Multidimensional Factors Associated With Depressive Symptoms in Colombian Centenarians.

OBJECTIVES: To identify multidimensional factors associated with depressive symptoms in Colombian centenarians and to compare the relative contribution of functional, cognitive, psychosocial, and nutritional domains to these symptoms. DESIGN, SETTING, PARTICIPANTS: Cross-sectional analytical study. Community-based centenarian cohort in Colombia, in which a total of 152 adults aged 100 years or older with comprehensive geriatric assessment were included. MEASUREMENTS: Positive depressive symptoms were defined as a score ≥6 on the 15-item Geriatric Depression Scale. Twelve candidate variables from functional, cognitive, psychosocial, and nutritional domains were examined. Variable selection was performed using least absolute shrinkage and selection operator (LASSO) regression, followed by domain-specific and combined multivariable logistic models. Model performance was assessed by discrimination, calibration, collinearity diagnostics, influence analysis, and sensitivity analyses using multiple imputation. RESULTS: The prevalence of depressive symptoms was 15.8% (n = 24/152). In bivariate analyses, clinically notable differences include higher sarcopenia prevalence, worse cognitive scores, lower life satisfaction, lower quality of life, loneliness, insomnia, anxiety, worse self-perception, and poorer nutritional status (p <0.05). Across modeling strategies, the psychosocial domain showed the strongest discriminative performance. In the final combined model, less overall life satisfaction and loneliness remained associated with depressive symptoms. The final model demonstrated good apparent discrimination (AUC = 0.836). CONCLUSIONS: In Colombian centenarians, depressive symptoms were primarily associated with lower life satisfaction and perceived loneliness. Future studies should determine the temporal relationships among psychosocial, cognitive, and functional domains, evaluate causal pathways underlying depressive symptoms, and establish whether these associations are consistent across diverse populations of individuals with extreme longevity.