THE MISSING LINK BETWEEN BACTERIAL CARRIAGE STRUCTURE AND PNEUMONIA
Although there is substantial evidence that viral infection influences the URT bacterial community, whether these changes are reflected in the LRT and ultimately in pneumonia etiology is unclear, which weakens our first hypothesis (i.e. respiratory viruses can influence the etiology of pneumonia by altering bacterial carriage structure in the URT). Studies that examine the joint effects of viral co-infection, bacterial carriage and bacterial pneumonia would provide one strategy for filling this gap. However, we found only two such studies. In a South African hospital-based surveillance study of severe acute respiratory illness, 969 nasopharyngeal-oropharyngeal specimens were tested for S. pneumoniae and a panel of respiratory viruses. A high pneumococcal colonization density in the nasopharynx and oropharynx was associated with both respiratory virus co-infection and pneumococcal pneumonia [86]. A second hospital-based case–control study compared nasopharyngeal carriage among 274 radiologically confirmed cases of pneumonia, 276 cases of other LRT infections and 350 controls in Vietnam. Their findings for S. pneumoniae were similar to that of the South African study. However, the investigators also studied H. influenzae and M. catarrhalis and found no clear association between viral co-infection, nasopharyngeal bacterial load and pneumonia for these species [87]. As noted above, M. catarrhalis rarely causes pneumonia, but H. influenzae is second only to S. pneumoniae. Although there appears to be a persuasive argument for a link between viral co-infection, carriage and pneumonia for S. pneumoniae, whether or why the interaction is not true for other URT bacteria needs further exploration. In particular, studies that can directly test whether viral infection led to bacterial colonization or overgrowth by a potential pathogen, which led to bacterial pneumonia by that pathogen, are in order. In conclusion, there is no definitive answer to our first hypothesis. Epidemiologic studies and experiments indicate viruses alter the bacterial community in the URT, but they do not yet adequately address whether these changes in the URT bacterial community play a significant role in pneumonia etiology.