https://orcid.org/0000-0002-3016-3287
Abstract
Acute otitis media (AOM) is the most common reason children are prescribed antibiotics. Bacteria that produce beta-lactamase are an increasingly frequent cause of AOM and may be resistant to amoxicillin, the currently recommended treatment for AOM. We aimed to evaluate the clinical outcomes of children treated with amoxicillin for AOM and assessed whether outcomes vary by infecting pathogen or beta-lactamase production.
205 children 6-35 months old diagnosed with AOM and prescribed amoxicillin were included. Bacterial culture and qualitative multiplex real-time polymerase chain reaction were performed on nasopharyngeal swabs collected at enrollment. Parents completed surveys assessing symptoms, antibiotic adherence, and potential adverse events. The primary outcome was treatment failure with amoxicillin. Secondary outcomes included recurrence, symptom improvement, resolution, and adverse drug events (ADE).
8 children (5.4%) experienced treatment failure and 14 (6.8%) had recurrence. By day 5, 152 (74.1%) children had symptom improvement and 97 (47.3%) had resolution. Parents reported ADE for 56 (27.3%) children. Among 149 children who did not take any amoxicillin before enrollment, 98 (65.8%) had one or more beta-lactamase-producing bacteria. Common bacterial otopathogens were Moraxella catarrhalis (79, 53.0%), Streptococcus pneumoniae (51, 34.2%), Haemophilus influenzae (30, 20.1%), and Staphylococcus aureus (21, 14.1%). Treatment failure did not differ between children that did (5, 5.1%) and did not (3, 5.9%) have beta-lactamase-producing otopathogens (p = .05).
Among children diagnosed with AOM treated with amoxicillin, treatment failure was uncommon and did not differ by pathogen or beta-lactamase production. These data support guidance recommending amoxicillin despite an increasing prevalence of beta-lactamase-producing bacteria.
INTRODUCTION
Acute otitis media (AOM) is a common childhood infection that affects over 60% of children by the age of 3 [1]. In the United States, AOM is the most frequent reason children are prescribed an antibiotic [2]. Widespread use of pneumococcal conjugate vaccines (PCV) 7 and 13 has resulted in fewer AOM episodes and changes in the epidemiology of bacterial otopathogens causing AOM [3]. Specifically, the proportion of cases caused by Streptococcus pneumoniae has decreased and the proportion caused by Haemophilus influenzae and Moraxella catarrhalis has increased [1, 3, 4]. AOM due to H. influenzae and M. catarrhalis may result in milder symptoms, a lower risk for tympanic membrane (TM) perforation, and increased likelihood of resolution without an antibiotic compared to S. pneumoniae [5–10]. However, some H. influenzae and most M. catarrhalis isolates produce beta-lactamase, resulting in potential resistance to amoxicillin, the antibiotic currently recommended by the American Academy of Pediatrics (AAP) to treat AOM [11]. Some therefore suggest that first-line treatment should be changed to amoxicillin-clavulanate or a cephalosporin, antibiotics unaffected by beta-lactamase production [4]. Routine use of broader-spectrum agents like amoxicillin-clavulanate has the potential to increase adverse drug events (ADE) [12, 13], Clostridioides difficile infections [14], and the development of antibiotic-resistant bacteria [15]. Thus, any potential benefit ineffectiveness must be weighed against potential harms.
No randomized trials in the PCV era have evaluated the efficacy of amoxicillin versus amoxicillin-clavulanate or other broader-spectrum antibiotics for treating AOM. Additionally, most prior randomized trials included children who met strict diagnostic criteria [16–18], the results of which may not be generalizable to children diagnosed in typical settings. Gerber and colleagues reported that, among a prospective cohort of children diagnosed with AOM in routine clinical practice, the failure rate was low and not significantly different with narrower- versus broader-spectrum antibiotics (3.7 vs 4.6%) [12]. We reported similar findings among over one million children in a national payer claims database [19]. However, due to the lack of microbiological data in both studies, neither assessed amoxicillin treatment failure by associated pathogens. We, therefore, aimed to determine outcomes in a prospective cohort of children treated with amoxicillin after AOM diagnosis in clinical practice and evaluate whether outcomes differed among children with potentially amoxicillin-resistant bacterial otopathogens.
METHODS
Study Design and Population
This prospective, observational cohort study took place at Denver Health and Hospital Authority (DHHA) in Denver, Colorado from January 2018 to May 2022. Children were enrolled in primary and urgent care clinics, and an emergency department by trained research staff located in clinical care areas or by contacting parents by phone within 24 hours of an encounter for AOM. Eligible children were 6-35 months of age, had an International Classification of Diseases, 10th Revision code diagnosis for AOM (Supplementary material), were prescribed amoxicillin by their clinician, and had taken two or fewer doses at enrollment. We aimed to evaluate treatment failure in routine clinical practice where it is estimated that 30-50% of children diagnosed with AOM do not meet AAP diagnostic criteria. Thus, any child with clinician-diagnosed and -treated AOM was eligible irrespective of whether AAP diagnostic criteria [11] were met. We also conducted sub-analyses to compare outcomes for children who did and did not meet AAP diagnostic criteria. Children who were given a delayed prescription were excluded as our primary objective was to evaluate outcomes in children taking amoxicillin. We excluded children with complicating factors including TM rupture, symptom onset over 10 days before enrollment, a concurrent bacterial diagnosis that warranted a systemic antibiotic (eg, pneumonia), treatment with a systemic antibiotic in the preceding 30 days, corticosteroid use, immunocompromise, underlying ear abnormalities (eg, Down syndrome, cleft palate, etc.), or current tympanostomy tubes.
Patient Reported Outcomes and Electronic Data Abstraction
Parents completed surveys at enrollment, and 5, 14, and 30 days after enrollment. Surveys were administered via REDCap [20, 21], were available in English or Spanish, and included questions on whether and how the prescribed amoxicillin was taken, severity and duration of symptoms, potential ADE, missed work and/or daycare, and potential complications of AOM (Supplementary material). At enrollment, day 5, and day 14, parents completed the Acute Otitis Media Severity of Symptoms (AOM-SOS) scale version 5.0 (Copyright 2016, University of Pittsburgh. All rights reserved) [22, 23]. Clinical data were abstracted from the electronic health record (EHR) at enrollment and Day 30, including medical history, physical examination findings, antibiotics prescribed, and encounters and diagnoses within the 30-day follow-up period.
Sample Collection and Microbiology Testing
All children had a nasopharyngeal (NP) flocked swab (Eswab®, Copan Diagnostics) collected at enrollment and before more than two doses of amoxicillin were administered. NP sampling was chosen over middle ear fluid sampling because (1) tympanocentesis to collect middle ear fluid can have a therapeutic effect, directly affecting outcome measures and (2) bacterial otopathogens in the middle ear during AOM are reliably present in the NP (negative predictive value > 92%), though use of NP samples can result in some overestimation of middle ear fluid bacterial otopathogens [24]. Samples were transported to the laboratory within one hour of collection and aliquoted for culture and storage at −80°C. Culture was performed by inoculating 10µL of specimen on blood, chocolate, and MacConkey agars, and streaking for isolation. Media were incubated at 35-37°C in 5-10% CO2 and examined at 24, 48, and 72 h for growth. Bacteria were identified by standard methods including biochemical and agglutination tests, supplemented when appropriate by the use of Vitek cards (Biomerieux, Durham, NC) and MicroScan panels (Beckman Coulter, Brea, California). Results were recorded semiquantitatively (rare, few, moderate, and many). S. pneumoniae isolates were tested for penicillin resistance and H. influenzae isolates were tested for beta-lactamase production. Beta-lactamase-producing bacteria were defined as H. influenzae isolates testing positive for beta-lactamase production, M. catarrhalis, or Staphylococcus aureus.
