Professional Documents
Culture Documents
37503755
0095-1137/11/$12.00 doi:10.1128/JCM.01186-11
Copyright 2011, American Society for Microbiology. All Rights Reserved.
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 2 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.
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 (Mannedorf, 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 2 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
Age at enrollment, mo
612
1218
1824
2436
Gender
Female
Male
Race
White
Black
Hispanic
Other
Day care
No
Yes
Breastfed for 3 months
No
Yes
Environmental tobacco
smoke exposure
No
Yes
Antibiotic use within 7
days of URI
None
Any
a
b
Total no.
(%)
No. (%)
diagnosed with
AOM (anya)
84 (43)
55 (28)
28 (14)
27 (14)
59 (70)
31 (56)
17 (61)
15 (56)
95 (49)
99 (51)
58 (61)
64 (65)
39 (20)
56 (29)
83 (43)
16 (8)
25 (64)
32 (57)
51 (61)
14 (88)
133 (69)
61 (31)
85 (64)
37 (61)
127 (66)
67 (34)
83 (65)
39 (58)
P valueb
0.30
0.60
0.17
0.66
0.33
0.37
141 (73)
53 (27)
86 (61)
36 (68)
0.003
171 (88)
23 (12)
101 (59)
21 (91)
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; 2 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
3751
3752
PETTIGREW ET AL.
J. CLIN. MICROBIOL.
No. (%) of
subjects
1 ...............................................................................................62 (32)
2 ...............................................................................................41 (21)
34 ...........................................................................................42 (21)
5............................................................................................49 (25)
a
Virus/bacteria
No. (%) of
specimens
No. (%) of
specimens with
AOM diagnosis
Total
640
239 (37)
Adenovirus by culture
Absent
Present
619 (97)
21 (3)
226 (36)
13 (62)
548 (86)
92 (14)
200 (36)
39 (42)
483 (75)
157 (25)
170 (35)
69 (44)
596 (93)
44 (7)
223 (37)
16 (36)
353 (55)
287 (45)
110 (31)
129 (45)
446 (70)
194 (30)
138 (31)
101 (52)
226 (35)
414 (65)
60 (27)
179 (43)
Virus by PCR
RSV
Absent
Present
hBOV
Absent
Present
hMPV
Absent
Present
Bacteria by culture
S. pneumoniae
Absent
Present
H. influenzae
Absent
Present
M. catarrhalis
Absent
Present
a
P valuea
0.02
0.28
0.05
0.89
0.0003
<0.0001
<0.0001
P values are from 2 tests. Significant P values (0.05) are shown in boldface.
We next evaluated the risk of AOM associated with categories of viral load for RSV, hBOV, and hMPV. The repeatedmeasure 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).
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.
distribution of URI episodes by child is in Table 2. Approximately 25% of the children contributed 55% of the specimens.
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.
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; 2 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; 2 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.
OR (95% CI)
Factor(s)
RSVa
None (reference)
Low
Med
High
548 (85.6)
31 (4.8)
30 (4.7)
31 (4.8)
1.00
1.16 (0.65, 2.04)
0.96 (0.42, 2.18)
2.60 (1.30, 5.20)
hBOVb
None (reference)
Low
Med
High
483 (75.5)
51 (8.0)
54 (8.4)
52 (8.1)
hMPVc
None (reference)
Low
Med
High
596 (93.1)
15 (2.3)
15 (2.3)
14 (2.2)
Adenovirus
Absent (reference)
Present
619 (96.7)
21 (3.3)
No. (%) of
samples
OR (95% CI)
345 (53.9)
264 (41.2)
8 (1.2)
23 (3.6)
1.00
1.27 (0.86, 1.86)
2.01 (0.24, 4.77)
4.40 (1.90, 10.19)
1.00
1.00 (0.56, 1.79)
1.50 (0.92, 2.45)
0.97 (0.51, 1.88)
346 (54.1)
137 (21.4)
100 (15.6)
57 (8.9)
1.00
2.04 (1.38, 3.02)
0.88 (0.54, 1.43)
3.61 (1.90, 6.86)
1.00
1.48 (0.53, 5.58)
0.75 (0.22, 2.53)
0.43 (0.15, 1.29)
hMPV
Absent (reference)
Present
596 (93.1)
44 (6.9)
1.00
0.81 (0.41, 1.59)
Adenovirus
Absent (reference)
Present
619 (96.7)
21 (3.3)
1.00
3.06 (1.36, 6.89)
M. catarrhalis
Absent (reference)
Present
226 (35.3)
414 (64.7)
1.00
1.85 (1.28, 2.68)
1.00
4.64 (1.42, 7.07)
S. pneumoniae
Absent (reference)
Present
353 (55.2)
287 (44.8)
1.00
1.25 (0.85, 1.84)
H. influenzae
Absent (reference)
Present
446 (69.7)
194 (30.3)
1.00
2.41 (1.66, 3.48)
M. catarrhalis
Absent (reference)
Present
226 (35.3)
414 (64.7)
1.00
1.91 (1.31, 2.78)
a
RSV (copies/ml): none, 0; low, 0 to 1.26 105; med (medium), 1.26 105
to 3.16 107; high, 3.16 107.
