Search

Present methods of monitoring for allograft dysfunction in pancreaticoduodenal transplantation (urinary amylase, scintigraphy) give indirect evidence and do not consistently allow recognition of early, potentially reversible rejection from other causes of allograft dysfunction. For better diagnosis of allograft dysfunction, adequate tissue specimens must be obtained without trauma to the allograft. We devised a cystoscopically directed needle biopsy technique to obtain tissue from the canine pancreaticoduodenal allograft. This technique is made feasible by a duodenocystostomy similar to that in human pancreaticoduodenal transplantation. The duodenum acts as a port of entry for the biopsy instruments, providing access to the pancreas. A 24F side-viewing nephroscope is used to view the duodenum. The tissue specimen is obtained with an 18-gauge, 350 mm Menghini aspiration needle inserted by way of the nephroscope through the duodenum and into the pancreas. Pancreaticoduodenal allotransplantation with a duodenocystostomy for exocrine drainage was performed in 18 dogs. Of 59 serial biopsy specimens obtained, 41 (69.5%) were adequate for pathologic evaluation. The principal complication associated with the technique was minimal bleeding at the biopsy site, occurring in 22 of the 59 biopsy procedures. This cystoscopically directed biopsy technique appears to be safe, often obtains adequate tissue for diagnosis, and may be directly applicable for distinguishing causes of allograft dysfunction in human pancreaticoduodenal transplantation.

Most techniques described in animal models of pancreatic transplantation use either segmental or autotransplants. We employ a technique of pancreaticoduodenal allotransplantation in the dog that closely resembles the operation used in humans. The arterial supply of the entire pancreatic graft is preserved by procuring a Carrel patch of aorta encompassing the origin of the celiac and the superior mesenteric arteries. Splenic, inferior pancreaticoduodenal, and superior pancreaticoduodenal arteries remain intact with the graft. Venous drainage is through a short segment of portal vein. A 6-cm cuff of duodenum is taken with the head of the pancreas. Engraftment proceeds by placing the allograft within the peritoneal cavity of the recipient. End-to-side vascular anastomoses are constructed to distal aorta and inferior vena cava. The duodenal cuff is anastomosed to the dome of the bladder for drainage and analysis of exocrine secretions and to provide a port of entry for cystoscopically directed needle biopsy. A total pancreatectomy is performed to induce a state of diabetes. The average operating time is 5 h. Twenty-two dogs have undergone allotransplantation using this technique. Six dogs had no complications and were sacrificed after meeting criteria of their study protocol. There were three technical failures, two arterial thromboses and one exsanguination, yielding an 86% rate of successful engraftment. Three other dogs died of intussusception and three dogs died of sepsis, one secondary to wound dehiscence and one due to inadvertent common bile duct ligation during pancreatectomy. Wound problems, four dehiscences and two superficial infections, occurred only in immunosuppressed dogs.

This canine pancreaticodunodenal autotransplantation model includes virtually the entire pancreas attached to a duodenal cuff (second portion). The blood supply is based on the superior pancreaticoduodenal artery and the venous outflow on the gastroduodenal vein. The vascular anastomoses are end-to-side to the external iliac artery and vein. Exocrine drainage is channeled through the bladder from the transplanted duodenum. This technique closely resembles whole-organ pancreas transplant in humans with a cystoduodenostomy. The bowel reconstruction consists of a Billroth I gastroduodenostomy and a cholecystoduodenostomy. The entire procedure is relatively free of major complications and may be performed in less than 4 hours.

To determine the nature and sequence of the histologic changes in the early rejection of pancreaticoduodenal allografts and to assess the correlation between pancreaticoduodenal biopsy findings and the pathologic changes in the graft, we performed serial cystoscopically directed needle biopsies of pancreaticoduodenal allografts in 18 dogs and compared the findings with the histologic changes in 16 autografts as controls. Tissue adequate for evaluation was obtained by the biopsy technique in 70% of instances. The earliest and most characteristic manifestation of rejection was diffuse mixed inflammatory infiltrates involving the pancreatic acinar tissue and duodenum. The biopsy findings correlated well with the changes in the resected pancreatic specimens. Cellular rejection in the duodenum correlated with rejection in the pancreatic graft. There were no changes in the autografts that resembled cellular rejection. We conclude that, in the canine model, cystoscopically directed needle biopsy of pancreaticoduodenal allografts consistently provides adequate tissue for the diagnosis of rejection; the status of the graft can be monitored by serial biopsies of pancreatic acinar tissue and, possibly, by serial biopsies of the duodenal wall alone.

A technique of diversion of the gastroduodenal vein in a canine model is described to compare long-term metabolic effects of systemic versus portal pancreatic endocrine drainage. The vein was transected at its entrance into the portal vein and either diverted to the inferior vena cava (systemic group) or reanastomosed to the portal vein (portal group). All remaining venous drainage of the pancreas was interrupted. An additional group of animals underwent laparotomy without manipulation of pancreatic vasculature (sham group). Fasting peripheral insulin and glucose values were determined 3 months postoperatively. Fasting insulin values were significantly higher in the systemic group (mean 10.7 +/- 1.06 U/ml) than in the portal (5.8 +/- 0.70, P = 0.002) and sham (6.4 +/- 0.68, P = 0.01) groups. Fasting glucose values were not significantly different in the three groups. At sacrifice, venous thrombosis was noted in one systemically diverted dog (6.7%). All other anastomoses were patent. No significant collateralization was apparent in any group. No significant complications were noted. This procedure simulates the hormonal milieu created by heterotopic pancreatic transplantation while preserving pancreatic innervation and exocrine function, and serves as an excellent model for investigating the effects of systemic hyperinsulinemia on protein, carbohydrate, and lipid metabolism.