What We Do and Donʼt Understand About Pancreas Graft Failure.

Abstract

The purpose of this study was to analyze patterns of graft failure (GF) following pancreas transplantation (PT) in an attempt to develop a better understanding of GF. Methods: We retrospectively reviewed outcomes in 202 consecutive PTs in 192 patients at our center; all were insulin-requiring pretransplant and received either r-ATG or alemtuzumab induction with tacrolimus/mycophenolate ± steroids. GF was defined as a return to daily scheduled exogenous insulin therapy irrespective of dose. Results: From 11/01 to 3/13, we performed 162 simultaneous pancreas-kidney transplants (SPKT) and 40 solitary PTs. In the SPKT group, 51 patients had measurable pretransplant C-peptide levels. With a mean follow-up of 6 years, 76 (37.6%) pancreas GFs occurred including 20 deaths with functioning grafts (DWFG). The death-censored pancreas graft failure rate was 30.8%; of these 56 pancreas GFs, 20 occurred early (within 3 months of PT) and 36 occurred late. Rates of early thrombosis were 8% in SPKT and 5% in solitary PT patients. Death-censored causes of pancreas GF were acute/chronic rejection (34), thrombosis (18), infection (3) and one primary nonfunction. Allograft pancreatectomy (AP) was performed in 22 cases; rates of AP were 95% for early (<3 months) and 11.8% for late pancreas GF (p<0.001). After excluding DWFGs and pancreas GFs that resulted in AP, a total of 36 pancreas GFs remained, 10 of which did not have C-peptide data at the time of pancreas GF. Of the remaining 26 pancreas GFs, 13 occurred in patients who were C-peptide positive (mean pretransplant C-peptide level 5.7 ng/ml). In these 13 cases, the mean C-peptide level at the time of pancreas GF (insulin resumption) was 3.3 ng/ml (range 0.9-7.9). In all but one case, the C-peptide level at the time of pancreas GF was less than the pretransplant C-peptide level. Of 13 pancreas GFs in pretransplant C-peptide negative patients, the mean C-peptide level at the time of pancreas GF was 1.5 ng/ml (range 0-3.8); 5 had C-peptide levels ≥2.0, 4 had C-peptide levels 1-2, and 4 had C-peptide levels <1.0 ng/ml. Conclusions: DWFG and AP account for more than half of cases of pancreas GF. Of the remaining cases of GF (defined as loss of insulin independence), C-peptide levels appear to be of limited predictive value whether or not the patient had detectable C-peptide levels pretransplant.