Incident Dementia in Trials of Antihypertensive Treatments

Abstract

Received May 13, 2019 Accepted for publication May 21, 2019 The Systolic Blood Pressure Intervention Trial (SPRINT) Memory and Cognition in Decreased Hypertension (MIND) study (1), failed to show a significant effect on the primary endpoint (probable dementia, hazard ratio [HR] 0.83; 95% CI, 0.67-1.04). Because this may be due to limited power, the results of SPRINT MIND were put in the context of major randomised and controlled hypertension trials (RCT) in which incident dementia had been adjudicated. The same search strategy as in an earlier meta-analysis (up to 2015) (2) was used to identify trials which fulfilled the following criteria: 1) RCT, 2) antihypertensive treatment for at least 2 years, 3) dementia (not only cognitive decline or clinical impression) as an adjudicated endpoint. Trials not fulfilling those criteria were excluded. Seven trials were identified, five trials were placebo-controlled, one compared intensive vs. usual treatment (1), and in one antihypertensives were part of multidomain prevention (3). HRs with 95% CI, and a meta-analysis of trials using random effects model and test for heterogeneity were performed with NCSS statistical software 8 (www.ncss.com). Clinical data and HRs with 95% CI for individual trials are shown in Table 1. In addition to total in-trial results, some subgroup and extension data, and mild cognitive impairment (MCI) for SPRINT MIND are presented. Except in the Study on Cognition and Prognosis in the Elderly (SCOPE) (small inter-group blood pressure differences), point estimates of HRs for dementia were generally below unity suggesting benefit. However, the results were statistically significant only for the Systolic Hypertension in Europe (Syst-Eur) trial extension and in SPRINT MIND when incident MCI was combined with probable dementia (1). When results of the seven trials were combined in a meta-analysis (only adjudicated dementia, n=1,297), active or a more intensive antihypertensive treatment was associated with a significant reduction of dementia by 13% (95% CI -3% to -23%, P=0.011). Although heterogeneity test (Cochran’s Q) was nonsignificant (P=0.60), the studies are admittedly very different, but this would rather give robustness to summary result. The result has also biological plausibility from observational studies (2). To the best of our knowledge, no other intervention has similar record of dementia reduction in randomized trials as antihypertensive treatment. This message should be actively promoted in ageing societies as a feasible, usually safe, and with generic drugs also inexpensive way of dementia prevention in younger-old people. However, this may not apply to frailest and oldest patients (4).