Silencing of acid-labile subunit (ALS) improved glucose metabolism in animal models. The aim of this study is to evaluate the effects of rosiglitazone (RSG) on ALS levels in individuals with type 2 diabetes. A randomized, double-blind, placebo-controlled trial was conducted. Subjects with type 2 diabetes mellitus were randomly distributed to an RSG-treated (n = 30) or a placebo (n = 31) group. Patients were evaluated prior to treatment at baseline and at 12 and 24 weeks after treatment. At baseline, ALS levels were negatively associated with low-density lipoprotein cholesterol (LDLc) levels and homeostatic model assessment version 2 insulin sensitivity (HOMA2-%S). Over 24 weeks, there was a significantly greater reduction in ALS levels in the nonobese RSG-treated individuals than placebo-treated group. The effect of RSG on ALS was not significant in obese individuals. Fasting plasma glucose and hemoglobin A1c were reduced, but total cholesterol and LDLc were increased, in patients on RSG. Change in ALS levels predicted changes in total cholesterol and HOMA2-%S over time. This study suggested a BMI-dependent effect of RSG treatment on ALS levels. Reduction of ALS by RSG increases the risk of atherosclerosis in individuals with type 2 diabetes.