Liver Transplantation Following Donation After Cardiac Death: An Analysis Using Matched Pairs

Transcription

1 LIVER TRANSPLANTATION 15: , 2009 ORIGINAL ARTICLE Liver Transplantation Following Donation After Cardiac Death: An Analysis Using Matched Pairs James K. Pine, Amer Aldouri, Alistair L. Young, Mervyn H. Davies, 2 Magdy Attia, Giles J. Toogood, Stephen G. Pollard, J. P. A. Lodge, and K. R. Prasad Departments of 1 Hepatobiliary/Transplant Surgery and 2 Hepatology, St. James s University Hospital, Beckett Street, United Kingdom Grafts from donation after cardiac death (DCD) donors are used to increase the number of organs available for liver transplantation. There is concern that warm ischemia may impair graft function. We compared our DCD recipients with a case-matched group of donation after brain death (DBD) recipients. Between January 2002 and April 2008, 39 DCD grafts were transplanted. These were matched with 39 DBD recipients on the basis of identified variables that had a significant impact on mortality. These were used to individually match DCD and DBD patients with similar predictive mortality. We compared patient/graft survival, primary non-function (PNF), and rates of complications. Of all liver transplants, 6.1% were DCD grafts. PNF occurred twice in the DCD group. The incidence of nonanastomotic biliary strictures (NABS; 20.5% versus 0%, P 0.005) and hepatic artery stenosis (HAS; 12.8% versus 0%, P 0.027) in the DCD group was higher. One-year (79.5% versus 97.4%, P 0.029) and 3-year (63.6% versus 97.4%, P 0.001) graft survival was lower in the DCD group. Three-year patient survival was also lower (68.2% versus 100%, P ). Our study is the first to use case-matched patients and compare groups with similar predictive mortality. There was a higher incidence of NABS and HAS in the DCD group. NABS were likely a result of warm ischemia. HAS may have been due to ischemia or arterial injury during retrieval. The DCD group had significantly poorer outcomes, but DCD grafts remain a valuable resource. With careful donor/recipient selection, minimization of ischemia, and good postoperative care, acceptable results can be achieved. Liver Transpl 15: , AASLD. Received December 21, 2008; accepted July 9, The success of liver transplantation has resulted in an increasing need for organs as more patients are being listed for transplantation. This increase in the size of the waiting list has not been accompanied by a corresponding increase in the number of organ donors. The UK transplant activity report of reported that 8% of people died while on the transplant list and 11% were removed from it as a result of failing health. 1 Similarly, the United Network for Organ Sharing (UNOS) showed in its 2006 annual report that 1752 (10.4%) people died while on the waiting list and 630 (3.7%) were deemed too sick to transplant. The organ gap has continued to grow, and an increased supply of donor grafts is required. Various methods have been identified to reduce the organ deficit. These include living donor liver transplantation 2 and split liver transplantation. 3 Donation after cardiac death (DCD), or non heart-beating donation, has been proposed as an alternative source of liver grafts following the increasing success of this approach in renal transplantation. 4,5 DCD involves organs obtained from patients in whom cardiopulmonary function has ceased prior to organ retrieval. This usually involves donors from the intensive care unit from whom Abbreviations: ALD, alcoholic liver disease; Alk Phos, alkaline phosphatase; ALT, alanine aminotransferase; DBD, donation after brain death; DCD, donation after cardiac death; HAS, hepatic artery stenosis; HAT, hepatic artery thrombosis; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; ICH, intracerebral hemorrhage; IVC, inferior vena cava; MELD, Model for End-Stage Liver Disease; NABS, nonanastomotic biliary strictures; PBC, primary biliary cirrhosis; PNF, primary nonfunction; PSC, primary sclerosing cholangitis; SAH, subarachnoid hemorrhage; TBili, total bilirubin; UNOS, United Network for Organ Sharing. Address reprint requests to K. R. Prasad, Department of Hepatobiliary/Transplant Surgery, St. James s University Hospital, Beckett Street, Leeds, United Kingdom LS9 7TF. Telephone: (0113) ; FAX: (0113) ; raj.prasad@leedsth.nhs.uk DOI /lt Published online in Wiley InterScience ( American Association for the Study of Liver Diseases.

