Higher intrinsic capacity linked to lower 3-year dementia risk in older Mexican adults

Dementia affects roughly 57 million people worldwide, and low- and middle-income countries face the steepest projected growth. Mexico has an estimated dementia prevalence of about 6% in adults 60 and older. Practical, low-cost tools to identify people at risk are needed. The WHO's intrinsic capacity (IC) framework captures the sum of an older person's physical and mental reserves across five domains: vitality, locomotion, sensory function, cognition and psychological well-being. IC has been linked to frailty, disability and mortality, but evidence on dementia is thin. Before this study, only a UK Biobank analysis had examined it, and no study had in a Latin American population.

The researchers analyzed data from the Mexican Health and Aging Study, a nationally representative cohort. They used the 2012 wave as baseline and the 2015 wave for follow-up. After excluding people under 60, proxy interviews, incomplete cognitive testing and baseline dementia, and accounting for deaths and attrition, 7,263 participants remained (mean age 68.9, 55.4% women). IC was built from proxy survey items using factor analysis and structural equation modeling. These included depressive symptoms, self-reported vision and hearing, weight and appetite loss, orientation and word recall, and mobility limitations. The result was a standardized continuous score. Dementia was defined as performance 1.5 SD or more below norms in two or more cognitive domains, plus at least one instrumental activity of daily living limitation. All analyses used survey weights, and logistic models were adjusted for age, sex, marital status, education, multimorbidity, smoking and alcohol use.

Over three years, 283 people developed dementia, a cumulative incidence of 4.15%, or roughly 14.2 cases per 1,000 person-years. Mean IC fell significantly (from −0.30 to −0.53), and impairment rose in every domain. Participants in the lowest IC tertile were older, more often female, and had more hypertension, diabetes and multimorbidity. Higher baseline IC was associated with lower odds of incident dementia (fully adjusted OR 0.62, 95% CI 0.41–0.94). Baseline cognitive status modified the association, so models were stratified. Among cognitively unimpaired participants the association was significant (age- and sex-adjusted OR 0.89; fully adjusted OR 0.62). Among those with baseline cognitive impairment it was not (OR 0.88, 95% CI 0.68–1.14). Of the individual domains, only locomotor capacity independently predicted lower dementia odds (OR 0.79, 95% CI 0.66–0.95). Excluding the cognitive domain, non-cognitive IC still predicted lower dementia risk overall (OR 0.71).

The results support IC as a multidimensional marker of cognitive vulnerability that may be useful where specialist assessment is scarce. They also hint that mobility is an especially informative component and a possible target for intervention.

Several caveats apply. This is an observational study, so causality cannot be inferred. Follow-up was only three years, and many people died or were lost to follow-up, with those lost generally being frailer. IC components were crude self-reported proxies rather than objective tests. The cognitive domain overlaps conceptually with the dementia outcome. Some results were inconsistent: the stratified non-cognitive analysis showed the opposite pattern by baseline cognition, and some table values appear internally inconsistent. Whether improving IC lowers dementia risk remains untested.