Tuesday, January 20, 2009

Treatment of Sepsis

I heard a talk on "Sepsis & Renal Failure" today, which reviewed much of the recent evidence-based medicine trials regarding the treatment of sepsis in the ICU.

Honestly, the main message I came away with was this: most of the purported "advances" in sepsis treatment I learned in residency are probably not that useful. Let me explain: As an internal medicine resident, there were a four hot ICU papers that were repeatedly emphasized. They were:

#1: "early goal-directed therapy" by Rivers et. al. published in NEJM 2001. This protocol-driven ER-based study showed that keeping the MAP >65 with fluids and inotropes, the CVP between 8-12, and transfusing blood to achieve an SvO2 >70% & Hct >30% was associated with improved mortality in sepsis patients. This is one of the few studies which hasn't been debunked, so it is still in use.

#2: steroid therapy. I was taught to perform a cosyntropin stim test to determine if sepsis patients were "relatively adrenally-insufficient", and if so to give a give low-dose steroids with hydrocortisone 50mg iv q6h and fludrocortisone as well. Most of the time we gave the steroids right off the bat and took them off only if the cosyntropin test came back negative the next day or so. However, a more recent NEJM 2008 study (the CORTICUS trial) shows a pretty convincingly negative result for this practice.

#3: Intensive insulin therapy. It was also en vogue to be pretty aggressive with insulin drips in the ICU in an attempt to achieve tight glucose control in the setting of sepsis, reasoning that hyperglycemia can have a cell-toxic effect. The recent VISEP study was a large trial which (among other things) compared an intensive glucose control strategy to a more conventional one: the intensive glucose control strategy showed no mortality benefit, while showing many more hypoglycemic episodes.

#4: Activated Protein C (Xigris). The possibility that derangement of the coagulation cascade plays a role in sepsis led to the development of activated protein C (drotrecogin alfa, or Xigris) for use in sepsis patients. An NEJM study showed a mild beneficial effect (31% versus 25% mortality) in the PROWESS trial, a randomized control trial. Even though the effect was small and the drug was expensive, as residents we were encouraged to give the drug under the rationale that these patients were super-sick and they needed any advantage they could get in order to make it through. However, a more recent trial, ADDRESS, looked at septic patients who were slightly less sick (e.g., lower APACHE thresholds for administering Xigris) and found no mortality benefit and an increased rate of hemorrhage in the Xigris group compared to placebo.

So of these four "breakthroughs" in the treatment of ICU sepsis, the efficacy of three have been seriously called into question through the use of large, seemingly well-designed randomized control trials. It appears that more creative approaches will still be needed to come up with better therapies for this high-mortality (and highly renal-toxic) condition.


Joel Topf said...

Outstanding post. I liked the more indepth perspective with useful links for EBM buffs to dig even deeper.

The early goal directed therapy (EGDT) is a hometown paper done just down the road at Henry Ford Hospital.

It has been interesting to see the difference in care at various ICUs. In Chicago, the ICU regularly used SVO2 to monitor perfusion and used dobutamine in sepsis as done in the EGDT, here at St John's the intensivists are less focused on it and never use dobutamine outside of cardiogenic shock.

Anonymous said...

EGDT is definitely the key. Early awareness and education of staff is imperative. Many places are using I-STATs for lactates that have a result time of two minutes at the bedside.

nathanhellman said...

Many of the ICU docs I've spoken with regarding the EGDT seem convinced that the large amount of fluids patients on these protocols get is the key, which makes inherent sense to nephrologists.