Microbial Interactions during Upper Respiratory Tract Infections

Competitive interactions between bacteria differ by number and species present; thus, vaccination and treatment strategies may alter nasopharyngeal flora and disease susceptibility.


S treptococcus pneumoniae, Haemophilus infl uenzae,
Moraxella catarrhalis, and Staphylococcus aureus often asymptomatically colonize the nasopharynx of young children and are also associated with disease. S. pneumoniae, H. infl uenzae, and M. catarrhalis are the 3 most common otitis media pathogens (1,2). S. pneumoniae are also common causes of pneumonia, sepsis, and meningitis in young children (3). The proportion of young children colonized with any of these 3 bacteria species can be >50% in certain populations (4)(5)(6). S. aureus strains colonize up to 35% of young children and are associated with a wide range of diseases including soft tissue infections, sepsis, and pneumonia (7,8). Increases in the incidence of disease caused by community-acquired methicillin-resistant S. aureus are of great concern (9).
Host factors have been shown to infl uence colonization with S. pneumoniae, H. infl uenzae, M. catarrhalis, and S. aureus. These include host immunity, age, gender, race, out-of-home daycare, breastfeeding, and environmental exposure to tobacco smoke (10). The magnitude of host effects may differ by bacteria species.

CME ACTIVITY
Medscape, LLC is pleased to provide online continuing medical education (CME) for this journal article, allowing clinicians the opportunity to earn CME credit. Medscape, LLC is accredited by the Accreditation Council for Continuing Medical Education (ACCME) to provide CME for physicians. Medscape, LLC designates this educational activity for a maximum of 1.0 AMA PRA Category 1 Credits™. Physicians should only claim credit commensurate with the extent of their participation in the activity. All other clinicians completing this activity will be issued a certificate of participation. To participate in this journal CME activity: (1) review the learning objectives and author disclosures; (2) study the education content; (3) take the post-test and/or complete the evaluation at http://www.medscape.com/cme/eid; (4) view/print certifi cate.

Learning Objectives
Upon completion of this activity, participants will be able to: Identify common bacterial isolates from children with upper respiratory infection • be positively associated; this occurs when they are found together more often than would be expected by chance. A negative association could occur when bacteria compete within same environment. Several studies have described a negative association between S. pneumoniae and S. aureus (12)(13)(14)(15)(16). Understanding of interactions between S. pneumoniae, H. infl uenzae, M. catarrhalis, and S. aureus is limited.
The nasopharyngeal fl ora change over time; the level of bacteria colonization is higher during upper respiratory infection (URI) (6,17). Knowledge is lacking regarding S. pneumoniae, H. infl uenzae, M. catarrhalis, and S. aureus interactions during URI because colonization studies either do not examine competitive interactions among all 4 pathogens or focus on healthy children (5,11,16,18). Children are susceptible to secondary bacterial infections during and after URI (19)(20)(21).
A better understanding of polymicrobial interactions in the nasopharynx is important for several reasons. Colonization is the initial step in the disease process (22,23). Colonized children serve as reservoirs for bacterial transmission to others in the community (24). Additionally, antibimicrobial drugs or vaccines, which target specifi c bacteria species, may alter polymicrobial interactions in the nasopharynx and have unanticipated consequences (25,26). The goals of our study were to 1) describe the prevalence of colonization with S. pneumoniae, H. infl uenzae, M. catarrhalis, and S. aureus; 2) evaluate interactions between S. pneumoniae, H. infl uenzae, M. catarrhalis, and S. aureus; and 3) estimate the effect of host factors on colonization with S. pneumoniae, H. infl uenzae, M. catarrhalis, and S. aureus after a URI in a prospective cohort of young children.

