Family Practice Advance Access originally published online on August 17, 2005
Family Practice 2006 23(1):40-45; doi:10.1093/fampra/cmi083
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Long-term prognosis of acute otitis media in infancy: determinants of recurrent acute otitis media and persistent middle ear effusion
Julius Center of Health Sciences and Primary Care, University Medical Center Utrecht, The Netherlands
Correspondence to Maroeska M Rovers, Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Stratenum 6.139, PO Box 85060, 3508 AB Utrecht, The Netherlands; Email: M.Rovers{at}umcutrecht.nl
Received 10 February 2005; Accepted 6 July 2005.
Damoiseaux RAMJ, Rovers MM, Van Balen FAM, Hoes AW and de Melker RA. Long-term prognosis of acute otitis media in infancy: determinants of recurrent acute otitis media and persistent middle ear effusion. Family Practice 2006; 23: 40–45.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
Objectives. Children under 2 years of age with acute otitis media are known to have a relatively poor prognosis. The objective of this study was to assess whether known determinants for recurrent acute otitis media and persistent middle ear effusion after an episode of acute otitis media during childhood also apply to children under 2 years.
Methods. Study design: prospective study of 210 children under 2, with identification of potential prognostic determinants. Univariate and multivariate logistic regression analyses were applied to evaluate which parameters independently contributed to the prediction of both outcome measures (recurrent acute otitis media and persistent middle ear effusion). A prognostic function was developed, and the area under the receiving operating characteristic (ROC) was used to estimate the predictive ability of the prognostic models. Population: children under 2 years of age with an episode of acute otitis media in family practice. Outcomes measured: recurrent acute otitis media and persistent middle ear effusion.
Results. For the outcome recurrent acute otitis media data from 210 children were used and winter season, male sex, passive smoking and persistent symptoms for more than 10 days at presentation were independent prognostic determinants. For the outcome persistent middle ear effusion data from 190 children were used and winter season, bilateral disease at entry, a sibling history of recurrent acute otitis media, and a previous episode of acute otitis media independently predicted the outcome. No sufficiently discriminatory prognostic model could be constructed for either outcome measure.
Conclusion. Prediction of recurrent acute otitis media or persistent middle ear effusion in individual young children remains poor.
Keywords. Children, otitis media, prognosis.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
Acute otitis media is one of the most common diseases in infants and children. After an uncomplicated episode of acute otitis media, a child can face two problems, recurrent acute otitis media or persistent middle ear effusion. Children under 2 are at increased risk of recurrent acute otitis media1–3 and persistent middle ear effusion after an episode of acute otitis media.4–6 Known determinants of recurrence after an episode of acute otitis media include, apart from young age, male sex, sibling history of recurrent acute otitis media, not being breastfed, primary clinical treatment failure, bilateral disease, passive smoking, day care, season, history of recurrent acute otitis media.1,3,7–11 Determinants for persistent middle ear effusion after acute otitis media include season, bilateral disease, type B tympanogram at diagnosis of acute otitis media and history of recurrent acute otitis media.5,6,12–14
Most studies on determinants for recurrent acute otitis media or persistent middle ear effusion after an episode of acute otitis media were carried out in children up to 12 years. Studies identifying prognostic factors in the high-risk group of children under age 2 are scarce. In this prospective study we examined determinants for persistent middle ear effusion or recurrent acute otitis media in the children under 2 years.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
Patients
The study was conducted between February 1996 and December 1998 in The Netherlands, within the framework of a placebo controlled, double blind, randomised clinical trial studying the effect of amoxicillin for acute otitis media.15 Children aged between 6 and 24 months were eligible if they presented with acute otitis media at the office of their family doctor. Diagnosis was based on otoscopy (red eardrum, bulging or otorrhea) and presence of acute signs of infection (fever, pain, irritability), according to the guidelines of the Dutch College of General Practitioners. Children with a known immunological disorder, craniofacial abnormality, or Down's syndrome were excluded from this study. The ethical committee of the Children's Hospital of the University Medical Center Utrecht approved the protocol, and all parents of the children gave written informed consent before enrolment.
