Antihypertensive Medication Class and Incident Alzheimers Disease and Related Dementias: External Replication Across Two Independent Healthcare Data Sources-the Vanderbilt EHR Synthetic Derivative and US Medicare Claims
Tiwari, A., Chen, R., Tan, Y., Yan, Y., Wen, Z., Wei, W.-Q., Zhong, X., Li, B.
neurology
Quick verdict
Replication: stimulating antihypertensives HR 0.92 vs inhibiting for ADRD in two cohorts; plausible, modest effect, residual confounding limits causality.
Summary
This retrospective pharmacoepidemiologic study replicates, in two independent data sources, the hypothesis that antihypertensive classes stimulating (versus inhibiting) angiotensin II type 2/type 4 receptor signaling carry lower Alzheimer's disease and related dementia (ADRD) risk. Adults 65 or older with hypertension and at least 365 cumulative antihypertensive exposure days were classified by predominant (>=80%) use: stimulating (ARBs, dihydropyridine CCBs, thiazides) versus inhibiting (ACEIs, beta-blockers, nondihydropyridine CCBs), plus a residual group. After propensity-score matching and Cox models with time-dependent covariates, stimulating predominance was associated with lower ADRD hazard versus inhibiting in both cohorts: Medicare MarketScan (n=1.6M; HR 0.92, 95% CI 0.90-0.93) and VUMC Synthetic Derivative EHR (n=80K; HR 0.92, 0.85-0.99). Comparisons with the residual group were null (HR 1.01 and 0.98), suggesting risk elevation among inhibiting users rather than benefit of stimulating agents. Expected covariate associations (atrial fibrillation, heart failure, CKD, depression) support internal validity. Limitations include residual confounding (authors concede the effect size is within the range explainable by unmeasured confounding), a post-match depression imbalance (3.0% vs 3.8% in Medicare), an inconsistently described 1:1 match for the much larger residual group, claims/EHR outcome misclassification, and single-center generalizability for VUMC. Causal claims are appropriately withheld pending randomized trials.
Review complete
Assessment at a glance
5
risk flags identified
● high 0
● medium 3
● low 2
5 strengths · 0 weaknesses · 0 author questions
AI text signal = Binoculars zero-shot detector score over 20 sampled passages (Fast-DetectGPT co-reported) — absence does not prove human authorship, and a flag does not imply misconduct.
Levels reflect verification complexity — not a quality score. Evolve never rates your paper.
Key claims under review
- Predominant use of receptor-stimulating antihypertensives (ARBs, dihydropyridine CCBs, thiazides) is associated with modestly lower incident ADRD risk versus inhibiting agents in both a Medicare claims cohort (HR 0.92, 95% CI 0.90-0.93) and the VUMC SD EHR cohort (HR 0.92, 95% CI 0.85-0.99).
- Concordant estimates across two structurally independent healthcare data environments externally replicate the prior single-database finding by Marcum and colleagues.
- The association appears driven by elevated risk among predominant inhibiting users rather than reduced risk among stimulating users, since comparisons against the residual group are null.
- Expected associations of cardiovascular and neuropsychiatric covariates (atrial fibrillation, heart failure, CKD, depression) with ADRD in both cohorts support the internal validity of the analytic approach.
Strengths
Genuine dual-source external replication
The same adjusted HR (0.92) emerges in MarketScan claims (n=1,596,034) and the VUMC SD EHR (n=80,267) under harmonized exposure, covariate, and outcome definitions, directly addressing the dataset-artifact concern raised for the single-source Marcum 2023 analysis (Discussion).
Methodologically sound core design
New-user, active-comparator cohorts with 365-day exposure accumulation, 365-day blanking against reverse causation, time-updated exposure and covariates, and a prespecified >=80% predominance threshold (Methods).
Honest mechanistic interpretation
Plausibility checks present
Very large samples with tight precision
Full Review Report
$1.99 · one-timeUnlock the complete preliminary analysis Evolve produced for this paper.
