Orthotopic liver transplantation (OLT) is a lifesaving modality for treating end-stage liver disease. The current practice of liver transplantation is limited by the significant disparity between organ availability and the number of patients waiting for OLT. An important source of livers that has been used to expand the donor pool is donation after cardiac death (DCD) donors; these are expanded criteria donors for whom death is declared on the basis of cardiopulmonary criteria rather than cessation of brain function. DCD livers are considered inferior grafts because of the higher risk of primary nonfunction (PNF), hepatic artery thrombosis (HAT), and ischemic cholangiopathy (IC). We previously reported our experience with 108 DCD liver recipients.1 Our liver transplant program now has performed over 200 OLT procedures with liver grafts from DCD donors between 1998 and 2010.
Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; CI, confidence interval; CIT, cold ischemia time; CMV, cytomegalovirus; DBD, donation after brain death; DCD, donation after cardiac death; DWIT, DCD warm ischemia time; HAT, hepatic artery thrombosis; HCV, hepatitis C virus; HCV+DBD+, hepatitis C virus–positive and donation after brain death–positive; HCV−DCD+, hepatitis C virus–negative and donation after cardiac death–positive; HCV+DCD+, hepatitis C virus–positive and donation after cardiac death–positive; HR, hazard ratio; IC, ischemic cholangiopathy; MELD, Model for End-Stage Liver Disease; OLT, orthotopic liver transplantation; PNF, primary nonfunction; pRBC, packed red blood cell; WIT, warm ischemia time.
Hepatitis C virus (HCV) infection is currently the most common indication for OLT in the United States.2 In our program, patients with HCV-related end-stage liver disease constitute 35% of all OLT cases. Although studies have addressed the use of expanded criteria donor organs in HCV+ patients, to date only 1 study has specifically addressed the use of liver grafts from DCD donors for HCV+ patients.3-5
This retrospective analysis, which represents the largest single-center experience to date, was undertaken to study the outcomes of DCD liver grafts used for HCV+ recipients in our program. Our goal was to determine whether the use of DCD liver grafts caused inferior graft and patient survival in HCV+ recipients. We matched the HCV+ patients who received DCD liver grafts with HCV+ patients who received donation after brain death (DBD) liver grafts. We also compared the study group to non-HCV patients who received DCD liver grafts.
PATIENTS AND METHODS
This was a retrospective, match-controlled study of HCV+ patients who received DCD liver grafts at Mayo Clinic Florida between January 2003 and September 2009. Data were gathered from a prospectively collected database. Approval for the study was obtained from the Mayo Clinic institutional review board. The study was performed by the retrospective chart review of all OLT cases with DCD grafts during the stated time period. The study group consisted of HCV+ patients who received DCD liver grafts [hepatitis C virus–positive and donation after cardiac death–positive (HCV+DCD+)]. These patients were matched to HCV+ patients who received DBD liver grafts [hepatitis C virus–positive and donation after brain death–positive (HCV+DBD+)]. The matching criteria were as follows: ±5 years for donor age, ±5 years for recipient age, ±2 hours for the cold ischemia time (CIT), and ±5 points for the Model for End-Stage Liver Disease (MELD) score (or both scores >30). The study group was also compared to a cohort of non-HCV patients who received DCD liver grafts [hepatitis C virus–negative and donation after cardiac death–positive (HCV−DCD+)] without the use of matching criteria.
All DCD donors were classified as Maastricht type 3 (controlled and awaiting cardiac death).6, 7 Detailed information regarding the DCD donors was obtained from the Mayo Clinic Florida procurement database and the organ procurement organization records. The gathered donor information included the following: geographic location, age, sex, ethnicity, cause of death, warm ischemia time (WIT), CIT, and liver function tests. The recipient information included the following: age, sex, cause of liver failure, cause of graft loss, presence of hepatocellular carcinoma, raw MELD score at the time of OLT, amount of steatosis in the liver graft, cytomegalovirus (CMV) infection needing treatment after OLT, diabetes mellitus status, median follow-up time, acute rejection episodes documented by liver graft biopsy, and treatment of rejection. For DCD donors, the donor warm ischemia time (DWIT) was defined as the time from the withdrawal of both ventilator and cardiac support to aortic cold perfusion of the organ. CIT was defined as the time from the infusion of the cold preservation solution until portal reperfusion of the liver in the recipient. Outcomes included the patient and graft survival rates, the retransplantation rate, and the incidence of PNF, IC, and HAT.
