Corticosteroid Treatment in Critically Ill Patients with Pandemic Influenza A/H1N1 2009 Infection: Analytic Strategy Using Propensity Scores

Rationale Administration of adjuvant corticosteroids to patients with pandemic influenza A/H1N1 2009 (pH1N1) may reduce inflammation and improve outcomes. Objectives To assess the effect of adjuvant corticosteroid treatment on the outcome of critically ill patients with pH1N1 infection. Methods All adult patients with confirmed pH1N1 admitted to the intensive care unit of 28 hospitals in South Korea from September 2009 to February 2010 were enrolled. Patients with and without adjuvant corticosteroid treatment were retrospectively compared by two risk stratification models: (1) a retrospective cohort study that used propensity score analysis to adjust for confounding by treatment assignment and (2) a propensity-matched case–control study. Measurements and Main Results A total of 245 patients were enrolled in the cohort study, 107 of whom (44%) received adjuvant steroid treatment. In the cohort study, the 90-day mortality rate of patients given steroids (58%, 62 of 107) was significantly higher than that of those not given steroids (27%, 37 of 138) (P < 0.001). The steroid group was more likely to have superinfection such as secondary bacterial pneumonia or invasive fungal infection, and had more prolonged intensive care unit stays than the no-steroid group. Multivariate analysis indicated that steroid treatment was associated with increased 90-day mortality when independent predictors for 90-day mortality and propensity score were considered (adjusted odds ratio, 2.20; 95% confidence interval, 1.03–4.71). In the case–control study, the 90-day mortality rate in the steroid group was 54% (35 of 65) and 31% (20 of 65) in the no-steroid group (McNemar test, P = 0.004). Conclusions Adjuvant corticosteroids were significantly associated with higher mortality in critically ill patients with pH1N1 infection.