Effect of preadmission glucocorticoid therapy on 30-day mortality in critically ill patients: a retrospective study of a mixed ICU population in a tertiary hospital

BACKGROUND: This study aimed to investigate the association between preadmission glucocorticoid (GC) therapy and 30-day mortality in critically ill patients following admission to an intensive care unit (ICU). We aimed to determine whether this association differed according to daily GC dosage and type. We conducted a retrospective cohort study of adult patients admitted to a single tertiary academic hospital ICU from January 2012 to December 2017. We classified the patients regularly undergoing oral GC therapy as preadmission GC users, and those with no history of GC use were classified as non-GC users. RESULTS: The study included 24,929 patients, of whom 816 (3.3%) were preadmission GC users. Thirty-day mortality in preadmission GC users (173 of 816 patients) was 21.2% compared to 8.8% (2113 of 24,113 patients) in non-GC users. Multivariate Cox regression analysis showed that preadmission GC users had a 1.62-fold increase in 30-day mortality compared to non-GC users [hazard ratio (HR) 1.62, 95% confidence interval (CI) 1.29–2.03, P < 0.001]. When comparing preadmission GC users with diabetes mellitus to non-GC users, a 2.29-fold increase in 30-day mortality was noted (HR 2.29, 95% CI 1.08–4.86, P = 0.031). In the sensitivity analysis, compared to non-GC users, daily dosages of ≤ 5 and > 5 mg of prednisolone in preadmission GC users showed 1.45-fold (HR 1.45, 95% CI 1.03–2.03, P = 0.033) and 1.67-fold (HR 1.67, 95% CI 1.25–2.24, P = 0.001) increases, respectively, in 30-day mortality after ICU admission. Moreover, prednisolone, methylprednisolone, and dexamethasone users in the preadmission GC users group showed 1.56-fold (HR 1.56, 95% CI 1.21–2.01, P = 0.001), 1.90-fold (HR 1.90, 95% CI 1.12–3.25, P = 0.018), and 1.30-fold (HR 1.30, 95% CI 1.05–1.50, P = 0.042) increases, respectively, in 30-day mortality compared to non-GC users. CONCLUSION: Preadmission GC use among critically ill patients was associated with an increased 30-day mortality after ICU admission compared to non-GC use. This association was more prevalent in preadmission GC users with diabetes mellitus and in preadmission GC users who took > 5 mg/day of prednisolone and methylprednisolone. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (10.1186/s13613-019-0489-8) contains supplementary material, which is available to authorized users.