Summary
A retrospective study using US health records found that dementia-risk associations varied among individual angiotensin receptor blockers (ARBs). The results are observational and the authors say randomized trials are needed for confirmation.
A retrospective study of electronic health records and claims data found that dementia-risk associations differed among individual angiotensin receptor blockers (ARBs). The medRxiv preprint compared the drugs with lisinopril, an ACE inhibitor, and also compared non-losartan ARBs with losartan. Its results point to variation within the drug class, but do not establish that one ARB prevents dementia better than another.
How the study compared treatments
Researchers used data from 44 US health care organizations in the TriNetX Linked Research Network. They screened 22,025,599 people and included 2,051,275 adults aged 55 to 80 who had not previously used an ACE inhibitor or ARB and were starting an ARB or lisinopril.
The team used a retrospective cohort design to emulate target trials: they structured comparisons in existing health records to resemble trials of different treatment choices. Separate 1:1 propensity-score matching was performed for each comparison. The primary outcome was newly recorded all-cause dementia. Secondary outcomes included Alzheimer disease, vascular dementia, cerebral infarction, death, and dementia or death. Median follow-up across the matched cohorts ranged from 4.7 to 7.2 years.
Dementia associations differed across drugs
When ARBs were pooled together, their use was associated with a lower dementia hazard than lisinopril use (hazard ratio 0.86; 95% confidence interval 0.85 to 0.87). A hazard ratio compares the rate at which an outcome occurs over time between groups; it is not an estimate of the absolute number of people who will develop dementia.
Estimates varied by individual ARB, with hazard ratios ranging from 0.66 to 0.93 in comparisons with lisinopril. Comparisons of non-losartan ARBs with losartan also showed variation, with estimates ranging from 0.68 to 0.94. Losartan, the most commonly prescribed ARB in the study, had the least favorable dementia associations. The authors also reported variation across secondary outcomes.
The researchers say negative-control associations within the ARB class were close to null and that sensitivity analyses supported the findings. These checks can help assess the stability of observational results, but treatment was not randomly assigned. Differences in prescribing and other factors could still influence the associations.
The study is a medRxiv preprint and its findings are hypothesis-generating. They are not a basis for choosing or switching medicines; the authors say the differences need confirmation in randomized trials.