Abstract
Acute otitis media (AOM) is a common complication of upper respiratory tract infection whose pathogenesis involves both viruses and bacteria. We examined risks of acute otitis media associated with specific combinations of respiratory viruses and acute otitis media bacterial pathogens. Data were from a prospective study of children ages 6 to 36 months and included viral and bacterial culture and quantitative PCR for respiratory syncytial virus (RSV), human bocavirus, and human metapneumovirus. Repeated-measure logistic regression was used to assess the relationship between specific viruses, bacteria, and the risk of acute otitis media complicating upper respiratory tract infection. In unadjusted analyses of data from 194 children, adenovirus, bocavirus, Streptococcus pneumoniae, Haemophilus influenzae, and Moraxella catarrhalis were significantly associated with AOM (P < 0.05 by χ2 test). Children with high respiratory syncytial virus loads (≥3.16 × 107 copies/ml) experienced increased acute otitis media risk. Higher viral loads of bocavirus and metapneumovirus were not significantly associated with acute otitis media. In adjusted models controlling for the presence of key viruses, bacteria, and acute otitis media risk factors, acute otitis media risk was independently associated with high RSV viral load with Streptococcus pneumoniae (odds ratio [OR], 4.40; 95% confidence interval [CI], 1.90 and 10.19) and Haemophilus influenzae (OR, 2.04; 95% CI, 1.38 and 3.02). The risk was higher for the presence of bocavirus and H. influenzae together (OR, 3.61; 95% CI, 1.90 and 6.86). Acute otitis media risk differs by the specific viruses and bacteria involved. Acute otitis media prevention efforts should consider methods for reducing infections caused by respiratory syncytial virus, bocavirus, and adenovirus in addition to acute otitis media bacterial pathogens.
INTRODUCTION
Acute otitis media (AOM) is one of the most common diseases seen in pediatric practice and is a major reason for antibiotic use in the United States (15, 33, 37). AOM pathogenesis involves complex interactions between bacteria, viruses, and the host inflammatory response. Streptococcus pneumoniae, Haemophilus influenzae, and Moraxella catarrhalis colonize the nasopharynx and are the three most frequent AOM pathogens (9). While AOM is often considered a bacterial infection, its pathogenesis involves complex interactions among viruses and bacteria (4). The majority of AOM episodes occur as complications of viral upper respiratory tract infection (URI) (10, 17). Viral URI increases the risk of bacterial AOM by promoting the replication of bacteria and increasing inflammation in the nasopharynx and Eustachian tube, which subsequently facilitates bacterial entry into the middle ear space (4). Respiratory viruses can coinfect the middle ear with bacterial AOM pathogens and have been identified as the sole causative agent of AOM (19, 29).
Viral URI is a major risk factor for AOM; however, specific viruses differ in their propensity to cause disease (10, 21). Respiratory syncytial virus (RSV) has been identified as being strongly associated with AOM (10, 21, 27). The risk of AOM differs depending on the number and type of bacterial AOM pathogens that colonize the nasopharynx (28). Epidemiologic studies have not consistently identified interactions between specific viruses and bacteria and increased risk of AOM (2, 21, 22). Inconsistent results may be due to differences in study design and methods used to detect viruses and bacteria. For example, the risk of AOM complicating URI due to specific viruses was most strongly associated with viruses detected by culture as opposed to molecular methods, perhaps because culture captures infections associated with higher viral loads (10).
The understanding of relationships between viruses, bacteria, and AOM risk is changing as new viruses are discovered and new methods are developed to quantify pathogen loads. Human bocavirus (hBOV) is a parvovirus that was first described in 2005 (1). Several studies have detected hBOV in nasal samples from children with URI and in middle ear fluid samples from children with AOM (5, 14, 29, 30). However, the role of hBOV as a causative URI pathogen has been debated (3, 23, 25). Human metapneumovirus (hMPV), described in 2001 (32), also has been isolated from middle ear fluid samples from children with AOM (29, 31). Retrospective analyses indicated that more than half of URI episodes due to hMPV were complicated by AOM (35). However, the risk of AOM due to hMPV did not differ significantly from that of other respiratory viruses, such as RSV (35). Additional studies are needed to accurately determine the risk of AOM complicating URI due to these relatively new viruses.
The prevalence of AOM has declined slightly during the past decade due in part to the use of pneumococcal conjugate vaccines (33, 37). AOM is a polymicrobial disease; further advances in AOM prevention will depend upon the identification of critical polymicrobial interactions that influence AOM risk to prioritize bacterial and viral targets for intervention strategies. We tested the hypothesis that specific combinations of respiratory viruses and bacteria were associated with an increased risk for AOM. Data are from a prospective cohort study of healthy children monitored for 1 year for AOM complicating URI (10). Culture-based data were used to assess infection by respiratory viruses and bacterial colonization. In addition, newly developed quantitative PCR (qPCR) assays were used to assess the relationship between viral loads of RSV, hBOV, and hMPV and the risk of AOM.
