Peters · Dementia and geriatric cognitive disorders 2021 · individual participant data meta-analysis of cohort studies · n=21,000+ participants across 8 studies

Evaluation of High Cholesterol and Risk of Dementia and Cognitive Decline in Older Adults Using Individual Patient Meta-Analysis.

Cited 52 times in the scientific literature.

Level 3 - non-randomized controlled study

Individual participant data meta-analysis of observational cohort studies

PubMed 34700321 · doi:10.1159/000519452 · record verified 2026-08-28

What was done

A two-stage individual participant data (IPD) meta-analysis harmonized data from 8 studies including over 21,000 participants aged 60 years and older (mean age 76 years). The authors evaluated associations between total, HDL, and LDL cholesterol (per mmol/L increase) and the risk of subsequent incident dementia or cognitive decline (assessed via a reliable change index method). Analyses were also stratified by statin use and APOE ε4 carrier status.

What was found

No relationship was found between cholesterol levels and cognitive outcomes: - Total cholesterol and cognitive decline: OR 0.93 (95% CI 0.86 to 1.01). - Total cholesterol and incident dementia: OR 1.01 (95% CI 0.89 to 1.13). - HDL and LDL cholesterol also showed no relationship with risk of cognitive decline (point estimates not provided in the abstract). - Stratification by statin use or APOE ε4 carrier status did not alter the results.

Why it matters

This IPD meta-analysis demonstrates that cholesterol levels measured in late life do not reliably predict subsequent cognitive decline or incident dementia, suggesting that late-life lipid profiles alone are not informative prognostic targets for dementia risk.

Limits

The study is restricted to older adults (mean age 76), precluding assessment of mid-life cholesterol exposure. As a synthesis of observational studies, residual confounding cannot be excluded. Specific risk estimates and confidence intervals for HDL and LDL cholesterol, as well as exact follow-up durations, were not reported in the abstract.

Cited by