Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Risk Factors of Early Otitis Media in the Danish National Birth Cohort

  • Asbjørn Kørvel-Hanquist ,

    Asbp@regionsjaelland.dk

    Affiliation Department of Otorhinolaryngology and Maxillofacial Surgery, Zealand University Hospital, Køge, Denmark

    ORCID http://orcid.org/0000-0002-6645-1926

  • Anders Koch,

    Affiliations Department of Epidemiology Research, Statens Serum Institut, Copenhagen, Denmark, Department of Infectious Disease Epidemiology, Statens Serum Institut, Copenhagen, Denmark

  • Janni Niclasen,

    Affiliations Department of Psychology, University of Copenhagen, Copenhagen, Denmark, Center for Collaborative Health, Aarhus University, Aarhus, Denmark

  • Jesper Dammeye,

    Affiliation Department of Psychology, University of Copenhagen, Copenhagen, Denmark

  • Jørgen Lous,

    Affiliation Research Unit for General Practice Institute of Public Health, University of Southern Denmark, Odense, Denmark

  • Sjurdur Frodi Olsen,

    Affiliation Department of Epidemiology Research, Centre for Fetal Programming, Statens Serum Institut, Copenhagen, Denmark

  • Preben Homøe

    Affiliation Department of Otorhinolaryngology and Maxillofacial Surgery, Zealand University Hospital, Køge, Denmark

Risk Factors of Early Otitis Media in the Danish National Birth Cohort

  • Asbjørn Kørvel-Hanquist, 
  • Anders Koch, 
  • Janni Niclasen, 
  • Jesper Dammeye, 
  • Jørgen Lous, 
  • Sjurdur Frodi Olsen, 
  • Preben Homøe
PLOS
x

Correction

6 Feb 2017: Kørvel-Hanquist A, Koch A, Niclasen J, Dammeyer J, Lous J, et al. (2017) Correction: Risk Factors of Early Otitis Media in the Danish National Birth Cohort. PLOS ONE 12(2): e0171901. https://doi.org/10.1371/journal.pone.0171901 View correction

Abstract

Objective

To assess risk factors of otitis media (OM) in six-months-old children.

Method

The sample consisted of 69,105 mothers and their children from the Danish National Birth Cohort. The women were interviewed twice during pregnancy and again 6 months after birth. The outcome “one or more” maternal reported episodes of OM at age six months. In total 37 factors were assessed, covering prenatal, maternal, perinatal and postnatal factors.

Results

At age six months 5.3% (95% CI 5.1–5.5) of the children had experienced one or more episodes of OM. From the regression analysis, 11 variables were associated with a risk of OM. When a Bonferroni correction was introduced, gender, prematurity, parity, maternal age, maternal self-estimated health, taking penicillin during pregnancy, and terminating breastfeeding before age six months, was associated with a risk of early OM. The adjusted ORs of OM for boys versus girls was 1.30 (95% CI 1.18–1.44). The OR having one sibling versus no siblings was 3.0 (95% CI 2.64–3.41). If the woman had been taking penicillin during pregnancy, the OR was 1.35 (95% CI 1.15–1.58). Children born before 38th gestational week had an increased OR for early OM of 1.49 (95% CI 1.21–1.82). Children of young women had an increased OR of early OM compared to children of older women. Additionally, children of women who rated their own health low compared to those rating their health as high, had an increased OR of 1.38 (95% CI 1.10–1.74). Finally, children being breastfeed less than 6 months, had an increased OR of 1.42 (95% CI 1.28–1.58) compared to children being breastfeed beyond 6 months.

Conclusion

These findings indicate that prenatal factors are of less importance regarding early OM before the age of six months. Postnatal risk factors seem to pose the main risk of early OM.

Introduction

Otitis media (OM) is one of the most common infections in early childhood and may be regarded as a complication of upper respiratory tract infections[1]. OM covers a diverse group of diseases, which contain some degree of inflammation or infection located in the middle ear and the tympanic membrane. Also, OM is the primary cause of narrow and broad spectrum antibiotic prescriptions for children in western countries[2,3]. At introduction of the conjugate pneumococcus vaccination the incidence of OM was expected to decline, however the decline in otitis media due to introduction of the pneumococcal vaccination has only been modest, in addition no decline in prescription of antibiotics has been observed, at least not in Denmark[4,5]. It is estimated that 60–70% of all children will experience one or more episodes of OM during childhood, and approximately 24% will experience more than 3 episodes of OM during childhood [6,7]. As a consequence, OM remains a significant childhood disease and can thus be considered a substantial ecological and economical health burden[810].

