Low-Carbohydrate Diets, but Not Low-Fat Diets, Increase Non-alcoholic Fatty Liver Disease Risk in the Framingham Heart Study

Abstract

ObjectivesNonalcoholic fatty liver disease (NAFLD) is a highly prevalent disease associated with morbidity and mortality. NAFLD prevention focuses on weight loss; however, information on the long-term effects of low-fat diets (LFD) or low-carbohydrate diets (LCD), two popular strategies for weight loss, on NAFLD risk are not known. We examined the prospective association of LFD and LCD with NAFLD risk in a US community setting.MethodsWe included participants from the Framingham Heart Study second and third-generation cohorts. Liver fat was assessed on computed tomography scan using the average liver fat attenuation referenced to a radiopaque phantom to create the liver phantom ratio (LPR) at baseline and after 6 years (median). Incident NAFLD was defined as LPR ≤ 0.33 on the follow-up CT scan after excluding baseline NAFLD. We quantified baseline adherence to LFD and LCD using scores that consider carbohydrate, fat, and protein energy percentages derived from food-frequency questionnaires. Modified poison regression models were used to compute NAFLD risk ratios (RR) and confidence intervals (CI) associated with score tertiles adjusting for age, sex, education, and lifestyle confounders, including baseline BMI and BMI change.ResultsOver 6 years, 18.6% of participants (mean age 51 years, 38% women) developed NAFLD. Those with a higher adherence to LCD (vs. lower) had lower carbohydrate, higher fat, and slightly higher protein intakes; their diets included less fiber and dairy and more meat products. Those with higher LFD scores (vs. lower) had the opposite dietary composition except protein intake was similar. Women with the highest adherence to LCD (vs. lowest) had an 89% increased risk of developing NAFLD (95% CI: 1.16–3.08) after adjusting for sociodemographic and lifestyle factors. No associations were found in men. Further adjustment for BMI but not BMI change slightly attenuated the results. Participants who had hypertension and followed an LCD were at highest risk of developing NAFLD. A LFD was not associated with NAFLD.ConclusionsIn our study, higher adherence to LCD associated with an increased risk for incident NAFLD among women. Additional studies are needed to determine the optimal diet to prevent NAFLD.Funding SourcesNational Institute of Diabetes and Digestive and Kidney Diseases.