BACKGROUND AND OBJECTIVE: Randomised, controlled trials and population studies have suggested a link between hypertension and the development of dementia (vascular dementia and Alzheimer's disease) although the results are not conclusive. The very elderly are at highest risk for both hypertension and dementia but have been underrepresented in studies to date. The Hypertension in the Very Elderly Trial (HYVET), an international, randomised, double-blind, placebo-controlled trial of antihypertensive medication for those > or =80 years of age, is currently underway. The protocol for the assessment of cognitive decline and detection of incident dementia cases is described in this article. The objective of this protocol is to determine whether treatment of hypertension in this very elderly group affects the incidence of dementia and decline in cognitive function during the period of the HYVET trial. METHODS: Cognitive function in all HYVET trial participants is assessed at baseline and annually thereafter using the Mini-Mental State Examination (MMSE). An MMSE score that decreases more than three points per year or decreases to <24 prompts further investigation for possible incident dementia. This includes use of the Diagnostic and Statistical Manual-IV criteria, the modified Hachinski scale and CT scanning. In cases where a CT scan cannot be obtained the full Hachinski scale is used. Additional information is provided when available from the clock drawing test and Geriatric Depression Scale. Data are also gathered from all trial participants with regard to education, blood pressure, alcohol use, co-morbidities and concomitant treatments. At the end of the trial, changes in cognitive function and incident dementia cases will be compared in those receiving active treatment (indapamide slow-release +/- perindopril) and those receiving matching placebo. RESULTS: The protocol has been applied successfully and the data collection phase is underway. DISCUSSION: Previous population studies have been unable to study antihypertensive use in a controlled manner and randomised, controlled trials suggest a positive or mixed picture regarding an association between use of antihypertensive therapy and reduction in incident cognitive decline and/or dementia. No such trial has yet been performed exclusively in subjects > or =80 years of age with hypertension. The HYVET trial is placebo controlled and while investigating the cardiovascular effect of treatment also assesses cognitive function and quality of life in both the placebo and actively treated arms of the trial. The HYVET cognitive substudy trial (HYVET-COG) design has been successfully applied and the instruments chosen appear to be appropriate, valid and pragmatic. CONCLUSION: HYVET-COG provides the first opportunity to examine the effect of antihypertensive treatment on incident dementia in a high-risk very elderly population.