Literature evidences about regression of nonalcoholic fatty liver disease (NAFLD) are not as high as its development or progression. Therefore this cohort study was conducted to reveal the factors influencing the regression of the NAFLD

Through a multistage randomized sampling, 844 adults older than 18 years from population of Shiraz, Iran were participated in this study in 2010-2011, while 554 of them were recruited in the study in 2015. Demographic, anthropometric, nutrition and medical characteristics of each interviewee was entered into a valid and reliable questionnaire. Interviewees were categorized into three groups according to the changing of their NAFLD status between two stages of this study; no change, regressed and progressed. Kruskal-Wallis test, chi-squared test and multi-nominal logistic regression analysis were used for data analysis in SPSS.

Of the 537 participants, 163 (30.3%) showed regression of their NAFLD compared to 93 (17.3%) that developed NAFLD or their NAFLD was progressed. Significant differences were observed among regressed, progressed and no change groups regarding their daily calories intake (p=0.026), carbohydrate intake (p=0.011), Body Mass Index (BMI) (p<0.001) and waist circumference (WC) (p<0.001) in univariate analysis. Logistic regression showed that each unit decreasing in BMI (Δ BMI= -1 kg/m2) augmented the odds of improvement of NAFLD by 11.5% (OR: 0.885; 95% CI: 0.793-0.987; P=0.028) and reduced the chances of NAFLD deterioration by 14.5% (OR: 1.145; 95% CI: 1.015-1.291; P=0.027). The reduction in BMI increased the regression of NAFLD and conversely. However, increasing BMI had a bit more effect on increasing chances of progression (OR=1.12) of NAFLD than reduction of BMI in increasing chances of its regression (OR=1.14).

A significant portion of patients showed regression of their NAFLD by decreasing their BMI. Therefore, BMI as a modifiable factor should be regarded in the management of NAFLD patients.