Aim: The goal of this study was to explore whether donor and recipient’s characteristics, and one-year biological parameters could have a predictive value for the diagnosis of late pancreas graft failure.
Methods: 354 type 1 diabetic patients underwent pancreas transplantation between 2000 and 2015. Pancreas graft failure was defined as return to oral antidiabetics, permanent insulin-dependence. Recipient and donor characteristics were analyzed at the time of transplantation to determine potential risk factors for late graft dysfunction (after one year). Biological parameters such as HbA1c, fasting blood glucose, C-peptide, insulin, amylase, lipase, OGTT, Matsuda index, Beta score, HOMA IR and HOMA B, were analyzed one year after transplantation.
Results: 280 SPK, 38 PTA and 36 PAK were performed in 139 men and 215 women, with a mean age of 40 years. Median follow-up was 6 years. Patient survival was 96% at 1 year, 93% at 3 years and 87.5% at 5 years, whithout any significant difference between the three pancreas categories. Cardio-vascular diseases were the first cause of late recipient death. 111 patients experienced graft failure, 50 of them after the first year of transplantation. Pancreas survival was 80.5%, 70.7% and 67.5%, respectively at 1, 3 and 5 years, without any difference between the three categories. After the first year post-transplantation, death with a functioning graft occurred in 35.5% of patients with graft dysfunction. Donor death from cardiovascular cause was the only risk factor for late graft failure (OR 1,82 [1.05 ; 3.75], p=0,0345). One-year fasting blood glucose was the best predictive marker for a future graft dysfunction, especially when this value was over 5,4 mmol/l (OR 3.44 [1.80; 6.58] p=0.0002). 2-hour OGTT glucose level was also a risk factor for graft failure (OR 1.19 [1.04; 1.37] p=0,0109) while high Beta score (OR 0.62 [0.41; 0.94] p=0.0231) and high values of HOMA-B (OR 0.996 [0.992; 0.999] p=0,0483) were protective factors. Multivariate analysis identified a value of fasting blood glucose over 5.4 mmol/l (OR 4.24 [1.28; 14.04] p=0.0180) as the only significant risk factor. Graft survival was significantly different between patients with one-year blood glucose over or below 5.5 mmol/l: 85.2 vs 94.5% at 3 years and 72.3% vs 91.5% at 5 years (p<0,05), respectively.
Conclusion: Donor death from cardiovascular cause was the single risk factor for late graft failure among all donor and recipient’s base-line data. One-year fasting blood glucose level was the most significant biochemical parameter among patients at risk for graft failure, especially when the value was over 5,4 mmol/l. Two-hour OGTT glucose level, Beta score and HOMA B performed one year after transplantation are additional useful tools for identification of patients at risk. Whether medical intervention (i.e. GLP-1 analogue) among patients with still “normal glucose metabolism” could improve long-term graft survival requires prospective evaluation.