Study Design and Participants
We used data from a prospective study of otitis media complications of URI in children at the University of Texas Medical Branch (UTMB) at Galveston (19,26). The study was reviewed and approved by the UTMB Institutional Review Board. The parents of healthy children 6 months through 3 years of age, who were receiving medical care at UTMB from January 2003 through March 2007, were invited to enroll their children. Children with chronic medical problems and anatomic or physiologic defects of the ear or nasopharynx were excluded.
At enrollment, we collected information about demographic and URI risk factors. Parents were asked to describe their child's race and ethnicity. We also obtained information regarding the number of weeks the child had been breast-fed and the number of hours and days/week the child currently attended day care. We ascertained environmental exposure to tobacco smoke based on self-reports of whether any household members smoked cigarettes in the home.
The children in our study were followed up for 1 year. We requested that parents notify study staff when the child began to exhibit URI symptoms including nasal congestion, rhinorrhea, cough, sore throat, or fever. A study physician saw children as soon as possible after the onset of URI symptoms. At each study visit, the study physician obtained information regarding specifi c URI symptoms and examined the child's ears. The children were then monitored closely for 3 weeks for the development of otitis media. The study physician collected a nasopharyngeal swab during the visit for each URI episode and when acute otitis media or sinusitis was diagnosed. URI episodes were categorized as the same episode if symptoms persisted. An episode of URI was considered new when symptoms of the previous episode subsided and the parents noted new symptoms of URI as described above. Given our prospective study design, many children had >1 URI episode and some had >1 visit/URI episode. We collected 1 swab/physician visit. Data regarding antimicrobial drug therapy during the past 7 days were collected by medical record review. A description of the methods is provided elsewhere (19,26).
A total of 294 children were enrolled in the original study (19,26). Included in these analyses are data from 212 (72%) children who experienced at least 1 URI, were seen by a study physician, and had a nasopharyngeal swab collected for bacterial culture. Thus, we excluded 82 children who did not have a URI and a swab for bacterial culture. Of these 82 children without URI visits, 35 (59%) were lost to follow-up in the fi rst 6 months, 13 (38%) were lost to follow-up in months 7-11, and 34 (17%) completed 1 year of follow up.
Mini-Tip Culturette kits (Becton Dickinson Microbiology Systems, Cockeysville, MD, USA) were used for sample collection. Each swab was streaked onto 1 blood and 1 chocolate agar plate. We subcultured and identifi ed suspected isolates of each species as follows: S. pneumoniae isolates were identifi ed by using the optochin disk susceptibility test (Taxo P, Becton Dickinson Microbiology Systems), H. infl uenzae isolates were identifi ed by the Haemophilus ID Quad Plate with Growth Factors (Becton Dickinson Microbiology Systems), M. catarrhalis isolates were identifi ed by the API QuadFerm assay (bioMérieux, Inc., Hazelwood, MO, USA), and S. aureus isolates were identifi ed by coagulase, catalase, and latex agglutination test (Staphaurex Plus, Remel, Lenexa, KS, USA).

Statistical Methods
The main outcomes of interest were the relationships between bacteria during URI. All statistical analyses were conducted by using SAS version 9.1 (SAS Institute, Inc., Cary, NC, USA). We examined colonization by S. pneumoniae, H. infl uenzae, M. catarrhalis, and S. aureus by using repeated measures logistic regression with generalized estimating equations and an autoregressive correlation structure (AR1) using the procedure PROC GENMOD (SAS Institute, Inc.). Because each child could potentially have multiple URI episodes and contribute multiple bacterial swabs to the analysis, we used a repeated measures design to take into account variability of multiple samples from each child. To examine the effect of covariates on each bacteria species, we modeled colonization by S. pneumoniae, H. infl uenzae, and M. catarrhalis separately. We did not separately model the outcome of colonization by S. aureus because of low numbers of isolates obtained. Each model included the presence or absence of other bacteria species, as well as potential sampling-time confounders comprising time of swab collection after URI onset, antimicrobial drug therapy within the past 7 days, and age of the child at the time of swab collection. Host factors included in the model were gender, race, day care, breast-fed for >4 months, and environmental exposure to tobacco smoke.

