ABSTRACT
BACKGROUND: The current study examines the association between the Dietary Diversity Score (DDS) and nonalcoholic fatty liver disease (NAFLD) in Iranian adults using structural equation modeling (SEM). METHODS: A sample of 3220 adults from the Amol Cohort Study was recruited for this cross-sectional study. Dietary acid load (DAL) and DDS were calculated using the data obtained from a validated food frequency questionnaire. Anthropometric parameters, blood pressure, biochemical measurements, and liver ultrasonography were evaluated according to standard protocols. RESULTS: DDS was neither directly nor indirectly associated with a greater prevalence of NAFLD. In the second model (DDS sub-scores model), the association of DAL with NAFLD was fully mediated through waist circumference (WC) (of DAL to WC: ß = 0.14, P < 0.0001, and of WC to NAFLD: ß = 0.50, P < 0.001). Vegetable and fruit diversity scores had a significant negative indirect relationship with NAFLD prevalence through DAL (ß = -0.06, P = 0.001, ß = -0.10, P < 0.001, respectively). Meat diversity score was positively associated with NAFLD prevalence in a full mediational process through DAL (ß = 0.12, P < 0.001). The SEM fit indices suggested a reasonably adequate fit of the data to the DDS model (Χ2/df = 4.76, GFI = 0.98, AGFI = 0.97, IFI = 0.97, CFI = 0.97, RMSEA = 0.03, and SRMR = 0.02) and its sub-scores model (Χ2/df = 4.72, GFI = 0.98, AGFI = 0.97, IFI = 0.95, CFI = 0.95, RMSEA = 0.03, and SRMR = 0.02). CONCLUSION: Meat diversity and lack of vegetable and fruit diversity were indirectly associated with NAFLD prevalence through DAL and WC mediators. Interventions for NAFLD may be more successful if they target a lower intake of animal protein sources and dietary diversity, particularly vegetable and fruit diversity.