Response to Pioglitazone in Patients With Nonalcoholic Steatohepatitis With vs Without Type 2 Diabetes

BACKGROUND & AIMS: Pioglitazone is effective for long-term treatment of patients with nonalcoholic steatohepatitis (NASH) with prediabetes or type 2 diabetes. However, it is not clear how the presence of type 2 diabetes affects the drug's efficacy. We compared metabolic and histologic responses to pioglitazone in patients with NASH and prediabetes vs type 2 diabetes. METHODS: Weperformed a prospective study of adults with biopsy-proven NASH (52with type 2 diabetes and 49 with prediabetes), enrolled from the general population of San Antonio, Texas, from 2008 through 2014. After a run-in period of approximately 4 weeks, when all baseline measurements were made (liver magnetic resonance proton spectroscopy, euglycemic insulin clamp with glucose turnover measurements, dual-energy absorptiometry, and liver biopsy), subjects were randomly assigned to groups given pioglitazone or placebo (45 mg/d) for 18 months; all procedures performed at baseline were then repeated. The primary outcomewas a reduction in nonalcoholic fatty liver disease activity score of 2 points or more (for at least 2 components) without worsening of fibrosis (and expressed as difference vs placebo). Secondary outcomes included NASH resolution, individual histologic components, intrahepatic triglyceride content (measured by H-1 magnetic resonance spectroscopy), and insulin sensitivity (measured by euglycemic insulin clamp). RESULTS: The primary outcome was met by 48% of patients with type 2 diabetes vs 46% without diabetes. Resolution of NASH was achieved in 44% of patients with type 2 diabetes vs 26% without diabetes. A significant reduction in fibrosis, from baseline, was observed only in patients with type 2 diabetes (P = .035). Intrahepatic triglyceride content was reduced by 11% +/- 2% in patients with diabetes vs a reduction of 9% +/- 2% in patients without diabetes (P = .62); the plasma level of alanine aminotransferase was reduced by 50 +/- 10 U/L in patients with diabetes vs a reduction of 36 +/- 5 U/L in patients without diabetes (P = .22). Pioglitazone was associated with a significantly greater insulin sensitivity in adipose tissue of patients with diabetes vs without diabetes (P < .001), but nonsignificant differences in responses in hepatic (P = .49) and skeletal muscle (P = .32) insulin sensitivity. CONCLUSIONS: In a prospective study, we found pioglitazone to be effective in patients with and without type 2 diabetes. However, pioglitazone reduced liver fibrosis and increased adipose tissue insulin sensitivity at significantly greater levels in patients with type 2 diabetes than in patients with prediabetes. Further studies are needed to determine the mechanisms by which pioglitazone reduces liver disease in patients with type 2 diabetes.