Qualitative multiplex real-time polymerase chain reaction (RT-PCR) for S. pneumoniae, H. influenzae, M. catarrhalis, S. aureus, and 11 respiratory viruses (including SARS-CoV-2) was completed using Lyra® (Quidel, San Diego, CA) and AnDiaTec® assay kits (Quidel Germany GmbH, Kornwestheim, Germany). Nucleic acids were extracted using the NucliSENS® easyMAG® system (Quidel, San Diego, CA). Nucleic acid amplification and detection were performed on the Applied Biosystems® (ABI) 7500 Fast Dx Real-Time PCR Instrument.
A Priori Outcome Measures
The primary outcome was treatment failure, defined as the need for a new antibiotic prescription for AOM 2-14 days after the index encounter [12, 19]. Secondary outcome measures included improvement in AOM-SOS scores by at least 20% from baseline or a score ≤2 at days 5 and 14, improvement in AOM-SOS scores by at least 50% from baseline or a score ≤2 at days 5 and 14, parent-reported time to symptom improvement and resolution, parent-reported ADE, and recurrence, defined as the need for a new antibiotic prescription for AOM 15-30 days after the index encounter [12, 19].
Sample Size Estimate
An a priori sample size estimate of 300 patients was based on an alpha of 0.05 and 90% power to detect a 10% delta around a predicted 30% treatment failure rate. Observed treatment failure rates were lower than predicted. Therefore, sample size and power estimates for the primary endpoint were re-evaluated using a 15% predicted treatment failure rate in the Wald one-sample proportions test with two-sided estimates, which resulted in a sample size target of 207 patients needed to power the study for the primary endpoint.
To compare treatment failure rates relative to beta-lactamase production, we hypothesized rates of 20% for children with beta-lactamase-producing organisms and 5% for children without. Using power analysis for a two-sample-proportion likelihood-ratio test and an assumption that enrollment of children relative to beta-lactamase production would be approximately equal, a sample size of 150 was calculated to provide 80% power at an alpha of 0.05.
Data Analysis
Analyses were conducted using Stata 15.1 software (StataCorp, TX, USA). Missing values were multiply imputed using an iterative Markov Monte Carlo method (Supplementary material). We calculated descriptive statistics for demographics, clinical characteristics, and duration of amoxicillin prescribed. These parameters were also assessed for association with treatment failure, recurrence, symptom improvement, and parent-reported ADE via multivariate logistic or Poisson regression and with a number of covariates (Supplementary material).
Primary and secondary outcomes were analyzed. Prevalence of specific otopathogens and treatment failure rate by pathogen were evaluated only in the subgroup of children who did not take any doses of amoxicillin prior to NP specimen collection. For the primary analysis of treatment failure by specific bacteria, only bacteria that grew in culture were included (ie, specimens positive by PCR only were excluded). We conducted secondary analyses to evaluate treatment failure by organisms detected with PCR. Odds ratios for treatment failure were calculated using logistic regression models for individuals and groups of otopathogens using referents that were negative for that otopathogen or group. Models were adjusted for age, whether AAP diagnostic criteria [11] were met, viral co-infection, language spoken, and laterality of infection. Smoke exposure, sex, race, ethnicity, and pre-COVID or COVID periods were not included in the models due to collinearity. Additional sub-analyses were conducted to evaluate thresholds for success using percent decreases in AOM-SOS scores (Supplementary material) and to compare outcomes between children who did or did not meet AAP diagnostic criteria (Pearson Chi-Squared tests for independence). For comparative analyses, significance was defined at an alpha of < 0.05 using two-tailed tests.
Informed consent was obtained from parents of participating children. The study was approved by the Colorado Multiple Institutional Review Board.
RESULTS
Of 1,604 children diagnosed with AOM, 944 were potentially eligible by initial EHR review. Of these, 334 could not be reached to verify eligibility and 395 declined to participate. In total, 205 children were enrolled, and an NP specimen collected (Figure 1). Survey completion rates were 95% on day 5, 92% on day 14, and 87% on day 30. The mean age of participating children was 17.4 months, and three-quarters were Hispanic (Table 1). In one-third of cases, AAP criteria for AOM were not met, primarily due to the absence of TM bulging.
Demographic and Clinical Characteristics of Participating Childrena
| Characteristic | All Children (N = 205) |
|---|---|
| Age (mean, SD) | 17.4 (8.4) |
| 6-11 months | 64 (31.2) |
| 2-23 months | 90 (43.9) |
| 24-35 months | 51 (24.88) |
| Sex (n (%)) | |
| Female | 118 (57.6) |
| Male | 87 (42.4) |
| Race (n (%)) | |
| White | 117 (57.1) |
| Unknown or not reported | 65 (31.7) |
| African American/Black | 14 (6.8) |
| Other or multiracial | 9 (4.4) |
| Ethnicity (n (%)) | |
| Hispanic/Latinx | 155 (75.6) |
| Not Hispanic/Latinx | 48 (23.4) |
| Unknown or not reported | 2 (1.0) |
| Language preferenceb (n (%)) | |
| English | 152 (74.1) |
| Spanish | 53 (25.8) |
| Attends school or daycare (n (%)) | 54 (26.3) |
| Diagnostic setting (n (%)) | |
| Emergency department or urgent care | 135 (65.8) |
| Pediatric primary care clinic | 39 (19.0) |
| Family medicine primary care clinic | 31 (15.1) |
| Time periodc (n (%)) | |
| Prior to emergence of SARS-CoV-2 | 101 (49.3) |
| After the emergence of SARS-CoV-2 | 104 (50.7) |
| Diagnostic criteria used (n (%))d,e | |
| AAP diagnostic criteria for AOM met | 137 (66.8) |
| AAP diagnostic criteria for AOM not met | 68 (33.2) |
| Number of ears infected (n (%)) | |
| Unilateral | 131 (63.9) |
| Bilateral | 56 (27.3) |
| Unspecified | 18 (8.8) |
| Severity of infectionf (n (%)) | |
| Severe | 103 (50.2) |
| Not severe | 102 (49.8) |
| AOM-SOS at diagnosis (mean (SD)) | 14.3 (5.1) |
| AOM-SOS categories at diagnosis (n (%)) | |
| 0-5 | 15 (7.3) |
| 6-10 | 31 (15.1) |
| 11-15 | 68 (33.2) |
| 16-20 | 74 (36.1) |
| 21-25 | 17 (8.3) |
| Antibiotics use (n (%))g | |
| Participants who took antibiotics | 198 (96.6) |
| Participants who did not take antibiotics | 7 (3.4) |
| Number of antibiotic doses taken (mean (SD))h | 15.7 (5.7) |
| Duration of antibiotics taken (median days (IQR)) | 8 (4.5) |
| Characteristic | All Children (N = 205) |
|---|---|
| Age (mean, SD) | 17.4 (8.4) |
| 6-11 months | 64 (31.2) |
| 2-23 months | 90 (43.9) |
| 24-35 months | 51 (24.88) |
| Sex (n (%)) | |
| Female | 118 (57.6) |
| Male | 87 (42.4) |
| Race (n (%)) | |
| White | 117 (57.1) |
| Unknown or not reported | 65 (31.7) |
| African American/Black | 14 (6.8) |
| Other or multiracial | 9 (4.4) |
| Ethnicity (n (%)) | |
| Hispanic/Latinx | 155 (75.6) |
| Not Hispanic/Latinx | 48 (23.4) |
| Unknown or not reported | 2 (1.0) |
| Language preferenceb (n (%)) | |
| English | 152 (74.1) |
| Spanish | 53 (25.8) |
| Attends school or daycare (n (%)) | 54 (26.3) |
| Diagnostic setting (n (%)) | |
| Emergency department or urgent care | 135 (65.8) |
| Pediatric primary care clinic | 39 (19.0) |
| Family medicine primary care clinic | 31 (15.1) |
| Time periodc (n (%)) | |
| Prior to emergence of SARS-CoV-2 | 101 (49.3) |
| After the emergence of SARS-CoV-2 | 104 (50.7) |
| Diagnostic criteria used (n (%))d,e | |
| AAP diagnostic criteria for AOM met | 137 (66.8) |
| AAP diagnostic criteria for AOM not met | 68 (33.2) |
| Number of ears infected (n (%)) | |
| Unilateral | 131 (63.9) |
| Bilateral | 56 (27.3) |
| Unspecified | 18 (8.8) |
| Severity of infectionf (n (%)) | |
| Severe | 103 (50.2) |
| Not severe | 102 (49.8) |
| AOM-SOS at diagnosis (mean (SD)) | 14.3 (5.1) |
| AOM-SOS categories at diagnosis (n (%)) | |
| 0-5 | 15 (7.3) |
| 6-10 | 31 (15.1) |
| 11-15 | 68 (33.2) |
| 16-20 | 74 (36.1) |
| 21-25 | 17 (8.3) |
| Antibiotics use (n (%))g | |
| Participants who took antibiotics | 198 (96.6) |
| Participants who did not take antibiotics | 7 (3.4) |
| Number of antibiotic doses taken (mean (SD))h | 15.7 (5.7) |
| Duration of antibiotics taken (median days (IQR)) | 8 (4.5) |
aNo missing data noted for ITT.
