Association between glucocorticoid administration and outcomes in patients with ARDS based on the MIMIC-III database

Abstract

This study aimed to investigate the association between glucocorticoid administration and outcomes in critically ill patients with ARDS using the Medical Information Mart for Intensive Care (MIMIC)-III database. Data were collected from the MIMIC-III database, which consists of critically ill participants between 2001 and 2012 in the USA. A total of 1831 adult critically ill patients with ARDS were enrolled from the MIMIC-III database. The 60-day and in-hospital mortality, were the primary endpoints. Secondary outcomes included length of stay (LOS) in the hospital and intensive care unit (ICU), 28-day ventilator-free days, ICU mortality, and 28-day mortality. A total of 1831 patients were included in the data analysis. After propensity score (PS) matching, 464 patients diagnosed with ARDS were matched between the glucocorticoid treatment and control groups. Glucocorticoids were associated with increased in-hospital mortality [hazard ratio (HR) 1.32; 95% CI 1.01–1.71; P = .039], longer ICU stay [HR 2.25; 95% CI 0.84–3.65; P = .002], and shorter ventilation-free days at 28 days in all ARDS patients [HR −2.70; 95% CI −4.28–-1.13; P = .001]. The 60-day mortality was higher in the glucocorticoid group (44.83% vs 35.34%; P = .154; HR 1.24; 95% CI 0.93–1.66). Excluding the impact of the glucocorticoid initiation time, from day 15 to day 60, mortality was significantly higher in the glucocorticoid group compared to the non-glucocorticoid group (27.16% vs 12.70%; P < .001; HR 1.75; 95% CI 1.32–2.32). Glucocorticoid administration was associated with worse 60-day and in-hospital survival, longer ICU stay, and shorter ventilator-free days on day 28 in patients with ARDS. Our findings suggest careful consideration of glucocorticoids for ARDS.