2008 NEJM: Steroid 治 septic shock 沒效

Steroids in Septic Shock: Not Looking Good

Mortality rates at 28 days did not differ between steroid and placebo groups, regardless of response to corticotropin testing.

Guidelines for septic shock therapy recommend the use of hydrocortisone largely on the basis of one well-conducted study in which benefit was seen only in patients who did not respond to a corticotropin test (Journal Watch Emergency Medicine Nov 13 2002). In the current multicenter, randomized, controlled, double-blind study, researchers included a broad population of 499 patients with sepsis who had evidence of shock that persisted for more than 1 hour during the 72 hours before enrollment, despite adequate fluid replacement. Patients received either 50 mg of intravenous hydrocortisone or placebo every 6 hours for 5 days, followed by dose tapering over 6 days.

Rates of death at 28 days (the primary outcome) did not differ between treatment groups overall, or for patients who did not have a response to a short corticotropin test, or for patients who did have a response. This finding also held true for those patients who were enrolled within 12 hours after an episode of shock. A second episode of infection (superinfection) was more common in the steroid group, as was hyperglycemia. Reversal of shock was more rapid in the steroid group, but earlier reversal had no effect on outcome. Post hoc analysis revealed significantly higher death rates among patients who received etomidate before randomization than among those who did not (45.1% vs. 31.5% in the steroid group and 40.0% vs. 29.6% in the placebo group).

An editorialist notes that the study was underpowered to allow a definitive statement on the role of steroids in septic shock and recommends a trial involving at least 2600 patients.

Comment: The story of steroids in septic shock is long, and the final chapter remains to be written. Presently, there is insufficient evidence to support steroid administration for sepsis with shock, regardless of a patient’s response to corticotropin testing. The authors cite "an absolute reduction in mortality of 11.2% in the hydrocortisone group" for patients with persistent systolic blood pressure <90 mm Hg 1 day after fluid and vasopressor resuscitation, but their own statistical analysis demonstrates that there was not a significant difference in mortality between the hydrocortisone and placebo groups. Although a post hoc analysis identified a higher mortality rate in patients who were given etomidate as an induction agent, the authors correctly did not make any cause-and-effect statement, likely because of selection bias (clinicians may use etomidate more often in more-unstable patients), conflicting findings from prior studies, and the inability of the study to determine whether etomidate had any real effect (Journal Watch Emergency Medicine Sep 21 2007). This study also delivers a serious blow to the present fascination with corticotropin testing as a guide to treatment for such patients.

— J. Stephen Bohan, MD, MS, FACP, FACEP

Published in Journal Watch Emergency Medicine January 9, 2008

Citation(s):
Sprung CL et al. Hydrocortisone therapy for patients with septic shock. N Engl J Med 2008 Jan 10; 358:111.