Browse the corpus

1 Introduction There is well-replicated evidence that in childhood those with permanent childhood hearing loss (PCHL) are at an increased risk of poor psychosocial functioning [1], [2]. These difficulties are apparent from the preschool age range [3]. Adolescents with PCHL face a number of challenges not experienced by their hearing peers and can find some aspects, such a friendship and peer relations, particularly daunting [4]. Children with PCHL are likely to be at risk of developing emotional and behaviour difficulties (EBD) as a result of a number of factors. Their social-emotional development may be adversely influenced by difficulties in communication and many have additional cognitive and physical impairments that are risk factors for EBD [5], [6], [7]. For this reason it is important to determine if, and to what extent, they show an elevated rate of mental health problems compared to children in the general population so that intervention can be targeted at this potentially vulnerable group. A narrative review linked hearing loss (HL) with mental health problems in children and adolescents, including depression, aggression, oppositional defiant disorder and conduct disorder, and, less consistently, anxiety, somatization, and delinquency [1]. That review was based on 35 papers and found that there were substantial differences between studies and marked heterogeneity in the HL population. The review commented on the absence of longitudinal studies on this issue.

isorder and conduct disorder, and, less consistently, anxiety, somatization, and delinquency [1]. That review was based on 35 papers and found that there were substantial differences between studies and marked heterogeneity in the HL population. The review commented on the absence of longitudinal studies on this issue. A quantitative review of studies on the mental health of children with HL presented results in two parts [2]. The first part identified 33 studies in which emotional and behaviour difficulties in children with HL could be compared to a hearing comparison group (HCG) using a variety of measures of EBD. The average effect size (standardised mean difference, SMD) for these studies was 0.36. The second part reported a meta-analysis on 12 studies of children with HL using the Strengths and Difficulties Questionnaire (SDQ) [8]. The estimated effect sizes were +0.23 (95%CI 0.07 to 0.40), +0.34 (95%CI 0.19 to 0.49) and −0.01 (95%CI −0.32 to 0.13) for parent, teacher and self-report ratings of Total Difficulties respectively. The most consistent differences between children with HL and a HCG were in the area of Peer Problems.

ionnaire (SDQ) [8]. The estimated effect sizes were +0.23 (95%CI 0.07 to 0.40), +0.34 (95%CI 0.19 to 0.49) and −0.01 (95%CI −0.32 to 0.13) for parent, teacher and self-report ratings of Total Difficulties respectively. The most consistent differences between children with HL and a HCG were in the area of Peer Problems. In that meta-analysis based on cross-sectional studies of children with HL at different ages, there was no evidence of age-related changes in the risk of EBD in children with HL. There is a paucity of evidence regarding age-related changes from longitudinally studied samples. One such study of children with “hearing problems”, identified on the basis of parental report either at age less than 1 year or at age 4–5 years [9], assessed children on the SDQ for EBD at ages up to 10–11 years. There was no clear pattern of an increasing or decreasing level of EBD over time. There is a need to extend such longitudinal studies into the adolescent years and on the basis of HL established using objective audiological evaluation.

t age 4–5 years [9], assessed children on the SDQ for EBD at ages up to 10–11 years. There was no clear pattern of an increasing or decreasing level of EBD over time. There is a need to extend such longitudinal studies into the adolescent years and on the basis of HL established using objective audiological evaluation. There are aspects of development and management that may be related to EBD score in those with PCHL. When the PCHL sample in the present study was assessed at age 6–10 years, there was a strong relationship between EBD scores and poor expressive and receptive language ability [10]. However we found, at that age, that while early confirmation of PCHL had a beneficial effect on receptive language development, it had no significant impact in reducing behaviour problems in children with PCHL [11]. We argued that this beneficial effect of early confirmation on language development at age 6–10 years was not sufficiently great to bring the language ability of this early confirmed group to the level of the hearing comparison group (HCG) and therefore the increased risk of behaviour problems remained. Early confirmation also had a beneficial effect at ages 13–20 years on reading comprehension [12], though significant benefits of early confirmation on receptive language ability were only detectable for those without cochlear implants [13]. There was no significant benefit for language development [13] or for reading [12] from exposure to Universal Neonatal Hearing Screening (UNHS). It is possible that by this later age any enhancement of language development by early confirmation may be sufficient to have an impact on behaviour.

t cochlear implants [13]. There was no significant benefit for language development [13] or for reading [12] from exposure to Universal Neonatal Hearing Screening (UNHS). It is possible that by this later age any enhancement of language development by early confirmation may be sufficient to have an impact on behaviour. Here we present the findings from a longitudinal study whose participants were assessed in their teenage years. We previously reported a number of findings in infancy and the first decade of life in this sample [14], [15], [16]. The first aim of the present study was to examine whether the pattern of elevated SDQ scores, indicating the presence of behaviour problems, demonstrated in the group with PCHL in childhood was still present in the teenage years. The second aim was to determine which factors, including early confirmation of PCHL, exposure to UNHS, cochlear implantation (CI), poorer receptive language ability, and/or the presence of other disabilities, were associated with high EBD scores within the PCHL group.

n childhood was still present in the teenage years. The second aim was to determine which factors, including early confirmation of PCHL, exposure to UNHS, cochlear implantation (CI), poorer receptive language ability, and/or the presence of other disabilities, were associated with high EBD scores within the PCHL group. 2 Methods 2.1 Participants The study was a population-based cohort study of children with bilateral PCHL that also included a HCG that was half the size of the group of participants with PCHL. The 183 adolescents aged 13–20 years (120 with PCHL, 63 in the HCG) who were eligible for this prospective follow-up study were drawn from a birth cohort of children born in eight districts of southern England and had participated in a previous phase of the study aged 6–10 years [14], [15], [16]. The birth cohort comprised two sub-cohorts: First, the 1993–1996 Wessex birth cohort of 54,000 babies enrolled in the Wessex controlled trial of UNHS [17], Second, the 1992–1997 Greater London birth cohort of 100,000 babies, born in four districts in Greater London, of which two were the only two districts in the UK (Waltham Forest; Hillingdon) that provided UNHS for PCHL in the early 1990s and the other two were districts geographically adjacent to them (Redbridge; Brent & Harrow, respectively). The adolescents with PCHL had all been diagnosed with PCHL ≥40 dB in the better ear that was not known to be acquired. The HCG was comparable to those with PCHL with respect to place and date of birth.

arly 1990s and the other two were districts geographically adjacent to them (Redbridge; Brent & Harrow, respectively). The adolescents with PCHL had all been diagnosed with PCHL ≥40 dB in the better ear that was not known to be acquired. The HCG was comparable to those with PCHL with respect to place and date of birth. At the earlier time point in childhood (Time 1), 120 participants with PCHL and 63 in the HCG took part at a mean (SD) age of 7.96 (1.23) years. At the second time point (Time 2) 76 teenagers with PCHL and 38 in the HCG participated in the current study at a mean age of 16.84 (1.42) years. The annual attrition rate among children with PCHL eligible for the present study was 3% over 17 years since their exposure (or not) to UNHS and 4% over the 9 years since their assessment at primary school. Attrition was largely attributable to the participants not responding to requests to participate in later phases of the study. Severity of hearing impairment was categorised from the most recent audiological evaluation at audiology and cochlear implant clinics as moderate (40–69 dB HL), severe (70–94 dB HL) or profound (≥95 dB HL) according to four-frequency averaging of the pure-tone thresholds at 0.5, 1, 2 and 4 kHz in the better ear. It should be noted that the PCHL group is an unselected population of all children with bilateral PCHL of ≥40 dB and that those with severe and profound PCHL make up only half of the sample with the remainder having been diagnosed with moderate PCHL.

of the pure-tone thresholds at 0.5, 1, 2 and 4 kHz in the better ear. It should be noted that the PCHL group is an unselected population of all children with bilateral PCHL of ≥40 dB and that those with severe and profound PCHL make up only half of the sample with the remainder having been diagnosed with moderate PCHL. Other disabilities in addition to PCHL included cerebral palsy, visual impairment, and learning disability. The latter was determined by a Ravens Progressive Matrices score equivalent to a non-verbal IQ less than 70 [18]. The presence of conditions was noted from medical records and parent report. This study was approved by the Southampton and SW Hampshire Research Ethics Committee. Written informed consent for participation in the study was obtained from principal caregivers and from the teenage participants themselves. 2.2 Measures of EBD EBD were measured with teacher, parent and self-report versions of the Strengths and Difficulties Questionnaire (SDQ) [8]. This is a widely used behaviour screening questionnaire on children and young people's behaviours, emotions, and relationships. It has been recommended as suitable for use with children with PCHL [19], [20]. Total Difficulties score reflecting EBD in the child was derived from summing the scores of four SDQ scales (Emotional Symptoms, Conduct Problems, Hyperactivity, and Peer Problems) in the parent, teacher and self-report questionnaires separately. Higher Total Difficulties scores indicate more EBD. A fifth scale measured Prosocial Behaviour on which higher scores indicate more prosocial behaviour.

he scores of four SDQ scales (Emotional Symptoms, Conduct Problems, Hyperactivity, and Peer Problems) in the parent, teacher and self-report questionnaires separately. Higher Total Difficulties scores indicate more EBD. A fifth scale measured Prosocial Behaviour on which higher scores indicate more prosocial behaviour. The self-report SDQ has been found to be a less good predictor of psychiatric diagnosis than parent and teacher ratings [21] and there have been reports that the psychometric characteristics of the self-rated SDQ are less than optimal in relation to scale reliabilities [22] and to item loadings [23]. Teacher rated SDQ scores were obtained on less than 75% of the participants. For these reasons the report of the findings on the teacher-rated and self-rated SDQ scores will be limited to an initial comparison between the HCG and PCHL groups at Time 2. 2.3 Measures of non-verbal ability and language For the purpose of comparisons within the group of teenagers with PCHL on non-verbal ability and language, we used norms obtained from the HCG [15]. Each participant was assessed by a trained researcher, unaware of their audiological history. The group mean score and standard deviation score for a particular measure in teenagers in the HCG were used to derive age-adjusted z scores for the teenagers with PCHL on that measure.

e used norms obtained from the HCG [15]. Each participant was assessed by a trained researcher, unaware of their audiological history. The group mean score and standard deviation score for a particular measure in teenagers in the HCG were used to derive age-adjusted z scores for the teenagers with PCHL on that measure. 2.3.1 Receptive language The Test for Reception of Grammar Version 2 (TROG) [24] was used to assess participants' receptive skills for spoken English grammar, and the British Picture Vocabulary Scale Third Edition (BPVS) [25] provided a measure of their receptive skills for spoken English vocabulary. Both of these assessments had also been used to measure the participants' receptive language skills at primary school age. An aggregate measure of receptive language was obtained by averaging the z scores for the TROG and the BPVS.

(BPVS) [25] provided a measure of their receptive skills for spoken English vocabulary. Both of these assessments had also been used to measure the participants' receptive language skills at primary school age. An aggregate measure of receptive language was obtained by averaging the z scores for the TROG and the BPVS. 2.3.2 Expressive language The Expression, Reception and Recall of Narrative Instrument (ERNNI) [26] provided a measure of participants' expressive spoken language skills. Participants were required to produce a narrative based on a series of picture cues, and then subsequently to reproduce that narrative without the support of the pictures. Their narrative outputs were scored according to the ERRNI manual to produce three scores: an Initial score for the quality of their initial narrative, a Recall score for the quality of their recalled narrative, and a Mean Length of Utterance (MLU) score which reflected the average length of their utterances across both the initial and recall narratives. An aggregate measure of expressive language was obtained by averaging the z scores derived from the initial storytelling and recall scores from the ERRNI. 2.3.3 Non-verbal ability At Time 1 we assessed Non-Verbal Ability using the Raven's Standard Progressive Matrices [27]. At Time 2 the 20 min timed version [28] was used. Participants were given twenty minutes to work their way through a series of progressively more complex matrix reasoning puzzles. Raw scores reflecting the total number of correct items out of a possible 60 were calculated.

he Raven's Standard Progressive Matrices [27]. At Time 2 the 20 min timed version [28] was used. Participants were given twenty minutes to work their way through a series of progressively more complex matrix reasoning puzzles. Raw scores reflecting the total number of correct items out of a possible 60 were calculated. 2.4 Statistical analysis The primary outcome measure was specified as the Total Difficulties scores on the SDQ and group differences were tested with effect sizes obtained as standardised mean differences (SMDs) and associated 95% confidence intervals (95%CI). The distributions of the SDQ scores were somewhat skewed, as is usually found with this questionnaire [29]. To address this issue bootstrapped estimates of the standard errors were obtained [30]. A post-hoc power analysis indicated that, in a two-group comparison using a two-sided t-test, these sample sizes had 80% power to detect an SMD of 0.56 with alpha at 5%. The relationships between continuous measures, such as language scores, and the Total Difficulties score on the SDQ were tested using correlations. A post-hoc power analysis indicated that, using a two-sided test, the sample size of n = 72 had 80% power to detect a correlation of 0.32 with alpha at 5%. 3 Results 3.1 Characteristics of the samples Of the 76 participants with PCHL and the 38 in the HCG in the current study, 73 (96%) and 37 (97%) provided parent-rated, 55 (72%) and 28 (74%) teacher-rated and 65 (85%) and 38 (100%) self-rated SDQ data respectively.

2.4 Statistical analysis The primary outcome measure was specified as the Total Difficulties scores on the SDQ and group differences were tested with effect sizes obtained as standardised mean differences (SMDs) and associated 95% confidence intervals (95%CI). The distributions of the SDQ scores were somewhat skewed, as is usually found with this questionnaire [29]. To address this issue bootstrapped estimates of the standard errors were obtained [30]. A post-hoc power analysis indicated that, in a two-group comparison using a two-sided t-test, these sample sizes had 80% power to detect an SMD of 0.56 with alpha at 5%. The relationships between continuous measures, such as language scores, and the Total Difficulties score on the SDQ were tested using correlations. A post-hoc power analysis indicated that, using a two-sided test, the sample size of n = 72 had 80% power to detect a correlation of 0.32 with alpha at 5%. 3 Results 3.1 Characteristics of the samples Of the 76 participants with PCHL and the 38 in the HCG in the current study, 73 (96%) and 37 (97%) provided parent-rated, 55 (72%) and 28 (74%) teacher-rated and 65 (85%) and 38 (100%) self-rated SDQ data respectively. The teenagers with PCHL and the HCG were drawn from the same birth cohort and had similar baseline characteristics but those with PCHL were on average 0.74 years older than those in the HCG at the time of the teenage assessment (Table 1). However there were no significant correlations between age and scores on parent, teacher or self-rated SDQ Total Difficulties score in either the PCHL or HCG group.Table 1 Characteristics of children in the PCHL group and the HCG.

e 0.74 years older than those in the HCG at the time of the teenage assessment (Table 1). However there were no significant correlations between age and scores on parent, teacher or self-rated SDQ Total Difficulties score in either the PCHL or HCG group.Table 1 Characteristics of children in the PCHL group and the HCG. Table 1 PCHLa N = 76 Mean (SD) HCGb N = 38 Mean (SD) Standardised mean difference (SMD) or odds ratio (95%CI) Age at assessment (years) 17.09 (1.45) 16.35 (1.24) SMD = 0.57 (0.20–1.29) Non-verbal ability −0.28 (0.83) 0.00 (1.00) SMD = −0.30 (−0.70 to 0.10) n (%) n (%) Female gender 37 (48.7) 13 (34.2) OR = 0.55 (0.24–1.23) English main language at home 70 (92.1) 36 (94.7) OR = 1.54 (0.30–8.03) Degree of hearing loss Moderate 33 (43.4) – n/ac Severe/Profound 43 (56.6) – n/a Other disability Cerebral palsy 2 (2.6) 0 n/a Visual disability 2 (2.6) 0 n/a Learning disability 13 (17.1) 1 OR = 7.69 (0.96–62.50) None 62 (81.6) 37 (97) OR = 8.35 (1.06–66.16) Mother's education Less than A level 43 (56.6) 19 (50) OR = 1.30 (0.60–2.85) First cochlear implant Under age 8 years 11 (14.4) – n/a After 8 years of age 3 (3.9) – n/a Born in periods with UNHSd 37 (48.7) – n/a PCHL confirmed ≤ 9 months 35 (46.1) – n/a a PCHL Permanent childhood hearing loss. b HCG Hearing comparison group. c Not applicable. d Universal neonatal hearing screening for PCHL. There were no significant differences between the PCHL and HCG in terms of gender, English as the main language spoken by the family at home as reported by the parents, non-verbal ability or mothers education (see Table 1).

Table 1 PCHLa N = 76 Mean (SD) HCGb N = 38 Mean (SD) Standardised mean difference (SMD) or odds ratio (95%CI) Age at assessment (years) 17.09 (1.45) 16.35 (1.24) SMD = 0.57 (0.20–1.29) Non-verbal ability −0.28 (0.83) 0.00 (1.00) SMD = −0.30 (−0.70 to 0.10) n (%) n (%) Female gender 37 (48.7) 13 (34.2) OR = 0.55 (0.24–1.23) English main language at home 70 (92.1) 36 (94.7) OR = 1.54 (0.30–8.03) Degree of hearing loss Moderate 33 (43.4) – n/ac Severe/Profound 43 (56.6) – n/a Other disability Cerebral palsy 2 (2.6) 0 n/a Visual disability 2 (2.6) 0 n/a Learning disability 13 (17.1) 1 OR = 7.69 (0.96–62.50) None 62 (81.6) 37 (97) OR = 8.35 (1.06–66.16) Mother's education Less than A level 43 (56.6) 19 (50) OR = 1.30 (0.60–2.85) First cochlear implant Under age 8 years 11 (14.4) – n/a After 8 years of age 3 (3.9) – n/a Born in periods with UNHSd 37 (48.7) – n/a PCHL confirmed ≤ 9 months 35 (46.1) – n/a a PCHL Permanent childhood hearing loss. b HCG Hearing comparison group. c Not applicable. d Universal neonatal hearing screening for PCHL. There were no significant differences between the PCHL and HCG in terms of gender, English as the main language spoken by the family at home as reported by the parents, non-verbal ability or mothers education (see Table 1). As expected, a higher percentage of children with PCHL (18.4%) than those in the HCG (3%) had one or more of cerebral palsy, visual impairment or learning disability (OR = 8.35, 95%CI 1.06 to 66.16) (Table 1). The SMD for Total Difficulties as rated by parents for teenagers with Other disabilities compared to those with none was +1.68, 95%CI 1.04 to 2.32. For this reason the results are reported below both for the entire PCHL group and separately for those without other disabilities.

R = 8.35, 95%CI 1.06 to 66.16) (Table 1). The SMD for Total Difficulties as rated by parents for teenagers with Other disabilities compared to those with none was +1.68, 95%CI 1.04 to 2.32. For this reason the results are reported below both for the entire PCHL group and separately for those without other disabilities. 3.2 Effects of attrition To check on the possible biasing effect of selective attrition, those retained at follow-up, defined as a having a Time 2 score on parent-rated SDQ, were compared to those lost to follow-up on a range of measures at Time 1. For the PCHL group there were no significant differences between those retained and those lost to follow-up in terms of gender, mother's educational qualifications or English as the main language at home. There was no significant difference between these groups on the severity of hearing loss at Time 1. There were no significant differences in expressive and receptive language scores and nonverbal ability in those in the PCHL group at Time 1 and those lost to follow-up. Lastly there were no significant differences in the group mean parent-rated Total Difficulties scores at Time 1 between those in the PCHL group who were followed up and those lost to follow-up (SMD = −0.17, 95%CI −0.55 to 0.20).

res and nonverbal ability in those in the PCHL group at Time 1 and those lost to follow-up. Lastly there were no significant differences in the group mean parent-rated Total Difficulties scores at Time 1 between those in the PCHL group who were followed up and those lost to follow-up (SMD = −0.17, 95%CI −0.55 to 0.20). For the HCG those lost to follow-up did not differ significantly from those retained on gender, mother's educational qualifications and English as the main language at home. The lost to follow-up and retained groups had similar parent-rated Total Difficulties scores at Time 1 (SMD = −0.12, 95%CI −0.63 to 0.38) and therefore on the variable of central concern in the analyses in the present paper there was no significant difference.

al qualifications and English as the main language at home. The lost to follow-up and retained groups had similar parent-rated Total Difficulties scores at Time 1 (SMD = −0.12, 95%CI −0.63 to 0.38) and therefore on the variable of central concern in the analyses in the present paper there was no significant difference. 3.3 Comparison of the PCHL group and HCG on total difficulties scores in adolescence On parent-rated SDQ, the PCHL group had significantly greater Total Difficulties scores than the HCG (SMD = +0.39, 95%CI 0.00 to 0.79) (F = 4.23, df = 1,106, p = 0.04) (Table 2). There were no main effects of gender (F = 0.03, df = 1,106, p = 0.87). The interaction between gender and group was not significant (F = 0.98, df = 1,106, p = 0.32). For teacher rated SDQ there was no significant difference between the PCHL group and the HCG (SMD = +0.17, 95%CI −0.28 to 0.62) (F = 1.04, df = 1,80, p = 0.31). On teacher ratings males scored more highly than females (F = 5.89, df = 1,80, p = 0.02). The interaction between group and gender for teacher ratings was not significant (F = 0.11, df = 1,80, p = 0.74).Table 2 Mean and SD of SDQ Total Difficulties scores for adolescents with PCHL and the HCG by gender.

