CT-assessed abdominal visceral adiposity and MASLD: a sex-stratified cross-sectional analysis

To test the hypothesis that CT-derived visceral fat and muscle measures are more strongly associated with MASLD than BMI, and that these associations are sex-specific and nonlinear, we examined abdominal body composition in relation to MASLD risk.

In this cross-sectional study of 7,805 Chinese adults who underwent non-contrast abdominal CT between September 2020 and June 2023, body composition was quantified, including visceral fat area (VFA), subcutaneous fat area (SFA), skeletal muscle area (SMA), L3 skeletal muscle index (L3SMI), and CT-derived waist circumference(CT-WC). Participants were drawn from hospitalized patients undergoing CT for clinical indications (n = 7,051) and individuals attending routine health examinations (n = 754). MASLD was defined according to the 2023 criteria. Associations with MASLD were assessed using logistic regression, spline modeling, and ROC analyses.

Among 7,805 participants (mean age 47.5 years; 62.2% men), MASLD prevalence was 25.8% (28.1% in men and 22.0% in women). In sex-specific fully adjusted models, VFA showed the strongest and most consistent association with MASLD. Men in the highest VFA quartile had an odds ratio (OR) of 11.5 (95% CI 8.3-16.1) compared with the lowest quartile, while women exhibited a markedly steeper risk gradient (OR 32.5, 95% CI 17.3-67.0). Waist circumference (WC) showed similar independent associations, whereas BMI was not independently associated with MASLD in men but remained significant in women. Subcutaneous fat and skeletal muscle measures were weakly associated or not significant after adjustment. Restricted cubic spline analyses revealed pronounced nonlinear and threshold effects for visceral fat, with risk rising sharply beyond sex-specific VFA thresholds. CT-derived body composition models improved MASLD discrimination compared with BMI alone (AUC 0.76 vs. 0.70 in men; 0.83 vs. 0.77 in women). Associations were consistent across metabolic subgroups and robust to sensitivity analyses, including E-value assessment for unmeasured confounding.

CT-based measures of visceral fat may provide significant incremental value over BMI in MASLD risk assessment. Findings suggest that MASLD risk may differ by sex and appears to be more closely related to fat distribution than to total adiposity, highlighting the potential need to incorporate central obesity assessment into clinical and public health strategies.