FEAST - Fluid Bolus Therapy in Septic African Children

Key Question : 

In a Sub-saharan African setting, does bolus albumin or saline in the resuscitation of febrile children with impaired perfusion improve 48-hour mortality compared with gentle saline resuscitation?

Paper Link

Mortality after fluid bolus in African Children with Severe Infection

The FEAST trial group (Maitland K et al)

New England Journal of Medicine (2011); 364:2483-95

At a glance

• Randomised - Yes
• Controlled - Yes
• Placebo used - No
• Blinded - No, though neuro assessment at 4 weeks was blinded
• Population / Sample size calculation - 3141 patients (terminated early)
• Groups same at randomisation - It appears so
• Groups treated the same after randomisation - Yes, with minor exceptions
• Intention to treat analysis - Yes
• Follow-up - Primary endpoint of 48-hour mortality was complete
• Significant primary endpoint - Yes, absolute risk of death at 48 hours was 7.3% in control group versus 10.5% and 10.6% in saline bolus and albumin bolus groups
• Precision of results - p=<0.01
• Harms analysed - Limited analysis, though these are addressed in a separate paper

Summary

This is an important trial in many ways, but mostly in that it demonstrates the way that some of the foundations of our practice are established with only limited data to support them.

Bolus fluid therapy in shocked patients is one such foundation.  We all do it - a patient arrives in the emergency department with a blood pressure of 60mmHg systolic, and the first thing we do is stick in an IV cannula and give them a litre or two of our finest crystalloid.

But what evidence is there to support this?  Almost none when you start looking for it.

Despite this lack of evidence, bolus fluid therapy is part of the resuscitation algorithms issues by countless authoritative bodies around the globe.

The research group in Kalifi, Kenya, led by prominent researcher Kath Maitland, challenged this belief at the time the world renowned Advanced Paediatric Life Support course was proliferating throughout the continent.  Concerned that this approach was problematic in the 3rd world environment, and noting that there was almost no evidence to support the practice, they sought to conduct a randomised controlled trial comparing 3 arms - albumin bolus therapy, saline bolus therapy, and a more gentle infusion based resuscitation approach with saline.

The study has been lauded as a fantastic achievement, particularly in that it was produced in a relatively austere environment.

The study demonstrated an increased mortality in both bolus arms compared with the control infusion arm.  This result stunned the investigators, who fervently believed that the bolus arm patients were responding to the therapy.

Many reviewers have questioned the applicability to modern intensive care practice in first world countries.  Rightly, they point out that these children were not supported with ventilation, inotropic support or dialytic therapies.  A large proportion were probably suffering cerebral malaria, and the use of high volumes of fluid may well have worsened outcomes by exacerbating cerebral oedema.

So what might have prompted this increase in mortality?  A post-hoc review of the data recorded during the study suggests that the increased mortality was related to irreversible cardiovascular shock.  There are a number of theories, but it appears a reperfusion type injury may be at play.

However, the study does serve to highlight the lack of evidence for bolus fluid therapy, and at the very least should prompt the research community to study this area further.

---

Additional Resources

Podcasts

• Podcast  of my interview with lead author Kath Maitland

Links

• The FEAST-TRIAL website
• Controlled trials website listing for FEAST
• Paper describing the principles for termination of the trial

Reviews and Editorials Evidence-Based Wiki Journal Club JICS editorial Life in the fast lane f1000 review Related Paper Exploration of mechanisms of excess mortality