bNone coded for “other.”
cBased on study enrollment date of pre-covid <March 1, 2020 and peri-covid ≥ March 1, 2020.
dAmerican Academy of Pediatrics (AAP) criteria used was bulging tympanum in either ear.
eMissing data for two patients were imputed by multiple imputation.
fUsing criteria: AOM-SOS version 5.0 vs 4.0: 0-25 on 5.0. 1. Score of 4 or 5 on ear tugging and/or score of 4 or 5 on fever.
gMissing data for 10 patients were imputed by multiple imputation.
hFor patients who took antibiotics.
Demographic and Clinical Characteristics of Participating Childrena
| Characteristic | All Children (N = 205) |
|---|---|
| Age (mean, SD) | 17.4 (8.4) |
| 6-11 months | 64 (31.2) |
| 2-23 months | 90 (43.9) |
| 24-35 months | 51 (24.88) |
| Sex (n (%)) | |
| Female | 118 (57.6) |
| Male | 87 (42.4) |
| Race (n (%)) | |
| White | 117 (57.1) |
| Unknown or not reported | 65 (31.7) |
| African American/Black | 14 (6.8) |
| Other or multiracial | 9 (4.4) |
| Ethnicity (n (%)) | |
| Hispanic/Latinx | 155 (75.6) |
| Not Hispanic/Latinx | 48 (23.4) |
| Unknown or not reported | 2 (1.0) |
| Language preferenceb (n (%)) | |
| English | 152 (74.1) |
| Spanish | 53 (25.8) |
| Attends school or daycare (n (%)) | 54 (26.3) |
| Diagnostic setting (n (%)) | |
| Emergency department or urgent care | 135 (65.8) |
| Pediatric primary care clinic | 39 (19.0) |
| Family medicine primary care clinic | 31 (15.1) |
| Time periodc (n (%)) | |
| Prior to emergence of SARS-CoV-2 | 101 (49.3) |
| After the emergence of SARS-CoV-2 | 104 (50.7) |
| Diagnostic criteria used (n (%))d,e | |
| AAP diagnostic criteria for AOM met | 137 (66.8) |
| AAP diagnostic criteria for AOM not met | 68 (33.2) |
| Number of ears infected (n (%)) | |
| Unilateral | 131 (63.9) |
| Bilateral | 56 (27.3) |
| Unspecified | 18 (8.8) |
| Severity of infectionf (n (%)) | |
| Severe | 103 (50.2) |
| Not severe | 102 (49.8) |
| AOM-SOS at diagnosis (mean (SD)) | 14.3 (5.1) |
| AOM-SOS categories at diagnosis (n (%)) | |
| 0-5 | 15 (7.3) |
| 6-10 | 31 (15.1) |
| 11-15 | 68 (33.2) |
| 16-20 | 74 (36.1) |
| 21-25 | 17 (8.3) |
| Antibiotics use (n (%))g | |
| Participants who took antibiotics | 198 (96.6) |
| Participants who did not take antibiotics | 7 (3.4) |
| Number of antibiotic doses taken (mean (SD))h | 15.7 (5.7) |
| Duration of antibiotics taken (median days (IQR)) | 8 (4.5) |
| Characteristic | All Children (N = 205) |
|---|---|
| Age (mean, SD) | 17.4 (8.4) |
| 6-11 months | 64 (31.2) |
| 2-23 months | 90 (43.9) |
| 24-35 months | 51 (24.88) |
| Sex (n (%)) | |
| Female | 118 (57.6) |
| Male | 87 (42.4) |
| Race (n (%)) | |
| White | 117 (57.1) |
| Unknown or not reported | 65 (31.7) |
| African American/Black | 14 (6.8) |
| Other or multiracial | 9 (4.4) |
| Ethnicity (n (%)) | |
| Hispanic/Latinx | 155 (75.6) |
| Not Hispanic/Latinx | 48 (23.4) |
| Unknown or not reported | 2 (1.0) |
| Language preferenceb (n (%)) | |
| English | 152 (74.1) |
| Spanish | 53 (25.8) |
| Attends school or daycare (n (%)) | 54 (26.3) |
| Diagnostic setting (n (%)) | |
| Emergency department or urgent care | 135 (65.8) |
| Pediatric primary care clinic | 39 (19.0) |
| Family medicine primary care clinic | 31 (15.1) |
| Time periodc (n (%)) | |
| Prior to emergence of SARS-CoV-2 | 101 (49.3) |
| After the emergence of SARS-CoV-2 | 104 (50.7) |
| Diagnostic criteria used (n (%))d,e | |
| AAP diagnostic criteria for AOM met | 137 (66.8) |
| AAP diagnostic criteria for AOM not met | 68 (33.2) |
| Number of ears infected (n (%)) | |
| Unilateral | 131 (63.9) |
| Bilateral | 56 (27.3) |
| Unspecified | 18 (8.8) |
| Severity of infectionf (n (%)) | |
| Severe | 103 (50.2) |
| Not severe | 102 (49.8) |
| AOM-SOS at diagnosis (mean (SD)) | 14.3 (5.1) |
| AOM-SOS categories at diagnosis (n (%)) | |
| 0-5 | 15 (7.3) |
| 6-10 | 31 (15.1) |
| 11-15 | 68 (33.2) |
| 16-20 | 74 (36.1) |
| 21-25 | 17 (8.3) |
| Antibiotics use (n (%))g | |
| Participants who took antibiotics | 198 (96.6) |
| Participants who did not take antibiotics | 7 (3.4) |
| Number of antibiotic doses taken (mean (SD))h | 15.7 (5.7) |
| Duration of antibiotics taken (median days (IQR)) | 8 (4.5) |
aNo missing data noted for ITT.
bNone coded for “other.”
cBased on study enrollment date of pre-covid <March 1, 2020 and peri-covid ≥ March 1, 2020.
dAmerican Academy of Pediatrics (AAP) criteria used was bulging tympanum in either ear.
eMissing data for two patients were imputed by multiple imputation.
fUsing criteria: AOM-SOS version 5.0 vs 4.0: 0-25 on 5.0. 1. Score of 4 or 5 on ear tugging and/or score of 4 or 5 on fever.
gMissing data for 10 patients were imputed by multiple imputation.
hFor patients who took antibiotics.
Study Schematic.
Antibiotic Usage
The median duration of amoxicillin taken was 8 days (4.5) (Table 1); 27 (13%), 16 (8%), and 162 (79%) of the prescriptions were written for 5, 7, or 10 days, respectively (Supplementary material). Parents reported giving the antibiotic for the number of days that were prescribed in 44% of cases. Children took the prescribed number of antibiotic doses for 5-, 7-, and 10-day prescriptions in 59.3% (median 5 days), 37.5% (median 7 days), and 42.0% (median 9.75 days) of cases, respectively (Supplementary material). Seven children were not given any of the prescribed antibiotics.
Primary and Secondary Outcomes
Treatment failure occurred in 11 (5.4%) children (Table 2). Treatment failure rate was similar among children who did and did not meet AAP diagnostic criteria (5.8% and 4.4%, respectively) (Supplementary material). Neither patient nor clinical covariates were associated with treatment failure in logistic regression analysis.
