Wu · Frontiers in endocrinology 2026 · systematic review and meta-analysis · n=11 studies (3,393,545 participants)

Association between diabetes mellitus and risk of Alzheimer's disease: a meta-analysis and systematic review.

Cited 3 times in the scientific literature.

Level 3 - non-randomized controlled study

Systematic review and meta-analysis of observational cohort and case-control studies

PubMed 41852471 · doi:10.3389/fendo.2026.1736410 · record verified 2026-08-29

What was done

A systematic review and meta-analysis searching PubMed, Web of Science, and Embase through October 2025 for cohort and case-control studies investigating diabetes mellitus (DM) and Alzheimer's disease (AD) risk. Risk of bias was evaluated using the Newcastle-Ottawa Scale. Pooled hazard ratios (HR) and 95% confidence intervals (CI) were calculated with RevMan 5.3, and heterogeneity was assessed using Cochran's Q and I² statistics.

What was found

Across 11 studies with 3,393,545 participants, DM was associated with increased AD risk overall (HR = 1.36, 95% CI: 1.19–1.55, P < 0.00001). Risk increases were reported in subgroups by study sample size (<100,000: HR = 1.33, 95% CI: 1.11–1.59; >100,000: HR = 1.39, 95% CI: 1.13–1.71) and geographic background (Asian: HR = 1.45, 95% CI: 1.20–1.76; non-Asian: HR = 1.29, 95% CI: 1.13–1.48). However, in studies adjusting for APOE ε4 mutations, the association was not statistically significant (HR = 1.07, 95% CI: 0.97–1.19, P = 0.177), whereas unadjusted studies showed a persistent association (HR = 1.42, 95% CI: 1.23–1.64, P < 0.00001).

Why it matters

This review demonstrates that while diabetes is associated with higher Alzheimer's disease incidence in aggregate observational data, genetic confounding by APOE ε4 status may account for much of the observed relationship.

Limits

The analysis relies entirely on observational cohort and case-control studies subject to residual confounding. The authors noted high heterogeneity across studies. Adjustment for APOE ε4 eliminated statistical significance, pointing to potential genetic confounding. The abstract does not provide details on diabetes subtypes (type 1 vs type 2), disease duration, glycemic control, or concurrent medications.

Cited by