Hydrocortisone plus Fludrocortisone for Adults with Septic Shock

Annane. NEJM 2018; 378: 809-818. DOI: 10.1056/NEJMoa1705716

Clinical Question

• In critically ill patients with septic shock, does the combination of hydrocortisone plus fludrocortisone therapy reduce 90 day mortality?

Background

• The use of steroids in critically ill patients continues to be controversial. Whilst there are signals for improved cardiovascular parameters, this did not translate to clear mortality benefits.
• The most recent of these trials (ADRENAL, 2018), concluded that among patients with septic shock undergoing mechanical ventilation, a continuous infusion of hydrocortisone did not result in lower 90-day mortality than placebo.

Design

• Multi-centre
• Randomised placebo controlled trial
• Double blinded
• Initially designed to have 4 parallel groups to evaluate benefits/risks of steroids and drotrecogin alfa (DAA) in a 2 by 2 factorial design
  • Original 4 groups
    • Group 1: Corticosteroids placebo + DAA placebo
    • Group 2: Corticosteroids + DAA placebo
    • Group 3: Corticosteroids placebo + DAA
    • Group 4: Corticosteroids + DAA
  • When drotrecogin alfa was withdrawn from the market, the trial was continued with 2 parallel groups
    • Group 1 and 3 combined
    • Group 2 and 4 combined
• Randomised by permuted blocks of 8
• Intention to treat analysis
• Sample size calculation
  • Based on anticipated 90 day mortality of 45% among pts with septic shock
  • 320 pts required in each of the original 4 groups (1280 pts total) for 95% power to detect an absolute difference of 10% in 90 day mortality (α=0.05)

Setting

• 34 French ICUs
• September 2008 – June 2015
  • Trial suspended twice: October 2011 – May 2012 (DAA withdrawal) and July 2014 – October 2014 (request of data and safety monitoring board to check the quality of the trial agents and the distribution of the serious adverse events)
• 1671 patients screened; 1241 randomised into 4 groups

Population

• Inclusion: Intensive care patients with indisputable or probable septic shock for less than 24 hours
  • Septic shocked defined as clinically or microbiologically documented infection
  • SOFA score of 3 or 4 for at least 2 organs and at least 6 hours in duration
  • Vasopressor therapy for at least 6 hours to maintain a systolic blood pressure of at least 90mmHg or a mean blood pressure of 65mmHg
• Exclusion:
  • Presence of septic shock for more than 24 hours
  • High risk of bleeding
  • Pregnancy or lactation
  • Underlying condition which would limit short-term survival
  • Known hypersensitivity to drotrecogin alfa (later removed)
  • Previous treatment with corticosteroids
• Baseline characteristics were similar: intervention vs placebo group
  • Mean age (years): 66 vs 66
  • Male sex (%): 65.5 vs 67.7
  • Medical admission (%): 82.4 vs 81
  • SOFA score: 12 vs 11
  • Site of infection (%)
    • Unknown: 1.8 vs 2.9
    • Lung: 60.7 vs 58.0
    • Abdomen: 12.1 vs 10.9
    • Urinary tract: 16.6 vs 18.8
  • Positive blood culture (%): 36.6 vs 36.6
  • Vasopressor administration
    • Epinephrine
      • Number of pts: 53 vs 58
      • Dose (mcg/kg/min): 2.31 vs 1.74
    • Norepinephrine
      • Number of pts: 534 vs 552
      • Dose (mcg/kg/min): 1.02 vs 1.14
  • Mechanical ventilation (%): 92.3 vs 91.3
  • RRT (%): 27.0 vs 28.1
  • % of pts who received DAA: 17.1 (105/614) vs 16.4 (103/627)

Intervention

• Hydrocortisone
  • 50mg IV bolus every 6 hours
• Fludrocortisone
  • 50 mcg tablet once in the morning
• Administered for 7 days without tapering