Frequency of Primary and Secondary Outcomes (Missing Data Imputed)
| Outcome | All Patients (N = 205) |
|---|---|
| Primary outcome | |
| Treatment failurea,b | 11 (5.4) |
| Secondary outcomes | |
| Symptom severity—day 5 | |
| AOM-SOS Score (mean, SD) | 4.3 (4.9) |
| Change from baseline (mean, SD) | −9.9 (6.5) |
| Score ≤ 2 (n (%)) | 97 (47.3) |
| Score ≤ 2 or decrease of > 20% from baseline (n (%)) | 180 (87.8) |
| Score ≤ 2 or decrease of > 50% from baseline (n (%)) | 152 (74.1) |
| Symptom severity—day 14 | |
| AOM-SOS Score (mean, SD) | 2.4 (4.2) |
| Change from baseline (mean, SD) | −11.9 (6.2) |
| Score ≤ 2 (n (%)) | 193 (94.1) |
| Score ≤ 2 or decrease of > 20% from baseline (n (%)) | 180 (87.8) |
| Score ≤ 2 or decrease of > 50% from baseline (n (%)) | 154 (75.1) |
| Days to symptom improvement (parent reported) (median, IQR)c | 3 (1) |
| Days to symptom resolution (parent reported) (median, IQR) | 7 (3) |
| Adverse drug eventsd | |
| Parent-reported adverse drug events (n (%)) | 56 (27.3) |
| Total number of adverse drug events (n) | 74 |
| Diarrhea (n (%)) | 33 (16.1) |
| Diaper dermatitis (n (%)) | 16 (7.8) |
| Upset stomach or vomiting (n (%)) | 13 (6.3) |
| Body rash (n (%)) | 12 (5.9) |
| Severe allergic reaction requiring medical intervention (n (%)) | 0 (0.0) |
| Need for a follow-up visit due to adverse event (n (%))e | 18 (8.8) |
| Recurrence (n (%))f | 14 (6.8) |
| Treatment failure (0-14 days) or Recurrence (15-30 days) (n (n%)) | 24 (11.7%) |
| Outcome | All Patients (N = 205) |
|---|---|
| Primary outcome | |
| Treatment failurea,b | 11 (5.4) |
| Secondary outcomes | |
| Symptom severity—day 5 | |
| AOM-SOS Score (mean, SD) | 4.3 (4.9) |
| Change from baseline (mean, SD) | −9.9 (6.5) |
| Score ≤ 2 (n (%)) | 97 (47.3) |
| Score ≤ 2 or decrease of > 20% from baseline (n (%)) | 180 (87.8) |
| Score ≤ 2 or decrease of > 50% from baseline (n (%)) | 152 (74.1) |
| Symptom severity—day 14 | |
| AOM-SOS Score (mean, SD) | 2.4 (4.2) |
| Change from baseline (mean, SD) | −11.9 (6.2) |
| Score ≤ 2 (n (%)) | 193 (94.1) |
| Score ≤ 2 or decrease of > 20% from baseline (n (%)) | 180 (87.8) |
| Score ≤ 2 or decrease of > 50% from baseline (n (%)) | 154 (75.1) |
| Days to symptom improvement (parent reported) (median, IQR)c | 3 (1) |
| Days to symptom resolution (parent reported) (median, IQR) | 7 (3) |
| Adverse drug eventsd | |
| Parent-reported adverse drug events (n (%)) | 56 (27.3) |
| Total number of adverse drug events (n) | 74 |
| Diarrhea (n (%)) | 33 (16.1) |
| Diaper dermatitis (n (%)) | 16 (7.8) |
| Upset stomach or vomiting (n (%)) | 13 (6.3) |
| Body rash (n (%)) | 12 (5.9) |
| Severe allergic reaction requiring medical intervention (n (%)) | 0 (0.0) |
| Need for a follow-up visit due to adverse event (n (%))e | 18 (8.8) |
| Recurrence (n (%))f | 14 (6.8) |
| Treatment failure (0-14 days) or Recurrence (15-30 days) (n (n%)) | 24 (11.7%) |
aTreatment failure is defined by requiring new abx for AOM by day 14—NOT for adverse drug events; either diffabx OR abxchange =1 at 0-14 days.
bMissing data for 17 patients were imputed by multiple imputation.
cCalculated from baseline visit date to date improvement in symptoms was observed by parents.
dDay 14 ADEs were not available.
eOut of all patients, parent-reported.
fCalculated from new antibiotic required for a new ear infection, per patient survey (newabx) or chart review (abxear30).
Frequency of Primary and Secondary Outcomes (Missing Data Imputed)
| Outcome | All Patients (N = 205) |
|---|---|
| Primary outcome | |
| Treatment failurea,b | 11 (5.4) |
| Secondary outcomes | |
| Symptom severity—day 5 | |
| AOM-SOS Score (mean, SD) | 4.3 (4.9) |
| Change from baseline (mean, SD) | −9.9 (6.5) |
| Score ≤ 2 (n (%)) | 97 (47.3) |
| Score ≤ 2 or decrease of > 20% from baseline (n (%)) | 180 (87.8) |
| Score ≤ 2 or decrease of > 50% from baseline (n (%)) | 152 (74.1) |
| Symptom severity—day 14 | |
| AOM-SOS Score (mean, SD) | 2.4 (4.2) |
| Change from baseline (mean, SD) | −11.9 (6.2) |
| Score ≤ 2 (n (%)) | 193 (94.1) |
| Score ≤ 2 or decrease of > 20% from baseline (n (%)) | 180 (87.8) |
| Score ≤ 2 or decrease of > 50% from baseline (n (%)) | 154 (75.1) |
| Days to symptom improvement (parent reported) (median, IQR)c | 3 (1) |
| Days to symptom resolution (parent reported) (median, IQR) | 7 (3) |
| Adverse drug eventsd | |
| Parent-reported adverse drug events (n (%)) | 56 (27.3) |
| Total number of adverse drug events (n) | 74 |
| Diarrhea (n (%)) | 33 (16.1) |
| Diaper dermatitis (n (%)) | 16 (7.8) |
| Upset stomach or vomiting (n (%)) | 13 (6.3) |
| Body rash (n (%)) | 12 (5.9) |
| Severe allergic reaction requiring medical intervention (n (%)) | 0 (0.0) |
| Need for a follow-up visit due to adverse event (n (%))e | 18 (8.8) |
| Recurrence (n (%))f | 14 (6.8) |
| Treatment failure (0-14 days) or Recurrence (15-30 days) (n (n%)) | 24 (11.7%) |
| Outcome | All Patients (N = 205) |
|---|---|
| Primary outcome | |
| Treatment failurea,b | 11 (5.4) |
| Secondary outcomes | |
| Symptom severity—day 5 | |
| AOM-SOS Score (mean, SD) | 4.3 (4.9) |
| Change from baseline (mean, SD) | −9.9 (6.5) |
| Score ≤ 2 (n (%)) | 97 (47.3) |
| Score ≤ 2 or decrease of > 20% from baseline (n (%)) | 180 (87.8) |
| Score ≤ 2 or decrease of > 50% from baseline (n (%)) | 152 (74.1) |
| Symptom severity—day 14 | |
| AOM-SOS Score (mean, SD) | 2.4 (4.2) |
| Change from baseline (mean, SD) | −11.9 (6.2) |
| Score ≤ 2 (n (%)) | 193 (94.1) |
| Score ≤ 2 or decrease of > 20% from baseline (n (%)) | 180 (87.8) |
| Score ≤ 2 or decrease of > 50% from baseline (n (%)) | 154 (75.1) |
| Days to symptom improvement (parent reported) (median, IQR)c | 3 (1) |
| Days to symptom resolution (parent reported) (median, IQR) | 7 (3) |
| Adverse drug eventsd | |
| Parent-reported adverse drug events (n (%)) | 56 (27.3) |
| Total number of adverse drug events (n) | 74 |
| Diarrhea (n (%)) | 33 (16.1) |
| Diaper dermatitis (n (%)) | 16 (7.8) |
| Upset stomach or vomiting (n (%)) | 13 (6.3) |
| Body rash (n (%)) | 12 (5.9) |
| Severe allergic reaction requiring medical intervention (n (%)) | 0 (0.0) |
| Need for a follow-up visit due to adverse event (n (%))e | 18 (8.8) |
| Recurrence (n (%))f | 14 (6.8) |
| Treatment failure (0-14 days) or Recurrence (15-30 days) (n (n%)) | 24 (11.7%) |
aTreatment failure is defined by requiring new abx for AOM by day 14—NOT for adverse drug events; either diffabx OR abxchange =1 at 0-14 days.
bMissing data for 17 patients were imputed by multiple imputation.
cCalculated from baseline visit date to date improvement in symptoms was observed by parents.
dDay 14 ADEs were not available.
eOut of all patients, parent-reported.
fCalculated from new antibiotic required for a new ear infection, per patient survey (newabx) or chart review (abxear30).
Mean AOM-SOS score at enrollment was 14.3. Symptom severity decreased to a mean AOM-SOS score of 4.3 at day 5 and 2.4 at day 14 (Table 2). Overall, 47% of children had symptomatic resolution (AOM-SOS ≤ 2) by day 5 and 75% by day 14. Using criteria for success of an AOM-SOS score 2 or a decrease of more than 20% from baseline, 88% of children had success at day 5 and 94% at day 14. Alternatively, using criteria of an AOM-SOS score ≤2 or a decrease of more than 50% from baseline, 74% and 88% had success at day 5 and day 14, respectively (Supplementary material). The median parent-reported days to improvement and resolution were 3 (IQR: 1) and 7 (IQR: 3), respectively.
Parent-reported ADE occurred in 56 (27.3%) children (Table 2). The most common adverse event was diarrhea (33, 16.1%). There were no severe allergic reactions; however, parents reported seeking medical care due to an ADE in 18 (8.8%) cases.
AOM recurrence occurred in 14 (6.8%) children. Children who met AAP diagnostic criteria were more likely to have a recurrent infection than those who did not (13/137, 9.5% vs 1/68, 1.5%) (Supplementary material). Male sex, children seen in a primary care setting rather than an emergent setting, meeting AAP diagnostic criteria, and a high day 14 AOM-SOS score (p-values < 0.05) were associated with increased risk of recurrence. When the children with a treatment failure during days 0-14 were removed from the analysis of recurrence, recurrence was 6.7% (13/194) as one child was both a treatment failure and had a recurrent infection. Over the course of 30 days participating in the study, 24 (11.7%) children required a new or different antibiotic for treatment failure or recurrence of AOM.
Clinical Outcomes by Otopathogen
Among 149 children who did not receive any amoxicillin prior to specimen collection, one or more bacteria or virus were identified in 123 (82.6%) and 71 (47.6%) specimens, respectively (Table 3). M. catarrhalis was the most frequently identified bacteria (78, 52.3%), followed by S. pneumoniae (51, 34.2%) and H. influenzae (30, 20.1%).
Treatment Failure by Microorganism in Patients who did not Receive Antibiotics Prior to Specimen Collectiona
| Organism | Organism Prevalence (N = 149) | Treatment Success (N = 148) | Treatment Failure (N = 8) | OR (95% CI)b | aOR (95% CI)b,c | aOR (95% CI)b,d |
|---|---|---|---|---|---|---|
| Microorganism types (n (%, row))a,e | ||||||
| No microorganism identified | 14 (9.4) | 13 (92.9) | 1 (7.1) | Referent | Referent | Referent |
| Any bacteria | 123 (82.6) | 117 (95.1) | 7 (5.6) | .67 (.07, 5.98) | .66 (.06, 7.43) | 1.05 (.10, 10.84) |
| Any virusf | 71 (47.6) | 68 (95.8) | 3 (4.2) | .58 (.06, 6.01) | .003 (3e-7, 28.71) | .52 (.04, 6.75) |
| Only bacteria | 64 (42.9) | 60 (92.3) | 4 (6.2) | .87 (.09, 8.40) | .68 (.05, 8.56) | 1.00 (.0.10, 10.69) |
| Only virus(es) | 12 (8.0) | 11 (91.7) | 1 (8.3) | 1.18 (.07, 21.17) | NCg | NC |
| Either bacteria or virus(es) | 135 (90.6) | 128 (94.8) | 7 (5.2) | .71 (.08, 6.24) | .57 (.05, 6.48) | 1.02 (.10, 10.52) |
| Both bacteria and virus(es) | 59 (39.6) | 57 (96.6) | 2 (3.4) | .46 (.04, 5.42) | .004 (3e-7, 42.46) | .51 (.03, 7.58) |
| Bacterial pathogens (n(%))h,i | ||||||
| No bacteria identifiedj | 26 (17.4) | 24 (92.3) | 2 (7.7) | Referent | Referent | Referent |
| S. pneumoniae | 51 (34.2) | 49 (96.1) | 2 (3.9) | .49 (.06, 3.69) | .43 (.03, 5.27) | .58 (.07, 4.83) |
| H. influenzae | 30 (20.1) | 29 (96.7) | 1 (3.3) | .41 (.04, 4.85) | .47 (.03, 8.48) | .63 (.04, 9.25) |
| No beta-lactamase productionk | 19 (63.3) | 18 (94.7) | 1 (5.3) | .67 (.056, 7.94) | .89 (.04, 18.48) | .87 (.06, 12.84) |
| Beta-lactamase productionk | 11 (36.7) | 11 (100.0) | 0 (0.0) | NC | NC | NC |
| M. catarrhalis | 78 (52.3) | 74 (94.9) | 4 (5.1) | .65 (.11, 3.77) | .41 (.03, 5.01) | 1.15 (.17, 7.65) |
| S. aureus | 21 (14.1) | 20 (95.2) | 1 (4.8) | .60 (.05, 7.11) | .68 (.02, 20.0) | .90 (.05, 14.75) |
| Beta-lactamase-producing bacteriall | 98 (65.8) | 93 (94.9) | 5 (5.1) | .65 (.12, 3.53) | .55 (.08, 4.07) | .98 (.16, 5.90) |
| Non-beta-lactamase-producing bacteria | 51 (34.2) | 48 (94.1) | 3 (5.9) | .50 (.04, 5.89) | .50 (.03, 9.7) | .41 (.03, 5.65) |
| Bacterial pathogens (n(%)) | ||||||
| No bacteria identified | 25 (16.8) | 24 (96.0) | 1 (4.0) | Referent | Referent | Referent |
| Only S. pneumoniae | 2 (1.3) | 2 (100.0) | 0 (0.0) | NC | NC | NC |
| Only H. influenzae | 9 (6.0) | 9 (100.0) | 0 (0.0) | NC | NC | NC |
| No beta-lactamase productionm | 5 (55.6) | 5 (100.0) | 0 (0.0) | NC | NC | NC |
| Beta-lactamase productionm | 4 (44.4) | 4 (100.0) | 0 (0.0) | NC | NC | NC |
| Only M. catarrahlis | 17 (1.4) | 15 (88.2) | 2 (11.8) | 1.6 (.20, 12.60) | NC | 21.13 (.18, 2397.91) |
| Only S. aureus | 5 (3.4) | 4 (80.0) | 1 (20.0) | 3.0 (.22, 41.35) | 3.80 (.06, 240.03) | 5.84 (.16, 219.18) |
| Organism | Organism Prevalence (N = 149) | Treatment Success (N = 148) | Treatment Failure (N = 8) | OR (95% CI)b | aOR (95% CI)b,c | aOR (95% CI)b,d |
|---|---|---|---|---|---|---|
| Microorganism types (n (%, row))a,e | ||||||
| No microorganism identified | 14 (9.4) | 13 (92.9) | 1 (7.1) | Referent | Referent | Referent |
| Any bacteria | 123 (82.6) | 117 (95.1) | 7 (5.6) | .67 (.07, 5.98) | .66 (.06, 7.43) | 1.05 (.10, 10.84) |
| Any virusf | 71 (47.6) | 68 (95.8) | 3 (4.2) | .58 (.06, 6.01) | .003 (3e-7, 28.71) | .52 (.04, 6.75) |
| Only bacteria | 64 (42.9) | 60 (92.3) | 4 (6.2) | .87 (.09, 8.40) | .68 (.05, 8.56) | 1.00 (.0.10, 10.69) |
| Only virus(es) | 12 (8.0) | 11 (91.7) | 1 (8.3) | 1.18 (.07, 21.17) | NCg | NC |
