Surviving Sepsis 2008

In a December 27th post (one of my top ten issues in hospital medicine for 2007) I offered some predictions on what the first update of the 2004 guidelines would look like. Little did I know that the new guidelines would be published in a matter of days. (I’m not sure exactly when they were posted but I first found them, as a special article in the January issue of Critical Care Medicine, on December 31).

Thus far I have only seen the abstract. The full text of the guidelines, which currently requires a subscription, should soon be published free on the SCCM web site. What changes were made from the old guidelines? How did I do with my predictions? The abstract lists the recommendations and their grades, so I’ll take a stab.

The changes were mainly in the recommendation grades. Recommendations were graded as strong (grade 1) or weak (grade 2). For grade 1 recommendations the risk/benefit ratio had to be clearly in favor of the recommendation. For grade 2 recommendations the risk/benefit ratio was “less clear.” Although the grading scheme differs from that of the 2004 A-E categories I was able get an idea of whether a given recommendation was strengthened or weakened.

The recommendations which were the focus of my post were activated protein C (apC or Xigris), corticosteroids and glycemic control. In my discussion of aPC I cited negative evidence that has come out since the 2004 guidelines and predicted that the guideline authors would weaken their recommendation for aPC and it appears that’s what they’ve done. The 2004 guidelines carried a strong recommendation (grade B on an A-E scale) whereas the 2008 recommendation is grade 2 (weak recommendation) on a 1-2 scale. Moreover, the authors added a strong recommendation (grade 1) against the use of aPC in children.

Concerning corticosteroids I predicted, citing the CORTICUS study, that the new guideline recommendations would be more restrictive. It appears that they are although the differences in wording are subtle. The grading of the recommendation went from C (A-E) to 2 (1-2, weak recommendation) and the indications became more limited. The old guidelines recommended corticosteroids for all patients who required pressor agents after volume repletion. The new ones restrict steroids to those patients whose blood pressures are poorly responsive to combined fluid and pressor therapy.

For glycemic control the guidelines strengthened the general recommendation that glycemic control should be implemented but weakened the recommendation for the specific target of 150mg/dl, in keeping with my previous remarks:

Just about the only thing everyone agreed on this year was that hyperglycemia should not be ignored in hospitalized patients. The unresolved questions relate to what target for what indication.

Other changes of note include the removal of a weak recommendation for vasopressin to be considered as an adjunct in patients refractory to fluid resuscitation and high dose pressors. This probably reflects recent papers suggesting difficulty in patient selection and adverse effects. This study from 2005 showed that, although outcomes may improve with the early use of vasopressin, its initiation after titration of norepinephrine to doses exceeding 0.6mcg/kg/min was associated with increased mortality. Moreover, the use of vasopressin was associated with elevated liver function tests and thrombocytopenia. The VASST study (unpublished as far as I know) showed no overall difference in mortality with vasopressin and again showed that the effect depended on the dose of norepinephrine administered at the time vasopressin was started.

The 2008 guidelines added a strong recommendation against routine use of PA catheters in the management of ALI and ARDS, in keeping with the findings of the Fluid and Catheter Treatment Trial.

I should make a final point about aPC (Xigris). With the benefit of time, additional evidence and an examination of the first update of the Surviving Sepsis Guidelines we are in a good position to revisit the question of whether the guidelines are, according to the accusations of some, merely marketing disguised as evidence based medicine. The available evidence in 2004, from a single randomized controlled trial, was that aPC improved mortality in patients with severe sepsis with a high risk of death. Though far from perfect the trial was the best and most current evidence the guideline authors had to work with. Its inclusion in the guidelines was appropriate according to the principles of evidence based medicine. But the naysayers who accused the guideline authors of being in the pockets of Pharma made up their own EBM principle, seemingly out of thin air: judge, and sometimes reject, evidence by the company it keeps. That, as I pointed out, was more of an attack on the authors than an appeal to evidence.

When newer evidence published 3 years after the SSS guidelines raised concerns about the risk/benefit ratio of Xygris I posted it here and suggested that the guideline authors would need to reconsider their recommendation in the update. A reasonable synthesis of the new evidence with what was known before is that Xigris saves lives, but in which patients? Patient selection is difficult and, according to the evidence synthesis, should take into account time to anticipated treatment (perhaps restricting it to patients who can be treated within 24 hours of the first organ failure), bleeding risks (perhaps employing stricter criteria for exclusion) and risk of death (in which we consider restricting it to patients in septic shock or with a combination of multiple organ failures and APACHE II scores above 25). Given this evidence, when it came time to update the guidelines for 2008, what choices did the guideline writers have? Given that Xigris can save lives in carefully selected patients, its elimination from the guidelines would have been inappropriate. What the writers seem to have done was give Xigris the lowest possible grade of recommendation without eliminating it. That, it seems to me, is the best they could do given the evidence at hand.

I’ll examine the full text of the guideline once I can access it and post a follow up if indicated. I hope it puts to rest once and for all the notion of marketing disguised as EBM.