Follow up
At entry the doctor recorded the medical history, the presence or absence of known determinants for otitis media, and the results of otoscopy. During the 10 days of treatment (amoxicillin or placebo) the child was seen twice at the doctor's office. Inquiry was made about remaining symptoms and the ear drum was examined. The first author (RD) visited all included children at home at 6 weeks and, when the child had unilateral or bilateral effusion at 6 weeks, again at 3 months after their episode of acute otitis media. At 6 months all children with effusion (uni- or bilateral) at 3 months were visited at home and otoscopy and tympanometry were repeated. At six months the parents of all other children were contacted by telephone and were asked whether the child had suffered from another episode of acute otitis media and whether this diagnosis was verified by their family doctor.
Outcome criteria
Two prognostic outcomes were distinguished: recurrent acute otitis media and persistent middle ear effusion. Recurrent acute otitis media was defined as at least one episode of acute otitis media within 6 months of their initial acute otitis media episode. Persistent middle ear effusion was defined as uni- or bilateral middle ear effusion at all follow-up visits at 6 weeks, 3 months and 6 months. Middle ear effusion was diagnosed by tympanometry. Type B and C2 tympanograms (modified Jerger's classification) were regarded as indicative of the presence of fluid in the middle ear.16 The positive predictive value of this combination, lies between 50 and 99 per cent, also in children aged 6 to 24 months.17,18 Children who had undergone an ENT operation (tympanostomy tubes, adenoidectomy) during the six months follow-up because of persistent MEE or recurrent acute otitis media, were considered to have persistent middle ear effusion. When tympanometry was unsuccessful (in both ears) because of a lack of co-operation by a child, the subjects were excluded from the persistent middle ear effusion analysis.
Statistical methods
The following prognostic parameters were considered: age and sex, history of at least one previous episode of acute otitis media, day-care attendance (care for at least 5 children, not from the same family, during at least one day of the week), history of recurrent upper respiratory tract infections (6 or more episodes during the previous 12 months), allergy, number of siblings (>2 versus 
2), siblings with recurrent acute otitis media (3 or more episodes during one year), smoking in the household, season, breastfeeding (>6 months), bilateral disease, duration of symptoms of presentation of acute otitis media (>10 days versus 10), treatment at entry. The associations between each of these factors and the outcome measures (recurrent acute otitis media, persistent middle ear effusion) were determined by applying univariate logistic regression analysis, using SPSS 11.0. Analyses were performed separately for the two outcome parameters. Results were expressed as odds ratio (OR) with 95% confidence interval (CI).
Predictors that were univariately associated with the outcome (P 0.10) were included in multivariate logistic regression analyses. The predictive accuracy of the models was estimated by their reliability (goodness-of-fit) using Hosmer-Lemeshow statistics, which evaluate the correspondence between a model's predicted probabilities and the observed frequencies over groups spanning the entire range of probabilities.19 The discriminating ability of these prognostic functions was assessed by using receiver operating characteristics (ROC) analysis. The ROC curve is a plot of the true positive rate (sensitivity) versus the false positive rate (1 – specificity) evaluated at consecutive cut-off points of the predicted probability. The area under the ROC curve (AUC) provides a quantitative summary of the discriminative ability of a predictive model. A useless predictive model, such as a coin flip, would yield an AUC of 0.5. When the AUC is 1.0, the model discriminates perfectly between those with and without developing a prognostic outcome.21 In our analyses a value over 0.8 was considered as good with respect to discriminative ability.
To obtain an easily applicable prediction rule, the adjusted regression coefficients of the model were multiplied by a factor 10 and rounded to the nearest integer. Scores for each individual patient were obtained by assigning points for each variable and adding the results. Patients were classified according to their risk score. To simplify the interpretation of the model, various cut-off points were chosen to show the number of children developing or not developing recurrent acute otitis media and persistent middle ear effusion, and corresponding sensitivity, specificity, positive and negative predictive values were calculated.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
In total 240 patients were enrolled in the trial, 15 had an initial treatment failure during the first 10 days, i.e. they received another antibiotic, and were not followed further.16 Another 15 were lost to follow-up, mainly due to possible side effects or because the children had been withdrawn from the study. Thus, 210 children were included in our prognostic study. Characteristics of these children are shown in Table 1. The baseline characteristics of the 30 excluded children did not differ from those of the participating children. Valid tympanograms on all occasions were obtained from 190 children, and their baseline characteristics were similar to those of the total group (Table 1).