| Either bacteria or virus(es) | 135 (90.6) | 128 (94.8) | 7 (5.2) | .71 (.08, 6.24) | .57 (.05, 6.48) | 1.02 (.10, 10.52) |
| Both bacteria and virus(es) | 59 (39.6) | 57 (96.6) | 2 (3.4) | .46 (.04, 5.42) | .004 (3e-7, 42.46) | .51 (.03, 7.58) |
| Bacterial pathogens (n(%))h,i | ||||||
| No bacteria identifiedj | 26 (17.4) | 24 (92.3) | 2 (7.7) | Referent | Referent | Referent |
| S. pneumoniae | 51 (34.2) | 49 (96.1) | 2 (3.9) | .49 (.06, 3.69) | .43 (.03, 5.27) | .58 (.07, 4.83) |
| H. influenzae | 30 (20.1) | 29 (96.7) | 1 (3.3) | .41 (.04, 4.85) | .47 (.03, 8.48) | .63 (.04, 9.25) |
| No beta-lactamase productionk | 19 (63.3) | 18 (94.7) | 1 (5.3) | .67 (.056, 7.94) | .89 (.04, 18.48) | .87 (.06, 12.84) |
| Beta-lactamase productionk | 11 (36.7) | 11 (100.0) | 0 (0.0) | NC | NC | NC |
| M. catarrhalis | 78 (52.3) | 74 (94.9) | 4 (5.1) | .65 (.11, 3.77) | .41 (.03, 5.01) | 1.15 (.17, 7.65) |
| S. aureus | 21 (14.1) | 20 (95.2) | 1 (4.8) | .60 (.05, 7.11) | .68 (.02, 20.0) | .90 (.05, 14.75) |
| Beta-lactamase-producing bacteriall | 98 (65.8) | 93 (94.9) | 5 (5.1) | .65 (.12, 3.53) | .55 (.08, 4.07) | .98 (.16, 5.90) |
| Non-beta-lactamase-producing bacteria | 51 (34.2) | 48 (94.1) | 3 (5.9) | .50 (.04, 5.89) | .50 (.03, 9.7) | .41 (.03, 5.65) |
| Bacterial pathogens (n(%)) | ||||||
| No bacteria identified | 25 (16.8) | 24 (96.0) | 1 (4.0) | Referent | Referent | Referent |
| Only S. pneumoniae | 2 (1.3) | 2 (100.0) | 0 (0.0) | NC | NC | NC |
| Only H. influenzae | 9 (6.0) | 9 (100.0) | 0 (0.0) | NC | NC | NC |
| No beta-lactamase productionm | 5 (55.6) | 5 (100.0) | 0 (0.0) | NC | NC | NC |
| Beta-lactamase productionm | 4 (44.4) | 4 (100.0) | 0 (0.0) | NC | NC | NC |
| Only M. catarrahlis | 17 (1.4) | 15 (88.2) | 2 (11.8) | 1.6 (.20, 12.60) | NC | 21.13 (.18, 2397.91) |
| Only S. aureus | 5 (3.4) | 4 (80.0) | 1 (20.0) | 3.0 (.22, 41.35) | 3.80 (.06, 240.03) | 5.84 (.16, 219.18) |
aCulture for bacteria, RVP for viruses (PCR); only bacterial pathogens recovered in culture reported in this table—see supplement for data from culture plus PCR.
bp-value less than 0.100 indicated by an asterisk (*).
cCovariates for logistic regression adjustment: age category, meets AAP criteria, co-infection with virus, and language (smoking status, sex, race, ethnicity, and pre- or peri-covid period were excluded due to collinearity). Additional covariates were included after confirming they were not colinear with the different organism categories: AOM-SOS scale at 5 and 14 days, laterality of infection, number of days antibiotic taken.
dCovariates for logistic regression adjustment: age category, meets AAP criteria, co-infection with virus, laterality of infection, and language (smoking status, sex, race, ethnicity, and pre- or peri-covid period were excluded due to collinearity).
eFor group “Organism types” referent used for odds ratio estimation is “No organisms.”
fIncludes patients positive for COVID-19.
gNC, Not calculable.
hThese categories include culture positive for specified bacteria, and viral PCR could be positive or negative. Reference is no bacteria.
iFive patients had positive cultures for other bacteria, which included diptheroids (1), few mixed oral flora (1), strep group A beta hemolytic (1), many mixed bacterial flora (1), type missing (1).
jOne patient had discrepant entries with data indicating organisms were present (cultureresult_1num = 0), but no individual organisms were identified in culture. Individual culture results were used to tally patients with no bacteria identified.
kIn patients with any H. influenzae, denominator of 30.
lIncludes all M. catarrhalis positive cultures, all S. aureus positive cultures, and all beta-lactamase-producing H. influenzae cultures.
mIn patients with only H. influenzae, denominator of 9.
Treatment Failure by Microorganism in Patients who did not Receive Antibiotics Prior to Specimen Collectiona
| Organism | Organism Prevalence (N = 149) | Treatment Success (N = 148) | Treatment Failure (N = 8) | OR (95% CI)b | aOR (95% CI)b,c | aOR (95% CI)b,d |
|---|---|---|---|---|---|---|
| Microorganism types (n (%, row))a,e | ||||||
| No microorganism identified | 14 (9.4) | 13 (92.9) | 1 (7.1) | Referent | Referent | Referent |
| Any bacteria | 123 (82.6) | 117 (95.1) | 7 (5.6) | .67 (.07, 5.98) | .66 (.06, 7.43) | 1.05 (.10, 10.84) |
| Any virusf | 71 (47.6) | 68 (95.8) | 3 (4.2) | .58 (.06, 6.01) | .003 (3e-7, 28.71) | .52 (.04, 6.75) |
| Only bacteria | 64 (42.9) | 60 (92.3) | 4 (6.2) | .87 (.09, 8.40) | .68 (.05, 8.56) | 1.00 (.0.10, 10.69) |
| Only virus(es) | 12 (8.0) | 11 (91.7) | 1 (8.3) | 1.18 (.07, 21.17) | NCg | NC |
| Either bacteria or virus(es) | 135 (90.6) | 128 (94.8) | 7 (5.2) | .71 (.08, 6.24) | .57 (.05, 6.48) | 1.02 (.10, 10.52) |
| Both bacteria and virus(es) | 59 (39.6) | 57 (96.6) | 2 (3.4) | .46 (.04, 5.42) | .004 (3e-7, 42.46) | .51 (.03, 7.58) |
| Bacterial pathogens (n(%))h,i | ||||||
| No bacteria identifiedj | 26 (17.4) | 24 (92.3) | 2 (7.7) | Referent | Referent | Referent |
| S. pneumoniae | 51 (34.2) | 49 (96.1) | 2 (3.9) | .49 (.06, 3.69) | .43 (.03, 5.27) | .58 (.07, 4.83) |
| H. influenzae | 30 (20.1) | 29 (96.7) | 1 (3.3) | .41 (.04, 4.85) | .47 (.03, 8.48) | .63 (.04, 9.25) |
| No beta-lactamase productionk | 19 (63.3) | 18 (94.7) | 1 (5.3) | .67 (.056, 7.94) | .89 (.04, 18.48) | .87 (.06, 12.84) |
| Beta-lactamase productionk | 11 (36.7) | 11 (100.0) | 0 (0.0) | NC | NC | NC |
| M. catarrhalis | 78 (52.3) | 74 (94.9) | 4 (5.1) | .65 (.11, 3.77) | .41 (.03, 5.01) | 1.15 (.17, 7.65) |
