Key points are not available for this paper at this time.
Background Depression in middle-aged and older adults with heart disease is clinically heterogeneous, but this heterogeneity is often assessed using total symptom scores or symptom-only approaches. We examined whether multimorbidity burden, caregiving status, and sex were associated with a joint depressive symptom-biomarker network and whether the main patterns were reproducible in an independent hospital cohort. Methods We analyzed 1,685 participants with heart disease from the China Health and Retirement Longitudinal Study as a discovery cohort and 506 patients from an independent hospital cohort for external replication. We estimated a joint depressive symptom-biomarker network and examined node-specific associations of the three external factors, as well as subgroup differences in network structure and global connectivity. Results The external factors were associated with heterogeneous local patterns in the joint symptom-biomarker network, whereas significant subgroup differences in overall network structure or global connectivity were not observed. The factor-specific patterns were broadest across domains for multi-morbidity burden, weaker and more localized for caregiving status, and clearest and most reproducible for sex, particularly involving high-density lipoprotein cholesterol, triglycerides, and cystatin C. Across both cohorts, depressed mood consistently showed the highest centrality among symptom nodes. Conclusions In middle-aged and older adults with heart disease, depressive heterogeneity was reflected more clearly in reproducible local symptom-biomarker associations than in broad network-wide differences. These findings support an assessment-oriented approach that integrates external factors and routinely available biomarkers when characterizing depressive heterogeneity in this population.
Shi et al. (Tue,) studied this question.