Abstract
OBJECTIVE: The aim of this study was to investigate the independent and combined effects of oxidative balance score (OBS) and composite dietary antioxidant index (CDAI) on the risk of frailty in patients with early cardiovascular-kidney-metabolic (CKM) and to evaluate their cumulative predictive value. METHODS: Publicly available data from National Health and Nutrition Examination Survey (NHANES) 2007-2018 were used, including 10,201 participants aged 20 years or older who met the criteria for early-stage CKM. Multivariable logistic regression was employed to examine the independent associations of OBS and CDAI with frailty risk. RCS models were used to analyze non-linear relationships and threshold effects. Stepwise logistic regression, the DeLong test, and Net reclassification improvement/Integrated discrimination improvement (NRI/IDI) metrics were applied to compare the predictive performance and cumulative effects of the two indices. RESULTS: Among the 10,201 early-stage CKM patients (mean age 46.30 [0.30] years; 51.40% female), multivariable logistic regression revealed that after adjusting for potential confounders, Quartile-based analyses indicated clear dose-response relationships: compared to the lowest quartile, the highest quartile of OBS was associated with a 54% lower risk of frailty (OR = 0.46, 95% CI: 0.36-0.60), and the highest quartile of CDAI was associated with a 39% lower risk (OR = 0.61, 95% CI: 0.50-0.74). Restricted cubic spline (RCS) analyses demonstrated significant non-linear associations for both OBS and CDAI with frailty risk (P for non-linearity = 0.002 and <0.001, respectively). An OBS threshold of approximately 9.05 was identified, beyond which frailty risk declined substantially. For CDAI, the threshold was about -2.39, below which frailty risk dropped markedly, while above this threshold, the risk plateaued with little further reduction. Cumulative effect analysis showed that the combined OBS and CDAI model (AUC = 0.577) did not offer a significant improvement over models including OBS alone (AUC = 0.577) or CDAI alone (AUC = 0.565), as indicated by DeLong test results (all p > 0.05). Additional analyses using continuous NRI and IDI metrics further confirmed the lack of significant additive effect when combining the two indices. CONCLUSION: This study found that both the OBS and the CDAI independently serve as protective factors against frailty in patients with early-stage CKM, each displaying a non-linear inverse association with clearly defined threshold effects.