Influenza pneumonia surveillance among hospitalized adults may underestimate the burden of severe influenza disease.

Studies seeking to estimate the burden of influenza among hospitalized adults often use case definitions that require presence of pneumonia. The goal of this study was to assess the extent to which restricting influenza testing to adults hospitalized with pneumonia could underestimate the total burden of hospitalized influenza disease.We conducted a modelling study using the complete State Inpatient Databases from Arizona, California, and Washington and regional influenza surveillance data acquired from CDC from January 2003 through March 2009. The exposures of interest were positive laboratory tests for influenza A (H1N1), influenza A (H3N2), and influenza B from two contiguous US Federal Regions encompassing the study area. We identified the two outcomes of interest by ICD-9-CM code: respiratory and circulatory hospitalizations, as well as critical illness hospitalizations (acute respiratory failure, severe sepsis, and in-hospital death). We linked the hospitalization datasets with the virus surveillance datasets by geographic region and month of hospitalization. We used negative binomial regression models to estimate the number of influenza-associated events for the outcomes of interest. We sub-categorized these events to include all outcomes with or without pneumonia diagnosis codes.We estimated that there were 80,834 (95% CI 29,214-174,033) influenza-associated respiratory and circulatory hospitalizations and 26,760 (95% CI 14,541-47,464) influenza-associated critical illness hospitalizations. When a pneumonia diagnosis was excluded, the estimated number of influenza-associated respiratory and circulatory hospitalizations was 24,816 (95% CI 6,342-92,624). The estimated number of influenza-associated critical illness hospitalizations was 8,213 (95% CI 3,764-20,799). Around 30% of both influenza-associated respiratory and circulatory hospitalizations, as well as influenza-associated critical illness hospitalizations did not have pneumonia diagnosis codes.Surveillance studies which only consider hospitalizations that include a diagnosis of pneumonia may underestimate the total burden of influenza hospitalizations.