A Low-Data-Burden Vascular-Behavioral Profile for First-Pass Stratification of Later-Life Cognitive Decline Across International Aging Cohorts.
Where this comes from
- Record sourced from PubMed, PMID 42503365.
- Also identified by DOI 10.1016/j.amepre.2026.108528.
- No licence information is recorded for this record.
- Because redistribution is not established, this page shows the abstract only. Follow the links below for the full text.
Abstract
Cognitive decline is a major prevention priority, but scalable approaches are needed to identify adults who may benefit from proportionate cognitive follow-up and vascular-behavioral prevention support. This study evaluated whether a low-data-burden vascular-behavioral profile supports first-pass cognitive-decline risk ordering across aging cohorts and where local recalibration is needed. Repeated cognition data from 2000-2025 were analyzed in UK Biobank, CHARLS, ELSA, HRS, NHATS, KLOSA, and CLHLS. Analyses were conducted in 2026. The 5-factor profile included hypertension, diabetes, smoking, physical inactivity, and obesity. Linear mixed models tested profile-by-time associations; random-slope models were fitted in major cohorts. A UK Biobank logistic reference model evaluated substantial decline, defined as a 5-year-equivalent decrease of at least 0.5 cohort-standardized SD. Discrimination and calibration were assessed. Higher baseline burden was associated with faster global cognitive decline in UK Biobank, CHARLS, ELSA, and HRS, with similar random-slope estimates. For substantial decline, AUCs were 0.576 in UK Biobank, 0.603 in CHARLS, 0.666 in ELSA, and 0.643 in HRS. Absolute risk was underpredicted in external cohorts; observed versus predicted risks were 22.3% versus 12.8%, 19.4% versus 13.0%, and 20.9% versus 13.6% in CHARLS, ELSA, and HRS, respectively. The vascular-only score showed lower heterogeneity (I<sup>2</sup>=53.4%) than the behavior-only score (I<sup>2</sup>=88.4%). This simple profile supports low-data-burden first-pass cognitive-decline risk ordering. It may help prioritize higher-burden adults for proportionate follow-up and prevention support, with local recalibration required before absolute probabilities guide individual decisions.