|
Clinical outcome
Recurrent acute otitis media (n = 210). One hundred and five (50 per cent) children developed at least one recurrent episode of acute otitis media within 6 months. Eighty per cent of all episodes reported by the parents were confirmed by their family doctor; the remainder were not seen by a doctor. In both the univariate and multivariate analyses recurrent acute otitis media was seen more often in boys (adjusted OR 1.8; 95% CI 1.0–3.3), children included in wintertime (October–March) (adjusted OR 2.4; 95% CI 1.3–4.4), and children with symptoms lasting longer than 10 days after enrolment (adjusted OR 2.3; 95% CI 1.2–4.1). Smoking in the household was associated with a lower risk of recurrence of acute otitis media (adjusted OR 0.5; 95% CI 0.3–0.9) (Tables 2 and 3). The items season, sex, passive smoking and persistent symptoms were included in the prognostic rule. The goodness-of-fit test indicated an acceptable fit of the prognostic model (P = 0.91). The AUC was 0.69; 95% CI 0.62–0.76.
|
|
Using the regression coefficients of the final predictive model, the probability of developing recurrent acute otitis media can be estimated for each child using the formula given in Table 3 (legend). For example, a boy (6 points), whose parents do not smoke (0 points), visiting a GP in the winter (9 points) with persistent symptoms (8 points), has a total score of –9 + 6 + 0 + 9 + 8 = 14 points, corresponding with a probability of recurrent acute otitis media of 67%. Table 5 shows the number of children in the cohort with and without recurrent acute otitis media across different categories of the risk score. For all score thresholds, the number of false–positive (those identified at high risk of developing the prognostic outcome, while the outcome did not occur) and/or false–negatives (those identified at low risk of developing the prognostic outcome, while in fact the outcome did occur) was too high to be of value in clinical practice.
|
|
Persistent middle ear effusion (n = 190). At 3 months 90 (47 per cent), and at 6 months 66 (35 per cent) children had persistent middle ear effusion. Twenty-six of the latter had an ENT operation during those 6 months. Persistent middle ear effusion at 6 months was seen more often in children enrolled in winter (adjusted OR 2.0; 95% CI 0.9–4.6), in those who presented with bilateral acute otitis media (adjusted OR 2.0; 95% CI 0.9–4.3), those with a history of a sibling with recurrent acute otitis media (adjusted OR 2.0; 95% CI 0.9–4.6), and in those who had had acute otitis media before (adjusted OR 2.1; 95% CI 0.9–4.3) (Tables 2 and 4).
The factors season, bilateral disease, sibling with recurrent acute otitis media, and a previous acute otitis media episode were included in the prognostic rule. The goodness-of-fit test indicated an acceptable fit of the prognostic model (P = 0.26) and the AUC was 0.69; 95% CI 0.60–0.79.
The probability of developing persistent MEE can be estimated for each child using the formula in Table 4 (legend), and the number of children in the cohort with and without persistent MEE across different categories of the risk score are shown in Table 6. Again, the number of false–positive and/or false–negatives was too high to be of value in clinical practice.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
Among children under 2 years of age with acute otitis media, determinants for recurrence of acute otitis media within 6 months include: winter season, persisting symptoms (>10 days) of the entry episode, and being male. Winter season, bilateral disease at entry, a sibling with recurrent acute otitis media, and a previous acute otitis media episode were independent predictors for persistent middle ear effusion at 6 months in these children. The performance of a prognostic model including these parameters in terms of the occurrence of recurrent acute otitis media or persistent middle ear effusion was rather poor (AUC <0.70), and it was not possible to find a threshold in the risk score that classified the children satisfactorily. For each arbitrary threshold too many children were misclassified to be of any clinical use.
Fifty per cent of the children in our study had at least one recurrence of acute otitis media within 6 months. This rate of recurrent acute otitis media after a previous episode of acute otitis media is in agreement with earlier studies.11,13,20,21 Season and sex were also identified as determinants for recurrent acute otitis media in studies in children of all ages.1,3,9,11,12 Jero et al. reported that primary clinical failure, defined as no improvement or worsening at 2 weeks, was a potential risk factor for recurrence of acute otitis media in the univariate analysis, but not the multivariate analysis.11 We also found that persistent symptoms, defined as lasting longer than 10 days after entry, were a determinant for recurrent acute otitis media in the multivariate analysis. Smoking was associated with a lower risk for recurrent acute otitis media in our study, contrary to the often accepted idea that passive inhalation of smoke can cause ear disease. Empirical evidence for the latter is, however, scarce.22 A paper by Stathis et al.23 also reported a lower risk of acute ear infections associated with passive smoking during early life. They even suggested a biological explanation, i.e. cigarette exposure may lead to histological changes in the epithelium of the middle ear and Eustachian tube. The finding may also be attributable to lower reporting rates of acute otitis media by parents who smoke.
Most studies on the occurrence of persistent middle ear effusion after an episode of acute otitis media in children under 2 included only a short follow-up period of 3 months or even less. Persistent middle ear effusion at 3 months was found in 47 per cent of the children in our study. Shurin et al. found persistent middle ear effusion at 3 months in 55 per cent of children aged under 2.4 Iino et al. found this condition in 50 per cent in this age group.5 Studies with older children reported a 0 to 10 per cent persistent middle ear effusion incidence at 6 months post acute otitis media.5,12,24 Bilateral disease, a history of acute otitis media and season were previously identified as determinants for persistent middle ear effusion.5,6,12–14 in studies including children from roughly 6 months to about 12 years. Our finding that a positive sibling history of recurrent acute otitis media increases the risk for persistent middle ear effusion after an episode of acute otitis media has, to out knowledge, not been reported before.
This study has a number of potential limitations. Recurrent acute otitis media was based on parental reports, and although 80 per cent of all episodes were confirmed by a doctor, 20 per cent was, thus, based on parental report only. A previous study showed that parents can report fairly accurately whether their child suffers from acute otitis media.25 Thus, the dilution of the risk estimates this may have caused is likely to be quite limited. Besides, restriction of the outcome definition to those with a physician-confirmed diagnosis yielded similar results. Diagnosis of middle ear effusion was based on tympanometry only, and therefore some children (about 10 percent) had to be excluded from the analysis because we could not obtain a valid tympanogram at every occasion. However, comparison of baseline characteristics of the remaining children with those of the total group, did not reveal relevant differences (Table 1) and any bias resulting from this exclusion would be modest. Furthermore almost 50% of the children with persistent middle ear effusion had undergone an ENT operation. We performed a sensitivity analysis excluding these children and the results were similar (data not shown). Treatment of the initial acute otitis media episode with amoxicillin did not influence the development of recurrent acute otitis media of persistent middle ear effusion.
Our study showed that some of the known indicators of recurrent acute otitis media or persistent middle ear effusion for childhood population at large with an acute otitis media, also determine prognosis in children aged under 2, a group with an established relatively poor prognosis.
A child without any of the items included in de prognostic function has a probability of developing recurrent acute otitis media of 25%. If the child has all 4 factors, this probability increases to 70%. These figures are 25% and 60% respectively for the outcome persistent middle ear effusion. Prediction of recurrent acute otitis media or persistent middle ear effusion in individual young children, on the basis of these determinants, as indicated by the area under the curve of the prognostic function, as well as by the sensitivity, specificity, positive and negative predictive values of different cut off points of the scoring formula remains, however, poor.
|
Declaration |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
Funding: Netherlands Organisation for Scientific Research (grant no 904–58–074).
Ethical approval: the ethical committee of the University Medical Center Utrecht approved the protocol.
Conflict of interest: none.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
1 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.[ISI][Medline]
2 Kvaerner KJ, Nafstad P, Hagen JA, Mair IWS, Jaakkola JJK. Recurrent acute otitis media: the significance of age at onset. Acta Otolaryngol 1997; 117: 578–584.[Medline]
3 Rovers MM, Schilder AG, Zielhuis GA, Rosenfeld RM. Otitis media. Lancet 2004; 363(9407): 465–473.[CrossRef][ISI][Medline]
4 Iino Y, Nakamura Y, Koizumi T, Toriyama M. Prognostic factors for persistent middle ear effusion after acute otitis media in children. Acta Otolaryngol 1993; 113: 761–765.[Medline]
5 Jero J, Virolainen A, Virtanen M, Eskola J, Karma P. Prognosis of acute otitis media: Factors associated with poor outcome. Acta Otolaryngol 1997; 117: 278–283.[Medline]
6 Monobe H, Ishibashi T, Fujishiro Y, Shinogami M, Yano J. Factors associated with poor outcome in children with acute otitis media. Acta Otolaryngol 2003; 123(5): 564–568.[Medline]
7 Froom J, Culpepper L. Otitis media in day-care children. A report from the international primary care network. J Fam Pract 1991; 32: 289–294.[ISI][Medline]
8 Alho OP, Laara E, Oja H. Public health impact of various risk factors for acute otitis media in northern Finland. Am J Epidemiol 1996; 143: 1149–1156.
