Author(s): GONZÁLEZ MARTOS, R., RODRÍGUEZ-GÓMEZ, I., RAMÍREZ-CASTILLEJO, C., GALEANO, J., ARA, I, ALEGRE, L.M., RODRÍGUEZ-MAÑAS, L., GARCÍA-GARCÍA, F.J., GUADALUPE-GRAU, A., Institution: UNIVERSIDAD POLITÉCNICA DE MADRID, Country: SPAIN, Abstract-ID: 1824

INTRODUCTION:
Aging is associated with health issues leading to functional decline and disability. Identifying key factors (biochemical, physical, or social) for successful aging is crucial for resource efficiency [1]. Using unsupervised clustering analysis on routine laboratory test data, we aim to uncover biochemical phenotypes in the spectrum of old age. This approach may enhance understanding of mechanisms influencing healthy aging. We also intend to explore associations between these phenotypes and key health outcomes, including mortality [2].
METHODS:
A final sample of n=1491 participants (65 to 99 years old, 57.7% women), from the Toledo Study for Healthy Aging (TSHA), was included. Hierarchical and k-means cluster analyses (Orange software) incorporated 39 variables related to white and red blood cells, and renal, liver lipid, and glucose metabolism. Health outcomes included body composition (DXA), physical functionality (SPPB test), physical activity (PASE questionnaire), disability (Katz questionnaire), frailty (FTS-5 test), comorbidity (medical history), and 12-year mortality (Spanish National Mortality Database). The associations between cluster membership and health outcomes and mortality hazard ratios (HR) were analyzed through general linear models (ANCOVA, SPSS software) and Cox regression (STATA software) respectively, including sex, age, and socioeconomic status as covariates.
RESULTS:
Three different clusters were identified: Cluster 1 (41.8%, Healthy), all biochemical values within the normal range, taken as reference for statistical analysis; Cluster 2 (44.5%, Metabolic), characterized by high but sub-clinically relevant levels of glucose, triglycerides, HOMA, and liver enzymes; and Cluster 3 (13.7%, Red Blood Cells), defined by lower levels of hematocrit, hemoglobin, and red blood cells. Our findings revealed that Cluster 2 (metabolic) showed a significantly higher BMI in men (28.70±0.21 vs. 27.46±0.35 Kg/m2) and women (31.01±0.35 vs. 29.61±0.26 Kg/m2), and a higher risk of death in women (HR=1.50; 95%CI= (1.08; 2.08)) respect to the healthy cluster. Cluster 3 (Red Blood Cells) participants were older (77.8±0.4 vs. 74.5±0.2 yrs.), and men were less physically active (PASE score: 57.46±6.03 vs 69.40±4.41) and showed worse physical functionality with less gait speed (0.85±0.04 vs 0.94±0.03 m/s) than healthy cluster. They were also less independent in their daily activities with a significantly lower Katz index in both sexes (men: 5.46±0.08 vs. 5.82±0.06; women: 5.56±0.07 vs. 5.77±0.04). Significant findings (p<0.05) were observed for all mentioned values.
CONCLUSION:
Our findings suggest that a biochemical phenotype characterized by impaired red blood cell function is associated with a decreased health-related quality of life, whereas impaired glucose and lipid metabolism is associated with a higher risk of death. This information could be helpful to develop exercise-based interventions oriented to healthy aging.
1. Urtamo et al. (2019) 2. O ‘Sullivan et al. (2011)