Otitis Media and Tympanostomy Tube Insertion During the First Three Years of Life: Developmental Outcomes at the Age of Four Years
Objective. In a long-term, prospective study, we set out to determine whether otitis media in the first 3 years of life persisting for periods currently considered developmentally threatening actually results in later impairments of children’s cognitive, language, speech, or psychosocial development; whether prompt insertion of tympanostomy tubes prevents or lessens any such impairments; and whether, irrespective of causality, associations exist between persistent early-life otitis media and later developmental impairments. This report describes findings in study participants at the age of 4 years.
Methods. We enrolled 6350 healthy infants from 2 to 61 days of age at urban hospitals and 2 small-town/rural and 4 suburban private pediatric practices. We regularly evaluated the children for the presence of middle-ear effusion (MEE) throughout their first 3 years of life by pneumatic otoscopy, supplemented by tympanometry; we monitored the validity of the otoscopic observations on an ongoing basis; and we treated children for otitis media according to specified guidelines. In the clinical trial component of the study, we randomly assigned 429 children who met specified minimum criteria regarding the persistence of MEE to undergo tympanostomy tube insertion either promptly or after a defined extended period if MEE remained present. In the associational component of the study, we selected a representative sample of 241 children who ranged from having no MEE to having MEE the cumulative duration of which fell just short of meeting randomization criteria for the clinical trial. In 397 (92.5%) of the children in the clinical trial and in 234 (97.1%) of the children in the representative sample, we assessed cognitive, language, speech, and psychosocial development at the age of 4 years, using formal tests, conversational samples, and parent questionnaires.
Results. In children in the randomized clinical trial, there were no statistically significant differences in mean (±standard deviation) scores (higher denotes more favorable) favoring the early-treatment group over the late-treatment group on the General Cognitive Index of the McCarthy Scales of Children’s Abilities (97 ± 14 and 98 ± 14, respectively); the Peabody Picture Vocabulary Test–Revised, a measure of receptive language (90 ± 15 vs 92 ± 16); the Nonword Repetition Test, a measure of phonological memory (66 ± 12 vs 70 ± 12); the Number of Different Words, a measure of word diversity (150 ± 34 vs 150 ± 31); the Mean Length of Utterance in Morphemes, a measure of sentence length and grammatical complexity (3.4 ± 0.8 vs 3.4 ± 0.7); or the Percentage of Consonants Correct–Revised, a measure of speech-sound production (92 ± 5 vs 93 ± 5). There were also no significant differences in ratings (higher denotes less favorable) on the Parenting Stress Index–Short Form (Total Stress scores: 68 ± 18 vs 65 ± 17) or the Child Behavior Checklist (Total Problem T scores: 50 ± 10 vs 49 ± 10). In the associational component of the study, correlations between the children’s durations of MEE and their developmental outcomes were generally weak and, in most instances, nonsignificant. Exceptions, after adjustment for sociodemographic variables and for hearing thresholds at the time of developmental testing, consisted of a significant negative correlation between children’s cumulative durations of MEE in their first 3 years of life and scores on the McCarthy Verbal subscale, and significant positive correlations between durations of MEE and scores on 2 measures of parent–child stress. The percentage of variance in these scores explained by time with MEE beyond that explained by sociodemographic variables ranged from 1.6% to 3.3%. In both the randomized clinical trial and the associational component, sociodemographic variables seemed to be the most important factors influencing developmental outcomes, and in both components, the results at 4 years of age were consistent with the results that had been obtained at 3 years of age.
Conclusions. In otherwise healthy children who are younger than 3 years and have persistent MEE within the duration limits that we studied, prompt insertion of tympanostomy tubes does not measurably improve developmental outcomes at 4 years of age. In such children, persistent MEE within the duration limits that we studied is negligibly associated with and probably does not affect developmental outcomes at 4 years of age.
- otitis media
- middle-ear effusion
- tympanostomy tube insertion
- parent–child stress
Two important, unresolved questions confront clinicians who care for children: whether persistent early-life otitis media results in lasting impairments of children’s development and whether insertion of tympanostomy tubes prevents or lessens any such impairments.1,2 The importance of the questions stems from a chain of factors. First is the high prevalence of otitis media with effusion (OME) in infants and young children3,4 and the fact that OME is usually accompanied by a variable degree of conductive hearing loss.5 Second is widespread professional concern that sustained hearing loss during supposedly critical or sensitive periods of children’s development may result in lasting impairments of the children’s speech, language, or cognitive skills or their psychosocial adjustment6,7—all crucial attributes for later-life success and well-being. Third, arising mainly from this concern, is the current common and officially recommended8 or sanctioned2 practice of subjecting infants and young children with middle-ear effusion (MEE) that has persisted for as long as 3 months to myringotomy with tube insertion. Fourth is the limited nature and the inconclusiveness of evidence concerning relations between early-life otitis media and children’s later development1,9–14 and the lack of evidence that tube insertion in children with persistent MEE has a favorable developmental effect.
