Insulin-like growth factor (IGF)-binding protein-1 (IGFBP-1) modulates the metabolic and mitogenic effects of IGFs. Although IGFBP-1 levels are abnormally high in insulin-dependent diabetes (IDDM), relatively little is known in NIDDM; conflicting data have suggested both high and low levels. We investigated whether treatment modifies IGFBP-1 levels in two groups of NIDDM patients. Study 1 examined fasting concentrations in groups of patients with NIDDM, comparable except for treatment type (sulfonylurea, n = 23; once daily insulin, n = 15; sulfonylurea plus once daily insulin, n = 14; multiple insulin injections, n = 9) and 25 nondiabetic subjects. In sulfonylurea-treated patients there were markedly reduced plasma IGFBP-1 concentrations (median, interquartile range in parentheses): control, 61.0 (36-96) micrograms/L; sulfonylureas alone, 31.5 (21-61) micrograms/L (P < 0.01); and sulfonylureas plus insulin, 31.5 (9-53) micrograms/L (P < 0.01). Once daily insulin was associated with values similar to those in the control group [62.0 (27-103) micrograms/L; P = NS], whereas IGFBP-1 levels were higher with multiple insulin injection therapy [156.0 (71-184) micrograms/L; P < 0.05]. Proinsulin levels were higher in sulfonylurea-treated patients, but there was no significant correlation between IGFBP-1 and proinsulin within any individual group. Study 2 examined the effects of treatment on the dynamics of IGFBP-1 levels between 0800-1900 h. In control subjects (n = 8), levels fell from 0800 h (mean +/- SEM, 22.4 +/- 5.2 micrograms/L) to 1000 h (14 +/- 5.2 micrograms/L), followed by a rise, more rapid after food, to a peak at 1240 h (20.6 +/- 3.7 micrograms/L). Levels then declined until 1500 h (10.7 +/- 2.9 micrograms/L), with a further postprandial peak at 1840 h (23.1 +/- 3.2 micrograms/L). Sulfonylurea therapy (n = 6) resulted in a complete loss of this pattern, with a marked fall in IGFBP-1 from 0800 h (22 +/- 2.7 micrograms/L) to less than 7 micrograms/L for the remainder of the study (area under the curve, 1150-1400 h, P < 0.001 vs. control). By contrast, in metformin-treated patients (n = 7), neither IGFBP-1 levels nor postprandial peaks were significantly different from those in the control group. Our findings suggest that in patients with NIDDM, the regulation of IGFBP-1 is markedly influenced by the choice of treatment.