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

Despite extensive research, effective causal therapies for sepsis remain elusive, highlighting a persistent translational gap between biological understanding and clinical outcomes. Traditional randomised clinical trials (RCTs) have largely failed, likely due to profound patient heterogeneity, dynamic immune trajectories, and reliance on broad syndromic definitions coupled with mortality-centric endpoints. Such designs risk diluting treatment effects and obscuring biologically meaningful signals. Recent studies, however, suggest that precision medicine approaches may overcome these limitations. Trials such as ANDROMEDA-SHOCK-2, IMMUNOSEP, INSPIRE, and TIGRIS have operationalised patient stratification based on physiological or immunological phenotypes and employed novel endpoints or statistical frameworks. ANDROMEDA-SHOCK-2 used capillary refill time guided haemodynamic resuscitation and a hierarchical win-ratio composite endpoint to personalise treatment and enhance sensitivity to benefit. IMMUNOSEP stratified patients by macrophage activation-like syndrome or immunoparalysis, demonstrating improved organ function with targeted interventions using the SOFA score as a primary endpoint. Similarly, INSPIRE focused on progression to organ dysfunction in high-risk pneumonia patients, while TIGRIS employed Bayesian analysis to enrich patients by endotoxin activity, increasing trial efficiency. These studies illustrate that aligning interventions with biological pathways, incorporating adaptive trial designs, and moving beyond short-term mortality can reveal clinically meaningful benefits obscured in conventional trials. Collectively, they signal a conceptual shift in sepsis research towards precision-guided therapies, adaptive protocols, and biologically informed endpoints, suggesting that the era of universal failure may be giving way to more targeted, effective approaches in critical care.