,80, p = 0.31). On teacher ratings males scored more highly than females (F = 5.89, df = 1,80, p = 0.02). The interaction between group and gender for teacher ratings was not significant (F = 0.11, df = 1,80, p = 0.74).Table 2 Mean and SD of SDQ Total Difficulties scores for adolescents with PCHL and the HCG by gender. Table 2 PCHLa HCGb n Mean SD n Mean SD Standardised mean difference (SMD) SMD 95%CI Parent-rated Both sexes 73 8.48 6.17 37 6.22 4.95 +0.39 0.00 to 0.79 Females 35 9.00 6.21 13 5.31 4.27 +0.64 0.00 to 1.29 Males 38 8.00 6.18 24 6.71 5.30 +0.22 −0.29 to 0.73 Teacher-rated Both sexes 55 6.25 5.46 29 5.38 4.62 +0.17 −0.28 to 0.62 Females 24 4.83 5.30 10 3.20 3.94 +0.33 −0.41 to 1.07 Males 31 7.35 5.42 19 6.53 4.26 +0.16 −0.40 to 0.73 Self-rated Both sexes 65 9.74 5.18 38 9.13 5.14 +0.12 −0.28 to 0.52 Females 32 8.94 5.80 13 9.00 5.31 −0.01 −0.66 to 0.63 Males 33 10.52 4.44 25 9.20 5.16 +0.28 −0.24 to 0.79 a PCHL Permanent childhood hearing loss. b HCG Hearing comparison group. There was no difference in self-rated group mean Total Difficulties between the PCHL and the HCG groups (SMD = +0.12, 95%CI −0.28 to 0.52) (F = 0.33, df = 1,99, p = 0.56) (Table 2). There was no significant effect of gender (F = 0.36, df = 1,99, p = 0.42) nor a significant interaction between gender and group (F = 0.40, df = 1,99, p = 0.53) on self-reported SDQ Total Difficulties scores. For those teenagers with PCHL and no disabilities the SMD compared to the HCG on parent-rated SDQ Total Difficulties was smaller and no longer significant (SMD = +0.21 95%CI −0.21 to 0.62).

There was no difference in self-rated group mean Total Difficulties between the PCHL and the HCG groups (SMD = +0.12, 95%CI −0.28 to 0.52) (F = 0.33, df = 1,99, p = 0.56) (Table 2). There was no significant effect of gender (F = 0.36, df = 1,99, p = 0.42) nor a significant interaction between gender and group (F = 0.40, df = 1,99, p = 0.53) on self-reported SDQ Total Difficulties scores. For those teenagers with PCHL and no disabilities the SMD compared to the HCG on parent-rated SDQ Total Difficulties was smaller and no longer significant (SMD = +0.21 95%CI −0.21 to 0.62). 3.4 Comparison of the PCHL group and national norms on total difficulties scores in adolescence The mean scores for the HCG in Table 2 are somewhat lower than those in a large normative sample [31]. One sample t-tests were carried out to test whether the PCHL group had scores that were significantly different from these norm values (11–15 year olds - parent rated SDQ mean = 8.2, SD = 5.8; - teacher rated SDQ mean = 6.3, SD = 6.1). Neither the parent rated mean score (t = 0.39, df = 72, p = 0.70) nor the teacher rated mean score (t = 0.06, df = 54, p = 0.95) showed a significant difference from the mean of the norm group. 3.5 PCHL and HCG differences on specific types of EBD in adolescence The differences in means between teenagers with PCHL and the HCG were not significant on the EBD sub-scales of the parent-rated SDQ (Multivariate F = 1.25, df = 4, 105, p = 0.30). This pattern of results remained unchanged when the analysis was limited to those without other disabilities.

ic types of EBD in adolescence The differences in means between teenagers with PCHL and the HCG were not significant on the EBD sub-scales of the parent-rated SDQ (Multivariate F = 1.25, df = 4, 105, p = 0.30). This pattern of results remained unchanged when the analysis was limited to those without other disabilities. There were no significant differences between the PCHL and HCG teenagers on the parent-rated SDQ Prosocial scale (SMD = +0.21, 95%CI −0.19 to 0.61). When the analysis was limited to those without other disabilities, those with PCHL had significantly higher Prosocial scores than the HCG (SMD = +0.64, 95%CI 0.22 to 1.06). Teenagers with PCHL were not reported to have more difficulties than the HCG teenagers on the EBD sub-scales of the teacher-rated SDQ (Multivariate F = 0.25, df = 4, 79, p = 0.91). This pattern of results remained unchanged when the analysis was limited to those without other disabilities. There were no significant differences between the PCHL and HCG teenagers on the teacher-rated SDQ Prosocial scale (SMD = −0.0.09, 95%CI −0.54 to 0.36). This analysis was unchanged when it was limited to those without other disabilities.

Teenagers with PCHL were not reported to have more difficulties than the HCG teenagers on the EBD sub-scales of the teacher-rated SDQ (Multivariate F = 0.25, df = 4, 79, p = 0.91). This pattern of results remained unchanged when the analysis was limited to those without other disabilities. There were no significant differences between the PCHL and HCG teenagers on the teacher-rated SDQ Prosocial scale (SMD = −0.0.09, 95%CI −0.54 to 0.36). This analysis was unchanged when it was limited to those without other disabilities. Teenagers with PCHL reported significantly higher overall scores than the HCG on the self-rated EBD sub-scales (Multivariate F = 3.32, df = 4, 98, p = 0.01). This arose from the high score on the Peer Problems sub-scale (SMD = +0.54, 95%CI 0.13 to 0.95) (F = 6.98, df = 1, 101, p = 0.01). The other sub-scales showed no significant differences. This pattern of results remained unchanged when the analysis was limited to those without other disabilities. There were no significant differences between the PCHL and HCG teenagers on the self-rated SDQ Prosocial scale (SMD = +0.07, 95%CI −0.33 to 0.47). This result remained unchanged when the analysis was limited to those without other disabilities.

Teenagers with PCHL reported significantly higher overall scores than the HCG on the self-rated EBD sub-scales (Multivariate F = 3.32, df = 4, 98, p = 0.01). This arose from the high score on the Peer Problems sub-scale (SMD = +0.54, 95%CI 0.13 to 0.95) (F = 6.98, df = 1, 101, p = 0.01). The other sub-scales showed no significant differences. This pattern of results remained unchanged when the analysis was limited to those without other disabilities. There were no significant differences between the PCHL and HCG teenagers on the self-rated SDQ Prosocial scale (SMD = +0.07, 95%CI −0.33 to 0.47). This result remained unchanged when the analysis was limited to those without other disabilities. 3.6 Changes with age in EBD as measured by parent-rating in the PCHL group and the HCG The most appropriate way to examine longitudinal changes is to investigate groups for whom parent report is available in both childhood and adolescence for the same participants. Applying this to the 72 participants with PCHL for whom parent report was available at both time points, the mean Total Difficulties score showed no significant change from 9.22 to 8.29 (SMD = −0.16, 95%CI −0.48 to 0.17) (Table 3). The stability of individual differences in the Total Difficulties score is indicated by the correlation (r) between time points of 0.52 (95%CI 0.32 to 0.68).Table 3 Mean and SD SDQ parent-rated Total Difficulties and sub-scale scores for adolescents with PCHL and the HCG at Time 1 (6–10 years) and Time 2 (13–20 years).

The stability of individual differences in the Total Difficulties score is indicated by the correlation (r) between time points of 0.52 (95%CI 0.32 to 0.68).Table 3 Mean and SD SDQ parent-rated Total Difficulties and sub-scale scores for adolescents with PCHL and the HCG at Time 1 (6–10 years) and Time 2 (13–20 years). Table 3 PCHLa HCGb Time 1 Time 2 Time 1 Time 2 n Mean SD Mean SD SMDc 95%CI Paireddt-test r n Mean SD Mean SD SMDc 95%CI Paireddt-test r Total difficulties 72 9.22 5.81 8.29 6.00 −0.16 −0.48 to 0.17 t = 1.36,df = 71,p = 0.18 0.52 37 6.49 3.73 6.22 4.95 −0.06 −0.51 to 0.39 t = 0.38,df = 36,p = 0.71 0.54 Emotional Symptoms 72 1.67 1.70 1.92 1.95 +0.14 −0.19 to 0.46 t = 1.05,df = 71,p = 0.30 0.40 37 1.54 1.68 1.30 1.61 −0.15 −0.60 to 0.31 t = 0.81,df = 36,p = 0.42 0.39 Conduct Problems 72 1.60 1.62 1.21 1.36 −0.26 −0.59 to 0.07 t = 1.86,df = 71,p = 0.08 0.24 37 0.86 0.95 1.03 1.28 +0.15 −0.31 to 0.61 t = 0.76,df = 36,p = 0.45 0.35 Hyperactivity 72 4.43 3.12 3.24 2.59 −0.42 −0.74 to −0.08 t = 3.66,df = 71,p < 0.001 0.54 37 3.03 2.24 2.59 2.17 −0.20 −0.66 to 0.26 t = 1.15,df = 36,p = 0.26 0.47 Peer problems 72 1.53 1.76 1.93 1.77 +0.23 −0.10 to 0.55 t = 1.91,df = 71,p = 0.06 0.49 37 1.05 1.60 1.30 1.85 +0.14 −0.31 to 0.60 t = 0.81,df = 36,p = 0.42 0.45 Prosocial behaviour 72 7.58 2.38 8.43 2.19 +0.37 0.04 to 0.70 t = 3.42,df = 71,p = 0.001 0.58 37 8.38 1.42 7.89 1.47 −0.34 −0.80 to 0.12 t = 2.64,df = 36p = 0.01 0.70 a PCHL Permanent childhood hearing loss. b HCG Hearing comparison group.

Table 3 PCHLa HCGb Time 1 Time 2 Time 1 Time 2 n Mean SD Mean SD SMDc 95%CI Paireddt-test r n Mean SD Mean SD SMDc 95%CI Paireddt-test r Total difficulties 72 9.22 5.81 8.29 6.00 −0.16 −0.48 to 0.17 t = 1.36,df = 71,p = 0.18 0.52 37 6.49 3.73 6.22 4.95 −0.06 −0.51 to 0.39 t = 0.38,df = 36,p = 0.71 0.54 Emotional Symptoms 72 1.67 1.70 1.92 1.95 +0.14 −0.19 to 0.46 t = 1.05,df = 71,p = 0.30 0.40 37 1.54 1.68 1.30 1.61 −0.15 −0.60 to 0.31 t = 0.81,df = 36,p = 0.42 0.39 Conduct Problems 72 1.60 1.62 1.21 1.36 −0.26 −0.59 to 0.07 t = 1.86,df = 71,p = 0.08 0.24 37 0.86 0.95 1.03 1.28 +0.15 −0.31 to 0.61 t = 0.76,df = 36,p = 0.45 0.35 Hyperactivity 72 4.43 3.12 3.24 2.59 −0.42 −0.74 to −0.08 t = 3.66,df = 71,p < 0.001 0.54 37 3.03 2.24 2.59 2.17 −0.20 −0.66 to 0.26 t = 1.15,df = 36,p = 0.26 0.47 Peer problems 72 1.53 1.76 1.93 1.77 +0.23 −0.10 to 0.55 t = 1.91,df = 71,p = 0.06 0.49 37 1.05 1.60 1.30 1.85 +0.14 −0.31 to 0.60 t = 0.81,df = 36,p = 0.42 0.45 Prosocial behaviour 72 7.58 2.38 8.43 2.19 +0.37 0.04 to 0.70 t = 3.42,df = 71,p = 0.001 0.58 37 8.38 1.42 7.89 1.47 −0.34 −0.80 to 0.12 t = 2.64,df = 36p = 0.01 0.70 a PCHL Permanent childhood hearing loss. b HCG Hearing comparison group. c Based on means and SD; this gives a smaller effect size and wider confidence intervals than those derived from the value of t in the paired-test (Dunlap et al., 1996). d p values based on bootstrapping with 1000 iterations.

Table 3 PCHLa HCGb Time 1 Time 2 Time 1 Time 2 n Mean SD Mean SD SMDc 95%CI Paireddt-test r n Mean SD Mean SD SMDc 95%CI Paireddt-test r Total difficulties 72 9.22 5.81 8.29 6.00 −0.16 −0.48 to 0.17 t = 1.36,df = 71,p = 0.18 0.52 37 6.49 3.73 6.22 4.95 −0.06 −0.51 to 0.39 t = 0.38,df = 36,p = 0.71 0.54 Emotional Symptoms 72 1.67 1.70 1.92 1.95 +0.14 −0.19 to 0.46 t = 1.05,df = 71,p = 0.30 0.40 37 1.54 1.68 1.30 1.61 −0.15 −0.60 to 0.31 t = 0.81,df = 36,p = 0.42 0.39 Conduct Problems 72 1.60 1.62 1.21 1.36 −0.26 −0.59 to 0.07 t = 1.86,df = 71,p = 0.08 0.24 37 0.86 0.95 1.03 1.28 +0.15 −0.31 to 0.61 t = 0.76,df = 36,p = 0.45 0.35 Hyperactivity 72 4.43 3.12 3.24 2.59 −0.42 −0.74 to −0.08 t = 3.66,df = 71,p < 0.001 0.54 37 3.03 2.24 2.59 2.17 −0.20 −0.66 to 0.26 t = 1.15,df = 36,p = 0.26 0.47 Peer problems 72 1.53 1.76 1.93 1.77 +0.23 −0.10 to 0.55 t = 1.91,df = 71,p = 0.06 0.49 37 1.05 1.60 1.30 1.85 +0.14 −0.31 to 0.60 t = 0.81,df = 36,p = 0.42 0.45 Prosocial behaviour 72 7.58 2.38 8.43 2.19 +0.37 0.04 to 0.70 t = 3.42,df = 71,p = 0.001 0.58 37 8.38 1.42 7.89 1.47 −0.34 −0.80 to 0.12 t = 2.64,df = 36p = 0.01 0.70 a PCHL Permanent childhood hearing loss. b HCG Hearing comparison group. c Based on means and SD; this gives a smaller effect size and wider confidence intervals than those derived from the value of t in the paired-test (Dunlap et al., 1996). d p values based on bootstrapping with 1000 iterations. In the 37 participants from the HCG in whom parent report was available at both time points, there was no significant change over time in the mean Total Difficulties score from 6.49 to 6.22 (SMD = −0.06, 95%CI −0.51 to 0.39) with r = 0.54 (95%CI 0.27 to 0.94).

c Based on means and SD; this gives a smaller effect size and wider confidence intervals than those derived from the value of t in the paired-test (Dunlap et al., 1996). d p values based on bootstrapping with 1000 iterations. In the 37 participants from the HCG in whom parent report was available at both time points, there was no significant change over time in the mean Total Difficulties score from 6.49 to 6.22 (SMD = −0.06, 95%CI −0.51 to 0.39) with r = 0.54 (95%CI 0.27 to 0.94). The only problem scale to show a significant decline was Hyperactivity for those with PCHL (SMD = −0.42, 95%CI −0.74 to −0.08). For the PCHL group there was a non-significant tendency (p < 0.06) to show an increase over time in Peer problems (SMD = +0.23, 95%CI −0.10 to 0.55). The Prosocial scale showed a significant increase in those with PCHL (SMD = +0.37, 95%CI 0.04 to 0.70) but no significant change in the HCG (SMD = −0.34, 95%CI −0.80 to 0.12). The pattern of the results presented in Table 3 was unchanged when the analysis was limited to those without other disabilities. The difference between the parent-rated Total Difficulties scores for the PCHL group and the HCG in this longitudinally studied sub-sample declined from SMD = +0.52 (95%CI 0.12 to 0.93) at Time 1 to SMD = +0.36 (95%CI −0.03 to 0.76) at Time 2. However this change in the difference is not significant in a repeated measures ANOVA (F = 0.37, df = 1,107, p = 0.54).

al Difficulties scores for the PCHL group and the HCG in this longitudinally studied sub-sample declined from SMD = +0.52 (95%CI 0.12 to 0.93) at Time 1 to SMD = +0.36 (95%CI −0.03 to 0.76) at Time 2. However this change in the difference is not significant in a repeated measures ANOVA (F = 0.37, df = 1,107, p = 0.54). 3.7 Cognitive and language abilities and EBD in the PCHL group and the HCG in adolescence The PCHL group had poorer receptive language abilities than the HCG [13]. The correlation between receptive language and parent-rated Total Difficulties scores was significant for both the PCHL group (r = −0.32, 95%CI −0.53 to −0.02) and HCG (r = −0.33, 95%CI −0.59 to −0.01). For expressive language the correlations with parent-rated Total Difficulties scores were lower and not significant (PCHL r = −0.02, 95%CI −0.28 to 0.24, HCG r = 0.12, 95%CI −0.21 to 0.43). For non-verbal ability there was a significant correlation for the HCG only (PCHL r = −0.01, 95%CI −0.26 to 0.24; HCG r = −0.44, 95%CI −0.66 to −0.13). Receptive language ability was therefore tested as a factor potentially accounting for the PCHL/HCG difference in Total Difficulties. An ANOVA showed a marginally non-significant effect of group (i.e. PCHL/HCG) on Total Difficulties score (F = 3.75, df = 1,108, p = 0.05). When receptive language was entered as a covariate the effect of group was no longer significant (F = 0.12, df = 1,95, p = 0.73).

for the PCHL/HCG difference in Total Difficulties. An ANOVA showed a marginally non-significant effect of group (i.e. PCHL/HCG) on Total Difficulties score (F = 3.75, df = 1,108, p = 0.05). When receptive language was entered as a covariate the effect of group was no longer significant (F = 0.12, df = 1,95, p = 0.73). Within the PCHL group those with severe/profound PCHL showed receptive language ability scores below those with moderate hearing impairment but this was not significant (p = 0.08) (SMD = +0.44, 95%CI −0.06 to 0.94). As a further check on the relationships between language, hearing loss and EBD, the correlation between Total Difficulties scores and receptive language ability scores was calculated within the moderate and severe/profound PCHL groups separately. These correlations were negative in each case; the lower the receptive language score, the higher the EBD score. The correlation with parent-rated Total Difficulties scores was significant for the moderate (r = −0.39, 95%CI −0.65 to −0.05) but not the severe/profound group (r = −0.21, 95%CI −0.53 to 0.17). Therefore the relationship between receptive language ability and EBD remains when severity of hearing impairment is taken into account, at least for the moderate hearing loss group. These results suggest that it is the difference in receptive language between the PCHL group and the HCG, rather than hearing impairment per se, that accounts for the effect of PCHL on EBD.

e ability and EBD remains when severity of hearing impairment is taken into account, at least for the moderate hearing loss group. These results suggest that it is the difference in receptive language between the PCHL group and the HCG, rather than hearing impairment per se, that accounts for the effect of PCHL on EBD. This analysis was repeated for those without other disabilities. Again the correlation with receptive language ability for parent-rated Total Difficulties was significant for the moderate hearing impairment group (r = −0.47, 95%CI −0.72 to −0.18) but not for the severe/profound impairment group (r = 0.13, 95%CI −0.26 to 0.49).

lysis was repeated for those without other disabilities. Again the correlation with receptive language ability for parent-rated Total Difficulties was significant for the moderate hearing impairment group (r = −0.47, 95%CI −0.72 to −0.18) but not for the severe/profound impairment group (r = 0.13, 95%CI −0.26 to 0.49). 3.8 Factors within the PCHL group related to total difficulties scores Parent-rated Total Difficulties scores were not significantly different between the participants with PCHL with and without a cochlear implant (SMD = +0.35, 95%CI −0.25 to 0.96). If the teenagers without a cochlear implant are compared on parent-rated Total Difficulties to other teenagers with severe/profound PCHL and a cochlear implant, the effect was more marked with those having cochlear implants having lower scores but was not significant (p = 0.08) (SMD = +0.61, 95%CI −0.08 to 1.29). There were 7% of those with other disabilities who received cochlear implants compared to 21% of those without other disabilities. This difference was not significant (OR = 3.45, 95%CI 0.41 to 28.85). This analysis was repeated for those without other disabilities and the pattern of results was unchanged. The difference between those with and without cochlear implants was not significant on parent-rated Total Difficulties scores (SMD = +0.14, 95%CI −0.50 to 0.77). The comparison with of those with a cochlear implant with other teenagers with severe/profound PCHL was also not significant (SMD = +0.35, 95%CI −0.38 to 1.07).

e difference between those with and without cochlear implants was not significant on parent-rated Total Difficulties scores (SMD = +0.14, 95%CI −0.50 to 0.77). The comparison with of those with a cochlear implant with other teenagers with severe/profound PCHL was also not significant (SMD = +0.35, 95%CI −0.38 to 1.07). The SMD for parent-rated Total Difficulties in teenagers with severe/profound hearing loss compared to those with moderate hearing loss did not differ significantly for either the whole sample (SMD = +0.19, 95%CI −0.26 to 0.66) or those without other disabilities (SMD = +0.20, 95%CI −0.30 to 0.72). Both receptive language ability and other disabilities were related to Total Difficulties scores. They are also closely related to each other. To test whether they are independently related to Total Difficulties scores, the correlations between receptive language ability and Total Difficulties score were calculated for the whole PCHL group (r = −0.32, 95%CI −0.53 to −0.07) and then for the group without other disabilities (r = −0.23, 95%CI −0.47 to 0.04). The latter correlation was corrected for restriction of range in the receptive language scores arising from this selection. The strength of relationship between receptive language ability and Total Behaviour score was reduced for those without other disabilities. This suggests that in part, the association between other disabilities and a high EBD score is mediated via poorer receptive language development.

ores arising from this selection. The strength of relationship between receptive language ability and Total Behaviour score was reduced for those without other disabilities. This suggests that in part, the association between other disabilities and a high EBD score is mediated via poorer receptive language development. 3.9 Factors within the PCHL group without other disabilities related to Prosocial behaviour scores Those without other disabilities in the PCHL group had higher Prosocial behaviour scores on parent ratings than the HCG and the mean Prosocial score for this group increased significantly from Time 1 to Time 2. As with Total Difficulties score, a number of factors were therefore examined to determine if they related to variation in Prosocial behaviour scores in this sub-group of the participants with PCHL. At Time 1 (SMD = 0.62, 95%CI 0.09 to 1.13) but not at Time 2 (SMD = 0.34, 95%CI −0.16 to 0.86) there were significantly higher parent rated Prosocial behaviour scores in females compared to males. A repeated measures ANOVA showed there to be significant effect of age (F = 9.55, df = 1,57, p = 0.003) and gender (F = 5.44, df = 1, 57, p = 0.023) but no significant interaction between age and gender (F = 1.24, df = 1,57, p = 0.269).

cantly higher parent rated Prosocial behaviour scores in females compared to males. A repeated measures ANOVA showed there to be significant effect of age (F = 9.55, df = 1,57, p = 0.003) and gender (F = 5.44, df = 1, 57, p = 0.023) but no significant interaction between age and gender (F = 1.24, df = 1,57, p = 0.269). At Time 1 (SMD = −0.13, 95%CI, −0.75 to 0.48) and at Time 2 (SMD = −0.37, 95%CI −1.00 to 0.27) there were no significant differences in the Prosocial behaviour scores rated by parents for those with and without cochlear implants. This pattern of results for cochlear implantation was unchanged if the analysis was restricted to those with other disabilities and severe/profound degrees of hearing loss. The mean parent rated Prosocial score at Time 1 for those with moderate was significantly greater than that for those with severe/profound hearing loss (SMD = 0.58, 95%CI 0.07 to 1.91) at Time 1 but not at Time 2 (SMD = 0.48, 95%CI −0.05 to 1.00). A repeated measures ANOVA showed there to be significant effect of age (F = 9.50, df = 1,57, p = 0.004) and degree of hearing loss (F = 5.98 df = 1, 57, p = 0.018) but no significant interaction between age and degree of hearing loss (F = 0.28, df = 1,57, p = 0.600). There was no significant correlation between receptive language score and parent rated Prosocial scores at Time 1 (r = 0.07, 95%CI −0.20 to 0.33) or Time 2 (r = −0.04, 95%CI −0.30 to 0.23). These analyses suggest the only factors related to high Prosocial behaviour scores in the PCHL without other disabilities group were age, female gender and less severe hearing loss.