2 DONATION AFTER CARDIAC DEATH 1073 treatment is being withdrawn. However, there are concerns about the quality of such organs, with evidence that a prolonged warm ischemic time causes a raised incidence of primary nonfunction and increased biliary complications. 6,7 Also, patient selection criteria for DCD liver transplantation are still to be clearly defined. Several studies, including single-center and registry data, have compared DCD patient outcomes with donation after brain death (DBD) patient outcomes. However, none of them have individually matched DCD and DBD cases on the basis of statistically validated risk factors. This study is a retrospective review providing information regarding the complication and graft survival rates of the 39 DCD liver transplants performed in Leeds and comparing them to 39 DBD liver transplants matched for a variety of preoperative donor/recipient factors statistically proven to influence postoperative mortality. 8 All liver transplants were performed at the same center. PATIENTS AND METHODS The Department of Transplantation at St. James s University Hospital has had a controlled DCD organ donation program for both liver and kidney transplants since January Between January 2002 and April 2008, 639 liver transplants were performed, of which 40 used DCD donors. One of the DCD procedures was an auxiliary transplant in an acute liver failure patient, and it was therefore excluded. The 39 DCD cases were matched to corresponding DBD liver transplants performed at the same center during the same period. The matching was based on a statistical study of 34,664 adult liver transplants from the European Liver Transplant Registry. In this study, a multivariable logistic regression model was used to generate mortality scores for each individual and assess model discrimination and calibration with an independent data set. 8 This study identified preoperative and perioperative variables that had a significant impact on 3- and 12-month mortality (Table 1). These variables were used to precisely match transplant patients in the DCD group with patients from the HB group and enable a comparison between 2 groups with similar predictive mortality. The examined parameters were patient/graft survival at 90 days, 1 year, and 3 years, primary nonfunction (PNF), overall graft survival, and rates of complications (biliary complications and hepatic/portal vessel complications). DCD Donors: Inclusion/Exclusion Criteria Any patient from whom active treatment is to be withdrawn is briefly assessed for suitability for donation. Absolute contraindications include age less than 16 or greater than 70 and human immunodeficiency virus/ Creutzfeldt-Jakob disease infection. Any evidence of pre-existing liver disease and premorbid alcohol consumption is reviewed. The time interval between the withdrawal of treatment and asystole is unpredictable. TABLE 1. Factors Associated with Mortality to Which the DCD and DBD Groups Were Matched Year of transplant Cause of liver failure Acute liver failure Hepatocellular carcinoma Alcoholic cirrhosis HCV cirrhosis Primary biliary cirrhosis Other Age of donor (years) Donor-recipient ABO Identical groups Compatible Incompatible Recipient age (years) Within 5 years Split or reduced Yes graft No UNOS status Mean total ischemic 13 hours time 13 hours Unknown Size of center (transplants per year) NOTE: This table was adapted from Burroughs et al. 8 Abbreviations: HCV, hepatitis C virus; UNOS, United Network for Organ Sharing. If asystole has not occurred within 90 minutes of treatment withdrawal or significant, prolonged hypotension or hypoxia occurs, the retrieval is abandoned. Finally, the macroscopic appearance of the organ is assessed by the retrieving surgeon. The color and texture are examined to determine any evidence of fatty change, and the perfusion quality is assessed. Any visible lesions are assessed, and biopsy samples are taken as appropriate. If the graft appearance is deemed moderate or severely fatty or if there is evidence of poor perfusion, the graft will not be used. Organ Retrieval Technique Once treatment is withdrawn, the donor is monitored via electrocardiogram and intra-arterial blood pressure to identify the onset of cardiorespiratory arrest. This is confirmed by a member of the critical care team by identifying the absence of cardiac output and respiration, a lack of response to supraorbital pressure, and the absence of papillary and corneal reflexes. The retrieval process will not begin until at least 10 minutes after the loss of cardiac output and 5 minutes after the certification of death. The donor is quickly transferred to the operating room. The retrieval technique used is

3 1074 PINE ET AL. super-rapid, as described by Casavilla et al. 9 The aorta is swiftly accessed via a full-length midline laparotomy, and a large-bore cannula is inserted. Perfusion is commenced immediately via the aortic cannula. The inferior vena cava is vented, and congestion on the liver is relieved. At the same time, topical cooling is achieved with chilled/frozen 0.9% sodium chloride. The inferior mesenteric vein is cannulated, and portal perfusion is achieved. Two liters of Marshall s solution is perfused through the aortic cannula. Each of the bags contains 5000 units of heparin. Three liters of University of Wisconsin solution is perfused through the portal cannula, each with 5000 units of heparin. The procedure then proceeds in a manner similar to the cold phase of standard retrieval. Recipient Selection The recipients were selected on the basis of clinical urgency. They were counseled regarding the problems associated with receiving a DCD liver transplant. Patients with stable cholestatic liver disease were not given DCD grafts. Patients with hepatocellular carcinoma were prioritized for receiving DCD grafts. Regrafts were excluded from the DCD program. DBD Group The DBD group against which we compare our DCD group was matched on the basis of data published by the European Liver Transplant Association. This study involved 34,664 adult liver transplants from the European Liver Transplant Registry. The authors used the data to identify significant risk factors and compile 3- and 12-month mortality models for cadaveric liver transplant patients. 