From August 1987 through February 1995 we performed 42 surgical procedures in 29 patients with occluded or stenotic radiocephalic arteriovenous fistulae. Operations were designed to preserve native veins for cannulation (Group I) or to preserve access in the same forearm, bypassing the failed fistula (Group II). For 27 procedures in 22 Group I patients, cumulative primary patency was 70%, 57%, and 47% at 6, 12, and 18 months, respectively. A subgroup of patients was identified, however, in whom excellent results could be reliably predicted. Among 19 hemodynamically stable patients with mature fistulae amendable to more proximal arteriovenous anastomoses, cumulative primary patency was 100%, 81%, and 67% at 6, 12, and 18 months, respectively. Secondary patency for 17 such patients was 100%, 89% and 89% for these same intervals. In Group II only two of ten patients required use of other access sites (9 1/2, 18 1/2 months). We believe that all occluded or stenotic radiocephalic arteriovenous fistulae should be considered for surgical salvage. Excellent results can be predicted for (1) hemodynamically stable patients with (2) mature fistulae that (3) fail near the arterial anastomosis and are (4) amendable to new more proximal arteriovenous anastomoses.

The quantitative insulin response to glucose stimulus can be drastically reduced by subtotal pancreatectomy. An 80% pancreatectomy was performed preserving the pancreatic duct in seven dogs. The insulin output into the portal vein and cephalic vein insulin after intravenous glucose challenge were measured. Output was recorded in 25 controls, and before and 2 wk after subtotal pancreatectomy in the seven animals. Histologic sections of the original resection were compared with the remnant of pancreas taken at the end of the study. In this model, which approaches islet cell failure in terms of glucose homeostasis, the tactics that permit enhancement of islet function can be discerned and to some degree quantitated. The pancreatic remnant does not oversecrete to approximate normal function, although glucose sensitivity is somewhat enhanced. No beta-cell hyperplasia was seen. Despite low insulin output into the portal vein, systemic insulin levels are conserved. This decrease in the plasma clearance of insulin supports glucose homeostasis. Accommodation to severe islet reduction occurs via both intrapancreatic and, more importantly, extrapancreatic mechanisms.

Successful intrasplenic islet autotransplantation in dogs requires an islet cell mass considerably greater than what might be expected based on studies of subtotal pancreatectomy. Grafts of marginal function ultimately fail, suggesting severe limitations in the capacity of an islet graft to adapt. Accommodation was tested in established intrasplenic grafts by either chronically stressing the graft with mild carbohydrate intolerance induced by exogenous corticosteroids or chronically suppressing the graft with exogenous insulin. After these manipulations, insulin output into the portal vein in response to intravenous (i.v.) glucose was measured and compared with that of normal dogs and dogs receiving islet autografts with no further treatment with either steroids or insulin. Transplanted islets tolerated the two manipulations well in that neither exogenous steroid nor insulin led to failure of the graft as a consequence of either stress or protracted diminished demand. The major determinant of successful islet grafting is the endocrine competence of the initial graft. If that competence is provided at the outset, the graft can adapt to a considerable range of demand for insulin secretion.

No abstract