This single-centre study investigated parameters that positively correlated with the success rate after islet allotransplantation in insulin-dependent diabetic (IDDM) patients. Twenty-one intrahepatic, fresh islet transplantations were performed in 20 IDDM patients (one patient had two transplants), after or simultaneous with kidney transplantation. The correlation between number and purity of transplanted islets and final outcome was investigated. One patient died of a cardiac arrest several hours after islet transplantation; this patient was not included in the follow-up analysis. Three patients (15 %) experienced acute, irreversible, early failure of islet function, which was considered as a ’presumed rejection'. Nine patients (45 %) achieved either complete insulin-independence (seven cases) or a reduction (> 50 %) of exogenous insulin requirement (two cases), with sustained serum C-peptide secretion (0.89 ± 0.04 nmol/l; duration: 21 ± 7 months, range 2–58 months). Liver biopsy, performed 3 years after transplantation in one successful case, showed normal islets within the hepatic parenchyma. Eight cases (40 %) did not show any metabolic effect of islet transplantation, with low serum C-peptide levels (’presumed function exhaustion'). Metabolic investigations performed in successful cases showed an early phase of insulin release after arginine, mild and reversible postprandial hyperglycaemia and normal HbA_1c levels. Success of islet transplantation positively correlates with the number (p < 0.05) of the transplanted islets. Islet transplantation is a safe procedure, with 45 % success rate, in terms of insulin-independence or relevant reduction of exogenous insulin requirement, although success can be transient. [Diabetologia (1997) 40: 225–231]