| S. aureus | 21 (14.1) | 20 (95.2) | 1 (4.8) | .60 (.05, 7.11) | .68 (.02, 20.0) | .90 (.05, 14.75) |
| Beta-lactamase-producing bacteriall | 98 (65.8) | 93 (94.9) | 5 (5.1) | .65 (.12, 3.53) | .55 (.08, 4.07) | .98 (.16, 5.90) |
| Non-beta-lactamase-producing bacteria | 51 (34.2) | 48 (94.1) | 3 (5.9) | .50 (.04, 5.89) | .50 (.03, 9.7) | .41 (.03, 5.65) |
| Bacterial pathogens (n(%)) | ||||||
| No bacteria identified | 25 (16.8) | 24 (96.0) | 1 (4.0) | Referent | Referent | Referent |
| Only S. pneumoniae | 2 (1.3) | 2 (100.0) | 0 (0.0) | NC | NC | NC |
| Only H. influenzae | 9 (6.0) | 9 (100.0) | 0 (0.0) | NC | NC | NC |
| No beta-lactamase productionm | 5 (55.6) | 5 (100.0) | 0 (0.0) | NC | NC | NC |
| Beta-lactamase productionm | 4 (44.4) | 4 (100.0) | 0 (0.0) | NC | NC | NC |
| Only M. catarrahlis | 17 (1.4) | 15 (88.2) | 2 (11.8) | 1.6 (.20, 12.60) | NC | 21.13 (.18, 2397.91) |
| Only S. aureus | 5 (3.4) | 4 (80.0) | 1 (20.0) | 3.0 (.22, 41.35) | 3.80 (.06, 240.03) | 5.84 (.16, 219.18) |
| Organism | Organism Prevalence (N = 149) | Treatment Success (N = 148) | Treatment Failure (N = 8) | OR (95% CI)b | aOR (95% CI)b,c | aOR (95% CI)b,d |
|---|---|---|---|---|---|---|
| Microorganism types (n (%, row))a,e | ||||||
| No microorganism identified | 14 (9.4) | 13 (92.9) | 1 (7.1) | Referent | Referent | Referent |
| Any bacteria | 123 (82.6) | 117 (95.1) | 7 (5.6) | .67 (.07, 5.98) | .66 (.06, 7.43) | 1.05 (.10, 10.84) |
| Any virusf | 71 (47.6) | 68 (95.8) | 3 (4.2) | .58 (.06, 6.01) | .003 (3e-7, 28.71) | .52 (.04, 6.75) |
| Only bacteria | 64 (42.9) | 60 (92.3) | 4 (6.2) | .87 (.09, 8.40) | .68 (.05, 8.56) | 1.00 (.0.10, 10.69) |
| Only virus(es) | 12 (8.0) | 11 (91.7) | 1 (8.3) | 1.18 (.07, 21.17) | NCg | NC |
| Either bacteria or virus(es) | 135 (90.6) | 128 (94.8) | 7 (5.2) | .71 (.08, 6.24) | .57 (.05, 6.48) | 1.02 (.10, 10.52) |
| Both bacteria and virus(es) | 59 (39.6) | 57 (96.6) | 2 (3.4) | .46 (.04, 5.42) | .004 (3e-7, 42.46) | .51 (.03, 7.58) |
| Bacterial pathogens (n(%))h,i | ||||||
| No bacteria identifiedj | 26 (17.4) | 24 (92.3) | 2 (7.7) | Referent | Referent | Referent |
| S. pneumoniae | 51 (34.2) | 49 (96.1) | 2 (3.9) | .49 (.06, 3.69) | .43 (.03, 5.27) | .58 (.07, 4.83) |
| H. influenzae | 30 (20.1) | 29 (96.7) | 1 (3.3) | .41 (.04, 4.85) | .47 (.03, 8.48) | .63 (.04, 9.25) |
| No beta-lactamase productionk | 19 (63.3) | 18 (94.7) | 1 (5.3) | .67 (.056, 7.94) | .89 (.04, 18.48) | .87 (.06, 12.84) |
| Beta-lactamase productionk | 11 (36.7) | 11 (100.0) | 0 (0.0) | NC | NC | NC |
| M. catarrhalis | 78 (52.3) | 74 (94.9) | 4 (5.1) | .65 (.11, 3.77) | .41 (.03, 5.01) | 1.15 (.17, 7.65) |
| S. aureus | 21 (14.1) | 20 (95.2) | 1 (4.8) | .60 (.05, 7.11) | .68 (.02, 20.0) | .90 (.05, 14.75) |
| Beta-lactamase-producing bacteriall | 98 (65.8) | 93 (94.9) | 5 (5.1) | .65 (.12, 3.53) | .55 (.08, 4.07) | .98 (.16, 5.90) |
| Non-beta-lactamase-producing bacteria | 51 (34.2) | 48 (94.1) | 3 (5.9) | .50 (.04, 5.89) | .50 (.03, 9.7) | .41 (.03, 5.65) |
| Bacterial pathogens (n(%)) | ||||||
| No bacteria identified | 25 (16.8) | 24 (96.0) | 1 (4.0) | Referent | Referent | Referent |
| Only S. pneumoniae | 2 (1.3) | 2 (100.0) | 0 (0.0) | NC | NC | NC |
| Only H. influenzae | 9 (6.0) | 9 (100.0) | 0 (0.0) | NC | NC | NC |
| No beta-lactamase productionm | 5 (55.6) | 5 (100.0) | 0 (0.0) | NC | NC | NC |
| Beta-lactamase productionm | 4 (44.4) | 4 (100.0) | 0 (0.0) | NC | NC | NC |
| Only M. catarrahlis | 17 (1.4) | 15 (88.2) | 2 (11.8) | 1.6 (.20, 12.60) | NC | 21.13 (.18, 2397.91) |
| Only S. aureus | 5 (3.4) | 4 (80.0) | 1 (20.0) | 3.0 (.22, 41.35) | 3.80 (.06, 240.03) | 5.84 (.16, 219.18) |
aCulture for bacteria, RVP for viruses (PCR); only bacterial pathogens recovered in culture reported in this table—see supplement for data from culture plus PCR.
bp-value less than 0.100 indicated by an asterisk (*).
cCovariates for logistic regression adjustment: age category, meets AAP criteria, co-infection with virus, and language (smoking status, sex, race, ethnicity, and pre- or peri-covid period were excluded due to collinearity). Additional covariates were included after confirming they were not colinear with the different organism categories: AOM-SOS scale at 5 and 14 days, laterality of infection, number of days antibiotic taken.
dCovariates for logistic regression adjustment: age category, meets AAP criteria, co-infection with virus, laterality of infection, and language (smoking status, sex, race, ethnicity, and pre- or peri-covid period were excluded due to collinearity).
eFor group “Organism types” referent used for odds ratio estimation is “No organisms.”
fIncludes patients positive for COVID-19.
gNC, Not calculable.
hThese categories include culture positive for specified bacteria, and viral PCR could be positive or negative. Reference is no bacteria.
iFive patients had positive cultures for other bacteria, which included diptheroids (1), few mixed oral flora (1), strep group A beta hemolytic (1), many mixed bacterial flora (1), type missing (1).
jOne patient had discrepant entries with data indicating organisms were present (cultureresult_1num = 0), but no individual organisms were identified in culture. Individual culture results were used to tally patients with no bacteria identified.
kIn patients with any H. influenzae, denominator of 30.
lIncludes all M. catarrhalis positive cultures, all S. aureus positive cultures, and all beta-lactamase-producing H. influenzae cultures.
mIn patients with only H. influenzae, denominator of 9.
