A cross-sectional population study was performed in a cohort of 890 non-insulin-dependent diabetes mellitus (NIDDM) patients residing in the greater Denver metropolitan region. Its purpose was to evaluate the relationship between insulin and oral hypoglycemic agents (OHAs) with regard to metabolic control and diabetic complications. The mean glycosylated hemoglobin for patients treated with insulin was 12.0 +/- 0.15% versus 11.4 +/- 0.14% (p < .03) for OHA. The difference in fasting blood sugar for the insulin-treated group (195.0 +/- 3.5 mg/dl) versus the OHA-treated group (194.0 +/- 2.9 mg/dl) was not statistically significant. Categorical increases in urinary albumin excretion were associated positively within insulin versus OHA therapy (p < .0001). Patients treated with insulin therapy had a higher frequency of peripheral vascular disease (insulin therapy, 14%; OHA therapy, 10%; p < .05); neuropathy (insulin therapy, 55%; OHA therapy, 37%; p < .0001); and retinopathy (insulin therapy, 71%; OHA therapy, 45%; p < .0001). The frequency of cardiovascular disease was equivalent in the two groups (17% versus 13%). In protocols correcting for diabetes duration, glycosylated hemoglobin, and gender in a multivariate model, the use of insulin still was related significantly to increases in urinary albumin excretion (p < .01), retinopathy (p < .0001), and neuropathy (p < .0008). In a subgroup of individuals with diabetes duration > 10 years (n = 211 for insulin treatment, n = 118 for OHA treatment), the frequency of neuropathy still was significantly higher in the insulin group (63% vs 49%; p < .016) as was retinopathy (85% vs 58%; p < .0001). Overt albuminuria also was more significant in the insulin-treated patients (p < .04). In summary, the NIDDM patients treated with insulin had more nephropathy, retinopathy, and neuropathy than did NIDDM patients treated with OHA, independent of duration of diabetes, fasting blood glucose, glycosylated hemoglobin, age, and blood pressure level. These results in NIDDM patients may be due to contributions from worse blood glucose control at an earlier stage in the patients' diabetes and/or the mitogenic, atherogenic, thrombogenic, and vascular permeability effects of insulin.