The effect of the magnitude of weight loss on non-alcoholic fatty liver disease: a systematic review and meta-analysis.

BACKGROUND: Trials show that weight loss interventions improve biomarkers of non-alcoholic fatty liver disease (NAFLD), but it is unclear if a dose-response relationship exists. OBJECTIVE: We aimed to quantify the dose-response relationship between the magnitude of weight loss and improvements in NAFLD. METHODS: Nine databases and trial registries were searched until October 2020. Single-arm, non-randomized comparative, or randomized trials of weight loss interventions (behavioral weight loss programs [BWLPs], pharmacotherapy, or bariatric surgery) in people with NAFLD were eligible for inclusion if they reported an association between changes in weight and changes in blood, radiological, or histological biomarkers of liver disease. The review followed Cochrane methods and the risk of bias was assessed using the Newcastle-Ottawa scale. Pooled unstandardized b coefficients were calculated using random-effect meta-analyses. RESULTS: Forty-three studies (BWMPs: 26, pharmacotherapy: 9, surgery: 8) with 2809 participants were included. The median follow-up was 6 (interquartile range: 6) months. The direction of effect was generally consistent but the estimates imprecise. Every 1 kg of weight lost was associated with 0.83-unit (95% CI: 0.53 to 1.14, p < 0.0001, I2 = 92%, n = 18) reduction in alanine aminotransferase (U/L), 0.56-unit (95% CI: 0.32 to 0.79, p < 0.0001, I2 = 68%, n = 11) reduction in aspartate transaminase (U/L), and 0.77 percentage point (95% CI: 0.51 to 1.03, p < 0.0001, I2 = 72%, n = 11) reduction in steatosis assessed by radiology or histology. There was evidence of a dose-response relationship with liver inflammation, ballooning, and resolution of NAFLD or NASH, but limited evidence of a dose-response relationship with fibrosis or NAFLD activity score. On average, the risk of bias for selection and outcome was medium and low, respectively. CONCLUSION: Clinically significant improvements in NAFLD are achieved even with modest weight loss, but greater weight loss is associated with greater improvements. Embedding support for formal weight loss programs as part of the care pathway for the treatment for NAFLD could reduce the burden of disease. PROSPERO: CRD42018093676.