At least one M. catarrhalis, S. pneumoniae, or H. influenzae was identified in 76% of specimens. Among H. influenzae isolates, 36.7% produced beta-lactamase. Overall, a beta-lactamase-producing bacteria was identified in 65.8% of specimens. In S. pneumoniae isolates, 16% displayed intermediate susceptibility to penicillin and 9% were resistant. Additional details of microbiological findings are provided in the Supplementary material.
In both univariable and multivariable logistic regression analyses, there was no significant association between treatment failure and individual or groups of bacterial pathogens (Table 3). Similarly, there was no association between treatment failure and beta-lactamase-producing bacteria (multivariable regression odds ratio: 0.98; 95% confidence interval: 0.16, 5.90; p-value: 0.980) versus non-beta-lactamase-producing bacteria. No differences in treatment failure rates were identified by individual pathogens when identification was evaluated by bacterial culture alone (Table 3) or by bacterial culture or PCR (Supplementary material).
DISCUSSION
Treatment failure was low among a cohort of children diagnosed with AOM in clinical practice and treated with amoxicillin. Treatment failure was not associated with individual bacterial pathogens nor beta-lactamase-producing organisms as a group. Parent-reported ADE occurred in over 25% of children.
Similar to other studies of children diagnosed with AOM in clinical practice [12, 19], we observed low treatment failure and recurrent infections (5.4% and 6.8%, respectively). Most children’s symptoms resolved within 7 days. Importantly, children diagnosed with AOM were included even if they did not meet AAP diagnostic criteria because we aimed to assess treatment outcomes in routine clinical practice. We found that one-third of children diagnosed with AOM by clinicians did not meet AAP criteria, primarily due to the absence of TM bulging. These children would typically be excluded from clinical trials [16–18], even though they comprise 30-50% of children in the US that are prescribed an antibiotic for AOM in clinical practice. This may partly explain the lower rate of treatment failure than in prior randomized trials [16–18]. The higher rate of treatment failure in these trials may reflect the exclusion of children with lower severity symptoms, differences in outcome measures, or reporting bias from parents enrolling their children in a clinical trial where they might receive a placebo. We did not find significant differences in the frequency of treatment failure, time to symptom improvement, or resolution between children who did and did not meet AAP diagnostic criteria for AOM. However, the recurrence rate was somewhat higher for those meeting AAP diagnostic criteria, though it is unclear if this is clinically meaningful given recurrence rates <10% in both groups.
We identified a high frequency of bacterial otopathogens with likely non-susceptibility to amoxicillin due to beta-lactamase production. Notably, M. catarrhalis accounted for more than half of the otopathogens detected. Since M. catarrhalis may be associated with less severe AOM or asymptomatic carriage [5], this finding may have important treatment implications and raises the question of whether antibiotic therapy improves outcomes in such cases. Consistent with prior data demonstrating an increasing proportion of beta-lactamase-producing H. influenzae [1, 3, 4], beta-lactamase was detected in over one-third of isolates. Importantly, we observed similar outcomes with amoxicillin treatment among children regardless of the presence of a beta-lactamase-producing otopathogen. This could be explained by inaccurate diagnosis or the self-limited nature of most AOM. Either way, symptoms would be expected to resolve without antibiotic therapy. Another potential explanation is that, despite the presence of beta-lactamase, its clinical relevance was limited due to low levels of expression. Regardless of the underlying explanation, these data suggest that given the low treatment failure rate of amoxicillin in clinical practice, there may not be a substantial benefit to using broader-spectrum antibiotics that are not inactivated by beta-lactamases, such as amoxicillin-clavulanate or cephalosporins. This supports AAP guidance for amoxicillin as first-line therapy in most cases of AOM when an antibiotic is indicated.
Several notable patient-level findings were observed. First, parents reported difficulty administering the antibiotic in 14.6% of cases and the prescribed antibiotic course was not adhered to in over half of cases. We speculate that some parents may have stopped the antibiotic because their child felt better and others due to ADE (27%). Second, similar to findings reported by Gerber and colleagues [12], parents reported that 27.3% of children experienced an ADE during the amoxicillin course. Though most ADEs were mild, a follow-up medical visit occurred in 8.8% of cases. In effect, antibiotic-related ADEs were about as likely to require follow-up care as treatment failure or recurrence. Routine use of broader-spectrum antibiotics like amoxicillin-clavulanate or cephalosporins may increase the potential for side effects. These findings highlight the importance of several aspects of AOM care: the need for accurate AOM diagnosis to avoid unnecessary antibiotic prescriptions, prioritization of observation or delayed prescriptions to reduce unnecessary antibiotic exposure, and when antibiotics are indicated, shared decision-making with discussion of the potential risks and benefits of antibiotic therapy.
Strengths of this study include the assessment of clinical outcomes among a cohort of children diagnosed with and treated for AOM in varied settings, the use of both objective clinical and parent-reported outcomes, and the coupling of microbiological and outcomes data. Additionally, enrolled children sought care almost exclusively at DHHA, which allowed for monitoring of failure, recurrence, and complications.
This study had limitations. First, it was performed within a single integrated healthcare system, so the results—in particular microbiological findings—may not be generalizable to other geographical areas. Second, our a priori sample size calculations were based on higher estimates of treatment failure (15% overall), thus we may have lacked power to detect a significant difference in the treatment failure rate between children with and without beta-lactamase-producing organisms. Furthermore, as children are often misdiagnosed with AOM in clinical practice and identification of a microorganism may represent colonization rather than infection, this also limits inferences that can be drawn between the association between treatment failure and beta-lactamase-producing organisms. However, given the overall low rate of treatment failure, it seems unlikely a significant difference, even if one was observed, would be clinically relevant. Third, we collected microbiologic data from NP specimens rather than from middle ear fluid. We chose NP sampling because tympanocentesis can have a treatment effect for AOM and middle ear fluid organisms can reliably be detected in NP specimens. We also aimed to use a replicable sampling technique that could be conducted by providers if found to be clinically useful. Though the positive predictive value of NP sampling for middle ear bacterial otopathogens varies, the negative predictive value for common bacterial otopathogens is high [24, 25]. Thus, our findings may overestimate the proportion of children where a bacterial pathogen was actually present in the middle ear. Finally, because all children were prescribed amoxicillin, we could not compare clinical outcomes between amoxicillin and other treatment approaches. Randomized trials are needed to evaluate these important questions.
CONCLUSION
The rate of treatment failure for children diagnosed with AOM in clinical practice and treated with amoxicillin was low and did not differ significantly based on potential bacterial otopathogens identified. When an antibiotic is indicated, these data support current AAP guidance recommending amoxicillin over broader-spectrum antibiotics despite the increasing prevalence of beta-lactamase-producing bacteria. In light of the high incidence of parent-reported ADE, clinicians should weigh the risks and benefits of prescribing antibiotics for AOM.
Author Contribution Dr Holly M. Frost acquired funding, conceptualized, and designed the study, developed study methodology, oversaw research planning and execution, drafted the initial manuscript, and critically reviewed and revised the manuscript. Ms Amy Keith assisted with data curation, investigation, drafted the initial manuscript, and critically revised the manuscript. Dr Dana R. Fletcher performed formal analysis, drafted the initial manuscript, and critically reviewed and revised the manuscript. Ms Melanie Kurtz assisted with data curation, investigation, and critically reviewed and revised the manuscript. Drs Sebastian, Dominguez, Parker, Wilson, and Jenkins provided subject matter expertise and critically reviewed and revised the manuscript. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
Data availability statement. Study data is not available as it includes patient identifiers.
Financial support. This work was supported by a Grant from The Gerber Foundation. H. F. received salary support from the Eunice Kennedy Shriver National Institute of Child Health & Human Development of the National Institutes of Health (K23HD099925). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The funder had no role in the design or interpretation of the study.
Potential conflicts of interest. H. F. serves as a senior scientific advisor for QuidelOrtho and holds a patent for diagnosing and treating otitis media #63/335,801. QuidoOrtho had no role in the design or interpretation of the study and Quidel staff were blinded to culture results. S. D. serves as a consultant and receives grant support from Biofire, serves as a consultant for Karius, and receives grant support from Pfizer for projects not pertaining to this project.
