Regev-Yochay G, Dagan R, Raz M, Carmeli Y, Shainberg B, Derazne E, Rahav G, Rubinstein E. Association Between Carriage of Streptococcus pneumoniae and Staphylococcus aureus in Children. JAMA. 2004;292(6):716–720. doi:10.1001/jama.292.6.716
Context Widespread pneumococcal conjugate vaccination may bring about epidemiologic
changes in upper respiratory tract flora of children. Of particular significance
may be an interaction between Streptococcus pneumoniae and Staphylococcus aureus, in view of the recent emergence
of community-acquired methicillin-resistant S aureus.
Objective To examine the prevalence and risk factors of carriage of S pneumoniae and S aureus in the prevaccination
era in young children.
Design, Setting, and Patients Cross-sectional surveillance study of nasopharyngeal carriage of S pneumoniae and nasal carriage of S
aureus by 790 children aged 40 months or younger seen at primary care
clinics in central Israel during February 2002.
Main Outcome Measures Carriage rates of S pneumoniae (by serotype)
and S aureus; risk factors associated with carriage
of each pathogen.
Results Among 790 children screened, 43% carried S pneumoniae and 10% carried S aureus. Staphylococcus aureus carriage among S pneumoniae carriers was 6.5% vs 12.9% in S pneumoniae noncarriers. Streptococcus pneumoniae carriage among S aureus carriers was 27.5% vs 44.8% in S aureus noncarriers. Only 2.8% carried both pathogens concomitantly vs 4.3%
expected dual carriage (P = .03). Risk factors for S pneumoniae carriage (attending day care, having young
siblings, and age older than 3 months) were negatively associated with S aureus carriage.
Conclusions Streptococcus pneumoniae carriage, specifically
of vaccine-type strains, is negatively associated with S aureus carriage in children. The implications of these findings in
the pneumococcal vaccine era require further investigation.
Streptococcus pneumoniae and Staphylococcus aureus are common inhabitants of the upper respiratory
tract in children and are responsible for common infections. Carriage of S aureus and S pneumoniae can
result in bacterial spread and endogenous infections.1- 3Streptococcus pneumoniae is carried in the nasopharynx
by most children at least once during early childhood1 but
not frequently by adults.4Staphylococcus aureus is carried by 10% to 35% of children5- 7 and by approximately
35% of the general adult population.3Staphylococcus aureus is carried most consistently in the
Various studies have explored bacterial interference—the suppression
of one species by another.8- 11 However,
studies examining possible interference between S aureus and S pneumoniae are noticeably absent. An
association between these 2 pathogens may suggest epidemiologic changes that
could follow widespread vaccination with pneumococcal conjugate vaccines.
We investigated the possible association between the 2 pathogens by
studying their prevalence and risk factors for carriage in young children
in a region where pneumococcal conjugate vaccination is not practiced.
The study was approved by the Sheba Medical Center Ethics Committee,
Ramat-Gan, Israel. Informed consent was obtained from all parents.
During February 2002, children aged 40 months or younger seen for any
reason in 53 participating primary care pediatric clinics of a major health
maintenance organization (Maccabi Healthcare Services, Tel-Aviv) were enrolled.
The clinics were located in 4 large cities in the central district of Israel,
inhabited by a middle-class Jewish population. Each child was included once;
accompanying adults (usually parents) were also screened. None of the children
or their contacts received pneumococcal vaccine.
Nasopharyngeal and nasal swabs were obtained from children and their
accompanying adults, who also responded to an interviewer-administered questionnaire
including demographic characteristics, number of young siblings (aged <6
years), day care attendance, prior antibiotic treatment, and smoking habits
of family members. The physician's diagnosis on the screening day and medical
and immunization histories were obtained from patients' files. The diagnoses
were categorized as respiratory infections, skin diseases (including infections),
other infections (including urinary tract infections and enteric infections),
and noninfectious diagnoses.
Nasopharyngeal cultures were obtained with a rayon-tipped wire swab
and nasal cultures of both nares were obtained with a sterile cotton polyester
swab. Swabs were placed in Amies transport medium (Copan, Brescia, Italy).
All specimens were processed within 6 hours.
Nasopharyngeal swabs for S pneumoniae isolation
were streaked onto tryptic soy agar plates with 5% sheep blood and 5 µg/mL
of gentamicin (HyLabs, Rehovot, Israel) and incubated aerobically at 35°C
in 5% CO2-enriched air. Suspect colonies were isolated and identified
according to National Committee for Clinical Laboratory Standards recommendations.12 Serotyping of S pneumoniae was
performed using antisera (Statens Serum Institute, Copenhagen, Denmark). Vaccine
types were defined as serotypes included in the current 7-valent conjugate
vaccine as well as the cross-reactive types (ie, serogroups 4, 6, 9, 14, 18,
19, and 23).
Nasal swabs for S aureus isolation were streaked
onto tryptic soy agar plates with 5% sheep blood. Staphylococcus
aureus was identified by morphology, β-hemolysis, catalase, DNAase,
and coagulase production.
