Daily home tympanometry to study the pathogenesis of otitis media

A study was conducted to define the temporal relationship between a parent-identified cold episode and the diagnosis of otitis media. Forty children were studied in their homes with the use of daily tympanometry, symptom diaries and weekly otoscopy. A total of 136 parents identified colds, and 43 episodes of otitis media were recorded. New episodes of otitis media were observed in 22% of all colds, and 63% of all otitis media episodes occurred during a cold. Current research methods to study the pathogenesis of otitis media (OM) in the natural environment are flawed. The three most common protocols to study the relationship of upper respiratory tract infection (URI) and OM are data acquisition during sick visits only, 1–6 prospective interval monitoring during well visits 7–9 and/or a short period of daily monitoring. 10 Three factors introduce significant bias to the interpretation of the results of such studies and limit with respect to intervention strategies their potential usefulness: dependence on symptom presentation to mark the onset of an event; neglect of asymptomatic but pathogenic events; and use of recall to enumerate symptoms. With these deficiencies in mind, early events leading to the development of Eustachian tube dysfunction and middle ear effusion can be defined only by a prospective monitoring plan that includes obtaining daily subjective and objective information on URI and middle ear status throughout the “cold season.” To determine the feasibility of such an approach, a study was conducted that prospectively monitored a group of children in their natural environment on a daily basis (during well and sick days) throughout a typical cold season and included both subjective (symptom diaries) and objective (physical examination and tympanometry) measures of middle ear status and upper respiratory events. Preliminary results of this study (the cohort followed during the first year) were previously published. 11 In this report we present the complete set of results for 2 years of study and discuss the methods and benefits of this novel approach to studying the pathogenesis of OM. Methods and materials. Subjects. The Institutional Review Board of Children’s Hospital of Pittsburgh approved this study. Forty children ages 2 to 6 years were recruited by advertisements in neighborhood newspapers and flyers posted in the Children’s Hospital. The children were screened for exclusion criteria such as immunologic dysfunction, seizure disorder, diabetes mellitus, asthma requiring daily medication, medical conditions with a predisposition for OM (e.g. cleft palate, Down’s syndrome), congenital malformations of the ear, previous ear surgery including tympanostomy tubes, cholesteatoma, chronic mastoiditis, severe retraction pockets, hearing loss (conductive or sensorineural), external otitis, perforation of the tympanic membrane, any history of complications of OM and severe upper respiratory obstruction. For this study families were enrolled if they had at least two children who met qualifying criteria. Twenty families were enrolled, and their children were followed through the months of a typical cold season. Otoscopy, tympanometry and audiometry were done on entry. At the first clinic visit parents were trained to become proficient at tympanometry with the Racecar tympanometer (American Electromedics Corp., Amhurst, NH). Symptom diary. Each day the parents answered the following questions and recorded yes or no in a symptom diary: Does your child have a: cold, runny nose, cough, irritability, earache or pulling at ears? They recorded oral temperatures every third day. They maintained the diary for weekly review and collection. Home tympanometry. Daily, the parents performed tympanometry and recorded middle ear pressures bilaterally for each study child at approximately the same time (bedtime). The Racecar tympanometer was used. Hard copies of the tympanograms were maintained by the parent until collected by the investigator weekly. Home visits. After enrollment the investigator visited each home to install the tympanometer, review its use and provide symptom diaries. Weekly visits by the investigator included reviewing the symptom diary for completeness, collecting hard copies of the tympanograms, examining the instrument and performing otoscopy. Analysis. The presence of a cold was defined by parents who used their usual criteria for making that assessment. For this study, a “cold episode” was defined as 2 consecutive days of cold reporting by the parent or 3 or more consecutive days of runny nose. Cold episodes must have lasted at least 3 days. A new episode was defined as being separated from the previous episode by at least 3 days. OM was defined by flat tympanograms (compliance, ≤0.2 ml) that lasted 3 or more days and separated by other episodes by at least 3 days of nonflat curves (compliance, >0.2 ml). Otoscopic findings consistent with OM included fluid bubbles or air/fluid levels in the middle ear, tympanic membrane discoloration, decreased mobility of the tympanic membrane or bulging of the tympanic membrane. Results. Forty children (20 families) were enrolled during the fall and continued the study during the winter and spring months. There were 19 boys and 21 girls, all Caucasian. Of the 40 children 14 were between the age of 18 months and 3 years and the others were 4 to 6 years old. Twenty-two children were in either school or day care for at least 2 