Many of the studies addressing possible relations between persistent early-life otitis media and later developmental outcomes have had important methodological limitations, and the results have been inconsistent.1,11,12,14 An invariable conceptual limitation of previous studies has been an exclusive focus on association, leaving unaddressed the additional and more clinically pertinent questions of causality and the effects of treatment. In 1991, we began a study designed to avoid both the methodological and the conceptual limitations of earlier studies. The main objectives of our study were to determine whether, in young children, the presence of MEE for periods sufficient to be considered developmentally threatening by many investigators and authorities actually results in impairments of cognitive, language, speech, or psychosocial development, and to determine whether prompt insertion of tympanostomy tubes protects against or minimizes such impairments. We also sought to determine, in the light of earlier research by others, the extent to which there exist associations—irrespective of causality—between persistent early-life otitis media and later developmental impairment. In previous reports, we have detailed findings in study participants during their first 3 years of life.4,13–15 The present report describes developmental findings in these children at the age of 4 years.
The study includes 2 main components. The centerpiece of the study is the randomized clinical trial component. In this component, to obtain 2 groups of children with presumably equal developmental potential but with differing cumulative durations of MEE, we randomly assigned children who met specified criteria regarding persistent MEE in their first 3 years of life to undergo insertion of tympanostomy tubes either promptly (the early-treatment group) or after a defined extended period if their MEE persisted (the late-treatment group). We then measured a range of developmental outcomes in these children at the ages of 3, 4, and 6 years. In the associational component, involving representative samples of the remaining study population, we tested associations between the children’s cumulative duration of MEE in their first 3 years of life and the same developmental outcomes as measured in the randomized clinical trial, at the same ages.13,14
As described in detail later in “Methods,” we enrolled a diverse sample of 6350 children within their first 2 months of life and followed them prospectively. In these children, the most important sociodemographic factor associated with the occurrence of otitis media in the first 2 years of life proved to be low socioeconomic status, as estimated from type of health insurance and level of maternal education.4 We also found that hearing, when tested, was abnormal (as defined later) in approximately one half of instances when children had unilateral MEE and three quarters of instances when children had bilateral MEE.14,15
Of the 6350 children, 588 eventually met the eligibility criteria for the clinical trial. A total of 429 of these children (73.0%) underwent randomization, and 402 of the 429 children (93.7%) received developmental testing at the age of 3 years. As anticipated from the known effects of tube insertion in resolving OME and maintaining middle-ear ventilation,2 after assignment of the children, there were large differences in the cumulative duration of MEE between the early-treatment group and the late-treatment group (only one third of whom actually underwent tube insertion by the age of 3 years). For example, during the first 12 months after randomization, 45% of the children in the late-treatment group had MEE for >50% of the days, compared with 14% of the children in the early-treatment group. Notwithstanding those differences in cumulative duration of MEE, we found no significant differences between the 2 groups at the age of 3 years in mean scores on any of a range of developmental outcome measures concerning cognition, receptive and expressive language, speech sound production, parent–child stress, and child behavior.15
In the associational component of the study, we studied a sociodemographically representative sample of 241 children who represented a spectrum of MEE experience from having no MEE to having MEE the cumulative duration of which fell just short of meeting our criteria for enrollment in the randomized clinical trial. At the age of 3 years in these children, referred to hereafter as the associational sample, we found statistically significant but modest negative associations (range: −0.13 to −0.21) between the children’s cumulative durations of MEE in their first year of life or in age periods that included their first year of life and their scores at the age of 3 years on formal tests of receptive vocabulary and verbal aspects of cognition. However, we found no significant associations between antecedent time with MEE and the children’s scores on other aspects of cognition or on measures of spontaneous expressive language and speech sound production.14 We also found, in a larger, unselected sample of 2278 of the enrolled children, no substantial relationships between the children’s cumulative antecedent time with MEE and their parents’ ratings, when the children reached the age of 3 years, of parent–child stress and of the children’s behavior problems.13 In contrast to the limited relations found at the age of 3 years between the duration of MEE and developmental outcome, we found strong relations in both components of the study between results on developmental outcome measures and children’s socioeconomic status, with results most favorable among the most socioeconomically advantaged children and least favorable among the least advantaged children.13–15
We obtained developmental findings at the age of 4 years in 397 (92.5%) of the 429 children in the randomized clinical trial and in 234 (97.1%) of the 241 children in the associational sample. Most of the measures that we used were the same as those we had used in the children at the age of 3 years. Findings at this later age are of interest for several reasons. First, 36 months is the lower age limit for administering many formal tests, the reliability and predictive validity of which tend to increase with age. Second, many developmental attainments, particularly in the domains of speech and language production, occur after the age of 3 years. Third, children’s responses when they are older may be more valid and reliable. Accordingly, it is possible that relations between MEE and aspects of development might be seen more fully and precisely in 4-year-old than in 3-year-old children.
Details concerning study procedures have been described previously.4,13–15 In brief, between June 1991 and December 1995, we enrolled 6350 healthy infants from 2 to 61 days of age at 8 sites in the Greater Pittsburgh area: Children’s Hospital of Pittsburgh, Mercy Hospital of Pittsburgh, and 2 small-town and rural and 4 suburban private pediatric group practices. The study was approved by the institutional review boards of the 2 hospitals, and written informed consent was obtained from 1 or both parents of each enrolled infant. We excluded infants who met any of the following criteria: birth weight <5 lb (2268 g); small for gestational age; a history of neonatal asphyxia or other serious illness; a major congenital abnormality or chronic illness; multiple birth; a sibling enrolled in the study; in foster care or adopted; mother dead, seriously ill, a known drug or alcohol abuser, or, in the judgment of study personnel, too limited socially or intellectually to give informed consent or adhere to the study protocol; mother younger than 18 years; and English not the only household language.
We used pneumatic otoscopy, supplemented in most instances by tympanometry, to evaluate children’s middle-ear status at least monthly until they were 3 years of age. We regularly monitored the validity of study clinicians’ otoscopic observations and found satisfactory levels of interobserver agreement. We used the term MEE to encompass all types of otitis media in which effusion is present: acute otitis media with or without otorrhea, OME, and otorrhea through a tympanostomy tube. We estimated the cumulative proportions of days each child had unilateral and bilateral effusion, respectively, on the basis of diagnoses made at individual visits and interpolations for intervals between visits. Following a uniform treatment protocol, we prescribed antimicrobial drugs routinely for episodes of acute otitis media and also routinely for episodes of OME until 1995 but selectively thereafter.
We conducted audiometric testing whenever practicable in all children who had unilateral or bilateral MEE continuously for 8 weeks, every 4 weeks thereafter so long as effusion remained present, and again once effusion had resolved. We also conducted audiometric testing in a sample of children to serve as a presumably normative comparison group. These children either had never had MEE or had not had MEE for at least 60 days and were selected opportunistically so as to represent a range of age groups. Finally, we conducted audiometric testing in all children who were about to undergo developmental testing and in any child in whom a parent or a clinician suspected hearing loss. We defined hearing levels using auditory brainstem response data in infants younger than 6 months and pure-tone data in children 6 months of age or older. For behavioral testing in children up to 2.5 to 3 years of age, we presented stimuli in the sound field because of time constraints; results obtained thus reflected hearing acuity mainly in the better-hearing ear. We presented stimuli by means of earphones to older children. We calculated pure-tone average thresholds from values at 500, 1000, 2000, and 4000 Hz or, when values at only 3 of those frequencies were available, from values at those 3 frequencies. On the basis of data obtained in study children who had no effusion,16 abnormal hearing tests were defined as an auditory brainstem response threshold >20 decibels hearing level (dB HL), or a pure-tone average threshold >25 dB HL up to the age of 10 months, >20 dB HL from 10 to 23 months, and >15 dB HL from the age of 2 years onward.
Children in the Randomized Clinical Trial
Children became eligible for the randomized clinical trial when, beginning at 2 months of age and within the first 3 years of life, they developed MEE that seemed substantial in degree and that persisted, despite treatment with antimicrobial drugs, for 90 days in the case of bilateral effusion or 135 days in the case of unilateral effusion. Children with intermittent bilateral or unilateral MEE for specified proportions of longer periods were also eligible, following criteria that are listed in Appendix 2 of the electronic version of reference 15 and that are available from the authors. For example, a child was eligible when he or she had had bilateral effusion during at least 67% of the preceding 180-day period or unilateral effusion during at least 67% of the preceding 270-day period. Children who met one of these criteria and whose parent(s) or guardian(s) gave written informed consent for randomization were stratified according to practice site, age in 6-month categories, and whether they met the eligibility criteria on the basis of bilateral or unilateral effusion. They were then assigned randomly, within these strata and in balanced blocks of 4 children, to undergo either early insertion of tympanostomy tubes or late insertion if MEE persisted. Assignments were made by designated nonclinical staff members using separate, computer-generated random-number lists. Children who were assigned to the early-treatment group were scheduled to undergo tube insertion as soon as practicable. Those who were assigned to the late-treatment group were to undergo the operation 6 months later if bilateral effusion remained present or 9 months later if unilateral effusion remained present, but children in the late-treatment group were able to receive tube insertion earlier at parental request. Children for whom consent for random assignment was withheld were offered tube insertion electively. Tube insertion was performed using conventional methods,17 and, in most cases, tubes were allowed to remain in place until extruded spontaneously.
