Diuretic antihypertensive drugs and incident dementia risk: a systematic review, meta-analysis and meta-regression of prospective studies

Abstract

<p><b>Objective:</b> Diuretic drugs have been a mainstay of hypertension treatment in the elderly however their dementia sparing effects are under-reported. The objective was to quantify dementia risk in relation to diuretic antihypertensive drugs.</p> <p><b>Methods:</b> Electronic databases were searched until June 2015. Eligibility criteria: population, adults without dementia from primary care, community cohort, residential/institutionalized, or randomized controlled trial; exposure, diuretic antihypertensive drug; comparison, no diuretic drug, other or no antihypertensive drug, placebo-control; outcome, incident dementia diagnosed by standardized criteria. Adjusted hazard ratios (HR) with 95% confidence intervals (CI) were pooled in fixed-effects models with RevMan 5.3 (The Nordic Cochrane Centre, Copenhagen, Denmark) and the findings rated according to The Grading of Recommendations Assessment, Development and Evaluation criteria.</p> <p><b>Results:</b> A total of 15 articles were included (52 599 persons, 3444 dementia cases, median age 76.1 years) and median follow-up was 6.1 years. Diuretics were associated with reduced dementia risk (HR 0.83; 95% CI 0.76-0.91, <i>P</i> < 0.0001, I² = 0) and Alzheimer's disease risk (HR 0.82; 95% CI 0.71-0.94, <i>P</i> = 0.004, I² = 0). Stratified analysis indicated a difference between potassium sparing, thiazide and loop diuretics (<i>P</i> = 0.01). Risk estimates were generally consistent comparing monotherapy vs. combination therapy, study design and follow-up. Meta-regression showed that demographics, stroke, heart failure, diabetes, liver disease, attrition, mortality rate, cognitive function, and apolipoprotein E allele did not moderate the results.</p> <p><b>Conclusion:</b> Diuretic antihypertensive drugs were associated with a consistent reduction in dementia risk without heterogeneity, pointing to generalizability of these findings.</p>