days per week, and 5 other children had a sibling in day care. Ten children were exposed to second hand smoke in the household. Parent compliance was 92% (median) for both tympanometry and symptom diaries. Based on the first cohort of 20 subjects, only 3.2% of the tympanograms were unusable as defined by a failure to achieve an adequate pressure scan range or by failure to avoid probe tip occlusion. One child had to be excluded because her tympanograms persistently showed low compliance and otoscopic findings suggested a thickened tympanic membrane. Children were studied for 66 to 204 days per child, a median of 142 days per child and a total of 6225 child days. A total of 136 cold episodes (1444 days) were diagnosed with a median of 4 (range, 0 to 8) colds per child. Sixty-seven episodes (17 bilateral) of OM (either with or without acute symptoms) were identified in 16 children. Eleven children were treated with antibiotics for 5 episodes of symptomatic OM and 6 episodes of asymptomatic OM. Because of missed home visits and interval examinations, otoscopic confirmation of OM was made in 66% of the OM episodes defined on the basis of the tympanograms. The average duration of an OM was 20.5 days with a range of 3 to 129 days. Of 67 ear episodes 16 (24%), 6 (9%) and 1(1.5%) lasted >30, 60 and 90 days, respectively. Of 136 cold episodes 14 occurred when the ears had not resolved a previous OM episode Of the remaining122 cold episodes, 27 (22%) were associated with the development of OM within 7 days before and 14 days after the first day of the identified cold episode. Ninety-five (78%) cold episodes had no temporally associated OM episode. Forty-three child-episodes of OM started when both ears were free of OM. Of the 43 child-episodes 27 (63%) were and 16 (37%) were not temporally related to a cold episode. Of those episodes of OM with a cold, 6 (22%) occurred on the same day as the onset of a cold episode, 6 (22%) occurred by 1 to 7 days, 7 (26%) occurred by 8 to 14 days and 8 (30%) occurred before a cold episode was identified (Figs. 1 and 2).Fig. 1: OM episodes and temporal relationship to onset of URI symptoms.Fig. 2: Temporal relationship between OM and identified onset of cold symptoms. FTE, flat tympanogram episode.Discussion. An unbiased format for study of the epidemiology and pathophysiology of OM caused by URIs would involve daily, subjective and objective surveillance on a large cohort of enrollees at high URI risk. In the study reported here, we incorporated these components. Specifically we enrolled 40 children and studied them in their natural environment (home). The surveillance period included the fall and winter months and ended in the spring. The children were observed daily for both subjective and objective measures of colds and OM, both of which were defined by specific criteria. The results showed that family cooperation was excellent. The families adjusted to the daily schedule, were highly compliant and reliably recorded tympanograms. We attempted weekly visits by an investigator, which provided many opportunities to monitor compliance and interview and examine the enrollees. Subject retention was good; only two families dropped early from the study, both due to family emergencies. Overall the families were satisfied that the study did not interfere with family life and often reported participation to be an enjoyable family experience. Assuming that OM caused by a cold occurs within 7 days before or 14 days after the onset of cold symptoms, the incidence of OM during a cold in the present study was 22%. Other investigators have found a similar incidence. 5, 12 Of the OM episodes associated with URI, 70% occurred on the same days as or up to 14 days after the onset of cold symptoms. This correlates well with other studies that found that >50% of OM could be temporally related to a preceding URI. 12, 13 An additional 30% of OM episodes occurred within 7 days before the onset of cold symptoms, a group of episodes that was not previously identified. The existence of this group suggests that important pathologic events for OM occur earlier than previously appreciated. Overall our results suggest that in this age group, ∼63% of OM episodes are attributable as a complication of a cold. The remaining 37% of new OM were not temporally associated with a cold as defined in this study. However, it is possible that mildly symptomatic (asymptomatic) colds or URIs caused by other upper respiratory viruses went unnoticed by the parents. Alternatively these or a subset of these episodes might be completely unrelated to viral URIs and instead be caused by allergy, systemic illness or local mucosal conditions. Limitations of this study include: the interval physician examination of the children (maximum visits every week); the exclusion because of feasibility and compliance issues of children <18 months of age who are more susceptible to OM; the possible variability among parents in their assignment of a cold to a cluster of signs/symptoms; the lack of confirmation of virus infection (by culture, antigen detection or PCR); and the heavy reliance on tympanometry for objective diagnosis of OM. Many of these issues are related to budgetary constraints because of the pilot nature of this study. Nonetheless our results show that this study format is feasible and can yield new insights into the epidemiology and pathogenesis of OM during a cold. Acknowledgments. This work was supported in part by National Institutes of Health Grant DC-02833.