Background and aimsLow serum 25-hydroxyvitamin D (25(OH)D) is associated with higher nonalcoholic fatty liver disease (NAFLD) risk in studies of mainly white participants. Significant racial/ethnic differences exist in serum 25(OH)D and NAFLD prevalence questioning extending this association to other racial/ethnic groups. We tested whether the association between serum 25(OH)D and NAFLD vary by race/ethnicity. Methods and resultsThis was a cross-sectional analysis from the Multi-Ethnic Study of Atherosclerosis (MESA) that included 3484 participants (44% male; 38.4% Whites, 27.8% African-Americans, 23.5% Hispanics, and 10.3% Chinese-Americans) who had serum 25(OH)D and upper abdominal CT images available at baseline. Serum 25(OH)D was measured by high-performance liquid chromatography–tandem mass spectrometry. NAFLD was identified if liver-to-spleen Hounsfield-Unit ratio was <1. Whites had the highest 25(OH)D level and African-Americans had the lowest level (mean ± SD: 29.5 ± 10.4 vs.19.9 ± 9.1, respectively). Six hundred and eleven (17.5%) participants had NAFLD; Hispanics had the highest prevalence (26.2%) followed by Chinese-Americans (19.8%), Whites (15.8%) and African-Americans (11.7%), P < 0.0001. In adjusted model, the association of 25(OH)D with NAFLD differed by race/ethnicity (P < 0.0001). Negative association was only evident in Causations (OR (95% CI):1.23 (1.03, 1.47) per 1 SD lower serum 25(OH)D). For other racial/ethnic groups, BMI, triglycerides, diabetic status and/or smoking, but not serum 25(OH)D, were common independent risk factors for NAFLD. ConclusionsThe negative association between serum 25(OH)D and NAFLD in Whites may not be broadly generalizable to other racial/ethnic groups. Modifiable risk factors including BMI, triglycerides, diabetic status and/or smoking associate with NAFLD risk in non-white racial/ethnic groups beyond 25(OH)D.