We expected 15% of children to carry S aureus and
50% to carry S pneumoniae. To detect a difference
of at least 7% in S aureus carriage rates among S pneumoniae carriers and noncarriers with α = .05
and 80% power, a sample size of 353 children in each group was needed.
Odds ratios (ORs) and Fisher exact tests were calculated to assess risk
factors for carriage of each organism including age, sex, young siblings,
dwelling density, passive smoking, day care attendance, respiratory tract
infection diagnosis, chronic and recurrent diseases, S pneumoniae or S aureus carriage by the acccompanying
adult, steroid treatment, number of clinic visits and hospitalization in the
last 6 months, and antibiotic treatment in the last month. Mantel-Haenszel
common ORs and the Breslow-Day test for homogeneity were used to control for
possible confounding variables (age and day care attendance).
A multivariate logistic regression model with stepwise backward elimination
was performed separately for each pathogen. Variables with P<.10 in the univariate analysis were included. Interactions of S pneumoniae with day care attendance, age, and having
young siblings were also included in the model for S aureus carriage. The criterion for entering into the model was a score statistic
of P = .05. A Wald statistic of P = .10 was used to remove a variable from the model. −2 Log
likelihood, the Nagelkerke R2, and the
Hosmer-Lemeshow test were used to assess goodness of fit. Risk factors were
checked for confounding and collinearity. Cross-validation was used to assess
overfitting. All tests used were 2-tailed, and P<.05
was considered statistically significant. Computations were performed with
SPSS software, version 11.0 (SPSS Inc, Chicago, Ill) and S-Plus, version 6.2
(Insightful Corp, Seattle, Wash).
A total of 790 children (90% of children approached) aged 5 days to
40 months (median, 1.3 years) were screened. Fifty-five percent were male.
A total of 6.1% came for healthy check-up visits and 80% were diagnosed as
having a respiratory tract infection. Chronic or recurrent disease (ie, asthma,
recurrent otitis media, recurrent pneumonia, or skin disorders) was present
Staphylococcus aureus and S pneumoniae were isolated in 80 children (10.1%) and 340 children
(43.0%), respectively. The proportion of vaccine-type strains among S pneumoniae carriers was 74.2%. Staphylococcus
aureus carriage among S pneumoniae carriers
was 6.5% vs 12.9% in S pneumoniae noncarriers. Streptococcus pneumoniae carriage among S aureus carriers was 27.5% vs 44.8% in S aureus noncarriers. If carriage of the 2 organisms were independent (ie,
occurring at random), the expected dual carriage would be 4.3%. However, dual
carriage was found in only 22 children (2.8%) (OR, 0.47; 95% confidence interval
[CI], 0.28-0.78; P = .03 by Fisher exact test).
Seven hundred four adults (621 mothers [88%], 77 fathers, and 6 other
family members) aged 18 to 45 years (median, 30 years) were screened for S aureus nasal carriage and 693 for both pathogens. Staphylococcus aureus and S pneumoniae were isolated from 182 (25.9%) of 704 and 35 (5.1%) of 693, respectively.
The proportion of vaccine-type strains among S pneumoniae carriers was 66.7%. Staphylococcus aureus carriage
among adults was similar in S pneumoniae carriers
and noncarriers (25.7% and 25.3%, respectively), and S pneumoniae carriage was similar in S aureus carriers
and noncarriers (5.1% and 5.0%, respectively). Dual carriage was found in
1.3%, the same as the expected prevalence (OR, 1.02; 95% CI, 0.47-2.23; P = .96 by Fisher exact test).
The highest S aureus carriage rate (30%) was
observed in children aged 3 months or younger, in whom S pneumoniae prevalence was lowest (9%) (Figure 1). The highest S pneumoniae carriage
rate (approximately 50%) was in children aged 7 to 40 months, in whom S aureus prevalence was the lowest (5%-9%).
In a univariate analysis, age older than 3 months, having young siblings,
day care attendance, respiratory tract infection at screening, and prior steroid
treatment were risk factors for S pneumoniae carriage
in children (Table 1). These same
factors were inversely associated with S aureus carriage.
Being an S pneumoniae carrier was inversely associated
with S aureus carriage and vice versa (P = .003). Antibiotic treatment during the prior month reduced carriage
of both pathogens. Having a parent who was an S aureus carrier
was a risk factor for carrying S aureus, but having
a parent who was an S pneumoniae carrier was not
associated with S pneumoniae carriage. To assess
for confounding effects, the inverse relation of carriage was analyzed while
controlling for age and day care attendance. In children, the prevalence of S aureus was lower among S pneumoniae carriers compared with S pneumoniae noncarriers
in all age groups (Mantel-Haenszel OR, 0.51; 95% CI, 0.29-0.89; P = .02) (Figure 2). There
was no evidence that this association was modified by age (Breslow-Day χ2 = 1.716; P = .63).
The association between S pneumoniae and S aureus carriage was stratified by day care attendance.
