Abstract
OBJECTIVE: Biomarker-specific interventions (e.g., for dementia) will necessitate an individualized approach to treatment. We have constructed a psychometric classifier to identify persons adversely impacted by plasma adipokines. METHOD: The subjects (N = 1,737) of the Alzheimer's Disease Neuroimaging Initiative were assigned to groups "afflicted" by versus "resilient" against the unique effect of plasma adipokines using a classifier derived by confirmatory factor analysis in a structural equation model framework. The impact of affliction class above and beyond observed biomarker levels and covariates was tested by multivariate regression using CDR "Sum of Boxes" as the dependent variable. The affliction class' moderation of adipokines' effect was tested by chi-square difference. The effect of affliction class on prospective conversion risk was tested by Cox's proportional hazards models. RESULTS: Seven hundred four out of the 1,737 subjects (40.53%) were assigned to the afflicted class. The afflicted subjects had greater dementia severity, lower (adverse) Adipokines factor composite scores (by analysis of variance, F(1, 1,735) 2619.68, p < .001) and higher observed levels of plasma adipokines (by Tukey's honestly significant difference test, all p < .001). Adipokines' association with dementia severity was moderated by affliction class. The effect persisted at 48 months. Afflicted cases were more likely to convert to Alzheimer's disease in that timeframe, by Cox's F: F(234, 286) = 3.89, p < .001. CONCLUSIONS: Our approach could guide precision interventions against specific biomarkers. This classifier could be administered by telephone, making class assignment feasible without direct patient contact or biomarker assessment. (PsycInfo Database Record (c) 2025 APA, all rights reserved).