Association of fat-to-muscle ratio with diabetic kidney disease: A nationwide NHANES analysis and real-world validation

Xiaohong Zhang^1,2Jiaqun Lin^1,2Wenfeng Wang^1,2Zishan Lin^1,2Yuan Fang^1,2Yanfang Xu^1,2Jianxin Wan^1,2*

• ¹Department of Nephrology, Blood Purification Research Center，Research Center for Metabolic Chronic Kidney Disease, First Affiliated Hospital of Fujian Medical University, Fuzhou, China
• ²Department of Nephrology, National Regional Medical Center, Binhai Campus, The First Affiliated Hospital of Fujian Medical University, Fuzhou, China

Background: Obesity and abnormal body composition are recognized contributors to diabetic kidney disease (DKD) development. The fat-to-muscle mass ratio (FMR), an indicator of body composition, remains insufficiently studied in relation to DKD risk. Methods: This study was a nationwide cohort analysis utilizing data from eight NHANES cycles. FMR was derived using Dual-Energy X-Ray Absorptiometry (DXA) and evaluated in both categorical and continuous forms. Given the cross-sectional design of NHANES for DKD status assessment, the association between FMR and DKD was analyzed as a prevalence association. Mortality outcomes were evaluated via retrospective linkage to the National Death Index, forming a retrospective mortality cohort among prevalent DKD cases. To evaluate the association between FMRs and DKD prevalence, we additionally analyzed an independent hospital-based clinical cohort. Results: After applying exclusion criteria, 680 DKD patients were included. Over a median follow-up of 97 months, 267 deaths (37.58%) were recorded. Logistic regression analysis identified arm-FMR, trunk-FMR, and total-FMR were independently associated with increased DKD risk (all P < 0.0001). Stratified subgroup analyses further confirmed significant associations between FMR and DKD, with notable interactions observed in arm-FMR and trunk-FMR when stratified by age and sex. Receiver operating characteristic curve analysis demonstrated that trunk-FMR exhibited the strongest predictive value for DKD (AUC = 0.812). Kaplan-Meier survival curves revealed that lower FMR quartiles were associated with better survival outcomes for both all-cause and CVD mortality in DKD patients (all log-rank P < 0.001). Moreover, non-linear associations were detected between FMR and DKD prevalence, as well as between FMR and mortality outcomes. In real-world validation cohort consisting of 94 patients, univariate logistics analysis revealed that all FMRs were identified as risk factors for the development of DKD. Multivariate logistics analysis showed that trunk FMR exhibited the highest predictive model value (OR=12.029, 95% CI 1.431-121.317, P=0.026,AUC=0.735). Conclusions: This NHANES-based study identifies a robust association between FMR and DKD prevalence, along with all-cause and CVD mortality. Importantly, these associations were supported by an independent real-world clinical cohort, underscoring the robustness and generalizability of our findings. Optimizing FMR may play a pivotal role in improving the prognosis of DKD patients.