Ex situ Perfusion of Pancreas for Whole-Organ Transplantation: Is it Safe and Feasible? A Systematic Review

Abstract

Introduction:

Pancreas transplantation is currently one of the best treatments proposed in highly selected patients with unstable and brittle type 1 diabetes. The objective of pancreas transplantation is to restore normoglycemia and avoid the occurrence of complications associated with diabetes. Graft pancreatitis and thrombosis, arising from ischemia reperfusion injuries, are major causes of graft loss in the postoperative period. Ex situ perfusion, in hypothermic or normothermic settings, allowed to improve ischemic reperfusion injury in other organ transplantations (kidney, liver, or lung). The development of pancreatic graft perfusion techniques would limit these ischemic reperfusion injuries.

Objective:

Evaluation of the safety and feasibility of ex situ perfusion of pancreas for whole-organ transplantation.

Methods:

English literature about pancreas perfusion was analyzed using electronic database Medline via PubMed (1950-2018). Exclusion criteria were studies that did not specify the technical aspects of machine perfusion and studies focused only on pancreas perfusion for islet isolation.

Results:

Hypothermic machine perfusion for pancreas preservation has been evaluated in nine studies and normothermic machine perfusion in ten studies. We evaluated machine perfusion model, types of experimental model, anatomy, perfusion parameters, flushing and perfusion solution, length of perfusion, and comparison between static cold storage and perfusion.

Conclusions:

This review compared ex vivo machine perfusion of experimental pancreas for whole-organ transplantation. Pancreas perfusion is feasible and could be a helpful tool to evaluate pancreas prior to transplantation. Pancreas perfusion (in hypothermic or normothermic settings) could reduce ischemic reperfusion injuries, and maybe could avoid pancreas thrombosis and reduce morbidity of pancreas transplantation.

Introduction

Pancreas transplantation is currently one of the best treatments proposed in highly selected patients with unstable and brittle type 1 diabetes. In 2015, 30.3 million Americans (9.4% of the population) presented diabetes and diabetes was the seventh leading cause of death in the United States, with 79 535 deaths. Approximately 1.25 million American children and adults suffer from type 1 diabetes.¹ The objective of pancreas transplantation is to restore insulin secretion, thus avoiding life-threatening hypoglycemia and limit the progression of macro- and micro-angiopathic complications. Simultaneous pancreas and kidney transplantations are commonly performed in selected patients with diabetic nephropathy and chronic renal dysfunction.

Improvement of surgical procedures and postoperative management has strongly reduced the morbidity and mortality of pancreas transplantation.^2,3 Recent data from UNOS (2010-2014) showed three-year graft survival rates of 82.2% for pancreas graft among simultaneous pancreas and kidney transplant recipients, 75.4% for pancreas after kidney, and 73.3% for pancreas transplantation alone in the United States.⁴ Whereas long-term outcomes have improved, graft losses during the first year (early graft failure) remained stable, between 10% and 15%. Graft pancreatitis and thrombosis are major causes of graft loss in the postoperative period, and these complications can occur in 25%-50% of cases.^5,6 Despite strong evidence for the implication of ischemia reperfusion injury in the occurrence of these complications, there has been no major improvement of pancreas transplant preservation technique. Standard method for human pancreas preservation is hypothermic preservation by static cold storage (SCS) for a limited period of time, usually less than 20 hours.⁷ Furthermore, there is no tool to evaluate pancreas viability before transplantation, except visual macroscopic evaluation.

Moreover, organ shortage and the growing demand for organs prompt clinicians to expand the donor pool by including extended criteria donors (ECD).^8,9 In kidney transplantation, the use of machine perfusion, in hypothermic or normothermic settings, has beneficial effects in term of delayed graft function and primary nonfunction, for transplant from ECD and/or donation from cardiac death (DCD).¹⁰ These techniques to preserve and perfuse organs in hypothermic and normothermic settings have been developed in liver, heart, and lung models also.

Improvement of early graft survival is a major issue in pancreas transplantation. Pancreas preservation appears to be critical to prevent thrombosis and reperfusion pancreatitis. Similarly to other organs, pancreas perfusion before transplantation may improve graft survival and could be useful to evaluate the organ. The development of pancreas transplantation is linked to the improvement of pancreatic graft perfusion techniques. Very few studies have focused on hypothermic pulsatile perfusion for pancreas preservation. The main objective of these experiments was not to transplant the organ but to isolate Langerhans islets.^11-13

The objective of this literature review is to evaluate the safety and feasibility of ex situ pancreas perfusion for whole-organ transplantation.

Methods

English literature about perfusion of pancreas was realized using electronic database Medline via PubMed (1950-2018).

Key terms used included “pancreas perfusion,” “pancreas preservation,” “normothermic pancreas perfusion,” “hypothermic pancreas perfusion,” and “ex vivo pancreas perfusion.”

Exclusion criteria were studies that did not specify the technical aspects of machine perfusion and studies focused only on pancreas perfusion for islet isolation. Review articles were screened for their references for further relevant studies. References in the identified articles were used to identify more relevant studies. The literature search was performed independently by two authors (TP, JB) and results were crosschecked. Data from included studies were extracted by the primary author (TP) (Figure 1).

Figure 1.

PRISMA flow diagram.

Results

Hypothermic machine perfusion for pancreas preservation was evaluated in nine studies and normothermic machine perfusion in ten studies. Hamaoui et al¹⁴ evaluated both hypothermic and normothermic perfusion.

Hypothermic Pancreas Perfusion

The summary of the experimental hypothermic perfusion of human and animals’ pancreas studies is summarized in Tables 1 and 2.

Table 1.

Experimental Hypothermic Perfusion of Human Pancreas Studies.

Author, journal, date	Temperature and machine model	Models	Experimental groups	Perfusion parameters	Flushing and perfusion solution	Length of perfusion	Results
Branchereau, J. et al, Cryobiology,15	4°C Waves	Human pancreas rejected for vascularized organ or islet transplantation DBD model	Control group (n = 2): pancreas preserved under SCS at (4°C) Split group (n = 2): pancreases were separated to compare the tail preserved under SCS and the head preserved under HPP Whole-organ group (n = 7): whole pancreases preserved under HPP	SP: 25 mmHg DP: 10 mmHg	Celsior solution for flushing IGL-1 solution for static storage Perf-Gen solution for perfusion	24 hours for three groups	Macroscopic examination: ES Control group: no edema (ES = 0) at 12 hours and slight edema at 24 hours (ES = 1) Split group: pancreas head: unintelligible/pancreas tail: ES = 1 at 24 hours Whole pancreas: at 24 hours, ES = 0 in 5 organs and ES = 1 in 2 organs Histological examination: Control group: at 12 hours, ischemic necrosis Split group: head (perfusion) and tail (static): no ischemic necrosis at 24 hours Whole pancreas: at 12 hours: no ischemic necrosis and at 24 hours: ischemic necrosis in three pancreases Immunohistochemical analysis: Insulin and glucagon: normal at 12 hours of HPP. “washed out” positive β-cells and α-cells in the ischemic islet after 24 hours of HPP Somatostatin: normal and similar after 12 and 24 hours of HPP PRI: Decline during the first 12 hours and then no variation of PRI
Leemkuil, M. et al, Transplant direct,16	4-7°C Dual pulsatile perfusion system with two centrifugal pumps (Deltastream DPII)	Human pancreas rejected for vascularized organ or islet transplantation DCD and DBD models	Two preservation groups: Experimental group (n = 10; DCD = 5; DBD = 5): six hours of HPP Control group (n = 10; DCD = 5; DBD = 5): six hours of SCS	SP: 25 mmHg	UW solution for flushing and perfusion AO (0.1 mg/L) was added to the perfusion solution in 6 out of 10 HPP pancreas	1) Period of static storage (no length of time reported) for all groups 2) Six hours of HPP (experimental group) or Six hours of SCS (control group)	Perfusion pressure: No significant difference in flow between the DCD and DBD pancreas Histological examination: AO staining was visible in all biopsies: uniform perfusion No significant differences: edema formation, acinar cell integrity loss between HPP and SCS in both DBD and DCD pancreas Immunohistochemical analysis: After SCS, ATP concentration decreased in DCD and DBD pancreas After HPP ATP concentration increased in DCD and DBD pancreas Biochemical analysis: Amylase, lipase, and LDH increased after HPP and SCS
Hamaoui, K. et al, J Surg Res,14	4°C or normothermia RM3 machine perfusion for all HPP NP machine was not detailed	Human pancreas rejected for vascularized organ or islet transplantation -DCD porcine models	Phase 1 (n = 4): three to seven hours of SCS and five hours of HPP of a segmental porcine pancreas (splenic lobe) Phase 2 (n = 6): Whole-organ porcine pancreases divided into two groups: -SCS-only group (n = 3): 27 hours of SCS and 2 hours of NP -SCS-HMP group (n = 3): 26 hours of SCS and 5 hours of HMP Phase 3 (n = 2): 26.8 and 56 hours of SCS and five hours of HPP and two hours of NP of whole-organ human pancreas	Phase 1: SP of 30 mmHg Phase 2: SCS-HMP group: SP <20 mmHg Phase 3: SP <20 mmHg NP SP: 45 mmHg for all groups	UW solution for flushing and perfusion	Period of static storage for all groups Phase 1: three to seven hours of SCS and five hours of HPP Phase 2: SCS-only group: 27 hours of SCS and 2 hours of NP SCS-HMP group: 26 hours of SCS and 5 hours of HPP Phase 3: 26.8 and 56 hours of SCS and five hours of HPP and two hours of NP	WG: -Phase 1: 46%-140% -Phase 2: 15.3%-27.6% -Phase 3: Minimal WG (HP1: 14.7% and HP2: 3.90%) Biochemical analysis: -Insulin Phase 2: SCS-only: peak insulin levels at 15 minutes and decline until the end of NP. SCS-HMP: peak insulin levels Phase 3: HP2: peak insulin levels at 15 minutes but none during NP. HP1: peak insulin levels at 15 and at 75 minutes; and then decrease. -Lactates: Phase 2: Levels stable during five hours of HPP and progressive rise during NP -Amylase: Phase 3: Increased during perfusion Histological examination: -Phase 2: post-SCS/pre-HPP: intact pancreatic cells. After HPP in SCS-HMP: histological examination = similar. After NP, injury in all pancreases: from mild moderate edema to severe necrosis -Phase 3: post-SCS/pre-HMP: normal pancreatic structures. After reperfusion: focal mild-moderate necrosis

Abbreviations: AO, Acridin orange; DBD, donation after brain death; DCD, donation from cardiac death; ES, edema scale; HPP, hypothermic pulsatile perfusion; PRI, pancreatic resistance index; SCS, static cold storage; SP, systolic pressure; UW, University of Wisconsin; WG, weight gain.

Table 2.

Experimental Hypothermic Perfusion of Animals’ Pancreas Studies.

