Disease-specific cognitive impacts in aging: a large-scale longitudinal analysis of chronic conditions and demographic modifiers

Cognitive decline represents a major public health concern among aging populations, yet the domain-specific impacts of chronic diseases remain poorly understood. This study examined associations between chronic diseases and specific cognitive domains, exploring heterogeneity across population subgroups. We analyzed longitudinal data from the China Health and Retirement Longitudinal Study across three waves (2015, 2018, 2020; ~2–3-year intervals), comprising 16,027 participants aged 45–90 years (35,247 observations). Continuous cognitive outcomes (orientation, memory, computation, and total score) were standardized (z-scores), so β represents the SD-unit difference in cognitive score for disease presence versus absence; drawing ability, a binary outcome, was analyzed via logistic regression and reported as an odds ratio. These outcomes were modeled against 14 chronic conditions using mixed-effects models adjusted for demographic, lifestyle, and health covariates and for co-occurring conditions. P-values were corrected for multiple comparisons (56 tests), and effect modification by education and sex was tested formally. Nine of fourteen conditions remained significant after correction. Memory-related disease showed the broadest impairment, affecting total cognition (β=−0.292, P < 0.001), orientation (β=−0.326), and memory (β=−0.190). Chronic lung disease (β=−0.062) and emotional/psychiatric disorders (β=−0.158) were also negatively associated with cognition, as was stroke (β=−0.127). Heart disease (β=+0.109) and liver disease (β=+0.087) showed unexpected positive associations, which we interpret cautiously given the absence of clinical severity data. Education significantly modified the associations for stroke, chronic lung disease, and emotional/psychiatric disorders (all P < 0.05), with greater vulnerability among more-educated participants; no disease-by-sex interaction was significant. Chronic diseases show distinct, domain-specific cognitive impacts, with education—but not sex—modifying a subset of these associations. These findings support disease-specific rather than uniform approaches to cognitive risk assessment in aging populations.