SEPSISPAM

SEPSISPAM: High versus low Blood-Pressure Target in Patients with Septic Shock

Asfar. NEJM. 2014;370:1583-1593

Clinical Question

  • In patients with septic shock does a high target mean arterial pressure (MAP), compared to a low target MAP, improve mortality?

Design

  • Randomised controlled trial
  • Stratified according to presence of chronic hypertension
  • Open label
  • Blinding of patients and research staff

Setting

  • 29 centres in France
  • March 2010 – December 2011

Population

  • Inclusion: adult patients with septic shock that was refractory to fluid resuscitation, and had been treated with >0.1mcg/kg adrenaline or noradrenaline for <6 hours at inclusion
    • Refractory to fluid resuscitation defined as lack of response to 30ml/kg of normal saline/colloids or clinician assessment based on invasive measures or echocardiography
  • Exclusion: Pregnancy, decision not to resuscitate
  • 776 patients enrolled

Intervention

  • target MAP 80-85
    • for maximum of 5 days or until weaned from vasopressor support
    • observed values were generally MAP 85-90

Control

  • target MAP 65-70
    • observed values generally 70-75

In both intervention and control group, fluid resuscitation performed as recommended by French intensive care societies; with noradrenaline administered as a 1st line vasopressor in 28 centres, and adrenaline in 1 centre. The use of hydrocortisone and activated protein C was at the discretion of the treating physician. Diuretics prohibited except for ‘compelling indications,’ such as hypoxaemia secondary to sodium and water overload, or life-threatening hyperkalaemia.

Outcome

  • Primary outcome: mortality at 28 days – no significant difference
    • 36.6% in high MAP group vs. 34% in low MAP group (hazard ratio 1.07, 95% C.I. 0.84-1.38, P=0.57)
  • Secondary outcomes:
    • serious adverse events – no significant difference
      • 19.1% vs. 17.8%, P=0.64
    • new atrial fibrillation – significantly greater in high MAP group
      • 6.7% vs. 2.8%, P=0.02
    • 90 day mortality – no significant difference
      • 43.7% vs. 42.3%
  • Sub-group analysis in patients with chronic hypertension (n=340)
    • doubling of serum creatinine – significantly lower in the high MAP group
      • 38.9% vs 52%, P=0.02, NNT 7.6
    • number that required renal replacement therapy from day 1-7 – significantly lower in the high MAP group
      • 31.7% vs. 42.2%, P=0.046, NNT 9.5

Authors’ Conclusions

  • Targeting a MAP of 80-85, compared with 65-70 did not affect mortality

Strengths

  • Randomised
  • Blinding of research staff
  • Standard protocol for use of renal replacement therapy
  • No significant difference in fluid intake between intervention and control group
  • Clearly defined thresholds for renal replacement therapy

Weaknesses

  • Due to a lower mortality than predicted the study was under powered
  • Frequent use of steroids and activated protein C may limit generalisability
  • Achieved MAP was higher than target MAP, although this is common practice in many ICUs
  • Less than 15 patients recruited per centre per year. For a common condition this is surprisingly low and may limit the external validity of the trial.
  • The authors report 16.5% of patients in the high MAP group vs. 10.3% (P=0.01) in the low MAP group failed to achieve target BP because the attending clinician decided to limit the vasopressor infusion. Whilst this is an ethically important safety factor, this compliance bias will favour the null hypothesis.

The Bottom Line

  • For the majority of patients in septic shock a target MAP of 65-70 is a good starting point. However, in patients with chronic hypertension I will target a higher MAP. This is because, even though this did not improve mortality, it did reduce the need for renal replacement therapy with a NNT of 9.5. A number of studies have previously demonstrated that the need for renal replacement therapy is associated with mortality.
  • In patients without hypertension further studies comparing a MAP of 65 with a lower MAP of e.g. 55 would be beneficial

Links

Full text only available with subscription / abstract / doi: 10.1056/NEJMoa1312173

Editorial, Commentaries or Blogs

Metadata

Summary author: @DavidSlessor
Summary date: 16 June 2014
Peer-review editor: @DuncanChambler

One comment

Leave a Reply