There was no significant correlation between receptive language score and parent rated Prosocial scores at Time 1 (r = 0.07, 95%CI −0.20 to 0.33) or Time 2 (r = −0.04, 95%CI −0.30 to 0.23). These analyses suggest the only factors related to high Prosocial behaviour scores in the PCHL without other disabilities group were age, female gender and less severe hearing loss. 3.10 UNHS, early confirmation of hearing impairment and the risk of EBD The cohort study from which these participants were drawn was originally designed to examine whether the provision of UNHS for hearing impairment improved the outcome for children with PCHL. Some of the children were born in periods when UNHS was available and others not. We found that in adolescence there was an effect of exposure to UNHS on EBD by parent-ratings (SMD = +0.48, 95%CI 0.01 to 0.95): the scores of those born during periods of UNHS were lower than the scores of those born in periods with no UNHS (Table 4). When severity of hearing loss was added as a factor the effect of UNHS remained significant (F = 4.26, df = 1, 73, p = 0.04). The effect of exposure to UNHS on parent-rated SDQ Total Difficulties was unaffected when the comparison was limited to those without other disabilities (SMD = +0.52, 95%CI 0.00 to 1.04). Those born in periods with UNHS tended to have higher receptive language scores (SMD = −0.25 95%CI −0.75 to 0.23) but this was not significant. For the PCHL group as a whole, receptive language scores were related to parent-rated SDQ Total Difficulties scores. When receptive language ability was added as a covariate the effect of UNHS was no longer significant (SMD = +0.42, 95%CI −0.08 to 0.93). The possible beneficial of exposure to UNHS on behaviour was therefore at least partially explained by the higher receptive language scores in this group.Table 4 Mean parent-rated SDQ Total Difficulties scores by exposure to Universal Newborn Hearing Screening and early/late confirmation of Permanent Childhood Hearing Loss (PCHL) in total PCHL sample and in those without other disabilities.

least partially explained by the higher receptive language scores in this group.Table 4 Mean parent-rated SDQ Total Difficulties scores by exposure to Universal Newborn Hearing Screening and early/late confirmation of Permanent Childhood Hearing Loss (PCHL) in total PCHL sample and in those without other disabilities. Table 4 No UNHS UNHS N Mean SD N Mean SD SMD 95%CI t df p Total PCHL sample 38 9.87 5.81 35 6.97 6.28 +0.48 0.01 to 0.95 2.05 71 0.04 No other disabilities 29 8.14 4.72 30 5.53 5.26 +0.52 0.00 to 1.04 2.00 57 0.05 Early confirmed Late confirmed N Mean SD N Mean SD SMD 95%CI t df p Total PCHL sample 34 9.50 7.21 39 7.59 5.03 +0.31 −0.15 to 0.77 1.33 71 0.19 No other disabilities 26 7.00 5.79 33 6.67 4.63 +0.06 −0.45 to 0.58 0.25 57 0.80 PCHL = permanent childhood hearing loss; SD = Standard Deviation; SMD = Standardised Mean Difference. UNHS= Universal newborn hearing screening.

N Mean SD N Mean SD SMD 95%CI t df p Total PCHL sample 34 9.50 7.21 39 7.59 5.03 +0.31 −0.15 to 0.77 1.33 71 0.19 No other disabilities 26 7.00 5.79 33 6.67 4.63 +0.06 −0.45 to 0.58 0.25 57 0.80 PCHL = permanent childhood hearing loss; SD = Standard Deviation; SMD = Standardised Mean Difference. UNHS= Universal newborn hearing screening. Early confirmation of PCHL (i.e. by age 9 months) compared to late confirmation did not have a significant effect on parent-rated Total Difficulties (SMD = +0.31, 95%CI −0.15 to 0.77) (Table 4). The percentage of early and late confirmed participants with severe/profound impairment was 47.4% and 49.2% respectively (OR = 1.19, 95%CI 0.45 to 1.90) and the percentage of early confirmed participants with other disabilities was 22.9% compared to 14.6% of late confirmed participants (OR = 1.72, 95%CI 0.54 to 5.57). The effect of early confirmation on parent rated SDQ Total Difficulties at Time 2 remained not significant when restricted to those without cochlear implants (SMD = 0.25, 95%CI −0.03 to 0.77). The effect of early confirmation remained non-significant when the analysis was restricted to those without other disabilities (SMD = +0.06 95%CI −0.45 to 0.58). It also remained not significant when severity of hearing loss was added as a factor (F = 1.91, df = 1,73, p = 0.171). The lack of an effect of early confirmation was therefore not attributable to a confound with severity of hearing impairment (i.e. more severe impairment leads to early confirmation).

I −0.45 to 0.58). It also remained not significant when severity of hearing loss was added as a factor (F = 1.91, df = 1,73, p = 0.171). The lack of an effect of early confirmation was therefore not attributable to a confound with severity of hearing impairment (i.e. more severe impairment leads to early confirmation). 3.11 Age related changes in the effects of age at confirmation and exposure to UHNS There is a relationship between age at confirmation and UNHS with early confirmation being more common when UNHS is in place (66% vs. 34%, OR = 3.69, 95%CI 1.42 to 9.52). To examine their joint relationship with parent-rated Total Difficulties a 2 × 2 ANOVA was conducted with age at confirmation and UNHS as factors and Total Difficulties score as the dependent variable. This analysis was repeated for behaviour at Time 1 (childhood) and at Time 2 (adolescence) for those participants with parent-rated SDQ scores at both time points (N = 72). The analysis was repeated excluding participants with other disabilities as the presence of such disabilities may have led to early confirmation and therefore distort the findings on the effects of early confirmation of hearing impairment per se.

e participants with parent-rated SDQ scores at both time points (N = 72). The analysis was repeated excluding participants with other disabilities as the presence of such disabilities may have led to early confirmation and therefore distort the findings on the effects of early confirmation of hearing impairment per se. These results are presented in Fig. 1. The pattern of means was similar in childhood (Fig. 1A) and in adolescence (Fig. 1B). The effect of UNHS was significant in adolescence for the Total sample (Fig. 1B) and for those with no other disabilities (Fig. 1D). The effect of age of confirmation was marginally non-significant in adolescence (p < 0.07) but not significant in childhood.Fig. 1 Mean parent-rated SDQ Total Difficulties score and 95%CI in childhood and adolescence for early/late confirmed and UNHS/No UNHS groups (UNHS = Universal newborn hearing screening). Fig. 1

These results are presented in Fig. 1. The pattern of means was similar in childhood (Fig. 1A) and in adolescence (Fig. 1B). The effect of UNHS was significant in adolescence for the Total sample (Fig. 1B) and for those with no other disabilities (Fig. 1D). The effect of age of confirmation was marginally non-significant in adolescence (p < 0.07) but not significant in childhood.Fig. 1 Mean parent-rated SDQ Total Difficulties score and 95%CI in childhood and adolescence for early/late confirmed and UNHS/No UNHS groups (UNHS = Universal newborn hearing screening). Fig. 1 4 Discussion Compared to teenagers with normal hearing, the teenagers with bilateral permanent childhood hearing loss in this study showed significantly greater emotional and behaviour difficulties scores in late adolescence, but only as reported by their parents. For teachers' and self reports these differences fell short of significance. It should be noted however that the study had less power to detect effects on teacher-rated behaviour as there were fewer participants for whom this measure was available. This suggests that the elevated rate of emotional and behavioural problems previously reported by parents in children with PCHL is also found in adolescence, although the mean Total Difficulties scores in both the PCHL group and the HCG were lower in adolescence compared to that in childhood in this longitudinal study; the difference between those with PCHL and the HCG group fell from 0.52 to 0.36 SDs between middle childhood and the late teenage years. Correlations indicated moderate stability in EBD scores between childhood and adolescence in those with PCHL.

in adolescence compared to that in childhood in this longitudinal study; the difference between those with PCHL and the HCG group fell from 0.52 to 0.36 SDs between middle childhood and the late teenage years. Correlations indicated moderate stability in EBD scores between childhood and adolescence in those with PCHL. The elevated scores relative to the HCG on the parent-rated EBD measure need to be put in the context of the number of teenagers with PCHL who show abnormally high SDQ Total Difficulties scores. On the parent-rated SDQ, borderline and abnormal scores are designated by a score of 12 or higher. In the PCHL group 21.9% had scores at this level. This suggests that the majority of teenagers with PCHL do not have EBD. Additionally, the presence of other disabilities substantially increased the Total Difficulties score. Once those with other disabilities were excluded the size of the difference in parent-rated EBD between the PCHL group and the HCG was greatly reduced and no longer significant. As well as factors directly associated with the disability, it should be noted that there are other factors, such as increased family stresses, possibly contributing to the vulnerability to EBD for those with other disabilities [32].

EBD between the PCHL group and the HCG was greatly reduced and no longer significant. As well as factors directly associated with the disability, it should be noted that there are other factors, such as increased family stresses, possibly contributing to the vulnerability to EBD for those with other disabilities [32]. On the individual SDQ subscales, the PCHL group showed significantly higher self-rated Peer Problems than the HCG, though this difference was not significant for parent ratings on this subscale. It is also interesting to note that at Time 2 those with PCHL showed a higher level of parent-rated Prosocial behaviour than the HCG, but only for those without other disabilities. The presence of moderate rather than severe/profound degrees of hearing loss was also related to higher Prosocial behaviour scores. These results for Prosocial behaviour are at variance with the findings from a meta-analysis of SDQ sub-scales scores in those with a hearing loss [2]. In that analysis of the results of 10 studies using the parent rated SDQ there was a significant difference in the opposite direction with those with hearing loss having lower Prosocial scores than hearing children (SMD = 0.30, 95%CI 0.08 to 0.52) (N.B. the SMD reported in the meta-analysis had a reversed sign to that used here). In that meta-analysis, norms for the SDQ from the UK population [31] were used for comparison with the HL groups. If this comparison is used for the group without other disabilities in the present data set, the effect size (95%CI) falls from a significant SMD = 0.64, (0.22–0.1.06) to a non-significant SMD = 0.16 (−0.09 to 0.42). This disparity might also in part be due to the inclusion of children with other disabilities in the HL group in the meta-analysis.

or the group without other disabilities in the present data set, the effect size (95%CI) falls from a significant SMD = 0.64, (0.22–0.1.06) to a non-significant SMD = 0.16 (−0.09 to 0.42). This disparity might also in part be due to the inclusion of children with other disabilities in the HL group in the meta-analysis. A meta-analysis of SDQ reports on children with HL reported a similar pattern to that seen in the present study with higher Total Difficulties mean scores on parent and teacher reports, but not on self-reports, and elevated scores were reported for Peer Problems by parents, teachers and self-ratings [2]. The standardised mean difference obtained here can be compared with the above meta–analysis which reported SMDs of 0.23 and 0.34 for the differences between children with PCHL and the HCG on parent- and teacher-rated SDQ in 12 studies of children with hearing loss and hearing children. The equivalent SMDs in the present study were 0.39 and 0.17.

rence obtained here can be compared with the above meta–analysis which reported SMDs of 0.23 and 0.34 for the differences between children with PCHL and the HCG on parent- and teacher-rated SDQ in 12 studies of children with hearing loss and hearing children. The equivalent SMDs in the present study were 0.39 and 0.17. A feature of the results of this study is the association of the presence of other disabilities with an increase in the EBD scores in those with PCHL. A study of 140 adolescents with cochlear implants also reported a similarly large effect of additional disabilities on EBD [33]. In that study the adolescents with CI had significantly higher scores than a normal hearing comparison group on the peer problems SDQ sub-scale as rated by parents, teachers and on self-ratings. The risk of EBD was highest if the adolescent also had “risks” additional to cochlear implantation. These risks included general learning disorders, visual impairment and inner ear malformations. On parent and teacher ratings those with additional risks had significantly higher scores than the normal hearing group on both hyperactivity and conduct problems. Peer problems were reported more frequently on self-ratings by adolescents with CI than by those with normal hearing both with and without additional risks.

On parent and teacher ratings those with additional risks had significantly higher scores than the normal hearing group on both hyperactivity and conduct problems. Peer problems were reported more frequently on self-ratings by adolescents with CI than by those with normal hearing both with and without additional risks. In the present study, as when the participants were examined in childhood [10], the presence of poor receptive language ability was a key risk for a high EBD score in those with PCHL. Indeed when an adjustment was made for the effect of receptive language ability, the PCHL group no longer had significantly higher SDQ scores compared to HCG. We conclude that the effects of poor receptive language ability accounted for the PCHL and HCG differences in EBD rather that hearing loss per se. The results also suggest that poor receptive language ability account at least in part for the high EBD score in those with other disabilities.

r SDQ scores compared to HCG. We conclude that the effects of poor receptive language ability accounted for the PCHL and HCG differences in EBD rather that hearing loss per se. The results also suggest that poor receptive language ability account at least in part for the high EBD score in those with other disabilities. The finding of a benefit of UNHS on behaviour in adolescence contrasts with the absence of benefit reported in a recent study by Wake et al. [34]. That study investigated a range of cognitive and behavioural outcomes at ages 5–7 years in three Australian populations with contrasting approaches to the detection of bilateral congenital hearing loss. They used the same measure of EBD as that adopted here and the normative value of parent-rated Total Difficulties used in their analysis (6.9) was close to that found for the HCG (6.2). However, the mean Total Difficulties score for those with UNHS was higher (9.6) than that obtained for those with UNHS in the present study (7.0). One feature that differentiates the two studies is age at follow-up. The Australian sample was assessed at 5–7 years of age, while in the present study the participants were 13–20 years old. To test whether this age difference might account for the difference found for the effect of UNHS on behaviour the EBD scores for the children in the present study at age 5 0.5–11.5 years were examined. At that age, the effect of UNHS on SDQ parent-rated total difficulties scores was SMD = 0.34 (95%CI −0.12 to 0.80) and not significant. This is lower than the significant value of SMD = 0.48 (95%CI 0.01 to 0.95) obtained during the teenage years. This raises the possibility that the effect of UNHS on behaviour may become more marked with age.

of UNHS on SDQ parent-rated total difficulties scores was SMD = 0.34 (95%CI −0.12 to 0.80) and not significant. This is lower than the significant value of SMD = 0.48 (95%CI 0.01 to 0.95) obtained during the teenage years. This raises the possibility that the effect of UNHS on behaviour may become more marked with age. However, the finding that the effect of UNHS on EBD in adolescence appears greater than that of early confirmation of PCHL is unexpected. We would predict that any effects of UNHS exposure on behaviour would be mediated via the benefits to language that result from early confirmation. An alternative explanation suggested by the absence of a significant benefit of early confirmation on behaviour in this sample may be that the benefit associated with UNHS that we observed was a sample-specific chance effect. Examination of the effects of UNHS and early confirmation on EBD outcomes when the Wake et al. cohort reaches adolescence would provide additional valuable insights into this issue.

onfirmation on behaviour in this sample may be that the benefit associated with UNHS that we observed was a sample-specific chance effect. Examination of the effects of UNHS and early confirmation on EBD outcomes when the Wake et al. cohort reaches adolescence would provide additional valuable insights into this issue. The study reported here had a number of limitations. First, the annual attrition rate of 4% per annum over the 9 years since assessment at primary school among children with PCHL eligible for the present study limited its power to detect potentially clinical important effects. This attrition rate is nevertheless low for a longitudinal study of a teenage population with a long-term medical condition [35]. Moreover the EBD scores at age 8 years in those whose parents and teachers provided ratings in the present study were similar to those seen at 8 years in those that provided ratings only on in the assessment at primary school, indicating that there was not selective attrition related to the outcome measures of interest. Second, given the finding that the presence of other disabilities was strongly related to EBD and that PCHL teenagers in this group were less likely to be in mainstream schools (19% vs. 77%), it was not possible to test for the suggested relationship between type of school attended and EBD [1] as this relationship would have been confounded by the presence of other disabilities. Third, the teenagers with PCHL were slightly older than the HCG at assessment. However age showed no significant relationship with SDQ scores.

Second, given the finding that the presence of other disabilities was strongly related to EBD and that PCHL teenagers in this group were less likely to be in mainstream schools (19% vs. 77%), it was not possible to test for the suggested relationship between type of school attended and EBD [1] as this relationship would have been confounded by the presence of other disabilities. Third, the teenagers with PCHL were slightly older than the HCG at assessment. However age showed no significant relationship with SDQ scores. Fourth, a difficulty in interpreting the results stems from the low SDQ scores on parent and teacher ratings for the HCG compared to national norms [30]. Teenagers with PCHL had parent-rated SDQ scores that were significantly higher than the HCG. However their mean score was not significantly higher than the national norms on this measure. The comparison of the PCHL group with the HCG has the advantage that the groups are matched for the date and place of birth and the context in which the SDQ scores were obtained, but from these two comparisons there is ambiguity concerning the extent to which adolescents with PCHL have elevated parent-rated SDQ scores.

The comparison of the PCHL group with the HCG has the advantage that the groups are matched for the date and place of birth and the context in which the SDQ scores were obtained, but from these two comparisons there is ambiguity concerning the extent to which adolescents with PCHL have elevated parent-rated SDQ scores. The study also had some strengths. The sample of children with PCHL was drawn from a geographically defined population base. The participants that were born in periods with UNHS were closely comparable with those born in periods without UNHS with respect to place and date of birth and audiological service provision other than UNHS. The participants were studied longitudinally from birth up to their late adolescent years. The HCG was identified from children born in the same hospitals as the PCHL and of similar age at assessment. Interviewers who undertook the assessments were blind to the UNHS status of the participants.

rovision other than UNHS. The participants were studied longitudinally from birth up to their late adolescent years. The HCG was identified from children born in the same hospitals as the PCHL and of similar age at assessment. Interviewers who undertook the assessments were blind to the UNHS status of the participants. 5 Conclusions The present study is consistent with the conclusion that children and adolescents with PCHL are at increased risk of emotional and behavioural difficulties, as previously suggested by a meta-analysis. It also extends that conclusion by identifying the factors predisposing children with PCHL to EBD and by showing their developmental trajectory over time. More specifically, it suggests that in addition to PCHL, the presence of other disabilities and poor receptive language abilities create this vulnerability to EBD. In addition, those with EBD previously identified nine years earlier, are particularly likely to show EBD in adolescence. The continuity of EBD over time in the PCHL sample indicates that the long-term mental health of this group of children may particularly benefit from interventions in middle childhood and that such interventions should focus on those with poor receptive language. However, it is important also to recognise that, like hearing adolescents, the majority of adolescents with PCHL will not show clinically significant emotional and behaviour difficulties. Trial registration ISRCTN03307358.

5 Conclusions The present study is consistent with the conclusion that children and adolescents with PCHL are at increased risk of emotional and behavioural difficulties, as previously suggested by a meta-analysis. It also extends that conclusion by identifying the factors predisposing children with PCHL to EBD and by showing their developmental trajectory over time. More specifically, it suggests that in addition to PCHL, the presence of other disabilities and poor receptive language abilities create this vulnerability to EBD. In addition, those with EBD previously identified nine years earlier, are particularly likely to show EBD in adolescence. The continuity of EBD over time in the PCHL sample indicates that the long-term mental health of this group of children may particularly benefit from interventions in middle childhood and that such interventions should focus on those with poor receptive language. However, it is important also to recognise that, like hearing adolescents, the majority of adolescents with PCHL will not show clinically significant emotional and behaviour difficulties. Trial registration ISRCTN03307358. Acknowledgements We thank members of the Hearing Outcomes in Teenagers steering group for their advice; Hazel Blythe, Merle Mahon, Janet Peacock, Steve Powers, Brian Yuen. Funding This work was supported by The Wellcome Trust [Grant Number 089251/Z/09/Z]. Conflict of interest On behalf of all authors, the corresponding author states that there are no conflicts of interest.

1 Introduction Acute and chronic respiratory diseases are highly frequent and associated with excessive morbidity and mortality, especially in children, thus having great impact on public health [1], [2]. Respiratory viruses play major roles as etiologic agents of acute respiratory infections (ARI) in children, but their association with chronic respiratory diseases, especially with adenotonsillar hypertrophy, has only recently become the focus of investigation [3], [4], [5]. Adenotonsillar hypertrophy is the most common cause of sleep apnea in children, resulting in craniofacial growth changes and, in severe cases, leading to right ventricular dysfunction and cor pulmonale [6], [7], [8]. As a consequence, adenotonsillectomy is the most frequent surgical procedure performed by otorhinolaryngologists [9]. The aetiologies of chronic hypertrophic adenotonsillar diseases have not been properly established, but are believed to be multifactorial, including allergies, bacterial colonisation and viral infections [3], [10], [11]. We have previously reported high rates of detection of respiratory virus genomes in tonsils and adenoids from patients with chronic adenotonsillar diseases, suggesting a significant association of viruses, particularly picornaviruses, with severe tonsillar hypertrophy [3]. However, no conclusive evidence of productive – acute or persisting – viral infection, as opposed to virus latency, has been established.

and adenoids from patients with chronic adenotonsillar diseases, suggesting a significant association of viruses, particularly picornaviruses, with severe tonsillar hypertrophy [3]. However, no conclusive evidence of productive – acute or persisting – viral infection, as opposed to virus latency, has been established. In general, peaks of respiratory virus detection in children with ARI occur with marked seasonal variations in temperate and subtropical regions [12], [13]. In tropical areas, the seasonal pattern of viral detection is more difficult to be analysed, due to the heterogeneity of data in several parts of the world. However, respiratory viruses have been mainly observed during the rainy seasons [14]. In southeast Brazil, a region of transition between tropical and subtropical climates, peaks of viral ARI tend to occur during cooler months [15], [16], [17]. The present paper reports analyses of over time variations in rates of detection of respiratory viruses in tissues and secretions removed from children undergoing tonsillectomy while in the absence of ARI symptoms. The rationale was that if detection of respiratory viruses in hypertrophic tonsillar tissues oscillated with variations in temperature and rainfall, in a way similar to what occurs among ARI patients, this would suggest an association with acute subclinical respiratory viral infections, rather than prolonged asymptomatic harbouring of viral nucleic acids in the tissues.

iruses in hypertrophic tonsillar tissues oscillated with variations in temperature and rainfall, in a way similar to what occurs among ARI patients, this would suggest an association with acute subclinical respiratory viral infections, rather than prolonged asymptomatic harbouring of viral nucleic acids in the tissues. 2 Patients and methods 2.1 Patients and sampling Respiratory virus genomes were searched in adenoids (AD), palatine tonsils (PT) and nasopharyngeal secretions (NPS) obtained from all 172 children (91 males) aged 1–13 years (mean 5.8 years) who underwent adenotonsillectomy to treat adenotonsillar hypertrophy with clinical evidence of obstructive sleep apnoea syndrome [6] or recurrent adenotonsillitis according to Paradise criteria [18]. Patients were treated at the division of Otorhinolaryngology of the School of Medicine of Ribeirão Preto, University of São Paulo, between May 2010 and June 2012. Patients with signs/symptoms of acute respiratory infections within the last four weeks prior to surgery and patients with immunodeficiencies were excluded from the study. Indeed, the exclusion of these patients was a safety criterion for surgery. All clinical samples obtained in this study were maintained in a preservative solution (RNA later – Invitrogen, Carlsbad, CA, USA) at −70 °C until nucleic acid extraction.

rgery and patients with immunodeficiencies were excluded from the study. Indeed, the exclusion of these patients was a safety criterion for surgery. All clinical samples obtained in this study were maintained in a preservative solution (RNA later – Invitrogen, Carlsbad, CA, USA) at −70 °C until nucleic acid extraction. 2.2 Ethics statement The study was conducted according to the principles expressed in the Declaration of Helsinki and was approved by the University Hospital Clinical Research Ethics Committee (file number 10466/2008). A written informed consent was obtained from all parents and guardians prior to enrolment. 2.3 Detection of respiratory viruses Nucleic acids from AD, PT and NPS were obtained from 30 mg of tissue or 200 μL of secretion using the AllPrep DNA/RNA mini kit (Qiagen GmbH, Hilden, Germany) or QIAamp Min Elute Virus Spin Kit (Qiagen GmbH, Hilden, Germany), respectively. All samples were tested for the presence of genomes of human adenovirus (HAdV), human enterovirus (HEV), human rhinovirus (HRV), human bocavirus (HBoV), human respiratory syncytial virus (HRSV), metapneumovirus (HMPV), human influenza virus (FLU), human parainfluenza virus (HPIV) and human coronavirus (HCoV) by TaqMan real-time PCR.

re tested for the presence of genomes of human adenovirus (HAdV), human enterovirus (HEV), human rhinovirus (HRV), human bocavirus (HBoV), human respiratory syncytial virus (HRSV), metapneumovirus (HMPV), human influenza virus (FLU), human parainfluenza virus (HPIV) and human coronavirus (HCoV) by TaqMan real-time PCR. The detailed description of each PCR, including the primers sequences, can be obtained in a previously published paper [3]. In the present study were included 51 patients, extending the observation period to two years, allowing that the seasonal pattern of viral circulation was determined in these patients. The analysis of the seasonality of respiratory viruses in adenotonsillar tissues was performed cross matching the virus presence to the temperature and rainfall.

included 51 patients, extending the observation period to two years, allowing that the seasonal pattern of viral circulation was determined in these patients. The analysis of the seasonality of respiratory viruses in adenotonsillar tissues was performed cross matching the virus presence to the temperature and rainfall. 2.4 Analysis of climate variations Ribeirão Preto is a city in the state of São Paulo, southeast Brazil with a population of 619,746, located at 21°10′40″ S and 47°48′36″ W, 500 m above sea level. The climate is a transition between tropical and subtropical conditions, with annual average temperature of 23 °C, with dry mild winters and hot rainy summers. During this study, the mean monthly minimal and maximal daily temperatures were respectively 18.3 °C (range, 13.5–20.5) and 25.3 °C (range, 23.5–26.8). Rainfall is the major climate variable, with yearly rainy seasons between November and March and dry season from June to September. The mean monthly accumulated rainfall throughout the study was 108 mm, ranging from 0 to 533 mm. Accumulated rainfall and mean seasonal temperatures were obtained from the site of the Integrated Center of Agrometeorological Information of São Paulo Sate (http://www.ciiagro.sp.gov.br).

dry season from June to September. The mean monthly accumulated rainfall throughout the study was 108 mm, ranging from 0 to 533 mm. Accumulated rainfall and mean seasonal temperatures were obtained from the site of the Integrated Center of Agrometeorological Information of São Paulo Sate (http://www.ciiagro.sp.gov.br). 3 Results Rates of detection of respiratory viruses in adenoids, tonsils and respiratory secretions were determined for 172 children. Genomes of at least one respiratory virus were detected in over 87% of the patients, without discernible seasonal variations (Fig. 1 ). Remarkably, high rates of virus detection were obtained from all three kinds of clinical samples throughout the study. The frequencies of virus detection ranged from 71.5% to 94.1% in adenoids, and from 37.5% to 86.6% in secretions and palatine tonsils. Of the three sample kinds, adenoid tissue yielded the highest frequencies of virus detection during almost the whole study, except for autumn months (March–June) of 2011, when nasal secretions yielded higher rates of positivity (Fig. 1).Fig. 1 Seasonal distribution of the rates of positivity for respiratory viruses in adenoids, palatine tonsils and respiratory secretions from patients with chronic adenotonsillar disease. Numbers of samples tested per season are informed in parenthesis.

when nasal secretions yielded higher rates of positivity (Fig. 1).Fig. 1 Seasonal distribution of the rates of positivity for respiratory viruses in adenoids, palatine tonsils and respiratory secretions from patients with chronic adenotonsillar disease. Numbers of samples tested per season are informed in parenthesis. The present analysis is based on results from a total of 172 patients, covering a 2-year period, and raises new issues made clear upon inclusion of additional cases to a study that was underway [3]. Overall, the most frequent viruses were human adenovirus (HAdV) detected in 52.8%, followed by human enterovirus (HEV) in 47.2%, human rhinovirus (HRV) in 33.8%, human bocavirus (HBoV) in 31.1%, human metapneumovirus (HMPV) in 18.3%, human respiratory syncytial virus (HRSV) in 17.2%, influenza virus (FLU) in 4.5%, human parainfluenzavirus (HPIV) in 4.5%, and human coronavirus (HCoV) in 2.7%. The frequencies of HAdV, HBoV and HRSV were higher in adenoids, whereas HRV was more frequently detected in nasal secretions and HEV in palatine tonsils. The rates of viral co-infections and the agreement between results from different tissues were high. In this 2-year study period, two or more viruses were detected in 62.2% of the patients, and 54% of them had the same virus detected in adenoids and palatine tonsils.

ted in nasal secretions and HEV in palatine tonsils. The rates of viral co-infections and the agreement between results from different tissues were high. In this 2-year study period, two or more viruses were detected in 62.2% of the patients, and 54% of them had the same virus detected in adenoids and palatine tonsils. Overall, HAdV detection rates fluctuated from summer troughs of approximately 37.5% up to peaks of greater than 70% in spring-2011 and autumn-2012, without clearly seasonal periods (Fig. 2 ). HBoV detection rates were usually above 20% (12.5–46%) without discernible seasonality (Fig. 3 ). Detection frequencies of HAdV and HBoV were consistently higher in adenoids than in other samples. Rates of detection of the picornaviruses HRV and HEV were opposite in the summer, while the rate of HEV detection was at its peak, HRV was at its lowest (Fig. 4, Fig. 5 ). The overall HEV detection rates were composed mostly by results obtained from adenoids and palatine tonsils, while detection in respiratory secretions was found in a smaller proportion of the HEV-positive patients (Fig. 4). Although peaks of HEV detection occurred in summer/autumn months, the rates of HEV positivity were always above 30%, indicating that a great proportion of tonsil tissues harbour HEV, independently of the season of the year. Contrary to HEV, HRV overall rates showed no clear seasonal variations, and were mostly composed by results obtained from secretions, with correspondingly lower rates of positivity in tonsillar tissues (Fig. 5).Fig. 2 Seasonal distribution of the rates of detection of human adenovirus (HAdV) in adenoids, palatine tonsils and respiratory secretions from patients with chronic adenotonsillar disease.

mostly composed by results obtained from secretions, with correspondingly lower rates of positivity in tonsillar tissues (Fig. 5).Fig. 2 Seasonal distribution of the rates of detection of human adenovirus (HAdV) in adenoids, palatine tonsils and respiratory secretions from patients with chronic adenotonsillar disease. Fig. 3 Seasonal distribution of the rates of detection of human bocavirus (HBoV) in adenoids, palatine tonsils and respiratory secretions from patients with chronic adenotonsillar disease. Fig. 4 Seasonal distribution of rates of detection of human enterovirus (HEV) in adenoids, palatine tonsils and respiratory secretions from patients with chronic adenotonsillar disease. Fig. 5 Seasonal distribution of rates of detection of human rhinovirus (HRV) in adenoids, palatine tonsils and respiratory secretions from patients with chronic adenotonsillar disease. Rates of detection of the paramyxoviruses HMPV and HRSV varied from 0% to 31% during the study period (Fig. 6, Fig. 7 ). HRSV detection reached maximum level in spring/winter months, mostly composed by detection in respiratory secretions, with summer nadirs in both years of the study (Fig. 6). Differently, HMPV overall detection rates did not vary significantly during the study period (Fig. 7), and the nadir observed in the summer of 2012 probably reflects the low number of samples analysed in that season.Fig. 6 Seasonal distribution of rates of detection of human respiratory syncytial virus (HRSV) in adenoids, palatine tonsils and respiratory secretions from patients with chronic adenotonsillar disease.

7), and the nadir observed in the summer of 2012 probably reflects the low number of samples analysed in that season.Fig. 6 Seasonal distribution of rates of detection of human respiratory syncytial virus (HRSV) in adenoids, palatine tonsils and respiratory secretions from patients with chronic adenotonsillar disease. Fig. 7 Seasonal distribution of rates of detection of human metapneumovirus (HMPV) in adenoids, palatine tonsils and respiratory secretions from patients with chronic adenotonsillar disease. FLU, HPIV and HCoV were detected in very low frequencies, with sporadic cases distributed during the 2-year study period without seasonal pattern. 4 Discussion Several studies have shown that some respiratory viruses circulate seasonally, with a typically increase of the viral incidence in colder months, mainly in temperate regions [12], but also in subtropical regions [15], [16], [17]. In tropical regions the results are more difficult to interpret, with several studies indicating higher viral circulation in rainy seasons [14], [19], [20], [21], while others show that respiratory viruses are prevalent year-round [22]. In Salvador for instance, a tropical city in the northeast of Brazil, the presence of viral infections was significantly associated with precipitation during the rainy season in patients with community-acquired pneumonia [19].

, [19], [20], [21], while others show that respiratory viruses are prevalent year-round [22]. In Salvador for instance, a tropical city in the northeast of Brazil, the presence of viral infections was significantly associated with precipitation during the rainy season in patients with community-acquired pneumonia [19]. It is broadly accepted that respiratory viruses spread by shedding in secretions from acutely symptomatic patients [12], but respiratory viruses are also frequently detected in asymptomatic individuals [23], [24], [25], raising the hypothesis that the viral shedding by people without acute symptoms can be important to the viral dissemination. In fact, high frequencies of detection of respiratory viruses have also been observed in secretions and tissues from patients with chronic adenotonsillar diseases [3], [5], [26], [27], but no analysis of seasonality had been done in that particular setting. Although the presently reported analyses confirmed previously published findings that found high rates of viral detection in patients with chronic adenotonsillar diseases, the analysis of the fluctuations in viral detection rate during these 2 years showed that most of respiratory viruses have no obvious seasonal pattern, supporting the notion that such high frequency of virus genome detection can be related to virus persistence in lymphoepithelial tissues of the upper respiratory tract.

the analysis of the fluctuations in viral detection rate during these 2 years showed that most of respiratory viruses have no obvious seasonal pattern, supporting the notion that such high frequency of virus genome detection can be related to virus persistence in lymphoepithelial tissues of the upper respiratory tract. The discovery of HAdV was consequence of its recovery from adenoid explants [28], and several studies have documented that several adenovirus species, especially adenovirus C, can persist in mucosal lymphoid tissues, possibly by maintenance of the quiescent viral genomes in non-dividing lymphocytes [29], [30]. In fact, HAdV causes persistent/latent infection in tonsillar T-cell subpopulations [29], and can infect continuous B-cell and myeloid cell lineages in vitro [30], suggesting that several different cell populations in tonsils may carry the virus genome. Therefore, it was no surprise that in the present study HAdV was the most frequent respiratory virus detected in adenoids, and the second most frequent in palatine tonsils. In addition, the lack of seasonal trends in rates of HAdV detection, both in tissues and secretions, is confirmatory that most of the high HAdV frequency is attributable to persistence. In tropical regions, adenovirus is frequently associated with the rainy season in patients with acute respiratory infections [19].

ls. In addition, the lack of seasonal trends in rates of HAdV detection, both in tissues and secretions, is confirmatory that most of the high HAdV frequency is attributable to persistence. In tropical regions, adenovirus is frequently associated with the rainy season in patients with acute respiratory infections [19]. HBoV is a parvovirus that occurs worldwide in association with respiratory and gastrointestinal disorders [31], [32]. In addition to the general propensity of parvoviruses to persist and even endogenise into host genomes [33], at least two lines of evidence support persistence of HBoV in humans. Only around 25% of patients who are PCR-positive for HBoV have mRNA for a viral structural protein detectable as a marker of active viral replication [34] and HBoV episomes have been found in human clinical samples, including tissue biopsies [35], [36], [37]. Therefore, it is not surprising that in the present study HBoV was frequently detected, without discernible seasonality, suggesting that, at least part of this high frequency could be attributed to virus persistence, mainly in adenoids.

e been found in human clinical samples, including tissue biopsies [35], [36], [37]. Therefore, it is not surprising that in the present study HBoV was frequently detected, without discernible seasonality, suggesting that, at least part of this high frequency could be attributed to virus persistence, mainly in adenoids. While persistence and latency of DNA viruses in lymphoepithelial tissues have long been known, the same is not the case for RNA viruses. In the present study, HEV was the second most frequently detected agent at overall frequencies consistently above 30%. Although there are very few studies about seasonality of HEV in tropical regions, the observed trend for an increase in HEV detection towards the summer is in keeping with what happens in temperate regions of the world [38]. However, the high rates of HEV positivity were consistently due to detection in tissue fragments, which was less frequently accompanied by shedding in secretions, supporting the idea that HEV can persist in adenoids and tonsils in a high proportion of patients. While confirmation of HEV persistence in tonsillar tissues will require detailed molecular investigation, the present findings of such consistently and non-seasonal HEV detection over time, coupled with the already reported higher frequencies of HEV detection in the most highly hypertrophic tonsils [3] indicates that perhaps HEV persistence can be associated with the pathogenesis of chronic adenotonsillar diseases.

gation, the present findings of such consistently and non-seasonal HEV detection over time, coupled with the already reported higher frequencies of HEV detection in the most highly hypertrophic tonsils [3] indicates that perhaps HEV persistence can be associated with the pathogenesis of chronic adenotonsillar diseases. HRV, the other picornavirus included in the present analyses, is frequently detected in acute respiratory infections of children and adults, usually with marked seasonal variation, especially in subtropical and temperate areas [39], [40], as well as in asymptomatic patients [24]. In tropical regions, the literature results are controversial. In Trinidad, West Indies, HRV was prevalent throughout the year, without seasonal association [41], whilst in Salvador HRV was associated with relative humidity (p = 0.05) [19]. In the present analysis HRV was frequently detected in all seasons, mostly in secretions, rather than in tissues. It is interesting that HRV and HEV are both picornaviruses, currently classified in the same genus, with very similar replication cycles, and yet showed dissimilar frequencies of detection in tissues. HRV was detected in lower frequencies in lymphoid tissues as compared to HEV, suggesting the existence of other sites of infection as sources of HRV shedding into secretions, such as nasal epithelium. This is remarkable, since the patients with hypertrophic adenotonsillar diseases had no acute nasal symptoms at the time of surgery.

ted in lower frequencies in lymphoid tissues as compared to HEV, suggesting the existence of other sites of infection as sources of HRV shedding into secretions, such as nasal epithelium. This is remarkable, since the patients with hypertrophic adenotonsillar diseases had no acute nasal symptoms at the time of surgery. Although the overall rates of detection of HMPV and HRSV in the present analyses were lower than those of other viruses, they were still frequently higher than 20%. Moreover, HRSV rates were frequently higher than 10%, with significant contribution of positivity in adenotonsillar tissues, again suggesting that these tissues may be regarded as sites of persistence of paramyxoviruses. Pertaining to this issue, it is interesting that phylogenetic studies of respiratory syncytial virus have pointed to the existence of reservoirs to maintain the virus during inter-seasonal periods, thus creating potential for its reintroduction in the susceptible population [42]. It is therefore reasonable to think that adenoids and tonsils of children with chronic adenotonsillar diseases would be natural reservoirs of respiratory viruses and that, by shedding viruses in respiratory secretions, these children would become sources of infection for their siblings and schoolmates.

population [42]. It is therefore reasonable to think that adenoids and tonsils of children with chronic adenotonsillar diseases would be natural reservoirs of respiratory viruses and that, by shedding viruses in respiratory secretions, these children would become sources of infection for their siblings and schoolmates. To the best of our knowledge, this has been the most comprehensive study so far conducted on the variations of respiratory virus detection rates over time in children with chronic adenotonsillar diseases in a subtropical/tropical region. Although further studies are in order to clarify whether these findings result from long term virus shedding consequent to persistence, or to current asymptomatic viral infections, the lack of obvious seasonal patterns of respiratory viruses in hypertrophic adenotonsillar tissues supports the view that some respiratory viruses may persist in adenoids and tonsils. In the absence of results from molecular markers of active viral replication, the mere detection of viral genomes is not enough to establish that such infections were productive. However, the existence of genomes of numerous viruses in adenotonsillar tissues is an exciting finding, and deserves further investigation. In addition to its obvious epidemiological importance as possible sources of community respiratory virus outbreaks, persistence of these viruses could have pathogenic potential in the development of tonsillar hypertrophy, functioning as chronic stimuli for inflammation. Alternatively, for reasons not yet understood, respiratory viruses could more readily establish an asymptomatic carrier states in patients with chronic tonsillar hypertrophy. However, the lack of control group, with samples of tonsils obtained from healthy patients, makes any inference about the development of disease difficult to be explored in the present study. Thereby, studies about viral persistence in adenoids and tonsils, mainly those including tissue samples from healthy subjects, can bring new insights to the understanding of poorly understood chronic tonsillar diseases that affect large numbers of children worldwide.

sease difficult to be explored in the present study. Thereby, studies about viral persistence in adenoids and tonsils, mainly those including tissue samples from healthy subjects, can bring new insights to the understanding of poorly understood chronic tonsillar diseases that affect large numbers of children worldwide. Conflict of interest The authors have no conflicts of interest to disclose and have no financial disclosures to make. Acknowledgments The authors thank Maria Cecília Onofre and Helder G. de Souza for secretarial assistance; Lúcia Lopes, Jamila Mendonça de Souza and Maria Lúcia Silva for expert technical support. In addition, the authors thank FAPESP (Fundação de Amparo à Pesquisa do Estado de São Paulo) for financial support (Grant number 2009/51818-8).

1 Introduction Croup develops in more than 80,000 Canadian children annually and accounts for 5% of emergency hospital admissions in children under 6 years of age [1]. In its current use, a diagnosis of croup encompasses a number of respiratory illnesses characterized by varying degrees of inspiratory stridor, cough, and hoarseness resulting from inflammation and narrowing of the larynx. Viral croup (acute laryngotracheitits) refers to the typical croup syndrome that occurs in children at the peak ages of six months to three years [2]. It is characterized by hoarseness, stridor, and a barky cough that occurs after a viral prodrome of low-grade fever and coryza lasting 12–72 h. Symptoms tend to worsen at night, and when the child is agitated or crying.

he typical croup syndrome that occurs in children at the peak ages of six months to three years [2]. It is characterized by hoarseness, stridor, and a barky cough that occurs after a viral prodrome of low-grade fever and coryza lasting 12–72 h. Symptoms tend to worsen at night, and when the child is agitated or crying. Viral croup is most commonly caused by parainfluenza virus type 1, as well as parainfluenza type 2 and 3, rhinovirus, coronavirus, adenovirus, and respiratory syncytial virus (RSV) [3,4]. This results in erythema and swelling of the lateral walls of the trachea and larynx, leading to epithelial necrosis and laryngeal and subglottic narrowing secondary to inflammation [3,5]. The disease process is self-limited with resolution typically within one week [6,7]. In mild cases, a single dose of oral dexamethasone may be administered. In moderate to severe cases, dexamethasone and nebulized epinephrine are the mainstay of therapy. In cases of failure of medical treatment, progression to exhaustion from increased work of breathing, hypercapnic or hypoxic respiratory failure, or imminent airway obstruction, treatment with endotracheal intubation is warranted. The incidence of intubation is approximately 3% for all patients admitted for croup [[8], [9], [10]].

failure of medical treatment, progression to exhaustion from increased work of breathing, hypercapnic or hypoxic respiratory failure, or imminent airway obstruction, treatment with endotracheal intubation is warranted. The incidence of intubation is approximately 3% for all patients admitted for croup [[8], [9], [10]]. In the medical literature, there have been attempts to sub-classify and define croup based on infectious etiology, clinical recurrence or severity. Patients with uncharacteristic presentations or abnormal natural history are commonly diagnosed with atypical croup. The use of this term leads to several difficulties in the management of these patients in the acute care setting. Given there is no definitive diagnostic or management pathway for atypical croup, patients are often subject to radiographic studies and operative assessments that require a general anesthetic. The primary objective of this case report and systematic review is to evaluate the published definitions of atypical croup. The secondary objectives are to summarize its diagnosis, etiologies, and management, and identify the patient characteristics that may result in intubation or tracheostomy and propose a management pathway.

ry objective of this case report and systematic review is to evaluate the published definitions of atypical croup. The secondary objectives are to summarize its diagnosis, etiologies, and management, and identify the patient characteristics that may result in intubation or tracheostomy and propose a management pathway. 2 Case report 2.1 Initial presentation A previously healthy 3-year-old male manifested a low-grade fever, progressive stridor, a barky cough, tracheal tugging and suprasternal in-drawing 7 days following an upper respiratory tract infection. Despite a 3-day course of amoxicillin and prednisone, his work of breathing and stridor increased, requiring nebulized epinephrine administered at a regional emergency department. He was subsequently transferred to the Alberta Children's Hospital (ACH) pediatric intensive care unit (PICU) due to airway concerns. Although he remained non-toxic and playful, his biphasic stridor only improved briefly with medical therapy. As a result, a lateral neck x-ray and computed tomography (CT) of the neck were performed on day 2 of admission to rule out a subglottic lesion. 2.2 Clinical diagnosis CT imaging demonstrated diffuse edema and narrowing of the glottis, subglottis, and upper trachea in keeping with croup. Operative laryngoscopy and bronchoscopy performed 2 days after admission revealed an inflamed subglottis approximately 4–5 mm in diameter (Cotton Myers System Grade III). No other abnormalities were found.

ging demonstrated diffuse edema and narrowing of the glottis, subglottis, and upper trachea in keeping with croup. Operative laryngoscopy and bronchoscopy performed 2 days after admission revealed an inflamed subglottis approximately 4–5 mm in diameter (Cotton Myers System Grade III). No other abnormalities were found. Despite repeated nebulized epinephrine and intravenous dexamethasone administrations, the patient's stridor remained biphasic with little to no improvement. On post-admission day 7, the patient's respiratory status deteriorated such that he required endotracheal intubation. This was performed in the OR allowing for repeat laryngoscopy and rigid bronchoscopy, which revealed a narrower subglottic airway than was seen during previous bronchoscopy (still Cotton Myers Grade III). Due to the degree of narrowing, a 3.0 endotracheal tube was placed. A mucosal biopsy performed intraoperatively showed respiratory epithelium with prominent lymphoid aggregates and abundant reticulin fibres in the subepithelial stroma. These lymphocytes were diffusely positive for CD21 and B-cell markers, CD45, and CD3. There was no significant cytological atypia. Tracheal swabs were negative for bacteria, fungus, and EBV. Nasopharyngeal swabs were positive for rhinovirus RNA only. IgM testing for mycoplasma pneumoniae was negative.

pithelial stroma. These lymphocytes were diffusely positive for CD21 and B-cell markers, CD45, and CD3. There was no significant cytological atypia. Tracheal swabs were negative for bacteria, fungus, and EBV. Nasopharyngeal swabs were positive for rhinovirus RNA only. IgM testing for mycoplasma pneumoniae was negative. Because the patient's subglottic narrowing was felt to be firm intraoperatively, magnetic resonance (MR) imaging of his neck was ordered post-operatively. This showed a narrow 3 mm subglottis without T2 hyperintensity. A 7 mm × 5 mm round, well-circumscribed focus of non-enhancement posterior to the upper trachea was appreciated also, suggestive of a proteinaceous cyst. This lesion was not seen on later endoscopies and may have been an artefact of biopsy. No additional abnormalities were seen. Due to the unusual nature of this presentation, rheumatologic work up was performed. This showed normal serum anti-nuclear antibody, glomerular basement membrane antibody, anti-neutrophil cytoplastic antibody, anti-MPO and proteinase 3 antibody (PR3), c-reactive protein, ferritin, lactate dehydrogenase, and quantitative rheumatoid factor levels. Similarly, immunologic investigations showed normal immunoglobulin E and IgE aspergillus antibody levels. 2.3 Management Despite 4 days of scheduled systemic steroids in the PICU, an endotracheal cuff leak was not achieved; therefore a tracheostomy was performed in anticipation of prolonged intubation 14 days after presentation to the emergency department and 21 days after the start of his symptoms.

Due to the unusual nature of this presentation, rheumatologic work up was performed. This showed normal serum anti-nuclear antibody, glomerular basement membrane antibody, anti-neutrophil cytoplastic antibody, anti-MPO and proteinase 3 antibody (PR3), c-reactive protein, ferritin, lactate dehydrogenase, and quantitative rheumatoid factor levels. Similarly, immunologic investigations showed normal immunoglobulin E and IgE aspergillus antibody levels. 2.3 Management Despite 4 days of scheduled systemic steroids in the PICU, an endotracheal cuff leak was not achieved; therefore a tracheostomy was performed in anticipation of prolonged intubation 14 days after presentation to the emergency department and 21 days after the start of his symptoms. Six days following operative tracheostomy, a repeat operative bronchoscopy showed a 5 mm midline lesion on the posterior tracheal wall that was 3 cm distal to his glottis, superior to the distal tip of the tracheostomy tube. Excisional biopsy showed chronic lymphocytic inflammation, fibrin exudate, and early granulation consistent with an inflammatory pseudotumour. Following one month of watchful waiting, suspension laryngoscopy and rigid bronchoscopy demonstrated significant improvement in subglottic patency. However, a tracheostomy was maintained due to continued narrowing of his airway. On follow-up bronchoscopy, a suprastomal granuloma was seen and obliterated with a YAG laser. Repeat rigid bronchoscopy performed 3 months later showed resolution of the posterior tracheal wall mass and subglottic stenosis. He was decannulated intraoperatively and admitted to the PICU for overnight observation. He was discharged the following day with no further concerns, 219 days after his initial presentation. These findings most likely resulted from gross inflammation secondary to RSV given the absence of any other anatomical abnormality on radiographic and operative assessment.

itted to the PICU for overnight observation. He was discharged the following day with no further concerns, 219 days after his initial presentation. These findings most likely resulted from gross inflammation secondary to RSV given the absence of any other anatomical abnormality on radiographic and operative assessment. 3 Systematic review We aimed to identify all full-text, peer-reviewed publications pertaining to atypical croup. The Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) reporting guideline was adapted for the current review. Published studies pertaining to atypical croup were found utilizing the Ovid MEDLINE® and EMBASE databases from inception to January 1st, 2019 (Fig. 1 ). All peer-reviewed studies in children less than 18 years of age and studies in the English language were included. The following search terms were used: croup, laryngitis, laryngotracheitis, and laryngotracheobronchitis. Results were combined with the term atypical to retrieve the articles. Articles were screened by two authors and the reference lists of chosen articles were searched to further identify relevant articles. Articles with a focus on atypical presentations of croup were included. Reviews, commentaries, and editorials were excluded. The information extracted consisted of author, year, level of evidence, study design, demographics of the patient sample, including age and symptom duration, diagnostic data, etiology, and medical treatment.Fig. 1 The search strategy and flow diagram as per PRISMA guidelines. Fig. 1

3 Systematic review We aimed to identify all full-text, peer-reviewed publications pertaining to atypical croup. The Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) reporting guideline was adapted for the current review. Published studies pertaining to atypical croup were found utilizing the Ovid MEDLINE® and EMBASE databases from inception to January 1st, 2019 (Fig. 1 ). All peer-reviewed studies in children less than 18 years of age and studies in the English language were included. The following search terms were used: croup, laryngitis, laryngotracheitis, and laryngotracheobronchitis. Results were combined with the term atypical to retrieve the articles. Articles were screened by two authors and the reference lists of chosen articles were searched to further identify relevant articles. Articles with a focus on atypical presentations of croup were included. Reviews, commentaries, and editorials were excluded. The information extracted consisted of author, year, level of evidence, study design, demographics of the patient sample, including age and symptom duration, diagnostic data, etiology, and medical treatment.Fig. 1 The search strategy and flow diagram as per PRISMA guidelines. Fig. 1 4 Results Two authors examined 41 articles and identified 12 that met the inclusion criteria (Table 1 ). The search strategy and flow diagram are presented using the PRISMA guidelines. Results focused on identifying definitions of atypical croup in the literature. The etiology, incidence, prevalence, diagnosis and management of atypical croup were secondary objectives.Table 1 Articles included in systematic review with a summary of the results.

flow diagram are presented using the PRISMA guidelines. Results focused on identifying definitions of atypical croup in the literature. The etiology, incidence, prevalence, diagnosis and management of atypical croup were secondary objectives.Table 1 Articles included in systematic review with a summary of the results. Table 1Article Level of Evidence Study Type Demographics Number of Patients Mean Age Average symptom duration Microbiology Investigations Treatment Definition of Atypical Croup Chauhan et al. (2007) 4 Case Series n = 2 Mean age 15 months Unknown symptom duration Herpes simplex virus 1 (n = 2) Laryngoscopy Systemic dexamethasone Nebulized epinephrine IV acyclovir Intubation Presentation < 6 months of age, or beyond toddler stage OR Unresponsive to supportive measures OR Recurrence or progression to fulminant course Cooper et al. (2012) 4 Case series with chart review. n = 80 Mean age 4.8 years Unknown symptom duration Unknown 80 children underwent laryngoscopy Unknown Unknown Croup episodes in a child 6 months old or younger, older than 3 years, or with recurrent episodes of croup (4 or more episodes in total) Hatherill et al. (2001) 4 Retrospective chart review n = 263 Mean age unknown (median age 14 months) Unknown symptom duration Herpes simplex virus (n = 3) Cytomegalovirus (n = 1) Haemophilus influenzae (n = 5) Staphylococcus aureus (n = 2) Streptococcus pneumoniae (n = 1) Providentia rettgeri (n = 1) mixed commensals (n = 6) 147 children underwent microlaryngos-copy 15 children received IV acyclovir 9 children underwent nasotracheal intubation 5 children required tracheostomy Increased severity or duration Inglis Jr. (1993) 4 Case series n = 2 Mean age 16 months Mean symptom duration 24 days Herpes simplex virus 1 (n = 2) Laryngoscopy Prednisolone Acyclovir sodium Epinephrine Amoxicillin/clavul-anate potassium Nafcillin sodium Dexamethasone sodium phosphate Intubation (n = 1) Prolonged course (>7 days) Harris et al. (1987) 4 Case series n = 1 Mean age 18 months Symptom Unknown symptom duration Herpes simplex virus (n = 1) Rigid bronchoscopy Ampicillin Methylprednisone Epinephrine Oral dexamethasone Prolonged course O'Niel et al.

illin sodium Dexamethasone sodium phosphate Intubation (n = 1) Prolonged course (>7 days) Harris et al. (1987) 4 Case series n = 1 Mean age 18 months Symptom Unknown symptom duration Herpes simplex virus (n = 1) Rigid bronchoscopy Ampicillin Methylprednisone Epinephrine Oral dexamethasone Prolonged course O'Niel et al. (2013) 4 Case series n = 2 Mean age 15 months Mean symptom duration 28 days Herpes simplex virus 1 (n = 1) Direct laryngoscopy with rigid bronchoscopy Heliox Nebulized epinephrine MethylprednisoloneIV acyclovir Unasyn Valacyclovir Augmentin Intubation Atypical if lasting more than 7 days or does not respond to appropriate treatments Low et al. (2012) 4 Case report n = 1 Mean age 8 years old Unknown symptom duration Curvularia species (n = 1) Laryngoscopy Bronchoscopy IV amphotericin B Oral voriconazole IV caspofungin Croup cause by an atypical pathogen Krause et al. (1997) 4 Case report n = 2 Mean age 12 months Mean symptom duration 3.5 weeks Herpes simplex virus 1 (n = 2) Fibre optic laryngotracheo-bronchoscopy IV ceftriaxone IV dexamethasone nebulized epinephrine Oral betamethasone Prolonged course Miller et al. (1982) 4 Case report n = 1 Mean Age 3.5 years Chlamydia trachomatis Straphylococcus aureus Rigid bronchoscopy IV Nafcillin Erythromycin Intubation *Describes bacterial tracheitis as an atypical croup like syndrome, consisting of a more severe and prolonged presentation, affecting older children, often requiring an artificial airway, and purulent subglottic exudate associated with culture of Staphylococcus aureus from tracheal secretions Barnes et al. (1999) 4 Case report n = 1 Mean age 15 months Unknown Direct laryngoscopy and bronchoscopy Nebulized albuterol Nebulized epinephrine Drainage and marsupialization of two large subglottic cysts with CO2 laser Prolonged symptoms Waki et al. (1995) 4 Retrospective chart review n = 262 (31 with recurrent croup) Mean age 9 months Unknown Endoscopy Barium esophagram Metoclopramide Ranitidine Albuterol 2 or more episodes necessitating inpatient care Farmer and Wohl (2001) 4 Retrospective chart review n = 53 Median age of 2 years Unknown Endoscopy Unknown Recurrent croup

etrospective chart review n = 262 (31 with recurrent croup) Mean age 9 months Unknown Endoscopy Barium esophagram Metoclopramide Ranitidine Albuterol 2 or more episodes necessitating inpatient care Farmer and Wohl (2001) 4 Retrospective chart review n = 53 Median age of 2 years Unknown Endoscopy Unknown Recurrent croup 4.1 Definitions of atypical croup Definitions of atypical croup varied in the examined studies. Several articles included a recurrent course within the definition. For instance, Cooper et al. (2012) defined croup as atypical if a child had more than 4 recurrent episodes, however Waki et al. (1995) suggested 2 or more episodes necessitating inpatient care [11,12]. Several studies also noted an atypical age of presentation as part of their definition. Chauhan et al. (2007) defined an age of presentation less than 6 months as atypical while Cooper et al. (2012) added an upper limit of greater than 3 years of age [11,13].

ted 2 or more episodes necessitating inpatient care [11,12]. Several studies also noted an atypical age of presentation as part of their definition. Chauhan et al. (2007) defined an age of presentation less than 6 months as atypical while Cooper et al. (2012) added an upper limit of greater than 3 years of age [11,13]. Several authors included the severity of presentation and unresponsiveness to medical therapy as defining features. Farmer and Wohl (2001) defined severity based on the need for any inpatient admission while Hatherill et al. (2001) defined severity based on the need for PICU admission due to severe subglottic stenosis [14,15]. Furthermore, prolonged duration of symptoms was a common criterion listed in definitions of croup in the literature search. Inglis Jr. (1993) and O'Niel et al. (2013) defined croup as atypical if symptoms lasted longer than seven days, although other studies did not provide a specific duration [16,17]. Finally, croup associated with an uncommon pathogen was also defined as atypical. In a study by Low et al. (2012), fungus was identified as the causative organism while Miller et al. (1982) reported a case of atypical croup caused by Chlamydia trachomatis [[17], [18], [19]].

did not provide a specific duration [16,17]. Finally, croup associated with an uncommon pathogen was also defined as atypical. In a study by Low et al. (2012), fungus was identified as the causative organism while Miller et al. (1982) reported a case of atypical croup caused by Chlamydia trachomatis [[17], [18], [19]]. 4.2 Etiology and incidence of atypical croup The etiology of atypical croup included several organisms not normally associated with croup. Several studies noted herpes simplex virus type 1 as the causative organism, associated with oral and laryngeal ulcers. Low et al. (2012) reported the fungus Curvularia in a single case presenting as an atypical croup-like syndrome in an immunosuppressed child while Miller et al. (1982) reported Chlamydia trachomatis and Staphylococcus aureus in atypical croup [18,19]. Hatherill et al. (2001) reported other bacterial species; however it was unclear if these pathogens were associated with cases of atypical croup in their chart review [15]. The incidence of atypical croup was not reported in any of the studies reviewed.

a trachomatis and Staphylococcus aureus in atypical croup [18,19]. Hatherill et al. (2001) reported other bacterial species; however it was unclear if these pathogens were associated with cases of atypical croup in their chart review [15]. The incidence of atypical croup was not reported in any of the studies reviewed. 4.3 Diagnosis of atypical croup Croup is primarily a clinical diagnosis, however the cases of atypical croup reviewed often required further investigation. In our review, endoscopy was used to aid diagnosis in all 12 studies. Findings on endoscopy varied based on etiology. Ulcerations were reported in cases of herpetic croup by O'Niel et al. (2013), Inglis Jr (1993), Chauhan et al. (2007), and Krause et al. (1998) [13], [16], [17], [20]. In a retrospective chart review, Hatherill et al. (2001) assessed the presence of ulcerative lesions on microlaryngoscopy in children with croup who were admitted to the PICU [15]. Laryngeal ulcerations were identified in 10% of children admitted while 18% had gingivostomatitis consistent with HSV infection. Furthermore, in a review performed by Cooper et al. (2012) of 80 children diagnosed with atypical croup, 33 had large airway lesions, which included subglottic stenosis, laryngeal clefts, subglottic hemangiomas, tracheomalacia, and laryngomalacia [11]. In addition, the authors reported associations between atypical croup and atopic conditions such as eosinophilic esophagitis. The authors recommended that endoscopy be coupled with allergy and gastrointestinal investigations in children with atypical croup. Barnes et al. (1999) described a case where laryngoscopy revealed two large subglottic cysts in addition to subglottic stenosis [21]. Finally, Waki et al. (1995) noted an association between gastoesophageal reflex disease (GERD) and recurrent croup [12]. Patients with GERD had shorter time periods between episodes of croup and younger age of presentation. Findings associated with GERD, including edema and erythema of the arytenoid and tracheal mucosa, were identified in 47% of children assessed for recurrent croup and 25% showed anatomical airway narrowing with direct laryngoscopy and bronchoscopy.

time periods between episodes of croup and younger age of presentation. Findings associated with GERD, including edema and erythema of the arytenoid and tracheal mucosa, were identified in 47% of children assessed for recurrent croup and 25% showed anatomical airway narrowing with direct laryngoscopy and bronchoscopy. 4.4 Management of atypical croup The cornerstone of croup management is nebulized epinephrine and systemic corticosteroid. In cases of atypical croup, management often differs. Chauhan et al. (2007) discussed treatment with corticosteroids in the setting of HSV-induced atypical laryngitis and suggested that corticosteroid use may delay diagnosis and initiation of appropriate therapy, resulting in prolongation of the clinical course [13]. Steroid administration was also discouraged with additional causes of atypical laryngitis including inflammatory processes, such as GERD and alternate viral processes, such as recurrent respiratory papillomatosis. Inglis Jr. (1993) proposed that prolonged use of corticosteroids resulted in susceptibility to herpetic croup and suggested that corticosteroid use be limited to 48 h [16]. Acyclovir was given following cessation of corticosteroid medication in their case report.

processes, such as recurrent respiratory papillomatosis. Inglis Jr. (1993) proposed that prolonged use of corticosteroids resulted in susceptibility to herpetic croup and suggested that corticosteroid use be limited to 48 h [16]. Acyclovir was given following cessation of corticosteroid medication in their case report. Atypical croup often necessitated definitive airway management in the studies reviewed. Hatherill et al. (2001) reported that 9 out of 148 children (6%) admitted to the PICU for croup needed intubation with a median duration of 4 days [15]. Five children required tracheostomy to prevent mucosal damage secondary to the mechanical trauma. The clinical decision to proceed to tracheostomy was based on severe subglottic ulceration and narrowing such that a size 3.0 endotracheal tube could not be passed. Intubation was deemed necessary due to subglottic edema in HSV laryngitis reported by Chauhan et al. (2007), Inglis Jr (1993), and O'Neil et al. (2013) [13], [16], [17]. Barnes et al. (1999) reported a case in which marsupialization of subglottic cysts lead to resolution of symptoms [21].

heal tube could not be passed. Intubation was deemed necessary due to subglottic edema in HSV laryngitis reported by Chauhan et al. (2007), Inglis Jr (1993), and O'Neil et al. (2013) [13], [16], [17]. Barnes et al. (1999) reported a case in which marsupialization of subglottic cysts lead to resolution of symptoms [21]. 5 Discussion Presented is a case of croup in a child that occurred secondary to RSV infection. His symptoms were relapsing and remitting in nature, with only short-lived responses to conventional treatments, requiring PICU management, intubation, and tracheostomy. A pseudotumor identified on bronchoscopy suggested an inflammatory etiology, however rheumatologic and immunologic workup were negative. Complete recovery was unusually prolonged necessitating continued maintenance of his tracheostomy and repeated operative bronchoscopy for airway assessment. This case report of croup is atypical in three notable ways: (1) the prolonged presentation and unresponsiveness to medical therapy (2) the need for a surgical airway despite 21 days of maximal medical therapy; and (3) identification of an inflammatory airway lesion suggesting a potential underlying primary inflammatory disease process. This case offers insight into the diagnostic challenges of atypical croup and highlights the role of surgical airway management in cases unresponsive to medical therapy.

ximal medical therapy; and (3) identification of an inflammatory airway lesion suggesting a potential underlying primary inflammatory disease process. This case offers insight into the diagnostic challenges of atypical croup and highlights the role of surgical airway management in cases unresponsive to medical therapy. Typical croup syndromes are usually self-limited in nature and responsive to medical therapy [6]. In unusual cases of croup presentations, a different approach to diagnosis and management may be necessary. However, there is no commonly accepted definition of atypical croup. Thus, we performed a systematic review on atypical croup to identify definitions available in the literature. Furthermore, we sought to identify diagnostic approaches used to investigate these atypical presentations and reported management strategies. Our review identified 12 English language articles on atypical croup in the pediatric population. In our search the term atypical croup was used in the literature in a variety of contexts. Recurrent episodes, croup in a child who lies outside the age range of typical viral croup, prolonged or severe episodes, or croup occurring due to uncommon etiology were all observed.

atypical croup in the pediatric population. In our search the term atypical croup was used in the literature in a variety of contexts. Recurrent episodes, croup in a child who lies outside the age range of typical viral croup, prolonged or severe episodes, or croup occurring due to uncommon etiology were all observed. Typical viral croup is most commonly associated with the parainfluenza viruses, but is also associated with adenovirus, RSV, rhinovirus and the influenza viruses [4]. Our search revealed several infectious etiologies associated with atypical croup, the most common of which was HSV. This association may be linked to immunosuppressive effects of corticosteroid administration due to treatment of an initial typical croup presentation [13,15,16,20]. In one immunosuppressed child, the fungus Curvularia was isolated [18]. This suggests that immune modulation plays a role in determining croup presentation and severity. Bacterial causes such as Staphylococcus and Chlamydiae species were also reported although distinctions made between bacterial tracheitis and atypical croup were unclear in those studies [15,19].

vularia was isolated [18]. This suggests that immune modulation plays a role in determining croup presentation and severity. Bacterial causes such as Staphylococcus and Chlamydiae species were also reported although distinctions made between bacterial tracheitis and atypical croup were unclear in those studies [15,19]. Moreover, the clinical presentations reported varied in duration and severity. Prolonged duration was a defining factor in assigning the definition of atypical croup in several reports reviewed. Two studies outlined a time period of presentation beyond 7 days as atypical [16,17]. However, several studies noted prolonged duration as a factor without specifying a specific period of time [15,[20], [21], [22]]. Severity was defined very broadly or not at all in most studies, with the most common definition being PICU admission. In studies where mean or median ages were reported, age of presentation was often older than 3 years of age, with the oldest being 11 years of age, or younger than 6 months of age. No cases were reported beyond that lower boundary. In the literature, the peak incidence of typical croup is cited between the ages of 6 months to 3 years, however presentations of croup up to 6 years of age were not uncommon. The incidence of croup by age reported by Denny et al. (1983) for example was greater than 1 in every 100 children every year until 6 years of age [23]. Based on the incidence found in typical croup literature, the upper bound of 3 years of age proposed for atypical croup appears too low.

ears of age were not uncommon. The incidence of croup by age reported by Denny et al. (1983) for example was greater than 1 in every 100 children every year until 6 years of age [23]. Based on the incidence found in typical croup literature, the upper bound of 3 years of age proposed for atypical croup appears too low. Overall, our search suggests that atypical croup is not a unique clinical entity. Rather, it is a term most often used as catch-all diagnosis for presentations that lie outside the common definition of croup. Some studies have proposed more strict definitions of atypical croup, mostly based on recurrent presentations with consideration of age at presentation; however, there remains a gap in knowledge on the incidence of these atypical episodes [11]. This discordance suggests that it may be beneficial to present a broader definition of atypical croup. We therefore propose that atypical croup may be defined as a child who presents with croup that is either: a) severe, necessitating definitive airway management, b) prolonged, symptoms persisting for 7 or more days despite medical therapy, c) untimely, presenting at an age younger than 6 months or older than 6 years, d) associated with an atypical pathogen, or e) presenting with an airway lesion other than the classically described steeple sign on x-ray imaging. Several of these features were present in this case study, in particular symptoms refractory to maximal medical therapy, prolonged duration, a pseudotumour, and airway narrowing persisting for 6 months after initial presentation.

with an airway lesion other than the classically described steeple sign on x-ray imaging. Several of these features were present in this case study, in particular symptoms refractory to maximal medical therapy, prolonged duration, a pseudotumour, and airway narrowing persisting for 6 months after initial presentation. In the case presented, operative endoscopy was utilized to identify a severely narrowed subglottis and inflammatory pseudotumor. Atypical findings on endoscopy were reported in several studies. Large airway lesions such as subglottic clefts, laryngeal ulcers and hemangiomas were found in a variety of cases labelled atypical croup. In one study the prevalence of these lesions was as high as 40% in patients with recurrent croup [11]. The question of who needs endoscopic evaluation was only addressed in one evaluated study by Farmer and Wohl (2001) who proposed endoscopy be undertaken if croup is severe, persists despite treatment, in cases of abnormal imaging or prior to elective surgery [1,14].

as high as 40% in patients with recurrent croup [11]. The question of who needs endoscopic evaluation was only addressed in one evaluated study by Farmer and Wohl (2001) who proposed endoscopy be undertaken if croup is severe, persists despite treatment, in cases of abnormal imaging or prior to elective surgery [1,14]. Lastly, our patient developed progressive, severe respiratory distress that required intubation and eventual tracheostomy. We sought to clarify the patient characteristics that place children at risk of intubation or tracheostomy. In one review by O'Niel et al. (2013) of HSV laryngitis, 50% of patients required intubation [17]. The rate of intubation in other forms of atypical croup was not well studied. Furthermore, specific characteristics associated with intubation were not reported. However, Hatherill et al. (2001) described the decision to proceed with tracheostomy based on the prevention of further mucosal damage due to severe ulceration and subglottic narrowing from prolonged intubation [15]. The study noted that 5 children of the 263 reviewed required tracheostomy. In our case the decision to proceed with tracheostomy was based on persistence of airway narrowing following intubation. Overall tracheostomy was only reported in one of the twelve studies reviewed. Thus, other than herpetic croup, there remains a question of the risk factors and patient characteristics that may result in intubation and tracheostomy. Future investigation is warranted to clarify this issue.

narrowing following intubation. Overall tracheostomy was only reported in one of the twelve studies reviewed. Thus, other than herpetic croup, there remains a question of the risk factors and patient characteristics that may result in intubation and tracheostomy. Future investigation is warranted to clarify this issue. The atypical presentation of croup is a common problem faced by emergency physicians, pediatricians and otolaryngologists, and it provides several diagnostic and management challenges. Given the findings from our literature review, we propose a management pathway for atypical cases of croup. We reviewed the Canadian Pediatric Society Position Statement on the “Acute management of croup in the emergency department”, the Seattle Children's Hospital Croup Pathway, and the “Towards Optimized Practice (TOP) Guidelines on the Diagnosis and Management of Croup” [[24], [25], [26]]. We then adapted our institutional croup pathway, and integrated recommendations from the practice guidelines listed and the conclusions from our literature review on atypical croup to formulate a pathway for atypical croup (Fig. 2 ).Fig. 2 Croup diagnosis and management pathway. Fig. 2

The atypical presentation of croup is a common problem faced by emergency physicians, pediatricians and otolaryngologists, and it provides several diagnostic and management challenges. Given the findings from our literature review, we propose a management pathway for atypical cases of croup. We reviewed the Canadian Pediatric Society Position Statement on the “Acute management of croup in the emergency department”, the Seattle Children's Hospital Croup Pathway, and the “Towards Optimized Practice (TOP) Guidelines on the Diagnosis and Management of Croup” [[24], [25], [26]]. We then adapted our institutional croup pathway, and integrated recommendations from the practice guidelines listed and the conclusions from our literature review on atypical croup to formulate a pathway for atypical croup (Fig. 2 ).Fig. 2 Croup diagnosis and management pathway. Fig. 2 6 Conclusions This case offers insight into the diagnostic pathway of atypical croup and highlights the role of surgical airway management in cases unresponsive to medical therapy. Currently, there is no consensus definition for atypical croup. The current literature contains definitions applied to recurrent episodes, yet without agreement on the frequency, duration, or severity. Moreover, atypical croup was used in multiple contexts including airway lesions, bacterial tracheitis, and laryngopharyngeal reflux. Finally, there exists a gap in knowledge of the incidence of atypical croup and factors that result in definitive airway management.

out agreement on the frequency, duration, or severity. Moreover, atypical croup was used in multiple contexts including airway lesions, bacterial tracheitis, and laryngopharyngeal reflux. Finally, there exists a gap in knowledge of the incidence of atypical croup and factors that result in definitive airway management. We propose a broad definition of atypical croup and propose a pathway for healthcare providers faced with atypical presentations of croup. At our institution otolaryngology was commonly being consulted for both croup, as well as atypical croup, with ‘atypical’ frequently being applied to cases that, based on our literature review, are, in fact, quite typical. Application of this paradigm within our own centre has not only reduced croup consultations, but provided emergency physicians, hospital pediatricians and intensivists with a common basis from which to approach seemingly unusual or atypical croup presentations. Declarations of competing interests Raphael Hanna: none. Francisco Lee: none. Derek Drummond: none. Warren K. Yunker: none. Submission declaration and verification The authors approve the publication of this manuscript and confirm that the work described has not been published elsewhere, is not under consideration for publication elsewhere, and that, if accepted, it will not be published elsewhere in the same form, in English or in any other language, including electronically without the written consent of the copyright holder. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

1 Introduction Tonsils and adenoids, the common sites for recurrent inflammation in the pediatric population, are continually exposed to antigens, hence hyperplasia of their lymphoid component accounts for the increase in their size [1]. Some authors have demonstrated a strong relation between viral infection [2], [3], [4] and recurrent pharyngotonsillitis. There are many viruses involved in the pathogenesis of pharyngotonsillitis including adenoviruses, parainfluenza viruses, rhinoviruses, herpes simplex viruses, respiratory syncytial viruses, Epstein–Barr viruses (EBV), influenza viruses, coxsackie A viruses, corona viruses, and cytomegaloviruses [2], [3], [5], [6]. In spite of high prevalence of viral infection in adenotonsillar tissue, the methods to detect viruses make this approach difficult in routine practice [2], [3], [5], [6].

Adenoviruses infrequently cause common colds, and respiratory infections caused by these viruses tend to be severe, characterized by high and prolonged fever and strong inflammatory response [40], [41]. Adenoviruses more often cause occasional epidemics in semi-closed communities, such as garrisons or orphanages [42]. 4 Clinical entities and complications of viral respiratory infections in children A given respiratory viruses may infect widely respiratory mucosa in humans, although it seems that some viruses are more prone to infect specific parts of respiratory tract than others. The various designated respiratory disease entities are in practise partly overlapping and all of them are associated with different respiratory viruses (Table 1 ). In children, the most common viral respiratory infections are simple URIs (the common cold) and AOM.Table 1 Respiratory tract infections and viral causative agents in children Disease Adenoviruses Coronaviruses Enteroviruses Influenza viruses Parainfluenza viruses RSVa Rhinoviruses Common cold + ++ ++ ++ + + +++ Tonsillitis +++ − ++ + + + − Laryngitis + − + ++ +++ + + Bronchitis + + + +++ ++ +++ + Bronchiolitis + + + ++ ++ +++ ++ Pneumonia + + + +++ ++ +++ ++ a RSV: respiratory syncytial virus.

ory syncytial viruses, Epstein–Barr viruses (EBV), influenza viruses, coxsackie A viruses, corona viruses, and cytomegaloviruses [2], [3], [5], [6]. In spite of high prevalence of viral infection in adenotonsillar tissue, the methods to detect viruses make this approach difficult in routine practice [2], [3], [5], [6]. Commonly causing infectious mononucleosis, EBV is a member of γ1-herpesvirus and has a genome comprised approximately 172 nucleotide base pairs [7]. EBV has a linear genome which is characterized by a distinctive sequence reiteration. At the termini there are 20 copies of a 500 bp repeated sequence that is complementary and therefore permits circularization of the linear genome to form the EBV episome. EBV episome is the molecular basis for latent infection. It is a circular intracellular form of genome that maintains a persistent relationship within the cell, like an autonomous piece of DNA situated in the chromatin [7], [8]. EBV is B-lymphotropic and has the ability to transform memory B-cells into blast cells, with permanent proliferation, leading to tonsillar enlargement [9]. EBV is associated not only with infectious mononucleosis but also with benign diseases, such as oral hairy leukoplakia, and malignancies, such as Hodgkin's lymphoma, non-Hodgkin's lymphomas, nasopharyngeal carcinoma, gastric carcinoma and breast carcinoma. It is recently being associated with autoimmune diseases, such as lupus erythematosus and multiple sclerosis [7], [10], [11]. For these reasons, our study is designed to identify the magnitude of EBV infection and types and pattern of distribution of EBV-infected cells in tonsils and adenoids among UAE population which will be of great help in future planning to prevent this infection.

pus erythematosus and multiple sclerosis [7], [10], [11]. For these reasons, our study is designed to identify the magnitude of EBV infection and types and pattern of distribution of EBV-infected cells in tonsils and adenoids among UAE population which will be of great help in future planning to prevent this infection. 2 Materials and methods 2.1 Review of cases In total, 46 cases of tonsillectomy with adenoidectomy, which were performed due to tonsillar and adenoidal enlargement, were randomly selected from surgical pathology archive in the department of pathology at Tawam hospital in Al Ain city for the period June 2004 through May 2005. Forty-six paraffin blocks of tonsillectomy specimens and another 46 paraffin blocks of adenoidectomy specimens were available for this study. The available Hematoxylin and eosin (H&E) and immunohistochemical stained sections were reviewed. The age, sex, and clinical presentation were obtained by reviewing all the histopathologic reports and request forms of all cases. For the detection of EBV, two methods were used to increase the specificity and sensitivity of detection, the streptavidin-biotin immunohistochemical immunoperoxidase method to detect Epstein–Barr virus latent membrane antigen type I (EBV-LMP-1) and in situ hybridization (ISH) for EBV encoded RNA (EBER).

of all cases. For the detection of EBV, two methods were used to increase the specificity and sensitivity of detection, the streptavidin-biotin immunohistochemical immunoperoxidase method to detect Epstein–Barr virus latent membrane antigen type I (EBV-LMP-1) and in situ hybridization (ISH) for EBV encoded RNA (EBER). 2.2 Immunohistochemistry (IHC) Immunohistochemical (IHC) staining was performed by standard streptavidin–biotin immunoperoxidase technique [12] using the following mouse antihuman monoclonal antibodies (DAKO Cytomation, Glostrup, Denmark); EBV-latent membrane protein-1(LMP1)(clone CS 1–4), CD3 (clone PC3/188A), CD20 (clone L26), cytokeratin (clone AE1/AE3), all diluted to 1:100, and visualized by a commercially available detection kit (DAKO EnVision Plus-HRP, DAKO, Glostrup, Denmark) and 3-3′-diaminobenzidine (DAKO, Glostrup, Denmark) as a chromogen substrate to obtain a brown end-product. Lymph node sections were used as positive controls for CD20, CD3. Skin epidermis was used as a positive control for cytokeratin AE1/AE3. Hodgkin Lymphoma (HL) LMP-1-positive sections were used as positive controls for LMP-1. For negative controls, primary antibody was replaced with normal mouse serum.

rown end-product. Lymph node sections were used as positive controls for CD20, CD3. Skin epidermis was used as a positive control for cytokeratin AE1/AE3. Hodgkin Lymphoma (HL) LMP-1-positive sections were used as positive controls for LMP-1. For negative controls, primary antibody was replaced with normal mouse serum. 2.3 In situ hybridization (ISH) In situ hybridization (ISH) was performed by standard techniques using a specific oligonucleotide probe (Novocastra–LEBV-K, UK) which hybridizes to EBV encoded RNA (EBER) transcripts concentrated in the nuclei of latently infected cells. With each batch of cases studied, positive and negative control slides were also run. The positive control slide was a known case of EBV positive HL to which a specific EBER oligonucleotide probe was added. The negative control slide was another section of the same case of known EBV positive HL to which a random probe consisting of fluorescein labeled oligonucleotide cocktail was added. In addition, for each case studied two sections were used; the EBER oligonucleotide probe was added to one section, and the random probe was added to the other. Using this random probe as a negative background control alongside the EBV probe contributes to the validation of staining obtained by the EBV probe. If this negative control slide showed significant background staining in a particular case, the slide having the EBER probe was considered non-interpretable and the test was repeated for that particular case.

Disease Adenoviruses Coronaviruses Enteroviruses Influenza viruses Parainfluenza viruses RSVa Rhinoviruses Common cold + ++ ++ ++ + + +++ Tonsillitis +++ − ++ + + + − Laryngitis + − + ++ +++ + + Bronchitis + + + +++ ++ +++ + Bronchiolitis + + + ++ ++ +++ ++ Pneumonia + + + +++ ++ +++ ++ a RSV: respiratory syncytial virus. 4.1 Common colds The common cold is almost entirely a viral disease. Mäkelä et al. found respiratory viruses to be associated with two-thirds of common cold events among young adults, while bacteria were cultured in only 4% of cases [35]. Rhinoviruses are the leading cause of common colds in all age groups [6], [25], [35], and by the age of 2 years, most of the children have rhinovirus-specific antibodies [27]. After introduction of the reverse transcription (RT)-PCR technique, it has become evident that enteroviruses are also frequently associated with common colds [11]. Although both RSV and influenza A virus are well known for causing lower respiratory infections, they also generate upper respiratory infections [43].

background control alongside the EBV probe contributes to the validation of staining obtained by the EBV probe. If this negative control slide showed significant background staining in a particular case, the slide having the EBER probe was considered non-interpretable and the test was repeated for that particular case. 2.4 Triple staining technique Five-micrometer sections were stained for EBER using in situ hybridization protocol described earlier. After nuclear visualization, mouse antihuman monoclonal antibodies for CD20 (DAKO Cytomation, Glostrup, Denmark) were added and visualized by a commercially available detection kit (DAKO EnVision Plus-HRP, DAKO, Glostrup, Denmark) and 3-3′-diaminobenzidine (DAKO, Glostrup, Denmark) as a chromogen substrate to obtain a brown end-product. Subsequently, mouse antihuman monoclonal antibodies for CD3 (DAKO Cytomation, Glostrup, Denmark) were added and visualized using the EnVision Plus-alkaline phosphatase kit (DAKO, Glostrup, Denmark), and New Fuchsin (Merck, Darmstadt, Germany) as a second substrate to yield a red end-product. Finally, sections were mounted by water soluble mounting media. 2.5 Statistical analysis The statistical analysis was performed using SPSS for windows version 18 (SPSS Inc, Chicago, USA) and analyze it (Analyze-it software Ltd., Leeds, UK). Student's t-test was used to compare continuous variables. Quantitative variables were analyzed with the χ 2-test and correlations of ordinal variables using the Spearman rank correlation coefficient. P value <0.05 was considered significant.

SPSS Inc, Chicago, USA) and analyze it (Analyze-it software Ltd., Leeds, UK). Student's t-test was used to compare continuous variables. Quantitative variables were analyzed with the χ 2-test and correlations of ordinal variables using the Spearman rank correlation coefficient. P value <0.05 was considered significant. 2.6 Research ethics The project has been approved by Al Ain district Human research Ethical committee (Protocol No. 07-145). 3 Results 3.1 Age and gender distribution In total, 46 cases of tonsillectomy and adenoidectomy, due to tonsillar and adenoid hypertrophy, were selected. All cases were in the 1st and 2nd decade of life with predominant clustering (63%) in the 1st decade. Twenty-seven cases were females and 19 cases were males. 3.1.1 EBV expression correlation between age and gender Although EBV-positive cases were more (n = 14) in the 1st decade than in the 2nd decade (n = 6), Cross tab and logistic regression show no significant association between EBV expression and the age (P = 0.324) and gender (P = 0.442) distributions (Table 1 ).Table 1 EBV expression and correlation with age and gender. Age EBVa positive cases EBVa negative cases Total cases Female + Male + Total + Female − Male − Total − Female Male Total 0–9 6 8 14 6 8 14 12 16 28 10–19 5 1 6 10 2 12 15 3 18 Total 11 9 20 16 10 26 27 19 46 a Epstein–Barr virus.