Children in the Associational Sample
We selected the associational sample randomly from among the children who failed to meet randomization criteria within the first 3 years of life. The sample consisted of 241 children who represented a spectrum of MEE experience from having no MEE to having MEE the cumulative duration and/or sequencing of which fell just short of meeting randomization criteria and who were sociodemographically representative of the study population as a whole. Among the children in this sample, the estimated duration of MEE ranged from none at any time to 65.6% of the first year of life and 44.8% of the first 3 years of life.14
Developmental Tests and Procedures
We attempted to conduct age 4 developmental testing of children as soon as possible after their fourth birthday and in any case within 2 months afterward. Whenever possible, the testing was conducted only when on the same day the child’s hearing-level thresholds in each ear were ≤15 dB at 1000, 2000, and 4000 Hz or in the sound field were ≤20 dB at 500, 1000, 2000, and 4000 Hz. Children who failed the hearing test were scheduled for later developmental testing when feasible. When later testing was not feasible or when children failed the hearing test at or near the end of the 2-month period for developmental testing, developmental testing was undertaken without additional delay. We used 3 methods of developmental assessment: formal tests, samples of conversation, and parent-reported inventories regarding parent–child stress and children’s behavior. Details concerning test conditions, examiners, and procedures and the recording, transcription, and analysis of conversational samples have been described previously.13,14,18 In addition to the outcome measures applied at the age of 3 years and described previously,13,14,18 at the age of 4 years, we administered a nonsense-word repetition task.19 This task assesses the ability to repeat standardized phonological strings of increasing length as a measure of phonological memory; substandard performance on such tasks has consistently been associated with impaired language skills20,21 and impaired reading skills22,23 at later ages. Table 1 summarizes the developmental tests that we performed. The examiners and transcriptionists were unaware of the children’s histories and health insurance status and the mothers’ level of education.
All analyses excluded data concerning middle-ear status within the first 2 months of life. In the clinical trial, we assumed a priori that only poorer outcomes in the late-treatment group would be clinically important and that a difference of 0.33 standard deviations (SD) or greater favoring the early-treatment group on any outcome measure could be clinically important. On the basis of one-tailed α set at 0.05 and a Bonferroni adjustment made for multiple outcome measures, we calculated that 182 children would be needed in each group to detect a difference of 0.33 SD at a power of 0.80. Results of the clinical trial were based on the intention-to-treat principle. In the associational component of the study, we used Pearson pairwise correlations to test for associations between children’s scores on developmental tests and their estimated cumulative proportions of days with MEE. The original sample size of 241 children was sufficient, based on 2-tailed α set at 0.05 and a Bonferroni adjustment, to detect correlations of 0.20 at a power level of 0.71 and correlations of 0.25 at a power level of 0.91.32 Because correlations involving days with bilateral MEE differed little from correlations involving total days with MEE (ie, bilateral plus unilateral), the data and analyses presented here are limited to those concerning total days with MEE. We used χ2 tests to evaluate differences between proportions of children in different groups. We used analysis of variance to test for differences between mean values and a modification33 to test for differences involving trends. We used linear regression analysis to adjust for potentially confounding variables and to test for interactions, and we used repeated-measures analysis of variance to compare across-time measurements of commensurable outcome measures. We used 2-tailed tests for all analyses, and we set statistical significance at P ≤ .05.
Figure 1 shows schematically the derivation of the samples of children whom we assessed at the ages of 3 and 4 years and, in the children who received tympanostomy tubes, summarizes the ages at which tubes were inserted. Selected demographic and clinical characteristics of the 631 children who received developmental testing at the age of 4 years are shown in Table 2 . Within the randomized clinical trial, there were no significant differences in characteristics between the 397 children who received developmental testing and the 32 children who were not tested or between the 204 tested children in the early-treatment group and the 193 tested children in the late-treatment group. In 183 (72.6%) of the 252 children who met the criteria on the basis of unilateral MEE, bilateral effusion had been present during >25% of the immediately preceding 6-month period. In comparison with the 397 children in the randomized clinical trial, larger proportions of the 234 children in the associational component were from suburban study sites, were white, had mothers who were college graduates, were enrolled in private health insurance plans rather than Medicaid, and met protocol-specified criteria regarding hearing levels at the time of developmental testing; and smaller proportions were from urban study sites and were black (all P < .001). In 375 (94.5%) of the 397 children in the randomized trial and in 227 (97.0%) of the 234 children in the associational component, developmental testing was completed within 2 months of their fourth birthday.
Findings in the Randomized Clinical Trial
Mean scores in the children in the early- and late-treatment groups at 4 years of age on measures of cognition, receptive language, expressive language, and speech sound production are shown in Table 3 . Higher values reflect more favorable results. With the exception of a modest but statistically significant difference in scores on the Nonword Repetition Test favoring the late-treatment group, there were no significant differences between the 2 groups on any measure. These relationships remained unchanged after adjustment for age in months at the time of testing.
Mean scores on parent-rated measures of parent–child stress and children’s behavior are shown in Table 4 . Higher values reflect less favorable results. There were no significant differences between the early- and the late-treatment groups, either before or after adjustment for age in months at the time of testing. Analysis of all of the developmental outcome data in the randomized children according to whether they actually received tube insertion within 90 days after randomization also showed no significant differences favoring the children who received tubes within that period over those who did not.
In the trial group as a whole, mean scores on all outcome measures were, with few exceptions, most favorable among the most socioeconomically advantaged children—as reflected by study site grouping, race, maternal educational level, and health insurance status—and least favorable among the least advantaged children; with few exceptions, the differences were statistically significant. Mean scores were also consistently more favorable in girls than in boys on all measures of cognition, language, and speech but not on measures of parent–child stress and children’s behavior.
We conducted tests for interaction to determine whether outcomes in the clinical trial differed in relation to specific subject-related variables, namely 1) whether children met the study’s randomization criteria during their first, second, or third year of life; 2) the children’s prerandomization illness patterns, ie, whether they met the randomization criteria on the basis of bilateral continuous MEE, unilateral continuous MEE, bilateral discontinuous MEE, or unilateral discontinuous MEE; and 3) in the 356 children (187 early treatment, 169 late treatment) who received hearing tests during 1 or more episodes of MEE before being randomized, whether 1 or more of those tests gave abnormal results (as defined earlier), or gave a pure-tone average threshold of ≥30 dB, or gave a pure-tone average threshold of ≥40 dB. (Ninety-one percent of the tests were conducted in the sound field. Because sound field results reflect hearing acuity mainly in the better-hearing ear, elevated thresholds found in sound field testing were indicative of bilateral hearing loss.) We found no statistically significant interactions between age at meeting randomization criteria and treatment or between prerandomization illness patterns and treatment. However, we found a number of significant interactions between the hearing-level categories and treatment. Additional analysis indicated that among children in the more normal (or less abnormal) categories (ie, no prerandomization hearing test abnormal, or no test result ≥30 dB, or no test result ≥40 dB), many of the formal test results and Child Behavior Checklist (CBCL) results were significantly more favorable in the late-treatment group than in the early-treatment group, whereas among children in the more abnormal hearing-level categories (ie, 1 or more tests abnormal, or 1 or more test results ≥30 dB, or 1 or more test result ≥40 dB), the 2 treatment groups were comparable regarding most developmental outcomes.
Children Whose Parents Withheld Consent for Randomization
Among the 159 children whose parents declined to consent to randomization, we obtained test results at the age of 4 years regarding cognition, language, and speech in 101 and regarding parent–child stress and children’s behavior in 83. With the exception of a modestly lower mean score on the McCarthy Quantitative Subscale (47.5 vs 50.0; P = .046), none of the mean scores in these children on the various developmental measures differed significantly from corresponding mean scores in the children in the randomized trial (early- and late-treatment groups combined).
Findings in the Associational Sample
Developmental test results at 4 years of age were available for 234 of the 241 children who composed the original associational sample and had been tested at 3 years of age. We calculated unadjusted correlations between individual children’s scores at age 4 on the various outcome measures and the children’s cumulative proportions of total days with MEE during their first year of life (2–12 months), their first 2 years of life (2–24 months), and their first 3 years of life (2–36 months). The correlations were generally weak and, in most instances, nonsignificant. Exceptions consisted of significant negative correlations, ranging from −0.13 to −0.18, between the percentage of days with MEE during 1 or more of these age periods and scores on the General Cognitive Index of the McCarthy Scales of Children’s Abilities and its Verbal subscale and the Peabody Picture Vocabulary Test–Revised, and significant positive correlations, ranging from 0.14 to 0.21, between antecedent days with MEE and scores on the Parent–Child Dysfunctional Interaction subscale and Total Stress scores of the Parenting Stress Index–Short Form (PSI-SF). After adjustment by stepwise regression for statistically significant sociodemographic variables (study site grouping, gender, race, maternal education, and health insurance status) and for hearing thresholds at the time of testing (whether within or above protocol-specified limits), both individually and in varying combinations, only the correlations between the cumulative proportion of days with MEE in the first 3 years of life and scores on the McCarthy Verbal subscale and on the 2 aforementioned PSI-SF measures remained statistically significant, and the percentage of variance in these scores explained by time with MEE beyond that explained by sociodemographic variables ranged from 1.6% to 3.3%.
Table 5 shows these children’s scores on measures of cognition, receptive language, expressive language, and speech sound production in relation to selected sociodemographic characteristics, and Table 6 shows their scores in relation to those characteristics on measures of parent–child stress and children’s behavior. Findings were similar to those in the randomized clinical trial. On most outcome measures, statistically significant trends or differences in mean scores were found in relation to every demographic characteristic (as listed earlier) that reflected children’s socioeconomic status, with scores most favorable among the most advantaged children and least favorable among the least advantaged children. Also as in the clinical trial, mean scores on measures of cognition, language, and speech were consistently higher in girls than in boys, but on measures of parent–child stress and children’s behavior, mean scores in girls and boys were generally comparable.