Control

• Placebo
  • Similar in appearance and manufactured for the trial

Management common to both groups

• Before randomisation, plasma total cortisol levels measured before, 30 and 60 minutes after IV bolus of 250 mcg of corticotrophin (Synacthen)
• Other interventions were harmonised across centres according to 2008 Surviving Sepsis Campaign guidelines
• National guidelines for the prevention of superinfection were followed

Outcome

• Primary outcome: Significant reduction in 90 day mortality in intervention compared to control group
  • Intervention group: 264 of 614 (43.0%) patients had died
  • Control group: 308 of 627 (49.1%) patients had died
  • Relative Risk (RR): 0.88 (95% CI 0.78 to 0.99; P=0.03)
  • Absolute Risk Reduction (ARR): 6.1% (95% CI 0.6% to 11.7%; P=0.03)
  • Number Needed to Treat (NNT): 17
  • Fragility Index (FI): 3
• Secondary outcome: Intervention vs control group
  • Significantly in favour of intervention group
    • All cause mortality at ICU discharge
      • 35% vs 41% (RR 0.86; 95% CI 0.75–0.99; P=0.04)
    • All cause mortality at hospital discharge
      • 39% vs 45% (RR 0.86; 95% CI 0.76–0.98; P=0.02)
    • All cause mortality at 180 days
      • 47% vs 52% (RR 0.89; 95% CI 0.79–0.99; P=0.04)
    • No of days that pts were alive and free from vasopressors up to 28 days
      • mean 17+/-11 vs 15+/-11, P<0.001
    • Organ-failure-free days up to day 28
      • mean 14+/-11 vs 12+/-11, P=0.003
    • % of pts weaned from vasopressors at 28 days
      • P<0.001
    • % of pts weaned from mechanical ventilation at 28 days
      • P=0.006
    • % of pts with SOFA score below 6 at day 28
      • P<0.001
  • No significant difference between groups
    • All cause mortality at 28 days
      • 34% vs 39% (RR 0.87; 95% CI 0.75–1.01; P=0.06)
    • % of pts from whom care was withheld or withdrawn
      • 10.4% vs 9.7%, P=0.69
    • No of days that pts were alive and free from mechanical ventilation up to 28 days
      • mean 11+/-11 vs 10+/-11, P=0.07
    • Safety outcomes/Incidence of serious adverse events
      • 53.1% vs 58%, P=0.08

Authors’ Conclusions

• In critically ill patients with septic shock, the addition of hydrocortisone and fludrocortisone compared to placebo was associated with a significant improvement in mortality at 90 days.

Strengths

• Randomised controlled trial
• Multi-centre
• Intention to treat analysis
• Appropriate primary outcome
• Appropriateness of antibiotic therapy recorded

Weaknesses

• The trial was initially designed and powered with DAA being part of the therapy. The withdrawal of DAA has impacted aspects of this trial including statistical power calculation
  • Statistical analysis published in supplementary material suggest no interaction with DAA and other treatments
  • The Fragility Index for several of the outcomes (including primary outcome) in favour of intervention is in single figures
• Outcome of Synacthen test conducted pre-randomisation was not mentioned or discussed in main paper
  • Results discussed in supplementary material section where there was no difference between responders and non-responders in those from which the Synacthen test was actually conducted
• The trial was conducted using the Surviving Sepsis Guidelines from 2008 which has since been updated
• Very sick patient population – the high doses of vasopressors used in the trial population may limit external validity
• Not all secondary endpoints included in original trial protocol reported on in final manuscript

The Bottom Line

• The addition of fludrocortisone and its effect is less well investigated compared to hydrocortisone by itself. It is not my current practice to administer this drug for refractory septic shock
• Primary and secondary outcomes, including safety profile, shows a trend in favour of the corticosteroid group (consistent with the findings of the ADRENAL trial)
• I will continue my current practice of using hydrocortisone IV (6 hourly) for refractory septic shock

External Links

• [article] Hydrocortisone plus Fludrocortisone for Adults with Septic Shock
• [further reading] ADRENAL trial review
• [further reading] TBL Steroid Review

Metadata

Summary author: Adrian Wong
Summary date: 15 March 2018
Peer-review editor: Segun Olusanya