,
10 tweets,
5 min read
Read on Twitter
Lots of prognostic tools for #sepsis exist, and many are well-validated. Although endotypes often show outcomes differences at baseline, they are NOT for outcomes prediction! Purpose-built tools and biomarkers are better for this. /2
Instead, the goal of #sepsis endotyping is PREDICTIVE enrichment, by which we mean, predicting which patients will respond favorably to a given drug (classically an immunotherapy). Since no drug has ever been shown to work in the broad sepsis population, the idea is that /3
Endotypes (aka subtypes) may be able to reduce #sepsis heterogeneity by identifying patients with different underlying biology (and hence different response to therapy). But, here’s the rub: development is a chicken-and-egg problem. /4
Endotypes can be defined by their clinical phenotype differences, but arguably aren’t “real” (or clinically useful) until they’ve identified a subgroup of patients differentially responsive to a therapy. But, clinical trialists rightly only want to test validated endotypes. /5
Successfully validated therapy/endotype pairings (variously termed as “theranostic” or as “companion diagnostics”) would undoubtedly unleash a wave of investment in trying to repeat the success. This would benefit patients eventually. So it’s in everyone’s interest to try! /6
So, please: for those studying novel therapies for #sepsis, work with us or others (eg Hector Wong, @KnightGenetics, @brendonp77, Carolyn Calfee, etc) to make a plan for post-hoc retrospective endotyping in your upcoming studies! Add a PAXgene RNA tube at randomization! /7
And best of luck in your #sepsis therapy trials! Just keep us in mind if/when you miss primary endpoints. Feel free to DM me for paper links, etc. Ours is here: journals.lww.com/ccmjournal/Abs…. @Inflammatix_Inc /8
Finally, a great recent review on endotypes is below. Thanks! ccforum.biomedcentral.com/articles/10.11… /end
Here’s Hector Wong’s interpretation just presented at #ISICEM19! 
