A pathogenetic role in thrombotic disease, particularly in young people, has been postulated for anticardiolipin antibody (ACA). We have carried out a prospective controlled study of 262 unselected patients with acute stroke and 226 controls to assess the prevalence and relation to age and vascular risk factors of ACA. Titres of IgG, IgA, or IgM ACA were above the upper normal limit in 38% of patients. The proportions of patients and controls with raised titres did not differ significantly (13 vs 8% IgG, 22 vs 29% IgA, 11 vs 7% IgM). IgG titres were higher among patients than among controls (mean 3.88 vs 2.86 u/mL [95% CI for difference 0.62-0.87], p = 0.0004), whereas IgA and IgM titres were lower in patients than in controls (IgA 4.82 vs 5.98 u/mL [1.12-1.60], p = 0.01; IgM 3.00 vs 3.64 u/mL [1.01-1.45], p = 0.04). However, within age tertiles the only significant difference between patients and controls for IgG ACA was in the oldest tertile. Analysis by number of risk factors for stroke showed a significant difference between the groups only for subjects with one risk factor. IgA and IgM ACA titres were higher among controls only in those with no vascular risk factors. We found no evidence to support the hypothesis that ACA is an independent risk factor for stroke in young people. The increase in IgG titre with age and number of vascular risk factors in stroke patients suggests that ACA may be a non-specific accompaniment of vascular disease. Routine testing for ACA in stroke patients is not justified.