PANCREAS TRANSPLANTATION - AN INITIAL EXPERIENCE WITH SYSTEMIC AND PORTAL DRAINAGE OF PANCREATIC ALLOGRAFTS

Pancreas transplantation has evolved dramatically since its introduction in 1966. As new centers for transplantation have developed, the evaluation of complications associated with pancreas transplantation has led to advances in surgical technique. Furthermore, surgical alterations of the pancreas resulting from transplantation (systemic release of insulin and denervation) are of unproven consequence on glucose metabolism. Since 1988, the authors have performed 21 transplants (16 combined pancreas/kidney, 3 pancreas alone, which includes 1 retransplantation, 1 pancreas after previous kidney transplant, and 1 "cluster") in 20 patients aged 18 to 49 years; mean, 35 +/- 1 years. Overall patient survival is 95%. Three pancreatic grafts failed within the first year because of technical failure; one additional pancreas was lost to an immunologic event on postoperative day 449, for an overall pancreatic graft survival of 81%. No renal grafts were lost. To evaluate causes of graft failure, demographic data were compared, which included age and sex of the donor and the recipient, operative time, intraoperative blood transfusion, and ischemic time of the graft. No statistically significant differences were found between groups except for ischemic time (11.7 +/- 6.4 hours for the technical success group versus 19.8 +/- 3.7 hours for the technical failure group; p < 0.05 by unpaired Student's t test). Quadruple immunosuppression was used, which included prednisone, cyclosporine, azathioprine, and antilymphoblast globulin. A mean of 1.2 (range, 0 to 3) rejection episodes per patient occurred. Mean hospital stay was 24 +/- 11 days. Surgical and infectious complications were evaluated by comparing the technical success (TS) group (n = 17) with the technical failure (TF) group. Surgical complications in the TS group revealed a mean of 1.3 episodes per patient, whereas the TF group had 3.7 episodes per patient. The TS also had a reduced incidence of infectious complications compared with the TF (1.7 versus 4.3 episodes per patient). Cytomegalovirus was common in both groups, accounting for 11 infectious episodes, and occurred on a mean postoperative day of 38. Mean postoperative HbA1C levels dropped to 5 +/- 1% from 11 +/- 3%. The authors developed a new technique that incorporates portal drainage of the pancreatic venous effluent in three recipients. Preoperative metabolic studies disclosed a mean fasting glucose of 211 +/- 27 mg/dL and a mean stimulated glucose value of 434 +/- 41 mg/dL for all patients; the mean fasting insulin was 23 +/- 4-mu-U/mL. After operation, the systemic drainage groups had mean fasting and stimulated glucose values of 87 +/- 7 mg/dL and 151 +/- 11 mg/dL, mean fasting and stimulated insulin values were 56 +/- 10-mu-U/mL and 190 +/- 26-mu-U/mL, respectively. The portal drainage group had mean fasting and stimulated glucose values of 91 +/- 5 mg/dL and 135 +/- 28 mg/dL; mean fasting and stimulated insulin values were 34 +/- 1-mu-U/mL and 88 +/- 27-mu-U/mL, respectively. In summary, pancreas-renal transplantation can be safely employed with an acceptable complication rate for the treatment of type I diabetes and may provide a euglycemic state. Systemic drainage of the pancreatic graft, however, is associated with hyperinsulinemia, which has unknown consequences. An alternative technique with portal drainage appears to reduce peripheral levels of insulin and has unproven long-term consequences on the potential detrimental effects of hyperinsulinemina.