Among day care attendees, S aureus carriage was significantly
lower among S pneumoniae carriers compared with S pneumoniae noncarriers (3.0% vs 9.8%; OR, 0.28; 95% CI,
0.11-0.73; P = .005). Among non–day care attendees, S aureus carriage rates were 11.3% and 11.7% in S pneumoniae carriers and noncarriers, respectively (OR, 0.96; 95%
CI, 0.48-1.94; P = .54; conditional independence
Mantel-Haenszel χ2 = 2.858, P = .09;
homogeneity Breslow-Day χ2 = 4.313; P =
Risk factors for S pneumoniae carriage in a
multivariate logistic analysis were day care attendance (OR, 2.18; 95% CI,
1.56-3.06; P<.001); having young siblings (<6
years) (OR, 1.86; 95% CI, 1.34-2.56; P<.001);
and age older than 3 months (OR, 5.93; 95% CI, 2.28-15.42; P<.001); antibiotic treatment during the previous month was inversely
related (OR, 0.60; 95% CI, 0.43-0.85; P = .004).
Sex, S pneumoniae carriage by parents, respiratory
tract infection at screening, and S aureus carriage
did not significantly affect S pneumoniae carriage
(−2 log likelihood = 868.59; Nagelkerke R2 = 0.131; Hosmer-Lemeshow χ25 = 1.199; P = .95).
The only significant risk factor for S aureus carriage
in a multivariate logistic analysis model was having an S aureus carrier parent (OR, 2.60; 95% CI, 1.48-4.55; P = .001), while several factors were inversely related: carrying S pneumoniae while attending day care (OR, 0.27; 95% CI,
0.10-0.72; P = .009); having young siblings (OR,
0.52; 95% CI, 0.30-0.91; P = .02); and age older
than 3 months (OR, 0.51; 95% CI, 0.21-1.26; P = .15
for age 4-6 months; OR, 0.31; 95% CI, 0.13-0.74, P =
.009 for 7-12 months; OR, 0.14; 95% CI, 0.05-0.36; P<.001
for 13-24 months; and OR, 0.35; 95% CI, 0.15-0.35, P =
.02 for 25-40 months). Sex and skin disease did not have a significant influence
(−2 log likelihood = 367.63; Nagelkirk R2 = 0.178; Hosmer = Lemeshow χ28 = 9.448; P = .31). When the analysis was repeated but restricted
to vaccine-type S pneumoniae carriage, the results
were almost identical (OR, 0.22; 95% CI, 0.07-0.74; P =
.02 for being a vaccine-type S pneumoniae carrier
and attending day care). Analysis of non–vaccine-type S pneumoniae did not yield any associations with S aureus carriage.
We have shown an inverse relation between S aureus and S pneumoniae carriage (specifically of
vaccine-type strains) in children. Factors positively associated with S pneumoniae carriage were negatively associated with S aureus carriage. The distribution of S aureus carriage by age was a mirror image of that of S pneumoniae, an observation also reported in older children.13 In addition, concurrent carriage of S pneumoniae and S aureus in children was
significantly lower than expected.
This apparent inverse relationship could be due to bacterial interference
or could be a consequence of confounding effects. Age was considered a possible
confounder because the association between the 2 pathogens was not demonstrated
in the adult group and the tendency to carry different pathogens at different
age groups is well known (eg, pharyngeal Streptococcus pyogenes at age 6-14 years14; nasopharyngeal
meningococcus in adolescents15). Nevertheless,
when controlling for age, the negative association between S pneumoniae and S aureus carriage among children
Although close contacts and poor hygienic conditions (eg, in day care
centers and among young siblings) are considered to increase S aureus transmission,16 in our study
day care attendance was surprisingly inversely related to S aureus carriage (when combined with S pneumoniae carriage). A possible explanation is that day care attendance may
indicate prolonged S pneumoniae carriage (the inhibitory
factor against S aureus), while this study measured
point prevalence solely. Alternatively, day care attendance could be an indicator
of the presence of another bacterial or viral pathogen that interferes with S aureus.
Bacterial interference could explain the inverse relation between S aureus and S pneumoniae. This
phenomenon has been reported between S aureus and
coagulase-negative staphylococci,9Corynebacterium,8 and viridans group
streptococci.10 Hydrogen peroxide has been
suggested to be the inhibitory factorof viridans streptococci on S aureus10 and of S pneumoniae on other respiratory tract bacteria.11
Pneumococcal conjugate vaccines reduce nasopharyngeal carriage of vaccine-type S pneumoniae. Our finding of an inverse relationship between
vaccine-type S pneumoniae and S aureus may imply an upcoming shift, not only toward nonvaccine S pneumoniae serotypes17 but
also toward higher S aureus carriage rates in children.
This would be particularly disturbing in light of the emergence of community-associated
methicillin-resistant S aureus.5,6,18 This
possibility is supported by a recent report of an increased rate of S aureus culture-positive draining ears in vaccinated children
compared with controls.19
Limitations of this study include being a point prevalence study of
children visiting primary care clinics that may underrepresent healthy pediatric
populations; other pathogens potentially involved in bacterial interference
were not studied and antibiotic use was examined only for the prior month.
Longitudinal studies including older children and additional pathogens could
further refine this association.
Our study suggests a protective role of S pneumoniae carriage against S aureus carriage. Studies
measuring the effect of vaccination on S pneumoniae epidemiology
should also examine concurrent changes in S aureus.