8 The variables identified by this study are shown in Table 1. We used this information to match the DCD recipients to equivalent DBD recipients with respect to the year of transplant, recipient etiology, age of the donor, donor-recipient ABO groups, recipient age at transplantation, split or reduced grafts, UNOS status, total ischemic time, and size of the transplantation center. This enabled us to compare 2 groups with similar predicted mortality and assess differences in outcome. Statistical Analysis Statistical analyses included the Student t test for continuous variables, Fisher s exact test and the chisquare test for categorical variables, Kaplan-Meier plots to illustrate survival, and the log-rank test to compare patient and graft survival. All analyses were performed with SPSS version 14 (SPPS, Inc., Chicago, IL). RESULTS TABLE 2. Comparison of the Two Groups via Their Matched Variables DCD DBD Year of transplant: Cause of liver failure Acute liver failure 0 0 Hepatocellular carcinoma 4 3 Alcoholic cirrhosis 3 3 HCV cirrhosis Primary biliary cirrhosis 1 1 Other Age of donor (years) Donor-recipient ABO group Identical Compatible 6 4 Incompatible 0 0 Recipient age (years) Split or reduced graft Yes 0 0 No UNOS status Mean total ischemic time (hours) Unknown 0 0 Size of center (transplants per year): Abbreviations: DBD, donation after brain death; DCD, donation after cardiac death; HCV, hepatitis C virus; UNOS, United Network for Organ Sharing. From January 2002 to April 2008, 40 DCD liver transplants were performed at St. James s University Hospital (Leeds, England). Thirty-six were retrieved by the Leeds transplant team, and 4 were imported from other centers. There were 40 livers from DCD donors transplanted in all; 39 of these were whole liver orthotopic transplants. One case was an auxiliary transplant for a superurgent patient, and it has not been included in this study. Thirty-nine liver transplants from DBD donors were matched with the DCD cases with variables described by the European Liver Transplant Registry (see Table 2). Twenty-eight of the matched DCD transplanted livers were retrieved by Leeds, with the remaining 11 imported. Donor Characteristics Over a 6-year period, 57 DCD retrievals were undertaken. Eleven retrieved organs were deemed unsuitable (poor graft appearance or prolonged warm ischemic time), 36 were transplanted at Leeds, and the remaining 10 were exported to other transplant centers. The

4 DONATION AFTER CARDIAC DEATH 1075 TABLE 3. Donor Characteristics DCD Donors DBD Donors P Value Mean age (years) * Sex (male/female) 22/17 25/ Donor cause of death ICH Trauma Aneurysm Subdural hematoma Extradural hematoma SAH Intracranial thrombosis Hypoxic brain injury Brain tumor Drug overdose Hepatic encephalopathy post-transplant Abbreviations: DBD, donation after brain death; DCD, donation after cardiac death; ICH, intracerebral hemorrhage; SAH, subarachnoid hemorrhage. *t test. Chi-square test. Fisher s exact test. mean donor age in the DCD group was years (range, years), and 22 were male. In the DBD group, the mean donor age was years (range, years), and 25 were male. The most common donor causes of death in the DCD donor group were intracerebral hemorrhage (16 donors) and traumatic head injury (9 donors; see Table 3). In the DBD group, the most common cause of death in donors was also intracerebral hemorrhage (22 donors), which was followed closely by subarachnoid hemorrhage (10 donors; see Table 3). A statistical comparison of donor characteristics showed no significant differences between the DCD and DBD groups in donor age or donor sex. The only statistically significant difference between the 2 groups was that more donors in the DBD group died as a result of subarachnoid hemorrhage (10 versus 2, P 0.025). Recipient Characteristics In the DCD group, 39 orthotopic whole liver transplants were performed. In the DBD group, all 39 transplants were whole liver transplants. The indications for transplant in the DCD and DBD groups are shown in Table 4. The mean age of the DCD recipients was years (range, years), and 25 were male. The mean age of the DBD recipients was years (range, years), and 32 were male. The mean Model for End- Stage Liver Disease (MELD) score was (range, 6-26) in the DCD group and (range, 6-32) in the DBD group. A statistical comparison of the 2 groups for age, sex, etiology, and MELD score revealed no significant differences. Graft Ischemia The mean cold ischemia times and the mean operating reperfusion times are shown in Table 5. Both ischemic periods were statistically significantly shorter in the DCD group (P and 0.007, respectively). Retrieving/Transplanting Surgeon The majority of organ retrievals were performed by experienced senior trainees (specialist registrars and clinical fellows). In the case of the early DCD retrievals, consultants performed the initial retrieval procedures as new surgical techniques were being introduced; consequently, the number of organ retrievals performed by consultants in the DCD group was significantly greater (9 versus 2, P 0.047). All transplants in both groups of patients were performed by 1 of 5 consultant surgeons employed by the Leeds Teaching Hospitals NHS Trust. At both retrieval and transplantation, the operating surgeon was asked to make a macroscopic assessment of the graft. There were no significant differences in perceived graft quality between the 2 groups. Posttransplant Liver Function Tests Laboratory values for alanine aminotransferase (ALT) on days 1, 3, 7, 14, and 28 are shown in Fig. 1. This demonstrates that ALT levels in the DCD group were consistently higher throughout the immediate postoperative course than those in the DBD group, although the differences were not statistically significant. The alkaline phosphatase levels were also generally higher in the DCD group in the immediate postoperative period, being significantly higher on day 7 (648.7 versus IU/L, P 0.049). However, on day 30, the DBD groups levels were higher. This is illustrated in Fig. 2. Bilirubin levels for days 1, 3, 7, 14, and 28 are shown in Fig. 3. Interestingly, bilirubin levels post-transplant were consistently lower in the DCD group. This difference achieved statistical significance on day 1 (66.49 versus mol/l, P 0.009).

5 1076 PINE ET AL. TABLE 4. Recipient Characteristics DCD Recipient DBD Recipient P Value Mean age (years) * Sex (male/female) 25/14 32/ Mean MELD * Etiology HCC ALD HCV PBC PSC HBV Autoimmune Cryptogenic Hemochromatosis Alpha-1 antitrypsin deficiency Neuroendocrine tumor Abbreviations: ALD, alcoholic liver disease; DBD, donation after brain death; DCD, donation after cardiac death; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; MELD, Model for End-Stage Liver Disease; PBC, primary biliary cirrhosis; PSC, primary sclerosing cholangitis. *t test. Fisher s exact test. Chi-square test. TABLE 5. Comparison of Periods of Graft Ischemia in the DCD and DBD Groups DCD DBD P Value Mean cold ischemic time in minutes (range) ( ) ( ) * Mean operative reperfusion time in minutes (range) (34 82) (23 65) 0.007* Abbreviations: DBD, donation after brain death; DCD, donation after cardiac death. *t test. Figure 1. Mean postoperative ALT for DBD and DCD recipients. Abbreviations: ALT, alanine aminotransferase; DBD, donation after brain death; DCD, donation after cardiac death. PNF Two patients developed PNF in the DCD group. Both died, one after an emergency regraft and the other without retransplantation. None of the grafts in the DBD Figure 2. Mean postoperative Alk Phos for DCD and DCD recipients. Abbreviations: Alk Phos, alkaline phosphatase; DBD, donation after brain death; DCD, donation after cardiac death.

6 DONATION AFTER CARDIAC DEATH 1077 Thirteen patients in the DCD group developed biliary complications. Eight were nonanastomotic biliary strictures (NABS), 4 were anastomotic biliary strictures, and 1 was a bile leak. In the DBD group, there were 4 biliary complications, all of them anastomotic strictures. NABS were significantly more prevalent in the DCD group (8 versus 0, P 0.005). There were 6 vascular complications in the DCD group. These consisted of 5 cases of hepatic artery stenosis (HAS) and 1 case of hepatic artery thrombosis (HAT). The incidence of HAS was significantly higher in the DCD group (5 versus 0, P 0.027). In the DBD group, there were 4 vascular complications: 2 cases of HAT and 2 inferior vena cava strictures. With respect to HAT, the first was recognized within 24 hours of transplant, and the patient underwent successful revascularization. The second involved an accessory left hepatic artery; although revascularization was attempted, it was unsuccessful, and the ischemic left lobe was resected. Other operative complications seen in both groups were postoperative bleeding (3 in the DCD group and 3 in the DBD group) and postoperative collection requiring drainage (2 in the DCD group and 1 in the DBD group). In the postoperative period, 8 of the DCD group suffered acute rejection requiring high-dose steroids versus 9 in the DBD group. This is shown in Table 6. The group with NABS was compared with the remaining DCD patients. The differences in various ischemic times are shown in Table 7. Both the warm ischemic and operative reperfusion times were longer in the NABS group, but this did not reach statistical significance. There were no differences in donor and recipient variables between the 2 groups. However, the mean donor age was higher in the NABS group (50.43 versus years), but this did not reach statistical significance. Recipient/donor sex and recipient MELD scores were also compared, and no significant differences were found. This is shown in Table 7. Similar comparisons were made with HAS patients and non-has patients in the DCD group, with no significant differences demonstrated. Three of the 7 patients with NABS also had HAS. Figure 3. Mean postoperative TBili for DBD and DCD recipients. Abbreviations: DBD, donation after brain death; DCD, donation after cardiac death; TBili, total bilirubin. group had PNF. The incidence of PNF between the 2 groups was not statistically significant. Complications Ninety-Day Graft Survival In the DCD group, 4 livers were lost in the perioperative period. Two of these were as a result of PNF (discussed previously). The other causes of graft loss at this stage were fatal intraoperative cardiac arrest at reperfusion in one case and death as a result of postoperative complications with a functioning graft in another case. In the DBD group, there were no livers lost in the immediate postoperative period. There was one case of graft loss inside 90 days due to chronic rejection. This patient was successfully retransplanted. One-Year Graft Survival Another 4 DCD grafts were lost inside 1 year. One loss was a result of HAS causing extensive necrosis (the patient was successfully retransplanted), 1 loss was secondary to ischemic cholangiopathy (the patient was not suitable for retransplantation), 1 loss was due to chronic rejection (patient retransplanted), and 1 patient died with the graft still functioning. There was no further graft loss in the DBD group inside 1 year. The 1-year graft survival was significantly lower in the DCD group (79.5% versus 97.4%, P 0.029). Causes of Graft Failure A breakdown of the etiology of graft failure is shown in Table 8. The incidence of loss of graft function was higher in the DCD group (Fig. 4). Overall graft survival is illustrated in Fig. 4. The 3-year graft survival for the DCD group was 63.6% versus 97.4% for the DBD group. This was statistically significant (P 0.001). Patient Survival The mean follow-up for the DCD group was 2.5 years. For the DBD group, it was 6.6 years. The difference in the mean follow-up was statistically significant (P ). There were 9 deaths to date in the DCD group. In the immediate postoperative period, 2 patients were lost as a result of intraoperative cardiac arrests; one immediately died, and the other died inside 1 week. Two patients died of multiorgan failure as a result of postoperative complications. Three patients died within a year of their operation: one following a second transplant, one of metastatic hepatoma, and the third of an unknown cause. Finally, 2 patients died more than 1 year post-transplant: one from a disseminated malignancy and the other from unknown causes. There were 4 fatalities in the DBD group, all more

