Exploratory Factor Analysis Defined Allostatic Load Indices and Cognition Among Older Adults: Variation Across Race and Sex
Abstract
Objective: Allostatic load (AL), a measure of cumulative biological dysregulation, is associated with negative cognitive outcomes. This study employed a data-driven approach to construct AL indices from multiple biomarkers and examined their associations with cognition among older adults, as well as whether these associations vary across sex and race. Method/Data Selection: Participants (N = 321; mean age = 69.18 years; 63.60% female; 57.00% Black) were from the University of Alabama at Birmingham Alzheimer's Disease Research Center Cohort. Using Exploratory Factor Analysis (EFA), AL indices were constructed from 14 biomarkers. A cognitive composite z-score was derived from 12 neuropsychological tests (Unified Data Set). Pearson correlations between AL indices and cognition were stratified by sex and race. Results/Data Synthesis: The EFA yielded a 5-factor solution (77.88% variance) representing nutritional deficit, adiposity, renal, glycemic, and metabolic burden. Among females, renal (r=−.20, p< 0.01), glycemic (r=−.17, p< 0.05), and metabolic (r=−.14, p< 0.05) burden correlated negatively with cognition, with no significant associations among males. Among White participants, glycemic (r=−.33, p< 0.001) and metabolic (r=−.20, p< 0.05) burden correlated negatively with cognition, whereas adiposity correlated positively among Black participants (r=.18, p< 0.05). Conclusions: AL indices were differentially associated with cognition across sex and race, suggesting physiological contributors to cognitive health may differ across demographic groups. Female and White participants demonstrated greater vulnerability, particularly through glycemic and metabolic burden. While the positive association between adiposity and cognition among Black participants supports the “obesity paradox”, further investigation is warranted. Replication in larger, longitudinal samples is needed before clinical application.