Mediators of dietary diversity score (DDS) on NAFLD in Iranian adults: a structural equation modeling study

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: β = 0.14, P < 0.0001, and of WC to NAFLD: β = 0.50, P < 0.001). Vegetable and fruit diversity scores had a significant negative indirect relationship with NAFLD prevalence through DAL (β = −0.06, P = 0.001, β = −0.10, P < 0.001, respectively). Meat diversity score was positively associated with NAFLD prevalence in a full mediational process through DAL (β = 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 (Χ 2 /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.