A meta-analysis to evaluate the effects of corticosteroids on the outcomes of patients with septic shock

Qing LI; Jianfeng XIE; Ling LIU

ABSTRACT

Objective:

To evaluate the effect of corticosteroids on the prognosis of patients with septic shock.

Method:

In order to compare administration of corticosteroids with placebo or standard supportive care in adults with septic shock, clinical randomized controlled trials (RCT) were searched and selected, according to inclusion and exclusion criteria. A systemic assessment and meta-analysis was performed using RevMan 5.3.

Result:

A total of 16 RCTs enrolling 6 896 patients were finally included in present analysis. The corticosteroids group included 3 448 patients, and the control group included 3 448 patients. The 28-day mortality in corticosteroids group and control group were 28.6% and 31.2%, respectively ( P=0.16). The 90-day mortality, the mortality in intensive care unit (ICU) and the mortality in the hospital between corticosteroids group and control group were 31.7% vs. 34.0% ( P=0.16), 37.5% vs. 37.5% ( P=0.87), and 41.0% vs. 43.9% ( P=0.35) respectively, which indicated that corticosteroids could not improve the mortality of patients with septic shock. Subgroup analyses showed that hydrocortisone combined with hydrocortisone could reduce the 28-day mortality, and the 28-day mortality in corticosteroids group and control group were 37.7% and 43.3%, respectively ( P=0.02). However, other types of corticosteroids had no influence on 28-day mortality. The incidence of gastrointestinal hemorrhage and super-infections showed no statistical difference in corticosteroids group and control group. However, incidence of hyperglycemia was significantly increased in corticosteroids group, 27.1% vs. 25% ( P<0.000 1).

Conclusion:

Corticosteroids could not improve the mortality of patients with septic shock, and simultaneously, significantly increase incidence of hyperglycemia. Corticosteroids have no influence on the incidence of gastrointestinal hemorrhage and super-infections. Subgroup analyses showed that hydrocortisone combined with hydrocortisone could reduce the 28-day mortality.