Phenotype-guided polymyxin B haemoadsorption in abdominal septic shock: a 2015–2026 evidence synthesis and proposed clinical pathway

PURPOSE: Polymyxin B haemoadsorption (HP-PMX) has generated a decade of contradictory evidence in septic shock. After the Tigris phase 3 Bayesian trial and the accumulation of Japanese and European cohorts, we synthesise the 2015-2026 literature and propose a clinical pathway for phenotype-guided indication at the bedside in European intensive care units (ICUs). MATERIALS AND METHODS: SANRA-aligned narrative review of PubMed/MEDLINE, Embase and Cochrane Library (January 2015-May 2026), with no language restriction and independent selection by two authors. The clinical pathway was derived from convergent domains identified in the literature without formal external consensus. RESULTS: Available evidence converges on three domains defining a responder phenotype: detectable endotoxaemia (Endotoxin Activity Assay [EAA] 0.60-0.89); intermediate organ dysfunction (Sequential Organ Failure Assessment [SOFA] 7-13); intra-abdominal source, with early initiation as a fourth, operationally derived criterion. Tigris reported a marginal reduction in 90-day mortality of 15.5% (95% credible interval 3.6-27.1), with a posterior probability of benefit of 99.4% at 90 days and 95.3% at 28 days. Japanese (SOFA 7-12; n=4,066) and European (SOFA 8-13) cohorts describe an operatively equivalent severity range. Tanaka reported that the survival benefit concentrates in intra-abdominal infection (adjusted HR 0.485, 95% CI 0.252-0.935; p=0.031). A five-step decision algorithm is proposed in which SOFA 7-13 acts as a clinical surrogate when the EAA is unavailable. CONCLUSIONS: HP-PMX is not a universal therapy for septic shock but an example of phenotype-guided therapy. The proposed pathway is a synthesis-derived framework intended to standardise bedside decision-making in European ICUs ahead of formal external validation, which we identify as the necessary next step.