Influenza virus infection: an approach to identify predictors for in-hospital and 90-day mortality from patients in Vienna during the season 2017/18.

Background: Seasonal influenza outbreaks are associated with increased mortality and hospitalisation rates. Herein we tried to identify predictors of mortality in hospitalised patients with influenza virus infection.

Materials/methods: In this exploratory retrospective observational single-centre-study we included all influenza-positive patients older than 18 years who were hospitalised and treated at the flu-isolation-ward during the influenza season 2017/18. Diagnosis was based on point-of-care-test with the Alere™ i. First we performed χ tests and Mann-Whitney U tests to identify predictors of mortality. Significant variables were used in a stepwise-forward-logistic-regression-model to predict in-hospital and 90-day mortality.

Results: Of the 396 patients who tested positive for influenza 96 (24.2%) had influenza A and 300 (75.8%) influenza B. Twenty-two (5.6%) died in hospital and the 90-day mortality rate was 9.4%. In the stepwise logistic regression older age (OR 1.1 per year 95% CI 1.03-1.17), history of atrial fibrillation (OR 5.91 95% CI 1.91-18.34), dementia (OR 3.98 95% CI 1.24-12.78), leucocyte count (OR 1.11 per G/L 95% CI 1.03-1.20), pneumonia (OR 4.39 95% CI 1.44-13.39) and acute heart failure (OR 23.15 95% CI 4.33-123.76) increased the risk of in-hospital mortality. The risk for 90-day mortality was increased by older age (OR 1.04 per year 95% CI 1.01-1.07), history of atrial fibrillation (OR 3.1, 95% CI 1.36-7.05), history of congestive heart failure (OR 4.7 95% CI 1.94-11.48), pneumonia (OR 3.2 95% CI 1.45-6.91) and decreased by statin use (OR 0.28 95% CI 0.10-0.78).

Conclusions: Older age, history of atrial fibrillation and pneumonia are associated with increased risk of influenza-associated in-hospital and 90-day mortality. Statin use may decrease 90-day mortality.