We compared the data obtained at age 4 with corresponding data that had been obtained at age 3 and subjected the data to repeated-measures analysis. Regarding the age-normed, formal measures of receptive language and cognition (McCarthy General Cognitive Index and its subscales and Peabody Picture Vocabulary Test–Revised), there were significant decreases in scores from age 3 to age 4 in the randomized clinical trial (all P ≤ .02) but not in the associational component (all P ≥ .22). Regarding the conversational variables (Number of Different Words, Mean Length of Utterance in Morphemes, and Percentage of Consonants Correct–Revised), which are not age normed, there were, as expected, significant increases in scores from age 3 to age 4 (P < .001) in both the clinical trial and the associational component. Regarding the psychosocial measures (PSI-SF and CBCL) that were the same at the 2 ages (the CBCL subscale measures used at age 4 differ from those at age 3), there were no statistically significant changes in scores from age 3 to age 4 in either the clinical trial or in the associational component. In the clinical trial, the repeated-measures analysis showed no statistically significant treatment effects (averaged across time), reinforcing the conclusion that there were no differences in outcome between the early-treatment group and the late-treatment group.
The findings reported here follow on previously reported findings in a continuing prospective study aimed at clarifying relations between persistent otitis media early in children’s lives and their later cognitive, language, speech, and psychosocial development. The main strengths of the study consist of the large, diverse study population, the careful monitoring of children’s middle-ear status from early infancy, the design that incorporated both a randomized clinical trial and an associational component, the considerable quantity of hearing data, the long-term follow-up, and the low rate of attrition. The present findings in children at 4 years of age affirm and reinforce the findings that we reported earlier in the same children at 3 years of age.14,15
In children in the randomized clinical trial, we found no statistically significant differences in scores at age 4 favoring the early-treatment group over the late-treatment group on a range of developmental outcome measures. The associated confidence intervals afforded assurance that any differences of 0.2 SD or greater favoring the early-treatment group, if present, would have been detected. Within various subgroups, to the limited extent that significant differences occurred, the differences tended to favor children in the late-treatment group. In children in the associational component, correlations generally were weak and, in most instances, nonsignificant. However, we found, after adjustment for significant sociodemographic factors, statistically significant, albeit weak, correlations involving a few outcome measures, suggesting that greater antecedent time with MEE was correlated with less favorable results on those measures. In both the randomized clinical trial and the associational component, sociodemographic variables seemed to be the most important factors influencing developmental outcomes.
The present findings in the children in the randomized clinical trial indicate that prompt tympanostomy-tube insertion for persistent MEE in these children in their first 3 years of life did not result in improved developmental outcomes as measured at 4 years of age. The finding that certain of the subgroup results favored the late-treatment group over the early-treatment group is difficult to explain, and because the early-versus-late differences were small, the finding seems likely attributable to chance. In the study’s associational component, in which most of the associations examined between children’s antecedent time with MEE and the various developmental outcomes were nonsignificant, an explanation for the few statistically significant associations found is also not evident. The findings in the clinical trial suggest strongly that those associations were not causal in nature but rather were the result of either chance or confounding.
Given the totality of the present findings, it seems reasonable to infer that persistent otitis media in the first 3 years of life within the duration limits that we studied does not adversely affect children’s developmental outcomes as measured at 4 years of age, an inference reinforced by the consistency of the findings at age 4 with findings in the same children at age 3. The study’s findings, however, cannot be generalized to children who are not otherwise healthy or who have potentially handicapping conditions such as sensorineural hearing loss, cleft palate, or Down Syndrome. Neither can the study’s findings be generalized, as we have noted previously,15 to children with periods of effusion longer than those that we studied or to children whose effusion is consistently accompanied by moderately severe (rather than the more usual mild to moderate) hearing loss. However, both clinical experience and findings in the present clinical trial suggest that relatively few children in comparable circumstances will have periods of effusion substantially longer than those experienced by the trial participants. Of the 161 children in the trial who underwent developmental testing at age 3 and had not received tympanostomy tubes by the time of testing, only 22 (13.7%) were found at the time of testing to have unilateral MEE and only 9 (5.6%) to have bilateral MEE. In the 150 children who were tested at age 4, the corresponding values were 14 (9.3%) and 5 (3.3%), respectively. Of the 356 children in the trial who received hearing tests during 1 or more episodes of MEE before being randomized, only 58 (16.3%) had a pure-tone average threshold ≥45 dB on any test. Our findings thus seem applicable to most children in primary care settings who develop persistent MEE within the first 3 years of life as a more or less isolated condition. For such typically affected children, our findings provide a basis for the exercise of restraint in recommending tympanostomy tube insertion.
It remains to be seen whether relations between antecedent MEE and later development that were not discernible in the study participants at the ages of 3 or 4 years might emerge at later ages. Analysis of data obtained from these children at 6 years of age is currently under way, and we have also recently begun testing the children at 9 to 11 years of age using measures of literacy, attentional abilities, and related skills.
This work was supported by Grant HD26026 from the National Institute of Child Health and Human Development and the Agency for Healthcare Research and Quality, by grants from the University of Pittsburgh Competitive Medical Research Fund and the Children’s Hospital of Pittsburgh Research Advisory Committee, and by gifts from GlaxoSmithKline and Pfizer Inc.
This work was presented in part at the Pediatric Academic Societies Annual Meeting, May 4, 2002; Baltimore, Maryland.
