In order to determine if intraperitoneal insulin infusion could improve the insulin resistance of type 1 diabetic patients we have used the englycaemic insulin clamp technique in order to study the effects of insulin on glucose disposal in four C peptide negative type 1 diabetic patients treated by continuous subcutaneous or intraperitoneal insulin infusion and in five control subjects. Compared to control subjects, the diabetic patients treated by subcutaneous insulin infusion had a decreased maximal capacity of glucose utilization (diabetics: 12.6 +/- 0.3 mg.kg-1.min-1; controls: 15.7 +/- 0.7 mg/kg-1.min-1, p less than 0.01) and a trend towards higher half-maximally effective insulin concentrations (diabetics: 70 +/- 11 mU/l-1, controls: 48 +/- 4 mU/l-1). Treatment of the diabetic patients by intraperitoneal insulin infusion for 2 months decreased their mean peripheral free insulin levels (during subcutaneous infusion: 23.5 +/- 2.2 mU/l-1; during intraperitoneal infusion: 18.4 +/- 1.4 mU/l-1, p less than 0.05). However, mean daily insulin requirements were not decreased (during subcutaneous infusion: 0.59 +/- 0.05 U/kg-1.day-1; during intraperitoneal infusion: 0.57 +/- 0.03 U/kg-1.min-1). Moreover, the diabetic patients had a consistently lower maximal capacity of glucose utilization (12.6 +/- 0.7 mg kg-1.min-1) than control subjects (p less than 0.01) without modification of the half-maximally effective insulin concentration (62 +/- 10 mU.l-1). In conclusion, the only benefit of intraperitoneal insulin infusion was a reduction of peripheral free insulin levels; this decrease of peripheral insulinaemia was not associated with an improvement in the insulin resistance of diabetic patients.