Blood Pressure Lowering Medication, Visit-to-Visit Blood Pressure Variability, and Cognitive Function in Old Age.

Abstract

Visit-to-visit blood pressure (BP) variability is associated with cognitive impairment. We assessed to what extent the association between BP variability and cognitive impairment is mediated by the association of BP lowering medication (BPLM) with both BP variability and cognition. We studied 5,606 participants from the PROspective Study of Pravastatin in the Elderly at Risk (PROSPER). BP was measured every 3 months during 3.2 years; BP variability was defined as the SD of BP measurements during follow-up. Cognitive function was assessed at baseline and during follow-up using the Stroop test, Letter-Digit Coding test, and immediate and delayed Picture-Word Learning tests. Multivariate regression models were used with and without adjustments for BPLM to calculate the percentage to which BPLM mediated the association between BP variability and cognition. Participants taking calcium antagonists had a higher score in baseline Letter-Digit Coding test (mean difference (95% confidence interval (CI) 0.45 (0.06; 0.88). Participants taking beta-blockers had a steeper decline in Stroop test (additional change per year (95% CI) 0.40 (0.09; 0.70) and Letter-Digit Coding test (0.08 (-0.15; -0.02)). Furthermore, a steeper decline in Stroop test was found in participants taking renin-angiotensin system (RAS) inhibitors (0.50 (0.16; 0.85). Systolic BP variability was higher in participants taking beta-blockers and RAS inhibitors (mean difference in systolic BP variability in mm Hg (95% CI) 0.75 (0.45; 1.04) and 1.37 (1.04; 1.71) respectively). Participants taking diuretics, calcium antagonists, and RAS inhibitors had a higher diastolic BP variability (mean difference in diastolic BP variability in mm Hg (95% CI) 0.27 (0.04; 0.49), 0.37 (0.12; 0.62) and 0.65 (0.37; 0.93) SD, respectively). Beta estimates remained essentially the same when we adjusted for BPLM in the association of BP variability with cognitive function. The association between BP variability and cognitive impairment was not mediated by BPLM.