Abstract
We evaluated whether drugs approved for other indications that also target metabolic drivers of Alzheimer's disease and related dementia (ADRD) pathogenesis are associated with delayed onset of ADRD. Using routinely collected healthcare data from two population-based data sources from the US (Medicare) and UK (CPRD), we conducted active comparator, new-user cohort studies. Four alternate analytic and design specifications were implemented: (1) an as-treated follow-up approach, (2) an as-started follow-up approach incorporating a 6-month induction period, (3) incorporating a 6-month symptom to diagnosis period to account for misclassification of ADRD onset, and (4) identifying ADRD through symptomatic prescriptions and diagnosis codes. Of the 10 drug pairs evaluated, hydrochlorothiazide vs. dihydropyridine CCBs showed meaningful reductions in 3 out of 4 analyses that addressed specific biases including informative censoring, reverse causality, and outcome misclassification (pooled hazard ratios [95% confidence intervals] across Medicare and CPRD: 0.81 [0.75-0.88] in Analysis 1, 0.98 [0.92-1.06] in Analysis 2, 0.83 [0.75-0.91] in Analysis 3, 0.75 [0.65-0.85] in Analysis 4). Amiloride vs. triamterene, although less precise, also suggested a potential reduction in risk in 3 out of 4 analyses (0.86 [0.66-1.11] in Analysis 1, 0.98 [0.79-1.23] in Analysis 2, 0.74 [0.54-1.00] in Analysis 3, 0.61 [0.36-1.05] in Analysis 4). Other analyses suggested likely no major differences in risk (probenecid, salbutamol, montelukast, propranolol/carvedilol, and anastrozole) or had limited precision precluding a definitive conclusion (semaglutide, ciloztozol, levetiracetam). Future replication studies should be considered to validate our findings.