MATERIALS AND METHODS
Study design and subjects.
Data are from a prospective cohort study of AOM complicating symptomatic URI (10). A detailed study description has been provided elsewhere (10, 28). Subjects were enrolled and specimens were collected at the University of Texas Medical Branch (UTMB) at Galveston. The study was approved by the UTMB institutional review board. Parents provided written informed consent for all subjects. Healthy children living in Galveston, between 6 months to 3 years of age, were enrolled between January 2003 and March 2007 and monitored for 1 year. Children with chronic medical problems or anatomical or physiological defects of the ear or nasopharynx were excluded.
At enrollment, demographic and AOM risk factor data were collected (e.g., antibiotic use, duration of breastfeeding, daycare attendance, and exposure to environmental tobacco smoke). Children were brought to the study office for a sick visit at the onset of URI symptoms (nasal congestion, rhinorrhea, cough, sore throat, or fever). At each study visit, data were recorded on URI symptoms and otoscopic findings. Within a few days of the sick visit, children had a repeat study visit to evaluate their middle ear status. Criteria for AOM diagnosis included the acute onset of symptoms (fever, irritability, or earache), signs of tympanic membrane inflammation, and the presence of fluid documented by pneumatic otoscopy and/or tympanometry. AOM episodes occurring within a month of URI onset were considered a complication of URI. In addition to parental reports of URI, study personnel called parents twice monthly for current URI symptoms and missed URI or AOM episodes. Medical records were reviewed when children completed the study to identify missed URI and AOM episodes.
Two hundred ninety-four children were enrolled in the original study (10). Included in these analyses are data from 194 children who were seen by the study group at the onset of URI, were monitored for AOM, and had the following microbiologic data available: bacterial culture, viral culture, and quantitative PCR (qPCR) (n = 640 episodes). Analyses were restricted to specimens collected during the first study visit of each URI episode.
Viral and bacterial specimens.
Nasopharyngeal secretion specimens were collected for viral studies, and nasopharyngeal swabs were collected for bacterial culture as previously described (10, 28). Viral and bacterial cultures were performed by standard methods (10, 28). qPCR was performed on thawed frozen specimens as described below.
Nucleic acids were extracted from nasopharyngeal specimens using a Tecan Evo (Männedorf, Switzerland) automated liquid handler within the Assay Development Service Division of the Galveston National Laboratory, followed by processing using MagMAX Total Nucleic Acid isolation kits (Ambion/Applied Biosystems, Austin, TX). For the quantification of hMPV and RSV, cDNA was synthesized using the iScript synthesis kit (Bio-Rad, Hercules, CA). A duplex qPCR, with primers amplifying RSV and hMPV targets, was completed on cDNA (6). hBOV DNA was quantified using the methods of Lu et al. (24). Separate qPCRs for human glyceraldehyde 3-phosphate dehydrogenase assessed sample quality (7). For all targets, starting quantity values were extrapolated from standard curves established by parallel amplification of known amounts of cloned PCR targets. The viral load was calculated as copies/ml of the original specimen considering all sample dilutions.
Statistical methods.
Statistical analyses were conducted using SAS version 9.2 (Cary, NC). Unadjusted associations between AOM and other variables were examined with χ2 analyses. Since each of the 194 subjects could contribute data from multiple separate URI episodes, we examined patterns of virus infection and bacterial colonization using repeated-measure logistic regression with generalized estimating equations (GEE) and specified an autoregressive correlation structure (AR1). Results of qPCR for RSV, hBOV, and hMPV were categorized into a four-level variable (no, low, medium, or high copies/ml) prior to analysis to determine dose-response relationships to AOM. Boundaries of viral load categories were determined by dividing the distribution of copies/ml (excluding 0 copies/ml) into thirds. Separate logistic regression analyses of each of the viruses identified by culture were used to determine which were associated with an increased risk of AOM and should be included in final models. Separate analyses also were performed to examine potential interactions between each of the viruses identified by qPCR (RSV, hBOV, and hMPV) and AOM bacterial pathogens (S. pneumoniae, H. influenzae, and M. catarrhalis). The final model included variables representing coinfections between RSV and S. pneumoniae (as a four-level categorical variable representing the presence or absence of a high viral load of RSV and S. pneumoniae) and hBOV and H. influenzae (as a four-level categorical variable representing the presence or absence of each) in addition to the presence or absence of hMPV, adenovirus, and M. catarrhalis. All models were adjusted for the following AOM risk-related covariates: age of the child at the time of nasopharyngeal specimen collection, ethnicity, daycare attendance, breastfed for ≥3 months, exposure to environmental tobacco smoke, and antibiotic use within the previous 7 days.
