CORTICUS: Hydrocortisone Therapy for Patients with Septic Shock

Sprung et al. N Engl J Med 2008; 358:111-24.

Clinical Question

• In adult patients with severe septic shock, does hydrocortisone versus placebo reduce mortality at 28 days?

Design

• Multi-centre, randomised controlled trial
• Double-blinded
• Computer generated, block randomisation stratified by study centre.
• All patients had short corticotropin test to identify “responders” from “non-responders” for a priori sub-group analysis.
• Adverse events data collected for safety analysis at blinded interim intervals.
• Power calculation: if 40% included were “non-responders” with a baseline 28-day mortality of 50%, then 400 patients per arm were required to detect 10% absolute reduction in mortality with 80% power and 5% accepted type 1 error.

Setting

• 52 participating ICUs
• March 2002 to November 2005

Population

• Inclusion: adults with severe sepsis, with all of the following:
  • clinical evidence of infection;
  • systemic response to inflammation;
  • shock (systolic BP < 90 mmHg despite 1 hour fluid resuscitation or the need for vasopressors);
  • organ dysfunction attributable to shock.
• Exclusion: chronic steroids within 6 months; acute steroids within 4 weeks; immunosupression;  less than 24 survival expectation
• 499 patients recruited

Intervention

• Hydrocortisone given as 50 mg in 6-hourly boluses
  • Tapering regime from day 6. Stopped on day 12.

Control

• Identical placebo prepared and administered in an identical way to intervention.

Outcome

• Primary outcome: no difference in 28-day mortality in short corticotropin non-responders (i.e. the sub-group that were more likely to benefit from exogenous corticosteroids)
  • 39.2% in hydrocortisone group vs. 36.1% in placebo group
  • Absolute difference 3.1% (95% C.I. -9.5% to 15.7%) favouring placebo.
  • p=0.69
• Secondary outcome:
  • No difference in 28-day mortality in short corticotropin responders (i.e. the sub-group that were less likely to benefit from exogenous corticosteroids)
    • 28.8% vs. 28.7%, P=1.0
  • No difference in 28-day mortality in all patients
    • 34.3% vs. 31.5%, P=0.51
  • Statistically significant reduction in the time to reversal of shock favouring the hydrocortisone group (seen in responders, non-responders and all patient groups).
    • median time until reversal of shock, in all patients, was 3.3 days (95% C.I. 2.9-3.9) in the hydrocortisone group vs. 5.8 days (95% C.I. 5.2-6.9) in the placebo group
  • Non-significant increase in rate of superinfections in hydrocortisone vs. placebo group
    • 33% vs. 26%, RR 1.27 (95% C.I. 0.96-1.68)

Authors’ Conclusions

• The use of hydrocortisone did not decrease mortality in a general population of patients with septic shock, even though the drug hastened reversal of shock

Strengths

• Excellent randomisation and blinding methodology
• Sensible primary outcome, and useful secondard outcomes
• Fair commentary of problems by authors in ‘Discussion’ section of paper

Weaknesses

• Despite 52 ICUs, this study failed to recruit required numbers (average of less than 10 patients per ICU over 3 years) despite common incidence of severe sepsis.
• Actually mortality (~38%) was much lower than estimated mortality used for power calculation (50%)