The glycaemic and lipid effects of treatment with pioglitazone in combination with a stable insulin regimen were evaluated in patients with type 2 diabetes. Patients (n=566) receiving stable insulin regimens for > or = 30 days yet who had HbA1c > or = 8.0% and C-peptide > 0.7 microg/l were randomised to receive once-daily 15 mg pioglitazone, 30 mg pioglitazone, or placebo in a 16-week multicentre, double-blind, placebo-controlled trial. Per study protocol, the insulin dose was to remain unchanged, but could be decreased in response to hypoglycaemia. At the end of double-blind treatment, patients receiving pioglitazone (15 mg or 30 mg) showed statistically significant mean decreases relative to baseline HbA1c (-1.0 and -1.3, respectively; p<0.0001) and fasting plasma glucose (FPG) (-34.5 mg/dl [-1.92 mmol/l] and -48.0 mg/dl [-2.67 mmol/l], respectively; p<0.0001); these differences compared with placebo were also significant (p<0.0001). Pioglitazone (15 or 30 mg) yielded significant increases in HDL-C levels (mean increases ranging from +7.1% to + 9.3%) compared with baseline or placebo (p<0.01). The 30 mg dose also significantly reduced mean triglyceride levels (-23.7%) compared with placebo (p=0.0218). No consistent changes in TC or LDL-C levels were observed. The incidence of adverse events was similar in all treatment groups, although the incidences of weight increase, hypoglycaemia and oedema were higher among patients receiving insulin plus pioglitazone. There was no evidence of hepatotoxicity or drug-induced elevations of serum ALT > or = 3 x ULN. Pioglitazone, when added to stable insulin regimens, significantly improved HbA1c and FPG in type 2 diabetes. Pioglitazone treatment also provided significant benefit with respect to plasma HDL-C and triglyceride levels. Whether these lipid changes have an impact on overall diabetic complications remains to be determined.