RESULTS
The subset of 194 children available for this study did not differ significantly from the larger original enrollment (n = 294) by age, gender, or race (P = 0.10, 0.99, and 0.52, respectively; χ2 test). Participant characteristics are in Table 1. Mean age at enrollment was 14.2 months (standard deviations [SD], 7.6). The children experienced a mean of 3.3 (SD, 2.7) URI episodes (range, 1 to 15) and 1.2 diagnoses of AOM (range, 0 to 7). AOM episodes occurred within 1 and 17 days of URI onset; the mean time between the onset of URI and the development of AOM symptoms was 5.6 days (SD, 3.2). The distribution of URI episodes by child is in Table 2. Approximately 25% of the children contributed 55% of the specimens.
Table 1.
Characteristics of subjects (n = 194) and acute otitis media diagnosis
| Characteristic | Total no. (%) | No. (%) diagnosed with AOM (anya) | P valueb |
|---|---|---|---|
| Age at enrollment, mo | 0.30 | ||
| 6-<12 | 84 (43) | 59 (70) | |
| 12-<18 | 55 (28) | 31 (56) | |
| 18-<24 | 28 (14) | 17 (61) | |
| 24-<36 | 27 (14) | 15 (56) | |
| Gender | 0.60 | ||
| Female | 95 (49) | 58 (61) | |
| Male | 99 (51) | 64 (65) | |
| Race | 0.17 | ||
| White | 39 (20) | 25 (64) | |
| Black | 56 (29) | 32 (57) | |
| Hispanic | 83 (43) | 51 (61) | |
| Other | 16 (8) | 14 (88) | |
| Day care | 0.66 | ||
| No | 133 (69) | 85 (64) | |
| Yes | 61 (31) | 37 (61) | |
| Breastfed for ≥3 months | 0.33 | ||
| No | 127 (66) | 83 (65) | |
| Yes | 67 (34) | 39 (58) | |
| Environmental tobacco smoke exposure | 0.37 | ||
| No | 141 (73) | 86 (61) | |
| Yes | 53 (27) | 36 (68) | |
| Antibiotic use within 7 days of URI | 0.003 | ||
| None | 171 (88) | 101 (59) | |
| Any | 23 (12) | 21 (91) |
One or more acute otitis media diagnoses per subject.
P values from χ2 tests. Significant P values (<0.05) are shown in boldface.
Table 2.
Distribution of subjects by number of URI episodesa
| No. of URI episodes | No. (%) of subjects |
|---|---|
| 1 | 62 (32) |
| 2 | 41 (21) |
| 3-4 | 42 (21) |
| ≥5 | 49 (25) |
A total of 194 subjects were examined.
The proportion of children with any AOM diagnoses was highest among children 6 to <12 months of age and those who had antibiotics within the 7 days prior to the URI study visit (Table 1). The proportion of children with at least one AOM episode did not significantly differ by gender, race, daycare, breastfeeding history, or environmental tobacco smoke exposure (Table 1). More than 99% (n = 635) of the 640 specimens were collected from study participants who received at least one dose of the 7-valent pneumococcal conjugate vaccine (PCV7) at the time of specimen collection.
The following viruses were detected by culture in 640 specimens: rhinovirus, n = 33 (5.2%); parainfluenza types 1 to 3, n = 20 (3.1%); adenovirus, n = 21 (3.3%); influenza A and B viruses, n = 21 (3.3%); enterovirus, n = 14 (2.2%); and RSV, n = 12 (1.9%). Respiratory viruses were not isolated by culture for 499 (78%) samples. Separate repeated-measure logistic regression models were used to evaluate the risk of AOM complicating URI with individual viruses detected by culture. Adjusted models included each AOM bacterial pathogen and AOM risk-related covariates. The presence of rhinovirus, parainfluenza virus, influenza virus, enterovirus, and RSV by culture were not significantly associated with an increased risk of AOM (data not shown). These viruses were not considered in further models. In contrast, adenovirus was significantly associated with an increased risk of AOM (odds ratio [OR], 2.90; 95% CI, 1.30 and 6.46). The distribution of adenovirus is in Table 3.
Table 3.
