Cross-system patterns of physical, psychological and cognitive multimorbidity and 9-year all-cause mortality among middle-aged and older Chinese adults: a prospective cohort study from the China Health and Retirement Longitudinal Study

Abstract

Cross-system multimorbidity—the co-occurrence of physical, psychological and cognitive disorders—is increasingly recognised as a defining feature of population ageing, yet its relationship with long-term mortality has rarely been examined at the level of specific combinations. Most studies have used chronic-disease counts or binary indicators that obscure cross-system patterning. We assessed the association of cross-system multimorbidity with 9-year all-cause mortality and identified high-risk patterns in a national sample of middle-aged and older Chinese adults. We analysed 16,479 adults aged 45 years or older from the China Health and Retirement Longitudinal Study (CHARLS), enrolled at the 2011 baseline and followed through 2020. Three system indicators were defined: physical multimorbidity (≥ 2 of 14 self-reported physician-diagnosed chronic conditions), depression (CES-D-10 score ≥ 10), and cognitive impairment (lowest quartile of the CHARLS cognitive composite). The exposure was operationalised both as the number of involved systems (0/1/2/3) and as eight combination patterns (000–111). Missing exposure and covariate data were handled with multiple imputation (50 imputed datasets), and participants with incomplete follow-up were censored at their last known contact rather than excluded. Pooled Cox proportional hazards models were fitted with progressive adjustment for demographic, socioeconomic, and lifestyle/biomarker covariates. Over a mean follow-up of 7.96 years (SD 2.23), 1,033 participants died (6.27%). After full adjustment, each additional involved system was associated with a 34% higher hazard of all-cause mortality (HR 1.34, 95% CI 1.24–1.44; p < 0.001); compared with no involved systems, hazards were 1.34 (95% CI 1.09–1.66), 1.68 (1.35–2.09) and 2.48 (1.93–3.17) for one, two and three systems, respectively. In pattern-level analyses, co-occurring physical multimorbidity and cognitive impairment yielded a hazard close to all three systems combined (HR 2.14, 95% CI 1.57–2.91 vs HR 2.45, 95% CI 1.89–3.17; ratio of hazard ratios 0.87, 95% CI 0.66–1.16), whereas isolated depression showed no clear independent association (HR 1.12, 95% CI 0.81–1.57). Findings were consistent across complete-case, fully observed exposure, alternative cognitive definitions, discrete-time and sleep-adjusted analyses. Cross-system multimorbidity exhibited a graded dose–response association with 9-year all-cause mortality among middle-aged and older Chinese adults, and the co-occurrence of physical multimorbidity and cognitive impairment alone reached a hazard close to that of triple-system involvement. This pattern identifies a group that may merit closer clinical attention, although the model was not evaluated for predictive performance, calibration or external validity and is not intended for individual risk stratification.