Thursday, December 27, 2007

Top ten issues in hospital medicine for 2007, issue 4: evidence based management of sepsis---is it time to unbundle?

The Surviving Sepsis Guidelines (SSG) are now over 3 years old. I’m going to look into the crystal ball of recent evidence and make some predictions about what the next version will look like, realizing, with apologies to Yogi Berra, that “predictions are risky, especially when they’re about the future.”

The SSG recommendations were put forth as a “bundle” by the Institute for Healthcare Quality (IHI) and other organizations. Recently some key components of the bundle are being challenged by new evidence.

Perhaps the most contentious measure in the guidelines is the recommendation for Drotrecogin Alfa (activated protein C). Drotrecogin Alfa was controversial from the time of its original FDA approval because that approval was based on a subset of patients in a single RCT. Nevertheless, at the time of publication of the SSG this was the best and most current evidence available. Although recent evidence calls the use of Drotrecogin Alfa into question the loudest criticism, a perspective piece in NEJM (really more of a rant containing little in the way of perspective) which preceded this evidence, was not based on evidence at all. As I pointed out in posts here and here it seemed more like an ad hominem attack on the guideline authors for their association with Eli Lilly.

But recent disturbing evidence on Drotrecogin Alfa has surfaced including this paper which raised concerns about real world use of the drug and called for additional clinical trials to assess the risks in relation to the benefits. There was also a meta-analysis I linked to and discussed here which suggested a lack of mortality benefit.

The NEJM perspective authors launched another salvo in a recent letter to the editor of Critical Care Medicine, reported by Hooked: Ethics, Medicine and Pharma and Health Care Renewal. I can’t access the letter, so I don’t know whether it’s another exercise in Pharma bashing, a “see I told you so” or a sober assessment of the new evidence. Nevertheless the new evidence calls into question the efficacy, safety and cost effectiveness of Drotrecogin Alfa. If the letter writers made an appeal to this evidence they’ve gained my respect. If the writers of the next version of the SSG moderate their recommendation for Drotrecogin (and I expect they will) it should help put to rest concerns about marketing being disguised as evidence based medicine.

The use of corticosteroids in stress doses, another component of the bundle, took a hit with the release of CORTICUS (discussed here) which showed no mortality benefit. The study was underpowered, having been halted early due to enrollment difficulties, but showed no signal of benefit. It contained less ill patients than those in earlier smaller studies which showed benefit. Although critical illness related corticosteroid insufficiency (CIRCI) is real, the results of CORTICUS tell us that, while some patients need to be treated, a blanket recommendation is not warranted. Patient selection is now unclear. The best candidates for steroid treatment may be those among the more severely ill (matching the populations evaluated in older studies) who fail to respond to cortrosyn. Expect a more restrictive corticosteroid recommendation in the next version of the SSG.

Glycemic control has been challenged by evidence published since the release of the guidelines. There is little question that glycemic control is important in critical illness but the targets are unclear. Glycemic control recommendations will remain in the guidelines but I expect some refinement in the next update.

Early goal directed therapy, timely initiation of appropriate antibiotics, source control and other components of the bundle remain intact, having withstood the test of time.

No comments: