OBJECTIVE—To characterize potential differences in glycemic control, plasma lipid level, and weight in a cohort of patients previously treated with troglitazone (TROG) who were switched to either pioglitazone or rosiglitazone.

RESEARCH DESIGN AND METHODS—After a 2-week washout from TROG, 186 patients were randomly assigned to receive either pioglitazone (PIO) or rosiglitazone (ROSI). Weight, HbA_1c, and fasting lipid profile were documented before discontinuing TROG and at 4 months after starting either pioglitazone or rosiglitazone. Secondarily, the effect of concurrent medications on study outcomes was assessed.

RESULTS—A total of 127 patients completed follow-up: 67 individuals in the PIO group (32 women, 35 men) and 60 individuals in the ROSI group (33 women, 27 men). There were no significant differences in gender mix, age, weight, fasting lipid profile, or HbA_1c between the ROSI and PIO groups. After 4 months of randomized treatment, no change in HbA_1c from baseline between or within groups was noted. Both groups experienced an equal and significant increase in weight from baseline of ∼2.0 kg. Thiazolidinedione and HMG-CoA reductase inhibitor therapy had significant and independent effects on lipid profile (P < 0.005). Significant improvements in lipid profile were noted in the PIO group (P < 0.01), whereas none were detected with conversion to ROSI. Specifically, the PIO group experienced an average decrease in total cholesterol of ∼20 mg/dl.

CONCLUSIONS—Differing effects on lipid profile were apparent after random conversion from TROG to either PIO or ROSI, despite similar weight increase and glycemic control. The clinical significance of these differences remains to be determined, and further comparative research is warranted.