Early-event pathways to 6-month death or dependency after ischemic stroke: Secondary analysis of the international stroke trial

Background Broad 6-month death-or-dependency outcomes after acute ischemic stroke can conceal opposing early benefit and harm pathways after antithrombotic treatment. We examined these pathways in the corrected public-use International Stroke Trial dataset. Methods We analyzed 19,285 randomized participants with assignable 6-month outcome classification; 150 with missing classification were excluded. The mediator was the first recorded selected event within 14 days or earlier death/discharge: early death, intracranial hemorrhage or hemorrhagic stroke, major extracranial bleeding, pulmonary embolism, recurrent stroke, or no selected event. Standardized risks, risk differences per 1000 patients, and risk ratios were estimated using robust Poisson regression and g-computation. Interventional direct and indirect effects were estimated with multinomial mediator and outcome models, marginalizing factorial co-allocation. Results Aspirin was associated with a small reduction in 6-month death or dependency (risk difference, −12.1 per 1000; risk ratio, 0.981). The interventional direct component was −9.8 per 1000 and the first-event-mediated component was −2.3 per 1000. Low-dose heparin had a near-null total effect (3.0 per 1000), with opposing direct (6.4) and indirect (−3.4) components. Higher-dose heparin also had a near-null composite effect (−1.9 per 1000), but mortality sensitivity analysis showed an unfavorable signal (15.6 per 1000). Conclusions In this historical megatrial, small or near-null disability-inclusive effects were compatible with divergent early ischemic-prevention and bleeding-hazard pathways. Recorded first early events explained only part of the treatment-outcome contrasts. These estimates are methodological and interpretive, not contemporary prescribing guidance.