Rstudio version 4.0.1

We (ZD and TA or DK) used the Cochrane risk of bias (RoB) assessment tool [20 ] to assess risk of bias in each study. This consists of seven domains (randomised sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selecting reporting and other bias). For each domain, we directly cited evidence from studies or their protocols to support judgement of assessment and rated each domain as low or high risk of bias or some concerns. Where there were missing data or uncertainties based on the information given in articles we contacted authors of studies for more information before assessing the domain as ‘some concerns’. When assessing for ‘other bias’ we considered the following: balance of baseline characteristics; participant receipt of concurrent treatments such as other psychological therapies or pharmaceutical treatments; contamination of intervention between groups were collected to support the judgement; small sample size; and (with respect to cluster RCTs), loss of clusters, unit of analysis and adjustment for clustering. We constructed a ‘risk of bias’ graph and summary figures using the robvis package in RStudio (Version 4.0.1) (RStudio, Inc, Boston, USA). This review is reported in accordance with the PRISMA guidelines [17 (link)].

All continuous variables were tested for normality and are described by medians (interquartile range) and proportions. The Mann–Whitney U test was performed to compare continuous variables due to their nonnormal distribution. For categorical variables, the chi-square test was performed to compare different variables. The specific prevalence of different age, BMI and female menopausal status groups and their 95% confidence intervals (CIs) were calculated. Binary logistic regression analysis was performed to explore the related risk factors for MAFLD. Odds ratios (ORs) and their 95% CIs were finally calculated. Binary logistic regression was performed using RStudio version 4.0.1, and other analyses were performed using SPSS 25. A two-tailed p value < 0.05 was considered statistically significant.