Influenza-induced acute respiratory distress syndrome during the 2010-2016 seasons: bacterial co-infections and outcomes by virus type and subtype.

OBJECTIVES

We aimed to describe bacterial co-infections and acute respiratory distress (ARDS) outcomes according to influenza type and subtype.

METHODS

A retrospective observational study was conducted from 2012 to 2016 in patients admitted to the respiratory intensive care unit (ICU) of Marseille university hospital for influenza-induced ARDS. Microbiological investigations, including multiplex molecular respiratory panel testing and conventional bacteriological cultures, were performed as part of the routine ICU care on the bronchoalveloar lavage collected at admission. Bacterial co-infections, ICU mortality and respiratory function were investigated according to virus type and subtype.

RESULTS

Among the 45 ARDS patients included, A(H1N1)pdm09 was the most frequent influenza virus identified (28/45 A(H1N1)pdm09, eight out of 45 A(H3N2) and nine out of 45 influenza B). Bacterial co-infections involving a total of 23 bacteria were diagnosed in 16/45 patients (36%). A(H1N1)pdm09 patients presented fewer bacterial co-infections (17.9% vs. 50.0% for A(H3N2) patients and 77.8% for B patients; p < 0.01). Overall, mortality at 90 days post admission was 33.3% (15/45), and there was no significant difference between influenza type and subtype. The need for extracorporeal membrane oxygenation was more frequent for A(H1N1)pdm2009 (20/28, 71.4%) and B patients (7/9, 77.8%) than the A(H3N2) subtype (1/8, 12.5%; p < 0.01). A(H1N1)pdm09-ARDS patients were associated with fewer ventilation-free days at day 28 (median (IQR): 0 (0-8) days) compared with other influenza-ARDS patients (15 (0-25) days, p < 0.05).

DISCUSSION

In a population of influenza-induced ARDS, A(H1N1)pdm09 was associated with fewer bacterial co-infections but poorer respiratory outcomes. These data underline the major role of A(H1N1)pdm09 subtype on influenza disease severity.