Across 2,051,275 eligible adults aged 55–80 in the TriNetX network, starting an angiotensin receptor blocker (ARB) rather than lisinopril was linked to 14% lower incident dementia (HR 0.86, 95% CI 0.85–0.87) over a median 4.7–7.2 years. Individual agents diverged sharply, with hazard ratios from 0.66 to 0.93. Head-to-head comparisons against losartan reproduced the same ranking, and losartan, the most prescribed ARB, showed the least favorable dementia signal.
The class-level ARB-versus-ACE-inhibitor advantage has appeared in earlier observational work; the new element is within-class heterogeneity, which would matter if some ARBs cross the blood-brain barrier or engage brain angiotensin signaling differently. Still, this is a retrospective electronic-records emulation, and propensity matching cannot remove confounding by indication, such as prescribers steering healthier or higher-income patients toward newer, costlier agents. Dementia diagnoses recorded in claims also miss undetected cases. Reassuringly, negative-control outcomes were near null within class and E-values were reported, but the wide spread between agents is hard to explain biologically and may partly reflect residual bias. This is a preprint, not yet peer reviewed, so figures may change. Nobody should switch blood-pressure medication on this basis; blood pressure control itself protects the brain, and any change belongs in a conversation with a physician. Treat the findings as hypothesis-generating and promising for randomized testing, not as practice-changing.