Results
Characteristics of the study participants are shown in Table 1. The median age of study participants was 12.0 months; mean age was 14.1 (SD 7.4) months. Most children were white, were cared for at home, and had not been breast-fed for >4 months. Children were followed up for a median of 12 months and a mean of 10.  Table 2.
Most swabs (849 [87.7%]) were collected within 7 days of URI onset; 119 (12.3%) were taken 8-30 days after URI onset. Of the 968 swab samples, only 54 (5.6%) were collected from children who had taken antimicrobial drugs within the past 7 days. Therefore, most swabs were collected from children who were not taking antimicrobial drugs at the time of swab collection (94.5%). Of the 212 children, 205 (>96%) had received at least 1 dose of the 7-valent pneumococcal conjugate vaccine (PCV7) at the time of enrollment. Most of the children had received all age-appro-priate scheduled PCV7 vaccinations at their URI visit, 666 (69%) of samples were collected from children who had received the age-appropriate number of PCV7 doses at the time of swab collection. There was no association between being up to date with PCV7 vaccination and colonization with S. pneumoniae (p = 0.71). We did not further examine the effect of the pneumococcal vaccine further because of the high level of coverage in our study population.
Repeated measures logistic regression models predicting colonization by S. pneumoniae, H. infl uenzae, or M. catarrhalis are shown in Table 3. A positive association between bacteria is indicated by an odds ratio (OR) ≥1; a negative association is indicated by an OR <1. An OR of 1.0, or any 95% confi dence interval that includes 1.0 indicates no signifi cant association. The model predicting colonization by S. pneumoniae indicated that colonization by H. infl uenzae was negatively associated with S. pneumoniae. However, when H. infl uenzae and M. catarrhalis colonized together, they were positively associated with S. pneumoniae colonization. Colonization by S. aureus resulted in *Data given for 212 participants who experienced at least 1 upper respiratory infection, were seen by a study physician, and had a nasopharyngeal swab collected for bacterial culture. An additional 82 enrollees were excluded from the study because they did not experience an upper respiratory infection and did not have a nasopharyngeal swab collected for bacterial culture. Some numbers do not add up to 212 because of missing data. †No. hours and days/week in day care were grouped into any or none. ‡Environmental exposure to tobacco smoke was based on parental selfreport.
a 40% reduction in the odds of S. pneumoniae colonization. Older children were less likely to be colonized with S. pneumoniae; each 1-month increase in age was associated with a 2% decrease in the odds of S. pneumoniae colonization (Table 3). Antimicrobial drug therapy in the past 7 days was associated with decreased odds of S. pneumoniae colonization. The timing of swab collection after onset of URI symptoms and host characteristics such as gender, race, daycare, breastfeeding, and environmental exposure to tobacco smoke were not associated with colonization by S. pneumoniae.
In our model examining H. infl uenzae colonization as the outcome, H. infl uenzae was negatively associated with S. pneumoniae, M. catarrhalis, and S. aureus (Table 3). In contrast to their association with S. pneumoniae colonization, age and antimicrobial drug therapy during the past 7 days were not signifi cantly associated with colonization by H. infl uenzae. Host characteristics were associated with colonization by H. infl uenzae. Male gender and out-ofhome daycare were associated with increased odds of H. infl uenzae colonization. White race was associated with decreased odds of H. infl uenzae colonization.
Our third model examined factors associated with colonization by M. catarrhalis (Table 3). H. infl uenzae was negatively associated with colonization by M. catarrhalis, but when H. infl uenzae and S. pneumoniae colonized together, they were positively associated with coloniza-tion by M. catarrhalis. Older children were less likely to be colonized with M. catarrhalis; each 1-month increase in age was associated with a 2% decrease in the odds of M. catarrhalis colonization (Table 3). Antimicrobial drug therapy in the past 7 days was associated with decreased odds of M. catarrhalis colonization. The timing of swab collection after onset of URI symptoms and host characteristics such as gender, race, daycare, breastfeeding, and environmental exposure to tobacco smoke were not associated with colonization by M. catarrhalis.

