge. An additional 3–5 patients per group was projected to allow for screen failures and patients who withdrew during the study (27–45%). The effect size for this outcome was set a priori at a threefold reduction in LAMA ratios between any test group and placebo. The sample size calculation was based on a t test for differences between two groups. The standard deviation for log-transformed fold-ratios was assumed to be 0.3, based on the results of a previous study (37).

for this outcome was set a priori at a threefold reduction in LAMA ratios between any test group and placebo. The sample size calculation was based on a t test for differences between two groups. The standard deviation for log-transformed fold-ratios was assumed to be 0.3, based on the results of a previous study (37). RESULTS Demography and patient disposition Eighty-six patients signed the informed consent, were randomly assigned to treatment, and were included in the intent-to-treat analysis (Figure 1b). Fifty-three percent were women, and the mean age was 46.3 years. All patients except one were Caucasian. Eighty patients (93.0%) completed the study. Six patients discontinued the study prematurely: 1 in the larazotide acetate 8 mg/placebo gluten arm (because of abdominal discomfort), 2 in the placebo drug/gluten-challenge arm (1 positive drug screen and 1 allergic reaction), 2 in the larazotide acetate 1 mg/gluten-challenge arm (headache, nausea, flatulence, and diarrhea), and 1 in the larazotide acetate 4 mg/gluten-challenge arm (non-compliance with intestinal permeability tests and questionnaires).

in the placebo drug/gluten-challenge arm (1 positive drug screen and 1 allergic reaction), 2 in the larazotide acetate 1 mg/gluten-challenge arm (headache, nausea, flatulence, and diarrhea), and 1 in the larazotide acetate 4 mg/gluten-challenge arm (non-compliance with intestinal permeability tests and questionnaires). Primary efficacy outcome: LAMA ratio The geometric mean change in the LAMA ratio in the gluten-challenge control group was greater than that of patients in the gluten-free control group, although the difference between the groups was not statistically significant (Figure 2). Among the patients who received the gluten challenge, no statistically significant differences in the LAMA ratios were observed between the larazotide acetate groups and the placebo group (Figure 2). The LAMA ratios varied widely. Geometric mean LAMA fold-ratios numerically decreased between baseline and day 7 in the gluten-free control groups and the gluten-challenge control group (see Supplementary Figure S1 online). Secondary efficacy outcomes The changes in total GSRS and the CeD-GSRS for the each treatment group from baseline to day 14 are shown in Figure 3. Gastrointestinal symptoms were stable in the two treatment groups that received placebo gluten, but symptoms grew more severe in the gluten-challenge control group.

Primary efficacy outcome: LAMA ratio The geometric mean change in the LAMA ratio in the gluten-challenge control group was greater than that of patients in the gluten-free control group, although the difference between the groups was not statistically significant (Figure 2). Among the patients who received the gluten challenge, no statistically significant differences in the LAMA ratios were observed between the larazotide acetate groups and the placebo group (Figure 2). The LAMA ratios varied widely. Geometric mean LAMA fold-ratios numerically decreased between baseline and day 7 in the gluten-free control groups and the gluten-challenge control group (see Supplementary Figure S1 online). Secondary efficacy outcomes The changes in total GSRS and the CeD-GSRS for the each treatment group from baseline to day 14 are shown in Figure 3. Gastrointestinal symptoms were stable in the two treatment groups that received placebo gluten, but symptoms grew more severe in the gluten-challenge control group. Gastrointestinal symptoms in patients who received the larazotide acetate and the gluten challenge did not become as severe as they did in the gluten-challenge control group. The 0.25 and 4.0 mg doses of larazotide acetate showed statistically significant prevention of gastrointestinal symptom severity worsening. The 4- and 8-mg doses of larazotide acetate did not show statistically significant prevention of gastrointestinal symptom severity worsening (P=0.067 and 0.329, respectively). Results for the CeD-GSRS were generally similar to those of the total GSRS.

gnificant prevention of gastrointestinal symptom severity worsening. The 4- and 8-mg doses of larazotide acetate did not show statistically significant prevention of gastrointestinal symptom severity worsening (P=0.067 and 0.329, respectively). Results for the CeD-GSRS were generally similar to those of the total GSRS. Because the individual treatment groups were small (≤13 patients per group) and there were similar trends among the groups receiving larazotide acetate and the gluten challenge, the groups were combined into a single active treatment group (n=48) to make possible further analysis of symptoms and simplify data presentation for subsequent exploratory analyses. Likewise, the two groups receiving placebo gluten along with either placebo drug or 8 mg larazotide acetate were pooled into an aggregate gluten-free control group (n=20). The gluten-challenge control group (n=13) indicates the response to gluten ingestion in the absence of larazotide acetate.

ent exploratory analyses. Likewise, the two groups receiving placebo gluten along with either placebo drug or 8 mg larazotide acetate were pooled into an aggregate gluten-free control group (n=20). The gluten-challenge control group (n=13) indicates the response to gluten ingestion in the absence of larazotide acetate. Figure 4 shows the time course of the change in the GSRS scores for these combined treatment groups. In the gluten-challenge control group, the severity of gastrointestinal symptoms increased substantially during the gluten-challenge period (baseline to day 14), as measured by both the total GSRS and the CeD-GSRS. The scores rapidly decreased after the challenge was over (day 14 to day 21). By contrast, there was no increase in severity in the groups that did not receive the gluten challenge (gluten-free control), and only a modest increase in the patients who received larazotide acetate and the gluten challenge. Symptoms were statistically significantly less severe in both the gluten-free control group and the group consisting of patients who received larazotide acetate and the gluten challenge compared with the gluten-challenge control group (P<0.05). Thus, an increase in symptom severity in response to gluten, as well as protection from the exposure to gluten provided by larazotide acetate, was demonstrated using both the GSRS and the CeD-GSRS in this study.

eived larazotide acetate and the gluten challenge compared with the gluten-challenge control group (P<0.05). Thus, an increase in symptom severity in response to gluten, as well as protection from the exposure to gluten provided by larazotide acetate, was demonstrated using both the GSRS and the CeD-GSRS in this study. To complement the analysis of the effect of larazotide acetate on symptom severity, each sub-domain of the GSRS (diarrhea, abdominal pain, indigestion, constipation, and reflux) was analyzed individually (see Supplementary Figure S2 online). Based on a comparison of the gluten-challenge control to the groups not receiving gluten, the gluten challenge resulted in an increase in the severity of diarrhea, abdominal pain, and indigestion, with the largest increase in symptom severity in the indigestion domain. Larazotide acetate provided a statistically significant protection against an increase in severity of abdominal pain, indigestion, and reflux, while the protection against an increase in the severity of diarrhea nearly reached statistical significance (P=0.051). Although there was an increase in the severity of both the constipation and reflux domains in the gluten-challenge control group, these increases did not appear to be specifically attributable to gluten, as there were similar increases in the gluten-free control group. No statistically significant difference between the larazotide acetate groups and the gluten-challenge control group was observed in changes in the total Psychological General Well-Being Index scores (see Supplementary Table 1 online).

To complement the analysis of the effect of larazotide acetate on symptom severity, each sub-domain of the GSRS (diarrhea, abdominal pain, indigestion, constipation, and reflux) was analyzed individually (see Supplementary Figure S2 online). Based on a comparison of the gluten-challenge control to the groups not receiving gluten, the gluten challenge resulted in an increase in the severity of diarrhea, abdominal pain, and indigestion, with the largest increase in symptom severity in the indigestion domain. Larazotide acetate provided a statistically significant protection against an increase in severity of abdominal pain, indigestion, and reflux, while the protection against an increase in the severity of diarrhea nearly reached statistical significance (P=0.051). Although there was an increase in the severity of both the constipation and reflux domains in the gluten-challenge control group, these increases did not appear to be specifically attributable to gluten, as there were similar increases in the gluten-free control group. No statistically significant difference between the larazotide acetate groups and the gluten-challenge control group was observed in changes in the total Psychological General Well-Being Index scores (see Supplementary Table 1 online). Changes in the weekly means of the number of bowel movements, number of episodes of diarrhea, stool consistency rating, and abdominal discomfort scores from daily bowel diaries are provided in Supplementary Figure S3 online. In general, daily bowel diary scores for patients in the larazotide acetate 1-mg and 0.25-mg groups did not increase as much as those in the other groups exposed to gluten.

des of diarrhea, stool consistency rating, and abdominal discomfort scores from daily bowel diaries are provided in Supplementary Figure S3 online. In general, daily bowel diary scores for patients in the larazotide acetate 1-mg and 0.25-mg groups did not increase as much as those in the other groups exposed to gluten. At two of four tested doses, larazotide acetate appeared to protect patients from the signs and symptoms of gluten toxicity, which were defined a priori: abdominal discomfort, dyspepsia, nausea, diarrhea, vomiting, flatulence, constipation, or rash/dermatitis herpetiformis. Compared with the gluten-challenge control, in which 50.0% of patients experienced symptoms of gluten toxicity, 20.8% of participants in both the aggregated gluten-free control and the larazotide acetate treatment arms exhibited these symptoms (P=0.11 and P=0.046, respectively; Fisher's exact test). After other signs and symptoms that were determined post hoc likely to be related to gluten toxicity were included (namely, elevated liver function tests and aphthous stomatitis), 64.3% of participants in the gluten-challenge control were affected, compared with 16.7% in the aggregated gluten-free control and 23.3% in the aggregated larazotide acetate treatment arms (P=0.021 and P=0.008, respectively; Fisher's exact test).

included (namely, elevated liver function tests and aphthous stomatitis), 64.3% of participants in the gluten-challenge control were affected, compared with 16.7% in the aggregated gluten-free control and 23.3% in the aggregated larazotide acetate treatment arms (P=0.021 and P=0.008, respectively; Fisher's exact test). Mean titers at screening for antibodies to tTG ranged from 1.6 to 3.0 U/ml, and no significant differences between the treatment groups were observed at screening. No significant differences between the treatment groups were observed in the mean changes from screening to day 21 in the titers of antibodies to tTG (−0.1 U/ml in the combined gluten-free control group, 2.9 U/ml in the gluten-challenge control group (P=0.204 vs. gluten-free control group), and 0.9 U/ml in the combined larazotide acetate/gluten-challenge groups (P=0.256 vs. gluten-challenge control group).

from screening to day 21 in the titers of antibodies to tTG (−0.1 U/ml in the combined gluten-free control group, 2.9 U/ml in the gluten-challenge control group (P=0.204 vs. gluten-free control group), and 0.9 U/ml in the combined larazotide acetate/gluten-challenge groups (P=0.256 vs. gluten-challenge control group). Tolerability The multiple oral doses of larazotide acetate used in this study were well tolerated. Of the 86 patients in the intent-to-treat population, 44 (51.2%) had at least one treatment-emergent adverse event ( Table 1). Fifteen of the 60 patients who received larazotide acetate (25.0%) had adverse events that the investigator considered to be related to study medication, while 7 of the 26 patients who received placebo drug (26.9%) had such events. Headache was the most common adverse event, reported by 17 patients, with no differences among treatment groups. None of the patients had serious adverse events. There were also no clinically significant findings of hepatic, bone, or renal toxicity. Vital signs, electrocardiogram, hematology, and clinical chemistry parameters showed no changes from baseline that were considered clinically significant.

nces among treatment groups. None of the patients had serious adverse events. There were also no clinically significant findings of hepatic, bone, or renal toxicity. Vital signs, electrocardiogram, hematology, and clinical chemistry parameters showed no changes from baseline that were considered clinically significant. Five (5.8%) patients had urinary tract infections, all of which occurred in patients who received larazotide acetate and the gluten challenge. There was no evidence of a dosage effect or an association to the duration of dosing at the onset of the event. Microbiological confirmation was attempted in only one of these events, and the result was negative. Only one patient received antibiotic medication. All cases were mild, and none were considered to be related to study drug. Plasma levels of larazotide acetate were measured at days 0, 7, and 14 and were below the limit of quantification (0.5 ng/ml) in all the groups. Metabolites were measured in the highest dose group (8 mg) and were also below the lower limit of quantification.

Five (5.8%) patients had urinary tract infections, all of which occurred in patients who received larazotide acetate and the gluten challenge. There was no evidence of a dosage effect or an association to the duration of dosing at the onset of the event. Microbiological confirmation was attempted in only one of these events, and the result was negative. Only one patient received antibiotic medication. All cases were mild, and none were considered to be related to study drug. Plasma levels of larazotide acetate were measured at days 0, 7, and 14 and were below the limit of quantification (0.5 ng/ml) in all the groups. Metabolites were measured in the highest dose group (8 mg) and were also below the lower limit of quantification. DISCUSSION The purpose of the study was to evaluate the efficacy and tolerability of multiple doses of larazotide acetate in preventing the exacerbation of symptoms after a gluten challenge in patients with celiac disease well controlled through a gluten-free diet. The primary efficacy outcome was the LAMA ratio, which was used as an experimental biomarker for intestinal permeability. The results indicate that the gluten challenge induced both an increase in symptoms as measured by the GSRS and intestinal permeability in patients who received placebo drug, as indicated by an increase in the LAMA ratio; however, this increase was not statistically significant. LAMA ratios for patients who received larazotide acetate and the gluten challenge were also not statistically significantly different from those who received placebo and the gluten challenge. The significant deterioration in gastrointestinal symptoms in the group receiving gluten and placebo compared with the group receiving no gluten suggests that 2.4 g of gluten per day is sufficient to induce measurable and clinically important symptoms in clinical trial settings.

ceived placebo and the gluten challenge. The significant deterioration in gastrointestinal symptoms in the group receiving gluten and placebo compared with the group receiving no gluten suggests that 2.4 g of gluten per day is sufficient to induce measurable and clinically important symptoms in clinical trial settings. LAMA ratio values varied markedly among all treatment groups, which may be the result of conducting the measurements in an outpatient setting. In addition, a decrease in LAMA was observed in most treatment groups at day 7, including the gluten-free control group, in which patients received both placebo drug and placebo gluten. These results suggest that there may have been a study effect, by virtue of which patients were more compliant with their gluten-free diets while participating in the study. Future studies should include a run-in period for diet stabilization. In addition, the significant variability in the LAMA ratios in the outpatient setting should be taken into account in future studies utilizing this outcome. Sample size calculations used in the design of this study were based on one preliminary inpatient trial in which the effects of gluten challenge and drug treatment were larger than those observed in this study. Although we cannot exclude the possibility that the primary efficacy outcome in this study was not reached because the treatment did not produce an effect, it is likely that the methods used did not fully detect an effect of both the gluten challenge and the larazotide acetate therapy.

han those observed in this study. Although we cannot exclude the possibility that the primary efficacy outcome in this study was not reached because the treatment did not produce an effect, it is likely that the methods used did not fully detect an effect of both the gluten challenge and the larazotide acetate therapy. In contrast to the LAMA findings, the results of gastrointestinal symptom severity assessments indicate that larazotide acetate provided protection against the increase in symptoms associated with a gluten challenge in patients with celiac disease. Both the prespecified total GSRS and the exploratory CeD-GSRS scores of patients who received placebo drug increased following the gluten challenge. However, at two of four tested doses (0.25 and 4 mg), patients who received larazotide acetate and the gluten challenge did not show such increases in GSRS and CeD-GSRS scores. Symptom severity was statistically significantly lower than placebo drug in these treatment groups. As expected, most patients who received the gluten challenge, including those in the gluten-challenge control group, did not develop antibodies to tTG. Thus, the effect of larazotide acetate on antibodies to tTG could not be assessed. Longer gluten challenge studies and/or higher gluten doses will be necessary to assess the effect of the larazotide acetate on the development of antibodies to tTG.

n the gluten-challenge control group, did not develop antibodies to tTG. Thus, the effect of larazotide acetate on antibodies to tTG could not be assessed. Longer gluten challenge studies and/or higher gluten doses will be necessary to assess the effect of the larazotide acetate on the development of antibodies to tTG. In conclusion, larazotide acetate was well tolerated in this population of patients with celiac disease. Larazotide acetate prevented the increase in gastrointestinal symptom severity induced by gluten challenge at two of the four doses tested. A gluten challenge of 2.4 g per day was sufficient to evaluate this outcome. Additional studies with appropriate run-in periods, larger study populations, additional measures such as histology, and longer duration may be warranted to further evaluate the effect of larazotide acetate in patients with celiac disease. STUDY HIGHLIGHTS

In conclusion, larazotide acetate was well tolerated in this population of patients with celiac disease. Larazotide acetate prevented the increase in gastrointestinal symptom severity induced by gluten challenge at two of the four doses tested. A gluten challenge of 2.4 g per day was sufficient to evaluate this outcome. Additional studies with appropriate run-in periods, larger study populations, additional measures such as histology, and longer duration may be warranted to further evaluate the effect of larazotide acetate in patients with celiac disease. STUDY HIGHLIGHTS We thank John Jiang, PhD, and Joel Verter for statistical analyses, and Betsy Abraham-Van Parijs, MD, PhD, for comments and suggestions. Scott Newcomer of Cephalon helped prepare the manuscript but did not meet the criteria for authorship. The study coordinators for the investigating co-authors were Carol Van Dyke, Melinda Dennis, BJ Bahl, and Roberta O'Shea. We thank the patients who participated in this trial and the following clinicians and study coordinators: DiMarino A, MD, Moretti D, Miller C, and Wilson K, Thomas Jefferson University, Philadelphia, PA; Kirby D, MD, and Tucker P, VCU Medical Center, Richmond, VA; Medoff J, MD, and Uhl A, Vital re:Search, Greensboro, NC; DeLegge M and Davis CT, Medical University of South Carolina, Charleston, SC; Schuman R, MD, and Dalcomine C, Affiliates in GastroEnterology, Morristown, NJ; Harris L, MD, and Menghini M, Mayo Clinic, Scottsdale, AZ; Pressman MD and Tuohy D, Medical Associates Research Group, San Diego, CA.

l re:Search, Greensboro, NC; DeLegge M and Davis CT, Medical University of South Carolina, Charleston, SC; Schuman R, MD, and Dalcomine C, Affiliates in GastroEnterology, Morristown, NJ; Harris L, MD, and Menghini M, Mayo Clinic, Scottsdale, AZ; Pressman MD and Tuohy D, Medical Associates Research Group, San Diego, CA. SUPPLEMENTARY MATERIAL is linked to the online version of the paper at http://www.nature.com/ajg Guarantor of the article: J.A. Murray, MD. Specific author contributions: Study design and execution, patient recruitment, analysis of data, and writing of manuscript: Daniel A. Leffler, C.P. Kelly, and J.A. Murray; patient recruitment and writing of manuscript: H.Z. Abdallah and A.M. Colatrella; study design and execution, analysis of data, and writing of manuscript: L.A. Harris, F. Leon, L.A. Arterburn, and Z.H. Lan; study design and execution, and writing of manuscript: B.M. Paterson. Financial support: This study was funded by Alba Therapeutics Corporation, Baltimore, Maryland, USA. Potential competing interests: Daniel A. Leffler is a consultant for Alba Therapeutics. C.P. Kelly is a consultant for Alba Therapeutics, Alvine Pharmaceuticals, and ImmunosanT. F. Leon, L.A. Arterburn, and Z.H. Lan were formerly employed by Alba Therapeutics. B.M. Paterson was the former CEO and founder and owns stock in Alba Therapeutics. The other authors declare no conflict of interest. Supplementary Material Supplementary Information Click here for additional data file.

Potential competing interests: Daniel A. Leffler is a consultant for Alba Therapeutics. C.P. Kelly is a consultant for Alba Therapeutics, Alvine Pharmaceuticals, and ImmunosanT. F. Leon, L.A. Arterburn, and Z.H. Lan were formerly employed by Alba Therapeutics. B.M. Paterson was the former CEO and founder and owns stock in Alba Therapeutics. The other authors declare no conflict of interest. Supplementary Material Supplementary Information Click here for additional data file. Figure 1 Schema of overall study design and participant allocation. (a) Study design and (b) disposition of patients. LAMA, lactulose/mannitol. Figure 2 Change from baseline LAMA levels in the individual treatment groups. Mean change from baseline in the urinary lactulose-to-mannitol (LAMA) ratio. Values are the geometric mean fold-ratio on day 14 over baseline (day 0). Vertical bars represent 95% confidence intervals. Dashed horizontal line indicates level for gluten-free control. G, gluten; P, placebo. Figure 3 Change from baseline in gastrointestinal symptoms in the individual treatment groups as measured by the GSRS and CeD-GSRS. Mean (95% confidence interval) changes from baseline to day 14 in the total (a) Gastrointestinal Symptom Rating Scale (GSRS) scores and (b) Celiac Disease GSRS (CeD-GSRS) scores. n=9–13 per group. P values comparing larazotide acetate/gluten-challenge (G) groups to the placebo (P) drug/gluten-challenge group were calculated using an analysis of covariance model, with treatment as a fixed effect and the corresponding baseline value as a covariate.

s and (b) Celiac Disease GSRS (CeD-GSRS) scores. n=9–13 per group. P values comparing larazotide acetate/gluten-challenge (G) groups to the placebo (P) drug/gluten-challenge group were calculated using an analysis of covariance model, with treatment as a fixed effect and the corresponding baseline value as a covariate. Figure 4 Change in gastrointestinal symptoms as measured by the GSRS and CeD-GSRS during and after gluten challenge in the gluten challenge/placebo drug cohort compared to the aggregated gluten-free control cohorts and the Larazotide-treated gluten challenge groups. Time course of mean change from baseline in the total (a) Gastrointestinal Symptom Rating Scale (GSRS) score and (b) Celiac Disease (CeD-GSRS) scores. The gluten control group includes patients who received placebo drug and the gluten challenge (n=13). Gluten-free control includes patients who received placebo drug or 8 mg larazotide acetate and gluten placebo (n=20). Patients who received larazotide acetate and the gluten challenge were also combined into one group (n=38). P values were calculated using an analysis of covariance model with treatment as a fixed effect and the corresponding baseline value as a covariate.

cebo drug or 8 mg larazotide acetate and gluten placebo (n=20). Patients who received larazotide acetate and the gluten challenge were also combined into one group (n=38). P values were calculated using an analysis of covariance model with treatment as a fixed effect and the corresponding baseline value as a covariate. Table 1 Summary of adverse events Category Gluten-free control Gluten challenge Placebo (n=12) 8 mg (n=12) Placebo (n=14) 0.25 mg (n=12) 1 mg (n=12) 4 mg (n=12) 8 mg (n=12) Patients with ≥1 adverse event 5 (41.7) 4 (33.3) 7 (50.0) 7 (58.3) 8 (66.7) 6 (50.0) 7 (58.3) Patients with ≥1 adverse event related to study medicationa 2 (16.7) 1 (8.3) 5 (35.7) 3 (25.0) 4 (33.3) 3 (25.0) 4 (33.3) Patients with ≥1 severe adverse event 1 (8.3) 0 (0.0) 0 (0.0) 0 (0.0) 2 (16.7) 0 (0.0) 0 (0.0) Patients with ≥1 serious adverse events 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) Patients who discontinued study medication because of an adverse event 0 (0.0) 0 (0.0) 1 (7.1) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) Adverse events that occurred in ≥5% of patients Headache 4 (33.3) 0 (0.0) 3 (21.4) 3 (25.0) 2 (16.7) 3 (25.0) 2 (16.7) Urinary tract infection 0 (0.0) 0 (0.0) 0 (0.0) 2 (16.7) 1 (8.3) 1 (8.3) 1 (8.3) a Events that were considered by the investigator to be possibly or probably related to the study medication.

