BACKGROUND: Observational epidemiological studies have shown a positive association between hypertension and risk of incident dementia; however, the effects of antihypertensive therapy on cognitive function in controlled trials have been conflicting, and meta-analyses of the trials have not provided clear evidence of whether antihypertensive treatment reduces dementia incidence. The Hypertension in the Very Elderly trial (HYVET) was designed to assess the risks and benefits of treatment of hypertension in elderly patients and included an assessment of cognitive function. METHODS: Patients with hypertension (systolic pressure 160-200 mm Hg; diastolic pressure <110 mm Hg) who were aged 80 years or older were enrolled in this double-blind, placebo-controlled trial. Participants were randomly assigned to receive 1.5 mg slow release indapamide, with the option of 2-4 mg perindopril, or placebo. The target systolic blood pressure was 150 mm Hg; the target diastolic blood pressure was 80 mm Hg. Participants had no clinical diagnosis of dementia at baseline, and cognitive function was assessed at baseline and annually with the mini-mental state examination (MMSE). Possible cases of incident dementia (a fall in the MMSE score to <24 points or a drop of three points in 1 year) were assessed by standard diagnostic criteria and expert review. The trial was stopped in 2007 at the second interim analysis after treatment resulted in a reduction in stroke and total mortality. Analysis was by intention to treat. The trial is registered with ClinicalTrials.gov, number NCT00122811. FINDINGS: 3336 HYVET participants had at least one follow-up assessment (mean 2.2 years) and were included: 1687 participants were randomly assigned to the treatment group and 1649 to the placebo group. Only five reports of adverse effects were attributed to the medication: three in the placebo group and two in the treatment group. The mean decrease in systolic blood pressure between the treatment and placebo groups at 2 years was systolic -15 mm Hg, p<0.0001; and diastolic -5.9 mm Hg, p<0.0001. There were 263 incident cases of dementia. The rates of incident dementia were 38 per 1000 patient-years in the placebo group and 33 per 1000 patient-years in the treatment group. There was no significant difference between treatment and placebo groups (hazard ratio [HR] 0.86, 95% CI 0.67-1.09); however, when these data were combined in a meta-analysis with other placebo-controlled trials of antihypertensive treatment, the combined risk ratio favoured treatment (HR 0.87, 0.76-1.00, p=0.045). INTERPRETATION: Antihypertensive treatment in elderly patients does not statistically reduce incidence of dementia. This negative finding might have been due to the short follow-up, owing to the early termination of the trial, or the modest effect of treatment. Nevertheless, the HYVET findings, when included in a meta-analysis, might support antihypertensive treatment to reduce incident dementia.