TEMPORAL ASSOCIATIONS BETWEEN VIRUSES AND BACTERIA
The 1918 Spanish flu pandemic resulted in ∼50 million deaths worldwide: most of the deaths were caused by secondary bacterial pneumonia [53, 54]. During the 2009 H1N1 pandemic, bacterial co-infection was detected in 18–34% of influenza cases (reviewed by Chertow and Memoli [8]) with vulnerability peaking ∼1 week after influenza infection [55]. The association of viral infection and bacterial pneumonia is not limited to influenza although that interaction has been most studied: adenovirus, human metapneumovirus, respiratory syncytial virus (RSV), and other viruses have been temporally associated with an increased risk of pneumococcal pneumonia and invasive pneumococcal disease (IPD), defined as the isolation of S. pneumoniae from a normally sterile site, in the USA (Table 2) [56–67]. The majority of US studies suggest strong associations between S. pneumoniae infections (both pneumonia and IPD) and influenza virus and RSV, with potential effect modification by age. Temporal associations with other viruses are less supported and limited to IPD. We did not find any studies in the USA that examined temporal associations between viruses and bacterial species other than S. pneumoniae. Six studies conducted in other developed countries examined temporal associations between respiratory viruses and IPD [68–73]. Three out of five studies that examined influenza virus found associations with IPD in UK, The Netherlands and Sweden [70, 71, 73]. Among four studies that examined RSV in other countries, two indicated associations with IPD in all age groups [70, 71], one found an association only among children [68] and the last observed an association only in individuals 2 years or older [69].
Table 2  Temporal associations between respiratory viruses and S. pneumoniae, the USA   Abbreviations: ADV (adenovirus), hMPV (human metapneumovirus), IV (influenza virus) and PCV (picornavirus)   Temporal associations provide evidence of virus–bacterial interactions, but do not necessarily prove these interactions exist. Many viral infections are seasonal, as is pneumonia infection, so the temporal associations may merely reflect the influence of other seasonal phenomena, environmental or host, that are shared by both viral infection and pneumonia [74]. However, evidence for true virus–bacterial interactions is supported by population studies that estimate a high prevalence of viral co-infection during pneumonia [5, 6] and animal models that suggest increased susceptibility to pneumonia and increased disease severity during viral co-infection [75]. In the USA, ∼47% of children and 19% of adults with bacterial pneumonia are co-infected with one or more viruses [5, 6]. Further, vaccination for S. pneumoniae reduced pneumonia associated with RSV, influenza A and parainfluenza (PIV) types 1-3 [76]. Influenza vaccine probe studies may provide additional insight to the burden of influenza co-infection on bacterial pneumonia.