9 Stenstrom C, Ingvarsson L. Otitis-prone children and controls: a study of possible predisposing factors.1. Heredity, family background and perinatal period. Acta Otolaryngol 1997; 117: 87–93.[Medline]
10 Pukander J, Luotonen J, Timonen M, Karma P. Risk factors affecting the occurrence of acute otitis media among 2–3-year-old urban children. Acta Otolaryngol 1985; 100: 260–265.[Medline]
11 Jero J, Karma P. Prognosis of acute otitis media. Factors associated with the development of recurrent acute otitis media. Acta Otolaryngol 1997; suppl 529: 30–33.
12 Claessen JQPJ, Appelman CLM, Touw-Otten FWMM, Melker RA de, Hordijk GJ. Persistence of middle ear dysfunction after recurrent acute otitis media. Clin Otolaryngol 1994; 19: 35–40.[ISI][Medline]
13 Mandel EM, Kardatzke D, Bluestone CD, Rockette HE. A comparative evaluation of cefaclor and amoxicillin in treatment of acute otitis media. Pediatr Infect Dis J 1993; 12: 726–732.[ISI][Medline]
14 Sakaguchi M, Taguchi K, Ishiyama T, Netsu K, Katsuno S. Tympanometric changes following acute otitis media in Japanese children. Eur Arch Otorhinolaryngol 1994; 251: 113–116.[Medline]
15 Damoiseaux RAMJ, Balen FAM van, Hoes AW, Verheij TJM, Melker RA de. Primary care based randomised, double blind trial of amoxicillin versus placebo for acute otitis media in children aged under 2 years. Br Med J 2000; 320: 350–354.
16 Jerger J. Clinical experience with impedance audiometry. Arch Otolaryngol 1970; 92: 311–324.[Medline]
17 Balen FAM van, Melker RA de. Validation of a portable tympanometer for use in primary care. Int J Pediatric Otorhinolaryngol 1994; 29: 219–225.
18 Takata GS, Chan LS, Morphew T, Mangione-Smith R, Morton SC, Shekelle P. Evidence assessment of the accuracy of methods of diagnosing middle ear effusion in children with otitis media with effusion. Pediatrics 2003; 112(6 Pt 1): 1379–1387.
19 Steyerberg EW. Prognostic modeling for clinical dicision making, theory and applications (thesis). Rotterdam: Erasmus Universiteit Rotterdam; 1996.
20 Kaleida PH, Casselbrant ML, Rockette HE, et al. Amoxicillin or myringotomy or both for acute otitis media: results of a randomized clinical trial. Pediatrics 1991; 87: 466–474.
21 Pichichero ME. Recurrent and persistent otitis media. Pediatr Infect Dis J 2000; 19: 911–916.[ISI][Medline]
22 Blakley BW, Blakley JE. Smoking and middle ear disease: are they related? A review article. Otolaryngol Head Neck Surg 1995; 112: 441–446.[CrossRef][ISI][Medline]
23 Stathis SL, O'Callaghan M, Williams GM, Najman JM, Andersen MJ, Bor W. Maternal cigarette smoking during pregnancy is an independent predictor for symptoms of middle ear disease at five years' postdelivery. Pediatrics 1999; 104: 16.
24 Wheeler MTK. Tympanometry in children with treated acute otitis media. Lancet 1986; i: 529–532.
25 Kontiokari T, Koivunen P, Niemela M, Pokka T, Uhari M. Symptoms of acute otitis media. Pediatr Infect Dis J 1998; 17: 676–679.[CrossRef][ISI][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||