Five central adiposity-based anthropometric indices and cognitive impairment in elderly populations: Development and validation of a risk prediction nomogram using NHANES and CHARLS cohort

BackgroundCognitive impairment is a primary contributor to disability among older adults, with growing evidence identifying central adiposity as an adjustable risk factor for neurodegeneration. This study aimed to establish and verify a model for predicting mild cognitive impairment (MCI) risk in aging individuals by incorporating central adiposity indices. MethodsWe calculated five central adiposity indices from anthropometric measurements using data from 2,464 United States adults aged [&ge;] 60 (National Health and Nutrition Examination Survey 2011-2014). Cognitive performance was assessed using three standardized neuropsychological tests. Multivariable logistic regression, incorporating restricted cubic splines to test for nonlinearity, was used to evaluate associations between central obesity measures and MCI, complemented by sensitivity analyses (subgroup stratification and multiple imputation). A random assignment placed participants into either a training (n = 1725) or a internal validation (n = 739) set. Furthermore, the data from participants in the 2011 wave of the China Health and Retirement Longitudinal Study served as an external validation cohort (n = 536). Least absolute shrinkage and selection operator-selected predictors were employed to inform multivariable logistic regression modeling. Assessment of the nomograms performance involved the area under the curve (AUC), calibration curve, decision curve analysis, clinical impact curve, and external validation. ResultsPositive linear relationships were found between three anthropometric indices of central obesity--a body shape index (ABSI), conicity index (CoI) and weight-adjusted-waist index (WWI)--with MCI risk (P < 0.05, P for nonlinearity > 0.05). Ten features were screened as MCI predictors, including ABSI, age, educational level, race, poverty income ratio, stroke, depression, diabetes mellitus, physical activity, and gender. The nomogram demonstrated strong discriminative capacity (training AUC = 0.861; internal validation AUC = 0.826; external validation AUC = 0.798), precise calibration, and good clinical utility. ConclusionThe risk of MCI was independently linked to central adiposity indices (ABSI, WWI, and CoI). The nomogram incorporating ABSI provided a validated, clinically applicable prediction model for initial screening of MCI in older populations.