The effects of metformin therapy on whole body and splanchnic amino acid turnover are not known. Therefore, we have studied fasting and postprandial phenylalanine kinetics in type 2 diabetic subjects (non-insulin-dependent diabetes mellitus), previously treated with diet only, both before and after 4 weeks of either metformin (850 mg twice a day) (n = 11) or placebo administration (n = 6). Phenylalanine kinetic was evaluated by means of a multiple isotope technique: tritiated phenylalanine was infused i.v., whereas carbon-labeled phenylalanine was incorporated into a chemically-defined meal. Compared with placebo, metformin administration decreased both fasting (from 162 +/- 17 to 141 +/- 20 mg/dl) and postprandial (from 217 +/- 20 to 164 +/- 20 mg/dl) glucose concentrations (P < 0.05-P < 0.01). Fasting insulin concentrations were unaffected, but postmeal insulin tended to be lower (P < 0.06) after metformin. Compared with the pretreatment period, metformin administration did not change total phenylalanine rate of appearance (fasted state, 0.74 +/- 0.10 vs. 0.71 +/- 0.08 mumol/kg.min; fed state, 0.77 +/- 0.10 vs. 0.75 +/- 0.08 mumol/kg.min, respectively), dietary and endogenous phenylalanine rate of appearance, dietary phenylalanine oxidation, and splanchnic uptake, similar to what was observed in the placebo group. Our data indicate that, at least after a 4-week treatment, metformin does not affect fasting and postprandial protein turnover, as indicated by phenylalanine data, in subjects with mild non-insulin-dependent diabetes mellitus.