7 1078 PINE ET AL. TABLE 6. Complications DCD DBD P Value Postoperative complication Anastomotic biliary stricture 4 (10.2%) 4 (10.2%) 1* Nonanastomotic biliary stricture 7 (17.9%) * Bile leak 1 (2.6%) 0 1* Hepatic artery stenosis 5 (12.8%) * Hepatic artery thrombosis 1 (2.6%) 2 (5.1%) 1* IVC stricture 0 2 (5.1%) 0.494* Postoperative bleeding 3 (7.7%) 3 (7.7%) 1* Postoperative collection 2 (5.1%) 1 (2.6%) 1* Recurrent cholangitis 1 (2.6%) 0 1* Perioperative hepatic rupture 1 (2.6%) 0 1* Liver abscess 1 (2.6%) 0 1* Primary nonfunction of 2 (5.1%) * graft Rejection Acute 8 (20.5%) 9 (23.1%) Chronic 2 (5.1%) 4 (10.2%) 0.675* Abbreviations: DBD, donation after brain death; DCD, donation after cardiac death; IVC, inferior vena cava. * Fisher s exact test. Chi-square test. TABLE 7. Comparison of Recipient and Donor Variables and Graft Ischemic Times in Patients with Evidence of NABS and in Patients without Evidence of NABS in the DCD Group NABS No NABS P Value Recipient age at transplant (years) * Recipient sex (male/female) 6/1 19/ Recipient MELD * Donor age (years) * Donor sex (male/female) 2/5 20/ Mean extubation to asystole time (minutes) * Mean warm ischemic time (minutes) * Mean operative reperfusion time (minutes) * Mean total warm ischemic time (minutes) * Mean cold ischemic time (minutes) * Steatotic according to the retrieving surgeon (yes/no) 4/3 12/ Suboptimal according to the transplanting surgeon (yes/no) 4/3 11/ Grade of retrieving surgeon (senior trainee/consultant) 4/3 26/ Abbreviations: DCD, donation after cardiac death; MELD, Model for End-Stage Liver Disease; NABS, nonanastomotic biliary strictures. * t test. Fisher s exact test. than 2 years post-transplant. One died from liver failure as a result of recurrent HCV, 1 died from esophageal squamous cell carcinoma, 1 died from metastatic rectal cancer, and 1 died from metastatic hepatoma. There was a higher incidence of death in the DCD group, as illustrated by the Kaplan-Meier survival curve in Fig. 5. The 3-year patient survival was 68.2% and 100% in the DCD and DBD groups, respectively. This difference was statistically significant (P ). DISCUSSION In recent years, the requirement for hepatic allografts has steadily increased, whereas the organ pool from which the grafts are drawn has remained relatively static. 10 Steps have been taken to increase the number of organs available for transplantation. Strategies include expanding the existing population of DBD cadaveric donors, using split liver transplantation, and introducing live-donor liver donation programs. The use of marginal grafts has also been explored. Included in this group are organs retrieved from patients following cardiac death; these are also known as DCD donors. Using DCD donor grafts is not a new concept. Initial experience with liver transplantation in the 1960s used DCD donors. 11 After the introduction of brain death legislation in 1968, 12 it was apparent that organs procured from brain dead donors produced improved graft

8 DONATION AFTER CARDIAC DEATH 1079 TABLE 8. Causes of Graft Loss DCD DBD P Value Cause of graft loss Primary nonfunction 2 (5.1%) * Chronic rejection 1 (2.6%) 1 (2.6%) 1* Recurrent disease 0 1 (2.6%) 1* Biliary complications 1 (2.6%) 0 1* Vascular complications 1 (2.6%) 0 1* Recipient death with graft still functioning at 4 (10.2%) 3 (7.7%) 1* the time of death Intraoperative death 1 (2.6%) 0 1* Total graft loss 10 (25.6%) 5 (12.8%) Abbreviations: DBD, donation after brain death; DCD, donation after cardiac death. * Fisher s exact test. Chi-square test. Figure 4. Kaplan-Meier plot of 5-year graft survival. Abbreviations: DBD, donation after brain death; DCD, donation after cardiac death. Figure 5. Kaplan-Meier plot of 5-year patient survival. Abbreviations: DBD, donation after brain death; DCD, donation after cardiac death. and patient survival, and interest in DCD donation waned. However, the disparity between the demand for liver grafts and the supply has prompted this donor pool to be re-explored. Following the renewed interest in DCD grafts, the preliminary results appeared promising. Casavilla et al. 9 demonstrated in 1995 that controlled DCD grafts could be used successfully but that PNF in uncontrolled DCD was unacceptably high (50%). D Alessandro et al. 13 in 2000 compared their experience with 19 DCD grafts against their experience with 364 DBD grafts. They noted no differences in biliary or vascular complications, but PNF was found to be higher (10.5% versus 1.3%, P 0.04), and allograft survival was lower (53.8% versus 80.9%, P 0.007). Foley et al. 14 in 2005 published updated results from the same center in Wisconsin. This time, they compared 36 DCD grafts with 553 DBD grafts. DCD grafts were shown to have significantly higher rates of biliary stricture, HAS, and abscess/biloma. They also showed significantly reduced patient and graft survival. Most recently, Chan et al. 15 published their results for 52 DCD donor grafts. They compared them with DBD grafts used over the same period and did not find any significant differences in survival or hepatic artery/biliary anastomotic complications. They did show that the incidence of NABS was greater in the DCD group. Large studies based on registry data have also been published. Abt et al. 6 in 2004 analyzed the UNOS database, identifying 144 DCD grafts and 26,856 DBD grafts. A greater incidence of PNF and retransplantation was seen in the DCD group versus the DBD group. They also identified increased cold ischemic time and recipient life support status as predictors of graft failure in DCD grafts, although only cold ischemic time remained

9 1080 PINE ET AL. significant on multivariate analysis. Merion et al. 7 in 2006 published data from the Scientific Registry of Transplant Recipients database: 472 DCD grafts and 23,598 DBD grafts were compared. Three-year graft survival was significantly lower in the DCD group (60.5% versus 75.5%). We present a single-center cohort of 39 DCD donor grafts. This is the first such study to compare 2 casematched groups. This enables a more accurate comparison of patient outcomes as DCD donor recipients are compared with DBD graft recipients matched with respect to preoperative predicted mortality. Donor characteristics were similar in the 2 groups, with the only notable difference being that subarachnoid hemorrhage as the donor cause of death was significantly higher in the DBD group. Recipient characteristics also displayed no significant differences. Despite the recipients not being specifically matched for MELD score, the mean values for both groups were not significantly different (14.2 for DCD versus 15.2 for DBD). The difference in cold ischemia time between the DBD and DCD groups (594 versus 352 minutes, P ) reflects the impact that ischemic times are recognized to have on graft function following DCD donation. Complication rates were higher in the grafts obtained from DCD donors. A more detailed analysis revealed that differences in NABS (17.9% in DCD versus 0% in DBD, P 0.012) and HAS (12.8% in DCD versus 0% in DBD, P 0.027) were statistically significant. Further analysis of donor/recipient characteristics and ischemia times in the recipients with HAS revealed no significant differences. The average donor age was greater (56.3 versus 49.3 years) and the warm ischemia time (13.4 versus 12.8 minutes) was longer in the group with NABS, but neither difference achieved statistical significance. A significant proportion of recipients with NABS also had HAS (42.9%, P 0.032). This reflects the association that HAS has with NABS as a result of further ischemic injury to the allograft. The development of NABS, also termed ischemic cholangiopathy, is a well-recognized complication following ischemia-reperfusion injury of the liver. 16 The mechanism by which it occurs involves damage to biliary epithelial cells during organ retrieval and preservation. Cellular injury in the initial warm ischemic phase prior to cold perfusion is thought to be significant. The rapid depletion of adenosine triphosphate stores from cells during the warm phase causes intracellular acidosis and a loss of membrane integrity. This causes significant damage to sensitive biliary epithelial cells. 17 The presence of hydrophobic bile salts is thought to be a further mechanism of injury. 18 Flushing of the biliary tree with perfusate has been advocated in some quarters as a way of minimizing this damage. 19 The cold preservation phase also plays a role. Although metabolism is slowed, considerable activity still occurs at 4 C. Bile duct epithelial cells have been shown to be more sensitive to reperfusion injury than hepatocytes. 20 Studies using DBD donors have shown that a prolonged cold ischemia time is associated with the formation of NABS. 21 It is thought that damage to the microvascular endothelium during this phase leads to further cell damage and may contribute to microvascular thrombosis that in turn prevents effective revascularization. 18 The overall effect is shedding of bile duct epithelium leading to ulceration. The subsequent fibrotic healing process causes the widespread stricture formation seen in the biliary tree following ischemic injury. The incidence of HAS was significantly elevated in the DCD group. This has previously been reported after DCD donation. 14 All HAS occurred within 12 weeks of transplant (range, 5-12 weeks; median, 6 weeks). Two of the 5 stenoses occurred at the anastomosis, and the remaining 3 were found in the donor hepatic artery, distal to the anastomosis. Angioplasty was performed for all cases of all HAS with good angiographic results. Early HAS is potentially attributable to technical factors. 22 The super-rapid technique used for retrieval may lead to unrecognized traumatic injury to the arterial intima resulting in nonanastomotic stenosis. The need to reduce the cold ischemia time when DCD organs are used often leads to surgeons operating at suboptimal times, potentially predisposing them to technical error. However, the fact that over half the cases of HAS were distal to the anastomosis suggests that the donor extrahepatic arteries are susceptible to warm ischemic injury. The mechanism for this remains unclear. The incidence of PNF in the DCD group was 5.1% versus 0% following DBD donation. Although this difference did not reach statistical significance, it is consistent with the trend reported in other studies. PNF is thought to be a significant risk following transplantation from a DCD donor. For grafts from uncontrolled (Maastricht type 2) DCD donors, PNF rates as high as 50% have been reported. 9 In controlled DCD donation, the rates are 0% to 10.5%. 9,13,14,23-26 Once again, ischemia-reperfusion injury plays a key role in reducing graft viability, with the initial period of warm ischemia playing a crucial role. It has been demonstrated that this warm ischemic period increases graft susceptibility to damage during cold ischemia; therefore, both periods of ischemia must be kept to a minimum. Graft survival and patient survival were significantly reduced in the DCD group. Graft survival at 1 year in the 2 groups was 79.5% (DCD) and 97.4% (DBD), and at 3 years, it was 63.6% (DCD) and 97.4% (DBD). Patient survival displayed a similar trend, with 1-year survival being 82.1% (DCD) and 100% (DBD) and 3-year survival being 68.2% (DCD) and 100% (DBD). These results are consistent with published data. Graft survival at 1 year has been reported to be 50% to 86.5%, and patient survival has been reported to be 50% to 89.6%. 6,7,9,13,14,25,27 The DBD survival figures are higher than one would expect for a random group of transplant patients. Selection for DCD graft recipients is determined by the transplant center, and it has been the policy to prioritize HCC patients (who often have low MELD scores) and to avoid giving DCD grafts to patients with high MELD scores as it is felt these fare less well.

10 DONATION AFTER CARDIAC DEATH 1081 Therefore, you would expect a case-matched DBD group to have higher-than-average survival. If DCD donation is to provide a viable and safe alternative supply of hepatic allografts, it is clear that certain precautions need to be taken. Potential donors need to be carefully selected. Donor age seems to be particularly important. Lee et al. 28 in 2006 found donor age to be a statistically significant risk factor for graft failure on multivariate analysis and suggested donors greater than 45 years old increase the risk of graft failure. Other reports have suggested that a donor weight of greater than 100 kg predisposes to complications. 15 Organ retrieval should be as controlled as possible. Prolonged extubation to cardiac arrest times or any periods of extended hypotension should result in the retrieval being stepped down. Both warm and cold ischemia times need to be minimized. This is achieved by good surgical technique and a well-coordinated and rapid retrieval/transplant team. Although the transfer of grafts between centers risks prolonging the cold ischemia time, this is still possible, provided that a fast and efficient referral system is in place. Potential recipients require careful selection. It has been reported that highrisk recipients (intensive care unit patients, retransplantation patients, and multiorgan failure patients) have inferior graft survival rates in comparison with recipients of DBD grafts. 28 In this institution, we omit patients with stable cholestatic disease and prioritize patients with HCC. Patients also require counseling regarding the potential risks of receiving a DCD graft versus the risk of dying of their underlying disease while still on the transplant waiting list. Experimental strategies have been introduced to attempt to limit the damage caused by ischemia and its deleterious effect on graft function. Recent experiences using crystalloid-based preservation solutions combined with a thrombolytic agent have demonstrated a degree of success. 29 It is thought to reduce ischemiareperfusion injury by improving microcirculation and reducing microcirculatory thrombosis. Cold machine perfusion is a technique that is being employed with some success in kidney transplantation. 30 This technique has been shown to improve survival following liver transplantation in an animal model and has the potential to reduce the incidence of PNF in human transplantation. 31 There are currently no reliable markers following cold preservation to indicate postoperative graft function. Some experimental work has been performed that looks at hepatic xanthine and microdialysate hypoxanthine levels, but this has not been assessed in the clinical setting. 32,33 Further work needs to be done to assess markers of warm ischemia and graft viability. This single-center study is the first to compare groups of patients who received DCD donor grafts and patients who received DBD donor grafts that were matched on the basis of predictive mortality. Our results show that grafts from DCD donors have an increased rate of complications, particularly NABS and HAS. Graft survival and patient survival are also significantly reduced in comparison with survival with DBD donor grafts. However, there is a rising demand for transplantable organs; therefore, strategies to increase available grafts need to be considered. With careful donor and recipient selection and acceptable periods of warm and cold ischemia, DCD donors provide a method of expanding the donor pool. Particular attention needs to be paid to signs of biliary or arterial complications in recipients of DCD grafts. As our understanding of the physiological basis of ischemia-reperfusion injury in DCD donor grafts grows, new techniques are being developed that may help reduce the incidence of complications and poor graft function. Over time, graft and patient survival rates with DCD donations have improved. Therefore, we conclude that with careful donor/ recipient selection, minimization of periods of ischemia, and good postoperative care, DCD donor grafts can continue to provide a valuable contribution to the donor pool. REFERENCES 1. UK transplant activity report 2006/07: liver activity. Available at: Accessed July Malaga M, Testa G, Frilling A, Nadalin S, Valentin-Gamazo C, Paul A, et al. Right living donor liver transplantation: an option for adult patients. Ann Surg 2003;238: Renz J, Emond J, Yersiz H, Ascher N, Busuttil R. Split liver transplantation in the United States: outcomes of a national survey. Ann Surg 2004;239: Cho Y, Terasaki P, Cecka J, Gjertson D. Transplantation of kidneys from donors whose hearts have stopped beating. N Engl J Med 1998;338: Droupy S, Blanchet P, Eschwege P, Hammoudi