In an attempt to assess child susceptibility to recurrent episodes of OM, studies have found early-onset OM to be an important risk factor for recurrent otitis media(ROM)[1116]. Studies indicate that recurrent childhood OM have prolonged effects on hearing, speech development, behavior and school performance [1723]. Successful delay in onset of the first episode of OM could therefore decrease the frequency of ROM and later health and developmental consequences due to early OM and ROM.

Few studies have evaluated prenatal risk factors. So far, a small number of studies have reported significant associations between smoking during pregnancy, alcohol consumption during pregnancy, gestational use of medication, maternal exposure to dioxin-like compounds, and exposure to organochlorines and early OM or ROM [7,2429]. Many existing studies however, are not conclusive regarding which prenatal and environmental factors contributes to increased risk of early OM. Previous studies may suffer from insufficient sample size and lack of some relevant variables [30,31].

The objectives of the present study were to examine environmental risk factors of early OM in offspring using data from a large national cohort. These environmental risk factors included socioeconomic factors, dietary habits, smoking, alcohol consumption, disease during pregnancy and use of medicine during pregnancy. Peri- and postnatal factors have also been taken into account.

Materials and Methods

The present study was based on data from the Danish National Birth Cohort (DNBC) collecting baseline data between 1996 and 2002. The women were recruited at their first antenatal visit with their GP. When the women were enrolled they were offered four interviews, two prenatally at approximately weeks 12 and 30, and twice after delivery at age six and 18 months of age of their child[32].

In this study it was not possible to differentiate between different kinds of early OM as it depends on parent-reported outcome measures. Physicians often do not distinguish between different infections in the middle ear, hence it was not expected that parents were able to separate one kind of middle ear infection from another.

Settings

The DNBC is a national multipurpose birth cohort. Approximately 60% of the invited pregnant women accepted participation. A total of 100,329 pregnancies were enrolled in the cohort equivalent to a participation rate of 30% of all eligible Danish pregnant women [33]. The same woman could be represented with more than one pregnancy, if she had multiple pregnancies during the enrollment period. The cohort has been described in more details elsewhere[32].

The DNBC has been approved by the Committee on Biomedical Research Ethics (KF) 01-471/94.

Exclusion/Inclusion parameters

A complete case analysis was carried out, in which all participants with full information on OM from the third interview (six months of age) were included. Exclusion criteria were: children born premature before gestational week 34 because of increased incidence of infections[3436]; children with lip-palate malformation due to cranial deformation resulting in increased OM incidence[37], and Down syndrome (Fig 1).

thumbnail
Fig 1. Flowchart of subject participation, loss to follow-up and exclusion in the Danish National Birth Cohort.

https://doi.org/10.1371/journal.pone.0166465.g001

Exposure variables.

Based on the literature, 37 variables were selected for the analyses and risk estimations, including pre- and postnatal risk factors. The variables were selected, if they were suspected to influence the health of the newborn child through epigenetic changes, immunological changes or if they were known to influence the risk of OM. Some were not environmental factors, but were included because they were important for adjustment in the multiple logistic risk estimates.

Outcome variables.

The primary outcome was maternal-reported episodes of OM within the first six months. In the analyses, the outcomes were dichotomized. The outcome groups were “no OM” and “1 or more episodes of OM” at age six months of the child. If the woman had completed the interview, but items from the interview regarding OM were missing or otherwise inconclusive, the woman and her child were defined as having missing data and included in the analyses as “missing”.

Statistical methods

All analyses were performed using SAS 9.4 software (SAS institute Inc. Cary, North Carolina). We performed Chi-square tests to examine the background characteristics of already known and potential risk factors for the “no OM” and the “1 or more OM” group. Significance level for the Chi-square test was set at the 5% level. To estimate an adjusted odds ratio, a multiple logistic regression analysis was performed. Bonferroni correction was used to estimate an adjusted p-value for the multiple logistic model, to counteract the effect of multiple testing.

To make the logistic model we firstly tested related variables for internal correlation. Variables were grouped with related variables in an attempt to find the most important exposures and possible latent variables (Table 1). Each group of variables was tested against the outcome in a multiple logistic regression model. If an exposure variable was found to be significantly associated with the outcome in the logistic model with a group of related variables, it was included in the final multiple logistic regression model as an independent variable. The final model included 22 different exposure variables. To test the strength of the results, descriptive statistics at baseline for those attending complete follow-up compared to those not-attending follow-up were made using Chi-square test.

thumbnail
Table 1. Chosen exposure variables shown in groups for internal correlation analyses.

https://doi.org/10.1371/journal.pone.0166465.t001

Results

Of the cohort of 100,329 pregnancies, a total of 94,745 (94.4%) pregnancies resulted in the birth of a live born child. Between birth and the first postnatal interview 24,619 (24.5%) were lost during follow-up. Only three women formally withdrew from the cohort, the rest were non-responders on the exposure items or the item regarding OM. We excluded 973 (1.0%) from the study cohort due to the exclusion criteria. Final cohort for analysis included a total of 69,105 pregnancies (68.9%) of the total cohort (Fig 1).

At age six months 5.3% (95% confidence interval 5.1% - 5.5%) of the children had suffered from at least one episode of early OM. A number of variables were unequally distributed between the “no OM” and the “1 or more OM” groups. These variables (displayed in Table 1) were tested for internal-correlation, with the exception of gender that was included without being tested for internal correlation. Only the variables singleton versus plural birth, paracetamol, vitamins, fish-oil and being vaccinated were not associated with change in risk of early OM in the crude univariate analysis (Table 2).

thumbnail
Table 2. Prevalence, characteristics and crude odds ratio (OR) of exposure for early otitis media within first 6 months in the Danish National Birth Cohort.

https://doi.org/10.1371/journal.pone.0166465.t002

From the adjusted multiple logistic model we found 11 variables to be associated with increased or decreased risk of early OM (Table 3). Prenatal and maternal characteristics associated with increased or decreased risk of early OM were gender of the child, gestational age at birth, maternal age at birth, if the woman ever had suffered from allergy, weight status before pregnancy, maternal self-estimated health, and use of antibiotics during pregnancy. Taking antibiotics other than penicillin during the period from the 12th until the 30th gestational week was significantly associated with an increased risk of early OM in the adjusted model but not significant after Bonferroni correction. As opposed to this, maternal report of having taken penicillin during the same period from the 12th until the 30th gestational week was significantly related to an increased risk of early OM with odds ratio (OR) at 1.35 (95% confidence interval 1.15–1.58) in the adjusted model corrected with Bonferroni. This would result in a population attributable fraction of 112 children, which could avoid having early OM equal to 3.2% of the early OM cases, if the exposure were removed from the population. The effect on early OM risk from intake of coffee and cola from the crude analysis, were not significant in the adjusted risk estimate. Tea consumption was associated with a decreased risk in the crude analysis. In the adjusted model, only the analysis for trend in the consumption of tea was still significant. In addition, taking iron supplement during the last part of pregnancy tended towards a decreased risk of early OM, however this was also insignificant in the adjusted model.

thumbnail
Table 3. Adjusted odds ratio (OR) for otitis media within the first 6 months in the Danish National Birth Cohort.

https://doi.org/10.1371/journal.pone.0166465.t003

Postnatal factors associated with increased risk of OM were living together with older siblings and terminating breastfeeding before the age of six months. The number of siblings was associated with a 3-fold increased risk of early OM. Early daycare attendance was defined as attending a day care center, with more than five children attending the same center. Early daycare attendance was not significantly associated with early OM, when the Bonferroni correction was introduced. Socioeconomic status did not affect the risk of early OM when adjusted for the rest of the risk variables (Table 3). In a test of strength and potential bias between responders and non-responders differences, it was found that birthweight, gestational age at birth, parity, socioeconomic status, single parenthood, maternal smoking and maternal alcohol consumption during pregnancy, all were unequally distributed. Among non-responders, more women were smoking between 1 and 10 cigarettes/day (67% versus. 11% in responders), and more women were smoking more than 10 cigarettes/day (25% versus 5% in responders). Among responders, the frequency of women living without another adult in the household, were higher that among non-responders (2% versus 0%) (Table 4).

thumbnail
Table 4. Prevalence and characteristics of responders and non-responders in the DNBC.

https://doi.org/10.1371/journal.pone.0166465.t004

Discussion

In line with the objectives, this study investigated the effect of a variety of risk exposures for OM at six months of age. The study showed that risk of maternal reported early OM was increased in children born by women who had taken penicillin during pregnancy. Other prenatal factors as male gender, young maternal age, low maternal self-estimated health and low gestational age at birth was also associated with increased risk of early OM. After birth, the most important factors associated with increased risk of early OM before the age of six months in the adjusted multiple logistic model were having siblings and early termination of breastfeeding.

Limitations

During follow-up 11,683 eligible mothers and their children were non-responders. Because decision on not to participate in the follow-up was made by the woman, on behalf of the child, characteristics of non-responders reflect maternal characteristics. The test for strength of the results pointed at this tendency. Women not attending follow-up were more prone to be smoking. In addition, child characteristics did not seem to be associated with not attending follow-up, see Table 4.

Additional analyses to assess the effect of missing data on exposure variables were not performed. Although we assume that the information on exposure were “missing completely at random” or “missing at random” and imputation would have been possible, we did not replace the missing data[38]. Furthermore, the present study did not assess interactions between variables, although interactions between exposure variables might have affected the associations found.

A major limitation in this study was the use of parental reported disease outcome. This limitation is described later in more detail.

Strengths

Analyses were based on 69,105 children. This is more than most other studies investigating diet and environmental effect on risk of early OM. Prior studies have been criticized for not sufficiently adjusting for relevant social factors[30,31]. The large size cohort with detailed information on maternal factors and information on life during pregnancy made it possible to include a wide range of socioeconomic confounding variables. Additionally, the present study included 14 relevant confounders to separate the effect of different exposure behaviors.

Bias.

The present study was susceptible to different sorts of potential bias.

The data were collected by telephone, and the aim of the study was unknown to the interviewers. The cohort was a multipurpose cohort, and the outcome was therefore independent of the interviewer and the interviewer’s health beliefs. Therefore, the study was unlikely to suffer from interviewer bias.

During enrollment, only Danish speaking women were eligible for inclusion, because participants needed to have some basic Danish language skills. Excluding immigrant women not capable of speaking Danish potentially introduced selection bias, due to exclusion of a subpopulation prone to marginalization and unemployment. A study by Jacobsen et al. concerning bias from non-responders in the DNBC, reported an underrepresentation in the DNBC by non-academic women, women not working as health professionals and from women being unemployed, single, on high social supports and below the age of 20[39]. Many of these characteristics might be associated with risk of early OM. This could subvert a possible effect of these variables on the risk of early OM and render an impact of the generalizability of our findings[40].

Imprecision.

A possible imprecision on both our outcome and exposure measures arise from recall bias. Items in the two prenatal interviews assessed exposure in the time interval prior to the interviews at 12 weeks and between 12th and 30th gestational week. Exposure measures during pregnancy were believed to be less susceptible to recall bias, as most pregnant women are very much aware of exposures during this period. However, the present prospective cohort design may be vulnerable to bias from differentiated recall of exposures between Women in different socioeconomic or educational strata. We believe that such exposure recall bias generally diminish an existing relation between the exposure being measured and early OM[41]. The outcome measure obtained 30 weeks after birth is more vulnerable to differentiated recall differences between groups because of the longer time period the women have to recall. This could potentially influence the result both positively or negatively. However, we do not believe that the outcome measure is prone to this bias. Todberg et al. explored the internal validity of the items regarding reported episodes of OM and found a high internal correlation between reporting episodes of OM at six month after birth, and reporting OM episodes at 18 month after birth[6]. This supports a statement of only minor influence of bias from recall on our results.

Outcome measure.

Data used in this study was exclusively questionnaire data and primarily of qualitative nature. The outcome measure of maternal-reported OM has been validated in other studies[4246]. Maternal-reported OM was not specified into acute OM, OM with effusion or ROM. Therefore, OM in this study covers the whole spectrum of different middle ear infections and inflammatory diseases.

As in other studies using parent-reported disease measures, this study was affected by imprecision in the outcome information due to mismatch between the true clinical diagnosis and maternal disease reports. Kvestad et al.[46] reported, that inconsistency between parent-reported disease and otological findings, were related to less severe disease. In contrast, additional questions regarding disease incidence and treatment increased accuracy of parent-reported OM[42]. In a study by Alho et al.[43], the parent-reported cumulative incidence between children younger than two years of age, was lower than the incidence found in medical records. Thus, the true incidence of OM can be expected to be even higher than reported in medical records as not all OM episodes results in a medical contact. Therefore, we expect that maternal reported episodes of early OM is an underestimate of the true incidence.

Because information about exposure was collected before birth, the imprecision on early OM reports were considered equally distributed among exposed and non-exposed. Although it was possible that women with good health insight would tend to report the numbers of OM closer to the actual numbers of episodes compared to women without the same health insight. In addition, women with health insight would tend to demonstrate a low-risk behavior during pregnancy. This would draw the risk estimation towards the null.

Exposure measures.

The exposure measures were afflicted with the same imprecision as the outcome measures in form of possible misclassification bias, recall bias and detection bias which could have underestimated subclinical or borderline exposure measures or risk-behavior. In addition, each exposure variable might have its own individual mechanisms of imprecision and bias.

Reports on medicine consumption might be affected by misclassification. Not only were the reports subject to recall bias, but imprecision also increased due to low pharmacological compliance. A study by Olesen et al., estimated the compliance of used antibiotics in the DNBC at 52% [47]. This misclassification might also draw the risk estimation of the effect of penicillin treatment during pregnancy on early age OM towards the null hypothesis of no effect. Furthermore, the exposure for penicillin was afflicted with confounding by indication, as an indication for prescription precede penicillin consumption, thus the effect of penicillin during pregnancy on child’s early OM experience could be a reflection of the true effect from an unknown causal factor.

