Laterality of Acute Otitis Media: Different Clinical and Microbiologic Characteristics

Abstract

Background: 

A large individual patient data meta-analysis recently showed that children aged less than 2 years with bilateral, as compared with unilateral, acute otitis media (AOM) were at higher risk for persistent symptoms without antibiotic treatment. Prior studies have shown a propensity for children with bilateral AOM to be infected with bacterial pathogens, specifically Haemophilus influenzae. The objectives of this study were to further characterize risk factors for bilateral AOM and to assess the propensity for specific viral and bacterial pathogens to predispose to bilateral versus unilateral AOM.

Methods: 

We performed a secondary data analysis on 1216 cases of AOM diagnosed and treated at our institution: 566 subjects underwent tympanocentesis and middle ear fluid (MEF) culture. We compared subjects with bilateral and unilateral AOM for demographic characteristics, clinical findings, parent/clinician perception of AOM severity, and MEF study results for bacteria and viruses.

Results: 

When compared with children who have unilateral AOM, children with bilateral AOM were more likely to be younger (P < 0.001), have H. influenzae isolated from one or both MEFs (P < 0.0001), and have more severe inflammation of the tympanic membrane on otoscopic examination (P < 0.0001).

Conclusion: 

Compared with children who have unilateral AOM, children with bilateral AOM are more likely to have bacteria in the MEF and have more severe inflammation of the tympanic membrane. This may help explain why children with bilateral AOM are more likely to experience persistent symptoms without antibiotic treatment. Laterality of AOM should be considered when discussing treatment options with parents.

Acute otitis media (AOM) is one of the most common diseases of infancy and childhood. It has been shown that many children with nonsevere, uncomplicated AOM do well with watchful waiting, rather than immediate antibiotic treatment.^1–6 Using a wait-and-see strategy for prescribing antibiotics for AOM, Marchetti et al² avoided antibiotic treatment in two-thirds of children, and Little et al³ reduced the use of antibiotic prescriptions in his patients by 76%. In our clinical trial of watchful waiting for children with nonsevere AOM,¹ we observed a 70% reduction in antibiotic use in the observation group compared with the immediate antibiotic group. The reduction was 80% in children aged >2 years and 59% in younger children. In general, clinicians find it difficult to determine which children would benefit the most from antibiotic treatment. It would be helpful to have prediction rules for the initiation of antibiotics versus watchful waiting for children with AOM.

Rovers et al⁷ have recently performed a meta-analysis of 1634 children from 6 randomized trials published between 1991 and 2005, showing that without antibiotic treatment children younger than 2 years of age and who have bilateral AOM are more likely to experience persistent symptoms. These children are more likely to benefit from antibiotic treatment than older children and/or children with unilateral AOM. They calculated that only 4 children younger than 2 years of age with bilateral AOM need to be treated with antibiotics to prevent pain and/or fever 3 to 7 days later in one child, whereas 15 children 2 years or older with unilateral AOM need to be treated to prevent a prolonged course of AOM symptoms in 1 child. Rovers hypothesized that in contrast to unilateral AOM, bilateral AOM in children aged <2 years more often has a bacterial cause.

To further assess the potential of AOM laterality to assist the clinician in deciding about treatment options, we analyzed the association of laterality with demographic, clinical, and microbiologic characteristics in 1216 cases of AOM studied at our institution.

