Conventional drug therapy and several anti-CD154 mAb-based regimens were tested in the nonhuman primate (NHP) islet allograft model and found to be inadequate because islets were lost to rejection. Short-term therapy with an optimized donor-specific transfusion (DST) + rapamycin (RPM) + anti-CD154 mAb regimen enables immunosuppression drug-free islet allograft function for months following cessation of therapy in the NHP islet allograft model. After a substantial period of drug-free graft function, these allografts slowly and progressively lost function. Pathologic studies failed to identify islet allograft rejection as a destructive islet invasive lymphocytic infiltration of the allograft was not detected. To evaluate the mechanism, immunologic versus nonimmunologic, of the late islet allograft loss in hosts receiving the optimized therapeutic regimen, we performed experiments with islet autografts and studied islet function in NHPs with partial pancreatectomy. The results in both experiments utilizing autologous islet allografts and partially pancreatectomized hosts reinforce the view that the presence of a marginal islet mass leads to slowly progressive nonimmunological islet loss. Long-term clinically successful islet cell transplantation cannot be realized in the absence of parallel improvements in tolerizing regimens and in the preparation of adequate numbers of islets.