The relationship between the CUN-BAE body fatness index and incident diabetes: a longitudinal retrospective study
Qing Peng, Zihao Feng, Zhuojian Cai, Dixing Liu, Jiana Zhong, Hejia Zhao, Xiuwei Zhang, Weikun Chen
Abstract
BACKGROUND: The Clínica Universidad de Navarra-Body Adiposity Estimator (CUN-BAE) index has been recommended as an ideal indicator of body fat and exhibited significant correlation with cardiometabolic risk factors. However, whether the CUN-BAE index correlates with incident diabetes in Asian populations is unknown. Therefore, this longitudinal study was designed to evaluate the association between baseline CUN-BAE index and type 2 diabetes mellitus (T2DM). METHODS: This retrospective longitudinal study involved 15,464 participants of 18-79 years of age in the NAGALA (NAfld in the Gifu Area Longitudinal Analysis) study over the period of 2004-2015. Cox proportional hazards regression was performed to test the relationship between the baseline CUN-BAE index and diabetes incidence. Further stratification analysis was conducted to ensure that the results were robust. The diagnostic utility of the CUN-BAE index was tested by the receiver operating characteristic (ROC) curve. RESULTS: Over the course of an average follow-up of 5.4 years, 373 (2.41%) participants developed diabetes. A higher diabetes incidence was associated with higher CUN-BAE quartiles (P for trend< 0.001). Each 1 unit increase in CUN-BAE index was associated with a 1.08-fold and 1.14-fold increased risk of diabetes after adjustment for confounders in males and females, respectively (both P < 0.001). Stratification analysis demonstrated a consistent positive correlation between baseline CUN-BAE and diabetes incidence. Moreover, based on ROC analysis, CUN-BAE exhibited a better capacity for diabetes prediction than both body mass index (BMI) and waist circumference (WC) in both sexes. CONCLUSIONS: The baseline CUN-BAE level was independently related to the incidence of diabetes. Increased adiposity determined by CUN-BAE could be used as a strong nonlaboratory predictor of incident diabetes in clinical practice.