Fasting total cholesterol (TC), triglycerides (TG), HDL cholesterol (HDL C), apolipoprotein A1 (apo A1) and apolipoprotein B (apo B) were measured in 35 non-insulin dependent diabetic patients treated by diet with or without sulphonylureas and 35 control subjects matched for age, sex, and body mass index. Ratios of apolipoprotein and lipid were calculated. The diabetics were well controlled with a mean (+/- SD) glycosylated haemoglobin (HbA1) of 8.5 +/- 1.3% (normal range less than 8%). Compared to non-diabetic control subjects apo A1: HDL C, apo B: TC, and apo B: calculated LDL C were significantly higher in the NIDDM patients, (112.9 +/- 26.3 vs 83.0 +/- 28.7, p less than 0.001, 15.89 +/- 1.68 vs 14.22 +/- 3.48, p less than 0.01, and 24.32 +/- 3.19 vs 22.33 +/- 5.49, p less than 0.05 respectively). These findings reflect differences in cholesterol content in the absence of differences in apolipoprotein concentrations between the NIDDM and control groups. The cardiovascular risk ratio HDL C: non HDL C was significantly lower in the NIDDM patients (0.25 +/- 0.09 vs 0.31 +/- 0.15, p less than 0.01), but there was no difference in apo A1:apo B (1.42 +/- 0.42 vs 1.43 +/- 0.52, NS). Although apo A1: apo B correlated well with HDL C:non HDL C in both NIDDM and controls (r = 0.88, 0.72, p less than 0.001 respectively) the slope of the relationships differed b = 4.01 NIDDM vs 2.50 controls (95% confidence intervals for difference is 0.22-2.78). Simple widely available methods can identify abnormalities of lipoprotein content in treated NIDDM patients. Both HDL and LDL contain less cholesterol.(ABSTRACT TRUNCATED AT 250 WORDS)