Association between novel anthropometric indices and bone loss in patients with diabetes and prediabetes: findings from a national population-based study

Previous studies have demonstrated a significant increase in the prevalence of bone loss among patients with diabetes and prediabetes, which is closely associated with obesity indicators such as waist circumference and body weight. The present study aims to investigate the relationship between novel anthropometric indices (A Body Shape Index [ABSI], Weight-adjusted Waist Index [WWI], Waist-to-Height Ratio [WHtR], Body Roundness Index [BRI]) and bone loss (osteopenia/osteoporosis) in patients with diabetes and prediabetes. This cross-sectional analysis was conducted using data from the National Health and Nutrition Examination Survey (NHANES) for patients aged 20 and above with diabetes and prediabetes. Weighted multivariable logistic regression was employed to explore the associations between novel anthropometric indices and bone loss. Restricted cubic spline (RCS) curves were utilized to evaluate potential nonlinear relationships, while receiver operating characteristic (ROC) curves assessed diagnostic accuracy. Additionally, subgroup analyses and interaction tests were performed. This study included a total of 4,683 patients with diabetes and prediabetes, among which 1,814 were identified with bone loss. After fully adjusting for covariates, the multivariable logistic regression model revealed that compared to participants in the first quartile of ABSI, WHtR, WWI, and BRI, the prevalence of bone loss among patients with diabetes and prediabetes increased in the fourth quartile for ABSI: OR = 1.07; WHtR: OR = 1.09; WWI: OR = 1.13; BRI: OR = 1.09. RCS analysis indicated a linear dose-response relationship between ABSI, WWI and bone loss. Additionally, there was a non-linear dose-response relationship between WHtR, BRI and bone loss. The areas under the ROC curve values: 0.778 for ABSI, 0.781 for WHtR, 0.783 for BRI, and 0.782 for WWI. Significant interactions were found for factors such as age, hypertension, and alcohol consumption. Our findings reveal a significant correlation between ABSI, WWI, BRI, and WHtR levels and the prevalence of bone loss in patients with diabetes and prediabetes. This suggests that ABSI, WWI, BRI, and WHtR can serve as valuable biomarkers for assessing bone loss in these populations. Not applicable.