Sepsis: Untangling Bundles, Goals, Measures

Probably one of the of the most paradigm-changing papers in the fields of emergency medicine, critical care medicine, and infectious disease was Rivers, Nguyen and colleagues early goal directed therapy for sepsis paper published in The New England Journal of Medicine in 2001. 

Sepsis, the common final pathway for infectious syndromes and what your grandma called "blood poisoning", is a dysregulated immune response to infection that causes organ dysfunction.  Severe sepsis carries a mortality rate of approximately 25% and is a cause or contributor to many deaths. 

The EGDT paper, with its spectacular results showing a 16% mortality reduction, spurred the development of international consensus guidelines, the Surviving Sepsis campaign, and elevated sepsis to the medical emergency that it is. The bundle of interventions included in EGDT (aggressive fluid resuscitation, prompt antibiotic administration, vasoactive medication administration, blood transfusions)  has quickly come to be regarded as the standard of care.

However, there has been growing concern regarding just how powerful the bundle is and which components are essential and which have less importance. The ARISE, ProCESS, and ProMISe trials failed to replicate the results of the initial trial and there is a strong argument that sepsis care has advanced so much, because of EGDT, that "usual" care incorporates much more than it did in 2001 (because of the results of the initial trial) and later trial patients got many of the interventions of EGDT before randomization. As quality measures and reporting increasingly hone in on the bundle, it is essential to unpack the bundle and not grade providers and facilities negatively if not all components of the bundle contribute positively.

A pair of new papers provide some insight that may help to disentangle the various component of the bundles. The first, by Khalil and colleagues, was published in Critical Care Medicine and is focused on explaining the discordant results between Rivers's study and observational studies that support EGDT and the three negative randomized trials. The conclusion reached was that in the observational studies time to antibiotic administration was substantially faster in the EGDT arms and in the randomized trials very similar arguing that timing of antibiotics explained the discordant results. Interestingly, they also noted increased mortality with EGDT in the sickest group of patients.

The second paper is a patient level meta-analysis of the ARISE, ProCESS, and ProMISe trials. This type of analysis adds increased power to detect effects. The results again show no effect of EGDT on mortality and in subgroup analyses controlling for how aggressive a given center is in resuscitation -- addressing the "usual care" conjecture -- no difference was found. The investigators did find EGDT being associated with higher costs, however. 

So what to make of all this ? A few thoughts:

  • The bundle needs to be unbundled: as third-party payors have focused on bundle compliance as a quality measure that may influence payment it will be crucial to make sure they are measuring something that unequivocally matters
  • Antimicrobials are of paramount importance: sepsis is a response to infection and taming the infection with antibacterials, antivirals, antifungals, and/or antiparasitic medications as quickly as possible may make the biggest difference (even the order in which one gives antibiotics may matter, depending on the inciting organism)
  • As the PRISM investigators note, individualized sepsis care based on variables particular to the patient being treated, rather than a one-size-fits-all approach, is likely the correct manner to treat sepsis

Sepsis is an important diagnosis to make and aggressive care, even if non-formulaic, should be the norm. The Rivers trial was pathbreaking and changed the dialogue regarding this condition. I am sure that these latest salvos in the medical literature will be answered and more elucidation will be forthcoming,