3.1.1 EBV expression correlation between age and gender Although EBV-positive cases were more (n = 14) in the 1st decade than in the 2nd decade (n = 6), Cross tab and logistic regression show no significant association between EBV expression and the age (P = 0.324) and gender (P = 0.442) distributions (Table 1 ).Table 1 EBV expression and correlation with age and gender. Age EBVa positive cases EBVa negative cases Total cases Female + Male + Total + Female − Male − Total − Female Male Total 0–9 6 8 14 6 8 14 12 16 28 10–19 5 1 6 10 2 12 15 3 18 Total 11 9 20 16 10 26 27 19 46 a Epstein–Barr virus. 3.2 EBV distribution 3.2.1 EBV prevalence in tonsillectomy specimens In total, 20 (43%) specimens were positive by in situ hybridization method for EBER and show the black nuclear stain. EBER-positive cells are seen within follicular (Fig. 1 ) and interfollicular (Fig. 2 ) regions. The frequency of EBV is shown in Table 2 . Only 5 (11%) specimens of tonsillectomy show cytoplasmic and membranous positivity for LMP-1 as shown in Fig. 3 . LMP-1-positive large cells are seen within the mantle zone of lymphoid follicle and within the interfollicular area, however, LMP-1 expression is seen also in small lymphocytes in the interfollicular area.Fig. 1 Expression of EBV in the lymphoid follicles of enlarged tonsils and adenoids. (A) EBER-positive cell (thick arrow) is seen within a lymphoid follicle with a germinal center (arrow head). (B) Many EBER-positive cells (thick arrows) are seen in lymphoid follicles (arrow heads), however, some EBER-positive cells are also seen in the interfollicular area. (C) EBER-positive cells (thick arrows) are seen in the center of lymphoid follicles or having a circumferential distribution in the interphase between germinal center (arrow head) and mantle zone cells (thin arrow). (D) Many EBER-positive cells (thick arrows) are seen in the center of lymphoid follicles or having a circumferential distribution in the interphase between germinal center (arrow head) and mantle zone cells (thin arrow). A large atypical EBER-positive cell (curved arrow) is seen in the germinal center.

ls (thin arrow). (D) Many EBER-positive cells (thick arrows) are seen in the center of lymphoid follicles or having a circumferential distribution in the interphase between germinal center (arrow head) and mantle zone cells (thin arrow). A large atypical EBER-positive cell (curved arrow) is seen in the germinal center. Fig. 2 Expression of EBV in the interfollicular areas of enlarged tonsils and adenoids. (A) There is no expression of EBV in the epithelial cell (arrow head), however, EBER positive cells are seen in the interfollicular areas (thick arrowed) in between lymphoid follicles (thin arrows), EBER in situ hybridization. (B) Many EBER positive cells are seen in the interfollicular areas (thick arrows). (C) A lymphoid follicle (arrow head) is surrounded by many EBER-positive cells in the interfollicular area (thick arrows). (D) Many EBER positive cells are seen in the interfollicular areas (thick arrows), close to a lymphoid follicle with a germinal center (arrow head). Table 2 The prevalence of EBV in tonsillectomy and adenoidectomy specimens. Specimen EBV (+) ISHa cases % EBV (−) ISHa cases % Total % Tonsils 20 43 26 57 46 100 Adenoids 7 15 39 85 46 100 a In situ hybridization.

[27]. After introduction of the reverse transcription (RT)-PCR technique, it has become evident that enteroviruses are also frequently associated with common colds [11]. Although both RSV and influenza A virus are well known for causing lower respiratory infections, they also generate upper respiratory infections [43]. In adults and especially in children, acute respiratory infections often spread into the paranasal sinuses, causing mucosal edema and accumulation of mucus. A recent study reported that 68% of children with uncomplicated upper respiratory infections had major abnormalities in the paranasal sinuses, which mainly resolved after 2 weeks without antimicrobial treatment [44]. In another study, 70% of children with purulent rhinorrhea as the only symptom had opacification of the paranasal sinuses on computer tomography scans [45]. Young children almost always have nasal secretions in paranasal sinuses during upper respiratory infections (so-called rhinosinusitis). Based on this, sinusitis in children could be considered a natural extension of the common cold. However, evidence of direct viral infection of the maxillary sinus in children is lacking. In adults with acute maxillary sinusitis, a virus can be found in 10% of maxillary secretion samples by viral culture [46] and in 40% by PCR [47].

Fig. 2 Expression of EBV in the interfollicular areas of enlarged tonsils and adenoids. (A) There is no expression of EBV in the epithelial cell (arrow head), however, EBER positive cells are seen in the interfollicular areas (thick arrowed) in between lymphoid follicles (thin arrows), EBER in situ hybridization. (B) Many EBER positive cells are seen in the interfollicular areas (thick arrows). (C) A lymphoid follicle (arrow head) is surrounded by many EBER-positive cells in the interfollicular area (thick arrows). (D) Many EBER positive cells are seen in the interfollicular areas (thick arrows), close to a lymphoid follicle with a germinal center (arrow head). Table 2 The prevalence of EBV in tonsillectomy and adenoidectomy specimens. Specimen EBV (+) ISHa cases % EBV (−) ISHa cases % Total % Tonsils 20 43 26 57 46 100 Adenoids 7 15 39 85 46 100 a In situ hybridization. Fig. 3 Expression of EBV-LMP-1 in enlarged tonsils. (A) LMP-1-positive large cells (thick arrows) are seen within the mantle zone of a lymphoid follicle and within the interfollicular area, however, LMP-1 expression is seen also in small lymphocytes in the interfollicular area (thin arrows), streptavidin–biotin immunoperoxidase method. (B) LMP-1-positive cells (thick arrows) are seen within the interfollicular area, close to a lymphoid follicle (arrow head), streptavidin–biotin immunoperoxidase method. (C) LMP-1-positive large cells (thick arrows) are seen within the mantle zone of a lymphoid follicle (arrow head), streptavidin–biotin immunoperoxidase method. (D) Many LMP-1-positive cells (thick arrows) are seen within the interfollicular area, streptavidin–biotin immunoperoxidase method.

tin immunoperoxidase method. (C) LMP-1-positive large cells (thick arrows) are seen within the mantle zone of a lymphoid follicle (arrow head), streptavidin–biotin immunoperoxidase method. (D) Many LMP-1-positive cells (thick arrows) are seen within the interfollicular area, streptavidin–biotin immunoperoxidase method. 3.2.2 EBV prevalence in adenoidectomy specimens In total, 7 (15%) specimens were positive by in situ hybridization method for EBER and show the black nuclear stain. EBER-positive cells are seen within follicular (Fig. 1) and interfollicular (Fig. 2) regions. The frequency of EBV is shown in Table 2. All those cases were also EBV positive in their tonsillectomy specimens. LMP-1 staining was negative in all 46 case adenoid specimens.

method for EBER and show the black nuclear stain. EBER-positive cells are seen within follicular (Fig. 1) and interfollicular (Fig. 2) regions. The frequency of EBV is shown in Table 2. All those cases were also EBV positive in their tonsillectomy specimens. LMP-1 staining was negative in all 46 case adenoid specimens. 3.3 Distribution of EBV among lymphoid and epithelial cells Nearly 90% of the EBER-positive cells were found within the T-cell-rich interfollicular regions (Fig. 2) while 10% were present within the B-cell-rich germinal centers of secondary follicles (Fig. 1). All EBER-positive cells were B cells. The EBER-positive cells show immunoreactivity to CD20 only (Fig. 4 ). There was no immunoreactivity to CD3 (Fig. 4) and cytokeratin (Fig. 2A). Individual EBER positive cells appeared to be randomly distributed within the interfollicular regions (Fig. 3). The distribution of EBER-postive cells within lymphoid follicles is variable. Some follicles have few positive cells within the germinal center (Fig. 1). Some of them were atypical large cells (Fig. 1D). Other follicles show their presence within the mantle zone layer (Fig. 1C and D). There is a third pattern where EBER-positive cells form a circumferential distribution at the interphase between the germinal center cells and the mantle zone cells (Fig. 2C and D).Fig. 4 Triple stain of EBER, CD20 and CD3 in enlarged tonsils and adenoids. B lymphocytes show membranous brown staining for CD20 (thin arrows) while the T lymphocytes show membranous red staining for CD3 (arrow heads). Only some of the B lymphocytes show blue nuclear staining for EBER in situ hybridization (thick arrows): A, 400× and B 1000×. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

T lymphocytes show membranous red staining for CD3 (arrow heads). Only some of the B lymphocytes show blue nuclear staining for EBER in situ hybridization (thick arrows): A, 400× and B 1000×. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.) 4 Discussion Enlargement of tonsils and adenoids is a common finding in children but the etiology remains controversial. In children, there is no correlation between hypertrophic tonsils and the body mass index [13]. The weight of normal tonsils in children is not well documented in anatomy literature; however, the mean weight of the tonsillectomy specimens from pathologic examination was shown to be 7.3 g in children between the ages of 2 and 12 years with 63% of it falls between 5 and 8 g [14]. Tonsillectomy is a common surgical procedure in children, and the most common indication is tonsillar hypertrophy, which leads to obstructive symptoms in the upper airway. There is a poor correlation between enlarged tonsils and the associated symptoms [14].

f 2 and 12 years with 63% of it falls between 5 and 8 g [14]. Tonsillectomy is a common surgical procedure in children, and the most common indication is tonsillar hypertrophy, which leads to obstructive symptoms in the upper airway. There is a poor correlation between enlarged tonsils and the associated symptoms [14]. Recurrent bacterial tonsillitis is considered to be the main reason for the enlargement of the tonsils; however, the bacteriological examinations of the tonsils showed that the relationship is not very clear [15]. Misuse of antibiotic therapy in acute tonsillitis may lead to change in tonsillar microflora and predisposing to recurrent tonsillitis and subsequent colonization by various bacteria and viruses [16]. The role of Actinomycosis is considered in the etiology of recurrent tonsillitis, not as an active infection but as a factor in the development of lymphoid hyperplasia and hypertrophy [17]. Many viruses are involved in the pathogenesis of pharyngotonsillitis [2], [3], [5], [6]. Colonization of the tonsils by EBV shows no correlation between EBV–DNA quantity and viral core antigen-IgG quantity in the autologous sera [18]. EBV infection is very common, with a seroprevalence rates in excess of 90% worldwide. [19]. Nearly all infections are acquired by oral contact with persons carrying EBV in saliva. Although the exact details of EBV transmission in the oral cavity remain unknown, it is likely that initial infection of either oral epithelial cells or tonsillar B cells is followed by a brief period of replication and lifelong persistence in B lymphocytes [20].

red by oral contact with persons carrying EBV in saliva. Although the exact details of EBV transmission in the oral cavity remain unknown, it is likely that initial infection of either oral epithelial cells or tonsillar B cells is followed by a brief period of replication and lifelong persistence in B lymphocytes [20]. The current study involves tonsils and adenoids of UAE nationals only. Cases were randomly selected from the surgical archive of the pathology department at Tawam hospital for a period June 2004 to May 2005. The study includes patients from different parts of the country. Tawam hospital is the main hospital that provides medical services for UAE nationals. Hence our samples are almost representative of the UAE population.

the surgical archive of the pathology department at Tawam hospital for a period June 2004 to May 2005. The study includes patients from different parts of the country. Tawam hospital is the main hospital that provides medical services for UAE nationals. Hence our samples are almost representative of the UAE population. Tonsils and adenoids were chosen to identify the prevalence of EBV in tissue since they are the initial sites of infection and may reflect the magnitude of EBV infection in the community. In addition, it may help in identifying the correlation between EBV infection and the prevalence of EBV-associated diseases in the community, which might help in the setting of plans to prevent these diseases. In this study we show that EBV is prevalent in 43% of tonsils and 15% of adenoids. It is interesting to see all our cases are children, a finding seen by Endo et al. too [21], indicating that tonsillar and adenoidal hypertrophy occur mainly during this age group. It is well known that EBV has a tropism for the oral and nasopharyngeal tissues, and leads to lymphocytic proliferation [21]. The role of oropharyngeal epithelial cells as a reservoir of EBV was already suggested [21]. We show EBV only inhabits B lymphocytes in tonsils and adenoids, but not T lymphocytes or epithelial cells indicating that B lymphocytes are the main reservoir for EBV. Hudnall et al. [22], showed EBV to be rarely found within epithelial cells and T lymphocytes in EBV-infected tonsils.

f EBV was already suggested [21]. We show EBV only inhabits B lymphocytes in tonsils and adenoids, but not T lymphocytes or epithelial cells indicating that B lymphocytes are the main reservoir for EBV. Hudnall et al. [22], showed EBV to be rarely found within epithelial cells and T lymphocytes in EBV-infected tonsils. Following primary infection in the oropharynx, EBV persists in numerous anatomical sites including pharyngeal tonsils, adenoids, lymph nodes, and peripheral blood. In peripheral blood, the virus is present in small resting memory B cells with latent gene expression limited to EBER, LMP2a, and perhaps EBNA1 [19]. It is well known that the interleukin 10 (IL-10) coding sequence is highly homologous to the EBV open reading frame BCRF1 [23]. BCRF1 protein, also termed viral IL-10 (vIL-10), inhibits the synthesis of T-helper 1 cytokines [24] and cytotoxic T lymphocytes (CTL) activity [25]. Therefore, EBV-associated antigen-specific CTL activity might be down-regulated by vIL-10 in EBV-infected areas of the tonsil. Hence, EBV is much more likely to survive in the face of immune surveillance in the tonsils, suggesting that the immune response to EBV in tonsils may be different from that in peripheral blood [25]. The prevalence of EBV infection in tonsils varies according to the detection method. Studies using the ISH for detecting EBER found 26%, [26] 29%, [3] and 65% [27] association of the EBV with tonsillitis. While in our study the prevalence of EBV infection in tonsils is 43%, which is intermediate between these studies and reflects a possible geographical difference in the prevalence of EBV.

tion method. Studies using the ISH for detecting EBER found 26%, [26] 29%, [3] and 65% [27] association of the EBV with tonsillitis. While in our study the prevalence of EBV infection in tonsils is 43%, which is intermediate between these studies and reflects a possible geographical difference in the prevalence of EBV. It is noteworthy to mention here that the prevalence of EBV (43%) in our specimens is close to its prevalence in Hodgkin lymphoma among UAE national (38%) [28], which may reflect a causal relationship between EBV and Hodgkin lymphoma.

tion method. Studies using the ISH for detecting EBER found 26%, [26] 29%, [3] and 65% [27] association of the EBV with tonsillitis. While in our study the prevalence of EBV infection in tonsils is 43%, which is intermediate between these studies and reflects a possible geographical difference in the prevalence of EBV. It is noteworthy to mention here that the prevalence of EBV (43%) in our specimens is close to its prevalence in Hodgkin lymphoma among UAE national (38%) [28], which may reflect a causal relationship between EBV and Hodgkin lymphoma. Identification of a high prevalence (43%) of EBV-EBER in tonsillectomy specimens in children suggests that the tonsils are the main reservoir for the EBV, and that this virus may be involved in tonsillar enlargement. On the other hand, the prevalence of EBV in adenoids (15%) is lower than in tonsils (43%), which indicate a lower association between adenoid hypertrophy and EBV infection. This finding is different from that found by Endo et al. [21] which found the prevalence of EBV to be higher in adenoids (57%) than tonsils (29%) by using similar method for detection (ISH), suggesting sampling and geographical differences. In this study, the immunohistochemical staining method was used to detect LMP-1 in tonsils and adenoids with the aim of strengthening our results in combination with EBER-ISH. The results are not encouraging since we detect EBV in only 5 cases (11%) and in few cells within the interfollicular area. This finding is less than that reported by Dias et al. [16], indicating a lesser expression of LMP-1 in EBV infected cells. In addition, it indicates that using EBER-ISH is the gold standard for detecting EBV in tissue. Moreover, the use of EBER-ISH in the detection of EBV is more suitable since we can use it on paraffin-embedded tissue, which allows us to test archive material when required. This makes ISH superior to immunohistochemistry in the detection of EBV in tissue.

ng EBER-ISH is the gold standard for detecting EBV in tissue. Moreover, the use of EBER-ISH in the detection of EBV is more suitable since we can use it on paraffin-embedded tissue, which allows us to test archive material when required. This makes ISH superior to immunohistochemistry in the detection of EBV in tissue. All detected EBV-infected cells in this study were CD20-positive B-cells and were located predominantly in the interfollicular regions. Nearly 90% of the EBER-positive cells were located within the T-cell-rich interfollicular regions of the tonsils, while a significant number (10%) were located within the lymphocyte-rich follicular region. These results are consistent with previous studies [22], [29].

cated predominantly in the interfollicular regions. Nearly 90% of the EBER-positive cells were located within the T-cell-rich interfollicular regions of the tonsils, while a significant number (10%) were located within the lymphocyte-rich follicular region. These results are consistent with previous studies [22], [29]. Almost 10% of EBER-positive cells, all of which stained with CD20 B-cells, were located within the B-cell-rich germinal centers of secondary follicles. Few follicles contained numerous EBER-positive cells. These results are consistent with previous reports of EBER-positive germinal center B-cells in tonsils [30]. It has been suggested that germinal center centroblasts may be a site of EBV persistence [31]. The presence of EBV-positive B-cells in germinal centers is reportedly more common in tonsils from areas endemic for EBV related lymphomas; leading to the suggestion that EBV-infected germinal center B-cells may be more prone to malignant transformation, perhaps due to somatic hypermutation [30]. In this study EBER-positive cells were found within germinal centers of tonsils and adenoids, and some of these cells were atypical and large and might be a precursor for malignant transformation. Although the pattern of distribution of EBER-positive B lymphocytes in the interfollicular regions is random, they have 3 distinct patterns within lymphoid follicles; either within germinal center cells, or mantle zone cells or in the interphase between them. The interphase pattern is very interesting and involves some follicles which have one or more EBER –positive atypical cell within the germinal center. The interphase EBER-positive cells are distributed circumferentially in a round shape at the interphase between germinal center cells and mantle zone cells (Fig. 1C and D) as if they were generated from cell division of EBV-infected germinal center cells and moving outside the lymphoid follicle. Whether this pattern has any role in EBV-persistent infection in tonsils and adenoids or carrying any risk for future malignant transformation needs to be determined in future studies.

1C and D) as if they were generated from cell division of EBV-infected germinal center cells and moving outside the lymphoid follicle. Whether this pattern has any role in EBV-persistent infection in tonsils and adenoids or carrying any risk for future malignant transformation needs to be determined in future studies. On the other hand, EBER was not identified within T-lymphocytes when we use triple staining for EBER, CD20 and CD3 (Fig. 4). Hudnall et al. [22], shows that EBV is rarely found in T lymphocytes in EBV-infected tonsils. There is some controversy regarding the role of oropharyngeal epithelial cells in EBV infection. It has been suggested that primary and persistent EBV infection might be mediated through oropharyngeal epithelial cells [32]. The support for this perception has come mainly from infection detected in desquamated oropharyngeal epithelial cells from patients with infectious mononucleosis by in situ hybridization [33]. However, there is now increasing evidence pointing to B lymphocytes as the likely site of persistent EBV infection, and also as the possible target of primary EBV infection [34]. Several studies have demonstrated EBV replication within the upper epithelial cell layers in oral hairy leukoplakia and are not accompanied by detectable latent infection of basal epithelial cells [35], [36]. Thus, these results do not support the idea that EBV persists in epithelial cells. In our study we do not identify EBV in epithelial cells of tonsils and adenoids whereas it is only detected in B lymphocytes. We think that EBV infection might be mediated through oropharyngeal epithelium at early stages of acute infection but later on, the virus might have left epithelial cells or the infected epithelial cells died as a result of host antiviral immunological reaction, hence we do not have EBV persistence and latency within epithelial cells in our samples. On the other hand, we believe that B lymphocytes are sites of primary infection and latency in tonsils and adenoids.

left epithelial cells or the infected epithelial cells died as a result of host antiviral immunological reaction, hence we do not have EBV persistence and latency within epithelial cells in our samples. On the other hand, we believe that B lymphocytes are sites of primary infection and latency in tonsils and adenoids. In conclusion, EBV is associated with tonsillar hypertrophy and is prevalent in 43% of our cases. EBV is only detected in B lymphocytes and we believe that B lymphocytes are sites of primary infection and latency. In situ hybridization is the gold standard for the detection of EBV in tissue.