INTRODUCTION The pharmacological management of Crohn's disease (CD), an immune-mediated inflammatory disorder characterized by recurrent inflammation in the gastrointestinal tract, is determined by the location, extent, and severity of disease; disease-related complications (1); and patient preference related to the frequency and route of medication administration (2). In the past, therapeutic options were generally limited to aminosalicylates; antibiotics; immunosuppressive agents such as corticosteroids; and immunomodulators such as 6-mercaptopurine, azathioprine, and methotrexate. Over the past decade, tumor necrosis factor alpha (TNF-α) antagonists have become available as treatment options. Infliximab, adalimumab, and most recently certolizumab pegol are effective in inducing clinical remission and mucosal healing in patients with CD (3,4). When used in combination with immunomodulators, the early use of these biological therapies has been shown to induce remission more rapidly than the conventional “step-up” treatment, resulting in a more durable remission, decreased need for treatment with corticosteroids, and improved rates of mucosal healing (5).

,4). When used in combination with immunomodulators, the early use of these biological therapies has been shown to induce remission more rapidly than the conventional “step-up” treatment, resulting in a more durable remission, decreased need for treatment with corticosteroids, and improved rates of mucosal healing (5). All available therapies for CD have potential adverse side effects. For instance, corticosteroid therapy is associated with diabetes, hypertension, infection, and osteoporosis (6), and treatment with immunomodulators may be complicated by rare but serious side effects that can include bone marrow suppression, hepatotoxicity, opportunistic infection, an increased risk of lymphoma, and pancreatitis (6,7,8). Also, as both corticosteroids and immunomodulators have broad effects on the immune system, patients receiving such treatment at higher doses and/or for prolonged periods of time may be at increased risk for serious or opportunistic infections (9,10).

infection, an increased risk of lymphoma, and pancreatitis (6,7,8). Also, as both corticosteroids and immunomodulators have broad effects on the immune system, patients receiving such treatment at higher doses and/or for prolonged periods of time may be at increased risk for serious or opportunistic infections (9,10). Potential side effects of TNF-α antagonists include serious opportunistic infections, such as tuberculosis and histoplasmosis, which occur at a higher incidence when concomitant immunosuppression (particularly corticosteroids) is employed (11,12). The absolute incidence of opportunistic infections, however, is generally low, as are the incidences of other potential side effects of anti-TNF-α therapy, including malignancies, lymphomas, demyelinating disorders, and lupus-like disorders (11,13,14). The rare occurrence of hepatosplenic T-cell lymphomas in inflammatory bowel disease patients is associated with thiopurine immunomodulatory therapy (azathioprine or 6-mercaptopurine), used either as monotherapy or in combination with TNF-α antagonists (7,11,12).

elinating disorders, and lupus-like disorders (11,13,14). The rare occurrence of hepatosplenic T-cell lymphomas in inflammatory bowel disease patients is associated with thiopurine immunomodulatory therapy (azathioprine or 6-mercaptopurine), used either as monotherapy or in combination with TNF-α antagonists (7,11,12). Based on the continued need to answer important safety questions related to anti-TNF-α therapy, the Crohn's Therapy, Resource, Evaluation, and Assessment Tool (TREAT™) registry was undertaken. The TREAT registry is a large, observational registry that was designed to examine the long-term clinical, economic, and humanistic outcomes of various treatment regimens, including infliximab, used in the management of CD in community-based and academic practice settings in North America. We previously reported the initial safety findings derived from the TREAT registry after an average of 1.9 years of follow-up (15). Results of that earlier analysis indicated that mortality rates were similar for infliximab- and other-treatments-only-treated patients (0.53 vs. 0.43 per 100 patient-years; relative risk (RR) ratio =1.24, 95% confidence interval (CI) =0.73, 2.10; P=0.43). While results of an unadjusted analysis showed an increased risk of serious infection with infliximab treatment, results of multivariate logistic regression analysis suggested that infliximab was not an independent predictor of serious infection (RR=0.99, 95% CI=0.64, 1.54; P=0.97). Factors independently associated with serious infection included moderate-to-severe disease activity, prednisone treatment, and narcotic analgesic treatment (15). Although the TREAT registry is ongoing, patient enrollment into the TREAT registry is complete, and all enrolled patients have been followed for at least 5 years. Results of analyses reflecting data collected as part of the TREAT registry through February 2010 are reported herein.

and narcotic analgesic treatment (15). Although the TREAT registry is ongoing, patient enrollment into the TREAT registry is complete, and all enrolled patients have been followed for at least 5 years. Results of analyses reflecting data collected as part of the TREAT registry through February 2010 are reported herein. METHODS Study design The TREAT registry is an ongoing, prospective, observational, multicenter, long-term registry of North American patients with CD. The registry was initiated in 1999 to evaluate the clinical safety outcomes of various treatment regimens, including infliximab, in the management of CD. Approximately 350 gastroenterologists from both community-based and academic practice settings were each to enroll up to 150 patients, for a target enrollment of at least 5,000 patients. Physicians were generally identified from the membership list of the Crohn's and Colitis Foundation Association. Patients enrolled in the registry are treated at the discretion of their physicians (i.e., a treatment protocol is not predefined). At the start of the registry, it was stipulated that physician participation in the registry could be withdrawn if patient enrollment requirements were not met, complete data were not submitted, or a physician left the practice or elected to discontinue registry participation.

, a treatment protocol is not predefined). At the start of the registry, it was stipulated that physician participation in the registry could be withdrawn if patient enrollment requirements were not met, complete data were not submitted, or a physician left the practice or elected to discontinue registry participation. The design of the TREAT registry was approved by the institutional review boards at each participating site, and all patients provided written informed consent before participation in the registry. Physicians or their designees have been paid a small honorarium on a per patient basis as compensation for administering the registry. Janssen Biotech, Inc. (Spring House, PA), the manufacturer of infliximab, is the sponsor of the TREAT registry. The registry operates under the supervision and guidance of an Advisory Committee comprising several authors of this publication (GRL, BGF, RDC, BAS, and WJS) and the TREAT medical monitor (previously RHD and now XX). Registry participants Patients enrolled into the TREAT registry must have had a diagnosis of CD and could not be participating in any clinical trial. Upon TREAT's inception, both pediatric and adult patients were enrolled. However, the protocol was subsequently amended to limit enrollment to patients 18 years of age or older. The findings reported herein include limited data for 80 pediatric patients (defined as patients <18 years of age at the time of enrollment) who were subsequently discontinued early from the registry.

were enrolled. However, the protocol was subsequently amended to limit enrollment to patients 18 years of age or older. The findings reported herein include limited data for 80 pediatric patients (defined as patients <18 years of age at the time of enrollment) who were subsequently discontinued early from the registry. This report contains a summary of data collected for 6,273 patients, while a previous report cited 6,290 patients (15). Seventeen patients included in the previous report were found to be duplicates, i.e., sites either enrolled the same patient twice or treated a patient who was being seen at another TREAT registry site. These 17 patients are now excluded from the TREAT registry database. Registry evaluations In this registry, patient data were collected at enrollment and then on a semi-annual basis (each January and July). Patients enrolled in the registry were to be followed for a minimum of 5 years. Patient demographic information and physicians' assessments of overall patient health and disease severity were documented upon registry enrollment. Physicians assessed disease severity according to the American College of Gastroenterology Guidelines (16) as remission, mild-to-moderate, moderate-to-severe, or severe-to-fulminant.

ographic information and physicians' assessments of overall patient health and disease severity were documented upon registry enrollment. Physicians assessed disease severity according to the American College of Gastroenterology Guidelines (16) as remission, mild-to-moderate, moderate-to-severe, or severe-to-fulminant. A wide array of patient data were collected at the semi-annual follow-up visits, including disease severity, medication use, adverse events, and the dates and outcomes of each infliximab infusion. Participating TREAT physicians were asked to provide information regarding an expansive list of specific adverse events, including infusion reactions, infections, and malignancies. The following infusion reaction adverse event terms were tracked: arthritis, cardiopulmonary symptoms, chills, headache, hypertension, hypotension, influenza-like illness, muscle spasm, nausea, pyrexia, and rash. All other infusion reactions were categorized as “other.” In addition, serious adverse events (SAEs), lymphoma, tuberculosis, congestive heart failure, hypersensitivity reaction, interstitial lung disease, amyotrophic lateral sclerosis, and changes in pregnancy status were to be documented as they occurred. Physicians were required to report SAEs within 24 h of occurrence. A detailed description of the SAE was also required, including dates of onset and resolution, designation of whether the SAE was expected or unexpected, the seriousness and intensity of the SAE, the relationship of the SAE to the underlying CD, concomitant CD medications or surgical procedures for CD related to the SAE, and SAE treatment and outcome.

E was also required, including dates of onset and resolution, designation of whether the SAE was expected or unexpected, the seriousness and intensity of the SAE, the relationship of the SAE to the underlying CD, concomitant CD medications or surgical procedures for CD related to the SAE, and SAE treatment and outcome. Data analysis Data from both active and discontinued (using data available through the last patient contact) patients are included in the data analyses. The cohort of 3,420 patients that received infliximab within 12 weeks before enrollment, who were scheduled to receive infliximab within 30 days of enrollment, or who received infliximab at some other point in the registry is termed “infliximab-treated patients.” As it is not possible to document infliximab administration before participation in the registry, only patients who met the criteria outlined above are considered “infliximab-treated” for data related to serious infections, infliximab dosing, infusion reactions, and baseline patient/disease characteristics. Patients who did not receive infliximab during the registry are grouped as having received “other-treatments-only” in these analyses. For analyses involving mortality, an additional 344 patients with infliximab exposure >3 but <12 months before registry enrollment were also included to yield a total of 3,764 patients “ever-exposed” to infliximab. Baseline demographic and disease characteristics are summarized using descriptive statistics. As such, means and s.d.'s are employed for continuous variables (age, body mass index, years between CD diagnosis and enrollment), and frequencies and percentages are employed for categorical outcomes (gender, race, involved intestinal segment(s), disease severity, immunomodulator treatment, prednisone treatment, narcotic analgesic therapy, and health resource utilization). Corticosteroid therapy represents systemic corticosteroids (i.e., prednisone or equivalent) exclusively, and immunomodulator therapy represents treatment with azathioprine, 6-mercaptopurine, methotrexate, and/or cyclosporine.

treatment, prednisone treatment, narcotic analgesic therapy, and health resource utilization). Corticosteroid therapy represents systemic corticosteroids (i.e., prednisone or equivalent) exclusively, and immunomodulator therapy represents treatment with azathioprine, 6-mercaptopurine, methotrexate, and/or cyclosporine. Serious infections were defined as any event in the Medical Dictionary for Regulatory Activities System-Organ Class “Infections and Infestations” that met the standard criteria for a SAE, including any infection reported as “serious” by the investigator and any infection that required hospitalization. Commonly observed serious infections were defined as those occurring at an overall incidence >0.01/100 patient-years of follow-up. The rates of adverse events per 100 patient-years of follow-up were calculated for each medication category (infliximab-treated vs. other-treatments-only-treated) as the quotient of the total number of on-therapy events and patient-years of exposure to medication multiplied by 100. Comparisons of the resulting incidences were accomplished via generation of an RR ratio and its corresponding 95% CI; P-values were generated using Generalizing Estimation Equation methodology. For other treatment group comparisons, the Student t-test was used to test for equality of means across treatment groups, and the χ2 test was employed to evaluate the association between treatment group and categorical variables.

onding 95% CI; P-values were generated using Generalizing Estimation Equation methodology. For other treatment group comparisons, the Student t-test was used to test for equality of means across treatment groups, and the χ2 test was employed to evaluate the association between treatment group and categorical variables. Before data analysis, the TREAT Registry Advisory Board prespecified that infections occurring within 3 months of an infliximab infusion would be considered potentially related to infliximab, while infections occurring outside of this timeframe would be considered not related to infliximab. Further, a Cox proportional hazards model with time-varying covariates was used to determine the relative contribution (in the form of hazard ratios [HR] and 95% CI) of different factors to the occurrence of death and serious infection. Factors assessed included age, sex, race, diseased segment(s), disease severity, years between diagnosis and enrollment in the TREAT registry, and medication use. For mortality analyses, CD medications designated as “ever used” were obtained from all available 6-month data collection periods that occurred between enrollment and the time of death. Medication use during the period in which the death occurred was included in this analysis. For analysis of serious infection, CD medication use was defined as any use in the 6-month data collection period before the onset of the serious infection. Medication use during the period in which the infection occurred was not included in these analyses because medication start and stop dates were not collected for any medications other than infliximab. Therefore, it was unknown whether the medication was given before or after the event. Note that for covariate analyses involving data collected at 6-month intervals (e.g., 1 January–30 June, 1 July–31 December), the most current collection period before the 23 February 2010 data cutoff would have pertained to data collected through 31 December 2009.

er the medication was given before or after the event. Note that for covariate analyses involving data collected at 6-month intervals (e.g., 1 January–30 June, 1 July–31 December), the most current collection period before the 23 February 2010 data cutoff would have pertained to data collected through 31 December 2009. The effect of length of exposure on the infliximab safety profile was assessed by grouping each adverse event according to the number of infliximab infusions received by the patient before adverse event onset, i.e., 1–3, 4–9, 10–24, or ≥25 infusions. Adverse events in patients not exposed to infliximab or in infliximab-treated patients but before the first on-registry infliximab infusion were categorized within the other-treatments-only group. The 3,156 patients with at least two 6-month data collection periods as of 23 February 2010 were included in these analyses. The registry data collection forms received through 23 February 2010 were analyzed using SAS software, Version 8.02 or higher of the SAS System for Windows (SAS Institute Inc., Cary, NC). Given the retrospective nature of several data analyses, no adjustments were made for multiple comparisons.

The effect of length of exposure on the infliximab safety profile was assessed by grouping each adverse event according to the number of infliximab infusions received by the patient before adverse event onset, i.e., 1–3, 4–9, 10–24, or ≥25 infusions. Adverse events in patients not exposed to infliximab or in infliximab-treated patients but before the first on-registry infliximab infusion were categorized within the other-treatments-only group. The 3,156 patients with at least two 6-month data collection periods as of 23 February 2010 were included in these analyses. The registry data collection forms received through 23 February 2010 were analyzed using SAS software, Version 8.02 or higher of the SAS System for Windows (SAS Institute Inc., Cary, NC). Given the retrospective nature of several data analyses, no adjustments were made for multiple comparisons. RESULTS Patient disposition and baseline characteristics This report contains a summary of data collected for 6,273 patients, all of whom have been followed in the TREAT registry for at least 5 years. Data presented in this report reflect the time period from registry initiation (July 1999) through 23 February 2010. Patient enrollment was complete and closed as of 31 March 2004. The average duration of follow-up for all patients enrolled is currently 5.2 years. Infliximab-treated patients were followed for 17,712 patient-years, compared with 13,251 patient-years of follow-up for the other-treatments-only group. As of 23 February 2010, 43.2% (n=2,710) of enrolled patients were actively participating in the registry. Among these actively participating patients, the average duration of follow-up is currently 7.5 years. Both active and discontinued patients are included in this report. The most common reasons for patient discontinuation have been “patient lost to follow-up” (1,602/3,563, 45.0%) and reasons categorized as “other” (1,539/3,563, 43.2%), most of which were related to patient changes in care and changes in the physician's participation in the registry. Patient discontinuations are detailed in Figure 1.

n reasons for patient discontinuation have been “patient lost to follow-up” (1,602/3,563, 45.0%) and reasons categorized as “other” (1,539/3,563, 43.2%), most of which were related to patient changes in care and changes in the physician's participation in the registry. Patient discontinuations are detailed in Figure 1. More patients were female (58.6%) than male (41.4%), and most patients were Caucasian (90.8%). The average patient age at the time of enrollment was 42.5±14.7 years. Baseline demographic characteristics were similar between the infliximab-treated and other-treatments-only groups, with the exception of infliximab-treated patients being younger than patients in the other-treatments-only group (P<0.001) (Table 1). Note that 80 patients who were younger than 18 years were enrolled and then subsequently discontinued when the registry entrance criteria were amended to include only patients age 18 or older.

he exception of infliximab-treated patients being younger than patients in the other-treatments-only group (P<0.001) (Table 1). Note that 80 patients who were younger than 18 years were enrolled and then subsequently discontinued when the registry entrance criteria were amended to include only patients age 18 or older. A somewhat higher proportion of patients had been treated with infliximab (n=3,420, 54.5%) than those who had received other-treatments-only (n=2,853, 45.5%). Approximately 90% (n=2,986) of infliximab-treated patients had received at least two infusions. Among the 53,003 infliximab infusions documented, most (81.5%) were at a dose of 5 mg/kg. Patients in the infliximab-treated group vs. the other-treatments-only group were more likely to have received prednisone (47.8% vs. 31.4%), immunomodulators (52.0% vs. 32.1%), antibiotics (32.0% vs. 23.4%), narcotic analgesics (17.3% vs. 9.1%), and antidepressants (13.5% vs. 8.3%) within 1 year of registry enrollment (P<0.001) (Table 1).

iximab-treated group vs. the other-treatments-only group were more likely to have received prednisone (47.8% vs. 31.4%), immunomodulators (52.0% vs. 32.1%), antibiotics (32.0% vs. 23.4%), narcotic analgesics (17.3% vs. 9.1%), and antidepressants (13.5% vs. 8.3%) within 1 year of registry enrollment (P<0.001) (Table 1). Patients in the infliximab-treated group also differed significantly from patients in the other-treatments-only group with regard to intestinal segment(s) affected by disease, disease severity, and health resource utilization in the year before enrollment (Table 1). Specifically, patients who received infliximab, compared with patients who received other-treatments-only, were more likely to have both their ileum and colon involved (44.4% vs. 36.7% P<0.001), and to have moderate-to-severe (30.6% vs. 10.7% P<0.001) or severe-to-fulminant (2.5% vs. 0.6% P<0.001) disease at enrollment. Hospitalization in the year before enrollment for either surgical (17.4% vs. 13.6% P<0.001) or medical (14.2% vs. 8.8% P<0.001) management was also more common among infliximab-treated patients compared with patients who received other-treatments-only. Infusion reactions Of the 3,420 registry patients in the infliximab-treated group, 3,322 had data available for assessment of infusion reactions. Three percent (1,571/53,003) of infliximab infusions were associated with an infusion reaction; 0.047% of reactions were serious. The most common infusion reactions were headache (0.5%) and arthritis (0.4%) (Table 2).

tients in the infliximab-treated group, 3,322 had data available for assessment of infusion reactions. Three percent (1,571/53,003) of infliximab infusions were associated with an infusion reaction; 0.047% of reactions were serious. The most common infusion reactions were headache (0.5%) and arthritis (0.4%) (Table 2). Mortality One hundred and ninety-one (3%) of the 6,273 patients died over the course of registry participation; causes of death are detailed in Table 3. Of these 191 patients, 109 had received infliximab and 82 had received other-treatments-only, yielding similar time-adjusted mortality rates for patients who had received infliximab (0.58 per 100 patient-years; RR (95% CI) of 0.96 (0.72, 1.28; P=0.81) and those who had not (0.59 per 100 patient-years) (Table 4).

ble 3. Of these 191 patients, 109 had received infliximab and 82 had received other-treatments-only, yielding similar time-adjusted mortality rates for patients who had received infliximab (0.58 per 100 patient-years; RR (95% CI) of 0.96 (0.72, 1.28; P=0.81) and those who had not (0.59 per 100 patient-years) (Table 4). Adjusted results of a multivariate regression analysis (Table 5) indicated that age was a significant predictor of death (HR=1.08, 95% CI=1.07, 1.09; P<0.001). Protective effects were observed among females (HR=0.71, 95% CI=0.53, 0.96; P=0.023), Caucasians (HR=0.64, 95% CI=0.39, 1.03, P=0.064), and patients with disease in only the ileum vs. those with disease in both the ileum and colon (HR=0.53, 95% CI=0.36, 0.77; P=0.001). In investigating medication use among patients who died relative to those who did not, neither infliximab (HR=0.83, 95% CI=0.60, 1.15; P=0.262) nor immunomodulator (HR=0.86, 95% CI=0.62, 1.18; P=0.338) treatment was a significant predictor of death. However, treatment with prednisone (HR=2.14, 95% CI=1.55, 2.95; P<0.001) or narcotic analgesics (HR=1.79, 95% CI=1.29, 2.48; P<0.001) significantly predicted death (Table 5).

b (HR=0.83, 95% CI=0.60, 1.15; P=0.262) nor immunomodulator (HR=0.86, 95% CI=0.62, 1.18; P=0.338) treatment was a significant predictor of death. However, treatment with prednisone (HR=2.14, 95% CI=1.55, 2.95; P<0.001) or narcotic analgesics (HR=1.79, 95% CI=1.29, 2.48; P<0.001) significantly predicted death (Table 5). Serious infection Unadjusted rates of serious infection within 3 months of treatment were 2.06 and 1.42 per 100 patient-years, respectively, for the infliximab and other-treatments-only groups, yielding a RR (95% CI) of 1.45 (1.10, 1.91) (P=0.008). In a broader analysis including serious infections that occurred outside of the 3-month timeframe, unadjusted rates of serious infection were 2.04 and 1.00 per 100 patient-years, respectively, for the infliximab and other-treatments-only groups, yielding an RR (95% CI) of 2.04 (1.45, 2.89) (P<0.001) (Table 4). With the exception of six serious fungal infections (five in the infliximab group and one in the other-treatments-only group) and four serious mycobacterial infections (all in the infliximab group) (Table 6), all other serious infections were viral or bacterial in nature. The most commonly observed serious infections, defined as those occurring at an overall incidence >0.01/100 patient-years of follow-up, are displayed in Table 7. Without adjustment for other potential confounding factors (e.g., severity of CD and use of other immunosuppressive medications), the incidences of pneumonia, cellulitis, and perirectal abscess were higher among patients who had received infliximab (0.24, 0.15, 0.11/100 patient-years, respectfully) within the previous 3 months than among patients who had not (0.14, 0.03, 0.03/100 patient-years, respectively; Table 7).

unosuppressive medications), the incidences of pneumonia, cellulitis, and perirectal abscess were higher among patients who had received infliximab (0.24, 0.15, 0.11/100 patient-years, respectfully) within the previous 3 months than among patients who had not (0.14, 0.03, 0.03/100 patient-years, respectively; Table 7). Adjusted results of multivariate regression analyses (Table 8) indicated that moderate-to-severe disease severity was the strongest predictor of serious infection (HR=2.24, 95% CI=1.57, 3.19; P<0.001), while colon only involvement (vs. both ileum and colon involvement) was protective against the development of serious infections (HR=0.73, 95% CI=0.54, 1.00; P=0.046). When investigating medications used by patients with a serious infection relative to patients who did not have a serious infection, infliximab treatment conferred a significant HR of 1.43 (95% CI=1.11, 1.84; P=0.006). Treatment with immunomodulators (HR=1.23, 95% CI=0.96, 1.57; P=0.10) was not a significant predictor of serious infection, while narcotic analgesic (HR=1.98, 95% CI=1.44, 2.73; P<0.001) and prednisone (HR=1.57, 95% CI=1.17, 2.10; P=0.002) use were both significant predictors of serious infection (Table 8).

4; P=0.006). Treatment with immunomodulators (HR=1.23, 95% CI=0.96, 1.57; P=0.10) was not a significant predictor of serious infection, while narcotic analgesic (HR=1.98, 95% CI=1.44, 2.73; P<0.001) and prednisone (HR=1.57, 95% CI=1.17, 2.10; P=0.002) use were both significant predictors of serious infection (Table 8). There was no evidence of an increase in the occurrence of serious infections with receipt of more infliximab infusions. Specifically, the incidences of serious infections were 2.58, 2.16, 1.65, and 1.92 per 100 patient-years in association with receipt of 1–3, 4–9, 10–24, and ≥25 infliximab infusions, respectively. Escalation of the infliximab dose from 5 to 10 mg/kg also had no important effect on the occurrence of serious infections (data not shown).

e incidences of serious infections were 2.58, 2.16, 1.65, and 1.92 per 100 patient-years in association with receipt of 1–3, 4–9, 10–24, and ≥25 infliximab infusions, respectively. Escalation of the infliximab dose from 5 to 10 mg/kg also had no important effect on the occurrence of serious infections (data not shown). Pregnancy Maternal and paternal (i.e., treated male partner) pregnancies were monitored for outcome and clinical condition of live births. For both female and male registry participants, the vast majority of pregnancies resulted in lives births, i.e., 83.1% (118/142) and 95.3% (41/43), respectively, for infliximab-treated patients and 90.7% (68/75) and 91.7% (11/12), respectively, for patients receiving only other treatments. Among the maternal and paternal live births documented, the vast majority of babies were healthy with no defect or other adverse event, i.e., 92.4% (109/118) and 90.2% (37/41), respectively, for infliximab-treated patients and 85.3% (58/68) and 90.9% (10/11), respectively, for patients receiving only other treatments. A more detailed analysis of pregnancy in the TREAT registry will be the subject of a separate publication.

efect or other adverse event, i.e., 92.4% (109/118) and 90.2% (37/41), respectively, for infliximab-treated patients and 85.3% (58/68) and 90.9% (10/11), respectively, for patients receiving only other treatments. A more detailed analysis of pregnancy in the TREAT registry will be the subject of a separate publication. Neoplasia The incidences of neoplasia (including benign, malignant, and unspecified growths) were similar between infliximab-treated patients (0.78 per 100 patient-years) and patients who received other-treatments-only (0.85 per 100 patient-years), yielding an RR (95% CI) of 0.90 (0.69, 1.18) (P=0.46) (Table 4). While a more detailed analysis of malignancy in the TREAT registry will be the subject of a separate publication, as a top-level summary, the overall incidences of solid tumors (0.42 vs. 0.45 events per 100 patient-years), non-melanoma skin cancer (0.16 vs. 0.18 events per 100 patient-years), and lymphoma (0.05 vs. 0.06 events per 100 patient-years) were similar in infliximab-treated patients and those who received other-treatments-only, respectively.

y, the overall incidences of solid tumors (0.42 vs. 0.45 events per 100 patient-years), non-melanoma skin cancer (0.16 vs. 0.18 events per 100 patient-years), and lymphoma (0.05 vs. 0.06 events per 100 patient-years) were similar in infliximab-treated patients and those who received other-treatments-only, respectively. DISCUSSION We are reporting on the updated safety information acquired throughout the TREAT registry, which is a large, prospective, observational research program designed to address the long-term safety of medications, including infliximab, used to treat CD. This report, which describes the analysis of the prespecified follow-up period of at least 5 years per patient, is thus the principal registry report and updates previously reported findings through ∼2 years of registry participation. Consistent with the initial registry report (15), with an average of 5.2 years of patient follow-up in this registry study, the occurrence of death has been similar between patients treated with infliximab and those who have received other-treatments-only (0.58 vs. 0.59 per 100 patient-years, respectively; RR=0.96, 95% CI=0.72, 1.28, P=0.81). For both treatment groups, death rates remain below annual death rates for CD patients documented in epidemiological studies (17,18,19,20). Results of multivariate logistic regression analyses indicated that age, prednisone use, and narcotic analgesic use were significant predictors of death, while neither infliximab treatment nor immunomodulator therapy predicted death in patients with CD. These findings are particularly notable given that patients treated with infliximab had a worse prognosis at baseline. Although the multivariate model adjusted the HR estimate for known predictors of mortality, such as clinical disease activity, it could not account for unknown confounders associated with disease severity that likely would have conveyed a greater risk of death in the infliximab group. However, improved control of disease activity in patients treated with infliximab might have offset mortality risks related to therapy from other causes such as serious infection.

nt for unknown confounders associated with disease severity that likely would have conveyed a greater risk of death in the infliximab group. However, improved control of disease activity in patients treated with infliximab might have offset mortality risks related to therapy from other causes such as serious infection. A higher incidence of infections in general was observed in the infliximab-treated patients relative to patients receiving other-treatments-only. However, infliximab-treated patients in the TREAT registry had more severe CD at registry entry and also were more likely to be using other immunosuppressive agents or prednisone. Multivariate regression analyses indicated that moderate-to-severe disease activity was the strongest significant predictor of serious infection (HR=2.24, P<0.001), followed by treatment with narcotic analgesics (HR=1.98, P<0.001), treatment with prednisone (HR=1.57, P=0.002), and then infliximab treatment (HR=1.43, P=0.006). The statistical significance of infliximab in this regard is a new finding relative to the previous registry report (15). It is of note that the increased risk of serious infection with infliximab did not lead to a higher mortality risk. As mentioned, it is likely that infliximab lessened the disease burden and thus counterbalanced any mortality risk that might have been associated with other causes such as serious infection. It should be noted that while this type of analysis allows the relative comparison of the risk associated with one immunosuppressive drug with that of another, it does not allow for the assessment of risk associated with various combinations of therapies. In addition, the analysis does not take into consideration contributing risk factors such as duration of use, dose, or cumulative exposure of a particular immunosuppressive agent or combination regimen. However, in an analysis that grouped serious infections according to the number of infliximab infusions the patient had received before the event, there was no evidence of an increase in the occurrence of serious infections with increasing number of infliximab infusions.

immunosuppressive agent or combination regimen. However, in an analysis that grouped serious infections according to the number of infliximab infusions the patient had received before the event, there was no evidence of an increase in the occurrence of serious infections with increasing number of infliximab infusions. The significant effect of narcotic analgesic use on the risk of serious infection and mortality could result from such therapy serving as a proxy for patients with the most severe cases of CD, although we would anticipate that differences in disease severity would have been corrected by the multivariate regression analyses conducted. In addition, it could be that narcotic analgesic use, which is known to slow intestinal motility, yields longer gut processing times and thus longer periods of patient exposure to infection-causing organisms. Finally, it may also be possible that narcotic analgesic use can mask the signs and symptoms of infections until they become serious. When assessed on a “per infusion” basis, the incidences of infusion reactions were similar across the TREAT registry (1,571/53,003 or 3%) and previously conducted clinical trials of infliximab in CD patients, i.e., 106/2,026 or 5% in ACCENT I (21) and 70/1,728 or 4% in ACCENT II (22).

The significant effect of narcotic analgesic use on the risk of serious infection and mortality could result from such therapy serving as a proxy for patients with the most severe cases of CD, although we would anticipate that differences in disease severity would have been corrected by the multivariate regression analyses conducted. In addition, it could be that narcotic analgesic use, which is known to slow intestinal motility, yields longer gut processing times and thus longer periods of patient exposure to infection-causing organisms. Finally, it may also be possible that narcotic analgesic use can mask the signs and symptoms of infections until they become serious. When assessed on a “per infusion” basis, the incidences of infusion reactions were similar across the TREAT registry (1,571/53,003 or 3%) and previously conducted clinical trials of infliximab in CD patients, i.e., 106/2,026 or 5% in ACCENT I (21) and 70/1,728 or 4% in ACCENT II (22). With an approximate average of 5.2 years of patient follow-up, based on the TREAT registry data, infliximab does not appear to be associated with an increased risk of development of neoplasia in general or with an increased risk for the development of malignant solid tumors, non-melanoma skin cancer, or lymphoma. There have been no signals indicating an increased risk of fetal malformation associated with infliximab therapy analyses to date. Note that results of more detailed analyses of TREAT registry malignancy and pregnancy data will be reported in separate publications.

t solid tumors, non-melanoma skin cancer, or lymphoma. There have been no signals indicating an increased risk of fetal malformation associated with infliximab therapy analyses to date. Note that results of more detailed analyses of TREAT registry malignancy and pregnancy data will be reported in separate publications. While registries generally allow a large number of patients to be followed prospectively in a “real-world” setting, they are not as rigorously monitored as randomized controlled studies. As such, it is possible that some SAEs occurred and were not captured in the registry and, thus, such events could be under-reported in the TREAT registry. However, most analyses performed involved comparisons between different CD treatment regimens, and there is no evidence indicating that SAE reporting would be inconsistent between treatment groups, outside of the fact that maintenance infliximab therapy requires regular office visits and thus a stronger likelihood for adverse event reporting. Registries can also be limited by comprising a population that is not as “sick” as the general CD population, possibly leading to an under-representation of certain adverse events. To address this point, the rate of hospitalizations and surgeries in TREAT was compared with an age-matched (18–65 years of age) general CD population (from the MedStat MarketScan 2003 insurance claims database). In this analysis, the proportions of TREAT registry participants who had at least one CD-related hospitalization was similar to (5.36% vs. 5.15%, respectively) and the proportion who had at least one CD-related surgery was higher than (3.95% vs. 1.90%, respectively) those observed in the MarketScan database (data not shown). These data suggest that patients in the TREAT registry were representative of the general CD population.

n was similar to (5.36% vs. 5.15%, respectively) and the proportion who had at least one CD-related surgery was higher than (3.95% vs. 1.90%, respectively) those observed in the MarketScan database (data not shown). These data suggest that patients in the TREAT registry were representative of the general CD population. The potential for patient attrition can also limit interpretation of registry data (23). In an evaluation of 6,346 patients with rheumatoid arthritis who were followed semi-annually across 11 long-running databases, results of multivariable analyses indicated that younger age, lower levels of education, and non-Caucasian race predicted attrition (23). In two other registries involving a total of 6,185 patients with rheumatoid arthritis who initiated etanercept treatment, patients with follow-up data available through year 5 ranged from 42 to 57% across the registries (24). We observed similar proportions of patients with available data through an average of 5.2 years of follow-up in the TREAT registry (∼50% of patients overall).

rheumatoid arthritis who initiated etanercept treatment, patients with follow-up data available through year 5 ranged from 42 to 57% across the registries (24). We observed similar proportions of patients with available data through an average of 5.2 years of follow-up in the TREAT registry (∼50% of patients overall). Another potential limitation of these data is that only medication use during the registry or during the period before enrollment was considered. It is possible that some patients categorized as other-treatments-only had received infliximab in the more distant past. However, with regard to the risk for serious infection, the assessment of whether or not serious infections occurred within 3 months of an infliximab infusion addresses this potential concern. In addition, the prior remote use of infliximab is unlikely in most patients as the registry was initiated soon after infliximab was approved and became available. In conclusion, the results presented herein, derived from CD patients enrolled in the TREAT registry with an overall average of more than 5 years of follow-up, indicate that infliximab does not confer an increased risk of mortality. Moderate-to-severe CD was the strongest predictor of serious infection, followed by treatment with narcotic analgesics, prednisone therapy, and infliximab treatment. There has been no indication that infliximab is associated with an increased risk of neoplasia or an increased risk of fetal malformation in the TREAT registry. STUDY HIGHLIGHTS

In conclusion, the results presented herein, derived from CD patients enrolled in the TREAT registry with an overall average of more than 5 years of follow-up, indicate that infliximab does not confer an increased risk of mortality. Moderate-to-severe CD was the strongest predictor of serious infection, followed by treatment with narcotic analgesics, prednisone therapy, and infliximab treatment. There has been no indication that infliximab is associated with an increased risk of neoplasia or an increased risk of fetal malformation in the TREAT registry. STUDY HIGHLIGHTS We thank Mark Molenda (Janssen Biotech, Inc.) for analytical support, and Michelle Perate, MS, and Mary Whitman, PhD (Janssen Services, LLC) for writing and editorial support. Guarantor of the article: Gary R. Lichtenstein, MD. Specific author contributions: Gary R. Lichtenstein, Brian G. Feagan, Russell D. Cohen, Bruce A. Salzberg, Robert H. Diamond, and William J. Sandborn were involved in the registry design and conduct; Robert H. Diamond, Samiyeh Price, Wayne Langholff, and Anil Londhe were involved in data collection and analysis; and all authors provided critical content review and final approval of this manuscript. Financial support: Funding for the TREAT registry is provided by Janssen Biotech Inc., a Johnson and Johnson (J&J) pharmaceutical company.

Specific author contributions: Gary R. Lichtenstein, Brian G. Feagan, Russell D. Cohen, Bruce A. Salzberg, Robert H. Diamond, and William J. Sandborn were involved in the registry design and conduct; Robert H. Diamond, Samiyeh Price, Wayne Langholff, and Anil Londhe were involved in data collection and analysis; and all authors provided critical content review and final approval of this manuscript. Financial support: Funding for the TREAT registry is provided by Janssen Biotech Inc., a Johnson and Johnson (J&J) pharmaceutical company. Potential competing interests: Dr. Lichtenstein has received research grants and/or has served as a consultant for Abbott, Alaven, Bristol-Myers Squibb, Elan, Ferring, Janssen, Meda Pharmaceuticals, Millenium Pharmaceuticals, Pfizer Pharmaceuticals, Proctor and Gamble, Prometheus Laboratories, Salix Pharmaceuticals, Santarus, Merck/Schering-Plough, Shire Pharmaceuticals, UCB, Warner Chilcotte, and Wyeth. Dr. Feagan has received research grants and/or has served as a consultant for Abbott, Actogenix, Alba Therapeutics, Albireo Pharma, Astra Zeneca, Athersys, Axcan, Berlex, Boehringer Engelheim, Bristol-Myers Squibb, Celgene, Cerimon Pharma, CombinatoRx, Elan/Biogen, Funxional Therapeutics, GeneLogic, Genentech, Given Imaging, Gilead, Glaxo Smithkline, ISIS, Janssen, Merck/Schering-Plough, Millennium, Napo Pharma, Nektar, Novartis, Novonordisk, Ore Pharm. (previously GeneLogic), Osiris, Otsuka, Pfizer, Proctor and Gamble, Prometheus Therapeutics and Diagnostics, Protein Design Labs, Salix Pharma, Santarus, Schering Canada, Serono, Shire, Synta, Teva Pharma, Tillotts Pharma AG, Tioga Pharma, UCB Pharma, Unity Pharma, Wyeth, and Zealand Pharma. Dr Cohen has received research grants and/or has served as a consultant for Janssen. Dr Salzberg has received consulting and/or speaker fees for Abbott Laboratories, Janssen, and Shire. Dr Sandborn has received consulting fees and research support from Abbott Laboratories, Janssen, and UCB Pharma, as well as consulting fees from Merck/Schering-Plough. R.H. Diamond and S. Price are employed by Janssen, a J&J company, and own stock in J&J. W. Langholff and A. Londhe are employees of J&J Pharmaceutical Research Division and own stock in J&J.

es and research support from Abbott Laboratories, Janssen, and UCB Pharma, as well as consulting fees from Merck/Schering-Plough. R.H. Diamond and S. Price are employed by Janssen, a J&J company, and own stock in J&J. W. Langholff and A. Londhe are employees of J&J Pharmaceutical Research Division and own stock in J&J. Figure 1 Flow of participation in the Therapy, Resource, Evaluation, and Assessment Tool registry. Infliximab-treated patients are those patients who received infliximab within 12 weeks before registration, who were scheduled to receive infliximab within 30 days of registration, or who received infliximab at some other point in the registry (i.e., exposed while participating in the registry). Note that one patient who died after 23 February 2010 is included.

nts who received infliximab within 12 weeks before registration, who were scheduled to receive infliximab within 30 days of registration, or who received infliximab at some other point in the registry (i.e., exposed while participating in the registry). Note that one patient who died after 23 February 2010 is included. Table 1 Baseline patient demographics and disease characteristics All patients Infliximab-treateda Other-treatments-only P valueb N=6,273 N=3,420 N=2,853 Total patients, N (%) 6,273 (100.0) 3,420 (54.5) 2,853 (45.5) Age (years) at enrollment, N 6,236 3,404 2,832 <0.001 Mean (s.d.) 42.5 (14.7) 40.5 (14.0) 44.9 (15.3) Gender, N (%) 6,163 3,372 2,791 0.46 Male 2,552 (41.4) 1,382 (41.0) 1,170 (41.9) Female 3,611 (58.6) 1,990 (59.0) 1,621 (58.1) Baseline BMI, N 5,822 3,180 2,642 0.84 Mean (s.d.) 25.8 (5.5) 25.8 (5.6) 25.8 (5.4) Race/ethnicity, N (%) 6,153 3,368 2,785 0.59 Caucasian 5,586 (90.8) 3,044 (90.4) 2,542 (91.3) Black 420 (6.8) 244 (7.2%) 176 (6.3) Asian 24 (0.4) 15 (0.4) 9 (0.3) Hispanic 81 (1.3) 43 (1.3) 38 (1.4) Other 42 (0.7) 22 (0.7) 20 (0.7) Years between diagnosis and enrollment, N 6,105 3,341 2,764 0.31 Mean (s.d.) 11.3 (10.2) 11.2 (9.8) 11.4 (10.7) Disease severity, N (%) 5,999 3,280 2,719 <0.001 Remissionc 1,528 (25.5) 465 (14.2) 1,063 (39.1) Mild–moderated 3,076 (51.3) 1,728 (52.7) 1,348 (49.6) Moderate–severee 1,296 (21.6) 1,004 (30.6) 292 (10.7) Severe–fulminantf 99 (1.7) 83 (2.5) 16 (0.6) Involved intestinal area, N (%) 6,034 3,308 2,726 <0.001 Ileum only 1,801 (29.8) 869 (26.3) 932 (34.2) Colon only 1,764 (29.2) 971 (29.4) 793 (29.1) Ileum and colon 2,469 (40.9) 1,468 (44.4) 1,001 (36.7) Health resource utilization in year before enrollment, N (%) Any admission 1,469 (23.4) 930 (27.2) 539 (18.9) <0.001 Surgical admission 983 (15.7) 596 (17.4) 387 (13.6) <0.001 Medical admission 736 (11.7) 485 (14.2) 251 (8.8) <0.001 Medication use at enrollment, N (%) Antibiotics 1,763 (28.1) 1,094 (32.0) 669 (23.4) <0.001 Antidepressants 698 (11.1) 462 (13.5) 236 (8.3) <0.001 Immunomodulators 2,696 (43.0) 1,780 (52.0) 916 (32.1) <0.001 Narcotic analgesics 853 (13.6) 593 (17.3) 260 (9.1) <0.001 Prednisone 2,530 (40.3) 1,635 (47.8) 895 (31.4) <0.001 BMI, body mass index.

lment, N (%) Antibiotics 1,763 (28.1) 1,094 (32.0) 669 (23.4) <0.001 Antidepressants 698 (11.1) 462 (13.5) 236 (8.3) <0.001 Immunomodulators 2,696 (43.0) 1,780 (52.0) 916 (32.1) <0.001 Narcotic analgesics 853 (13.6) 593 (17.3) 260 (9.1) <0.001 Prednisone 2,530 (40.3) 1,635 (47.8) 895 (31.4) <0.001 BMI, body mass index. a Infliximab-treated patients are those who received infliximab within 12 weeks before enrollment, were scheduled to receive infliximab within 30 days of enrollment, or received infliximab at some other time during the registry. b P value from t-test (continuous variables) or χ2 test (categorical variables). c Remission refers to patients who are asymptomatic or without inflammatory sequelae and refers to patients who have responded to acute medical intervention or have undergone surgical resection without evidence of residual disease. Patients requiring corticosteroids to maintain well-being are considered to be “steroid-dependent” and are not in “in remission.” d Mild-moderate disease applies to ambulatory patients who are able to tolerate oral alimentation without manifestations of dehydration, toxicity (high fevers, rigors, and prostration), abdominal tenderness, painful mass, obstruction, or >10% weight loss. e Moderate-severe disease applies to patients who have failed to respond to treatment for mild-moderate disease or those with more prominent symptoms of fevers, significant weight loss, abdominal pain or tenderness, intermittent nausea or vomiting (without obstructive findings), or significant anemia.

d Mild-moderate disease applies to ambulatory patients who are able to tolerate oral alimentation without manifestations of dehydration, toxicity (high fevers, rigors, and prostration), abdominal tenderness, painful mass, obstruction, or >10% weight loss. e Moderate-severe disease applies to patients who have failed to respond to treatment for mild-moderate disease or those with more prominent symptoms of fevers, significant weight loss, abdominal pain or tenderness, intermittent nausea or vomiting (without obstructive findings), or significant anemia. f Severe-fulminant disease applies to patients with persistent symptoms despite the introduction of corticosteroids, or individuals presenting with high fever, persistent vomiting, evidence of intestinal obstruction, rebound tenderness, cachexia, or evidence of an abscess.

e Moderate-severe disease applies to patients who have failed to respond to treatment for mild-moderate disease or those with more prominent symptoms of fevers, significant weight loss, abdominal pain or tenderness, intermittent nausea or vomiting (without obstructive findings), or significant anemia. f Severe-fulminant disease applies to patients with persistent symptoms despite the introduction of corticosteroids, or individuals presenting with high fever, persistent vomiting, evidence of intestinal obstruction, rebound tenderness, cachexia, or evidence of an abscess. Table 2 Summary of infusion reactions among all infliximab-treated patients Total infusions N=53,003 Total number of patients receiving an infliximab infusiona 3,420 Total number of patients with available data 3,322 Total number of infusions 53,003 Average infusion dose, N (%) 5 mg/kg 43,195 (81.5) 10 mg/kg 9,094 (17.2) Unknown 714 (1.3) Patients receiving ≥2 infusions 2,986 (89.9) Medications administered before infusion, N (%) Acetaminophen 28,146 (53.2) Antihistamines 29,870 (56.5) None 18,930 (35.8) Steroids 7,453 (14.1) Infusion reaction status, N (%) Delayed (>4 h to 14 days) 594 (1.1) Immediate (≤4 h) 920 (1.7) None 51,432 (97.2) Unknown 57 (0.1) Infusion reactions, N (%) Total number of infusion reactions/number of infusions 1,571/53,003 (3.0) Arthritis 232 (0.4) Cardiopulmonary symptomsb 145 (0.3) Chills 76 (0.1) Headache 268 (0.5) Hypertension 35 (0.1) Hypotension 31 (0.1) Influenza-like illness 127 (0.2) Muscle spasm 130 (0.2) Nausea 177 (0.3) Other 815 (1.5) Pyrexia 44 (0.1) Rash 182 (0.3) a Refers to treatment only during participation in the Crohn's Therapy, Resource, Evaluation, and Assessment Tool registry; this includes patients who received infliximab infusion(s) within 12 weeks before registration, were scheduled to receive infliximab within 30 days of registration, or received infliximab at some other point in the Registry.

uring participation in the Crohn's Therapy, Resource, Evaluation, and Assessment Tool registry; this includes patients who received infliximab infusion(s) within 12 weeks before registration, were scheduled to receive infliximab within 30 days of registration, or received infliximab at some other point in the Registry. b Case report form (CRF) Verbatim Term—no equivalent Medical Dictionary for Regulatory Activities term available. This was one of the several terms provided to investigators to select from the CRF.

uring participation in the Crohn's Therapy, Resource, Evaluation, and Assessment Tool registry; this includes patients who received infliximab infusion(s) within 12 weeks before registration, were scheduled to receive infliximab within 30 days of registration, or received infliximab at some other point in the Registry. b Case report form (CRF) Verbatim Term—no equivalent Medical Dictionary for Regulatory Activities term available. This was one of the several terms provided to investigators to select from the CRF. Table 3 Reported causes of death in the TREAT registry Infliximab-treated Other-treatments-only Cause of death Incidents, N=109 Average days since last infliximab infusion Cause of death Incidents, N=83a Acute lymphocytic leukemia 1 2334.0 Accident 1 Anemia, aplastic 1 556.0 Anemia, severe 1 Aneurysm, thoracic 1 88.0 Aneurysm, cerebral 1 Anorexia 1 104.0 Aneurysm, ruptured 1 Aspiration into lungs 1 2573.0 Ascending aortic dissection 1 Aspiration pneumonia 1 1517.0 Cardiac arrest 7 Cardiac arrest 2 513.5 Cardiac arrhythmia 1 Cardiac arrhythmia 1 417.0 Cirrhosis of the liver 1 Cardiopulmonary event, acute 1 398.0 Colitis, ischemic 1 Cerebral amyloidosis 1 788.0 Decompensated cirrhosis 1 Cerebral vascular event 1 1235.0 Gastrointestinal bleeding 1 Drug overdose 2 597.0 Heart failure 3 Encephalopathy 1 1224.0 Heart failure, congestive 2 Gastroenteritis 1 1754.0 Human immunodeficiency virus 1 Gastrointestinal bleeding 2 347.5 Hypotension 1 Heart failure, congestive 2 316.5 Intracerebral hemorrhage 1 Ischemic ileocolitis 1 420.0 Leukemia 1 Liver failure 2 284.5 Lymphadenopathy 1 Liver transplant 1 1237.0 Lymphoma 4 Lymphoma, non-Hodgkin's 2 757.5 Malignancy, breast 2 Malignancy, brain 1 134.0 Malignancy, lung 3 Malignancy, colon 4 391.3 Malignancy, metastatic 8 Malignancy, lung 4 459.0 Malignancy, pancreatic 1 Malignancy, metastatic 4 476.8 Malignancy, renal 2 Malignancy, pancreatic 2 1506.0 Malignancy, squamous cell 1 Malignancy, peritoneal 1 46.0 Malignancy, throat 1 Malignancy, renal 1 184.0 Malignancy, tongue 1 Malignancy, small bowel 1 405.0 Malignancy, unspecified 1 Malnutrition 1 211.0 Malnutrition 2 Meningitis 1 76.0 Motor vehicle accident 1 Mesenteric venous thrombosis 1 14.0 Methicillin-resistant Staphylococcus aureus infection 1 Motor vehicle accident 1 46.0 Myeloma 1 Myocardial infarction 5 563.6 Myocardial infarction 3 Natural causes 1 46.0 Natural causes 1 Pneumonia 5 1027.6 Pneumonia 1 Pulmonary disease, chronic obstructive 4 649.3 Pulmonary complications, postsurgical 1 Pulmonary embolism 1 55.0 Renal failure 3 Renal failure 4 1482.0 Renal failure, chronic 1 Respiratory distress 1 155.0 Respiratory distress syndrome, acute 1 Respiratory failure 7 771.1

Natural causes 1 Pneumonia 5 1027.6 Pneumonia 1 Pulmonary disease, chronic obstructive 4 649.3 Pulmonary complications, postsurgical 1 Pulmonary embolism 1 55.0 Renal failure 3 Renal failure 4 1482.0 Renal failure, chronic 1 Respiratory distress 1 155.0 Respiratory distress syndrome, acute 1 Respiratory failure 7 771.1 Respiratory failure 2 Sepsis 9 355.7 Sepsis 6 Stroke 2 187.5 Small bowel perforation 1 Sudden cardiac arrest 1 3801.0 Suicide 2 Sudden death 1 20.0 Unknown 5 Surgical complications—cholecystectomy 1 476.0 — — Surgical complications—unspecified 2 443.0 — — Unknown 15 825.7 — — Urinary sepsis 1 316.0 — — Urosepsis 1 110.0 — — TREAT, Therapy, Resource, Evaluation, and Assessment Tool. a Includes an additional patient who died after 23 February 2010 and is not included in Table 4.

Respiratory failure 2 Sepsis 9 355.7 Sepsis 6 Stroke 2 187.5 Small bowel perforation 1 Sudden cardiac arrest 1 3801.0 Suicide 2 Sudden death 1 20.0 Unknown 5 Surgical complications—cholecystectomy 1 476.0 — — Surgical complications—unspecified 2 443.0 — — Unknown 15 825.7 — — Urinary sepsis 1 316.0 — — Urosepsis 1 110.0 — — TREAT, Therapy, Resource, Evaluation, and Assessment Tool. a Includes an additional patient who died after 23 February 2010 and is not included in Table 4. Table 4 Unadjusted incidences of neoplasia, mortality, and serious infection Treatment groupa Number of patientsb Number of eventsc Number of patient-years Unadjusted rate per 100 patient-years Unadjusted RR (95% CI) ratio P valued Neoplasia Other-treatments-only 4,010 113 13,251 0.85 1.00 (reference) Infliximab-treated 3,764 139 17,712 0.78 0.90 (0.69, 1.18) 0.46 Mortality Other-treatments-only 4,113 82e 13,979 0.59 1.00 (reference) Infliximab-treated 3,764 109 18,825 0.58 0.96 (0.72, 1.28) 0.81 Serious infection Other-treatments-only 4,557 147 14,710 1.00 1.00 (reference) Infliximab-treated 3,420 333 16,296 2.04 2.04 (1.45, 2.89) <0.001 Serious infection according to infliximab exposure within the previous 3 months Other-treatments-onlyf 5,597 317 22,344 1.42 1.00 (reference) Infliximab-treated 2,942 163 7,923 2.06 1.45 (1.10, 1.91) 0.008 CI, confidence interval; RR, relative risk.

nce) Infliximab-treated 3,420 333 16,296 2.04 2.04 (1.45, 2.89) <0.001 Serious infection according to infliximab exposure within the previous 3 months Other-treatments-onlyf 5,597 317 22,344 1.42 1.00 (reference) Infliximab-treated 2,942 163 7,923 2.06 1.45 (1.10, 1.91) 0.008 CI, confidence interval; RR, relative risk. a For neoplasia and mortality, patients in the infliximab-treated group received infliximab during the registry before event onset, including the year before registration (i.e., patients “ever exposed” to infliximab). For serious infections, patients in the infliximab-treated group received infliximab during the registry participation and before event onset, including the 12 weeks before registration. Patients could contribute data to both treatment groups depending on the date of first infliximab use. b Over the course of the registry, a patient can contribute patient-years at risk to more than one medication or dosing category. c Data are available from the beginning of the registry (neoplasia and mortality) or beginning in 2002 (serious infections). d P value derived from Generalized Estimating Equations. e Note that one additional patient “ever exposed” (see footnote a of Table 3) to infliximab died after 23 February 2010 and is not included in this table but is shown in Table 3. f The other-treatments-only group included all patients without infliximab exposure within the previous 3 months.

d P value derived from Generalized Estimating Equations. e Note that one additional patient “ever exposed” (see footnote a of Table 3) to infliximab died after 23 February 2010 and is not included in this table but is shown in Table 3. f The other-treatments-only group included all patients without infliximab exposure within the previous 3 months. Table 5 Predictors of mortality Unadjusted results Adjusted results All patients (N=6,080) Deaths (N=187) Hazard ratio (95% CI) P valuea Hazard ratio (95% CI) P valuea Age at enrollment Mean (s.d.) 42.5 (14.7) 59.1 (15.5) 1.074 (1.063, 1.084) <0.001 1.078 (1.067, 1.090) <0.001 Sex Male/unknown 2,519 88/2,519 (3.49%) —   — Female 3,561 99/3,561 (2.78%) 0.806 (0.605, 1.074) 0.14 0.714 (0.534, 0.955) 0.023 Race Non-Caucasian 573 19/573 (3.32%) —   — Caucasian 5,507 168/5,507 (3.05%) 0.779 (0.485, 1.253) 0.30 0.635 (0.394, 1.026) 0.06 Involved intestinal area(s)b Ileum and colon 2,436 93/2,436 (3.82%) —   — Ileum only 1,784 39/1,784 (2.19%) 0.540 (0.371, 0.785) 0.001 0.528 (0.362, 0.770) 0.001 Colon only 1,742 54/1,742 (3.10%) 0.767 (0.548, 1.073) 0.12 0.741 (0.522, 1.052) 0.09 Unknown 118 1/118 (0.85%) 0.316 (0.044, 2.273) 0.25 0.387 (0.054, 2.787) 0.35 Years between diagnosis and enrollment Mean (s.d.) 10.8 (10.2) 16.7 (13.7) 1.043 (1.031, 1.055) <0.001 1.011 (1.000, 1.022) 0.05 Disease severityc Remission 569 19/569 (3.34%) —   — Mild 3,127 85/3,127 (2.72%) 1.072 (0.648, 1.774) 0.79 1.078 (0.643, 1.807) 0.78 Moderate/severe/fulminant 2,336 83/2,336 (3.55%) 1.699 (1.019, 2.835) 0.042 1.612 (0.910, 2.856) 0.10 Unknown 48 0/48 (0.00%) 0.000 (0.000, 0.000) 0.98 0.000 (0.000, 0.000) 0.98 Infliximabd No 2,766 94/2,766 (3.40%) —   — Yes 3,314 93/3,314 (2.81%) 0.865 (0.648, 1.154) 0.33 0.828 (0.595, 1.152) 0.26 Prednisoned No 3,249 77/3,249 (2.37%) —   — Yes 2,831 110/2,831 (3.89%) 1.973 (1.470, 2.648) <0.001 2.140 (1.550, 2.954) <0.001 Immunomodoulatorsd,e No 2,097 82/2,097 (3.91%) —   — Yes 3,983 105/3,983 (2.64%) 0.713 (0.533, 0.955) 0.023 0.856 (0.623, 1.176) 0.34 Narcotic analgesicsd No 4,586 124/4,586 (2.70%) —   — Yes 1,494 63/1,494 (4.22%) 1.855 (1.366, 2.520) <0.001 1.789 (1.290, 2.479) <0.001 CI, confidence interval.

40 (1.550, 2.954) <0.001 Immunomodoulatorsd,e No 2,097 82/2,097 (3.91%) —   — Yes 3,983 105/3,983 (2.64%) 0.713 (0.533, 0.955) 0.023 0.856 (0.623, 1.176) 0.34 Narcotic analgesicsd No 4,586 124/4,586 (2.70%) —   — Yes 1,494 63/1,494 (4.22%) 1.855 (1.366, 2.520) <0.001 1.789 (1.290, 2.479) <0.001 CI, confidence interval. Patients were eligible for this analysis if they had non-missing baseline covariates. Only data representing patient status between registration and 31 December 2009 are used within this analysis. a P value from Wald χ2 test. b This represents diseased area at baseline, as disease area was not collected longitudinally. c This represents time-varying maximum severity between enrollment and the 6-month data collection period of the event or censoring. d This represents time-varying medication use, and is defined as any use between enrollment and the 6-month data collection period before the event or censoring. e Immunomodulators include azathioprine, methotrexate, and 6-mercaptopurine.

c This represents time-varying maximum severity between enrollment and the 6-month data collection period of the event or censoring. d This represents time-varying medication use, and is defined as any use between enrollment and the 6-month data collection period before the event or censoring. e Immunomodulators include azathioprine, methotrexate, and 6-mercaptopurine. Table 6 Summary of serious Mycobacterial and fungal infections Recenta use of Serious infection Patient age at onset (years) Treatment group Immunomodulatorsb Corticosteroids Fatal? Mycobacterial Mycobacterium avium complex infection 47.3 Infliximab Yes Yes No Mycobacterial TB 71.9 Infliximab No No No Mycobacterial pulmonary TB 52.0 Infliximab Yes No No Mycobacterial pulmonary TB 33.3 Infliximab Yes No No Mean 51.1   Yes/No: 3/1 Yes/No: 1/3 Fungal Pneumocystis jiroveci pneumonia 69.5 Infliximab No No Yes Candida glabrata fungaemia 36.4 Infliximab No Yes No Candida tropicalis candidiasisc 58.8 Infliximab Yes Yes No Systemic candidiasis 40.1 Other-treatments-only Yes No No Systemic fungaemiac 59.3 Inflixmab Yes Yes No Candida sepsis 26.4 Infliximab No No No Mean 48.4   Yes/No: 3/3 Yes/No: 3/3 TB, tuberculosis. a Recent use is defined as use within the 6-month data collection period before serious adverse event (SAE) onset and/or use within the same data collection period as SAE onset. b Includes azathioprine, 6-mercaptopurine, and/or methotrexate. c Two fungal infections occurred in the same patient.

Table 6 Summary of serious Mycobacterial and fungal infections Recenta use of Serious infection Patient age at onset (years) Treatment group Immunomodulatorsb Corticosteroids Fatal? Mycobacterial Mycobacterium avium complex infection 47.3 Infliximab Yes Yes No Mycobacterial TB 71.9 Infliximab No No No Mycobacterial pulmonary TB 52.0 Infliximab Yes No No Mycobacterial pulmonary TB 33.3 Infliximab Yes No No Mean 51.1   Yes/No: 3/1 Yes/No: 1/3 Fungal Pneumocystis jiroveci pneumonia 69.5 Infliximab No No Yes Candida glabrata fungaemia 36.4 Infliximab No Yes No Candida tropicalis candidiasisc 58.8 Infliximab Yes Yes No Systemic candidiasis 40.1 Other-treatments-only Yes No No Systemic fungaemiac 59.3 Inflixmab Yes Yes No Candida sepsis 26.4 Infliximab No No No Mean 48.4   Yes/No: 3/3 Yes/No: 3/3 TB, tuberculosis. a Recent use is defined as use within the 6-month data collection period before serious adverse event (SAE) onset and/or use within the same data collection period as SAE onset. b Includes azathioprine, 6-mercaptopurine, and/or methotrexate. c Two fungal infections occurred in the same patient. Table 7 Most common (>0.01/100 patient-years among all patients) serious infections All patients Infliximab-treateda Other-treatments-onlyb N=6,273 N=2,942 N=5,597 Total patient-years of follow-up 30,268 7,923 22,344 Common adverse events Rate/100 patient-years (no. of events) Pneumonia 0.17 (50) 0.24 (19) 0.14 (31) Abdominal abscess 0.09 (28) 0.11 (9) 0.09 (19) Catheter sepsis 0.09 (27) 0.06 (5) 0.10 (22) Sepsis 0.07 (21) 0.03 (2) 0.09 (19) Cellulitis 0.06 (18) 0.15 (12) 0.03 (6) Central line infection 0.06 (17) 0.04 (3) 0.06 (14) Gastroenteritis 0.05 (16) 0.05 (4) 0.05 (12) Perirectal abscess 0.05 (16) 0.11 (9) 0.03 (7) Diverticulitis 0.05 (15) 0.04 (3) 0.05 (12) Pelvic abscess 0.05 (14) 0.04 (3) 0.05 (11) Intestinal abscess 0.04 (13) 0.01 (1) 0.05 (12) Viral gastroenteritis 0.04 (11) 0.04 (3) 0.04 (8) Appendicitis 0.03 (10) 0.06 (5) 0.02 (5) Wound infection 0.03 (10) 0.05 (4) 0.03 (6) Abscess 0.03 (9) 0.03 (2) 0.03 (7) Urinary tract infection 0.03 (9) 0.06 (5) 0.02 (4) Catheter bacteremia 0.02 (7) 0.03 (2) 0.02 (5) Staphylococcal infection 0.02 (7) 0.03 (2) 0.02 (5) Clostridium difficile colitis 0.02 (6) 0.01 (1) 0.02 (5) Postoperative abscess 0.02 (6) 0.03 (2) 0.02 (4) Anal abscess 0.02 (5) 0.04 (3) 0.01 (2) Infection (not otherwise specified) 0.02 (5) 0.04 (3) 0.01 (2) Rectal abscess 0.02 (5) 0.01 (1) 0.02 (4) Urosepsis 0.02 (5) 0.03 (2) 0.01 (3) Viral infection 0.02 (5) 0.01 (1) 0.02 (4) a Infliximab-treated patients are those patients who received infliximab within 12 weeks before the serious infection.

.04 (3) 0.01 (2) Infection (not otherwise specified) 0.02 (5) 0.04 (3) 0.01 (2) Rectal abscess 0.02 (5) 0.01 (1) 0.02 (4) Urosepsis 0.02 (5) 0.03 (2) 0.01 (3) Viral infection 0.02 (5) 0.01 (1) 0.02 (4) a Infliximab-treated patients are those patients who received infliximab within 12 weeks before the serious infection. b Includes all patients who did not receive infliximab within the previous 3 months.

.04 (3) 0.01 (2) Infection (not otherwise specified) 0.02 (5) 0.04 (3) 0.01 (2) Rectal abscess 0.02 (5) 0.01 (1) 0.02 (4) Urosepsis 0.02 (5) 0.03 (2) 0.01 (3) Viral infection 0.02 (5) 0.01 (1) 0.02 (4) a Infliximab-treated patients are those patients who received infliximab within 12 weeks before the serious infection. b Includes all patients who did not receive infliximab within the previous 3 months. Table 8 Predictors of serious infections Unadjusted results Adjusted results All patients (N=5,394) Serious infections (N=297) Hazard ratio (95% CI) P valuea Hazard ratio (95% CI) P valuea Age at enrollment Mean (s.d.) 43.0 (14.6) 45.3 (15.5) 1.010 (1.002, 1.017) 0.017 1.013 (1.004, 1.023) 0.007 Sex Male/unknown 2,222 121/2,222 (5.45%) — Female 3,172 176/3,172 (5.55%) 1.044 (0.828, 1.316) 0.72 1.006 (0.789, 1.283) 0.96 Race Non-Caucasian 501 30/501 (5.99%) — Caucasian 4,893 267/4,893 (5.46%) 0.776 (0.532, 1.132) 0.19 0.764 (0.511, 1.142) 0.19 Involved intestinal area(s)b Ileum and colon 2,121 141/2,121 (6.65%) — Ileum only 1,631 85/1,631 (5.21%) 0.762 (0.582, 0.997) 0.048 0.858 (0.646, 1.139) 0.29 Colon only 1,579 67/1,579 (4.24%) 0.613 (0.459, 0.821) 0.001 0.729 (0.535, 0.995) 0.046 Unknown 63 4/63 (6.35%) 0.821 (0.302, 2.229) 0.70 1.412 (0.513, 3.891) 0.50 Years between diagnosis and enrollment Mean (s.d.) 10.9 (10.2) 14.0 (11.9) 1.025 (1.015, 1.035) <0.001 1.018 (1.007, 1.030) 0.002 Severityc Remission 2,468 99/2,468 (4.01%) — Mild 2,113 122/2,113 (5.77%) 1.197 (0.917, 1.562) 0.19 0.976 (0.731, 1.302) 0.87 Moderate/severe/fulminant 459 71/459 (15.47%) 3.473 (2.553, 4.723) <0.001 2.239 (1.569, 3.194) <0.001 Unknown 62 0/62 (0.00%) 0.000 (0.000, 0.000) 0.96 0.000 (0.000, 0.000) 0.97 Infliximabd No 2,283 146/2,283 (6.40%) — Yes 827 129/827 (15.60%) 1.749 (1.380, 2.219) <0.001 1.431 (1.110, 1.844) 0.006 Prednisoned No 4,496 217/4,496 (4.83%) — Yes 599 75/599 (12.52%) 2.317 (1.778, 3.019) <0.001 1.571 (1.173, 2.103) 0.002 Immunomodoulatorsd,e No 3,255 148/3,255 (4.55%) — Yes 1,849 146/1,849 (7.90%) 1.296 (1.030, 1.629) 0.027 1.227 (0.961, 1.566) 0.10 Narcotic analgesicsd No 4,721 237/4,721 (5.02%) — Yes 374 55/374 (14.71%) 2.906 (2.166, 3.899) <0.001 1.980 (1.436, 2.729) <0.001 CI, confidence interval.

<0.001 1.571 (1.173, 2.103) 0.002 Immunomodoulatorsd,e No 3,255 148/3,255 (4.55%) — Yes 1,849 146/1,849 (7.90%) 1.296 (1.030, 1.629) 0.027 1.227 (0.961, 1.566) 0.10 Narcotic analgesicsd No 4,721 237/4,721 (5.02%) — Yes 374 55/374 (14.71%) 2.906 (2.166, 3.899) <0.001 1.980 (1.436, 2.729) <0.001 CI, confidence interval. Patients were eligible for this analysis if they had non-missing baseline covariates. Only data representing patient status between registration and 31 December 2009 are used within this analysis. a P value from Wald χ2 test. b This represents diseased area at baseline, as disease area was not collected longitudinally. c This represents time-varying maximum severity between enrollment and the 6-month data collection period of the event or censoring. d This represents time-varying medication use and is defined as any use in the 6-month data collection period before the event or censoring. e Immunomodulators include azathioprine, methotrexate, and 6-mercaptopurine.

INTRODUCTION Hepatocellular carcinoma (HCC) ranks as the fifth most common cancer worldwide (1). In Japan, ∼35,000 patients die from HCC every year (2), and the main cause of HCC is hepatitis C virus infection. In chronic hepatitis patients, screening of HCC is usually performed by ultrasonography, and the diagnosis is confirmed by contrast-enhanced dynamic computed tomography (CT). Hyperattenuation in the arterial phase and hypoattenuation in the equilibrium phase are considered to be definitive signs of HCC ((3,4,5,6,7)). Hyperattenuation in the arterial phase is more emphasized when contrast material is injected from the hepatic artery through a catheter, because dilution of contrast material in the systemic circulation is avoided, thus keeping a high concentration of contrast material in the liver. This technique is called CT during hepatic arteriography (CTHA) ((6,8,9,10)). Similarly, hypoattenuation in the equilibrium phase is accentuated after injection of contrast material into the superior mesenteric artery, which is referred to as CT during arterial portography (CTAP) ((11,12,13,14)). The combination of CTHA and CTAP gives higher sensitivity and specificity for HCC detection than conventional dynamic enhanced CT (8).

equilibrium phase is accentuated after injection of contrast material into the superior mesenteric artery, which is referred to as CT during arterial portography (CTAP) ((11,12,13,14)). The combination of CTHA and CTAP gives higher sensitivity and specificity for HCC detection than conventional dynamic enhanced CT (8). If new HCC nodules are detected with CTHA/CTAP, in addition to those detected with dynamic CT, the treatment of choice may be changed ((15,16)). For example, surgical resection and liver transplantation are usually contraindicated for multinodular HCC; that is, exceeding three nodules. Percutaneous tumor ablation methods, such as ethanol injection and microwave coagulation, have played an important role as nonsurgical treatments that can achieve high local cure rates without affecting background liver function ((17,18,19,20)). Radiofrequency ablation (RFA) is currently considered to be the most effective first-line percutaneous ablation protocol because of its greater efficacy in terms of local cure as compared with ethanol injection ((21,22,23,24)). However, even after complete ablation, patients frequently encounter intrahepatic tumor recurrence at a rate of 50% in 2 years, the majority of which occurs at locations distant from the primary ablated site (25). Considering the tumor doubling time, many nodules diagnosed as recurrent within 2 years were probably present at the time of first ablation. If nodules that are undetectable by conventional dynamic CT could be detected and ablated, the recurrence rate would be decreased.

ions distant from the primary ablated site (25). Considering the tumor doubling time, many nodules diagnosed as recurrent within 2 years were probably present at the time of first ablation. If nodules that are undetectable by conventional dynamic CT could be detected and ablated, the recurrence rate would be decreased. Although CTHA/CTAP is one of the most sensitive techniques available for detection of small HCC, its disadvantages include invasiveness, high cost, and a high false-positive rate (26). The indication for CTHA/CTAP can be justified only when the expected benefits exceed the risk and cost of the procedure. We conducted a single-center, randomized, open-label, controlled trial to assess the utility of CTHA/CTAP before RFA in patients with early-stage HCC by comparing recurrence-free and overall survival.

he indication for CTHA/CTAP can be justified only when the expected benefits exceed the risk and cost of the procedure. We conducted a single-center, randomized, open-label, controlled trial to assess the utility of CTHA/CTAP before RFA in patients with early-stage HCC by comparing recurrence-free and overall survival. METHODS Patients The study population consisted of patients with early-stage HCC with an indication for RFA. Those who met the following criteria were enrolled between September 2004 and February 2009: (i) diagnosis of typical HCC on dynamic CT performed within 2 weeks, i.e., hyperattenuation during the arterial phase and hypoattenuation during the equilibrium phase ((5,6)); (ii) tumor size ≤3.0 cm and no more than three tumor nodules; (iii) Child–Pugh class A liver function; and (iv) age >20 years. Exclusion criteria were: allergy to contrast media; portal or hepatic vein tumor thrombosis; extrahepatic metastasis; diffuse and infiltrative tumors; renal failure (serum creatinine >2.0 mg/dl, or serum urea nitrogen >30 mg/dl); impaired coagulation (e.g., platelet count <50 × 103/μl, or prothrombin activity <50%); pregnancy; or past history of choledochojejunostomy. We included those with previous treatments as well as treatment-naive cases provided that there was no local recurrence at enrollment. These inclusion criteria and the study design did not change till the study completely ended. The study design conformed to the Declaration of Helsinki Principles and was approved by the ethics committee of our institution. The study was registered at the University Hospital Medical Information Network (UMIN) Clinical Trial Registry (UMIN-CTR000000070). Written informed consent was obtained from each patient. This study complied with the CONSORT guidelines for reporting of clinical trials (27).

by the ethics committee of our institution. The study was registered at the University Hospital Medical Information Network (UMIN) Clinical Trial Registry (UMIN-CTR000000070). Written informed consent was obtained from each patient. This study complied with the CONSORT guidelines for reporting of clinical trials (27). Study design Before receiving RFA, patients were randomly assigned to undergo CTHA/CTAP or not in equal numbers. Patient registration and randomization were performed by computer-generated allocation at a web-based data center (Internet Data and Information Center for Medical Research) administered by UMIN. At the time of randomization, patients were stratified either as treatment naive, for whom RFA was planned as an initial treatment for HCC, or recurrent, for whom RFA was planned for recurrent HCC. The randomization was based on the Efron's biased-coin design (28). In principal, the assignment was not blinded to the investigators and the participants. The interval between random assignment and implementation of treatment for HCC was <4 weeks. CTHA/CTAP was performed on the assigned patients on the second day of admission, and RFA was performed 2 or 3 days later, given that the total number of HCC nodules remained <4. When ≥4 HCC nodules were detected on CTHA/CTAP, patients first received transarterial chemoembolization (TACE) immediately after CTHA/CTAP, followed later by RFA to achieve complete ablation of the tumor nodules.

admission, and RFA was performed 2 or 3 days later, given that the total number of HCC nodules remained <4. When ≥4 HCC nodules were detected on CTHA/CTAP, patients first received transarterial chemoembolization (TACE) immediately after CTHA/CTAP, followed later by RFA to achieve complete ablation of the tumor nodules. Radiographic procedures For the diagnosis of HCC at study entry, intravenous contrast-enhanced dynamic CT was performed on an outpatient basis using an X-ray CT device with 4, 8, or 16 detector rows (Aquilion 4/16; Toshiba, Tokyo, Japan; LightSpeed Qx/I, LightSpeed Ultra; GE Healthcare, Milwaukee, WI). Images were obtained during the early arterial, late arterial, and equilibrium phases at 28, 40, and 120 s after starting the intravenous bolus injection of iopamidol (Iopamiron; Nihon Schering, Osaka, Japan) or iohexol (Omnipaque; Daiichi Sankyo, Tokyo, Japan) at a rate of 2.3–3.3 ml/s with a power injector. The total dose of iodine was 0.7 g/kg body weight, with an upper limit of 37 g iodine. The injection time for the contrast material was 30 s. Images were reconstructed with a section thickness of 2.5 mm and a reconstruction interval of 1.5 mm, and were reviewed by experienced radiologists.

3 ml/s with a power injector. The total dose of iodine was 0.7 g/kg body weight, with an upper limit of 37 g iodine. The injection time for the contrast material was 30 s. Images were reconstructed with a section thickness of 2.5 mm and a reconstruction interval of 1.5 mm, and were reviewed by experienced radiologists. CTHA/CTAP was performed on an inpatient basis. First, a 4-Fr modified Shepherd-hook catheter and a 4-Fr hepatic-curve catheter were placed in the celiac artery and superior mesenteric artery, respectively, through bilateral femoral arteries, according to Seldinger's method. Digital subtraction angiography was performed from the celiac artery to evaluate hepatic artery anatomy. A microcatheter was inserted through the 4-Fr catheter and placed in the proper or common hepatic artery for hepatic arteriography.

respectively, through bilateral femoral arteries, according to Seldinger's method. Digital subtraction angiography was performed from the celiac artery to evaluate hepatic artery anatomy. A microcatheter was inserted through the 4-Fr catheter and placed in the proper or common hepatic artery for hepatic arteriography. The CTAP catheter was placed in the superior mesenteric artery in all cases. In the case of a replaced or accessory right hepatic artery, the catheter was inserted well beyond the origin of the hepatic artery to prevent contrast medium overflow into the hepatic artery. Less than 30 ml of contrast agent, which was diluted to 100 mg I/ml, was used before the CTHA/CTAP study. First, CTAP was performed using 90 ml nonionic contrast medium diluted to 100 mg I/ml, and then CT scanning was performed 30 s after the start of the injection at a rate of 3.0 ml/s. Multidetector-row CT images were obtained during a single breath hold in a longitudinal direction with collimation of 1 mm, table speed of 30 mm/s, 120 kVp, and 300 mAs. CTHA was performed at least 5 min after CTAP, using the same parameters. CT scanning was performed at 10 and 45 s after the start of contrast medium injection into the microcatheter at a rate of 2.0–2.5 ml/s. A total of 30–50 ml contrast agent diluted to 100 mg I/ml was used. When the liver was perfused by two or more hepatic arteries such as a replaced right hepatic artery, accessory right hepatic artery, or left hepatic artery downstream of the left gastric artery, CTHA was performed from each of the respective arteries. A diagnosis of typical HCC on CTHA/CTAP was defined as a round hypervascular nodule on CTHA with a defect on CTAP, accompanied by corona enhancement during the second phase of CTHA or hypoattenuation during the equilibrium phase of prior dynamic CT ((10,29)).

ry, CTHA was performed from each of the respective arteries. A diagnosis of typical HCC on CTHA/CTAP was defined as a round hypervascular nodule on CTHA with a defect on CTAP, accompanied by corona enhancement during the second phase of CTHA or hypoattenuation during the equilibrium phase of prior dynamic CT ((10,29)). TACE was additionally performed when ≥4 HCC nodules were detected on CTHA/CTAP, as evaluated at the time by the operating radiologist. The procedure used 3.0 ml contrast medium, 30 mg doxorubicin (Adriacin; Kyowahakko Kirin, Tokyo, Japan), and 3.0 ml iodized oil (Lipiodol Ultra-Fluid; Guerbet Japan, Tokyo, Japan). The amounts of contrast medium and iodized oil in this suspension were arbitrarily adjusted according to tumor size. This agent was injected into each feeder of the HCC, followed by infusion of 2-mm-diameter gelatin sponge particles (Gelpart; Nihonkayaku, Tokyo, Japan). CTHA/CTAP images were scrutinized by two experienced radiologists, who made the final diagnosis. The radiologists were not blinded to information regarding the preceding conventional dynamic CT. Preceding intravenous contrast-enhanced dynamic CT was retrospectively reviewed for nodules newly diagnosed by CTHA/CTAP to determine whether the nodules could have been detected on dynamic CT.

ho made the final diagnosis. The radiologists were not blinded to information regarding the preceding conventional dynamic CT. Preceding intravenous contrast-enhanced dynamic CT was retrospectively reviewed for nodules newly diagnosed by CTHA/CTAP to determine whether the nodules could have been detected on dynamic CT. Radiofrequency ablation RFA was performed on an inpatient basis. The precise procedure of RFA is described elsewhere (30). All RFA procedures were performed percutaneously under ultrasonographic guidance. We used a 17-gauge cooled-tip electrode (Cool-Tip; RF Ablation System, Covidien, Boulder, Colombia, CO) for RFA. Radiofrequency energy was delivered for 6–12 min for each application. For large tumors, the electrode was repeatedly inserted into different sites, such that the entire tumor could be enveloped by assumed necrotic volumes. A CT scan with a 5-mm section thickness was performed 1–3 days after RFA to evaluate technical effectiveness. Complete ablation was defined as hypoattenuation of the entire tumor. We intended to ablate not only the tumor but also some of the liver parenchyma surrounding it. When we suspected that some portion of tumor remained nonablated, RFA was repeated. We did not predefine the procedure number in a treatment: treatment was generally continued until CT imaging demonstrated necrosis of the entire tumor.

d to ablate not only the tumor but also some of the liver parenchyma surrounding it. When we suspected that some portion of tumor remained nonablated, RFA was repeated. We did not predefine the procedure number in a treatment: treatment was generally continued until CT imaging demonstrated necrosis of the entire tumor. Follow-up The follow-up regimen after RFA consisted of blood tests and monitoring of tumor markers in an outpatient setting. Ultrasonography and dynamic CT were performed every 4 months. Tumor recurrence was defined as a newly developed lesion on a dynamic CT that showed hyperattenuation in the arterial phase with washout in the late phase. Recurrent site was categorized as intrahepatic recurrence distant from ablated nodules, local tumor progression defined as the appearance of viable cancer tissue touching the ablated nodules, and extrahepatic metastasis (31). The follow-up was censored in February 2011 when 2 years had passed after the enrollment of patient 280. No interim analysis was specified in the protocol.

tant from ablated nodules, local tumor progression defined as the appearance of viable cancer tissue touching the ablated nodules, and extrahepatic metastasis (31). The follow-up was censored in February 2011 when 2 years had passed after the enrollment of patient 280. No interim analysis was specified in the protocol. End points The primary end point was recurrence-free survival, where both recurrence and death were treated as an event. We intended to ablate all detected nodules in both groups. When additional nodules were detected by CTHA/CTAP, the newly detected nodules were also ablated. When >3 nodules were diagnosed as HCC by CTHA/CTAP, we performed TACE and subsequently intended to ablate all of the nodules. When nonablated viable tumors were detected by CT for treatment evaluation, those cases were treated as an event 120 days after randomization. Even when newly detected nodules showed dense Lipiodol deposits after TACE, the nodules were considered as viable if the nodules were nonablated. Secondary end points were the number of additional nodules detected by CTHA/CTAP, the proportion of patients with complete ablation, overall survival, and safety of CTHA/CTAP and RFA. Complications were defined according to the guidelines of the Society of Interventional Radiology (32). According to the guidelines, major complications were defined as those that required therapy or prolonged hospitalization, or left permanent adverse sequelae, or death.

rall survival, and safety of CTHA/CTAP and RFA. Complications were defined according to the guidelines of the Society of Interventional Radiology (32). According to the guidelines, major complications were defined as those that required therapy or prolonged hospitalization, or left permanent adverse sequelae, or death. Statistical analysis This study was designed to detect a 15% increase in 2-year recurrence-free survival in the CTHA/CTAP group from an anticipated 35% in the control group. To detect this difference with a power of 80% and type I error of 5% (two-sided test), we needed 280 patients (140 for each arm). Differences between groups for each characteristic were tested for significance with Fisher's exact test for categorical variables and t-test for continuous variables. All data necessary for analysis was corrected in the main computer server system of University of Tokyo, Department of Gastroenterology.

for each arm). Differences between groups for each characteristic were tested for significance with Fisher's exact test for categorical variables and t-test for continuous variables. All data necessary for analysis was corrected in the main computer server system of University of Tokyo, Department of Gastroenterology. Recurrence-free survival and overall survival were calculated using the Kaplan–Meier method and were compared by the log-rank test. Cox proportional hazard regression was used to calculate hazard ratios with 95% confidence interval (CI) between the groups in univariate and multivariate settings. The primary end point was evaluated in subgroups according to the following characteristics: age, sex, body mass index, treatment naivety, hepatitis B surface antigen (HBsAg) positivity, hepatitis C virus antibody positivity, tumor size, tumor number, platelet count, tumor marker positivity for α-fetoprotein (AFP), lens culinaris agglutinin-reactive fraction of AFP, and des-γ-carboxy prothrombin. An adjusted hazard ratio comparing the groups was calculated using multivariate Cox regression with factors that showed significance in univariate analysis. Data at entry were used for the analyses. A post hoc analysis comparing the recurrence-free survival of those with and without newly diagnosed HCC in the CTHA/CTAP group was performed.

tio comparing the groups was calculated using multivariate Cox regression with factors that showed significance in univariate analysis. Data at entry were used for the analyses. A post hoc analysis comparing the recurrence-free survival of those with and without newly diagnosed HCC in the CTHA/CTAP group was performed. All analyses were performed on an intention-to-treat basis. Differences with a two-sided P value of <0.05 were considered statistically significant. Data processing and analysis were performed with S-PLUS ver. 7 (TIBCO Software, Palo Alto, CA). Finally, all authors had access to the study data and had reviewed and approved the final manuscript. RESULTS Patient enrollment According to the study protocol, the registration started from September 2004 for 5 years and the follow-up was censored in February 2011 when 2 years had passed after the enrollment of patient 280. During the study period, 280 of 591 (47.4%) eligible patients agreed to participate in the trial, and 140 of these were randomly assigned to undergo CTHA/CTAP before RFA. Three patients declined to undergo CTHA/CTAP after assignment. A total of 140 patients were randomly assigned to the control group. One patient assigned to the control group received CTHA/CTAP because of strong preference (Figure 1).

in the trial, and 140 of these were randomly assigned to undergo CTHA/CTAP before RFA. Three patients declined to undergo CTHA/CTAP after assignment. A total of 140 patients were randomly assigned to the control group. One patient assigned to the control group received CTHA/CTAP because of strong preference (Figure 1). Treatment In 45 (32.4%) patients, 75 nodules with a median diameter of 8 mm (range, 2–20) were additionally diagnosed by experienced radiologists as definite HCC on CTHA/CTAP. The detailed characteristics of newly diagnosed nodules have been reported previously (33). In 17 patients, the number of HCC nodules exceeded 3 after CTHA/CTAP, and TACE was performed subsequently. We intended to ablate all nodules by RFA including additionally detected nodules. In 122 patients, there were ≤3 HCC nodules, and complete ablation was obtained in 121 patients (99.2%). Among 17 patients treated with TACE, 14 (82.4%) subsequently underwent RFA and complete ablation was obtained in 13 (92.9%) patients. The remaining 3 patients (17.6%) did not undergo RFA because of tumor nodule multiplicity in 2 patients and simultaneously diagnosed malignant B-cell lymphoma in the third patient. Among 140 patients who were assigned to the control group, 137 (97.9%) were treated with RFA, and complete ablation was obtained in 136 (99.3%) patients. One patient withdrew consent and underwent hepatic resection. Two patients refused to receive any treatment and were lost to follow-up. Finally, 139 (99.3%) patients in the CTHA/CTAP group and 138 (98.6%) patients in the control group were included in the analysis.

omplete ablation was obtained in 136 (99.3%) patients. One patient withdrew consent and underwent hepatic resection. Two patients refused to receive any treatment and were lost to follow-up. Finally, 139 (99.3%) patients in the CTHA/CTAP group and 138 (98.6%) patients in the control group were included in the analysis. Patient characteristics There was no statistically significant difference in patient characteristics between the groups ( Table 1). Median age at enrollment was 70 years, and approximately two-thirds of patients were male. Approximately 55% of patients were treatment-naive cases and the remaining patients had a history of previous treatment. Among those previously treated patients, the median interval between the initial treatment and the study enrollment was 42 (interquartile range, 22–65) months in the CTHA/CTAP group and 30 (20–61) months in the control group. There was no statistically significant difference between the two groups (P=0.72). The total number of HCC nodules detected in original contrast-enhanced dynamic CT was 197 (101 patients were uninodular and the rest were multinodular) in the CTHA/CTAP group and 196 (98 patients were uninodular and the rest were multinodular) in the control group.

ignificant difference between the two groups (P=0.72). The total number of HCC nodules detected in original contrast-enhanced dynamic CT was 197 (101 patients were uninodular and the rest were multinodular) in the CTHA/CTAP group and 196 (98 patients were uninodular and the rest were multinodular) in the control group. Recurrence By the end of the follow-up, tumor recurrence was identified in 109 patients (78.4%) in the CTHA/CTAP group and 112 patients (81.2%) in the control group. The distribution of recurrent site was intrahepatic distant recurrence (N=98), local tumor progression (N=7), both (N=1), and extrahepatic metastasis (N=3) in the CTHA/CTAP group and intrahepatic distant recurrence (N=103), local tumor progression (N=4), both (N=2), and extrahepatic metastasis (N=3) in the control group. Five patients (3.6%) in the CTHA/CTAP group and 1 patient (0.7%) in the control group in whom complete ablation could not be obtained by RFA were treated as recurrence on 120 days after randomization when the first follow-up CT would have been scheduled. In each group, four patients died without recurrence. The cumulative recurrence-free survival rates at 1, 2, and 3 years were 60.1, 29.0, and 18.9% in the CTHA/CTAP group and 52.2, 29.7, and 23.1% in the control group, respectively (Figure 2a). The difference between the two groups was not statistically significant (P=0.66 by log-rank test; hazard ratio, 0.94 for CTHA/CTAP vs. control; 95% CI, 0.73–1.22). The CTHA/CTAP group showed better recurrence-free survival with marginal statistical significance in the subgroups with higher AFP or AFP-L3 values (Figure 3).

difference between the two groups was not statistically significant (P=0.66 by log-rank test; hazard ratio, 0.94 for CTHA/CTAP vs. control; 95% CI, 0.73–1.22). The CTHA/CTAP group showed better recurrence-free survival with marginal statistical significance in the subgroups with higher AFP or AFP-L3 values (Figure 3). Univariate Cox regression analysis identified older age (P=0.01), hepatitis C virus antibody positivity (P=0.001), lower albumin level (P=0.04), recurrent cases (P<0.001), multinodular HCC (P<0.001), and higher AFP level (P=0.02) as significant predictors for recurrence-free survival ( Table 2). Adjusted hazard ratio of the CTHA/CTAP group vs. the control group by multivariate Cox regression analysis was 0.86 (95% CI, 0.67–1.12; P=0.27, Table 3). Overall survival By the end of the follow-up, 51 patients (36.7%) in the CTHA/CTAP group and 45 patients (32.6%) in the control group died. The cumulative overall survival rates at 3 and 5 years were 79.7 and 56.4% in the CTHA/CTAP group and 86.8 and 60.1% in the control group, respectively (Figure 2b). There was no statistically significant difference between the groups (P=0.50 by log-rank test; hazard ratio, 1.15, 95% CI, 0.77–1.71). Safety No procedural complications attributable to CTHA/CTAP or TACE were observed. Major complications related to RFA were observed in 2 patients (1.4%) in the CTHA/CTAP group (2 with neoplastic seeding) and in 3 patients (2.2%) in the control group (1 each with hepatic infarction, hemothorax, and neoplastic seeding). There was no procedure-related death.

ributable to CTHA/CTAP or TACE were observed. Major complications related to RFA were observed in 2 patients (1.4%) in the CTHA/CTAP group (2 with neoplastic seeding) and in 3 patients (2.2%) in the control group (1 each with hepatic infarction, hemothorax, and neoplastic seeding). There was no procedure-related death. Recurrence-free survival between those with and without additional nodules in CTHA/CTAP group As a post hoc analysis, we compared the recurrence-free survival between those with (N=45) and without (N=92) additional HCC nodules diagnosed by CTHA/CTAP. As compared with those in whom additional HCC nodules were not detected by CTHA/CTAP, those with additional nodules included more HBsAg-negative patients (97.7 vs. 78.3%, P=0.002), previously treated patients (62.2 vs. 23.9%, P=0.006), and patients with multiple HCC nodules on dynamic CT (44.4 vs. 17.4%, P=0.002). Patients with additional nodule by CTHA/CTAP showed significantly poorer recurrence-free survival than those without additional nodules (P=0.003, Figure 2c).

(97.7 vs. 78.3%, P=0.002), previously treated patients (62.2 vs. 23.9%, P=0.006), and patients with multiple HCC nodules on dynamic CT (44.4 vs. 17.4%, P=0.002). Patients with additional nodule by CTHA/CTAP showed significantly poorer recurrence-free survival than those without additional nodules (P=0.003, Figure 2c). DISCUSSION An advance in diagnostic technology generally indicates improved sensitivity or specificity, which corresponds to the detection of smaller lesions with a clearer view in imaging modalities. In our previous study, we showed that 75 nodules with a mean diameter of 8.7 mm (range, 2–20 mm) in 45 (33%) of 139 patients who underwent CTHA/CTAP were additionally diagnosed as definite HCC, compared with dynamic CT examination (33). However, no significant difference was observed in terms of recurrence-free survival between those who did and did not undergo CTHA/CTAP before RFA. One reason for this discrepancy may be that the impact of CTHA/CTAP on recurrence reduction was diluted by a long-term follow-up of >2 years. It is unlikely that CTHA/CTAP could detect small nodules that would be detected ≥2 years later by conventional dynamic CT. In fact, the number of recurrences identified within 1 year after enrollment was lower in the CTHA/CTAP group than the control group (54 vs. 65, data not shown).

a long-term follow-up of >2 years. It is unlikely that CTHA/CTAP could detect small nodules that would be detected ≥2 years later by conventional dynamic CT. In fact, the number of recurrences identified within 1 year after enrollment was lower in the CTHA/CTAP group than the control group (54 vs. 65, data not shown). Another reason could be that fewer patients achieved complete ablation of target nodules in the CTHA/CTAP group than in the control group. The additionally diagnosed HCC nodules were small, and detection of these nodules by ultrasonography was difficult. Recent technologies such as contrast ultrasonography or fusion imaging, which can improve the accuracy of ablation techniques ((34,35,36)), may increase the probability of detection of smaller nodules before RFA. Precise evaluation of the stage of progression is important for deciding on treatment procedures in HCC management. Seventeen patients in the CTHA/CTAP group were diagnosed with ≥4 nodules by CTHA/CTAP, which is not considered suitable for RFA according to widely used criteria.

Another reason could be that fewer patients achieved complete ablation of target nodules in the CTHA/CTAP group than in the control group. The additionally diagnosed HCC nodules were small, and detection of these nodules by ultrasonography was difficult. Recent technologies such as contrast ultrasonography or fusion imaging, which can improve the accuracy of ablation techniques ((34,35,36)), may increase the probability of detection of smaller nodules before RFA. Precise evaluation of the stage of progression is important for deciding on treatment procedures in HCC management. Seventeen patients in the CTHA/CTAP group were diagnosed with ≥4 nodules by CTHA/CTAP, which is not considered suitable for RFA according to widely used criteria. In our previous study, we showed that recurrence as opposed to initial occurrence, multinodularity on dynamic CT, and HBsAg negativity were significant predictors for finding additional HCC by CTHA/CTAP (33). In fact, the CTHA/CTAP group showed better outcomes in the subgroups with HBsAg-negative cases, previously treated patients, and multinodular HCC. However, post hoc analysis comparing recurrence-free survival of those with and without additional nodules detected by CTHA/CTAP showed that those with a higher probability of additional nodules were also at a higher risk of recurrence. The advantage of CTHA/CTAP in finding more HCC nodules might be counter balanced by the higher risk of recurrence.

aring recurrence-free survival of those with and without additional nodules detected by CTHA/CTAP showed that those with a higher probability of additional nodules were also at a higher risk of recurrence. The advantage of CTHA/CTAP in finding more HCC nodules might be counter balanced by the higher risk of recurrence. This study has several limitations. First, the additional nodules detected by CTHA/CTAP were not confirmed histologically. Therefore, we cannot exclude the possibility of overdiagnosis. Second, 45% of the patients had a history of previous treatment including resection, RFA, and TACE. Those previous treatments might substantially alter the hemodynamic status in the liver and affect the accuracy of CTHA/CTAP. On the other hand, in the previously treated cases, the radiologists could refer to the past series of dynamic CT during performing CTHA/CTAP, which might improve the accuracy of CTHA/CTAP as compared with treatment-naive cases. Third, 17 patients in the CTHA/CTAP group underwent TACE as a salvage treatment because total number of HCC nodules exceeded 3 after CTHA/CTAP. This might affect the recurrence-free and overall survival in the CTHA/CTAP group. Our results may be extrapolated to other imaging modalities including gadoxetic acid–enhanced magnetic resonance imaging and second-generation contrast ultrasonography ((37,38)). These newly developed modalities also make possible the detection of small nodules that are invisible by dynamic CT. However, better diagnosis does not necessarily lead to better primary outcome.

ities including gadoxetic acid–enhanced magnetic resonance imaging and second-generation contrast ultrasonography ((37,38)). These newly developed modalities also make possible the detection of small nodules that are invisible by dynamic CT. However, better diagnosis does not necessarily lead to better primary outcome. In conclusion, CTHA/CTAP before RFA resulted in improved HCC diagnosis and detection of additional nodules in one-third of the study participants. However, it did not improve recurrence-free survival. The indications for CTHA/CTAP should be evaluated carefully. Study protocol URL: https://upload.umin.ac.jp/cgi-open-bin/ctr/ctr.cgi?function=brows&action=brows&recptno=R000000117&type=summary&language=E. STUDY HIGHLIGHTS Guarantor of the article: Ryosuke Tateishi, MD, PhD. Specific author contributions: Conception and design: R.T., M.A., N.Y., T.G., S.S., H.Y., Y.M., and M.O.; analysis: R.T. and Y.M.; treatment and data collection: T.O., R.T., M.A., S.M., M.S., K.U., T.A., K.E., Y.K., T.G., and S.S.; drafting article: T.O.; critical revision: R.T., M.A., H.Y., K.O., and K.K. Financial support: This work was supported by Health Sciences Research grants of The Ministry of Health, Labour and Welfare of Japan (Research on Hepatitis). No additional external funding was received for this study. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Potential competing interests: None.

of identifying patients who are most likely to benefit from aggressive treatment is imperative, and poses great challenge for the clinicians involved in the care of these patients (5,7). Unfortunately, the quest for an accurate prognostic score applicable to these patients in clinical practice has remained elusive (5). The Child-Pugh score and the Model for End-Stage Liver Disease (MELD) are widely utilized for grading of the severity of liver disease and for liver graft allocation for patients with decompensated cirrhosis. They are also used to assess prognosis for patients with cirrhosis admitted to ICU. However, general ICU prognostic scores, such as the Acute Physiology and Chronic Health Evaluation (APACHE) and the Sequential Organ Failure Assessment (SOFA) scores, have proven more accurate than the currently used liver-specific models in predicting mortality, despite the fact that they are not derived specifically from populations of patients with cirrhosis (3,8,9,10,11,12,13,14,15,16). This finding reinforces the contribution of multi-organ dysfunction in determining outcome, irrespective of the nature of underlying disease, and holds true even for patients with cirrhosis.

Financial support: This work was supported by Health Sciences Research grants of The Ministry of Health, Labour and Welfare of Japan (Research on Hepatitis). No additional external funding was received for this study. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Potential competing interests: None. Figure 1 Patient enrollment and outcomes. CTAP, computed tomography during arterial portography; CTHA, computed tomography during hepatic arteriography; RFA, radiofrequency ablation; TACE, transarterial chemoembolization. Figure 2 Kaplan–Meier estimate of the recurrence-free survival and overall survival. (a) The cumulative recurrence-free survival rates at 1, 2, and 3 years were 60.1, 29.0, and 18.9% in the CTHA/CTAP group and 52.2, 29.7, and 23.1% in the control group, respectively. (b) The cumulative overall survival rates at 3 and 5 years were 79.7 and 56.4% in CTHA/CTAP group and 86.8 and 60.1% in the control group, respectively. (c) Patients with an additional nodule detected by CTHA/CTAP showed significantly poorer recurrence-free survival than those without an additional nodule. CTAP, computed tomography during arterial portography; CTHA, computed tomography during hepatic arteriography.

and 86.8 and 60.1% in the control group, respectively. (c) Patients with an additional nodule detected by CTHA/CTAP showed significantly poorer recurrence-free survival than those without an additional nodule. CTAP, computed tomography during arterial portography; CTHA, computed tomography during hepatic arteriography. Figure 3 Recurrence-free survival of subgroups by Cox proportional hazard regression according to clinical characteristics at study entry. AFP, α-fetoprotein; BMI, body mass index; CI, confidence interval; CT, computed tomography; CTAP, computed tomography during arterial portography; CTHA, computed tomography during hepatic arteriography; DCP, des-γ-carboxy prothrombin; HBsAg, hepatitis B surface antigen; HCVAb, hepatitis C virus antibody; yr, year.

try. AFP, α-fetoprotein; BMI, body mass index; CI, confidence interval; CT, computed tomography; CTAP, computed tomography during arterial portography; CTHA, computed tomography during hepatic arteriography; DCP, des-γ-carboxy prothrombin; HBsAg, hepatitis B surface antigen; HCVAb, hepatitis C virus antibody; yr, year. Table 1 Baseline characteristics of the patientsa Characteristics CTHA/CTAP (N=139) Control (N=138) P value Age, years 70 (63–74) 70 (64–75) 0.43 Male, n (%) 93 (67) 86 (62) 0.42 Alcohol >80 g/day, n (%) 23 (17) 20 (15) 0.82 BMI (kg/m2) 23.1 (21.4–25.1) 23.4 (21.2–25.3) 0.48 Viral markers HCVAb positive, n (%) 104 (75) 99 (72) 0.59 HBsAg positive, n (%) 21 (15) 20 (14) 1 Serum albumin (g/dl) 3.8 (3.6–4.1) 3.9 (3.6–4.1) 0.20 Total bilirubin (mg/dl) 0.8 (0.6–1.0) 0.8 (0.6–1.0) 0.31 AST (IU/l) 56 (34–69) 57 (33–70) 0.84 ALT (IU/l) 54 (29–63) 57 (27–73) 0.61 Platelet count (× 103/μl) 128 (89–163) 130 (91–159) 0.88 Prothrombin activity (%) 80 (72–90) 81 (74–87) 0.39 Treatment-naive case, n (%) 77 (55) 74 (54) 0.81 Previously treated case, n (%) 62 (45) 64 (46) Resection, n (%)b 15 (24) 16 (25) 0.27 RFA, n (%)b 46 (74) 45 (70) Ethanol injection, n (%)b 10 (16) 3 (4.6) TACE, n (%)b 11 (18) 7 (11) Tumor size (cm) 1.6 (1.2–2.0) 1.7 (1.2–2.0) 0.91 Single nodule, n (%) 101 (73) 98 (71) 0.76 AFP >100 ng/ml, n (%) 23 (17) 24 (17) 0.85 DCP >100 mAU/ml, n (%) 16 (12) 22 (16) 0.28 AFP-L3 >15%, n (%) 16 (12) 15 (11) 0.86 AFP, α-fetoprotein; AFP-L3, lens culinaris agglutinin-reactive fraction of AFP; ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; CTHA/CTAP, computed tomography during hepatic arteriography and arterial portography; DCP, des-γ-carboxy prothrombin; HBsAg, hepatitis B surface antigen; HCVAb, hepatitis C virus antibody; RFA, radiofrequency ablation, TACE, transarterial chemoembolization.

ferase; AST, aspartate aminotransferase; BMI, body mass index; CTHA/CTAP, computed tomography during hepatic arteriography and arterial portography; DCP, des-γ-carboxy prothrombin; HBsAg, hepatitis B surface antigen; HCVAb, hepatitis C virus antibody; RFA, radiofrequency ablation, TACE, transarterial chemoembolization. a Data are expressed as median (25th–75th percentiles) or number (percent). b Including overlap. Table 2 Univariate Cox's proportional hazard regression analysis of the risk for recurrence-free survival Variable Hazard ratio (95% CI) P value CTHA/CTAP vs. control 0.94 (0.73–1.22) 0.66 Age (per year) 1.02 (1.00–1.04) 0.01 Female vs. male 1.02 (0.78–1.34) 0.88 Alcohol >80 g/day 1.02 (0.88–1.17) 0.81 HCVAb positive 1.69 (1.23–2.31) 0.001 BMI (per 1.0 kg/m2) 1.02 (0.98–1.06) 0.35 Albumin (per 1.0 g/dl) 0.72 (0.52–0.98) 0.04 Total bilirubin (per 1.0 mg/dl) 1.02 (0.97–1.07) 0.51 AST >40 IU/l 1.14 (0.99–1.31) 0.07 ALT >40 IU/l 1.05 (0.92–1.20) 0.45 Platelet count >10 × 103/μl 0.89 (0.78–1.01) 0.08 Recurrent case 2.33 (1.79–3.02) <0.001 Tumor size of maximal nodule >2.0 cm 0.97 (0.85–1.10) 0.62 Multinodular 1.38 (1.20–1.59) <0.001 AFP >100 ng/ml 1.21 (1.03–1.43) 0.02 DCP >100 mAU/ml 0.99 (0.82–1.20) 0.93 AFP-L3 >15% 1.20 (0.99–1.46) 0.07 AFP, α-fetoprotein; AFP-L3, lens culinaris agglutinin-reactive fraction of AFP; ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; CI, confidence interval; CTHA/CTAP, computed tomography during hepatic arteriography and arterial portography; DCP, des-γ-carboxy prothrombin; HCVAb, hepatitis C virus antibody.

-L3, lens culinaris agglutinin-reactive fraction of AFP; ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; CI, confidence interval; CTHA/CTAP, computed tomography during hepatic arteriography and arterial portography; DCP, des-γ-carboxy prothrombin; HCVAb, hepatitis C virus antibody. Table 3 Multivariate Cox's proportional hazard regression analysis of the risk for recurrence-free survival Variable Hazard ratio (95% CI) P value CTHA/CTAP vs. control 0.86 (0.0.67–1.12) 0.27 Age (per year). 1.01 (0.99–1.02) 0.36 HCVAb positive 1.36 (0.98–1.89) 0.07 Albumin (per 1.0 g/dl) 0.75 (0.53–1.07) 0.11 Recurrent case 2.21 (1.69–2.89) <0.001 Multinodular 1.69 (1.27–2.25) <0.001 AFP >100 ng/ml 1.41 (0.996–1.98) 0.052 AFP, α-fetoprotein; CI, confidence interval; CTHA/CTAP, computed tomography during hepatic arteriography and arterial portography; HCVAb, hepatitis C virus antibody.

despite the fact that they are not derived specifically from populations of patients with cirrhosis (3,8,9,10,11,12,13,14,15,16). This finding reinforces the contribution of multi-organ dysfunction in determining outcome, irrespective of the nature of underlying disease, and holds true even for patients with cirrhosis. There are only three prognostic models (3,9,15) that have been developed from cohorts (n=111, 196, and 312) of ICU patients with cirrhosis. Some incorporate parameters such as serum sodium and lactate levels, which are highly predictive of outcome in the context of acute deterioration of chronic liver disease. Although most of these models demonstrate good discriminative ability, their calibration, i.e., the concordance between predicted and observed outcome, is modest at best. Therefore, to date none of the proposed models have been widely used.

INTRODUCTION Ulcerative colitis is a chronic inflammatory disorder of the colon defined by relapsing and remitting episodes (1). Mild-to-moderate disease symptoms are treated with mesalamine (2,3). Patients who fail to respond to mesalamine are treated with systemic steroids, azathioprine, and infliximab (2,3), which are all associated with serious toxicities (4). Additional treatments that can be used as first-line or salvage therapy in patients failing mesalamine are needed. Andrographis paniculata, a member of the plant family Acanthaceae, is an herbal remedy used in China, India, Thailand, and other Asian countries. A. paniculata has been used to treat upper respiratory tract infections (5,6,7). An ethanol extract of A. paniculata yields an herbal mixture containing mainly diterpene lactones including andrographolide. Andrographolide, which is used as a marker compound, comprises <10% of the mixture. A. paniculata extract has in vitro inhibitory activity against TNF-α, IL-1β, and NF-κB (8,9). A pilot study of A. paniculata extract (HMPL-004) 1,200 mg/day suggested similar efficacy to mesalamine for ulcerative colitis (10). Our study was an 8-week, placebo-controlled trial comparing A. paniculata extract (HMPL-004) with placebo in patients with mildly to moderately active ulcerative colitis.

Andrographis paniculata, a member of the plant family Acanthaceae, is an herbal remedy used in China, India, Thailand, and other Asian countries. A. paniculata has been used to treat upper respiratory tract infections (5,6,7). An ethanol extract of A. paniculata yields an herbal mixture containing mainly diterpene lactones including andrographolide. Andrographolide, which is used as a marker compound, comprises <10% of the mixture. A. paniculata extract has in vitro inhibitory activity against TNF-α, IL-1β, and NF-κB (8,9). A pilot study of A. paniculata extract (HMPL-004) 1,200 mg/day suggested similar efficacy to mesalamine for ulcerative colitis (10). Our study was an 8-week, placebo-controlled trial comparing A. paniculata extract (HMPL-004) with placebo in patients with mildly to moderately active ulcerative colitis. METHODS The study was designed by two academic investigators William Sandborn and Stephan Targan, Vera Byers from Immunology Inc., and Tom Tang from Hutchison MediPharma Ltd. Data were collected by Clinical Research Management Inc. (Agawam, MA) and Omnicare Clinical Research (Frankfurt, Germany), and were analyzed by Everest Clinical Research Services Inc. (Ontario, Canada). William Sandborn wrote the first draft of the manuscript. The academic authors vouch for the veracity and completeness of the data and data analyses.

