Source: Annane D, Bellissant E, Bollaert PE, Briegel J, Keh D, Kupfer Y. Corticosteroids for treating severe sepsis and septic shock. Cochrane Database of Systematic Reviews 2004, Issue 1. Art. No.: CD002243. DOI: 10.1002/14651858.CD002243.pub2.
Sprung CL, Annane D, Keh D, Moreno R, Singer M, Freivogel K, Weiss YG, Benbenishty J, Kalenka A, Forst H, Laterre PF, Reinhart K, Cuthbertson BH, Payen D, Briegel J; CORTICUS Study Group. Hydrocortisone therapy for patients with septic shock. N Engl J Med. 2008 Jan 10;358(2):111-24.
Efficacy Endpoints: 28-day all cause mortality, hospital mortality, ICU mortality
Harm Endpoints: Gastrointestinal bleeding, superinfection, hyperglycemia
Narrative: Sepsis is a common and often fatal condition. A 2001 study found an incidence of 176 per 100,000 in the United States1, while mortality is approximately 60%2. Studies on the biological mechanisms of shock have suggested that corticosteroids may potentially be beneficial.
This review examined the effect of corticosteroid therapy for severe sepsis and septic shock. 'Severe sepsis' meant documented infection with signs of systemic inflammatory response, hypotension, and at least one sign of organ dysfunction. 'Septic shock' included the above, with hypotension refractory to fluid resuscitation and requiring the use of vasopressors. The Cochrane Review included data from 15 trials of 2022 subjects performed between 1955 and 2002, including a variety of treatment protocols in terms of dosage and duration. In addition to the studies included in the Cochrane Review our analysis incorporates the results of the CORTICUS trial, which included 499 patients and examined the use of low dose corticosteroids in patients with septic shock. This trial was published in January 2008, after the most recent Cochrane Review was published.
Corticosteroids yielded no patient-oriented benefit. There was no statistically significant difference in mortality between the control group and those who received corticosteroids. This was true regardless of the dosing regimen or duration of treatment. Likewise, corticosteroids were not associated with harm. There was no statistically significant increase in mortality or measured adverse events (superinfection, GI bleeding, hyperglycemia).
Caveats: The Cochrane Review includes a subgroup analysis of 5 studies involving 465 subjects that tested 'low dose' corticosteroids (< 300mg hydrocortisone or equivalent) used over a 'long course' (> 5 days). In this subgroup there were statistically significant improvements in 28-day all cause mortality (ARR 9%, NNT = 11), ICU mortality (ARR 10.7%, NNT = 9) and hospital mortality (ARR 10.8%, NNT = 9) with corticosteroids. On the basis of this analysis the Cochrane Review concluded sufficient evidence to recommend low dose, long course corticosteroid therapy for septic shock. The trials, however, were underpowered with 4 of 5 involving less than 45 subjects each. The Cochrane Review also did not include results from the more recent CORTICUS trial of low dose, long course therapy in 499 subjects. The CORTICUS trial found no statistically significant benefit in patients who received corticosteroids and no improvements in mortality outcomes. When we incorporate the CORTICUS trial data into the Cochrane Review subgroup analysis for low dose, long course corticosteroids, using the same fixed effect model to calculate for a weighted treatment effect, we find no statistically significant benefit in this subgroup. Although the CORTICUS trial includes more patients than the sum total of the 5 studies in the Cochrane analysis it was also underpowered. While the study intended to enroll 800 participants loss of funding and slow enrollment led to early stoppage. It remains conceivable that a large, well-designed, randomized trial may detect a benefit. However, current evidence does not support the use of corticosteroids, regardless of treatment regimen, in the treatment of septic shock.
Author: Dan Runde, MD
Published/Updated: September 7, 2010
The title bar is color-coded with our overall recommendation.
If you have suggestions, requests, or questions about a particular NNT review, please send us a message and we’ll try to address it as soon as possible.