Unadjusted association of AOM diagnosis with presence of viruses and bacteria on 640 nasal specimens taken from 194 subjects
| Virus/bacteria | No. (%) of specimens | No. (%) of specimens with AOM diagnosis | P valuea |
|---|---|---|---|
| Total | 640 | 239 (37) | |
| Adenovirus by culture | 0.02 | ||
| Absent | 619 (97) | 226 (36) | |
| Present | 21 (3) | 13 (62) | |
| Virus by PCR | |||
| RSV | 0.28 | ||
| Absent | 548 (86) | 200 (36) | |
| Present | 92 (14) | 39 (42) | |
| hBOV | 0.05 | ||
| Absent | 483 (75) | 170 (35) | |
| Present | 157 (25) | 69 (44) | |
| hMPV | 0.89 | ||
| Absent | 596 (93) | 223 (37) | |
| Present | 44 (7) | 16 (36) | |
| Bacteria by culture | |||
| S. pneumoniae | 0.0003 | ||
| Absent | 353 (55) | 110 (31) | |
| Present | 287 (45) | 129 (45) | |
| H. influenzae | <0.0001 | ||
| Absent | 446 (70) | 138 (31) | |
| Present | 194 (30) | 101 (52) | |
| M. catarrhalis | <0.0001 | ||
| Absent | 226 (35) | 60 (27) | |
| Present | 414 (65) | 179 (43) |
P values are from χ2 tests. Significant P values (<0.05) are shown in boldface.
qPCR analyses detected the presence of RSV in 14%, hBOV in 24%, and hMPV in 7% of specimens (Table 3). Of the respiratory viruses identified by culture or qPCR, 49% of the specimens were negative for all viruses, 40% were positive for one virus, 10.5% were positive for two viruses, and 0.5% were positive for three respiratory viruses (data not shown). Of 640 specimens, 45, 30, and 65% were positive for S. pneumoniae, H. influenzae, and M. catarrhalis, respectively (Table 3). The presence of hBOV by PCR and adenovirus, S. pneumoniae, H. influenzae, and M. catarrhalis by culture were associated with an AOM diagnosis (P < 0.05; χ2 test) (Table 3).
Among samples positive by qPCR, the viral load for each ranged from 3.99 × 103 to 1.20 × 1011 copies/ml for RSV, 3.23 × 103 to 1.73 ×1012 copies/ml for hBOV, and 2.73 × 104 to 1.03 ×1010 copies/ml for hMPV. There was a significant association between the season of sampling and viral load of RSV (P = 0.0005; χ2 test) but not for hBOV (P = 0.32) or hMPV (P = 0.13). During June, July, and August, 112 samples were collected from children with URI. Of these, 11 (9.8%) specimens were positive for RSV, but none were in the highest viral load category.
We next evaluated the risk of AOM associated with categories of viral load for RSV, hBOV, and hMPV. The repeated-measure logistic regression model included these three viruses as well as variables representing the presence or absence of adenovirus and each AOM bacterial pathogen (Table 4). The model was adjusted for the AOM risk-related covariates. The boundaries for the viral load categories for each virus are in the footnote of Table 4. For RSV, there was a threshold effect: low and intermediate levels of RSV viral load were not associated with AOM any more frequently than the level of no RSV. However, children with an RSV viral load of greater than 3.16 × 107 copies/ml experienced a significantly increased risk of AOM (Table 4). Increasing viral loads of hBOV and hMPV were not significantly associated with an increased risk of AOM complicating URI. The presence of adenovirus, H. influenzae, and M. catarrhalis was independently associated with increased risk of AOM (Table 4). Of the AOM risk-related covariates, increasing age and having been breastfed for ≥3 months were associated with decreased risk of AOM. Each 1 month of age increase was associated with a 4% decreased risk of AOM (OR, 0.96; 95% CI, 0.94 and 0.99), and breastfeeding was associated with a 40% decreased risk of AOM (OR, 0.60; 95% CI, 0.36 and 0.99). Antimicrobial therapy within the previous 7 days was associated with an increased risk of AOM complicating URI (OR, 2.65; 95% CI, 1.14 and 6.15).
Table 4.
Association between viral load of RSV, hBOV, and hMPV and risk of AOM complicating URId
| Factor | No. (%) of samples | OR (95% CI) |
|---|---|---|
| RSVa | ||
| None (reference) | 548 (85.6) | 1.00 |
| Low | 31 (4.8) | 1.16 (0.65, 2.04) |
| Med | 30 (4.7) | 0.96 (0.42, 2.18) |
| High | 31 (4.8) | 2.60 (1.30, 5.20) |
| hBOVb | ||
| None (reference) | 483 (75.5) | 1.00 |
| Low | 51 (8.0) | 1.00 (0.56, 1.79) |
| Med | 54 (8.4) | 1.50 (0.92, 2.45) |
| High | 52 (8.1) | 0.97 (0.51, 1.88) |
| hMPVc | ||
| None (reference) | 596 (93.1) | 1.00 |
| Low | 15 (2.3) | 1.48 (0.53, 5.58) |
| Med | 15 (2.3) | 0.75 (0.22, 2.53) |
| High | 14 (2.2) | 0.43 (0.15, 1.29) |
| Adenovirus | ||
| Absent (reference) | 619 (96.7) | 1.00 |
| Present | 21 (3.3) | 4.64 (1.42, 7.07) |
| S. pneumoniae | ||
| Absent (reference) | 353 (55.2) | 1.00 |
| Present | 287 (44.8) | 1.25 (0.85, 1.84) |
| H. influenzae | ||
| Absent (reference) | 446 (69.7) | 1.00 |
| Present | 194 (30.3) | 2.41 (1.66, 3.48) |
| M. catarrhalis | ||
| Absent (reference) | 226 (35.3) | 1.00 |
| Present | 414 (64.7) | 1.91 (1.31, 2.78) |
RSV (copies/ml): none, 0; low, 0 to ≤1.26 × 105; med (medium), 1.26 × 105 to ≤3.16 × 107; high, >3.16 × 107.
