OBJECTIVE: The risk of stroke rises after episodes of herpes zoster and chickenpox, which are caused by varicella zoster virus (VZV). We conducted a pilot case-control study of stroke, nested within a long-standing cohort in Uganda (the General Population Cohort), to examine antibodies against VZV prior to diagnosis. METHODS: We used stored sera to examine the evolution of IgG and IgM antibodies against VZV among 31 clinically confirmed cases of stroke and 132 matched controls. For each participant, three samples of sera were identified: one each, taken at or near the time of (pseudo)diagnosis, between 5 and 10 years prior to diagnosis and at 15 years prior to diagnosis. RESULTS: All participants had detectable antibodies against VZV, but there were no significant differences between cases and controls in the 15 years prior to diagnosis. As a secondary finding, 16% (5/31) of cases and 6% (8/132) of controls had HIV (OR 3.0; 95% CI 0.8-10.1; P = 0.06). CONCLUSIONS: This is the first prospective study to examine a biological measure of exposure to VZV prior to diagnosis of stroke and although we identified no significant association, in this small pilot, with limited characterisation of cases, we cannot exclude the possibility that the virus is causal for a subset. The impact of HIV on risk of stroke has not been well characterised and warrants further study.