Background: Glucagon-like peptide-1 (GLP-1) analogues reduce hepatic steatosis, concentrations of liver enzymes, and insulin resistance in murine models of fatty liver disease. These analogues are licensed for type 2 diabetes, but their efficacy in patients with non-alcoholic steatohepatitis is unknown. We assessed the safety and efficacy of the long-acting GLP-1 analogue, liraglutide, in patients with non-alcoholic steatohepatitis.
Methods: This multicentre, double-blinded, randomised, placebo-controlled phase 2 trial was conducted in four UK medical centres to assess subcutaneous injections of liraglutide (1·8 mg daily) compared with placebo for patients who are overweight and show clinical evidence of non-alcoholic steatohepatitis. Patients were randomly assigned (1:1) using a computer-generated, centrally administered procedure, stratified by trial centre and diabetes status. The trial was designed using A'Hern's single-group method, which required eight (38%) of 21 successes in the liraglutide group for the effect of liraglutide to be considered clinically significant. Patients, investigators, clinical trial site staff, and pathologists were masked to treatment assignment throughout the study. The primary outcome measure was resolution of definite non-alcoholic steatohepatitis with no worsening in fibrosis from baseline to end of treatment (48 weeks), as assessed centrally by two independent pathologists. Analysis was done by intention-to-treat analysis, which included all patients who underwent end-of-treatment biopsy. The trial was registered with ClinicalTrials.gov, number NCT01237119.
Findings: Between Aug 1, 2010, and May 31, 2013, 26 patients were randomly assigned to receive liraglutide and 26 to placebo. Nine (39%) of 23 patients who received liraglutide and underwent end-of-treatment liver biopsy had resolution of definite non-alcoholic steatohepatitis compared with two (9%) of 22 such patients in the placebo group (relative risk 4·3 [95% CI 1·0-17·7]; p=0·019). Two (9%) of 23 patients in the liraglutide group versus eight (36%) of 22 patients in the placebo group had progression of fibrosis (0·2 [0·1-1·0]; p=0·04). Most adverse events were grade 1 (mild) to grade 2 (moderate) in severity, transient, and similar in the two treatment groups for all organ classes and symptoms, with the exception of gastrointestinal disorders in 21 (81%) of 23 patients in the liraglutide group and 17 (65%) of 22 patients in the placebo group, which included diarrhoea (ten [38%] patients in the liraglutide group vs five [19%] in the placebo group), constipation (seven [27%] vs none), and loss of appetite (eight [31%] vs two [8%]).
Interpretation: Liraglutide was safe, well tolerated, and led to histological resolution of non-alcoholic steatohepatitis, warranting extensive, longer-term studies.
Funding: Wellcome Trust, National Institute of Health Research, and Novo Nordisk.
Copyright © 2016 Elsevier Ltd. All rights reserved.