Comparison

Only a few studies have assessed risk exposure during pregnancy related to early life OM. The overall prevalence of OM before six months of age was estimated to be 5.3% and thus in the lower end compared to a prevalence between 2% and 20% found by Engel and colleagues[42]. Another study that assessed early onset OM in a high-risk population of American Indians found that 63% of the children had experienced at least one episode of OM before the age of six months. This study however, did not use maternal reported OM, but instead clinical findings[15]. Finally, Daly and colleagues estimated the percentage of children who had suffered from episodes of either AOM or OME at age six months in an American birth cohort of 596 participants at 48%[48].

The individual risk factor estimates also differ from some existing publications. The elevated risk found for male children is in line with some of the literature[27,49], although the majority of published studies have been unable to find any association between early age OM and gender[11,15,28,48,50,51].

Furthermore, this study’s findings of peri- and postnatal risk exposures from low birthweight, preterm birth, high parity and early termination of breastfeeding support findings in other studies. Having siblings seems to constitute a considerable risk of acquiring early age OM, and demonstrates a dose-response relationship. In the study by Daly et al. the relative risk (RR) of AOM at six months was 1.2 having one sibling compared to no siblings. The RR was 1.4 having two or more siblings in the household versus no siblings[14]. Also risk modification of early termination of breastfeeding resemble results from earlier studies[11,52].

Smoking during pregnancy and early OM have been carefully examined in previous studies. We did not include smoking during pregnancy in our final model, as the effect of smoking was closely correlated with other social variables, and the independent effect of smoking during pregnancy became insignificant, when other socioeconomic variables were included in the final model. However, our crude risk estimate of OR equal to 1.23 (95% confidence interval 1.12–1.37) for smoking less than 10 cigarettes/day, is in accordance with results from Haaberg and colleagues who reported a 1.27 crude OR for children having experienced one or more reported episodes of acute OM before the age of six months[53]. Lieu et al. estimated an adjusted OR of the child having OM before six month of age at 1.08 (95% confidence interval 1.02–1.14) if the woman had been smoking during pregnancy[7]. However, a meta-analysis of six studies on prenatal smoking exposure, found no significant relation between maternal smoking during pregnancy and middle ear diseases in general[29].

Alcohol consumption was not associated with an increase in reported early OM in our study. This finding is in line with other studies on effect of alcohol during pregnancy and early OM[15]. In a study from 2014, Brennan-Jones et al. found an increased OR of parent-reported ROM at three years after birth in children of women drinking alcohol during pregnancy[28]. However, when controlling for social variables in the study, the risk became insignificant.

Maternal use of penicillin during pregnancy and risk of OM in offspring have to our knowledge not been studied before. We found an OR of 1.35 in risk of early OM (Table 3). Whether this finding is due to penicillin use itself or an underlying infection during pregnancy cannot be determined from this study.

Generalizability

Although this study reflects an era before introduction of the pneumococcal vaccination in 2007 in Denmark, children before the age of six months are not fully protected by the vaccine[5456]. Hence, these results can be extrapolated to most western societies and health systems. Still many risk exposures may differ between rural and urban populations, and may not be applicable in a homogenous subpopulation with different characteristics compared to the present studied population.

Conclusion

The present study evaluated already known risk factors together with new potential factors not previously examined in relation to early OM. We did not in this large study find prenatal factors that could be used as predictors for the risk of early OM in childhood. However, we found a number of maternal characteristics as maternal age at birth, maternal self-estimated health and early life variables as low gestational age, high parity, and length of breastfeeding that predicts a higher risk of early OM. This knowledge may be important to parents and physicians in their speculations of the reasons why some children attracts early OM.

Acknowledgments

The Danish National Research Foundation has established the Danish Epidemiology Science Centre that initiated and created the Danish National Birth Cohort. The cohort is furthermore a result of a major grant from this Foundation. Additional support for the Danish National Birth Cohort is obtained from the Pharmacy Foundation, the Egmont Foundation, the March of Dimes Birth Defects Foundation, the Augustinus Foundation, and the Health Foundation.

Author Contributions

  1. Conceptualization: AK JL SFO PH AKH.
  2. Data curation: SFO.
  3. Formal analysis: AK AKH PH.
  4. Investigation: AK JL SFO PH.
  5. Methodology: AK JL SFO PH AKH JN JD.
  6. Supervision: AK JL SFO PH.
  7. Validation: AK PH AKH JN.
  8. Writing – original draft: AK JL PH AKH JN JD.
  9. Writing – review & editing: AK JL SFO PH AKH JN JD.

References

  1. 1. Chonmaitree T, Revai K, Grady JJ, Clos A, Patel JA, Nair S, et al. Viral Upper Respiratory Tract Infection and Otitis Media Complication in Young Children. Clin Infect Dis. 2008;46: 815–823. pmid:18279042
  2. 2. Sarpong EM, Miller GE. Narrow- and Broad-Spectrum Antibiotic Use among U.S. Children. Health Serv Res. 2015;50: 830–846. pmid:25424240
  3. 3. Otters HBM. Trends in prescribing antibiotics for children in Dutch general practice. J Antimicrob Chemother. 2004;53: 361–366. pmid:14729760
  4. 4. Fortanier AC, Venekamp RP, Boonacker CW, Hak E, Schilder AG, Sanders EA, et al. Pneumococcal conjugate vaccines for preventing otitis media. In: The Cochrane Collaboration, editor. Cochrane Database of Systematic Reviews. Chichester, UK: John Wiley & Sons, Ltd; 2014. Available: http://doi.wiley.com/10.1002/14651858.CD001480.pub4
  5. 5. Pottegård A, Broe A, Aabenhus R, Bjerrum L, Hallas J, Damkier P. Use of Antibiotics in Children: A Danish Nationwide Drug Utilization Study. Pediatr Infect Dis J. 2015;34: e16–e22. pmid:25144795
  6. 6. Todberg T, Koch A, Andersson M, Olsen SF, Lous J, Homøe P. Incidence of Otitis Media in a Contemporary Danish National Birth Cohort. Schooling CM, editor. PLoS ONE. 2014;9: e111732. pmid:25545891
  7. 7. Lieu JEC, Feinstein AR. Effect of gestational and passive smoke exposure on ear infections in children. Arch Pediatr Adolesc Med. 2002;156: 147–154. pmid:11814376
  8. 8. Jackson C, Chandler W. Households’ unpaid work: measurement and valuation. Ottawa: Statistics Canada; 1995.
  9. 9. Speets AM, Wolleswinkel JH, Forsgren A, Sobocki PA. Use of medical resources and indirect costs of otitis media in Sweden. Scand J Public Health. 2011;39: 137–146. pmid:21257647
  10. 10. Ventola CL. The Antibiotic Resistance Crisis: Part 1: Causes and Threats. Pharm Ther. 2015;40: 277.
  11. 11. Pettigrew MM, Gent JF, Triche EW, Belanger KD, Bracken MB, Leaderer BP. Association of early-onset otitis media in infants and exposure to household mould. Paediatr Perinat Epidemiol. 2004;18: 441–447. pmid:15535820
  12. 12. Lasisi AO, Olayemi O, Irabor AE. Early onset otitis media: risk factors and effects on the outcome of chronic suppurative otitis media. Eur Arch Otorhinolaryngol. 2008;265: 765–768. pmid:18046567
  13. 13. Teele DW, Klein JO, Rosner B. Epidemiology of otitis media during the first seven years of life in children in greater Boston: a prospective, cohort study. J Infect Dis. 1989;160: 83–94. pmid:2732519
  14. 14. Daly KA, Brown JE, Lindgren BR, Meland MH, Le CT, Giebink GS. Epidemiology of otitis media onset by six months of age. Pediatrics. 1999;103: 1158–1166. pmid:10353923
  15. 15. Daly KA, Pirie PL, Rhodes KL, Hunter LL, Davey CS. Early otitis media among Minnesota American Indians: the Little Ears Study. Am J Public Health. 2007;97: 317–322. pmid:17194873
  16. 16. Niclasen J, Obel C, Homøe P, Kørvel-Hanquist A, Dammeyer J. Associations between otitis media and child behavioural and learning difficulties: Results from a Danish cohort. Int J Pediatr Otorhinolaryngol. 2016;84: 12–20. pmid:27063746
  17. 17. Aarhus L, Tambs K, Kvestad E, Engdahl B. Childhood Otitis Media: A Cohort Study With 30-Year Follow-Up of Hearing (The HUNT Study). Ear Hear. 2014; Available: http://pdfs.journals.lww.com/ear-hearing/9000/00000/Childhood_Otitis_Media___A_Cohort_Study_With.99377.pdf
  18. 18. Winskel H. The effects of an early history of otitis media on children’s language and literacy skill development. Br J Educ Psychol. 2006;76: 727–744. pmid:17094883
  19. 19. Hogan A, Phillips RL, Howard D, Yiengprugsawan V. Psychosocial outcomes of children with ear infections and hearing problems: a longitudinal study. BMC Pediatr. 2014;14: 65. pmid:24593675
  20. 20. Yiengprugsawan V, Hogan A, Strazdins L. Longitudinal analysis of ear infection and hearing impairment: findings from 6-year prospective cohorts of Australian children. BMC Pediatr. 2013;13: 28. pmid:23432915
  21. 21. Azab SN, Rhman S, Shokr A. Otitis media: Is precursor of delayed reading in Arabic speaking children? Int J Pediatr Otorhinolaryngol. 2014;78: 670–673. pmid:24582077
  22. 22. Rvachew S, Slawinski EB, Williams M, Green CL. The impact of early onset otitis media on babbling and early language development. J Acoust Soc Am. 1999;105: 467–475. pmid:9921670
  23. 23. Maruthy S, Mannarukrishnaiah J. Effect of early onset otitis media on brainstem and cortical auditory processing. Behav Brain Funct. 2008;4: 17. pmid:18384677
  24. 24. Engel JAM, Straetemans M, Zielhuis GA. Birth characteristics and recurrent otitis media with effusion in young children. Int J Pediatr Otorhinolaryngol. 2005;69: 533–540. pmid:15763293
  25. 25. Miyashita C, Sasaki S, Saijo Y, Washino N, Okada E, Kobayashi S, et al. Effects of prenatal exposure to dioxin-like compounds on allergies and infections during infancy. Environ Res. 2011;111: 551–558. pmid:21324443
  26. 26. Jensen RG, Koch A, Homøe P, Bjerregaard P. Tobacco smoke increases the risk of otitis media among Greenlandic Inuit children while exposure to organochlorines remain insignificant. Environ Int. 2013;54: 112–118. pmid:23434818
  27. 27. Macintyre EA, Karr CJ, Koehoorn M, Demers P, Tamburic L, Lencar C, et al. Otitis media incidence and risk factors in a population-based birth cohort. Paediatr Child Health. 2010;15: 437–442. pmid:21886448
  28. 28. Brennan-Jones CG, Whitehouse AJ, Park J, Hegarty M, Jacques A, Eikelboom RH, et al. Prevalence and risk factors for parent-reported recurrent otitis media during early childhood in the Western Australian Pregnancy Cohort (Raine) Study: Risk factors for otitis media. J Paediatr Child Health. 2014; n/a–n/a. pmid:25303240
  29. 29. Jones LL. Parental Smoking and the Risk of Middle Ear Disease in Children: A Systematic Review and Meta-analysis. Arch Pediatr Adolesc Med. 2012;166: 18. pmid:21893640
  30. 30. Braveman PA, Cubbin C, Egerter S, Chideya S, Marchi KS, Metzler M, et al. Socioeconomic status in health research: one size does not fit all. Jama. 2005;294: 2879–2888. pmid:16352796
  31. 31. Daly KA, Hoffman HJ, Kvaerner KJ, Kvestad E, Casselbrant ML, Homoe P, et al. Epidemiology, natural history, and risk factors: Panel report from the Ninth International Research Conference on Otitis Media. Int J Pediatr Otorhinolaryngol. 2010;74: 231–240. pmid:19836843
  32. 32. Olsen J, Melbye M, Olsen SF, Sørensen TI, Aaby P, Andersen A-MN, et al. The Danish National Birth Cohort-its background, structure and aim. Scand J Public Health. 2001;29: 300–307. pmid:11775787
  33. 33. Olsen J, Meder K. Better health for mother and child–The Danish National Birth Cohort (DNBC), its structure, history and aims. Nor Epidemiol. 2014;24. Available: http://www.ntnu.no/ojs/index.php/norepid/article/viewFile/1756/1753
  34. 34. Rodríguez DA, Rodríguez-Martínez CE, Cárdenas AC, Quilaguy IE, Mayorga LY, Falla LM, et al. Predictors of severity and mortality in children hospitalized with respiratory syncytial virus infection in a tropical region: Epidemiology of RSV in a Tropical Region. Pediatr Pulmonol. 2014;49: 269–276. pmid:23401345
  35. 35. Simon A, Ammann RA, Wilkesmann A, Eis-Hübinger AM, Schildgen O, Weimann E, et al. Respiratory syncytial virus infection in 406 hospitalized premature infants: results from a prospective German multicentre database. Eur J Pediatr. 2007;166: 1273–1283. pmid:17943313
  36. 36. Gunville CF, Sontag MK, Stratton KA, Ranade DJ, Abman SH, Mourani PM. Scope and Impact of Early and Late Preterm Infants Admitted to the PICU with Respiratory Illness. J Pediatr. 2010;157: 209–214.e1. pmid:20338574
  37. 37. Flynn T, Möller C, Jönsson R, Lohmander A. The high prevalence of otitis media with effusion in children with cleft lip and palate as compared to children without clefts. Int J Pediatr Otorhinolaryngol. 2009;73: 1441–1446. pmid:19709760