hBOV (copies/ml): none, 0; low, 0 to ≤5.01 × 105; med, 5.01 × 105 to ≤2.51 × 107; high, >2.51 × 107.
hMPV (copies/ml): none, 0; low, 0 to ≤3.16 × 106; med, 3.16 × 106 to ≤6.31 × 108; high, >6.31 × 108.
OR and 95% CI are from repeated-measure logistic regression analysis of 640 specimens from 194 subjects. Significant ORs and 95% CIs are in boldface. The model included viral loads of RSV, hBOV, and hMPV as well as variables representing the presence or absence of adenovirus, each AOM bacterial pathogen by culture, and AOM risk-related covariates.
Repeated-measure logistic regression evaluation of statistical interactions between RSV, hBOV, and hMPV and the three AOM bacterial pathogens was significant for RSV with S. pneumoniae (P = 0.01) and hBOV with H. influenzae (P = 0.003). The model examined the independent risk of AOM associated with the highest level of RSV viral load and S. pneumoniae, hBOV, and H. influenzae and the presence of hMPV, adenovirus, and M. catarrhalis (Table 5). The model was adjusted for AOM risk-related covariates. S. pneumoniae had a synergistic relationship with RSV; children whose samples revealed a high viral load of RSV and the presence of S. pneumoniae experienced a 4-fold increased risk of AOM compared to that of samples with neither high-viral-load RSV nor S. pneumoniae. Compared to samples with neither H. influenzae nor hBOV, increased risk for AOM complicating URI was independently associated with H. influenzae (OR, 2.04; 95% CI, 1.38 and 3.02), and the risk was even higher for hBOV and H. influenzae together (OR, 3.61; 95% CI, 1.90 and 6.86).
Table 5.
Polymicrobial interactions and risk of AOM complicating URIb
| Factor(s) | No. (%) of samples | OR (95% CI) |
|---|---|---|
| High-viral-load RSV and S. pneumoniae | ||
| Neither present (reference)a | 345 (53.9) | 1.00 |
| S. pneumoniae alone | 264 (41.2) | 1.27 (0.86, 1.86) |
| High-viral-load RSV alone | 8 (1.2) | 2.01 (0.24, 4.77) |
| Both present | 23 (3.6) | 4.40 (1.90, 10.19) |
| hBOV and H. influenzae | ||
| Neither present (reference) | 346 (54.1) | 1.00 |
| H. influenzae alone | 137 (21.4) | 2.04 (1.38, 3.02) |
| hBOV alone | 100 (15.6) | 0.88 (0.54, 1.43) |
| Both present | 57 (8.9) | 3.61 (1.90, 6.86) |
| hMPV | ||
| Absent (reference) | 596 (93.1) | 1.00 |
| Present | 44 (6.9) | 0.81 (0.41, 1.59) |
| Adenovirus | ||
| Absent (reference) | 619 (96.7) | 1.00 |
| Present | 21 (3.3) | 3.06 (1.36, 6.89) |
| M. catarrhalis | ||
| Absent (reference) | 226 (35.3) | 1.00 |
| Present | 414 (64.7) | 1.85 (1.28, 2.68) |
The reference category includes samples where S. pneumoniae was not present and the viral load for RSV was not high (i.e., ≤3.16 × 107 copies/ml).
OR and 95% CI are from repeated-measure logistic regression analysis of 640 specimens from 194 subjects. Significant ORs and 95% CIs are in boldface. The model included significant interactions between RSV and S. pneumoniae and that between human bocavirus hBOV and H. influenzae. hMPV, adenovirus, M. catarrhalis, and AOM risk covariates also were included in the model.
Specimens were collected during the first study visit of each URI episode and linked to the development of AOM within a 1-month period. Some children may have acquired a new AOM pathogen between the time of sample collection and the development of AOM. Of the 239 AOM episodes, 80.3% (n = 192) occurred within 7 days of URI onset. Repeated-measure logistic regression was used to evaluate interactions between RSV, hBOV, and hMPV and the three AOM bacterial pathogens using data restricted to children who had specimens collected within 7 days of URI onset and diagnoses made within 7 days of URI onset. The associations between viruses, bacteria, and AOM (n = 614 specimens) were similar to data presented in Table 5. The significant associations were the following: high viral load of RSV and the presence of S. pneumoniae (OR, 6.19; 95% CI, 2.47 and 15.49), H. influenzae alone (OR, 1.86; 95% CI, 1.23 and 2.88), hBOV and H. influenzae together (OR, 3.65; 95% CI, 1.86 and 7.14)], adenovirus (OR, 3.27; 95% CI, 1.47 and 7.28)], and M. catarrhalis (OR, 1.66; 95% CI, 1.13 and 2.44).
DISCUSSION
Researchers have proposed that combined bacterium-virus vaccine formulations would have the most significant and long-term impact on AOM (12). Our data suggest that viruses and bacteria differ in their association with risk of AOM complicating URI. RSV had a synergistic interaction with S. pneumoniae when children experienced higher RSV viral loads. Since the introduction of PCV7, the proportion of AOM episodes associated with S. pneumoniae vaccine serotypes has decreased while the proportion with H. influenzae has increased (9, 13). The increased risk of AOM during hBOV infection and H. influenzae colonization is of concern given their high prevalence and the lack of vaccines for either pathogen.
Consistently with another study in the southern United States (16), we detected RSV infections year round. However, the highest RSV viral load occurred in winter months. Some studies have identified RSV as the most common virus in middle ear fluid samples (8, 19), and RSV is frequently detected as a middle ear pathogen in the absence of AOM bacterial pathogens (22). Our finding that high RSV load in the presence of S. pneumoniae is associated with increased risk of AOM suggests that RSV detection methods that quantify viral load will help identify subjects at the highest risk of AOM. The expanded use of the 13-valent pneumococcal conjugate vaccine may be associated with further reductions in AOM due to S. pneumoniae. However, the number of AOM episodes due to S. pneumoniae nonvaccine serotypes increased after the introduction of PCV7, and concerns remain regarding the potential for serotype replacement (9, 13).
The precise role of hBOV as a respiratory pathogen has been debated (11, 23, 25). hBOV has been identified in approximately 3 to 4% of middle ear fluid samples from children with AOM (9, 16). Martin et al. studied URI in young children attending daycare and identified hBOV in 44% of asymptomatic children at enrollment (25). hBOV viral load did not differ based on whether or not the children had URI symptoms (25). Moreover, hBOV-infected children can shed virus for several months (25, 34). Longtin et al. postulated that hBOV plays an important pathogenic role in AOM and that the high rates of hBOV in asymptomatic children could be due in part to prior hBOV-associated AOM episodes (23). In children experiencing symptomatic URI, we found that, regardless of the viral load, the presence of hBOV was associated with an increased risk of AOM among children colonized with H. influenzae.
We did not identify an increased risk of AOM complicating URI associated with hMPV. A prospective study in Finland indicated that AOM developed in more than 60% of children less than 3 years of age with hMPV (18). Our findings are consistent with a retrospective analysis of hMPV in children <5 years of age covering a 20-year period. While half of the children with hMPV were diagnosed with AOM, it was not associated with excess risk of AOM compared to that of other common respiratory viruses (35).
Previous studies have not conclusively identified interactions between specific viruses and bacteria and an increased risk of AOM. Investigators analyzing combined data from the Finnish OM cohort study and the Finnish OM vaccine trial were unable to detect significant associations between specific viruses and bacteria (22). All children in this study had AOM at the time of virus and bacterium sampling, whereas our study examined children with URI and monitored them for the development of AOM. Alper et al. prospectively followed children from 1 to 5 years of the age for URI symptoms and incident OM (2). While 56% of RSV infections were associated with OM, differences between specific viruses and OM were not statistically significant; data on bacterial colonization were not included in this analysis.
Our incidental finding of increased AOM risk associated with antibiotic use within the 7 days prior to URI is interesting. We previously showed that AOM risk was higher in children with one or more AOM bacterial pathogens in the nasopharynx than in those with no AOM bacterial pathogens (28). The use of antibiotics prior to URI development could lead to a reduction of nasopharyngeal colonization with bacterial AOM pathogens, thereby lowering the risk of AOM complicating URI. Other researchers showed that 7 days of amoxicillin-clavulanate, given early in the course of URI, did not prevent AOM (20). We hypothesized that the apparent increased AOM risk associated with antibiotic use is due in part to these children generally being more prone to URI and AOM. Post hoc analyses were done to compare the mean number of URI and AOM episodes in children who did and did not use antibiotics within the 7 days prior to swab collection; the children who used antibiotics averaged more URI episodes (means of 6.57 versus 2.86; P < 0.0001; t test) and more AOM episodes (means of 2.61 versus 1.05 AOM; P = 0.001; t test) compared to those who did not. The number of children who received antibiotics within the 7 days prior to URI was low (n = 23). Further studies with a larger number of children are required to determine whether antibiotic use prior to URI could increase the risk of AOM complicating URI.
Our study had several strengths, including the longitudinal prospective study design, careful documentation and diagnosis of AOM by team members, and our analyses of two recently identified viruses, hBOV and hMPV. Multiple viruses are frequently detected within children (10, 18, 26). We detected more than one respiratory virus in 11% of our specimens and controlled for the presence of AOM-associated viruses in our multivariate models. A recent study identified OM during asymptomatic URI (36). Since the children in our study all were experiencing URI symptoms, a limitation of our study is that we may have missed AOM associated with asymptomatic URI episodes.
In summary, we have identified specific bacterial and viral interactions associated with an increased risk of AOM complicating URI. These data will be useful for the targeting and design of effective AOM prevention and intervention strategies.
ACKNOWLEDGMENTS
This work was supported by the National Institute on Deafness and other Communication Disorders at the National Institutes of Health (R01 DC005841 to T.C.) and the National Institute of Allergy and Infectious Diseases at the National Institutes of Health (R01 AI068043 to M.M.P.). The study was conducted at the Clinical Research Center at the University of Texas Medical Branch, which is supported by the Clinical and Translational Science Award from the National Center for Research Resources at the National Institutes of Health (UL1 RR029876). The Galveston National Laboratory's assay development services division received support from the National Institute of Allergy and Infectious Diseases at the National Institutes of Health (AI057156).
We thank Janak Patel and Krystal Revai for their clinical contribution; Syed Ahmad, Michelle Tran, M. Lizette Rangel, Kyralessa B. Ramirez, Liliana Najera, Rafael Serna, and Carolina Pillion for assistance with study subjects; and Sangeeta Nair and Ying Xiong for assistance with study specimens.
Footnotes
Published ahead of print on 7 September 2011.
REFERENCES
- 1. Allander T., et al. 2005. Cloning of a human parvovirus by molecular screening of respiratory tract samples. Proc. Natl. Acad. Sci. U. S. A. 102:12891–12896 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Alper C. M., Winther B., Mandel E. M., Hendley J. O., Doyle W. J. 2009. Rate of concurrent otitis media in upper respiratory tract infections with specific viruses. Arch. Otolaryngol. Head Neck Surg. 135:17–21 [DOI] [PubMed] [Google Scholar]
- 3. Arnold J. C. 2010. Human bocavirus in children. Pediatr. Infect. Dis. J. 29:557–558 [DOI] [PubMed] [Google Scholar]
- 4. Bakaletz L. O. 2010. Immunopathogenesis of polymicrobial otitis media. J. Leukoc. Biol. 87:213–222 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Beder L. B., et al. 2009. Clinical and microbiological impact of human bocavirus on children with acute otitis media. Eur. J. Pediatr. 168:1365–1372 [DOI] [PubMed] [Google Scholar]
- 6. Bonroy C., Vankeerberghen A., Boel A., De Beenhouwer H. 2007. Use of a multiplex real-time PCR to study the incidence of human metapneumovirus and human respiratory syncytial virus infections during two winter seasons in a Belgian paediatric hospital. Clin. Microbiol. Infect. 13:504–509 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Bourne N., et al. 2005. Screening for hepatitis C virus antiviral activity with a cell-based secreted alkaline phosphatase reporter replicon system. Antiviral Res. 67:76–82 [DOI] [PubMed] [Google Scholar]
- 8. Bulut Y., et al. 2007. Acute otitis media and respiratory viruses. Eur. J. Pediatr. 166:223–228 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Casey J. R., Adlowitz D. G., Pichichero M. E. 2010. New patterns in the otopathogens causing acute otitis media six to eight years after introduction of pneumococcal conjugate vaccine. Pediatr. Infect. Dis. J. 29:304. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Chonmaitree T., et al. 2008. Viral upper respiratory tract infection and otitis media complication in young children. Clin. Infect. Dis. 46:815–823 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Christensen A., Nordbo S. A., Krokstad S., Rognlien A. G., Dollner H. 2010. Human bocavirus in children: mono-detection, high viral load and viraemia are associated with respiratory tract infection. J. Clin. Virol. 49:158–162 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Cripps A. W., Otczyk D. C. 2006. Prospects for a vaccine against otitis media. Expert Rev. Vaccines 5:517–534 [DOI] [PubMed] [Google Scholar]
- 13. Dupont D., et al. 2010. Evolving microbiology of complicated acute otitis media before and after introduction of the pneumococcal conjugate vaccine in France. Diagn. Microbiol. Infect. Dis. 68:89–92 [DOI] [PubMed] [Google Scholar]
