In non-insulin-dependent diabetes mellitus (NIDDM), cardiovascular risk factors improve during treatment, but whether insulin (I) differs from sulfonylurea (SU) therapy is unclear. To separate the contributions of improved diabetic control versus treatment regimen to risk factors, we examined the effects of SU and I on insulin sensitivity, basal and post-glucose load levels of insulin-like molecules, fibrinolysis, and lipid concentrations. Twenty poorly controlled, diet-treated NIDDM subjects were given I or SU each for a period of 16 weeks in a randomized crossover study, with a 4-week washout period between each treatment. Subjects were studied at the baselines (B1 and B2) and after each treatment. Treatment with I or SU produced similar improvements in glycemia (hemoglobin A1 [HbA1] B1, 11.7% +/- 2.1%; SU, 8.5% +/- 0.9%; I, 8.6% +/- 1.2%) and the metabolic clearance rate of glucose ([MCR-G] B1, 1.86 x/divided by 1.4; SU, 2.36 x/divided by 1.4 (P = .005 vB1); I, 2.27 x/divided by 1.4 (P = .07 vB1) ml x kg(-1) x min(-1)). On SU therapy, subjects had higher fasting and post-glucose load levels of intact proinsulin compared with B1 and I (fasting, 13.9 x/divided by 2.6 v 9.5 x/divided by 2.2 (P = .004) and 9.1 x/divided by 2.4 pmol x L(-1) (P = .01), respectively). Plasminogen activator inhibitor-1 (PAI-1) activity and antigen were higher than at B1 on SU therapy (23.7 v 19.9 AU x mL(-1) (P = .02) and 47.6 v 32.2 ng x mL(-1) (P = .006), respectively), but not on I. There were no changes compared with B1 and no differences between the two therapies in total, very-low-density lipoprotein (VLDL), and intermediate-density lipoprotein (IDL) cholesterol and triglyceride, low-density lipoprotein (LDL), high-density lipoprotein 2 (HDL2) and HDL3 cholesterol, apolipoprotein (apo) A1, A2, and B1, or lipoprotein (a) [Lp(a)] levels. In conclusion, (1) treatment with SU or I resulted in equal improvement in glycemia and insulin sensitivity, (2) intact proinsulin and PAI-1 antigen and activity were higher on SU, and (3) there were no differences in lipid concentrations with improved glycemia or between therapies.