Y, Joseph L, Kriaa F, et al. Long-term results of renal transplantation using kidneys harvested from non-heartbeating donors: a 15-year experience. J Urol 2003;169: Abt P, Desai N, Crawford M, Forman L, Markmann J, Olthoff K, Markmann J. Survival following liver transplantation from non-heartbeating donors. Ann Surg 2004;239: Merion R, Pelletier S, Goodrich N, Englesbe M, Delmonico F. Donation after cardiac death as a strategy to increase deceased donor liver availability. Ann Surg 2006;244: Burroughs A, Sabin C, Rolles K, Delvart V, Karam V, Buckels J, et al. 3-month and 12-month mortality after first liver transplant in adults in Europe: predictive models for outcome. Lancet 2006;367: Casavilla A, Ramirez C, Shapiro R, Nghiem D, Miracle K, Bronsther O, et al. Experience with liver and kidney allografts from non-heart-beating donors. Transplantation 1995;59: Howard R, Schold J, Cornell D. A 10-year analysis of organ donation after cardiac death in the United States. Transplantation 2005;80: D Alessandro A, Hoffman R, Belzer F. Non-heart-beating donors: one response to the organ shortage. Transpl Rev 1995;9: A definition of irreversible coma: report of the Ad Hoc Committee of the Harvard Medical School to examine the definition of brain death. JAMA 1968;205: D Alessandro A, Hoffman R, Knechtle S, Odorico J, Becker Y, Musat A, et al. Liver transplantation from controlled non-heart-beating donors. Surgery 2000;128: Foley D, Fernandez L, Leverson G, Chin L, Krieger N,

11 1082 PINE ET AL. Cooper J, et al. Donation after cardiac death: the University of Wisconsin experience with liver transplantation. Ann Surg 2005;242: Chan E, Olsen S, Kisthard J, Perkins J, Bakthavasalam R, Halldorson J, et al. Ischemic cholangiopathy following liver transplantation from donation after cardiac death donors. Liver Transpl 2008;14: Cameron A, Busuttil R. Ischaemic cholangiopathy after liver transplantation. Hepatobiliary Pancreat Dis Int 2005;4: Alpini G, McGill J, Larusso N. The pathobiology of biliary epithelia. Hepatology 2002;35: Kukan M, Haddad P. Role of hepatocytes and bile duct cells in preservation-reperfusion injury of liver grafts. Liver Transpl 2001;7: Maheshwari A, Maley W, Li Z, Thuluvath P. Biliary complications and outcomes of liver transplantation from donors after cardiac death. Liver Transpl 2007;13: McKeown C, Edwards V, Phillips M, Harvey P, Petrunka C, Strasberg S. Sinusoidal lining cell damage: the critical injury in cold preservation of liver allografts in the rat. Transplantation 1988;46: Rull R, Garcia Valdecasas J, Grande L, Fuster L, Lacy A, Gonzalez F. Intrahepatic biliary lesions alter orthotopic liver transplantation. Transpl Int 2001;14: Da Silva R, Raphe R, Felicio H, Rocha M, Duca W, Arroyo P, et al. Prevalence, treatment, and outcomes of the hepatic artery stenosis after liver transplantation. Transpl Proc 2008;40: Reich D, Munoz S, Rothstein K. Controlled non-heartbeating donor liver transplantation: a successful single centre experience, with topic update. Transplantation 2000;70: Manzarbeitia C, Ortiz J, Jeon H. Long-term outcome of controlled, non-heart-beating donor liver transplantation. Transplantation 2004;78: Muiesan P, Girlanda R, Jassem W, Melendez H, O Grady J, Bowles M, et al. Single-center experience with liver transplantation from controlled non-heartbeating donors: a viable source of grafts. Ann Surg 2005;242: Monbaliu D, Van Gelder F, Troisi R, de Hemptinne B, Lerut J, Reding R, et al. Liver transplantation using nonheart-beating donors: Belgian experience. Transpl Proc 2007;39: Fujita S, Mizuno S, Fujikawa T, Reed A, Kim R, Howard R, Hemming A. Liver transplantation from donation after cardiac death: a single center experience. Transplantation 2007;84: Lee K, Simpkins C, Montgomery R, Locke J, Segev D, Maley W. Factors affecting survival after liver transplantation from donation after cardiac death donors. Transplantation 2006;82: Fung J, Eghtesad B, Patel-Tom K. Using livers from donation after cardiac death donors a proposal to protect the true Achilles heel. Liver Transpl 2007;13: St Peter S, Imber C, Friend P. Liver and kidney preservation by perfusion. Lancet 2002;359: Lee C, Jain S, Duncan H, Zhang J, Jones J, Southard J, Clemens M. Survival transplantation of preserved nonheart-beating donor rat livers: preservation by hypothermic machine perfusion. Transplantation 2003;76: Net M, Valero R, Almenara R, Rull R, Gonzalez F, Taura P. Hepatic xanthine levels as viability predictor of livers procured from non-heart-beating-donor pigs. Transplantation 2001;71: Nagayama M, Katsuramaki T, Kimura H, Isobe M, Meguro M, Matsuno T. Prediction of graft viability from non-heartbeating donor pigs using hepatic microdialysate hypoxanthine levels. J Surg Res 2002;107: