Impact of Corticosteroids on Short- and Long-Term Mortality in Patients with Sepsis and Septic Shock: A Meta-Analysis with Trial Sequential Analysis

Abstract

Background Findings on the benefits and harms of corticosteroids for sepsis and septic shock from published meta-analyses are conflicting. With new randomized evidence emerging, an updated analysis of best available evidence on the effects of corticosteroids in sepsis and septic shock is warranted. Methods We did a conventional and cumulative meta-analysis and trial sequential analysis. Ovid MEDLINE, Ovid Embase, the Cochrane Central Register of Controlled Trials (CENTRAL) and LILACS were searched until May 2, 2018. Two reviewers selected randomized controlled trials (RCTs) reporting benefits and harms of corticosteroids as compared to placebo, standard therapy or no steroids in patients with sepsis and septic shock and independently collected data of interest. Primary outcomes were short-term (14-day and 28-day) mortality and long-term (90-day, 180-day, 1-year) mortality. Secondary outcomes included intensive care unit (ICU)- and hospital mortality, shock reversal by day 7 and 28, length of ICU and hospital stay, sequential organ failure assessment (SOFA) score at day 7 and adverse events. Fixed-effect and random-effects models were used to provide pooled estimates on predefined outcomes. This study is registered with PROSPERO, number CRD 42018092849. Findings Thirty-four RCTs including 9,942 participants were included in this meta-analysis. Results showed that long course of low-dose corticosteroids could improve 28-day mortality (relative ratio (RR), 0.91; 95% confidence interval (CI), 0.86 to 0.97; I2=0%; high-quality evidence), ICU mortality (RR, 0.87; 95%CI, 0.79 to 0.95; I2=21%; moderate-quality evidence) and hospital mortality (RR, 0.89; 95%CI, 0.79 to 0.99; I2=40%; low-quality evidence). Meta-regression analyses confirmed the benefits of corticosteroids in14-day and 28-day mortality and indicated that the use of hydrocortisone at least 135mg per day for at least 87h was associated with significant benefits. However, this meta-analysis found no benefits for 90-day mortality from low-quality evidence (RR, 0.96; 95%CI, 0.85 to 1.09; I2=44%), 180-day mortality and 1-year mortality and about 38,340 participants were required for clarification of the effect of 90-day mortality. In terms of adverse events, long course of low-dose corticosteroids were associated with hyperglycaemia (RR, 1.20; 95%CI, 1. 09 to 1.32; I2=40%) and hypernatraemia (RR, 1.60; 95%CI, 1.26 to 2.02; I2=0%). Short course of high-dose corticosteroids showed no significant benefits in any of the predefined outcomes. Interpretation Long course (at least 87h) of low-dose (at least 135 mg per day) hydrocortisone could improve 28-day mortality of patients with sepsis and septic shock as well as ICU- and hospital mortality, with varied quality of evidence. However, it could not reduce long-term (90-day, 180-day and 1-year) mortality. Funding: None. Declaration of Interest: We declare no competing interests.