BACKGROUND: The optimal treatment for patients with ischaemic stroke with a high risk of cerebral haemorrhage is unclear. We assessed the efficacy and safety of cilostazol versus aspirin, with and without probucol, in these patients. METHODS: In this randomised, controlled, 2 × 2 factorial trial, we enrolled patients with ischaemic stroke with a history of or imaging findings of intracerebral haemorrhage or two or more microbleeds from 67 centres in three Asian countries. Patients were randomly assigned (1:1:1:1) to receive oral cilostazol (100 mg twice a day), aspirin (100 mg once a day), cilostazol plus probucol (250 mg twice a day), or aspirin plus probucol with centralised blocks stratified by centre. Cilostazol versus aspirin was investigated double-blinded; probucol treatment was open-label, but the outcome assessor was masked to assignment. The co-primary outcomes were incidence of the composite of stroke, myocardial infarction, or vascular death (efficacy) and incidence of haemorrhagic stroke (safety), which were assessed in intention-to-treat and modified intention-to-treat populations. Efficacy was analysed with a non-inferiority test and a superiority test if non-inferiority was satisfied. Safety was assessed with a superiority test only. This trial is registered with ClinicalTrials.gov, NCT01013532. FINDINGS: Between Aug 1, 2009, and Aug 31, 2015, we randomly assigned 1534 patients to one of the four study groups, of whom 1512 were assessed for the co-primary endpoints. During a median follow-up of 1·9 years (IQR 1·0-3·0), the incidence of composite vascular events was 4·27 per 100 person-years in patients who received cilostazol and 5·33 per 100 person-years in patients who received aspirin (HR 0·80, 95% CI 0·57-1·11; non-inferiority p=0·0077; superiority p=0·18). Incidence of cerebral haemorrhage was 0·61 per 100 person-years in patients who received cilostazol and 1·20 per 100 person-years in those who received aspirin (HR 0·51, 97·5% CI 0·20-1·27; superiority p=0·18). The incidence of vascular events was 3·91 per 100 person-years in the probucol group compared with 5·75 per 100 person-years in the non-probucol group (HR 0·69, 95% CI 0·50-0·97; superiority p=0·0316). The incidence of cerebral haemorrhage was 0·72 per 100 person-years in the probucol group and 1·11 per 100 person-years in the non-probucol group (HR 0·65, 97·5% CI 0·27-1·57; p=0·55). Adverse events were similar across the four study groups; the most common events were dizziness, headache, diarrhoea, and constipation. INTERPRETATION: In patients with ischaemic stroke at high risk of cerebral haemorrhage, cilostazol was non-inferior to aspirin for the prevention of cardiovascular events, but did not reduce the risk of haemorrhagic stroke. Addition of probucol to aspirin or cilostazol could be beneficial for reducing the incidence of cardiovascular events. FUNDING: Korea Otsuka Pharmaceutical.