Author, Journal, Date	Temperature and machine model	Models	Experimental groups	Perfusion parameters	Flushing and perfusion solution	Length of perfusion	Results
Karcz, M. et al, Exp Clin Transplant,17	4-10°C RM3 pulsatile perfusion machine	Porcine models (Landrace-cross pigs: 60-90 kg) DCD model	Whole-organ pancreas (n = 15): 135-165 minutes of SCS and 315 minutes of HPP	First 60 minutes of HPP: SP between 5 and 13 mmHg Maintenance period of perfusion (255 minutes): SP between 15 and 23 mmHg	UW solution for flushing and perfusion	Total perfusion period: 315 minutes WIT: between 23.2 and 27.5 minutes	PRI: -Time to best PRI: 3-65 minutes -PRI: decreased during HPP -Mean flow: 65 mL/min/100 g -WG after perfusion: 3.2%-18.3% Histological examination: Preperfusion pancreas: lobular damage (grade 3) and moderate islet cell damage Postperfusion pancreas: improvement in the tissue: lobular (grade 2) acinar cell damage, mild islet cell damage Mean damage: Significant postperfusion reduction in islet and acinar cell damage
Brynger, H. et al, Eur Surg Res,18	6-8°C Gambro perfusion machine with oxygenation	Adult mongrel dogs (11-22.5 kg) Pancreatectomy performed two to eight days before transplantation DBD model	Recipients (n = 20) and donors (n = 20) Recipients divided into three groups: -Group I (n = 9): recipients underwent allotransplantation with pancreas preserved under 24 hours of SCS -Group II (n = 7): recipients underwent allotransplantation with pancreas preserved under 24 hours of HP -Group III (n = 4): recipients receiving organs immediately transplanted	SP: 50/36-44 mmHg	Flushing solution: 20 mL of 1% solution of lidocainhydrochloride with 5000 IU heparin, 200 mL of dextran, 10% invert sugar, 1.4% sodium bicarbonate Group I: 125 mL of 20% human serum albumin, 310 mL of 0.9% NaCl, 60 mL of 5.5% glucose, 2.6 mEq KCl, 12.4 mEq MgSO4, 7-12 mEq NaHCO3, 100 mg hydrocortisone sodium succinate, 1 M IU benzylpenicillin sodium Group II: Buffered invert sugar solution	Group I: 24 hours of albumin perfusion	Edema evaluation: Group I: WG: 135%-275% Edema decreased after revascularization Group II and III: No edema Survival evaluation: -Group I (n = 3) and group II (n = 3): death due to bleeding -Group II (n = 1): death due to hypoglycemia -Group I (n = 1): death due to volvulus -Group III (n = 1): death due to rejection Biochemical analysis: Glucose: normal within 24 hours after transplantation. Lower blood glucose levels in group III. Glucose tolerance test: No differences between three groups
Florack, G. et al, J Surg Res,19	4°C Mox-100 pulsatile perfusion machine	Adult mongrel dogs (14-30 kg) DBD model	Dogs underwent total pancreatectomy and heterotopic segmental pancreatic (tail) autotransplantation Group I: (n = 20) Autotransplantation with fresh unpreserved pancreatic segmental autografts Group II: Collins’ solution: 24 hours (n = 12) or 48 hours (n = 10) of SCS prior to autotransplantation SGF-I solution: 24 hours (n = 12), 48 hours (n = 12) and 72 hours (n = 10) of SCS prior to autotransplantation Group III: HPP prior to autotransplantation with SGF-I during 24 hours (n = 12) and 48 hours (n = 8) and with SGF II during 24 hours (n = 12) and 48 hours (n = 10)	SP: 30 mmHg	Group I: Ringer’s solution for flushing Group II: Collins’ solution or with SGF-I for flushing and preservation Group III: SGF-I or SGF-II for flushing and preservation SGF-I: 400 mL SGF plasma + 100 mL human albumin (25%) + 8 mEq/L MgSO4 + 20 mEq/L KCl + 250 mg methylprednisolone + 5 g dextrose + 2 mL PSP dye SGF-II: 400 mL SGF plasma + 50 mL human albumin (25%) + 8 mEq/L MgSO4 + 20 mEq/L KCl + 250 mg methylprednisolone + 12.5 g mannitol + 50 UI insulin + 2 mL PSP dye	Group III: 2-4 and 48 hours with SGF-I and SGF-II solutions	Success rate and preservation failure rate: Long-term function rate Group I: 80% Group II: -24 hours Collins: 67% -48 hours Collins: 40% -24 hours SGF-I: 75% -48 hours SGF-I: 75% -72 hours SGF-I: 30% Group III: -24 hours SGF-I: 50% -48 hours SGF-I: 12% -24 hours SGF-II: 58% -48 hours SGF-II: 50% Preservation failure rate Group II: -24 hours Collins: 20% -48 hours Collins: 50% -24 hours SGF-I: 0% -48 hours SGF-I: 0% -72 hours SGF-I: 57% Group III: -24 hours SGF-I: 40% -48 hours SGF-I: 83% -24 hours SGF-II: 30% -48 hours SGF-II: 37% Amylase levels evaluation: Mean peak amylases (UI/L): -3953 ± 365 in Group I -4226 ± 327 in Group II -2988 ± 228 in Group III Lower peak with HPP Histologic examination in Group III: -Interstitial edema -At 4 and 12 weeks: acinar cells replaced by fibrous tissue/islets: normal
Tersigni, R. et al, Ann Surg,20	6°C Mox-100 pulsatile perfusion machine	Adult mongrel dogs (15-25 kg) DBD model	Donors (n = 25) and recipients (n = 25) Model of whole-organ pancreatic allotransplantation Group I (control) (n = 5): allotransplantation with fresh transplant Group II: (n = 5) 24 hours of HPP at 5 mmHg with CPP prior to allotransplantation Group III (n = 5): 24 hours of HPP at 10 mmHg with CPP prior to allotransplantation Group IV (n = 5): 24 hours of HPP at 25 mmHg with CPP prior to allotransplantation Group V (n = 5): 24 hours of HPP at 10 mmHg with MCPP prior to allotransplantation	Group II: 5 mmHg Group III: 10 mmHg Groups IV and V: 25 mmHg	Flushing solution: Ringer’s lactate Perfusate solution: cryoprecipitated CPP and MCPP CPP: 200 mL distilled water + 2 mg/L methylprednisolone + 170 IU/L insulin + 1M IU penicillin + 8 mL/L MgSO4 20% + 2.5 mL/L KCl 10% + 30-50 mL mannitol MCPP: same composition than CPP with 200 mL/L of human albumin and 5-7 mg/L of dextrose and without mannitol	24 hours of perfusion in all groups	PRI: Groups IV and V: progressive decrease in PRI Groups II and III: progressive increase in PRI Biochemical analysis: Significantly higher secretion in Group V. Minimal secretions from Groups II, III, and IV Mean survival analysis: Group I: 14.6 days Groups II, III, and IV: shorter mean survival Group V: same mean survival than in Group I
de Gruyl, J. et al, Br J Surg,21	6-10°C Laboratory model of the Belzer machine	Beagles dogs mismatched for the determinants of the DL-A system DBD model	Recipients (n = 19) underwent pancreatectomy 3 weeks before transplantation Donors (n = 19) Group A (n = 9): Allotransplantation with fresh pancreas Group B (n = 5): 24 hours of HP prior to allotransplantation Group C (n = 5): 24 hours of SCS prior to allotransplantation	Group B: SP of 60 mmHg	Group B: Flushing and preservation with cryoprecipitated plasma + 0.115 L distilled water + 104 IU/L insulin + 575 000 IU/L penicillin + 0.007 g/L decadron + 0.115 g/L KCl + 0.870 g/L MgSO47H2O + 3.34 g/L mannitol Group C: Flushing and preservation with Collins’ solution	24 hours of perfusion	All allografts showed immediate function Biochemical analysis: Group A: Increased insulin peak level after transplantation, and not in Group B and C Mean survival time: Group A: 9.6 days Group B: 10.6 days Group C: 7.2 days Histologic analysis: -Interlobular edema and necrosis -In grafts survival greater than nine days: replacement of acinar by fibrous
Kenmochi, T. et al, Transplantation,22	ORPH3000C organ perfusion machine	Mongrel dogs (6-16 kg) DCD model	Donors (n = 25) Model of whole-organ pancreatic allotransplantation Group I: (n = 4) WIT = 0 minute Group II: (n = 5) WIT = 15 minutes Group III: (n = 9) WIT = 30 minutes Group IV: (n = 7) WIT = 60 minutes 13 allotransplantation with 13 grafts (Group II, n = 2; Group III, n = 6, Group IV, n = 5)	50 mmHg for all groups	Flushing solution: Collins’ solution Preservation solution: 500 mL cryoprecipitated fibrinogen-free plasma	1 hour of perfusion in all groups	WG: Group IV: 53% ± 11% and higher than in other groups Biochemical analysis: Amylase release (IU/g pancreas): -6.13 in Group I -4.23 in Group II -2.83 in Group III -2.23 in Group IV Histologic examination: Edema and dilatation of pericapillary space were detected in both Groups III and IV after perfusion Fasting blood sugar: Group A (<100 mg/dL): 7/13 dogs Group B (>100 mg/dL): 6/13 dogs Length of WIT: not predicted grafts’ prognosis

Abbreviations: CPP, pooled canine plasma; DBD, donation after brain death; DCD, donation from cardiac death; HPP, hypothermic pulsatile perfusion; MCPP, modified cryoprecipitated plasma; PRI, pancreatic resistance index; SCS, static cold storage; SGF, silica gel–filtered plasma solution; SP, systolic pressure; UW, University of Wisconsin; WG, weight gain; WIT, warm ischemia time.

Human experimental hypothermic pancreas perfusion studies

Branchereau et al,¹⁵ Leemkuil et al,¹⁶ and Hamaoui et al¹⁴ reported human experimental hypothermic perfusion studies.

Branchereau et al¹⁵ used human pancreas rejected for vascularized organ or islet transplantation. They compared three groups: control group (n = 3) with whole-organ pancreases preserved under conventional SCS during 24 hours; split group (n = 2) with pancreases separated in order to compare the tail, preserved under SCS during 24 hours, and the head, preserved under hypothermic pulsatile perfusion (HPP) during 24 hours and the whole-organ group (n = 7) with whole-organ pancreases preserved under HPP during 24 hours. The HPP machine was a Waves machine perfusion. The systolic pressure (SP) was 25 mmHg in all groups. The SCS solution was IGL-1 and the HP solution was Perfgen. Concerning macroscopic examination, with edema scale, they reported no edema at 12 hours and slight edema at 24 hours in the control group. In the split group, slight edema occurred at the head of the pancreas after 24 hours. In the whole pancreas group, after 24 hours of HP, no edema occurred in five pancreases and two transplants presented slight edema. Concerning histological examination, in the control group, after 12 hours, multiple ischemic necrosis occurred. In the split group, no ischemic necrosis occurred, after 24 hours, in the head and the tail of the pancreases. In the whole-organ group, after 12 hours, no ischemic necrosis occurred and after 24 hours, ischemic necrosis was present in three pancreases. Concerning immunohistochemical analysis, insulin and glucagon concentration was normal at 12 hours of HP. At 24 hours of HPP, a “washed out” positive β-cells and α-cells in the ischemic islet occurred. Somatostatin concentration was normal after 12 and 24 hours of HPP. Concerning pancreatic resistance index (PRI), during the first 12 hours, PRI declined and then no variation was reported.

Leemkuil et al¹⁶ compared six hours of SCS (n = 10) and six hours of HPP (n = 10) of human pancreases rejected for vascularized organ or islet transplantation. They had both DCD and donation after brain death (DBD) pancreases. The HPP machine was a dual pulsatile perfusion system (Deltastream). The SP was 25 mmHg in all groups. The SCS and HPP solution was UW and they added Acridin orange, in the solution, to evaluate the perfusion. Concerning perfusion pressure, they reported no significant differences in flow between DCD and DBD pancreases. Concerning histological examination, they reported no significant differences in terms of edema formation, and acinar cell integrity loss between HPP and SCS in both DBD and DCD pancreases. Acridin orange staining was visible in all biopsies, confirming the uniform perfusion. Concerning immunohistochemical examination, after six hours of SCS, ATP concentration decreased in DCD and DBD pancreas, while, after six hours of HPP, ATP concentration increased in DCD and DBD pancreas. Concerning biochemical examination, amylase, lipase, and LDH increased after HPP and SCS.

Hamaoui et al¹⁴ used both human pancreas rejected for vascularized organ or islet transplantation and DCD porcine pancreas. They compared three groups: phase 1 (n = 4) with segmental porcine pancreases (splenic lobe) preserved under HPP during five hours after a period of three to seven hours of SCS; phase 2 (n = 6): whole-organ porcine pancreases divided into two subgroups: SCS-only group (n = 3) with whole-organ porcine pancreases preserved under two hours of NP after 27 hours of SCS and SCS-HMP group (n = 3) with whole-organ porcine pancreases preserved under five hours of HPP after 26 hours of SCS; phase 3 (n = 2) with whole-organ human pancreas preserved under five hours of HPP and then two hours of NP after an initial period of 26.8 and 56 hours of SCS. The HPP machine was a RM3 and the NP machine was not detailed. The SP of HPP was 30 mmHg in phase 1, and <20 mmHg in SCS-HMP group (phase 2) and in phase 3. The SP of NP was 45 mmHg in all groups. The SCS and HPP solution was UW. The warm ischemia time (WIT) in phase 1 and 2 were 30-55 and 30 minutes. The WG was lower in phase 3 (HP1: 14.7% and HP2: 3.9% versus 46%-140% in phase 1 and 15.3%-27.6% in phase 2). Concerning biochemical analysis, both phase 2 and phase 3 pancreases presented peak insulin levels after the beginning of the HPP. In phase 2, lactates levels remained stable during HPP and raised during NP. In phase 3, amylase levels increased during HPP and NP. Concerning histological examination, in phase 2, after 27 and 26 hours of SCS, no pancreatic cell damages occurred. After HPP, no pancreatic cell damages were reported while after NP, moderate edema to severe necrosis was reported. In phase 3, after 26.8 and 56 hours of SCS, a normal pancreatic structure was reported. After HPP and NP, moderate necrosis occurred.

Animal experimental hypothermic pancreas perfusion studies

Machine perfusion system

Different types of machine perfusion were used: RM3 Machine Perfusion Unit (Waters Medical Systems, Rochester, MN, USA),^14,17 Gambro perfusion system with oxygenator,¹⁸ Mox-100 machine perfusion (Waters Medical System, Rochester, MN, USA),^19,20 ORPH3000C organ perfusion machine (Senko Medical Mfg. Co., Tokyo, Japan),²² and an experimental model of Belzer machine.²¹

Experimental models

Experimental models were porcine in one study,¹⁷ mongrel dog in four studies,^18-20,22 and bagel dog in one study.²¹ Five studies were conducted with DBD model, and four studies with DCD model.

Perfusion parameters

Karcz et al¹⁷ perfused porcine pancreas at an SP between 15 and 23 mmHg. They reported minimal WG and an improvement in histologic islet and acinar cell damages after 315 minutes of perfusion. Higher SP (50/36-44 mmHg) was used by Brynger et al. WG was more important in these dogs’ pancreases (135%-275%).¹⁸ Florack et al¹⁹ reported that use of high SP (45 and 60 mmHg) of dogs’ pancreases resulted in severe edema. At 30 mmHg, all grafts developed interstitial edema during perfusion (at 24 hours and more important at 48 hours). Tersigni et al²⁰ compared SP of 5-10 and 25 mmHg in dogs’ pancreases. Pancreatic secretions increased in pancreases perfused at SP of 25 mmHg. de Gruyl et al and Kenmochi et al^21,22 perfused dogs’ pancreases with high SP (60 and 50 mmHg). They reported an immediate function²¹ and a minimal weight gain with edema.²²

Flushing and perfusion solution

Karcz et al¹⁷ used University of Wisconsin solution for perfusion of porcine pancreases. Brynger et al¹⁸ developed an experimental perfusion solution with albumin, NaCl, glucose, KCl, MgSO₄, NaHCO₃, hydrocortisone sodium succinate, and benzylpenicillin. Likewise, Florack et al¹⁹ developed an experimental perfusion solution (Silica Gel-Filtered plasma solution: SGF-I and SGF-II). SGF-I contained plasma, human albumin, MgSO₄, KCl, methylprednisolone, dextrose, and PSP dye. SGF-II had the same composition as SGF-I with adjunction of mannitol and insulin. Tersigni et al²⁰ used experimental pooled canine plasma (CPP) and modified cryoprecipitated plasma (MCPP) perfusion solution. CPP contained water, methylprednisolone, insulin, penicillin, MgSO₄, KCl, and mannitol. MCPP had the same composition as CPP with human albumin and dextrose and without mannitol. Pancreatic secretions (amylase, trypsin, and lactic acid) were significantly higher in pancreases perfused with MCPP. These pancreases presented less edema. de Gruyl et al²¹ used an experimental (plasma, water, insulin, penicillin, decadron, KCl, MgSO₄7H₂O, and mannitol) solution in Beagles dog model. The mean survival of dogs’ recipients was longer with grafts preserved under hypothermic perfusion using experimental solution (10.6 days versus 7.2 days).

Length of perfusion

Florack et al¹⁹ compared 24 and 48 hours of perfusion in a dog model. They reported a long-term function rate of 50%, 58%, 12%, and 50% in groups perfused during 24 hours with SGF-I, 24 hours with SGF-II, 48 hours with SGF-I, and 48 hours avec SGF-II, respectively. The preservation failure rate was 40%, 30%, 83%, and 37% in groups perfused during 24 hours with SGF-I, 24 hours with SGF-II, 48 hours with SGF-I, and 48 hours avec SGF-II, respectively. Kenmochi et al²² compared different warm ischemia time (0, 15, 30, and 60 minutes) prior to 1 hour of HP in dog models. Long warm ischemia period was correlated with higher WG and more severe histologic lesions.

Pancreas transplantation

Brynger et al and Tersigni et al reported dogs pancreases allotransplantation experiences.^18,20 Brynger et al compared dogs pancreases allotransplantation after 24 hours of perfusion with albumin or buffered sugar solution, in hypothermic conditions.¹⁸ Tersigni et al compared pancreases allotransplantation after 24 hours of HP, at different SP values: 5, 10, and 25 mmHg, in dog models. Mean survival time was shorter for recipients transplanted with pancreases preserved in HP compared to recipients transplanted with fresh transplant without perfusion.²⁰ Florack et al published a model of dog’s pancreases autotransplantation. They compared pancreases autotransplantation, after 24 and 48 hours of HP or SCS of grafts. Long-term function rate was 75% after 24 hours of SCS and 50% after 24 hours of HP. After 48 hours of SCS, long-term function rate was 75%, and 12% after 48 hours of HP.¹⁹ de Gruyl et al compared pancreas allotransplantation, in dog models, after 24 hours of HP or 24 hours of SCS. Mean survival time was longer for recipients transplanted with grafts preserved with HP (10.6 days versus 7.2 days).²¹ Kenmochi et al compared pancreas allotransplantation, of grafts with different WIT (0, 15, 30, and 60 minutes), prior to one hour of HP. Grafts pancreases with long WIT (30 and 60 minutes) presented edema after one hour of perfusion.²²

Normothermic Pancreas Perfusion

The summary of the experimental normothermic perfusion of human and animals’ pancreas studies is presented in Tables 3 and 4.

Table 3.

Experimental Normothermic Perfusion of Human Pancreas Studies.

Author, Journal, Date	Temperature and machine model	Models	Experimental groups	Perfusion parameters	Flushing and perfusion solution	Length of perfusion	Results
Barlow, AD. et al, Am J Transplant,23	37°C NP: pediatric cardiopulmonary bypass technology + organ chamber + venous reservoir + centrifugal blood pump + fiber membrane oxygenator and a heat exchanger	Human pancreases (n = 5) unsuitable for transplantation by all UK pancreas transplant centers DBD model	NP group (n = 5): 13 hours of SCS and 2 hours of NP	SP: 50-55 mmHg	SCS solution: UW solution Flushing solution: gelofusine solution Perfusion solution: Blood-based donor ABO-compatible+ Gelofusine solution + 8.4% sodium bicarbonate + 20 mL 10% mannitol + 5 mL 5% glucose + 2500 UI heparin	Two hours of perfusion	One pancreas was excluded due to necrosis (CIT = 30 hours) Perfusion parameters: All pancreases appeared well perfused during NP Mean arterial flow: 35 ± 2.8 mL/min/100 g Biochemical parameters: Amylase and lipase: Increased in all pancreases during NP Insulin: Secretion was higher in the pancreas from young donors (27 years old) Histologic analysis: Three pancreases: focal acinar cell necrosis of body and tail One pancreas: extensive parenchymal and fat necrosis
Nassar, A. et al, Artif Organs,24	37°C Modified liver perfusion model	Human pancreases (n = 3) unsuitable for transplantation DBD model	NP group (n = 3): four to six hours of SCS and 6 hours (n = 2) or 12 hours (n = 1) of NP	SP: 60 mmHg Perfusion rate: 55 mL/min/100 g tissue	SCS solution: HTK solution Flushing solution: unspecified Perfusion solution: red blood corpuscles and plasma in a 1:3 ratio	6 hours (n = 2) and 12 hours (n = 1)	Biochemical analysis: C peptide levels increased from 2.7 to 18 ng/mL at six hours Histologic analysis: At six hours: healthy acini and islets Chromogranin: normal staining At 12 hours: healthy acini and islets

Abbreviations: CIT, cold ischemia time; DBD, donation after brain death; SCS, static cold storage; SP, systolic pressure; UW, University of Wisconsin.

Table 4.

Experimental Normothermic Perfusion of Animals’ Pancreas Studies.

Author, Journal, Date	Temperature and machine model	Models	Experimental groups	Perfusion parameters	Flushing and perfusion solution	Length of perfusion	Results
Kuan, K.G. et al, Artif Organs. 25	37°C Extracorporeal roller pump + venous reservoir + membrane oxygenator + metal organ chamber + water bath temperature regulator	White pig (n = 4) (40-50 kg) DBD model	Pancreases alone (n = 2) preserved under NP Pancreases and kidneys (n = 2) preserved under NP	SP: 70-80 mmHg for pancreas and 90-100 mmHg for kidney Flow rate: 0.37 L/min for pancreas only and 0.46 L/min for pancreas and kidney perfusion.	Flushing solution: Hartmann’s solution Perfusion solution: autologous whole blood + 1 g cephazolin + 500 µog epoprostenol sodium + 5000 IU heparin	First pancreas only: four hours Three other experiments: two hours WIT = 8.25 ± 6.55 minutes CIT = 34 ± 7.78 minutes	Macroscopic evaluation: All pancreases: moderate to severe edema after 90 minutes of NP Histologic analysis: All pancreases: acinar damages, inflammation and thrombosis after NP Islet cells: few cytoplasmic microvacuolation (progressive with perfusion period)
Eckhauser, F et al, J Surg Res,26	Perfusion apparatus: heating chamber + water bath + oxygenator + humidifier + pulsatile infusion pump + dialysis unit	Dogs (20-35 kg) (n = unspecified) DBD model	NP group: four to five hours of NP for whole-organ pancreas	SP: 90-110 mmHg initially, then declined to 30-50 mmHg Flow rate: 20 ± 5 mL/min	Flushing solution: unspecified Perfusion solution: balanced electrolyte solution + 4 g/dL bovine serum albumin + 150 mg/dL glucose + 12.5% mannitol + 1000 IU heparin + autologous red blood cells	Perfusion: four to five hours	Histologic analysis: No alteration in acinar pattern, nuclei, or cytoplasm Edema: constant occurrence Biochemical analysis: Infusion of 3 U/min of Boots secretin: rate of pancreatic secretion flow increased from 0.035 to 0.28 mL/min
Eloy, M.R. et al, Eur Surg Res,27	39-40°C Perfusion apparatus: blood reservoir + rotating disc oxygenator + roller pump + heat exchanger + small balloon immersed in a modulator (obtain pulsatile flow) + pressure transducer + ultrasonic flowmeter + electrothermometer	Dogs (15-25 kg) (n = 15) DBD model	NP group: 100 minutes of NP for whole-organ pancreas	Flow rate: 38-46 mL/min/100 g SP: unspecified	Flushing solution: unspecified Perfusion solution: Homologous erythrocytes + culture medium + 1 g/L Earles’ salts and glucose + 40 g/L bovine albumin + 50 g/L dextran	Perfusion: 100 minutes	Endocrine response to increased glucose concentrations: Different glucose concentrations: 420, 800, and 1700 mg/100 mL Insulin released at a constant rate (17 000 μU IRI/min), and not increased until arterial glucose concentration reached 140 mg/mL Exocrine response to prolonged hormonal stimulation: Infusion of secretin and cholecystokinin: Increase of secretion to 2.4 mL/10 min post stimulation
O’Malley, V.P. et al, J Surg Res,28	37°C Perfusion apparatus: heated perfusion chamber + oxygenator + pressure transducers	Dogs (18-25 kg) (n = 24) DBD model	Group I: Blood-perfused controls (n = 6) Group II: Blood-perfused injury (pancreatitis) (n = 6) Main pancreatic duct cannulated small-caliber cannula Group III: Fluosol-perfused controls (n = 6) Group IV: Fluosol-perfused injury (pancreatitis) (n = 6) Main pancreatic duct cannulated small-caliber cannula	SP: Group I: 61 mmHg at four hours Group II: 95 mmHg at four hours Group III: 66 mmHg at four hours Group IV: 76.6 mmHg at four hours Flow rate: 20 mL/min	Flushing solution: unspecified Perfusion solution: Fluosol-perfused groups: Fluosol (FC-43) + 500 mg glucose + 2.5 g albumin Blood-perfused groups: 300 mL autologous blood + 6000 IU heparin + 2.5 g albumin + 500 mg glucose	Perfusion: -30 minutes for stabilization -Then, four hours	Gross appearance: -Group I: normal -Group II: edema and hemorrhage -Group III: mild edema -Group IV: mild edema Amylase concentration at four hours: -Group I: 4253 ± 442 IU/L -Group II: 18 543 ± 3434 IU/L -Group III: 1285 ± 245 IU/L -Group IV: 1961 ± 834 IU/L Histologic analysis: Differences between control and injury, but not between blood and Fluosol-perfused glands
Wahlberg, J. et al, Transpl Int,29	37°C Ex vivo perfusion system: roller pump + heat exchanger + membrane oxygenator	Dogs (15-25 kg) (n = 26) DBD model	Group I (controls; n = 6): NP immediately after procurement Group IIA (n = 4): 24 hours of SCS and NP without allopurinol Group IIB (n = 6): 24 hours of SCS and NP with allopurinol Group IIIA (n = 4): 48 hours of SCS and NP without allopurinol Group IIIB (n = 6): 48 hours of SCS and NP with allopurinol	SP: 30 mmHg Flow rate: 1 mL/min per gram of tissue	Flushing solution: UW solution Perfusion solution: UW solution + 5 g dextran	Perfusion: two hours CIT: 24-48 hours	WG: -Group I: 35% ± 7% -Group IIA: 61% ± 8% -Group IIB: 76% ± 13% -Group IIIA: 80% ± 8% -Group IIIB: 81% ± 9% Amylase concentration at two hours: Group I: 2215 ± 330 IU/dL -Group IIA: 4373 ± 622 IU/dL -Group IIB: 4130 ± 665 IU/dL -Group IIIA: 6260 ± 1120 IU/dL -Group IIIB: 5420 ± 1181 IU/dL
Loubatières-Mariani, M.M. et al, Diabetologia,30	37.5°C or 28°C Perfusion system unspecified	Wistar rats (350 g) (n = unspecified) DBD model	Effects of hypothermia on insulin secretion examined: Glucose stimulation with low glucose concentration (1.5 g/L) or high glucose concentration (3 g/L) under hypothermia (28°C) and normothermia (37.5°C)	SP: 35 cmH2O in 37°C group 37.5 cmH2O in 28°C group Flow rate: 2.4 mL/min	Perfusion solution: Glucose solution (1.5 g/L) for 45 minutes And Krebs–Ringer bicarbonate buffer + 2 g/L purified bovine albumin + 0-3 g/L glucose	Perfusion: 90 minutes	Insulin secretion induced by 1.5 g/L glucose: 8.2 ng/mL at 28°C −41.4 ng/mL at 37°C Insulin secretion induced by 3 g/L glucose: -25 ng/mL at 28°C -240 ng/mL at 37°C
Pegg, D.E et al, J Surg Res,31	38°C Single-pass method: reservoir + pump + filter + buble trap duplicated: one system containing perfusate with 5 mM glucose and the other 30 mM glucose Circuits terminated in a two-way tap: connected to pancreatic cannula or to be returned to its own reservoir	WAG rats (130-170 g) (n = 45) DBD model	Recirculation method: (n = 5) Pancreases preserved under recirculation method for two hours Single-pass method: (n = 40) Pancreases preserved under single-pass method for two hours	Pressure in recirculation method: 60 mmHg Flow rate in single-pass method: 2 mL/min	Perfusion solution recirculation method: Gelatin polypeptide Haemaccel + 5.4 mM calcium + 5 mM glucose + 3 M glucose at one hour of perfusion Perfusion solution single-pass method: same solution with 5 mM glucose the first hour and solution with 30 mM for the remaining hour	Perfusion: two hours	Recirculation method: WG: wet/dry weight ratio = 8.2 ± 0.8 Single-pass method: WG: wet/dry weight ratio = 6.6 ± 0.3

Abbreviations: CIT, cold ischemia time; DBD, donation after brain death; SCS, static cold storage; SP, systolic pressure; UW, University of Wisconsin; WG, weight gain; WIT, warm ischemia time.

Human experimental normothermic pancreas perfusion studies

Barlow et al²³ and Nassar et al²⁴ reported human experimental normothermic perfusion studies.

Barlow et al evaluated the impact of 13 hours of SCS followed by 2-hour NP in human pancreases (n = 5) unsuitable for transplantation. They used a pediatric cardiopulmonary bypass technology with organ chamber, venous reservoir, centrifugal blood pump, and a heat exchanger. The SP was 50-55 mmHg. They used UW as SCS solution, gelofusine as flushing solution, and a typical NP solution (blood-based donor ABO-compatible, gelofusine, sodium bicarbonate, mannitol, glucose, and heparin) as perfusion solution. One pancreas was excluded due to prolonged cold ischemia time (CIT; 30 hours). They reported a good perfusion during NP, with a mean arterial flow of 35 ± 2.8 mL/min/100 g. Concerning biochemical parameters, amylase and lipase levels increased in all pancreases during NP. Insulin secretion was higher in the pancreas from young donors (27 years old). Concerning histologic analysis, three pancreases presented focal acinar cell necrosis of the body and the tail and one pancreas presented extensive parenchymal and fat necrosis.

Nassar et al evaluated the impact of 4-6 hours of SCS followed by 6 hours (n = 2) or 12 hours (n = 1) of NP in human, pancreases unsuitable for transplantation. They used a modified liver perfusion model. The SP was 60 mmHg, with a perfusion rate of 55 mL/min/100 g. They employed HTK as SCS solution and a perfusion solution containing red blood corpuscles and plasma. Concerning biochemical analysis, C-peptide levels increased from 2.7 to 18 ng/mL at six hours. Concerning histologic analysis, at 6 and 12 hours of NP, they reported healthy acini and islets.

Animal experimental normothermic pancreas perfusion studies

Machine perfusion system

The first experimental normothermic pancreas perfusion was realized in 1926 by Babkin and Starling, using a dog model. Several normothermic machine perfusions were presented in the studies included. All normothermic machine perfusions had perfusion pump (pulsatile or centrifugal), oxygenator, blood reservoir, and heat exchanger (e.g. circulating water bath). Dialysis unit was added in three studies, and in three studies, the data were missing. Pegg et al³¹ used a single-pass method for their experiments, which included two systems of normothermic perfusion, which allowed to use two different perfusion solutions.

Experimental models

Experimental models were porcine models in two studies,^14,25 adult mongrel dogs in four studies,^26-29 and rat models in two studies.^30,31 Nine studies were conducted with DBD model and one study with DCD model.

Perfusion parameters

Eckhauser et al²⁶ started dogs’ pancreas perfusion at an SP of 90-110 mmHg and then declined to 30-50 mmHg. O’Malley et al²⁸ compared different SP in dogs’ pancreases models (61, 66, 76.6, and 95 mmHg). WG was not correlated with SP. Pegg et al³¹ used WAG rats’ models, perfused using an SP of 60 mmHg.

Flushing and perfusion solution

Kuan et al²⁵ used Hartmann’s solution as SCS and autologous whole blood associated with cephazolin, epoprostenol sodium, and heparin for perfusion. O’Malley et al²⁸ evaluated adjunction of Fluosol, an oxygen-carrying emulsion in perfusion solution used in their dogs’ model. WG was more important in the Fluosol group and no histologic differences have been highlighted between Fluosol- and blood-perfused groups.

Length of perfusion

Studies included reported perfusion durations from 90 minutes to 12 hours.

Discussion

Despite almost a 50-year history of pancreatic perfusion, SCS remains the gold standard conservation. This review included studies of pancreas perfusion, in hypothermic and normothermic settings, in the objective of whole-organ transplantation. We have shown that there are important hurdles to overcome for pancreatic perfusion. Of utmost importance is the use of a low perfusion pressure during a short perfusion period (<12 hours). As pancreas is a low vascularized organ and very sensitive of barotraumatism lesions, the use of low perfusion pressure is essential to avoid training of edema.

The contribution of the perfusion would be to assess the organ prior to transplantation with regard to steatosis and the likelihood of post-transplant pancreatitis and to estimate β-cell viability and function of the organ.

Thus, these experimental studies provide elements of answers about adequate viability markers of the pancreas. Weight gain was evaluated to evaluate the edema grafts.

Histologic examination using a conventional staining method (hematoxylin eosin saffron) was performed to assess the degree of ischemic necrosis or congestion of exocrine and endocrine tissues and duodenum. Immunohistochemical analysis was performed to assess the islets viability, with insulin, glucagon, somatostatin, and ATP.

In kidney transplantation, some authors^32-34 considered high renal resistance during perfusion as an independent risk factor for the development of delayed graft function. PRI was monitored^{14,15,17,19-20,22-24,28,31} and decreased during the perfusion. This reflects initial and evolving status of pancreatic microcirculation and PRI should be used as a marker of pancreas transplant viability. Moreover, the pancreatic microvascularization preservation during the storage could help to reduce the risk of pancreas thrombosis. Biochemical analysis of the perfusion solution was performed to assess markers of pancreatic suffering, including lipase, amylase, lactate dehydrogenase, lactates, insulin, glucagon, glucose, and C-peptide.

We reported results of three human experimental hypothermic pancreas perfusion studies. These are feasibility studies of pancreas perfusion, in hypothermic and normothermic settings. Thus, none of these pancreases has been transplanted to assess the reperfusion lesions. Moreover, these pancreases were all rejected for vascularized organ or islet transplantation because of their poor quality. Then, the results presented in this paper did not reflect the possible consequences of the perfusion of an optimal pancreas. Hamaoui et al¹⁴ used porcine and human pancreases with long period of SCS, probably due to logistical problems. However, they reported no pancreatic damages after SCS and higher pancreatic damages after NP than HPP. Thus, in this study, NP would seem to induce more pancreatic lesions.

Concerning experimental hypothermic animals pancreas perfusion studies, the results seem to indicate that a low SP (<25 mmHg) should be used. Therefore, perfusion period longer than 24 hours provoked ischemic necrosis and cell damages of these pancreases.

We reported results of two human experimental normothermic perfusion studies. These two studies have been realized with human pancreases rejected for vascularized organ or islet transplantation. As in hypothermic settings, the results presented in this paper did not reflect the possible consequences of the normothermic perfusion of an optimal pancreas. However, there is a variability in the results presented by Barlow et al²³ and Nassar et al.²⁴ Indeed, Barlow et al presented important histologic pancreatic damages after six hours of NP, while Nassar et al reported healthy acini and islets after 6 and 12 hours. Thus, histologic lesions reported by Nassar et al could be due to longer SCS period (13 hours versus 4-6 hours).

Human and animal experimental studies suggest that perfusion of pancreatic allografts results in improved β-cell function. These results are interesting because the rate of pancreas discarded was still 32.8% in 2016 in France,³⁵ due to increasing age of donors, and graft thrombosis occurred in approximately 15%-20% of recipients. The CIT, one of the factors thought to increase graft pancreatitis, is still not able to be extended. Thus, the only option able to reduce CIT and, therefore, pancreatitis and vascular thrombosis, is NP of the pancreas. However, NP of the pancreas is still an experimental application and needs improvement for clinical application.

Potential benefits of machine perfusion include improvement in circulation of oxygen and nutrients, elimination of metabolic waste and toxins, maintenance of vasculature and endothelial protection, opportunity for viability assessment and therapeutic manipulation or gene therapy, and allow long-distance organ sharing.³⁶ Moreover, use of HP machine with pulsatile flow also enabled maintenance of vasculature and endothelial protection. Indeed, pulsatile flow allowed to maintain shear stress, known as an important factor in regulating endothelial cells’ inflammatory responses.³⁷ Moreover, this particular pulsatile action may activate endothelial protective genes such as Kruppel-like factor 2.³⁸ This gene, overexpressed by the endothelium during pulsatile perfusion, is thought to inhibit proinflammatory responses and protect endothelial cells.^39,40 The protective role of hypothermic machine perfusion for pancreas transplant may be evaluated by exploration of major ischemic reperfusion injury pathways such as hypoxia-inducible factor 1-alpha, inflammation, and endothelial nitric oxide synthase activation.

In our mind, due to decrease of standard donor (>18 or <50 years old, body mass index <30 kg/m², no medical history of alcoholism, diabetes, and cardiac arrest), the target of pancreas perfusion should be ECD.

A major limitation of most of these studies is the absence of in vivo evaluation of the pancreas. Further experimentation in animal models is needed to confirm the protective effect of pancreas perfusion, to select precise perfusion parameters useful for pancreas evaluation, and to explain the mechanisms of tissue protection. We could then consider the use of pancreas perfusion for human pancreas to broaden the organ pool and improve outcomes.

For the future, it should be important to evaluate ischemia-reperfusion injuries by model of transplantation and to invest in NP technology as has been done in liver and kidney transplantation. Indeed, with future development of portable NP machine devices, CIT can be minimized as the organ does not need to travel. Moreover, advantage of NP machine is the possibility to treat the organ (accommodation) prior to transplantation, instead of treating the recipient.

Conclusion

This review evaluated the feasibility of ex vivo machine perfusion of experimental pancreas for whole-organ transplantation. Pancreas perfusion is feasible and could be a helpful tool to evaluate pancreas transplant prior to transplantation. Pancreas perfusion (in hypothermic or normothermic settings) could reduce ischemic reperfusion injuries, and maybe could avoid pancreas thrombosis and reduce the morbidity of pancreas transplantation.