Reversal of late septic shock with supraphysiologic doses of hydrocortisone

Abstract

Objectives: Preliminary studies have suggested that low doses of corticosteroids might rapidly improve hemodynamics in late septic shock treated with catecholamines. We examined the effect of hydrocortisone on shock reversal, hemodynamics, and survival in this particular setting.

Design: Prospective, randomized, double-blind, placebo-controlled study.

Setting: Two intensive care units of a University hospital.

Patients: Forty-one patients with septic shock requiring catecholamine for >48 hrs.

Interventions: Patients were randomly assigned either hydrocortisone (100 mg i.v. three times daily for 5 days) or matching placebo.

Measurements and main results: Reversal of shock was defined by a stable systolic arterial pressure (>90 mm Hg) for > or =24 hrs without catecholamine or fluid infusion. Of the 22 hydrocortisone-treated patients and 19 placebo-treated patients, 15 (68%) and 4 (21%) achieved 7-day shock reversal, respectively, a difference of 47% (95% confidence interval 17% to 77%; p = .007). Serial invasive hemodynamic measurements for 5 days did not show significant differences between both groups. At 28-day follow-up, reversal of shock was higher in the hydrocortisone group (p = .005). Crude 28-day mortality was 7 (32%) of 22 treated patients and 12 (63%) of 19 placebo patients, a difference of 31% (95% confidence interval 1% to 61%; p = .091). Shock reversal within 7 days after the onset of corticosteroid therapy was a very strong predictor of survival. There were no significant differences in outcome in responders and nonresponders to a short corticotropin test. The respective rates of gastrointestinal bleeding and secondary infections did not differ between both groups.

Conclusions: Administration of modest doses of hydrocortisone in the setting of pressor-dependent septic shock for a mean of >96 hrs resulted in a significant improvement in hemodynamics and a beneficial effect on survival. These beneficial effects do not appear related to adrenocortical insufficiency.