Abstract
This study examined the relationship between fat mass index (FMI) and fat-free mass index (FFMI) and the incidence rate of prediabetes in the United States. This study analyzed a total of 4240 participants from the National Health and Nutrition Examination Survey. Logistic regression is the main method to analyze the correlation between FMI and FFMI and the incidence rate of prediabetes, including single-factor analysis, multiple-factor regression analysis, smooth curve fitting analysis and subgroup analysis. In addition, we further analyzed the role of intermediary variables in FMI and FFMI and the risk of prediabetes. Finally, the ROC analysis was used to confirm the predictive value of FMI and FFMI for prediabetes. After adjustment for covariates, each additional unit of FMI is associated with an 187% increase in the risk of prediabetes (OR: 2.87, 95% CI: 2.07-3.98); Each additional unit of FFMI was associated with a 292% increase in the risk of prediabetes (OR: 3.92, 95% CI: 2.68-5.76). Compared with individuals in the reference group, people with high FMI scores (Q4) are more likely to suffer from prediabetes (OR: 2.77, 95% CI: 1.98-3.87); People with high FFMI scores (Q4) are also more likely to suffer from prediabetes (OR: 3.01, 95% CI: 2.12-4.29). Smooth curve fitting analysis reveals a linear trend. The results of other subgroups were consistent with the overall results, except that gender factors were different in the relationship between FFMI and prediabetes. The ROC curve shows that FFMI is the best predictor of prediabetes, superior to FMI and body mass index.