Plasma insulin, intact proinsulin and 32-33 split proinsulin measured by specific immunoradiometric assays and insulin and C-peptide measured by radioimmunoassay were measured during a constant infusion of glucose test in ten diet-treated subjects with a history of Type 2 (non-insulin-dependent) diabetes (termed diabetic subjects), mean fasting plasma glucose 6.0 +/- 1.0 mmol/l (mean +/- SD), and 12 non-diabetic control subjects. Immunoreactive insulin concentrations measured by radioimmunoassay were 33% higher than insulin and 16% higher than the sum of insulin and its precursors by immunoradiometric assay. The diabetic and non-diabetic subjects had similar fasting concentrations of insulin, intact proinsulin and 32-33 split proinsulin. The ratio of fasting intact proinsulin to total insulin was greater in the diabetic than the non-diabetic group 12.0% (6.8-21.0%, 1 SD range) and 6.3% (4.0-9.8%), respectively, p less than 0.01), though the groups overlapped substantially. After glucose infusion, diabetic and non-diabetic subjects had similar intact proinsulin concentrations (geometric mean 4.9 and 5.2 pmol/l, respectively), but the diabetic group had impaired insulin secretion by immunoradiometric assay (geometric means 55 and 101 pmol/l, p less than 0.05) or by radioimmunoassay C-peptide (geometric means 935 and 1410 pmol/l, p less than 0.05), though not by radioimmunoassay insulin (87 and 144 pmol/l, p = 0.12), respectively. Individual immunoradiometric assay insulin responses to glucose expressed in terms of obesity were subnormal in nine of ten diabetic subjects. Radioimmunoassay insulin and C-peptide gave less complete discrimination (subnormal responses in six of ten and eight of ten, respectively).(ABSTRACT TRUNCATED AT 250 WORDS)