Hydrocortisone may not decrease 90-day mortality in adults with septic shock
EBM Focus - Volume 13, Issue 9
- There is limited evidence to support the recommendation to give intravenous hydrocortisone to patients with septic shock who do not respond to fluid resuscitation and vasopressor therapy.
- In the ADRENAL trial, 3,800 adults with septic shock having mechanical ventilation and ≥ 4 hours of vasopressor or inotropic therapy were randomized to adjunct continuous IV infusion of hydrocortisone 200 mg/day vs. placebo for 7 days.
- No significant difference in 90-day mortality was observed between the groups (27.9% vs. 28.8% with placebo, odds ratio 0.95, 95% CI 0.82-1.1), but the wide confidence interval includes the possibility of benefit or harm.
The Surviving Sepsis Campaign International Guidelines suggest considering intravenous hydrocortisone in patients with septic shock who do not respond to fluid resuscitation and vasopressor therapy (Intensive Care Med 2017). A systematic review of 22 trials with almost 2,300 patients with sepsis suggests that a ≥ 3-day course of corticosteroids compared to placebo or supportive treatment may reduce 28-day mortality (Cochrane Database Syst Rev 2015), but the two 2 largest trials in the review have inconsistent results (N Engl J Med 2008, JAMA 2002). In order to further evaluate the effect of hydrocortisone on mortality, the recent ADRENAL trial randomized 3,800 adults (mean age 62 years) with septic shock having mechanical ventilation to adjunct continuous IV infusion of hydrocortisone 200 mg/day vs. placebo for 7 days or until discharge from intensive care. All patients had documented or strong clinical suspicion of infection, met ≥ 2 criteria of systemic inflammatory response syndrome (SIRS), and were treated with vasopressors or inotropic agents for ≥ 4 hours until randomization.
Comparing hydrocortisone vs. placebo, there was no significant difference in 90-day mortality (27.9% vs. 28.8% with placebo, odds ratio 0.95, 95% CI 0.82-1.1) or 28-day mortality. Hydrocortisone therapy was associated with lower rates of blood transfusion (in 37% vs. 41.7%, p = 0.004, NNT 22), and reduced median time to resolution of shock (3 days vs. 4 days, p < 0.001), discharge from the intensive care unit (10 days vs. 12 days, p < 0.001), and cessation of initial mechanical ventilation episode 6 days vs. 7 days (p < 0.001). There were no significant differences in other secondary outcomes. The rates of adverse events were 1.1% with hydrocortisone vs. 0.3% with placebo (p = 0.009, NNH 125). Serious adverse events in the hydrocortisone group included myopathy in 3 patients, and ischemic bowel and circulatory shock in 1 patient each, and in the placebo group, bleeding and abdominal wound dehiscence in 1 patient each. The other adverse events in the hydrocortisone group were mostly minor.
The ADRENAL trial failed to find a mortality benefit with use of adjunct continuous infusion of low-dose hydrocortisone for 7 days in patients with septic shock who received vasopressors or inotropic agents and were on mechanical ventilation. The 95% confidence interval includes the possibility of moderate benefit or harm, and it is possible that there are subgroups of patients for whom corticosteroids may be more beneficial. A subgroup analysis of 229 patients from an earlier trial (JAMA 2002) suggests that corticosteroids might reduce 28-day mortality in patients with septic shock and inadequate adrenal reserve, but the ADRENAL trial did not address this question. There may also be some benefits for other outcomes such as reducing the number of days in the ICU and time to cessation of mechanical ventilation, but this would need to be weighed against the slightly higher rate of adverse events with hydrocortisone. Overall, the findings of the ADRENAL trial, the largest study to date, do not support the use of hydrocortisone to improve mortality in patients with septic shock.