Insulin sensitizers are attractive candidate therapies for nonalcoholic steatohepatitis mainly because of the strong association between this disease and insulin resistance. This review provides a critical overview of the mechanisms of action, clinical trial results, and safety issues of metformin and glitazones in nonalcoholic steatohepatitis. It also highlights methodological challenges for trial design and proposes endpoints for future proof of principle, registrational, and postmarketing trials. (HEPATOLOGY 2010;52:2206-2215 N onalcoholic steatohepatitis (NASH) is emerging as the next big therapeutic challenge in hepatology. 1 Currently the top cause of newly identified chronic liver diseases, NASH has potential for fibrosis, cirrhosis decompensation, and hepatocellular carcinoma. Not all individuals with metabolic risk factors will experience these adverse outcomes and identifying those at risk is critical. Prognostic markers for progression have mostly been derived from histological studies. Compared to steatosis alone or to steatosis and minimal inflammation, the coexistence of inflammation, cell injury (ballooning), and steatosis, a pathological aggregate labeled steatohepatitis, is associated with a significant increase in overall mortality and in liver-related mortality. The degree of inflammation is the best predictor of fibrosis progression, 2 a widely accepted surrogate for hard endpoints such as cirrhosis, endstage liver disease, and possibly hepatocellular carcinoma.Insulin resistance (IR) is consistently found in patients with NASH in both cross-sectional and longitudinal studies. Moreover, there is a stepwise increase in the severity of IR and its phenotypic complications (clustered as elements of the metabolic syndrome) between normal liver, isolated steatosis, and steatohepatitis that has prompted speculation that IR might contribute to liver damage. Logically, most therapeutic trials focused on insulin sensitizers (IS) as candidate therapies.The aim of this report is to critically review the results of trials of metformin and glitazones for NASH. We highlight methodological challenges for trial design and propose endpoints for future proof of principle, registrational, and postmarketing trials.