Antihypertensive treatment associated with lower risk of dementia
Blood pressure lowering and prevention of dementia: an individual patient data meta-analysis
Introduction and methods
Previous observational studies have demonstrated a strong association between elevated blood pressure (BP) and increased risk of dementia and cognitive decline, particularly in individuals aged 40–65 years . However, RCTs have shown mixed results regarding the effects of antihypertensive treatment on the prevention of dementia . Possibly, a meta-analysis can solve this issue. Yet, prior meta-analyses in this area were at risk of bias in their estimates because, for example, analysis and data handling could not be standardized. This limitation can be overcome with a patient-level meta-analysis in which data from sufficiently similar studies are combined and analyzed as a single dataset.
Aim of the study
The aim was to better analyze the effects of antihypertensive treatment on the prevention of dementia by performing a single-stage individual participant data meta-analysis of 5 randomized double-blind placebo-controlled trials.
For the current metanalysis, individual participant data of the following 5 double-blind placebo-controlled RCTs (>2000 participants each) that compared the effect of antihypertensive medication with that of placebo only in participants with increased vascular risk and that included incident dementia as the endpoint were pooled:
- Hypertension in the Very Elderly Trial (HYVET) [3,4];
- SYSTolic Hypertension in EURope trial (SYST-EUR) [5,6];
- Perindopril Protection Against Recurrent Stroke Study (PROGRESS) [7,8];
- Action in Diabetes and Vascular disease: preterAx and diamicroN-MR Controlled Evaluation (ADVANCE) [9-11];
- Systolic Hypertension in the Elderly Program (SHEP) .
The total sample consisted of 28,008 individuals (mean age: 69.1 years; SD: 9.3), who had been recruited from 20 countries. The primary endpoint of the current meta-analysis was incident all-cause dementia. Incident cognitive decline, based on Mini-Mental State Examination (MMSE) scores, was also assessed.
The a priori statistical analysis plan included multilevel logistic regression to evaluate the treatment effect on incident dementia; multilevel multinomial logistic regression to account for death as a competing risk; effect modification by age, sex, baseline systolic BP, prior stroke, or baseline MMSE score; and mediation analysis to quantify the contribution of trial medication and changes in the systolic and diastolic BP to incident dementia.
Effect of antihypertensive treatment on incident dementia
- At 12 months, the mean differences between the placebo and antihypertensive treatment groups were 9.6 mmHg (SD: 20.3) in systolic BP and 3.7 mmHg (SD: 10.4) in diastolic BP.
- After a median follow-up duration of 4.3 years, there were 861 cases of incident dementia: 403 (2.9%) in the antihypertensive treatment group and 458 (3.3%) in the placebo group.
- Multilevel logistic regression showed an unadjusted odds ratio (OR) of 0.868 (95%CI: 0.756–0.996) for the risk of incident dementia in favor of antihypertensive treatment compared with placebo. After adjustment for age, sex, and prior stroke, the OR was 0.865 (95%CI: 0.752–0.994), and it was 0.860 (95%CI: 0.748–0.989) after additional adjustment for BMI and DM.
- Similar results were seen with multilevel multinomial regression in a model adjusted for age and sex. Compared with placebo, antihypertensive treatment reduced the risk of dementia (OR: 0.853; 95%CI: 0.742–0.980) or death (OR: 0.876; 95%CI: 0.805–0.954) compared with achieving neither outcome.
- There were no modifications of treatment effect by baseline systolic BP (P=0.18; estimate: 0.006; standard error (SE): 0.004), age (P=0.80; estimate: −0.002; SE: 0.009), sex (P=0.72; estimate: −0.060; SE: 0.163), prior stroke (P=0.22; estimate: −0.219; SE: 0.180), or baseline MMSE score (P= 0.18; estimate: −0.025; SE: 0.019).
- After adjustment for baseline systolic BP, there was a positive linear relationship between treatment effect and age.
- Mediation analysis showed 53% (95%CI: 27–76) of the difference in dementia risk between the antihypertensive treatment and placebo groups could be attributed to the effect of the treatment on systolic BP rather than to any other aspects of trial participation or pleiotropic antihypertensive drug effects.
- At 12 months, there was a positive linear relationship between systolic BP and dementia risk in the antihypertensive treatment group.
Effect of antihypertensive treatment on incident cognitive decline
- At 24 months, the mean change in MMSE score was 0.006 (SD: 2.18; median: 0) in the antihypertensive treatment group and –0.05 (SD: 2.18; median: 0) in the placebo group. Multilevel linear regression (accounting for study and adjusting for age and sex) showed no significant difference between the 2 groups (P=0.15).
- There was also no effect of antihypertensive treatment on overall cognitive decline (i.e., a sustained decrease in MMSE score) compared with placebo (OR: 0.905; 95%CI: 0.695–1.179).
Classification of evidence
- The analyses provided Class I evidence in favor of antihypertensive treatment in late-midlife to later life reducing the risk of incident dementia compared with placebo.
In this patient-level meta-analysis of pooled raw data from 5 large double-blind placebo-controlled RCTs, antihypertensive treatment was associated with a reduction of incident dementia compared with placebo. A previous meta-analysis of 7 population-based cohorts showed a U-shaped relationship between BP and dementia risk in participants over 80 years of age . The authors of the current meta-analysis emphasized that in their study, “[t]here was no evidence of a U-shaped relation of the effect at any age , nor any increase in risk of dementia with treatment in the oldest age” group (>80 years).