Vitamin D status, metabolic determinants, and association with MASLD: retrospective health-screening cohort mediation analysis

Abstract

Background The relationship between vitamin D and metabolic dysfunction–associated steatotic liver disease (MASLD) remains controversial due to confounding by metabolic risk factors. We aimed to evaluate the association between vitamin D status and MASLD and examine the extent to which metabolic factors statistically contribute to this association. Methods In this large retrospective cross-sectional health-screening cohort study, 35,468 adults (mean age 46.3 years [SD 11.8]) undergoing routine health examinations were included. Serum 25-hydroxyvitamin D [25(OH)D] levels were analysed in relation to ultrasonography-defined MASLD. Multivariable logistic regression, partial correlation, and mediation analyses were used to assess independent associations and estimate the proportion of the observed association statistically explained by metabolic factors. Results Mean 25(OH)D concentration was 65.4 nmol/L (SD 26.4), and MASLD prevalence was 36.1%. Higher vitamin D levels were associated with lower odds of MASLD after adjustment for age and sex (OR 0.90 [95% CI 0.89–0.91]) and remained significant after additional adjustment for metabolic factors (OR 0.93 [0.92–0.95]). In contrast, associations between vitamin D and most metabolic traits were attenuated after adjustment, indicating non-independent relationships. Mediation analysis showed that 37.6% of the observed association between vitamin D and MASLD was statistically explained by metabolic factors, predominantly triglycerides (39.2%), HDL cholesterol (14.7%), and BMI (13.0%), whereas the remaining proportion was not explained by the measured mediators included in the model. The inverse association between vitamin D and MASLD was consistent across age and sex subgroups. Conclusion Vitamin D is independently associated with lower odds of MASLD beyond traditional metabolic factors. While metabolic factors contribute to part of the observed association, the remaining association was not explained by measured metabolic factors and should not be interpreted as evidence of a causal biological pathway. These findings support further investigation of vitamin D status as an associated marker of metabolic liver health and highlight the need for prospective studies evaluating its clinical relevance.