Successful intraportal islet transplantation normalizes glucose metabolism in diabetic humans. To date, full function is not routinely achieved after islet transplantation in humans, with most grafts being characterized by only partial function. Moreover, the duration of full function is variable and cannot be sufficiently predicted with available methods. In contrast, most grafts retain partial function for a long time. We hypothesized that partial function can restore normal protein and lipid metabolism in diabetic individuals. We studied 45 diabetic patients after islet transplantation. Labeled glucose and leucine were infused to assess whole-body glucose and protein turnover in 1) 6 type 1 diabetic patients with full function after intraportal islet transplantation (FF group; C-peptide > 0.6 nmol/l; daily insulin dosage 0.03 +/- 0.02 U x kg(-1) body wt x day(-1); fasting plasma glucose < 7.7 mmol/l; HbA1c < or = 6.5%), 2) 17 patients with partial function (PF group; C-peptide > 0.16 nmol/l; insulin dosage < 0.4 U x kg(-1) body wt x day(-1)), 3) 9 patients with no function (NF group; C-peptide < 0.16 nmol/l; insulin dosage > 0.4 U x kg(-1) body wt x day(-1)), and 4) 6 patients with chronic uveitis as control subjects (CU group). Hepatic albumin synthesis was assessed in an additional five PF and five healthy volunteers by means of a primed-continuous infusion of [3,3,3-2H3]leucine. The insulin requirement was 97% lower than pretransplant levels for the FF group and 57% lower than pretransplant levels for the PF group. In the basal state, the PF group had a plasma glucose concentration slightly higher than that of the FF (P = 0.249) and CU groups (P = 0.08), but was improved with respect to the NF group (P < 0.01). Plasma leucine (101.1 +/- 5.9 micromol/l) and branched-chain amino acids (337.6 +/- 16.6 micromol/l) were similar in the PF, FF, and CU groups, and significantly lower than in the NF group (P < 0.01). During insulin infusion, the metabolic clearance rate of glucose was defective in the NF group versus in the other groups (P < 0.01). Both the basal and insulin-stimulated proteolytic and proteosynthetic rates were comparable in the PF, FF, and CU groups, but significantly higher in the NF group (P = 0.05). In addition, the PF group had a normal hepatic albumin synthesis. Plasma free fatty acid concentrations in the PF and FF groups were similar to those of the CU group, but the NF group showed a reduced insulin-dependent suppression during the clamp. We concluded that the restoration of approximately 60% of endogenous insulin secretion is capable of normalizing the alterations of protein and lipid metabolism in type 1 diabetic kidney recipients, notwithstanding chronic immunosuppressive therapy. The results of the present study indicate that "success" of islet transplantation may be best defined by a number of metabolic criteria, not just glucose concentration/metabolism alone.