Introduction: Hypertension is highly prevalent in the older adult population and is associated with an increased risk of dementia. Although the treatment of hypertension is evidence-based in the cognitively intact population, this is not the case for those with dementia, partly because of their exclusion from trials and potential harms such as orthostatic hypotension. Our aim was to establish the evidence for treating hypertension in people with dementia.
Search methods: Data sources: Ovid Medline(R) 1966–2011, EMBASE 1988–2011 Week 41, Cochrane Library and National research register archives. One reviewer screened studies for eligibility. Thereafter, two reviewers independently assessed for inclusion and graded using the van Tulder criteria. Inclusion criteria: randomised-controlled trials of hypertension treatment; Participants >or =65 years; Diagnosis of dementia (global cognitive decline for >6 months affecting function); Cognitive outcomes assessed using validated tools. Exclusion criteria: poor quality (van Tulder <9/19); Mild cognitive impairment; insufficient English language content to extract meaningful information.
Results: From an initial return of 1,178 papers, 24 were selected for review and 7 met full inclusion criteria. The review population was 3,190, followed up for 3–47 months. Only the mild-moderate spectrum of dementia was represented. Four placebo-controlled trials were identified, with three more comparing alternative medication regimes. There was reasonable evidence for BP reduction from antihypertensives, but no clear evidence of benefit (or harm) on cognition or dependence. The marked heterogeneity of design prevented meta-analysis of the data.
Conclusions: The risk–benefit ratio for the use of antihypertensive drugs in people with dementia is not known. Treatment decisions for this group, therefore, depend on extrapolation from evidence found in cognitively intact populations. This extrapolation may not be valid in those, particularly, with severe dementia, at greater risk of adverse drug effects.