The FEAST (Fluid Expansion As Supportive Therapy) Study, is a landmark study of fluid management in sepsis conducted with high quality in difficult conditions of Africa. It was referred to in our FOAM Eye Catchers 4: Time for a Rethink on IV Fluids in Sepsis.
The study demonstrated that in sick, febrile, septic children in Africa, giving a 20ml/kg fluid bolus increased mortality compared to just providing maintenance fluids, which has fundamentally challenged our current approach to sepsis care. The 20ml/kg dose is in keeping with international guidelines in fluid resuscitation.
Of note the study excluded the pure hypovolaemic condition of gastroenteritis as the intent was clearly to study the effect of fluids on septic shock.
An excellent summary of the details of the initial FEAST trial is provided by Gerard Fennessy on LITFL. In addition Colin Parker‘s team at empem reviewed FEAST and other fluid controversies in a great podcast.
Earlier this year a further analysis of the FEAST data was released by the authors accompanied by an associated analysis by Myburgh and Finfer, which has revealed further insights and refuted some of the claimed limitations of the study.
- That just over half the patients had malaria and about a third had anaemia, conditions where ADH may be raised and haemodilution unhelpful, resulting in potentially more harmful effects of fluid administration. This limits the applicability of the findings to patients without these conditions, more often seen in the 1st world. However further analysis showed that the harm of boluses occurred in patients with and without malaria or anaemia.
- 18 months into the 2 year trial they changed the protocol to double the fluid boluses in the treatment arm. However analysis showed fluid bolus harm before and after the protocol change.
- The clinical definition of shock in the study resulting in inclusion was based on the finding of any single clinical sign of out of 4 WHO criteria rather than using them collectively. As each sign individual has low specificity for shock, many of the patients may not have actually had shock. However when classified into “severe shock or acidosis” (based on any one of “lactate > 5, base excess <-8, moderate hypotension, or presence of ALL 4 WHO clinical indicators of shock), this group had the greatest mortality increase from boluses (n=698, bolus=10%, control 3%). Further I would argue that fluid resuscitation in the 1st world frequently occurs in patients without the full gamut of clinical or investigative features of shock so the practice in FEAST may actually accord with 1st world practice.
- These African centres were without access to 1st world intensive care facilities where the mortality difference may not be seen. This limitation is obviously can not be refuted. However while it is possible that 1st world intensive care may have been able to reduce or mitigate some of the harmful effects of fluid boluses, if anything this only serves to illustrate how valuable the FEAST data is. An equivalent 1st world data set may have masked the harmful effects of IV fluid boluses with our use of other intensive care therapies such as vasopressors, non invasive and invasive ventilation. The fact that we have the facilities and drugs to bail us out after we have harmed our patients should not stop us from considering a change in our practice that may prevent this harm from occurring in the first place. It is quite possible fluid boluses, result in greater use of intensive care such as higher doses of vasopressors, ventilation etc and adjusting our approach could be beneficial, though this is unproven.
- Despite an expected increase in short term haemodynamics due to fluid boluses, mortality was still increased and this increased mortality was seen in both those whose haemodynamics improved and those who did not.
- The authors noted: “Remarkably, in every subgroup we examined, there was consistent evidence of harm by boluses”.
- The Terminal Clinical Event that was increased primarily by fluid boluses was “cardiovascular collapse” not fluid overload conditions. Myburgh & Finfer theorise (as do the authors) that: “although there was evidence of improved short-term hemodynamic effects with boluses in these patients, bolus administration may have resulted in a rapid reduction in sympathetically mediated compensatory mechanisms, leading to cardiovascular dysfunction and death.” Further the authors suggest: “Alternatively, this may indicate lethal reperfusion injury or a surge of cytokines in cases with advanced shock. As previously suggested, shock may be an adaptive, time-dependent response sustaining children through a prolonged period prior to hospital admission – only to die within hours of reperfusion”
However an important alternative theory that the authors and Myburgh/Finfer discuss is that the normal saline or albumin solutions (in saline carrier solutions) used in the treatment arm have high chloride content that may worsen acidosis and contribute to cardiovascular collapse. It is possible the harm observed may have been reduced or mitigated by the use of balanced solutions.
Given there is an absence of high quality evidence supporting the use of fluid boluses in septic shock, this study should at the very least give us pause to reconsider their use. It is however quite possible that smaller (5-10ml/kg) boluses in selected patients may provide benefit without such harm. Additionally reasonably liberal fluid boluses are likely still appropriate in dehydration conditions such as gastroenteritis. Finally this study was in kids so may not be totally applicable to adults but we should be wary of ignoring it on these grounds alone.