Multimorbidity patterns and cognitive trajectories in older adults: functional, behavioral, and environmental pathways—a longitudinal analysis of CHARLS

Abstract Multimorbidity may affect cognitive aging differently according to disease patterns rather than disease counts alone. This study analyzed 2011–2020 data from the China Health and Retirement Longitudinal Study, including 5979 Chinese adults aged 45 years or older. Fifteen chronic conditions were used to identify multimorbidity patterns through latent class analysis. Latent growth curve models examined associations with baseline cognition and cognitive decline, while mediation and moderation analyses assessed activities of daily living, social and intellectual engagement, living environment quality, and adverse childhood experiences. Five patterns were identified: minimal morbidity, sensory-psychiatric-musculoskeletal multimorbidity, cardiometabolic multimorbidity, extensive multisystem multimorbidity, and respiratory-psychiatric-sensory multimorbidity. In fully adjusted models, sensory-psychiatric-musculoskeletal multimorbidity (β = −0.757; 95% CI, − 0.92 to − 0.59) and extensive multisystem multimorbidity (β = −0.469; 95% CI, − 0.73 to − 0.21) were associated with lower baseline cognition. Both showed slower subsequent decline, suggesting a floor effect. Time-varying models showed faster cognitive decline in the cardiometabolic group (β = −0.028; 95% CI, − 0.052 to − 0.005; P = 0.019). Activities of daily living mediated associations across nonreference patterns, whereas social and intellectual engagement mediated only the sensory-psychiatric-musculoskeletal pattern. Poor living environment quality was associated with lower baseline cognition (β = −0.532; 95% CI, − 0.849 to − 0.214; P = 0.002). These findings support phenotype-aware cognitive risk screening and interventions targeting function, engagement, and living environments.