Research Management Inc. (Agawam, MA) and Omnicare Clinical Research (Frankfurt, Germany), and were analyzed by Everest Clinical Research Services Inc. (Ontario, Canada). William Sandborn wrote the first draft of the manuscript. The academic authors vouch for the veracity and completeness of the data and data analyses. Patients This multicenter, randomized, double-blind, placebo-controlled trial was conducted at 52 centers in 5 countries (United States, Canada, Germany, Romania, and Ukraine) between February 2008 and October 2009 (4 sites in Germany were activated but failed to recruit any patients). The institutional review board at each center approved the protocol. Patients gave written informed consent. Eligible patients were at least 18 years of age and had a confirmed diagnosis of ulcerative colitis. Patients had a Mayo Score of 4–10 points (11) and mildly to moderately active disease on sigmoidoscopy (endoscopic subscore of at least 1) while receiving either oral mesalamine (or equivalent medications sulfasalazine, balsalazide, and olsalazine) for at least 4 weeks or no medical therapy.

osis of ulcerative colitis. Patients had a Mayo Score of 4–10 points (11) and mildly to moderately active disease on sigmoidoscopy (endoscopic subscore of at least 1) while receiving either oral mesalamine (or equivalent medications sulfasalazine, balsalazide, and olsalazine) for at least 4 weeks or no medical therapy. Patients with Crohn's disease or indeterminate colitis, severe ulcerative colitis (Mayo Score of 11 or 12 points, toxic mega-colon, toxic colitis), previous colonic surgery or probable requirement for intestinal surgery within 12 weeks, enteric infection within 2 weeks, a history of tuberculosis, a positive chest X-ray or tuberculin protein-purified derivative skin test, active infection with hepatitis B or any infection with hepatitis C, infection with human immunodeficiency virus, cancer within 5 years, inadequate bone marrow, hepatic, or renal function, a history of alcohol or drug abuse that would interfere with the study, significant concurrent medical diseases, allergy to plants in the Acanthacea family, and women who were pregnant or breastfeeding were not eligible. Patients receiving oral or rectal steroids within 1 month, rectal mesalamine within 1 week, antibiotics within 2 weeks, or azathioprine, 6-mercaptopurine, anti-tumor necrosis factor agents, or immunosuppressive therapy within 6 weeks were also excluded.

y, and women who were pregnant or breastfeeding were not eligible. Patients receiving oral or rectal steroids within 1 month, rectal mesalamine within 1 week, antibiotics within 2 weeks, or azathioprine, 6-mercaptopurine, anti-tumor necrosis factor agents, or immunosuppressive therapy within 6 weeks were also excluded. Study design Patients were randomly assigned to receive oral capsules containing A. paniculata ethanol extract (HMPL-004; Hutchison MediPharma Ltd., Shanghai, China) at doses of 1,200 mg or 1,800 mg or placebo, administered in three divided doses. Patients were treated for 8 weeks and followed through week 12. Randomization was performed centrally using a block randomization schedule stratified by concurrent mesalamine use (yes or no) and country/geographic region (North East USA, Mid-East USA, South East USA, Western USA, Canada, Ukraine, and Romania). Oral mesalamine was continued at a stable dose.

ollowed through week 12. Randomization was performed centrally using a block randomization schedule stratified by concurrent mesalamine use (yes or no) and country/geographic region (North East USA, Mid-East USA, South East USA, Western USA, Canada, Ukraine, and Romania). Oral mesalamine was continued at a stable dose. Follow-up, efficacy, and safety evaluations A colonoscopy or flexible sigmoidoscopy was performed and the Mayo Score was determined at weeks 0 and 8. A partial Mayo Score (Mayo Score without endoscopy) was determined at all visits between weeks 0 and 8 inclusive. Clinical response was defined as a decrease from baseline in the total Mayo Score by at least 3 points and at least 30% with an accompanying decrease in rectal bleeding subscore of at least 1 point or a absolute rectal bleeding subscore of 0 or 1 point (12,13). Clinical remission was defined as a total Mayo Score of 2 points or lower, with no individual subscore exceeding 1 point (12,13). Mucosal healing was defined as a decrease from baseline in the endoscopy subscore by at least 1 point and an absolute endoscopy subscore of 0 or 1 point (12,13). Adverse events and concomitant medications were followed through week 12. Blood samples were collected at weeks 0 and 8 for C-reactive protein (CRP) concentrations.

Follow-up, efficacy, and safety evaluations A colonoscopy or flexible sigmoidoscopy was performed and the Mayo Score was determined at weeks 0 and 8. A partial Mayo Score (Mayo Score without endoscopy) was determined at all visits between weeks 0 and 8 inclusive. Clinical response was defined as a decrease from baseline in the total Mayo Score by at least 3 points and at least 30% with an accompanying decrease in rectal bleeding subscore of at least 1 point or a absolute rectal bleeding subscore of 0 or 1 point (12,13). Clinical remission was defined as a total Mayo Score of 2 points or lower, with no individual subscore exceeding 1 point (12,13). Mucosal healing was defined as a decrease from baseline in the endoscopy subscore by at least 1 point and an absolute endoscopy subscore of 0 or 1 point (12,13). Adverse events and concomitant medications were followed through week 12. Blood samples were collected at weeks 0 and 8 for C-reactive protein (CRP) concentrations. Statistical analysis The primary efficacy end point was clinical response at week 8. Secondary efficacy end points included clinical remission at week 8; mucosal healing at week 8; time to partial Mayo Score response (defined as the time point at weeks 2, 4, 6, or 8 at which there was a decrease from baseline in the partial Mayo Score by at least 2 points); change from baseline in the partial Mayo Score at weeks 2, 4, 6, or 8; and the mean change from baseline in the total Mayo Score at week 8. Safety assessments on adverse events were conducted through week 12.

eeks 2, 4, 6, or 8 at which there was a decrease from baseline in the partial Mayo Score by at least 2 points); change from baseline in the partial Mayo Score at weeks 2, 4, 6, or 8; and the mean change from baseline in the total Mayo Score at week 8. Safety assessments on adverse events were conducted through week 12. To control for a type I error of 0.05 or less, the primary end point analyses were conducted in a prespecified, sequential manner. The global null hypothesis was that the proportions of patients with clinical response at week 8 would not be different between the three treatment groups at a 0.05 (two-sided) significance level. If the global null hypothesis was rejected, then the combined A. paniculata dose groups, the A. paniculata 1,200 mg dose group, and the A. paniculata 1,800 mg dose group were compared with placebo at a 0.05 (two-sided) significance level. If the global null hypothesis was not rejected, then the combined dose group and individual dose group comparisons with placebo were considered as not statistically significant. Given the large number of prespecified secondary efficacy variables evaluated at multiple time points during the study, the P values for all secondary efficacy variables should be considered as nominal, as no adjustments were made for multiple comparisons.

with placebo were considered as not statistically significant. Given the large number of prespecified secondary efficacy variables evaluated at multiple time points during the study, the P values for all secondary efficacy variables should be considered as nominal, as no adjustments were made for multiple comparisons. Demographic and baseline characteristics were compared with the use of the χ2 or Fisher's exact test for categorical variables and with analysis of variance on van der Waerden normal scores for continuous variables. The statistical analysis plan stated that clinical response, clinical remission, and mucosal healing would be analyzed using logistic regression, and that last observation carried forward methodology would be used to handle missing data. At the request of regulatory authorities, the data were instead analyzed using the Cochran–Mantel–Haenszel test and missing data were handled using worst case methodology as described below. A two-sided Cochran–Mantel–Haenszel test with concurrent mesalamine use/non-use as strata was used to compare clinical response, clinical remission, and mucosal healing. Missing data were handled using a “worst case” intention-to-treat analysis in which patients with any missing component of the Mayo Score were considered not to be in clinical response, clinical remission, or to have mucosal healing. Given a significant overall treatment difference for clinical response, the specific comparisons of placebo vs. A. paniculata 1,200 mg, placebo vs. A. paniculata 1,800 mg, and placebo vs. the combined A. paniculata dose groups were analyzed. Time to partial Mayo Score response was compared with Cox regression and the results expressed as a hazard ratio (HR), interpreted as an odds ratio with 95% confidence interval (CI). A cumulative incidence Kaplan–Meier plot was also created. The change from baseline in partial Mayo Score at weeks 2, 4, 6, and 8 by treatment group and the mean treatment difference (placebo minus A. paniculata with respect to change from baseline) were plotted by post-baseline visit for each treatment group along with 95% CI at these assessment times. The earliest time at which the lower limit on the CI for the treatment difference above zero was defined to be the time to first significant partial Mayo Score difference. The mean treatment differences of Mayo Score change from baseline between placebo and A.

reatment group along with 95% CI at these assessment times. The earliest time at which the lower limit on the CI for the treatment difference above zero was defined to be the time to first significant partial Mayo Score difference. The mean treatment differences of Mayo Score change from baseline between placebo and A. paniculata treatment groups were tested using ANCOVA, with treatment as a fixed effect, country/region as a random effect, and age, gender, race, baseline value, and concomitant mesalamine use as covariates. Safety comparisons used the Fisher's exact test. All patients receiving at least one dose of study medication were analyzed for safety according to the treatment actually received. To evaluate the consistency of treatment effect on clinical response between placebo, A. paniculata 1,200 mg, and A. paniculata 1,800 mg, 2 prespecified subgroup analyses (concurrent mesalamine and country/region) and 7 post hoc subgroup analyses (gender, race, age, weight, disease duration, elevated CRP, baseline Mayo Score) were performed. P values were calculated based on the Pearson χ2 test and 95% CI's were based on the normal approximation to the binomial. For the primary end point of clinical response at week 8, it was estimated that 202 patients would allow 80% power to detect a difference in response rates of 21% between the combined A. paniculata dose groups and the placebo group using the logistic regression analysis, assuming a 51% rate of response to A. paniculata and a 30% rate of response to placebo.

esponse at week 8, it was estimated that 202 patients would allow 80% power to detect a difference in response rates of 21% between the combined A. paniculata dose groups and the placebo group using the logistic regression analysis, assuming a 51% rate of response to A. paniculata and a 30% rate of response to placebo. RESULTS Characteristics of the patients In total, 224 patients were randomized to treatment: 75 to placebo, 75 to the A. paniculata 1,200 mg daily, and 74 to A. paniculata 1,800 mg daily. One patient was incorrectly randomized and was excluded from the efficacy analyses. The baseline disease characteristics were similar in the three groups ( Table 1). In all, 180 of 224 patients (80.7%) completed the 8-week trial (Figure 1).

75 to the A. paniculata 1,200 mg daily, and 74 to A. paniculata 1,800 mg daily. One patient was incorrectly randomized and was excluded from the efficacy analyses. The baseline disease characteristics were similar in the three groups ( Table 1). In all, 180 of 224 patients (80.7%) completed the 8-week trial (Figure 1). Efficacy Primary end point The overall treatment comparison of A. paniculata with placebo yielded a P value of 0.0465. In all, 52% of patients receiving A. paniculata (78 of 148) were in clinical response at week 8 as compared with 40% receiving placebo (30 of 75) (P=0.092; Table 2). A dose response for A. paniculata was demonstrated. In all, 45% of patients receiving A. paniculata 1,200 mg daily (33 of 74) were in clinical response at week 8 as compared with 40% receiving placebo (30 of 75) (P=0.592). In all, 60% of patients receiving A. paniculata 1,800 mg daily (44 of 74) were in clinical response at week 8 as compared with 40% receiving placebo (30 of 75) (P=0.018). The results of logistic regression analysis using last observation carried forward methodology were similar (Supplementary Table 1 online). The efficacy of A. paniculata was generally consistent among demographic and baseline disease characteristics (see subgroup logistic regression analyses in Supplementary Figure 1A (A. paniculata vs. placebo), Supplementary Figure 1B (A. paniculata 1,200 mg vs. placebo), and Supplementary Figure 1C (A. paniculata 1,800 mg vs. placebo)). The following subgroups showed trends toward a greater A. paniculata response compared with placebo: Mayo Score <6 points (primarily in the 1,200-mg group); endoscopy subscore <2 points; CRP >0.8 mg/dl; weight >85 kg; patients from Europe (primarily in the 1,800-mg group); concomitant mesalamine use (primarily in the 1,800-mg group); disease duration >5 years; white patients; patients aged 18–45 years; and male patients. All subgroup comparisons in the pooled A. paniculata vs. placebo analysis showed distinct divergence with the exception of the baseline CRP subgroup. In the 1,200-mg vs. placebo subgroup analysis only the endoscopy subscore <2 points (P=0.0241) and race (P=0.0780) subgroups were observed to be divergent. However, the 1,800-mg vs. placebo subgroup analysis showed distinct divergence in all subgroups with the exception of the baseline CRP subgroup. Disease duration was not confounded within age group as there was little difference in disease duration across the age group categories.

bgroups were observed to be divergent. However, the 1,800-mg vs. placebo subgroup analysis showed distinct divergence in all subgroups with the exception of the baseline CRP subgroup. Disease duration was not confounded within age group as there was little difference in disease duration across the age group categories. Secondary end points In total, 36% of patients receiving A. paniculata (53 of 148) were in clinical remission at week 8 as compared with 25% receiving placebo (19 of 75) (P=0.1173; Table 2). The rates of clinical remission for the 1,800- and 1,200-mg doses were not significantly greater than placebo, although there was a trend toward significance for the 1,800-mg dose, P=0.1011. The results of logistic regression analysis using last observation carried forward methodology were generally similar to the worst case methodology, but using this analysis the difference between A. paniculata 1,800 mg and placebo was significant (Supplementary Table 1). In all, 44% of patients receiving A. paniculata (65 of 148) achieved mucosal healing at week 8 as compared with 33% receiving placebo (25 of 75) (P=0.1309; Table 2). A dose response for A. paniculata was demonstrated; the rate of mucosal healing for the 1,800-mg dose was significantly greater than placebo whereas the 1,200-mg dose was not. The results of logistic regression analysis using last observation carried forward methodology were similar (Supplementary Table 1).

; Table 2). A dose response for A. paniculata was demonstrated; the rate of mucosal healing for the 1,800-mg dose was significantly greater than placebo whereas the 1,200-mg dose was not. The results of logistic regression analysis using last observation carried forward methodology were similar (Supplementary Table 1). The times to first significant difference in partial Mayo Score response and first significant partial Mayo Score response were both 4 weeks in the combined A. paniculata groups, 4 weeks in the A. paniculata 1,800 mg group, and 4 weeks in the A. paniculata 1,200 mg group. The HR for the time to first significant difference in partial Mayo Score response was 1.57 (P=0.0096) for the combined A. paniculata groups over placebo; the HRs for the 1,800-mg and 1,200-mg groups were 1.72 (P=0.0060) and 1.44 (P=0.0679), respectively. The cumulative incidence Kaplan–Meier plot is shown in Figure 2. The mean improvement (95% CIs) from baseline in the total Mayo score at week 8 between placebo and A. paniculata from the ANCOVA model was 0.51 (−0.33, 1.35) points for the combined A. paniculata groups (P=0.2329), 0.58 (−0.24, 1.41) points, and 0.43 (−0.42, 1.29) points for the 1,800-mg and 1,200-mg groups (P=0.1652 and P=0.3208, respectively). Safety Through week 8, the incidence of adverse events was generally similar among groups ( Table 3). A rash occurred in 8% of patients receiving A. paniculata and 1% of patients receiving placebo. The rashes were mostly mild (with the rest moderate), reversible, and did not cause treatment discontinuation.

The times to first significant difference in partial Mayo Score response and first significant partial Mayo Score response were both 4 weeks in the combined A. paniculata groups, 4 weeks in the A. paniculata 1,800 mg group, and 4 weeks in the A. paniculata 1,200 mg group. The HR for the time to first significant difference in partial Mayo Score response was 1.57 (P=0.0096) for the combined A. paniculata groups over placebo; the HRs for the 1,800-mg and 1,200-mg groups were 1.72 (P=0.0060) and 1.44 (P=0.0679), respectively. The cumulative incidence Kaplan–Meier plot is shown in Figure 2. The mean improvement (95% CIs) from baseline in the total Mayo score at week 8 between placebo and A. paniculata from the ANCOVA model was 0.51 (−0.33, 1.35) points for the combined A. paniculata groups (P=0.2329), 0.58 (−0.24, 1.41) points, and 0.43 (−0.42, 1.29) points for the 1,800-mg and 1,200-mg groups (P=0.1652 and P=0.3208, respectively). Safety Through week 8, the incidence of adverse events was generally similar among groups ( Table 3). A rash occurred in 8% of patients receiving A. paniculata and 1% of patients receiving placebo. The rashes were mostly mild (with the rest moderate), reversible, and did not cause treatment discontinuation. DISCUSSION Treatment with A. paniculata extract (HMPL-004) was more effective than placebo for induction of clinical response and mucosal healing at week 8 among patients with mildly to moderately active ulcerative colitis, most of whom were failing first-line therapy with mesalamine. The best efficacy was observed with the 1,800-mg dose. However, there were no significant differences between the A. paniculata and placebo groups in the worst case scenario analyses of clinical remission rates or mean improvement in total Mayo score at week 8. The end point of clinical remission was significant with the last observation carried forward analysis. Subgroup analysis indicated trend toward greater clinical response rates in subgroups of patients treated with A. paniculata who had baseline Mayo score <6 points and those with an endoscopy subscores <2 points; this was most apparent in those patients treated with the 1,200-mg dose. Age, gender, and geographic region were also assessed as covariates of clinical response. There was significant subgroup-by-treatment interaction observed in the 1,200-mg group for the baseline endoscopy subscore and race. There was also a trend toward a better clinical response to patients treated with A. paniculata in subgroups of patients with younger age, male sex, and European geographic location. A statistically significant improvement in clinical response was not observed in patients with North American geographic location. These subgroup analysis results should be interpreted with caution due to limited power based on small number of study subjects. The HRs for the time to first significant difference in partial Mayo Score response were greater in the A. paniculata-treated patients, but in the Kaplan–Meier analysis, the response rates equalized at week 8.

lysis results should be interpreted with caution due to limited power based on small number of study subjects. The HRs for the time to first significant difference in partial Mayo Score response were greater in the A. paniculata-treated patients, but in the Kaplan–Meier analysis, the response rates equalized at week 8. A previous pilot study demonstrated that treatment with A. paniculata 1,200 mg daily for 8 weeks resulted in a similar reduction from baseline in disease activity to that observed with oral mesalamine in patients with mildly to moderately active ulcerative colitis (10). We also found preliminary evidence of efficacy for A. paniculata 1,200 mg daily for 8 weeks as compared with placebo in patients with mildly to moderately active Crohn's disease (14). In the current study, we demonstrated dose response with A. paniculata 1,200 mg and 1,800 mg daily and no dose-dependent toxicity. Additional clinical trials to evaluate the safety and efficacy of even higher doses of A. paniculata extract (HMPL-004) should be undertaken in both ulcerative colitis and Crohn's disease.

In the current study, we demonstrated dose response with A. paniculata 1,200 mg and 1,800 mg daily and no dose-dependent toxicity. Additional clinical trials to evaluate the safety and efficacy of even higher doses of A. paniculata extract (HMPL-004) should be undertaken in both ulcerative colitis and Crohn's disease. Subgroup analysis showed a larger therapeutic effect for A. paniculata extract (HMPL-004) relative to placebo in patients who were currently failing oral mesalamine, and a smaller and non-significant effect relative to placebo among patients not receiving mesalamine. This latter result appears largely due to a greater rate of placebo response among patients not receiving mesalamine. This subgroup of patients was relatively small, and the study lacked sufficient statistical power to draw any meaningful conclusions regarding the efficacy of A. paniculata in this patient population. Additional adequately powered studies to assess the efficacy of A. paniculata extract (HMPL-004) as monotherapy in patients with ulcerative colitis are warranted. It should also be noted that the efficacy of A. paniculata extract (HMPL-004) for induction therapy in patients with ulcerative colitis who are failing various combinations of corticosteroids, azathioprine, and anti-tumor necrosis factor therapy with infliximab is unknown, as its efficacy as a maintenance agent. Additional studies in these patient populations are also needed.

extract (HMPL-004) for induction therapy in patients with ulcerative colitis who are failing various combinations of corticosteroids, azathioprine, and anti-tumor necrosis factor therapy with infliximab is unknown, as its efficacy as a maintenance agent. Additional studies in these patient populations are also needed. The 40% placebo response rate in our study was slightly higher than expected. This study was patterned after two randomized double-blind placebo-controlled studies of infliximab in ulcerative colitis (12). In the first of these studies, the week 8 response rate was about 40% and in the second study the analogous response rate was 30%. The placebo response in our study was in this same general range. In the infliximab trials, the entry criteria specified baseline Mayo Scores of 6–12 points. In our study, the baseline Mayo Scores were 4–10 points, which may also have influenced the placebo response rate. Experience with the use of real-time central reading of endoscopy in clinical trials to reduce placebo rates is evolving, and in future trials central reading could potentially reduce the relatively high placebo rates that we observed in this trial.

Carol Stanciu, Institutul de Gastroenterologie si Hepatologie, Lasi, Romania; Ira Stein, Sarah Cannon Research Institute, Nashville, TN, USA; Jawahar Taunk, Advanced Gastroenterology Associates, Palm Harbor, FL, USA; James Underwood, Gastrointestinal Associates, Jackson, MS, USA; Lawrence Wruble, Memphis Gastroenterology Group, Germantown, TN, USA; June Yong, Hutchinson Clinic, Hutchinson, KS, USA; Oren Zaidel, South Bay GI, Torrance, CA, USA; Iurii Mykhailovich Zakharash, Central Clinical Hospital, National Medical University, Kyiv, Ukraine; Salam Zakko, Connecticut Gastroenterology Institute, Bristol Hospital, Bristol, CT, USA. SUPPLEMENTARY MATERIAL is linked to the online version of the paper at http://www.nature.com/ajg Guarantors of the article: William J. Sandborn, MD and Stephan R. Targan, MD. Specific author contributions: William J. Sandborn and Stephan R. Targan served on the executive committee of study investigators, and worked with Vera Byers of Immunology Inc., Dean Rutty of Everest Clinical Research Services, Inc., and Tom Tang of Hutchison Medipharma Ltd. to design the study. All authors have been involved in analysis and interpretation of the data. William J. Sandborn drafted the manuscript. All authors have critically reviewed the manuscript draft for intellectual content, and approved the final version of the manuscript for publication Financial support: This work was supported by a research grant from Hutchison Medipharma Ltd., Shanghai, China.

