Another Journal club, another fruity (or vegetable-y) discussion today.
The recent large prospective RCT by the Scandinavian 6S trial group on HydroxyEthyl Starch vs Ringer’s acetate in severe sepsis was published in the NEJM this month here. A cracking read which pops the lid off the crystalloid vs. colloid debate once again for old times sake.
So what did we think?
800 patients meeting criteria for severe sepsis within the last 24 hours and randomised on the ICU to receive double blinded HES or Ringers as resuscitation fluid. Reasonably powered. Nice allocation concealment. TSC. DSMB. Most of the boxes ticked.
What happened? Well, the HES group didn’t fare very well at all. An 8% increase in all cause mortality at 90 days. A NNH of 13 – Wow! A Relative Risk (if that is your bag) of 1.17; 95% confidence interval [CI], 1.01 to 1.36; P=0.03.
This was a sick group. Mortality in the control (Ringers) group was 43%, which is fairly high for a developed world ICU treating severe sepsis. Baseline characteristics were pretty evenly distributed, including SOFA and SAPS II scores. Randomisation was stratified, accounting for Acute Kidney Injury (AKI), University hospitals and a few other bits and bobs. Most of the common confounders seemed to have been negated. And with double blinding and an objective endpoint (You are either dead or you’re not) they have also sidestepped the most likely sources of bias in a trial of this nature.
We couldn’t really fault it.
Interesting that the Kaplin Meier curves digress at about day 10, implying that it is a sequential phenomenon that causes problems rather than acute toxicity. Some were postulating within the group whether the increased mortality was all due to the incidence of renal failure and subsequent withdrawal of care. I don’t buy that from the stats presented in the paper, but I would be happily convinced if someone could demonstrate the proof to me.
What I do buy, is that there is now compelling evidence against the use of starch based colloids in acute resuscitation. They are off my shelf until someone can show me some evidence why they should go back on.