We thank the following pediatricians, who made the decisions, participated in the planning, and assisted in the efforts to incorporate this study into their practices and who, at no small inconvenience and cost, provided unflagging support for study activities: at Beaver: David J. Cahill, MD; James Scibilia, MD; and Julius A. Vogel, Jr., MD; at Brentwood: Mark Diamond, MD; and Thomas D. Skelly, MD; at Gibsonia: Amelia V. Agustin, MD; and Eva A. Vogeley, MD; at Kittanning: Harold A. Altman, MD; James K. Greenbaum, MD; Kenneth R. Keppel, MD; and Donald J. Vigliotti, MD; at Mt. Lebanon: Scott L. Tyson, MD; and Celeste J. Welkon, MD; at Pleasant Hills: K. Gopalkrishna Pai, MD; and Harvey M. Rubin, MD; and at Mercy Hospital of Pittsburgh: Bradley J. Bradford, MD.
We also thank to Lawrence Shriberg, PhD, and Chad Allen of the Phonology Project at the University of Wisconsin–Madison for conducting the computer analyses of the speech samples.
In addition to the clinicians named above or as authors, the following people served as study-team clinicians. At Children’s Hospital of Pittsburgh: Irene Fabian, CRNP; Nancy J. Guerra, CRNP; Lisa M. Hakos-Zoffel, CRNP; Alejandro Hoberman, MD; and Phillip H. Kaleida, MD; at Beaver: Allen H. Chamovitz, MD; Sharon N. Cowden, MD; Valentina E. DiCenzo, MD; Verda S. Graf, BS, PA; S. Nasrin Ghorbanian, MD; George R. Haddad, MD; Janet D. Liljestrand, MD; and Jennifer J. Momen, MD; at Brentwood: Norman L. Cohen, MD; Kristin L. Frederick, MD; Joan Schiebel, RN; and Brenda E. Watkins, MD; at Kittanning: Tracy Balentine, RN; Shirley Baum, CRNP; Lawrence J. Butler, MD; Thomas G. Lynch, MD; and JoAnn Nickleach, MD; at Mt. Lebanon: Barbara J. Bahl, CRNP; Barbara Braman, CRNP; M. Bridgetta Devlin, CRNP; Holly A. Frost, MD; Thelma L. Herlich, MD; and Elizabeth H. Michael, CRNP; at Pleasant Hills: Todd H. Wolynn, MD; and at Mercy Hospital of Pittsburgh: Barbara L. Ayars, MD; Kimberly Brown, MD; Michael J. Daly, MD; Karla Falcon, MD; Pamela Heald, CRNP; Cynthia M. Hoess, MD; Barbara L. McNulty, CRNP; Yolanda Moore-Forbes, MD; Charles A. Pohl, MD; Sharon M. Roncevich, MD; Sherrill J. Rudy, CRNP; Evelyn J. Schmidt, RN; Sarah H. Springer, MD; and Karen S. Vargo, MD. We are also grateful to the many Children’s Hospital of Pittsburgh and Mercy Hospital of Pittsburgh pediatric house officers who served as primary care clinicians for study subjects and whose collaboration was essential for the successful conduct of the study.
The following people also assisted in the study: Ida Smith (clinic support coordinator); Miki Rakay (volunteer patient liaison specialist); Robin L. Lavelle and Valerie S. Quickley (schedulers); Jennifer A. Aliberti, Sandra Barnett, Nanci Barrett, Susan Braden, Cindy Brown, Nancy Ciaburri, Christin E. Costella, Cyndi Getty, Sue Ellen Hall, Jerome Hill, Anthony Heard, Karen Horox, Isabel Hunter, Beverly Joyce, Andrew R. Kaleida, Deborah Klemm, Judy Lazzeri, Janet R. Marshall, Jean Martin, Sue Musser, Karen M. Noto, Shirley Petrie, Deborah M. Pettibon, Kathleen A. Rafferty, Dawn M. Rone, Leslie A. Schropp, Brenda Shaffer, Sunitha Somanath, Dana Wingard, and Sheila Vasbinder (study technicians); Victoria M. Carrigan, MA; Kristin S. Carson, MA; Judith A. Cercone, MA; Craig E. Coleman, MA; Sarah E. Delano, MEd; Kathleen H. Fitch, MA; Diane L. Frank, MA, MEd; Melanie D. Naylor, MA; Julia Hauser, MS; Jeanne V. Manganaro, MS; LaVonne E. Milisits, MS; George A. Modreck, MA; Kathleen D. Post, EdD; Gretchen E. Probst, MAT; Tammy Sobek, MA; Barbara A. Vento, PhD; and James D. White, MA (audiologists); Robert D. Allen, MA; Julie M. Arrowsmith, MA; Sandra A. Cook, MA; Kari A. Copper, MA; Kristen A. Dambach, MA; Christine G. Colantoni, MA; Lisa A. Gamrat, BA; Melissa J. Glath, MA; Shane E. Goldberg, MA; Rachel E. Goodman, MA; Stephanie A. Hackett, MA; Maribeth Hayes, MA; Tara M. Jackson, MA; Anita A. Katzman, BS; Sheryl M. Kaufhold, BA; Tara L. Kotfis, MA; Laura M. Kriniske, MA; Heather L. Leavy, MA; Jennifer R. Mason, MA; Robert J. Masterson, MA; Jean A. Molleca, MA; Stephanie Nixon, MA; Nandini Ramakrishna, MA; Dana L. Raubenstrauch, MA; Karen G. Rizzo, MA; Tonia A. Sacca, MA; Emily J. Saner, MA; Diana G. Saveriano, MA; Andrea R. Schwartz, MA; Gina R. Shongo, MA; Beth A. Simari, MA; Lakeya C. Smith, MA; Deborah J. Speicher, MA; Leslie K. Szwarc, MA; Brenda M. Tossi, MA; James D. White, MA; Amy D. Wyatt, MA; Dawn M. Zeis, MA; and Melanie M. Zygowski, MA (transcriptionists); Kathleen A. Cecotti, BA; Sharon A. DiBridge, BS; Charlotte Heller; Toni L. McKeever; Karen S. Pourboghrat, BA; and Jennifer L. Schiebel (research/data assistants); Sharon M. Caputo, BS; Ron F. Hollis, BS; Sekip Firinciogullari, BS; Robert J. Molnar; Stephen Sefcik, BS; and Lingshi Tan, PhD (programmers); Hsiao-Lan Wei, MS (statistician); Jennifer S. Dietrich; L. Annabelle Kyle; and Robin E. Rice, BS (administration); and Margaretha L. Casselbrant, MD, PhD; Kenny H. Chan, MD; Joseph E. Dohar, MD; Margaret A. Kenna, MD; J. Christopher Post, MD; Sylvan E. Stool, MD; and Robert F. Yellon, MD (otolaryngologist consultants).
- Received November 21, 2002.
- Accepted January 31, 2003.
- Reprint requests to (J.L.P.) Children’s Hospital of Pittsburgh, 3705 Fifth Ave, Pittsburgh, PA 15213-2583. E-mail:
- ↵Stool SE, Berg AO, Berman S, et al. Otitis Media With Effusion in Young Children. Clinical Practice Guideline, Number 12. Rockville, MD: Agency for Health Care Policy and Research, Public Health Service, US Department of Health and Human Services; 1994 (AHCPR Publication No. 94-0622)
- ↵Klein JO. Risk factors for otitis media in children. In: Kavanagh JF, ed. Otitis Media and Child Development. Parkton, MD: York Press; 1986:45–51
- ↵Paradise JL, Rockette HE, Colborn DK, 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
- ↵American Speech-Language-Hearing Association. Guidelines for acoustic immitance screening of middle-ear function. ASHA.1979;21 :283– 288
- ↵Joint Committee on Infant Hearing 1994 Position Statement. Pediatrics.1995;95 :152– 156
- ↵2000 Clinical Indicators Compendium. Alexandria, VA: American Academy of Otolaryngology–Head and Neck Surgery; 2000;10 (bulletin).
- ↵Rapin I. Conductive hearing loss effects on children’s language and scholastic skills: a review of the literature. Ann Otol Rhinol Laryngol.1979;88(suppl 60) :3– 12
- ↵Paradise JL. Otitis media during early life: how hazardous to development? A critical review of the evidence. Pediatrics.1981;68 :869– 873
- ↵Paradise JL, Rogers KD. On otitis media, child development, and tympanostomy tubes: new answers or old questions? Pediatrics.1986;77 :88– 92
- ↵Paradise JL, Feldman HM, Colborn DK, et al. Parental stress and parent-rated child behavior in relation to otitis media in the first three years of life. Pediatrics.1999;104 :1264– 1273
- ↵Paradise JL, Dollaghan CA, Campbell TF, et al. Language, speech sound production, and cognition in 3-year-old children in relation to otitis media in their first 3 years of life. Pediatrics.2000;105 :1119– 1130
- McCarthy D. Manual for the McCarthy Scales of Children’s Abilities. San Antonio, TX: The Psychological Corporation; 1972
- Dunn LM, Dunn LM. Peabody Picture Vocabulary Test–Revised: Manual for Forms L and M. Circle Pines, MN: American Guidance Service; 1981
- Miller JF. Quantifying productive language disorders. In: Miller JF, ed. Research on Child Language Disorders: A Decade of Progress. A Volume Marking the Tenth Anniversary of the Wisconsin Symposium for Research in Child Language Disorders. Austin, TX: Pro-Ed; 1991:211–220
- Brown R. A First Language: The Early Stages. Cambridge, MA: Harvard University Press; 1973
- Abidin RR. Parenting Stress Index: Professional Manual. 3rd ed. Odessa, FL: Psychological Assessment Resources; 1995
- Achenbach TM. Manual for the Child Behavior Checklist/4–18 and 1991 Profile. Burlington, VT: University of Vermont Department of Psychiatry; 1991
- Copyright © 2003 by the American Academy of Pediatrics