NHANES cross-sectional analysis links visceral fat mass index with diabetes prevalence
image: Red lines show odds ratios (OR) with 95% confidence intervals (CIs) (dashed). Gray bars indicate the distribution of log(vFMI). Estimated inflection points are marked for diabetes (−1.380) and DN (−1.423). P-overall values represent overall associations, and P-nonlinear values test for deviation from linearity. Adjusted for age, sex, race, education level, poverty income ratio (PIR), alcohol consumption, smoking, HDL cholesterol, and systolic blood pressure (SBP). The solid and dotted lines represent the estimated values and their corresponding 95% CIs, respectively.
Credit: Zheng Shi
Researchers from Chengdu University, Nanfang Hospital, and collaborating institutions have published a cross-sectional study in Health Data Science assessing associations of visceral fat mass index (vFMI) with the prevalence of diabetes, diabetic nephropathy, and cardiovascular disease. The analysis indicates that higher vFMI is associated with higher prevalence of these outcomes.
Visceral fat has been linked to insulin resistance and cardiovascular impairment, yet the relationship of dual-energy x-ray absorptiometry (DXA)-derived vFMI to diabetes and its vascular complications has not been systematically assessed. The team analyzed data from the 2011–2018 National Health and Nutrition Examination Survey (NHANES).
The study included 4,847 participants aged 18 years and older, with a mean age of 40 years. Prevalence of diabetes, diabetic nephropathy, and cardiovascular disease was 11.6% (567 participants), 2.8% (136), and 4.3% (208), respectively. In fully adjusted models, each 1-standard-deviation increase in vFMI was associated with odds ratios of 1.90 (95% CI, 1.66 to 2.17) for diabetes and 1.78 (95% CI, 1.35 to 2.34) for diabetic nephropathy. Comparing the highest with the lowest vFMI quartile, the odds ratio was 5.85 (95% CI, 3.25 to 10.55) for diabetes and 2.55 (95% CI, 1.10 to 5.93) for cardiovascular disease. The corresponding diabetic-nephropathy comparison was 2.84 (95% CI, 0.89 to 9.08) and was not statistically significant. Nonlinear analysis of log-transformed vFMI identified thresholds above which odds rose more sharply: 0.252 for diabetes and 0.241 for diabetic nephropathy on the original vFMI scale.
The findings can inform further study of body-composition markers in metabolic risk identification. Because the analysis is cross-sectional, it cannot establish temporal order or causality and should not be used alone for clinical diagnosis or risk stratification.