METHODS

Of 1216 AOM cases included in this analysis, we performed tympanocentesis and MEF culture on 566 subjects (tap study subjects). The remaining 650 were nontap study subjects for whom clinical and otoscopic severity scores were available. Data reported here were all obtained on the day the child first presented with acute symptoms of AOM. AOM was defined as the presence of: (1) acute onset of symptoms such as fever, irritability, or otalgia, (2) middle ear effusion, and (3) tympanic membrane (TM) erythema or fullness/bulging.^8,9 The tap studies (1989–1998) focused on the relation of laterality to bacterial and viral cultures of MEF obtained by tympanocentesis. The tap studies controlled for previous antibiotic use by excluding from enrollment any child who had received an antibiotic during the week before enrolment. The nontap study data (2000–2003) were obtained in a clinical trial assessing the safety, efficacy, acceptability, and costs to parents of immediate antibiotic treatment versus watchful waiting for children with nonsevere AOM.¹ The aforementioned Rovers meta-analysis⁷ included 223 subjects from the 650 nontap study subjects presented here. The present study data include 427 additional subjects who were not enrolled in the clinical trial, most of whom were experiencing severe AOM. These data included AOM laterality, demographic data, risk factors, parent/clinician perceptions of AOM severity and clinical findings.

Tap Studies.

The data for one or both ears obtained from these studies included the following: (a) laterality of AOM, (b) age, gender, and ethnicity of the subjects, (c) the number of subjects with bacteria, virus, both bacteria and virus, or no pathogen, and (d) the number of subjects with specific bacterial species isolated from one or both MEFs. Methods describing the tympanocentesis and MEF culture procedures used in these studies have been previously published.^10–13

Nontap Study.

To assess TM severity and confirm the diagnosis of AOM, trained and validated otoscopic examiners used an 8-grade otoscopic severity scale (OS-8).^14–16 The grades are as follows: 0 indicates normal, or effusion only, without erythema; 1, erythema only, no effusion; 2, erythema, air-fluid level, no opacification, meniscus noted; 3, erythema, complete effusion, no opacification; 4, erythema, air bubble(s) observed, purulent (cloudy) fluid visualized through TM, no bulging; 5, erythema, complete effusion, complete opacification, no bulging; 6, erythema, bulging, opacification, rounded, doughnut appearance of TM; 7, erythema, complete effusion and opacification with bulla formation. OS-8 scores were assessed for both the left and right TMs, and only the greater of the 2 scores was recorded for each subject. To be categorized as bilateral, subjects had to have symptoms and were required to have an OS-8 score >3 for both ears. A body temperature for each subject was also recorded at initial examination. Each subject's symptoms were assessed for severity by parents who filled out a previously published 5-item ear treatment group symptom questionnaire (ETG-5).¹⁶ We have previously shown that ETG-5 scores correlate well with OS-8 scores for a large group of children representing a wide spectrum of AOM disease severity. Using ETG-5, parents rated AOM symptoms such as fever, ear ache (tugging), irritability, feeding, and sleeping on a scale of 1–7 for a maximum score of 35. The median of all ETG-5 scores was 16; symptoms scores <16 were considered to be nonsevere, and scores ≥16 were considered to be severe. Of 792 subjects in the original dataset, 142 were excluded because they lacked criteria for a diagnosis of AOM or were lacking sufficient data to analyze, leaving a final total of 650 subjects. Subjects excluded were OS-8 <4 in both ears (42 subjects), OS-8 data missing for one or both ears (50 subjects), ETG-5 = 0, child asymptomatic per parent (29 subjects), and ETG-5 data missing (21 subjects).

Statistics.

Data were analyzed using SAS statistical software, version 9.1 (SAS Institute, Cary, NC). We used χ² analysis and odds ratios to evaluate univariate associations between laterality and demographic characteristics, risk factors, clinician/parent perception of severity, clinical observations, and MEF culture results. Subsequently, the association between each risk factor and laterality was examined by multivariate logistic regression analysis using the risk factors that demonstrated univariate relations with laterality at a significance: P ≤ 0.10.

RESULTS

Demographics.

Demographics are summarized in Table 1. The nontap study included several older subjects to age 16 years. A total of 183 subjects in the nontap study had received 1 or more doses of heptavalent pneumococcal vaccine before being seen for their AOM episode.

TABLE 1: Demographic and Laterality Data from 1216 Study Subjects

AOM Laterality Versus Demographic Characteristics.

Table 2 compares the demographic characteristics of 1216 subjects by laterality of AOM. Children age <2 years were more likely to be diagnosed with bilateral AOM (P < 0.0001). There were no associations between laterality and gender or ethnicity.

