Abstract
Objective: To determine the association between discrete antihypertensive drug classes and incident dementia controlling for blood pressure variability (BPV) in the preceding 4 years. Methods: A total of 6,537 participants (mean age 79 years, 62% women) in a prospective population-based cohort were followed up for incident dementia. A 4-year time lag period was created to classify drug exposure and measure blood pressure. BPV (percent coefficient of variation [CV]) was regressed against 9 antihypertensive drug classes (BPV reg). Cox regression models were employed to determine hazard ratios (HRs) for incident dementia thereafter according to drug class, adjusted for mean blood pressure, covariates, and BPV or BPV reg. Results: Over a median 8.4 years follow-up (interquartile range 6.7-9.0), lower dementia risk was associated with nondihydropyridine calcium channel blocker (HR 0.56; 95% confidence interval [CI] 0.31-1.00, p 5 0.05) and loop diuretics (HR 0.45; 95% CI 0.22-0.93, p 5 0.03) after adjusting for CV-BPV. Similar findings were obtained in analyses restricted to antihypertensive drug users for nondihydropyridine calcium channel blocker (HR 0.52; 95% CI 0.28-0.95, p 5 0.03) and loop diuretics (HR 0.40; 95% CI 0.19-0.83, p 5 0.01). All systolic BPV 3 antihyper-tensive drug interaction terms were not different from p , 0.05. Conclusions: Nondihydropyridine calcium channel blocker and loop diuretics were associated with a reduced dementia risk independent of CV-BPV in the preceding 4 years. Systolic BPV was not the primary mechanism through which antihypertensive drug classes lower dementia risk. Neurology ® 2016;87:1-8 GLOSSARY 3C 5 Three-City; ACEI 5 angiotensin-converting enzyme inhibitor; AD 5 Alzheimer disease; ARB 5 angiotensin-2 receptor blocker; BMI 5 body mass index; BP 5 blood pressure; BPV 5 blood pressure variability; CCB 5 calcium channel blocker; CI 5 confidence interval; CKD 5 chronic kidney disease; CV 5 coefficient of variation; DSM-IV 5 Diagnostic and Statistical Manual of Mental Disorders, 4th edition; HR 5 hazard ratio; ICD-10 5 International Classification of Diseases-10; VaD 5 vascular dementia. There are conflicting reports as to whether intervention with specific antihypertensive drug classes , or antihypertensive drugs in general, can delay or mitigate dementia risk. 1-4 One mechanism that could explain prior discrepancies in findings are the between-medication class differences in visit-to-visit blood pressure (BP) variability (BPV). Accumulating evidence suggests that visit-to-visit BPV is associated with adverse cardiovascular outcomes independent of mean BP. 5,6 Indeed, a parallel and emerging body of work also corroborates that BPV is associated with cognitive decline 7,8 and incident dementia risk, 9 thus broadly implicating BPV in neurologic outcomes. To date, however, it remains unknown whether the purported links between discrete antihypertensive drug classes and dementia risk remain after controlling for BPV, and each drug's direct effect on BPV. Additional methodologic weaknesses in the extant literature concerning dementia and anti-hypertensive drug use exacerbate the lack of clarity on this subject. Specifically, limitations From Neuroepidemiology, UMR897 (P.J.T.