- 14. Franz A., et al. 2010. Correlation of viral load of respiratory pathogens and co-infections with disease severity in children hospitalized for lower respiratory tract infection. J. Clin. Virol. 48:239–245 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Grijalva C. G., Nuorti J. P., Griffin M. R. 2009. Antibiotic prescription rates for acute respiratory tract infections in US ambulatory settings. JAMA 302:758–766 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Halstead D. C., Jenkins S. G. 1998. Continuous non-seasonal epidemic of respiratory syncytial virus infection in the southeast United States. South. Med. J. 91:433–436 [DOI] [PubMed] [Google Scholar]
- 17. Heikkinen T., Chonmaitree T. 2003. Importance of respiratory viruses in acute otitis media. Clin. Microbiol. Rev. 16:230–241 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Heikkinen T., Osterback R., Peltola V., Jartti T., Vainionpaa R. 2008. Human metapneumovirus infections in children. Emerg. Infect. Dis. 14:101–106 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Heikkinen T., Thint M., Chonmaitree T. 1999. Prevalence of various respiratory viruses in the middle ear during acute otitis media. N. Engl. J. Med. 340:260–264 [DOI] [PubMed] [Google Scholar]
- 20. Heikkinen T., Ruuskanen O., Ziegler T., Waris M., Puhakka H. 1995. Short-term use of amoxicillin-clavulanate during upper respiratory tract infection for prevention of acute otitis media. J. Pediatr. 126:313–316 [DOI] [PubMed] [Google Scholar]
- 21. Henderson F. W., et al. 1982. A longitudinal study of respiratory viruses and bacteria in the etiology of acute otitis media with effusion. N. Engl. J. Med. 306:1377–1383 [DOI] [PubMed] [Google Scholar]
- 22. Kleemola M., et al. 2006. Is there any specific association between respiratory viruses and bacteria in acute otitis media of young children? J. Infect. 52:181–187 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Longtin J., Gubbay J. B., Patel S., Low D. E. 2010. High prevalence of asymptomatic bocavirus in daycare: is otitis media a confounder? J. Infect. Dis. 202:1617. [DOI] [PubMed] [Google Scholar]
- 24. Lu X., et al. 2006. Real-time PCR assays for detection of bocavirus in human specimens. J. Clin. Microbiol. 44:3231–3235 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Martin E. T., et al. 2010. Frequent and prolonged shedding of bocavirus in young children attending daycare. J. Infect. Dis. 201:1625–1632 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Moore H. C., et al. 2010. The interaction between respiratory viruses and pathogenic bacteria in the upper respiratory tract of asymptomatic aboriginal and non-aboriginal children. Pediatr. Infect. Dis. J. 29:540–545 [DOI] [PubMed] [Google Scholar]
- 27. Patel J. A., Nguyen D. T., Revai K., Chonmaitree T. 2007. Role of respiratory syncytial virus in acute otitis media: implications for vaccine development. Vaccine 25:1683–1689 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Revai K., Mamidi D., Chonmaitree T. 2008. Association of nasopharyngeal bacterial colonization during upper respiratory tract infection and the development of acute otitis media. Clin. Infect. Dis. 46:e34–e37 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Ruohola A., et al. 2006. Microbiology of acute otitis media in children with tympanostomy tubes: prevalences of bacteria and viruses. Clin. Infect. Dis. 43:1417–1422 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Song J. R., et al. 2010. Novel human bocavirus in children with acute respiratory tract infection. Emerg. Infect. Dis. 16:324–327 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Suzuki A., et al. 2005. Detection of human metapneumovirus from children with acute otitis media. Pediatr. Infect. Dis. J. 24:655–657 [DOI] [PubMed] [Google Scholar]
- 32. van den Hoogen B. G., et al. 2001. A newly discovered human pneumovirus isolated from young children with respiratory tract disease. Nat. Med. 7:719–724 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Vergison A., et al. 2010. Otitis media and its consequences: beyond the earache. Lancet Infect. Dis. 10:195–203 [DOI] [PubMed] [Google Scholar]
- 34. von Linstow M. L., Hogh M., Hogh B. 2008. Clinical and epidemiologic characteristics of human bocavirus in Danish infants: results from a prospective birth cohort study. Pediatr. Infect. Dis. J. 27:897–902 [DOI] [PubMed] [Google Scholar]
- 35. Williams J. V., et al. 2006. The role of human metapneumovirus in upper respiratory tract infections in children: a 20-year experience. J. Infect. Dis. 193:387–395 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Winther B., Alper C. M., Mandel E. M., Doyle W. J., Hendley J. O. 2007. Temporal relationships between colds, upper respiratory viruses detected by polymerase chain reaction, and otitis media in young children followed through a typical cold season. Pediatrics 119:1069–1075 [DOI] [PubMed] [Google Scholar]
- 37. Zhou F., Shefer A., Kong Y., Nuorti J. P. 2008. Trends in acute otitis media-related health care utilization by privately insured young children in the United States, 1997-2004. Pediatrics 121:253–260 [DOI] [PubMed] [Google Scholar]