Report preview
Scope & Claims
This preprint (medRxiv, posted September 17, 2026; NIH-funded R01AG069900) asks whether antihypertensive classes that 'stimulate' versus 'inhibit' angiotensin II type 2/type 4 receptor signaling differ in incident Alzheimer's disease and related dementias (ADRD) risk among adults 65+ with hypertension. It explicitly replicates Marcum et al. (JAMA Netw Open 2023, ref 26), which used a 5% Medicare sample, in two parallel cohorts: MarketScan Medicare claims 2015-2023 (1,596,034 beneficiaries) and the VUMC Synthetic Derivative EHR (80,267 patients). Three central claims: (1) predominant (>=80% of exposure days) stimulating-class use is associated with lower ADRD hazard versus inhibiting-class use (HR 0.92, 95% CI 0.90-0.93 Medicare; 0.92, 0.85-0.99 VUMC); (2) concordance across structurally different data environments constitutes external replication; (3) the contrast is driven by elevated risk among inhibiting users, not benefit among stimulating users, since comparisons to the residual 'nonuser' group are null (HR 1.01 and 0.98).
The remaining sections cover Methodology Assessment, Statistical Evidence, Results & Reproducibility, Limitations & Risks, and Verdict & Recommendations.
- All 5 strengths & 0 weaknesses, analyzed in detail
- 5 risk flags with severity, section & reviewer questions
- Methods & statistical evidence assessment
- 0 questions a peer reviewer would ask
- Full six-dimension evaluation report & recommendations
One-time purchase · this paper only
One-time purchase, this paper only. AI-generated preliminary analysis — not peer review.
Frequently asked questions
Do ARBs or other 'stimulating' antihypertensives prevent Alzheimer's disease?
This preprint reports an 8% lower hazard of incident Alzheimer's disease and related dementias (HR 0.92) among older hypertensive adults predominantly taking stimulating-class drugs (ARBs, dihydropyridine calcium channel blockers, thiazide diuretics) versus inhibiting-class drugs (ACE inhibitors, beta-blockers, nondihydropyridine CCBs), with concordant results in a 1.6M-person Medicare claims cohort and an 80K-person Vanderbilt EHR cohort. It is observational: it cannot prove causation, and the authors concede the effect size falls within the range explainable by unmeasured confounding.
What data sources were used and how large were the cohorts?
The Merative MarketScan Medicare Database (2015-2023, about 1.6 million beneficiaries) and the Vanderbilt University Medical Center Synthetic Derivative electronic health record (about 80,000 patients), both analyzed with a new-user active-comparator design, propensity-score matching, and roughly 7 years of follow-up.
Is this the first study linking antihypertensive receptor pathways to dementia risk?
No. The paper explicitly replicates Marcum and colleagues (JAMA Network Open, 2023), which reported the same receptor-pathway association in a 5% random Medicare sample. The new contribution is external validation in a structurally different EHR environment plus a roughly 20-fold larger contemporary claims cohort.
Should patients switch blood pressure medications based on this paper?
No. This is a non-peer-reviewed preprint of an observational study with a modest effect size and acknowledged residual-confounding limits. The authors themselves call for randomized comparative-effectiveness trials before any change to prescribing practice.
About this review
This is an AI-generated preliminary review by Evolve.Pub, produced from the openly available text of the paper. It is not peer review and does not represent the authors' views. The canonical record of the paper is at the publisher link above.
- Text extraction garbles portions of the Results narrative (scrambled fragments around Figure 2); key numbers remain readable but some sentences were reconstructed from fragments.
- Supplement 1 (eTables 1-2: ICD and drug code lists) and Figures 1-2 are not analyzable from the provided text.
- Matching algorithm details for the third (nonuser) group and proportional-hazards diagnostics are not visible in the main text; judgments there rest on reported numbers only.
- Preprint not peer reviewed; all figures quoted as printed, without independent verification.
Go deeper on this paper
Wrote a paper? Stress-test it before reviewers do.
Upload your manuscript for the same structured scrutiny — reasoning, statistics, evidence, figures, and reproducibility.