4–10 points, which may also have influenced the placebo response rate. Experience with the use of real-time central reading of endoscopy in clinical trials to reduce placebo rates is evolving, and in future trials central reading could potentially reduce the relatively high placebo rates that we observed in this trial. In this short-term study, the overall incidence of adverse events observed in patients treated with A. paniculata extract (HMPL-004) was similar to patients receiving placebo, with the exception of rash which was higher in the A. paniculata extract (HMPL-004) groups. Longer term studies are needed to further assess the safety and tolerability of A. paniculata extract (HMPL-004) in patients with ulcerative colitis. If such studies continue to demonstrate a favorable safety profile, then it may be possible to combine A. paniculata extract (HMPL-004) with other drugs of known efficacy such as corticosteroids, azathioprine, and anti-tumor necrosis factor agents, with the goal of achieving synergistic efficacy without incurring synergistic toxicity. The mechanism of action of A. paniculata extract should be further explored. In vitro inhibitory activity against TNF-α, IL-1β, and NF-κB has been reported (8,9). Additional preclinical studies in animal models of colitis should be undertaken to further explore the efficacy and biologic effects of A. paniculata extract and its major active components the diterpene lactones such as andrographolide across a range of doses

ty against TNF-α, IL-1β, and NF-κB has been reported (8,9). Additional preclinical studies in animal models of colitis should be undertaken to further explore the efficacy and biologic effects of A. paniculata extract and its major active components the diterpene lactones such as andrographolide across a range of doses In conclusion, patients with mildly to moderately active ulcerative colitis treated with A. paniculata extract (HMPL-004) were more likely to achieve clinical response than those receiving placebo. STUDY HIGHLIGHTS APPENDIX The study Principal Investigators are listed below in alphabetical order:Orest Ostapovich Abrahamovych, Danylo Halytsky Lviv National Medical University, Lviv, Ukraine; Mohammed Azzouz, Gastroenterology and Liver Diseases, Warwick, RI, USA; Daniel Baumgart, Charité - Campus Virchow, Berlin, Germany; William J. Cohn, Gastrointestinal Research Associates, Setauket, NY, USA; Hanns-Gerd Dammann, Klinische Forschung Hannover Mitte, Hannover, Germany; Oleksiy Borisovich Datsenko, Municipal Clinical Hospital #2, Proctology department, Kharkov, Ukraine; Frank Drouven, Bethlehem Krankenhaus, Stolberg, Germany; J. Daniel Davis, Hope Research Institute, Phoenix, AZ, USA; Steven Desautels, Physicians Research Option, Sandy, UT, USA; Ion Dina, SC Endocenter Medicina Integrativa, Bucharest, Romania; Dan Dumitrascu, Universitatea Iuliu Hatieganu, Cluj Napoca, Romainia; Galyna Dmytrivna Fadieienko, Institute of Therapy by name L.T.Maloy of Ukrainian AMS, Kharkov, Ukraine; Blanche Fung-Liu, Long Island Clinical Research Associates, Great Neck, NY, USA;

Steven Desautels, Physicians Research Option, Sandy, UT, USA; Ion Dina, SC Endocenter Medicina Integrativa, Bucharest, Romania; Dan Dumitrascu, Universitatea Iuliu Hatieganu, Cluj Napoca, Romainia; Galyna Dmytrivna Fadieienko, Institute of Therapy by name L.T.Maloy of Ukrainian AMS, Kharkov, Ukraine; Blanche Fung-Liu, Long Island Clinical Research Associates, Great Neck, NY, USA; Syam P. Gaddam, Digestive and Liver Disease Specialists, Garden Grove, CA, USA; B. Matthew Garner, St Bernards Research Center, Jonesboro, AR, USA; Glenn L. Gordon, Center for Digestive and Liver Diseases, Inc., Mexico, MO, USA; Robert Hardi, Metropolitan Gastroenterology Group, Chevy Chase Clinical Research, Chevy Chase, MD, USA Robert Holmes, Piedmont Medical Research, Winston-Salem, NC, USA; Steven Hong, Spokane Digestive Disease Center, Spokane, WA, USA; Oleksandra Mykhaylivna Iemets, Medychnyy Centr Mriya, Kyiv, Ukraine; Robert Kaplan, LeBauer Research Associates, Greensboro, NC, USA; James H. Kopp, Hartwell Research Group, Anderson, SC, USA; Bernard Leman, Iowa Digestive Disease Center, Clive, IA, USA; Olena Mykhaylivna Levchenko, Odes′kyy oblasnyy centr gastroenterologiyi, Odessa, Ukraine; John Lowe, Advanced Research Institute, Ogden, UT, USA; Mitchell Mah'Moud, Boice-Willis Clinic, Clinical Research Department, Rocky Mount, NC, USA; Nasrullah Manji, Pioneer Research Solutions, Houston, TX, USA; Stefano Marcuard, Carolina Research, Greenville, NC, USA; Hipolito G. Mariano, Research Center of Fresno, Fresno, CA, USA; Ravi K. Moparty, Spring Clinical Research, Houston, TX, USA;

Mitchell Mah'Moud, Boice-Willis Clinic, Clinical Research Department, Rocky Mount, NC, USA; Nasrullah Manji, Pioneer Research Solutions, Houston, TX, USA; Stefano Marcuard, Carolina Research, Greenville, NC, USA; Hipolito G. Mariano, Research Center of Fresno, Fresno, CA, USA; Ravi K. Moparty, Spring Clinical Research, Houston, TX, USA; Michael J. Ostro, Toronto Digestive Disease Associates, Toronto, Ontario, Canada; Daniel Pambianco, Charlottesville Medical Research, Charlottesville, VA, USA; Edward Pensa, Pharma Resource, East Providence, RI, USA; Corina Popa, SC Pelican Impex SRL, Oradea Jud Bihor, Romania; Natarajan Ravendhran, Digestive Disease Associates, Baltimore, MD, USA; Mark Reichelderfer, University of Wisconsin Hospital & Clinics, Madison, WI, USA; Donato R. Ricci, Accord Clinical Research, Port Orange, FL, USA; Dennis Riff, Advanced Clinical Research Institute, Anaheim, CA, USA; Barry Sanders, North Texas Gastroenterology Consultants, Lewisville, TX, USA; Omer Secil, Sp. Clinic de Urgenta Sf. Ioan, Bucharest, Romania; Ursula Seidler, MHH, Hannover, Germany; Ira Shafran, Shafran Gastroenterology Center, Winter Park, FL, USA; Andreas Stallmach, Universitätsklinikum Jena, Jena Lobeda-Ost, Germany; Carol Stanciu, Institutul de Gastroenterologie si Hepatologie, Lasi, Romania; Ira Stein, Sarah Cannon Research Institute, Nashville, TN, USA; Jawahar Taunk, Advanced Gastroenterology Associates, Palm Harbor, FL, USA; James Underwood, Gastrointestinal Associates, Jackson, MS, USA; Lawrence Wruble, Memphis Gastroenterology Group, Germantown, TN, USA; June Yong, Hutchinson Clinic, Hutchinson, KS, USA;

Specific author contributions: William J. Sandborn and Stephan R. Targan served on the executive committee of study investigators, and worked with Vera Byers of Immunology Inc., Dean Rutty of Everest Clinical Research Services, Inc., and Tom Tang of Hutchison Medipharma Ltd. to design the study. All authors have been involved in analysis and interpretation of the data. William J. Sandborn drafted the manuscript. All authors have critically reviewed the manuscript draft for intellectual content, and approved the final version of the manuscript for publication Financial support: This work was supported by a research grant from Hutchison Medipharma Ltd., Shanghai, China. Potential competing interests: Dr Sandborn reports having received consulting fees from Abbott Laboratories, ActoGeniX NV, AGI Therapeutics, Inc., Alba Therapeutics Corporation, Albireo, Alfa Wasserman, Amgen, AM-Pharma BV, Anaphore, Astellas Pharma, Athersys, Inc., Atlantic Healthcare Limited, Axcan Pharma (now Aptalis), BioBalance Corporation, Boehringer-Ingelheim Inc, Bristol Meyers Squibb, Celegene, Celek Pharmaceuticals, Cellerix SL, Cerimon Pharmaceuticals, ChemoCentryx, CoMentis, Cosmo Technologies, Coronado Biosciences, Cytokine Pharmasciences, Eagle Pharmaceuticals, Eisai Medical Research Inc., Elan Pharmaceuticals, EnGene, Inc., Eli Lilly, Enteromedics, Exagen Diagnostics, Inc., Ferring Pharmaceuticals, Flexion Therapeutics, Inc., Funxional Therapeutics Limited, Genzyme Corporation, Genentech (now Roche), Gilead Sciences, Given Imaging, Glaxo Smith Kline, Human Genome Sciences, Ironwood Pharmaceuticals (previously Microbia Inc.), Janssen (previously Centocor), KaloBios Pharmaceuticals, Inc., Lexicon Pharmaceuticals, Lycera Corporation, Meda Pharmaceuticals (previously Alaven Pharmaceuticals), Merck Research Laboratories, MerckSerono, Millennium Pharmaceuticals (subsequently merged with Takeda), Nisshin Kyorin Pharmaceuticals Co., Ltd., Novo Nordisk A/S, NPS Pharmaceuticals, Optimer Pharmaceuticals, Orexigen Therapeutics, Inc., PDL Biopharma, Pfizer, Procter and Gamble, Prometheus Laboratories, ProtAb Limited, Purgenesis Technologies, Inc., Receptos, Relypsa, Inc., Salient Pharmaceuticals, Salix Pharmaceuticals, Inc., Santarus, Schering Plough Corporation (acquired by Merck), Shire Pharmaceuticals, Sigmoid Pharma Limited, Sirtris Pharmaceuticals, Inc. (a GSK company), S.L.A. Pharma (UK) Limited, Targacept, Teva Pharmaceuticals, Therakos, Tillotts Pharma AG (acquired by Zeria Pharmaceutical Co., Ltd.), TxCell SA, UCB Pharma, Viamet Pharmaceuticals, Vascular Biogenics Limited (VBL), Warner Chilcott UK Limited, Wyeth (now Pfizer). He has received lecture fees from Abbott Laboratories, Bristol Meyers Squibb, and Janssen (previously Centocor).

ls, Therakos, Tillotts Pharma AG (acquired by Zeria Pharmaceutical Co., Ltd.), TxCell SA, UCB Pharma, Viamet Pharmaceuticals, Vascular Biogenics Limited (VBL), Warner Chilcott UK Limited, Wyeth (now Pfizer). He has received lecture fees from Abbott Laboratories, Bristol Meyers Squibb, and Janssen (previously Centocor). He has received research support from Abbott Laboratories, Bristol Meyers Squibb, Genentech, Glaxo Smith Kline, Janssen (previously Centocor), Millennium Pharmaceuticals (now Takeda), Novartis, Pfizer, Procter and Gamble Pharmaceuticals, Shire Pharmaceuticals, and UCB Pharma. Dr Targan reports having received consulting fees from Amgen, Biogen, Esai, GSK, Hutchison Medipharma, Lycera, Merck, Prometheus, Symbiotix, Takeda Pharmaceutical and Zyngenia. In addition, he reports having received research and meeting/travel support from Hutchison Medipharma. He also reports having received a grant for researching the mechanism of action for the drug published in this paper from Hutchison Medipharma. Dr Byers reports having received the following from Hutchison Medipharma: support for travel to meetings for the study and other purposes, receiving fees for review activity related to data monitoring and analysis, and for writing and reviewing the manuscript. Mr. Rutty reports having received the following from Hutchison Medipharma: support for travel to meetings for the study and other purposes, receiving fees for review activity related to data monitoring and analysis, and for writing and reviewing the manuscript. In addition, he reports Board membership for Steba Biotech SA and reports consulting for the following companies: Schering Corporation, Roche, Methylgene, Steba Biotech SA, Aderans Research Institute Inc., Stem Cell Theraputics, Genentech, Pearly Therapeutics, Sundise Chinese Medicine Technology Development Corp., Endocyte, Inc, and Generon Corporation Ltd. Mr Rutty is an employee of Everest Clinical Research Services Inc. Dr Tang reports being an employee of Hutchison Medipharma at the time of the clinical trial described herein and during the drafting of this manuscript. He is currently an employee of Generon (Shanghai). Dr Mu reports being an employee of Hutchison Medipharma. Dr Zhang reports being an employee of Hutchison Medipharma.

Dr Tang reports being an employee of Hutchison Medipharma at the time of the clinical trial described herein and during the drafting of this manuscript. He is currently an employee of Generon (Shanghai). Dr Mu reports being an employee of Hutchison Medipharma. Dr Zhang reports being an employee of Hutchison Medipharma. Supplementary Material Supplementary Figure 1a Click here for additional data file. Supplementary Figure 1b Click here for additional data file. Supplementary Figure 1c Click here for additional data file. Supplementary Figure Legends Click here for additional data file. Figure 1 Enrollment and treatment through week 8. Figure 2 Kaplan–Meier estimate of the proportion of patients free of clinical response as defined by the partial Mayo Score through 8 weeks for the combined Andrographis paniculata group, the A. paniculata 1,200 mg group, the A. paniculata 1,800 mg group, and the placebo group. ITT, intention-to-treat.

Figure 1 Enrollment and treatment through week 8. Figure 2 Kaplan–Meier estimate of the proportion of patients free of clinical response as defined by the partial Mayo Score through 8 weeks for the combined Andrographis paniculata group, the A. paniculata 1,200 mg group, the A. paniculata 1,800 mg group, and the placebo group. ITT, intention-to-treat. Table 1 Demographic and baseline disease characteristics Placebo (N=75) Andrographis paniculata 1,200 mg (N=74) Andrographis paniculata 1,800 mg (N=74) Total (N=223) P valuea Male, no. (%) 41 (54.7) 40 (54.1) 41 (55.4) 122 (54.7) 0.9864 Caucasian, no. (%) 67 (89.3) 64 (86.5) 64 (86.5) 195 (87.4) 0.7448 Age (years) Mean±s.d. 44.7±15.2 44.3±14.5 45.6±13.6 44.9±14.4 0.8454 Median 45.0 43.5 46.5 45.0 Body weight (kg) Mean±s.d. 76.8±14.8 77.0±19.9 78.6±20.1 77.5±18.4 0.8030 Median 75.8 74.2 74.4 74.8 Disease duration (months)         0.5615 Mean±s.d. 63.4±70.3 68.6±83.4 72.6±74.7 68.2±76.0 Median 42.9 32.4 43.4 40.7 C-reactive protein (mg/dl) Mean±s.d. 0.58±0.82 0.64±0.70 1.14±2.70 0.79±1.69 0.5282 Median 0.30 0.31 0.30 0.30 Elevated CRP,b n (%) 14 (18.7) 21 (28.4) 20 (27.0) 55 (24.7) 0.3290 Mayo Score Mean±s.d. 6.1±1.8 6.3±1.8 6.1±1.7 6.2±1.8 0.6748 Median 6.0 6.0 6.0 6.0 Concomitant mesalamine, n (%) 52 (69.3) 50 (67.6) 51 (68.9) 153 (68.6) 0.9710 CRP, C-reactive protein. a P values for all categorical variables are based on a χ2 test. P values for continuous variables are based on analysis of variance if normal distribution assumption is met; otherwise P values are based on the Kruskal–Wallis test.

Table 1 Demographic and baseline disease characteristics Placebo (N=75) Andrographis paniculata 1,200 mg (N=74) Andrographis paniculata 1,800 mg (N=74) Total (N=223) P valuea Male, no. (%) 41 (54.7) 40 (54.1) 41 (55.4) 122 (54.7) 0.9864 Caucasian, no. (%) 67 (89.3) 64 (86.5) 64 (86.5) 195 (87.4) 0.7448 Age (years) Mean±s.d. 44.7±15.2 44.3±14.5 45.6±13.6 44.9±14.4 0.8454 Median 45.0 43.5 46.5 45.0 Body weight (kg) Mean±s.d. 76.8±14.8 77.0±19.9 78.6±20.1 77.5±18.4 0.8030 Median 75.8 74.2 74.4 74.8 Disease duration (months)         0.5615 Mean±s.d. 63.4±70.3 68.6±83.4 72.6±74.7 68.2±76.0 Median 42.9 32.4 43.4 40.7 C-reactive protein (mg/dl) Mean±s.d. 0.58±0.82 0.64±0.70 1.14±2.70 0.79±1.69 0.5282 Median 0.30 0.31 0.30 0.30 Elevated CRP,b n (%) 14 (18.7) 21 (28.4) 20 (27.0) 55 (24.7) 0.3290 Mayo Score Mean±s.d. 6.1±1.8 6.3±1.8 6.1±1.7 6.2±1.8 0.6748 Median 6.0 6.0 6.0 6.0 Concomitant mesalamine, n (%) 52 (69.3) 50 (67.6) 51 (68.9) 153 (68.6) 0.9710 CRP, C-reactive protein. a P values for all categorical variables are based on a χ2 test. P values for continuous variables are based on analysis of variance if normal distribution assumption is met; otherwise P values are based on the Kruskal–Wallis test. b The normal range for CRP concentration is <0.8 mg/dl.

Table 1 Demographic and baseline disease characteristics Placebo (N=75) Andrographis paniculata 1,200 mg (N=74) Andrographis paniculata 1,800 mg (N=74) Total (N=223) P valuea Male, no. (%) 41 (54.7) 40 (54.1) 41 (55.4) 122 (54.7) 0.9864 Caucasian, no. (%) 67 (89.3) 64 (86.5) 64 (86.5) 195 (87.4) 0.7448 Age (years) Mean±s.d. 44.7±15.2 44.3±14.5 45.6±13.6 44.9±14.4 0.8454 Median 45.0 43.5 46.5 45.0 Body weight (kg) Mean±s.d. 76.8±14.8 77.0±19.9 78.6±20.1 77.5±18.4 0.8030 Median 75.8 74.2 74.4 74.8 Disease duration (months)         0.5615 Mean±s.d. 63.4±70.3 68.6±83.4 72.6±74.7 68.2±76.0 Median 42.9 32.4 43.4 40.7 C-reactive protein (mg/dl) Mean±s.d. 0.58±0.82 0.64±0.70 1.14±2.70 0.79±1.69 0.5282 Median 0.30 0.31 0.30 0.30 Elevated CRP,b n (%) 14 (18.7) 21 (28.4) 20 (27.0) 55 (24.7) 0.3290 Mayo Score Mean±s.d. 6.1±1.8 6.3±1.8 6.1±1.7 6.2±1.8 0.6748 Median 6.0 6.0 6.0 6.0 Concomitant mesalamine, n (%) 52 (69.3) 50 (67.6) 51 (68.9) 153 (68.6) 0.9710 CRP, C-reactive protein. a P values for all categorical variables are based on a χ2 test. P values for continuous variables are based on analysis of variance if normal distribution assumption is met; otherwise P values are based on the Kruskal–Wallis test. b The normal range for CRP concentration is <0.8 mg/dl. Table 2 Efficacy resultsa End point Dose Andrographis paniculata Placebo P valueb P valuec Clinical response at week 8 Combined 1,800+1,200 mg 78/148 (52.0%) 30/75 (40.0%) 0.0922 0.0465 1,800 mg 44/74 (59.5%)   0.0183 1,200 mg 33/74 (44.6%)   0.5924 Clinical remission at week 8 Combined 1,800+1,200 mg 53/148 (35.8%) 19/75 (25.3%) 0.1173 0.2516 1,800 mg 28/74 (37.8%)   0.1011 1,200 mg 25/74 (33.8%)   0.2718 Mucosal healing at week 8 Combined 1,800+1,200 mg 65/148 (43.9%) 25/75 (33.3%) 0.1309 0.1025 1,800 mg 37/74 (50.0%)   0.0404 1,200 mg 28/74 (37.8%)   0.5821 a Efficacy determined using worst case handling of missing data in which missing data were set to failure.

g 28/74 (37.8%)   0.1011 1,200 mg 25/74 (33.8%)   0.2718 Mucosal healing at week 8 Combined 1,800+1,200 mg 65/148 (43.9%) 25/75 (33.3%) 0.1309 0.1025 1,800 mg 37/74 (50.0%)   0.0404 1,200 mg 28/74 (37.8%)   0.5821 a Efficacy determined using worst case handling of missing data in which missing data were set to failure. b P values for the pair wise comparisons are based on the Cochran–Mantel–Haenszel with concomitant mesalamine use as strata. c P value is for the overall treatment comparison based on the Cochran–Mantel–Haenszel.

g 28/74 (37.8%)   0.1011 1,200 mg 25/74 (33.8%)   0.2718 Mucosal healing at week 8 Combined 1,800+1,200 mg 65/148 (43.9%) 25/75 (33.3%) 0.1309 0.1025 1,800 mg 37/74 (50.0%)   0.0404 1,200 mg 28/74 (37.8%)   0.5821 a Efficacy determined using worst case handling of missing data in which missing data were set to failure. b P values for the pair wise comparisons are based on the Cochran–Mantel–Haenszel with concomitant mesalamine use as strata. c P value is for the overall treatment comparison based on the Cochran–Mantel–Haenszel. Table 3 Summary of safety findings through week 12a P valueb   P valueb Placebo (N=75) Andrographis paniculata 1,200 mg (N=75) Andrographis paniculata 1,200 mg vs. placebo Andrographis paniculata 1,800 mg (N=74) Andrographis paniculata 1,800 mg vs. placebo Mean duration of treatment, days 52 49   48 Any adverse event, no. (%) 45 (60) 45 (60) 1.0000 39 (53) 0.3692 Adverse events occurring in 4% of any treatment group, n (%) Abdominal pain 6 (8) 4 (5)   4 (5) Diarrhea 2 (3) 3 (4)   4 (5) Dyspepsia 1 (1) 3 (4)   1 (1) Flatulence 1 (1) 1 (1)   4 (5) Nausea 2 (3) 4 (5)   3 (4) Ageusia 0 (0) 3 (4)   2 (3) Dysgeusia 0 (0) 0 (0)   3 (4) Headache 5 (7) 8 (11)   4 (5) Influenza 4 (5) 2 (3)   2 (3) Nasopharyngitis 3 (4) 2 (3)   2 (3) Alanine aminotransferase increased 0 (0) 3 (4)   0 (0) Blood alkaline phosphatase increased 1 (1) 3 (4)   0 (0) Blood glucose increased 0 (0) 0 (0)   3 (4) Gamma-glutamyl transferase increased 2 (3) 3 (4)   1 (1) Rash 1 (1) 3 (4)   3 (4) Fatigue 3 (4) 2 (3)   0 (0) Anemia 3 (4) 0 (0)   1 (1) Basophilia 3 (4) 0 (0)   0 (0) Back pain 3 (4) 0 (0)   3 (4) Adverse events leading to study drug discontinuation, n (%) 3 (4) 7 (9) 0.3268 6 (8) 0.3268 Serious adverse events, n (%) 2 (3) 2 (3) 1.000 2 (3) 1.000 a A patient's treatment group is based on treatment received.

(4) 2 (3)   0 (0) Anemia 3 (4) 0 (0)   1 (1) Basophilia 3 (4) 0 (0)   0 (0) Back pain 3 (4) 0 (0)   3 (4) Adverse events leading to study drug discontinuation, n (%) 3 (4) 7 (9) 0.3268 6 (8) 0.3268 Serious adverse events, n (%) 2 (3) 2 (3) 1.000 2 (3) 1.000 a A patient's treatment group is based on treatment received. b P values are based on Fisher's exact test if the frequency of at least 1 cell is <5; otherwise P values are based on a χ2 test.