Innovation in sepsis trials: rethinking endpoints, statistics and patient stratification

Authors: Sascha David, Marc Leone, Massimo Girardis, Mattia M. Müller, Srdjan Gavrilovic, Roberta Domizi, Elisa Damiani, Ignacio Martin-Loeches, Ricard Ferrer, Christian Bode, Benjamin Chousterman & Lene Russell

Members of the ESICM Systemic Inflammation and Sepsis (SIS) Section have published a viewpoint article in Intensive Care Medicine Experimental examining why so many sepsis trials have failed and how emerging methodological approaches may finally be changing this picture.

The paper, authored by Sascha David, Marc Leone, Massimo Girardis, Mattia Müller, Srdjan Gavrilovic, Roberta Domizi, Elisa Damiani, Ignacio Martin-Loeches, Ricard Ferrer, Christian Bode, Benjamin Chousterman, and Lene Russell, argues that decades of non-conclusive trials reflect not a lack of therapeutic targets, but a fundamental mismatch between the biological complexity of sepsis and the trial designs used to study it.

The authors discuss four recent trials that illustrate a conceptual and methodological shift in the field:

ANDROMEDA-SHOCK-2 used capillary refill time-guided haemodynamic resuscitation and a hierarchical win-ratio composite endpoint to personalise treatment in septic shock and capture clinically meaningful outcomes beyond short-term mortality.

IMMUNOSEP stratified patients by immune phenotype (macrophage activation-like syndrome or immunoparalysis) and demonstrated significantly improved organ function with targeted immunotherapy, using SOFA score trajectory as a biologically meaningful primary endpoint.

INSPIRE tested IL-1 receptor blockade in a biomarker-selected population at the transition between infection and sepsis, targeting progression to organ dysfunction as an endpoint and demonstrating a significant benefit over placebo.

TIGRIS employed Bayesian analytical methods and enriched enrolment by endotoxin activity, increasing trial efficiency and providing probabilistic estimates of treatment benefit rather than binary p-value thresholds.

Together, these trials signal a shift from defining sepsis to deconstructing it, and from uniform treatment strategies to stratified care aligned with dominant pathophysiology.