TABLE 2: Relationship Between AOM Laterality and Demographic Characteristics From 1216 Subjects

Nontap Subjects: AOM Laterality Versus Risk Factors and Clinical Findings.

Table 3 shows no relationship between AOM laterality and breast-feeding, passive smoke exposure, number of children in the subjects' home, daycare attendance, or history of prior AOM infections. In Table 3 we also see that laterality did not correlate with the parent perception of illness severity, as measured by ETG-5. Relatively small numbers of subjects had elevated body temperatures when seen by the physician at the time of enrolment: bilaterality was not associated with an elevated body temperature. TM inflammation was recorded as severe (OS-8 = 6 or 7) more often in subjects with bilateral AOM. In the regression model, age (OR = 1.11, 95% CI: 1.04–1.19) and severity of TM inflammation (OS-8) predicted bilaterality (OR = 0.63, 95% CI: 0.52–0.76), whereas neither prior AOM (OR = 1.04, CI: 0.96–1.12) nor daycare attendance (OR = 0.91, CI: 0.76–1.08) showed significant associations.

TABLE 3: Relationship Between AOM Laterality, Risk Factors and Clinical Findings of 650 Nontap Study Subjects

Tap Studies: AOM Laterality Versus MEF Pathogens.

Table 4 compares MEF culture results obtained from ears with bilateral and unilateral AOM. In general, bacterial pathogens without viruses were cultured most often from subjects' MEF (307/566, 54%). H. influenzae, S. pneumoniae, M. catarrhalis, and S. aureus were cultured from 382 subjects with bilateral AOM and 91 subjects with unilateral AOM. Pathogens less commonly isolated were K. pneumoniae, P. aeruginosa, S. pyogenes, Group B streptococcus, Acinetobacter, C. perfringens, and E. coli. These organisms were isolated from 15 subjects with bilateral and 1 subject with unilateral AOM. Bacterial pathogens were more often isolated from subjects with bilateral than unilateral AOM (Table 4, P = 0.03). Nine types of virus were detected from MEF of subjects with AOM. Respiratory syncytial virus (RSV) was detected by either culture or antigen detection from MEF of 63/566 (11%) of subjects overall; 45/63 (71%) of subjects with RSV detected had bilateral AOM. Viruses did not show a significant relation to laterality.

TABLE 4: Relationship Between AOM Laterality and MEF Microbiologic Results

Table 5 displays, by laterality, the number of subjects with any versus no bacterial or viral pathogen, and the specific bacterial pathogens most commonly isolated. Here we see again that bacterial pathogens were more likely to be isolated from subjects with bilateral AOM as compared with those with unilateral AOM (P = 0.009). An association with laterality was not observed among those subjects from whom a viral pathogen was isolated. H. influenzae was the only bacterial pathogen demonstrating a propensity to be isolated from subjects with bilateral versus unilateral AOM (P < 0.0001).

TABLE 5: Relationship Between AOM Laterality and Type of Bacteria

DISCUSSION

We have shown that children with bilateral AOM were more likely than children with unilateral AOM to grow H. influenzae from cultures of MEF. Children with bilateral AOM were also younger, and they had more severe inflammation of the TM. These results agree with Howie's observations on bilateral AOM,¹⁷ in which H. influenzae was cultured more often than S. pneumoniae or M. catarrhalis. Howie et al cultured H. influenzae from 26% of subjects with bilateral AOM but only 15% of subjects with unilateral AOM. In our study, 91% (131/144) of subjects with H. influenzae isolated from one or both ears had bilateral AOM. Palmu et al,¹⁸ did not compare bilateral with unilateral AOM, but did report that H. influenzae was the most common bacterial species isolated from the MEF of children with bilateral AOM. Whereas S. pneumoniae was the most common species isolated from the MEFs of all 573 subjects in Palmu's study. Viruses play an important role in the etiology of AOM,^19–21 but prior studies have not reported associations between laterality and viral infection. We have not shown that viral etiology predisposes to either unilateral or bilateral AOM.

