We assessed the metabolism of the two KBs, AcAc and 3-BOH; the relationships between ketogenesis and FFA inflow rate; and the effect of chronic sulfonylurea treatment in mild NIDDM patients (plasma glucose less than 10 mM). We studied 10 nonobese NIDDM patients in a crossover, randomized, double-blind, placebo-controlled fashion. Each patient was studied 4 times: after a run-in period with placebo, after 3 mo of placebo treatment, after 3 mo of glibenclamide treatments, respectively, and after 3 mo of sulfonylurea treatment during an acute exogenous Intralipid infusion. Ten normal, nondiabetic subjects served as the control group. Glibenclamide treatment decreased plasma FFAs. When these substrates were exogenously increased, plasma FFAs were comparable with placebo and baseline concentrations. In NIDDM patients, baseline and placebo blood total KB concentration was significantly higher than in control subjects (216 +/- 22 and 244 +/- 25, respectively vs. 127 +/- 18 microM; P less than 0.01). Glibenclamide treatment significantly decreased total KBs to 177 +/- 19 microM (P less than 0.05). When FFAs were exogenously increased, total KBs were similar to the placebo and baseline period. In the baseline study, the AcAc/3-BOH ratio was 0.72 +/- 0.06 in control subjects, whereas in NIDDM patients, the ratio was 1.61 +/- 0.13 at baseline (P less than 0.001 vs. control subjects), 1.66 +/- 0.15 during placebo, 1.57 +/- 0.09 during glibenclamide (NS vs. baseline), and 1.51 +/- 0.23 during glibenclamide plus placebo FFAs. Both the AcAc interconversion rate to 3-BOH and the 3-BOH interconversion rate to AcAc were significantly lower in NIDDM patients than in control subjects.