Modeling the progression of NAFLD: a multi-state Markov approach based on BMI classification in the Chinese population

Aim: To investigate the natural history and progression of non-alcoholic fatty liver disease (NAFLD) in Chinese adults using a community-based longitudinal cohort. Methods: We analyzed data from 24,893 adults in the Beijing Health Management Cohort (2016–2021) with annual follow-up, including questionnaires, physical examinations, and laboratory tests. Participants were categorized into three states: healthy (S1, n = 17,906), NAFLD with body mass index (BMI) < 24 kg/m2 (S2, n = 1,139), and NAFLD with BMI ≥ 24 kg/m2 (S3, n = 5,848). Transition probabilities, mean sojourn times, and determinants of progression were estimated using a multi-state Markov model, with analyses stratified by sex. Results: Most individuals remained in their baseline state. Females in S1 were more likely to stay healthy than males (93.76% vs. 83.52%), while males had a higher risk of progression to S3 (12.59% vs. 4.23%). Females in S3 had a greater chance of reverting to S1 (21.33% vs. 18.18%). Mean sojourn time in S1 was longer for females (18.58 vs. 6.97 years), whereas males spent more time in S3 (5.37 vs. 4.18 years). Age, hyperuricemia, abdominal obesity, high triglycerides, and low high-density lipoprotein cholesterol (HDL-C) significantly increased the risk of progression. Conclusions: Sex differences strongly affect NAFLD progression in Chinese adults. Males are more likely to deteriorate, while females show higher recovery. Metabolic factors and obesity are key targets for early prevention and intervention.