Modelled mortality benefits of reducing fibrinolysis resistance in sepsis: integrating observational and ex vivo data

Background. Suppressed fibrinolysis (fibrinolysis resistance) is increasingly recognised as a driver of organ failure in sepsis, yet trials in sepsis-induced coagulopathy have largely targeted coagulation rather than fibrinolysis. This study combined an observational ICU cohort with nested analyses of plasma fibrinolysis proteins and ex vivo reversibility, modelling the potential mortality reduction from restoring fibrinolytic capacity to inform trial design. Methods. In 188 adults admitted to ICU with sepsis, repeated tissue plasminogen activator (tPA)-enhanced viscoelastometric testing quantified fibrinolysis resistance as the fibrin adjusted lysis time (TPA-LT/FIBA10) and as ordinal Grades 1-3 of worsening resistance. Adjusted logistic regression and time-varying Cox models related fibrinolysis resistance to death within 28 days of admission, with unmeasured confounding assessed by E-values. In nested sub-studies, plasma fibrinolysis proteins were measured and whole blood was tested ex vivo for reversal of resistance by supplemental tPA, plasminogen, or an inhibitory antibody to a2-antiplasmin (a2AP iAb). Ex vivo effects were modelled onto the observed fibrinolysis resistance-mortality relationship using a hypothetical intervention distribution framework. Results. Each e-fold (~2.72) increase in TPA-LT/FIBA10 was associated with a 2.2-fold (95% CI 1.5-3.3) increase in the odds of death in the fully adjusted logistic model and a 2.3-fold (1.6-3.2) increased hazard in the time-varying Cox model. Comparing the least (Grade 1) with the worst (Grade 3) resistance state, the odds ratio was 7.4 (2.4-23) and the hazard ratio 6.7 (3.0-15). Associations were robust to unmeasured confounding, with E-values of 3-13. Grade 3 resistance showed the highest PAI-1/tPA and a2AP/plasminogen ratios. Ex vivo supplementation with plasminogen or a2AP iAb reversed fibrinolysis resistance from Grade 3 to Grade 1. Modelling these effects yielded hypothetical absolute mortality reductions of 18-21% across Grades 2 and 3, and 32-36% in Grade 3. Conclusions. Fibrinolysis resistance measured at the bedside in sepsis reflects mechanistically reversible changes in fibrinolytic balance ex vivo. Modelling reversal by plasminogen supplementation or a2AP inhibition yielded hypothetical absolute mortality reductions of up to 32% in the most severe resistance state. These hypothesis-generating estimates identify candidate interventions and provide a rationale for a clinical trial.