b
hBOV (copies/ml): none, 0; low, 0 to 5.01 105; med, 5.01 105 to
2.51 107; high, 2.51 107.
c
hMPV (copies/ml): none, 0; low, 0 to 3.16 106; med, 3.16 106 to
6.31 108; high, 6.31 108.
d
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.
a
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).
b
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.
No. (%) of
samples
Factor
3753
3754
PETTIGREW ET AL.
J. CLIN. MICROBIOL.
3.
4.
5.
6.
7.
8.
9.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
3755
Short-term use of amoxicillin-clavulanate during upper respiratory tract infection for prevention of acute otitis media. J. Pediatr. 126:313316.
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.
306:13771383.
Kleemola, M., et al. 2006. Is there any specific association between respiratory viruses and bacteria in acute otitis media of young children? J. Infect.
52:181187.
Longtin, J., J. B. Gubbay, S. Patel, and D. E. Low. 2010. High prevalence of
asymptomatic bocavirus in daycare: is otitis media a confounder? J. Infect.
Dis. 202:1617.
Lu, X., et al. 2006. Real-time PCR assays for detection of bocavirus in human
specimens. J. Clin. Microbiol. 44:32313235.
Martin, E. T., et al. 2010. Frequent and prolonged shedding of bocavirus in
young children attending daycare. J. Infect. Dis. 201:16251632.
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. 29:540545.
Patel, J. A., D. T. Nguyen, K. Revai, and T. Chonmaitree. 2007. Role of
respiratory syncytial virus in acute otitis media: implications for vaccine
development. Vaccine 25:16831689.
Revai, K., D. Mamidi, and T. Chonmaitree. 2008. Association of nasopharyngeal bacterial colonization during upper respiratory tract infection and
the development of acute otitis media. Clin. Infect. Dis. 46:e34e37.
Ruohola, A., et al. 2006. Microbiology of acute otitis media in children with
tympanostomy tubes: prevalences of bacteria and viruses. Clin. Infect. Dis.
43:14171422.
Song, J. R., et al. 2010. Novel human bocavirus in children with acute
respiratory tract infection. Emerg. Infect. Dis. 16:324327.
Suzuki, A., et al. 2005. Detection of human metapneumovirus from children
with acute otitis media. Pediatr. Infect. Dis. J. 24:655657.
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.
Vergison, A., et al. 2010. Otitis media and its consequences: beyond the
earache. Lancet Infect. Dis. 10:195203.
von Linstow, M. L., M. Hogh, and B. Hogh. 2008. Clinical and epidemiologic
characteristics of human bocavirus in Danish infants: results from a prospective birth cohort study. Pediatr. Infect. Dis. J. 27:897902.
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.
193:387395.
Winther, B., C. M. Alper, E. M. Mandel, W. J. Doyle, and J. O. Hendley.
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 119:10691075.
Zhou, F., A. Shefer, Y. Kong, and J. P. Nuorti. 2008. Trends in acute otitis
media-related health care utilization by privately insured young children in
the United States, 19972004. Pediatrics 121:253260.
10.