Discussion
We describe nasopharyngeal colonization of children with S. pneumoniae, H. infl uenzae, M. catarrhalis, and S. aureus alone or in combination during URI. Our models predicting S. pneumoniae colonization indicated that H. infl uenzae is negatively associated with S. pneumoniae. However, when H. infl uenzae was present with M. catarrhalis, odds of S. pneumoniae colonization increased by >2-fold. Models predicting H. infl uenzae colonization indicated a negative association with S. pneumoniae, M. catarrhalis, and S. aureus. Competitive interactions between bacteria are complex after URI and may shift from negative to positive when additional bacteria species are present. Modeling S. pneumoniae, H. infl uenzae, and M. catarrhalis colonization separately showed that gender, race, and daycare were associated with colonization by H.  (11). Their study differed from ours in that they examined the relationship between S. pneumoniae, H. infl uenzae, M. catarrhalis, and S. aureus in pairwise combinations. These researchers also examined healthy children; we examined children who had a URI. Jacoby et al. observed positive associations between pairwise combinations of S. pneumoniae and H. infl uenzae and between S. pneumoniae and M. catarrhalis. They did not identify an association between S. pneumoniae and S. aureus or between H. infl uenzae and S. aureus (11).
Our results confi rm a recent report describing a negative association between H. infl uenzae and S. aureus in HIV-negative children (12). Our data also support a grow-
An in vivo mouse model of competitive interactions between S. pneumoniae and H. infl uenzae has suggested mechanisms to explain our observations (27). Both S. pneumoniae and H. infl uenzae successfully colonized BALBc/ SCID mice when each bacteria species was injected separately. However, S. pneumoniae was cleared rapidly when H. infl uenzae was present in a co-colonization model. The competitive interaction between H. infl uenzae and S. pneumoniae was dependent on complement and neutrophils (27). These researchers proposed that H. infl uenzae cellular components activate the host innate immune response, thus killing S. pneumoniae (27). M. catarrhalis was not examined in this model, but our data suggest that the additional presence of M. catarrhalis might alter the competitive balance between S. pneumoniae and H. infl uenzae and that all 3 bacteria species would successfully colonize.
In vitro studies have also demonstrated competition between H. infl uenzae and S. pneumoniae but predicted that S. pneumoniae should inhibit the growth of H. infl uenzae. Neuraminidase A is produced by S. pneumoniae and cleaves sialic acid. It has been shown to remove sialic acid from lipopolysaccharides of H. infl uenzae strains (28), potentially giving pneumococci a competitive advantage by making H. infl uenzae more susceptible to complement-mediated clearance. Furthermore, in vitro co-culture experiments indicate that S. pneumoniae can inhibit H. infl uenzae through the action of hydrogen peroxide (29). Interference between S. pneumoniae and S. aureus may also be caused by hydrogen peroxide production by S. pneumoniae (30).
Our results indicate that antimicrobial drug therapy in the past 7 days was associated with a lower prevalence of colonization with S. pneumoniae or M. catarrhalis. In contrast, antimicrobial drug therapy in the past 7 days was not associated with colonization by H. infl uenzae. Varon et al. studied the effect of antimicrobial drugs on colonization with S. pneumoniae, H. infl uenzae, and M. catarrhalis in a cohort of young children with URI (31). Children in this study received antimicrobial drugs for a mean treatment period of 8 days. Swab samples were taken before treatment and on days 2 through 6 after treatment. Results showed that colonization by S. pneumoniae, H. infl uenzae, and M. catarrhalis decreased after antimicrobial drug therapy (31). The magnitude of the effect differed by bacteria species and the specifi c antimicrobial drug prescribed. In general, antimicrobial drugs were less effective for reducing colonization with H. infl uenzae than with S. pneumoniae and M. catarrhalis (31).
The effect of age, gender, race, and breastfeeding on colonization differs by population studied (10). Daycare has consistently been associated with increased levels of colonization with S. pneumoniae, H. infl uenzae, and M. catarrhalis (10), as has exposure to other children in the household (32,33). Our study was limited by lack of data on age and number of siblings or other potential confounders such as household crowding and socioeconomic status.
Our study had additional limitations. A cross-sectional study of S. aureus and S. pneumoniae colonization indicated a negative association between PCV7 vaccine serotypes and S. aureus (15). No association was found between S. pneumoniae nonvaccine types and S. aureus. We were unable to examine the association between S. pneumoniae serotype and colonization. Along these lines, we did not have data regarding H. infl uenzae type B vaccination status and did not serotype our H. infl uenzae strains. Therefore, we were also unable to evaluate the effect of this vaccination on polymicrobial colonization.
Nasopharyngeal colonization likely involves a complex combination of factors including host characteristics that infl uence exposure to specifi c bacterial species, host immune responses that may result in killing the bacteria, and direct competitive interactions between bacteria species. In addition to the inhibiting effects of neuraminidase A and hydrogen peroxide already described, competitive interactions between bacteria may also include the secretion of small peptide inhibitors, competition for nutrients, and competition for receptor binding sites. It is also possible that the presence of 1 bacteria species could create a more hospitable niche for another bacteria species. We were unable to evaluate the precise molecular mechanisms that mediate these complex polymicrobial interactions, an important area for future research.
Our study had several strengths, including its longitudinal, prospective design. We examined nasopharyngeal carriage during URI, a time when children are at risk for secondary bacterial infections. In addition, we took advantage of repeated measures analytic techniques to examine microbe-level factors infl uencing bacterial colonization while controlling for host factors.
Results from our study have public health implications. Scientists have debated whether they should seek to eradicate disease by preventing nasopharyngeal colonization (34). Vaccines targeting nasopharyngeal carriage of S. pneumoniae, H. infl uenzae, and M. catarrhalis may be needed to prevent otitis media because simultaneous carriage of these 3 bacteria may increase risk for otitis media (35). Our data indicate that the elimination of nasopharyngeal colonization with bacteria such as S. pneumoniae and H. infl uenzae may increase risk for colonization with S. aureus. Scientists conducting a randomized trial of the effectiveness of pneumococcal vaccines noted an in-crease in S. aureus when spontaneously draining infected middle ears of vaccinated children (25). Factors that may increase the risk of colonization with S. aureus are of special concern given the spread of methicillin-resistant S. aureus (9). Researchers are attempting to develop an S. pneumoniae vaccine containing pneumococcal choline binding protein A, which would protect against sepsis and pneumonia without interfering with pneumococcal colonization (36). Although this type of vaccination strategy may eventually decrease the incidence of potentially fatal invasive pneumococcal disease, it is unlikely to prevent otitis media. Thus, the public health impact of a given intervention strategy may be hard to predict, and caution should be used when designing control strategies that target nasopharyngeal colonization.
The opinions expressed by authors contributing to this journal do not necessarily refl ect the opinions of the Centers for Disease Control and Prevention or the institutions with which the authors are affi liated.
The print journal is available at no charge to public health professionals YES, I would like to receive Emerging Infectious Diseases.
Please print your name and business address in the box and return by fax to 404-639-1954