Background Asundexian (Bayer AG, Leverkusen, Germany), an oral small molecule factor XIa (FXIa) inhibitor, might prevent thrombosis without increasing bleeding. Asundexian's effect for secondary prevention of recurrent stroke is unknown. Methods In this randomised, double-blind, placebo-controlled, phase 2b dose-finding trial (PACIFIC-Stroke), patients with acute (within 48 h) non-cardioembolic ischaemic stroke were recruited from 196 hospitals in 23 countries. Patients were eligible if they were aged 45 years or older, to be treated with antiplatelet therapy, and able to have a baseline MRI (either before or within 72 h of randomisation). Eligible participants were randomly assigned (1:1:1:1), using an interactive web-based response system and stratified according to anticipated antiplatelet therapy (single vs dual), to once daily oral asundexian (BAY 2433334) 10 mg, 20 mg, or 50 mg, or placebo in addition to usual antiplatelet therapy, and were followed up during treatment for 26-52 weeks. Brain MRIs were obtained at study entry and at 26 weeks or as soon as possible after treatment discontinuation. The primary efficacy outcome was the dose-response effect on the composite of incident MRI-detected covert brain infarcts and recurrent symptomatic ischaemic stroke at or before 26 weeks after randomisation. The primary safety outcome was major or clinically relevant non-major bleeding as defined by International Society on Thrombosis and Haemostasis criteria. The efficacy outcome was assessed in all participants assigned to treatment, and the safety outcome was assessed in all participants who received at least one dose of study treatment. This study is registered with ClinicalTrials.gov, NCT04304508, and is now complete. Findings Between June 15, 2020, and July 22, 2021, 1880 patients were screened and 1808 participants were randomly assigned to asundexian 10 mg (n=455), 20 mg (n=450), or 50 mg (n=447), or placebo (n=456). Mean age was 67 years (SD 10) and 615 (34%) participants were women, 1193 (66%) were men, 1505 (83%) were White, and 268 (15%) were Asian. The mean time from index stroke to randomisation was 36 h (SD 10) and median baseline National Institutes of Health Stroke Scale score was 2.0 (IQR 1.0-4.0). 783 (43%) participants received dual antiplatelet treatment for a mean duration of 70.1 days (SD 113.4) after randomisation. At 26 weeks, the primary efficacy outcome was observed in 87 (19%) of 456 participants in the placebo group versus 86 (19%) of 455 in the asundexian 10 mg group (crude incidence ratio 0.99 [90% CI 0.79-1.24]), 99 (22%) of 450 in the asundexian 20 mg group (1.15 [0.93-1.43]), and 90 (20%) of 447 in the asundexian 50 mg group (1.06 [0.85-1.32]; t statistic -0.68; p=0 center dot 80). The primary safety outcome was observed in 11 (2%) of 452 participants in the placebo group versus 19 (4%) of 445 in the asundexian 10 mg group, 14 (3%) of 446 in the asundexian 20 mg group, and 19 (4%) of 443 in the asundexian 50 mg group (all asundexian doses pooled vs placebo hazard ratio 1.57 [90% CI 0.91-2.71]). Interpretation In this phase 2b trial, FXIa inhibition with asundexian did not reduce the composite of covert brain infarction or ischaemic stroke and did not increase the composite of major or clinically relevant non-major bleeding compared with placebo in patients with acute, non-cardioembolic ischaemic stroke.