An unexpected result was the possible relation between pathogen and severity. In this study we have shown that, when compared with unilateral AOM, children with bilateral AOM had more inflammation of the TM, were more likely to have H. influenzae isolated, but not to have higher parent-reported symptom scores. The association with higher TM severity scores will always occur when a large enough sample is studied and TM severity is recorded as the greater both TM scores. Children with unilateral AOM, by definition, have 1 normal TM. An association of H. influenzae with bilaterality was reported in 1970 by Howie et al.¹⁷ Prior studies have attempted to relate AOM severity to pathogen. Howie et al also reported that H. influenzae AOM tended to be associated with little fever or pain.¹⁷ Several reports have shown that AOM associated with S. pneumoniae colonization results in more inflammation when compared with other bacterial pathogens, as indicated by higher mean leukocyte counts, higher mean absolute neutrophil counts, higher body temperature levels, more pain and fever, higher levels of inflammatory mediators in the MEF, and greater likelihood of bulla formation on the TM.^17,22–24 Hotomi et al²⁵ reported in a clinical AOM trial that higher symptom scores on day 5 after enrolment were more often associated with S. pneumoniae colonization on day 1. Leibovitz et al²⁶ reported that AOM caused by H. influenzae was more often associated with older age and that H. influenzae was not clinically distinguishable from S. pneumoniae. Regarding symptoms, our results agree with those of Leibovitz et al since we did not show significant differences in clinical symptom scores of children based on the pathogen isolated.

Our results lend support to Rovers' hypothesis⁷ that bilateral AOM in children is caused more often by a bacterial pathogen. This may help explain why children with bilateral AOM appear to benefit more from antibiotics than children with unilateral AOM. In addition, our studies confirm Rovers et al's observations that children with bilateral AOM are more likely to be younger than children with unilateral AOM. This finding also agrees with the findings of Howie et al,¹⁷ who reported a mean age of 2.4 years for children with bilateral AOM as compared with a mean age of 3.3 and 3.5 years in left and right unilateral AOM, respectively. Palmu et al¹⁸ reported that “culture-positive samples were more frequently associated with bulging, immobile, and fully affected tympanic membranes than were the samples with negative cultures”. Compared with unilateral AOM, our analysis showed that bilateral AOM was more frequently associated with erythema, opacification, and bulging of a TM.

Some strengths of our study are the large number of subjects evaluated, reliable clinical assessment by trained investigators, availability of risk factor data, and culture data on both bacteria and viruses. One limitation is the inherent difficulty in interpreting subjective severity of AOM symptoms in younger children, who are often unable to verbalize an exact description of their discomfort. The ETG-5 score given by the parent or guardian may not be a completely accurate interpretation of their child's symptoms. Also, measuring body temperatures at the time of presentation in the office may not be helpful, since many subjects received antipyretic medication before presentation. Similarly, parental report of fever may be inaccurate because of variations in thermometers and techniques of taking temperatures (oral, rectal, axillary) or estimating the presence of fever by touch.

In conclusion, our analysis demonstrates that bilateral AOM has significant associations with presence of bacteria in the MEF, young age, and more severe inflammation of the TM. H. influenzae appears to be the most common organism isolated in children with bilateral AOM. These findings may help to explain why children with bilateral AOM are more likely to experience persistent symptoms without antibiotic treatment.

ACKNOWLEDGMENTS

Financial support was provided in part by the National Institutes of Health, Grants RO1 DC 02620 & DC5841; the General Clinical Research Center at the University of Texas Medical Branch at Galveston, Grant M01 RR 00073, the National Center for Research Resources, NIH, USPHS; the Agency for Healthcare Research and Quality grant R01 HS10613-02; and the Clinical Research Education Program at UTMB, National Institutes of Health. K30-RR022276-08. Maroeska Rovers, PhD, provided valuable review and comment on the manuscript. Mary Whitby assisted with manuscript preparation.