1 Introduction Acute upper respiratory infections (URI) are the most common acute diseases in children. Large community studies conducted decades ago already showed that the mean annual number of acute respiratory infections is as high as 5 in children less than 5 years of age and about 3 in older children [1], [2], [3], [4]. Respiratory infections cause significant economic losses. In the United States, acute respiratory infections in 1998 resulted in an estimated 84 million visits to physicians; of which 25 million were due to upper respiratory infection and 13 million due to acute otitis media (AOM) [5]. Every fourth respiratory infection results in a visit to a physician; in infants, the proportion is rising to approximately one-half of all respiratory infections [6]. In children attending day-care centers in Finland, infectious diseases caused more than 90% of total costs of illness (deficient utilization of day-care centers, parents’ lost working capacity, hospitalization, visits to physician, antibiotics) [7]. Viral respiratory infections can also lead to bacterial diseases, and mixed viral–bacterial infections are often associated with antibiotic treatment failure [8]. Previously, viral diagnoses and etiologic studies of URI have had little relevance to the clinical management of individual cases. However, prevention and therapies for viral infections are developing, and both vaccines and antivirals for respiratory viruses other than influenza may be available in the near future [9], [10]. This poses new requirements for viral diagnostics, particularly rapid and easy detection of respiratory viruses.

dual cases. However, prevention and therapies for viral infections are developing, and both vaccines and antivirals for respiratory viruses other than influenza may be available in the near future [9], [10]. This poses new requirements for viral diagnostics, particularly rapid and easy detection of respiratory viruses. 2 Viral etiology of upper respiratory infections In the 1960s, the development of virus detection methods made it possible to identify most of the known respiratory viruses or virus groups by cell culture or serologic methods. Over the years, the reported proportion of virus-positive respiratory infections has increased due to improved detection methods, e.g. from 22% (cell culture) [3] up to 86% (antigen detection and polymerase chain reaction (PCR) methods) [11]. The proportion of virus-positivity can vary between studies depending on many different factors; the type of samples, detection methods, varying epidemiology of viruses and the study settings all may affect the results [3], [11], [12], [13]. In addition, many viruses have typical annual seasonality (Fig. 1 ) [14].Fig. 1 Monthly occurrence of virus-positive acute otitis media events in young children. Finnish Otitis Media Vaccine Trial, in which children were followed from 2 to 24 months of age.

udy settings all may affect the results [3], [11], [12], [13]. In addition, many viruses have typical annual seasonality (Fig. 1 ) [14].Fig. 1 Monthly occurrence of virus-positive acute otitis media events in young children. Finnish Otitis Media Vaccine Trial, in which children were followed from 2 to 24 months of age. New molecular techniques have changed the picture about viruses in several ways. For example, based on genome organization and sequence similarity the enterovirus genus (including polioviruses, coxsackieviruses and echoviruses) have recently been reclassified and divided into five different species (polioviruses and human enteroviruses A–D) [15], and only in few years many new enterovirus types have been characterized [16], [17]. Also new respiratory viruses have been discovered; e.g. human metapneumovirus [18], two previously unrecognized human coronaviruses [19], SARS-associated coronavirus [20], [21] and human bocavirus [22]. The use of PCR methods as a diagnostic tool for respiratory infections is increasing. PCR methods are generally more sensitive than traditional virus detection methods [23], and this has raised the question of the clinical relevance of virus-positive PCR findings. Asymptomatic viral infections do occur [24] and viral shedding in the nasopharynx may continue for up to 3 weeks after onset of infection [25]. Respiratory viral RNA can be detected relatively often from the nasopharynx of apparently healthy children, but most of the virus-positive findings can be linked to previous or future respiratory symptoms [26].

ccur [24] and viral shedding in the nasopharynx may continue for up to 3 weeks after onset of infection [25]. Respiratory viral RNA can be detected relatively often from the nasopharynx of apparently healthy children, but most of the virus-positive findings can be linked to previous or future respiratory symptoms [26]. 3 Specific features of respiratory viruses There are over 200 different types of viruses that can cause upper respiratory infections in children. Rhinoviruses are the largest group of respiratory viruses, including at least 100 different serotypes. Rhinoviruses are the predominant cause of the common cold all over the world and in all age groups. In a prospective study, 91% of children had antibodies against rhinoviruses by the age of 2 years and 79% of children had experienced culture- or RT-PCR-confirmed rhinovirus infection [27]. Human rhinovirus infections typically occur in the early fall and in the spring. Although rhinoviruses are generally thought to cause only mild common cold, they are also associated with acute lower respiratory tract infections, wheezing, bronchiolitis and pneumonia in children [28], [29]. In addition, rhinoviruses often cause exacerbations of pre-existing airways diseases, such as asthma [29].

ng. Although rhinoviruses are generally thought to cause only mild common cold, they are also associated with acute lower respiratory tract infections, wheezing, bronchiolitis and pneumonia in children [28], [29]. In addition, rhinoviruses often cause exacerbations of pre-existing airways diseases, such as asthma [29]. Enteroviruses, which belong to the same virus family Picornaviridae as rhinoviruses, are very common worldwide, and most primary infections occur in childhood. Often enteroviruses have been thought to cause mild diseases with characteristic signs (e.g. hand, foot, and mouth disease and herpangina) or more severe diseases such as meningitis. However, recently with modern detection methods, it has been shown that enteroviruses are a common cause of upper respiratory infections and acute otitis media in children [11], [14]. Respiratory syncytial virus (RSV) infections are present in all age groups, but they predominate in children and especially in infants. Up to 70% of infants have been reported to be infected with RSV during the first year of life, and the rest are infected by the age of 2 years [30]. RSVs are the leading cause of bronchiolitis and acute wheezing in young children [31]. Typically, RSV epidemics occur during the winter months, sometimes starting in the late fall and continuing until early spring. In Finland, RSV epidemics occur every 2 years, and during the epidemic year 2 separate epidemic peaks can be seen [13]. The pattern of 2-year epidemics is different from that in, for example, France [32].

SV epidemics occur during the winter months, sometimes starting in the late fall and continuing until early spring. In Finland, RSV epidemics occur every 2 years, and during the epidemic year 2 separate epidemic peaks can be seen [13]. The pattern of 2-year epidemics is different from that in, for example, France [32]. Human metapneumovirus is a recently discovered respiratory virus [18]. The clinical symptoms caused by metapneumovirus are reported to be similar to those due to RSV, and this virus is an important cause of lower respiratory infections and wheezing in children [33]. Metapneumovirus infections occur mainly between winter and early spring [18], [34]. There are three types of influenza viruses in humans, types A, B and C. However, most data about influenza viruses are for types A and B because, due to detection difficulties, influenza C viruses have not been included in the studies. Influenza type A and B cause infections that can range from asymptomatic infections and common colds to serious illnesses with systemic complications, such as pneumonia [35], [36]. Influenza A and B infections typically occur in a seasonal pattern and in winter epidemics, which have variable intensities.

n the studies. Influenza type A and B cause infections that can range from asymptomatic infections and common colds to serious illnesses with systemic complications, such as pneumonia [35], [36]. Influenza A and B infections typically occur in a seasonal pattern and in winter epidemics, which have variable intensities. In addition, many other viruses or virus groups cause respiratory infections in children; e.g. parainfluenza viruses can cause a broad spectrum of respiratory diseases, ranging from mild upper respiratory infections to pneumonia, but are most often associated with laryngitis [37]. It seems that human coronavirus infections are not very common cause of respiratory infections in young children [38], whereas later in life, clinical and subclinical infections occur more often [39]. Adenoviruses infrequently cause common colds, and respiratory infections caused by these viruses tend to be severe, characterized by high and prolonged fever and strong inflammatory response [40], [41]. Adenoviruses more often cause occasional epidemics in semi-closed communities, such as garrisons or orphanages [42].

sinusitis in children could be considered a natural extension of the common cold. However, evidence of direct viral infection of the maxillary sinus in children is lacking. In adults with acute maxillary sinusitis, a virus can be found in 10% of maxillary secretion samples by viral culture [46] and in 40% by PCR [47]. 4.2 Acute otitis media AOM is one of the most common infectious diseases among children. As many studies during the past decades have shown, viral infection is an important predisposing factor for development of AOM, and viruses are significantly associated with AOMs (Table 2 ) [14], [48], [49], [50], [51], [52], [53], [54], [55], [56], [57]. In a prospective study of children less than 2 years of age, about 43% of all upper respiratory infections were associated with AOM [43]. There is little doubt that viruses have an important role in AOM, but it is not yet clear which factors are crucial for the development of AOM. Does the virus need to invade the middle ear and infect the middle ear mucosa, or would inflammation at the nasopharyngeal end of the Eustachian tube be sufficient to interfere with the innate defences of the middle ear mucosa and facilitate bacterial colonization? Viral infections have been shown to cause dysfunction of the Eustachian tube. Two-thirds of children develop abnormal middle ear pressure when they have a common cold [58]. On the other hand, viral infection might facilitate colonization of the nasopharynx with pathogenic bacteria.Table 2 Selected data from studies of viruses associated with acute otitis media (AOM)

ction of the Eustachian tube. Two-thirds of children develop abnormal middle ear pressure when they have a common cold [58]. On the other hand, viral infection might facilitate colonization of the nasopharynx with pathogenic bacteria.Table 2 Selected data from studies of viruses associated with acute otitis media (AOM) Study No. of children No. of MEF Virus detection methoda Virus infection associated with AOMb (%) Proportion of virus-positive MEF (%) Yoshie (1955) 10 10 Culture, serology 40 40 Grönroos (1964) 322 399 Culture NR 0 Berglund (1966) 27 44 Culture, serology 37 33 Tilles (1967) 90 NR Culture, serology 27 3 Klein (1982) 53 53 Ag 34 25 Chonmaitree (1986) 84 84 Culture 39 20 Sarkkinen (1985) 137 137 Ag 42 18 Pitkäranta (1998) 92 92 RT-PCR 75 48 Heikkinen (1999) 456 815 Culture, Ag, serology 41 17 Chonmaitree (2000) 40 65 Culture, PCR NR 74 Nokso-Koivisto (2004) 940 3210 Ag, RT-PCR 63 38 NR: not reported. a Ag: antigen detection, RT: reverse transcription, PCR: polymerase chain reaction. b Specific virus detected in nasopharyngeal aspirate (NPA) and/or middle ear fluid (MEF) specimen(s), and/or a viral infection documented serologically from paired serum samples.

Study No. of children No. of MEF Virus detection methoda Virus infection associated with AOMb (%) Proportion of virus-positive MEF (%) Yoshie (1955) 10 10 Culture, serology 40 40 Grönroos (1964) 322 399 Culture NR 0 Berglund (1966) 27 44 Culture, serology 37 33 Tilles (1967) 90 NR Culture, serology 27 3 Klein (1982) 53 53 Ag 34 25 Chonmaitree (1986) 84 84 Culture 39 20 Sarkkinen (1985) 137 137 Ag 42 18 Pitkäranta (1998) 92 92 RT-PCR 75 48 Heikkinen (1999) 456 815 Culture, Ag, serology 41 17 Chonmaitree (2000) 40 65 Culture, PCR NR 74 Nokso-Koivisto (2004) 940 3210 Ag, RT-PCR 63 38 NR: not reported. a Ag: antigen detection, RT: reverse transcription, PCR: polymerase chain reaction. b Specific virus detected in nasopharyngeal aspirate (NPA) and/or middle ear fluid (MEF) specimen(s), and/or a viral infection documented serologically from paired serum samples. In studies using antigen detection, RSV has usually been the most common virus associated with AOM, and RSV has been suggested to be one of the most potent viruses to cause AOM since it is most often detected concurrently in the nasopharyngeal aspirate and middle ear fluid (MEF) during AOM [14], [53]. The development of the PCR method enables detection of respiratory viruses for which antigen detection tests have neither been suitable nor available. By PCR, respiratory viruses have been found to be associated with over 70% of AOM events [51], [55] and by this sensitive method rhinoviruses have been the most common virus to be detected in association with AOM in children [14], [55].

viruses for which antigen detection tests have neither been suitable nor available. By PCR, respiratory viruses have been found to be associated with over 70% of AOM events [51], [55] and by this sensitive method rhinoviruses have been the most common virus to be detected in association with AOM in children [14], [55]. According to recent studies respiratory virus can be detected in more than one-half on AOMs, in 19% of cases virus was the only pathogen and in 25% no pathogen was detected [59]. Ongoing development of detection methods will probably diminish the proportion of pathogen-negative AOMs since bacterial culture techniques are not sensitive enough for low bacterial concentrations [60], and all respiratory viruses are not included in the detection panels. It has been suggested that some virus types or species associate more strongly to the development of AOM than others. However, in a recent study of young children no distinct species-specific associations were observed between the viral and bacterial findings at the time of AOM [59]. For a physician, it would be very beneficial to differentiate viral AOM from bacterial disease. However, thus far, viral and bacterial AOM cannot be separated from each other by clinical symptoms or signs. Only a bulging tympanic membrane has been associated with detection of bacteria or bacterial–viral combinations from the MEF during AOM [61], [62].

beneficial to differentiate viral AOM from bacterial disease. However, thus far, viral and bacterial AOM cannot be separated from each other by clinical symptoms or signs. Only a bulging tympanic membrane has been associated with detection of bacteria or bacterial–viral combinations from the MEF during AOM [61], [62]. 4.3 Pharyngitis and tonsillitis Although often antibiotics are described to children with acute pharyngitis or tonsillitis, many studies have shown that a major part of these infections are due to respiratory viruses. Respiratory viruses have been detected in approximately one-third of children with acute pharyngitis; adenoviruses and RSV being the most common viruses to be found [63]. In a study by Chi et al., among 416 young children (the mean age 52.9 months) with acute pharyngitis respiratory viruses were detected in 30% of patients as group A streptococci were isolated in only 2% of patients [64]. Over 40% of acute tonsillitis cases in children are associated with a respiratory virus, and in one-third a virus may be the sole pathogen. Children younger than 3 years of age have rarely bacterial tonsillitis [65], and usually the younger the child the more common is viral etiology of tonsillitis [65]. Adenovirus is the predominant cause of viral tonsillitis, with other common causative agents being Epstein-Barr viruses, influenza viruses and enteroviruses [65], [66].

er than 3 years of age have rarely bacterial tonsillitis [65], and usually the younger the child the more common is viral etiology of tonsillitis [65]. Adenovirus is the predominant cause of viral tonsillitis, with other common causative agents being Epstein-Barr viruses, influenza viruses and enteroviruses [65], [66]. 4.4 Lower respiratory tract infections Viruses are common causative agents of lower respiratory tract infections in children [67], with respiratory syncytial virus (RSV), parainfluenza virus type 3, and influenza viruses occurring most frequently [68], [69], [70]. In addition, rhinoviruses have been shown to cause pneumonia and bronchiolitis in infants and children [29]. The recently discovered human metapneumovirus has been reported to induce RSV-like respiratory infections [18], and this virus also produces lower respiratory tract infections in children [71]. Viral infections have been shown to trigger wheezing and exacerbate asthma in children [72], and virus-induced asthma exacerbations may be severe and increase the need of hospitalization [73]. In addition, there is evidence that RSV and especially rhinovirus-induced wheezing in infancy poses increased risk of childhood asthma even until the teenage years [74]. Acute laryngitis is a viral disease most frequently caused by parainfluenza viruses, and less frequently by RSV and influenza viruses [31], [75].

73]. In addition, there is evidence that RSV and especially rhinovirus-induced wheezing in infancy poses increased risk of childhood asthma even until the teenage years [74]. Acute laryngitis is a viral disease most frequently caused by parainfluenza viruses, and less frequently by RSV and influenza viruses [31], [75]. 5 Viral–bacterial interactions Although viruses cause most of the respiratory infections, virus infections can also lead to bacterial infections. It has been recognized that during viral epidemics, e.g. influenza virus epidemics, the incidence of bacterial pneumonia and acute otitis media is increased [8], [76]. Viral infections facilitate bacterial colonization, adherence and translocation through the epithelial barrier of respiratory cells [8]. For example, in an animal model with chinchillas, otitis media developed in 67% of the animals inoculated with Streptococcus pneumoniae and influenza virus type A, whereas otitis media developed in 21% of the animals inoculated with S. pneumoniae alone and in 4% inoculated with influenza virus type A alone [77]. There are multiple potential mechanisms, which lead to increased bacterial adherence to respiratory epithelial cells during viral respiratory disease [8]. Firstly, viral infection causes physical damage to respiratory tract epithelium leading to impaired local defence mechanisms (e.g. loss of cilia and impaired function of Eustachian tube) and basement membrane exposure. Common respiratory viruses are able to cause defects in both cell ciliary structure and function, leading to loss of cilia and loss of ciliated cells from respiratory epithelium. Viral respiratory infections are also known to impair the cough reflex, which can together with the impaired ciliary function lead to accumulation of secretions and to an increased risk for bacterial superinfection [78]. In addition, viral infection in the nasopharynx causes inflammation in the Eustachian tube, which will interfere with the clearing functions of the middle ear mucosa, and hence, facilitate bacterial growth there. Secondly, during virus infection bacterial adherence to epithelium is increased by viral glycoproteins expressed on the host cell, virus-induced changes on host cell membrane and proteins absorbed from extracellular matrix. Inflammatory response to viral infection may up-regulate expression of molecules that bacteria utilize as receptors [8], [79].

on bacterial adherence to epithelium is increased by viral glycoproteins expressed on the host cell, virus-induced changes on host cell membrane and proteins absorbed from extracellular matrix. Inflammatory response to viral infection may up-regulate expression of molecules that bacteria utilize as receptors [8], [79]. Mixed viral–bacterial respiratory infections are relatively common. On average one-fourth of children with community-acquired pneumonia have mixed viral–bacteria infection, and mixed infections seem to be especially common in children less than 2 years of age [78]. In a recent study, viruses and bacteria were detected simultaneously in more than one-third of children with acute otitis media [59]. Mixed viral–bacterial infections often confuse the clinical picture of the disease; prolongation of viral disease may be due to bacterial co-infection or failure of antibiotic treatment of bacterial disease may be due to viral co-infection. For example, viral infection has been suggested to cause prolonged symptoms of AOM [80], and rhinovirus-positive culture in MEF has been associated with persistence of bacteria or occurrence of new bacteria in the MEF [81]. Mixed viral–bacterial infections are also associated with antibiotic treatment failure [8]. It has been reported that in children even with adequate drug compliance virus infection was associated with an increased risk of bacteriologic failure of treated AOM [82], and that penetration of amoxicillin to MEF is lower in children with viral infection [83]. Controversial results have also been published, including a report that the duration of symptoms was longer in children with no detected pathogens in MEF than in children with bacteria or virus [50]. Acute viral tonsillitis or pharyngitis can also lead to a bacterial disease. However, it seems that especially in children, most often viruses or bacteria solely cause acute tonsillitis and mixed viral–bacterial infections are rare [65].

en with no detected pathogens in MEF than in children with bacteria or virus [50]. Acute viral tonsillitis or pharyngitis can also lead to a bacterial disease. However, it seems that especially in children, most often viruses or bacteria solely cause acute tonsillitis and mixed viral–bacterial infections are rare [65]. 6 Prevention and treatment of viral respiratory infections Amantadine and the related drug rimantadine can be used for treatment of influenza A virus infections. The newer group of drugs are neuraminidase inhibitors zanamivir and oseltamivir, which are effective against influenza viruses A and B [84]. Both inactivated and live attenuated vaccines have been developed for influenza viruses, but so far in annual vaccinations almost exclusively the intramuscularly administered inactivated vaccines have been used. The intranasally administered influenza vaccine contains live attenuated, cold-adapted influenza A and B viruses, and the vaccine has been approved to be marketed in the United States. A novel adjuvanted inactivated influenza vaccine administered intranasally has been associated with development of Bell's palsy [85]. It has been suggested that the Escherichia coli enterotoxin adjuvant in the vaccine might have been the risk-inducing factor, and that an induced response, rather than some direct toxic effect, led to the palsy [86].

influenza vaccine administered intranasally has been associated with development of Bell's palsy [85]. It has been suggested that the Escherichia coli enterotoxin adjuvant in the vaccine might have been the risk-inducing factor, and that an induced response, rather than some direct toxic effect, led to the palsy [86]. Rhinoviruses are the most common cause for URI in children and therefore prevention or treatment of rhinovirus infections would be the most beneficial. Due to its many different serotypes, developing a vaccine for rhinoviruses is unlikely. Treatment of rhinovirus disease is also problematic because infection often proceeds quickly and medication should be started as soon as the first symptoms occur or shortly thereafter. A capsid-binding agent, pleconaril, also effective against enteroviruses, has been the most promising drug to date and has demonstrated that developing a treatment for the main causes of the common cold is possible [87]. Unfortunately, since further studies revealed that pleconaril might have potential drug interactions, regulatory approval was not granted for the oral formulation of the drug. Other different treatment methods have been investigated, e.g. intranasal interferon [88], intranasally administered receptor decoy soluble intercellular adhesion molecule-1 (tremacamra), 3C protease inhibitors rupintrivir and pyridone, and oral anti-picornavirus capsid-binder pirodavir [10], [87]. However, none of these medications are thus far in clinical use.

investigated, e.g. intranasal interferon [88], intranasally administered receptor decoy soluble intercellular adhesion molecule-1 (tremacamra), 3C protease inhibitors rupintrivir and pyridone, and oral anti-picornavirus capsid-binder pirodavir [10], [87]. However, none of these medications are thus far in clinical use. Ribavirin has been used as a specific antiviral treatment for RSV infections. It is recommended as a possible treatment only for a selected group of infants at high risk for serious RSV disease [30]. Intravenous, enriched RSV immunoglobulin can be used as prophylaxis for RSV infections in high-risk children or neonates [89]. Prophylactic treatment with palivizumab, a humanized RSV monoclonal antibody, during the epidemic season has been reported to be associated with a decreased rate of hospitalization for RSV lower respiratory tract infections [90]. VP14637 is a new anti-RSV fusion inhibitor that is much more effective than ribavirin in in vitro tests [10]. Candidate vaccines against RSV infections are under development in clinical studies [91]. Live, enteric-coated adenovirus type 4 and 7 vaccines have been developed and used, but the sole manufacturer of these vaccines ceased production, and thus these vaccines are no longer available [4]. Live attenuated human parainfluenza virus 3 vaccine has been studied, and in a clinical trial, it has appeared to be safe and efficient in children [92].

ype 4 and 7 vaccines have been developed and used, but the sole manufacturer of these vaccines ceased production, and thus these vaccines are no longer available [4]. Live attenuated human parainfluenza virus 3 vaccine has been studied, and in a clinical trial, it has appeared to be safe and efficient in children [92]. 7 Conclusion Viral respiratory infections are very common in children and are an enormous burden to the families and society. Present knowledge of the natural course and etiology of respiratory infections mainly comes from large community studies conducted several decades ago. Although study settings and virus detection methods have varied between different studies, general characteristics and the main causative viral agents have remained similar over the years. However, recent development of modern microbiologic methods is also changing the world of virology; the knowledge of the respiratory viruses is broadening and even new viruses are being found [19]. Epidemiologic studies of respiratory viruses are important because of prevention and treatments for virus infections are emerging. In addition, new treatment possibilities are posing requirements for fast and sensitive diagnostic methods for respiratory viruses.

is broadening and even new viruses are being found [19]. Epidemiologic studies of respiratory viruses are important because of prevention and treatments for virus infections are emerging. In addition, new treatment possibilities are posing requirements for fast and sensitive diagnostic methods for respiratory viruses. As the use of the very sensitive PCR-based detection method has increased, the question has been posed as to whether PCR methods are too sensitive, detecting small remnants of viruses that probably have no clinical relevance. However, according to recent studies, the vast majority of viruses detected by PCR in patients can be assumed to be involved in the initiation of the observed respiratory infection [26], [93]. Detection of a respiratory viral nucleic acid by PCR in the nasopharynx must also be interpreted cautiously because the presence of a virus alone does not establish causality of the concurrent illness. The introduction of PCR methods into clinical practice has resulted in the same questions and problems of clinical relevance as for other diagnostic methods. Laboratory findings must always be interpreted in association with clinical signs and symptoms.

irus alone does not establish causality of the concurrent illness. The introduction of PCR methods into clinical practice has resulted in the same questions and problems of clinical relevance as for other diagnostic methods. Laboratory findings must always be interpreted in association with clinical signs and symptoms. Antibiotics are commonly used to treat AOMs, although a part of AOMs in children could resolve without treatment. Viral respiratory infection precedes the development of AOM, and respiratory virus can be detected in more than one-half on AOMs [14], [51], [55]. Especially rhinovirus, enterovirus and RSV have been found to associate in significant proportion of AOMs in young children [14]. In the future, these virus groups most commonly associated with AOM should be considered when aiming at prevention and treatment of AOM in children.