Abstract Body

The ongoing debate regarding nonalcoholic fatty liver disease (NAFLD)/metabolic-associated fatty liver disease (MAFLD) should consider risk of progression of fatty liver disease (FLD). We aimed to describe transition of NAFLD and MAFLD states towards significant fibrosis in people with HIV (PWH).

This was a longitudinal study of two prospective cohorts of PWH on stable antiretroviral therapy. FLD was assessed at least twice with controlled attenuation parameter (CAP ?248 dB/m) by transient elastography. NAFLD was defined as FLD in absence of significant alcohol intake or HBV or HCV co-infections. MAFLD was defined as the presence of FLD and at least one criteria: 1) BMI?25 kg/m2; 2) diabetes; or 3) lean (BMI<25 kg/m2) with at least two immune-metabolic alterations [Eslam M. JHepatol.2020;73(1):202-209]. Significant liver fibrosis was defined as liver stiffness ?7.1 kPa. A continuous-time multi-state Markov model was used to describe the process in which a study patient moved through a series of states allowing joint analysis of care length, incidence of FLD or fibrosis progression or reversion. The probabilities to switch from one state to another were modelled according to an exponential distribution for time-to-event data, considering censored follow-up times. The events were the transitions between the states. The analyses were performed separately for NAFLD and three MAFLD categories in which minimum two and maximum four assessments for FLD were considered.

A total of 888 PWH were screened for FLD, with a mean follow-up of 2 years, mean age 54.4 years, 77% males. At the first visit, after excluding PWH with alcohol intake and viral co-infections, prevalence of NAFLD was 42.9% (285/664), while the overall prevalence of the MAFLD was 34.3% (305/888). In detail, MAFLD with BMI?25 kg/m2 was present in 244 (27.5%), MAFLD with diabetes in 86 (9.7%) and lean MAFLD in 33 (3.7%). Figure 1 shows alluvial plots of state transitions in NAFLD (panel A) and in MAFLD with BMI?25 kg/m2 (Panel B), with diabetes (Panel C) and lean MAFLD (Panel D). Each panel is accompanied by table that summarizes probabilities to move from one state to another.

Use of Markov models depicts dynamic changes of FLD with or without fibrosis over time. The highest risk of liver fibrosis progression was observed in PWH with MAFLD with BMI >25. MAFLD categories offer the possibility to stratify PWH at highest risk of hepatic and extra-hepatic adverse outcomes.