  38. 38. Bhaskaran K, Smeeth L. What is the difference between missing completely at random and missing at random? Int J Epidemiol. 2014;43: 1336–1339. pmid:24706730
  39. 39. Jacobsen TN, Nohr EA, Frydenberg M. Selection by socioeconomic factors into the Danish National Birth Cohort. Eur J Epidemiol. 2010;25: 349–355. pmid:20349116
  40. 40. Greene N, Greenland S, Olsen J, Nohr EA. Estimating Bias From Loss to Follow-up in the Danish National Birth Cohort: Epidemiology. 2011; 1. pmid:21918455
  41. 41. Hassan E. Recall Bias can be a Threat to Retrospective and Prospective Research Designs. Internet J Epidemiol. 2005;3: 1–7.
  42. 42. Engel J, Anteunis L, Volovics A, Hendriks J, Marres E. Predictive value of parent-reported symptoms in the assessment of otitis media with effusion during infancy. Scand J Prim Health Care. 2000;18: 25–29. pmid:10811039
  43. 43. Alho OP. The validity of questionnaire reports of a history of acute otitis media. Am J Epidemiol. 1990;132: 1164–1170. pmid:2260548
  44. 44. Vernacchio L, Vezina RM, Ozonoff A, Mitchell AA. Validity of parental reporting of recent episodes of acute otitis media: a Slone Center Office-Based Research (SCOR) Network study. J Am Board Fam Med. 2007;20: 160–163. pmid:17341752
  45. 45. Engel JA, Anteunis LJ, Volovics A, Hendriks JJ, Manni JJ. Chronic otitis media with effusion during infancy, have parent-reported symptoms prognostic value? A prospective longitudinal study from 0 to 2 years of age. Clin Otolaryngol Allied Sci. 1999;24: 417–423. pmid:10542922
  46. 46. Kvestad E, Kvaerner KJ, Røysamb E, Tambs K, Harris JR, Magnus P. The reliability of self-reported childhood otitis media by adults. Int J Pediatr Otorhinolaryngol. 2006;70: 597–602. pmid:16143406
  47. 47. Olesen C, Søndergaard C, Thrane N, Nielsen GL, de Jong-van den Berg L, Olsen J, et al. Do pregnant women report use of dispensed medications? Epidemiol Camb Mass. 2001;12: 497–501.
  48. 48. Daly KA, Brown JE, Lindgren BR, Meland MH, Le CT, Giebink GS. Epidemiology of otitis media onset by six months of age. Pediatrics. 1999;103: 1158–1166. pmid:10353923
  49. 49. Alho OP, Koivu M, Sorri M, Rantakallio P. Risk factors for recurrent acute otitis media and respiratory infection in infancy. Int J Pediatr Otorhinolaryngol. 1990;19: 151–161. pmid:2373599
  50. 50. Bentdal YE, Håberg SE, Karevold G, Stigum H, Kværner KJ, Nafstad P. Birth characteristics and acute otitis media in early life. Int J Pediatr Otorhinolaryngol. 2010;74: 168–172. pmid:19959246
  51. 51. Paradise JL, Rockette HE, Colborn DK, Bernard BS, Smith CG, Kurs-Lasky M, et al. Otitis media in 2253 Pittsburgh-area infants: prevalence and risk factors during the first two years of life. Pediatrics. 1997;99: 318–333. pmid:9041282
  52. 52. Uhari M, Mäntysaari K, Niemelä M. Meta-analytic review of the risk factors for acute otitis media. Clin Infect Dis. 1996;22: 1079–1083. pmid:8783714
  53. 53. Håberg S, Bentdal Y, London S, Kvaerner K, Nystad W, Nafstad P. Prenatal and postnatal parental smoking and acute otitis media in early childhood. Acta Paediatr. 2009; pmid:19764924
  54. 54. Käyhty H, Ahman H, Eriksson K, Sörberg M, Nilsson L. Immunogenicity and tolerability of a heptavalent pneumococcal conjugate vaccine administered at 3, 5 and 12 months of age. Pediatr Infect Dis J. 2005;24: 108–114. pmid:15702037
  55. 55. Eskola J, Kilpi T, Palmu A, Jokinen J, Haapakoski J, Herva E, et al. Efficacy of a pneumococcal conjugate vaccine against acute otitis media. N Engl J Med. 2001;344: 403–409. pmid:11172176
  56. 56. Pavia M, Bianco A, Nobile CGA, Marinelli P, Angelillo IF. Efficacy of Pneumococcal Vaccination in Children Younger Than 24 Months: A Meta-Analysis. PEDIATRICS. 2009;123: e1103–e1110. pmid:19482744