In DCD donors, the withdrawal of support and the declaration of death were in strict compliance with donor hospital policies. The transplant team was not involved in the withdrawal process. An independent physician from the donor hospital, who was separate from the organ procurement organization and the transplant center, was assigned to withdraw artificial life support and provide end-of-life care to the patient. After the declaration of death by the independent physician, another 2 to 5 minutes of mandatory observation was performed as described in the 1997 Institute of Medicine guidelines.7 During the 2- to 5-minute waiting period, the patient was transported to the operating room (if he or she was not already there) and prepared for organ recovery. Heparin was administered to the patient according to the donor hospital policy. A rapid retrieval technique was used: the abdomen was opened with a cruciate incision; the aorta and portal systems were cannulated; and a cold preservation fluid, which consisted of University of Wisconsin solution, heparin, and glutathione, was then flushed through the abdominal aorta and the inferior mesenteric vein. The intrathoracic descending aorta was cross-clamped by the opening of the chest or left hemidiaphragm. Finally, the suprahepatic inferior vena cava was opened to allow venting. The liver was then removed, and the biliary system was flushed on the back table. The retrieval of organs from DBD donors was performed with University of Wisconsin solution according to a standard technique.
All transplants were performed with the piggyback technique without a portocaval shunt or caval clamping. Standard triple-drug immunosuppression with tacrolimus, mycophenolate mofetil, and corticosteroids was used in all cases. Recipients underwent percutaneous or transjugular liver biopsy at 4 months, at 1 year, and annually after OLT (protocol biopsy) or whenever it was clinically indicated. Pathologists interpreting liver biopsy samples were blinded to the type of liver graft. Banff criteria were used for the diagnosis of rejection.8 All patients with biopsy-proven moderate to severe acute rejection episodes were treated with intravenous steroid boluses (methylprednisolone: day 0, 1000 mg; day 2, 500 mg; and day 4, 500 mg). Liver biopsy was repeated to ensure the resolution of rejection. The CMV prophylaxis was oral valganciclovir hydrochloride (900 mg/day for 100 days post-transplantation) in CMV serology mismatch pairs (a positive donor and a negative recipient).
Categorical variables were examined with the chi-square test, and continuous variables were analyzed with the Kruskal-Wallis test. In the event of significant differences between the groups on the Kruskal-Wallis test, a pairwise comparison with the Mann-Whitney U test with a Bonferroni adjustment was performed; the P value to be achieved for significance was divided by the number of paired comparisons made. Patient survival, graft survival, and progression to fibrosis ≥ stage 2 among the patient groups were compared with Kaplan-Meier plots and log-rank tests. Graft survival was timed from the date of transplantation to the date of retransplantation or death (whichever came first) and was censored for the date of the end of the study period or for the date of the last correspondence for patients lost to follow-up. Patient survival was timed from the date of transplantation to the date of death and was censored for the date of the end of the study period or for the date of the last correspondence for patients lost to follow-up. Progression to fibrosis ≥ stage 2 was timed from the date of transplantation to the date of occurrence on protocol biopsy and was censored for the date of the end of the study period or for the date of the last correspondence for patients lost to follow-up, the date of death, or the date of retransplantation. Cox regression was used in the univariate and multivariate analyses of predictors of patient survival, graft survival, and fibrosis progression. Variables significant at P < 0.25 on the univariate analysis and variables that might have confounding effects or were clinically interesting were entered into the multivariate model. Multicollinearity among covariates was examined through correlations of regression coefficients. None of the variables entered into the multivariate model were collinear. The assumption of proportional hazards was assessed with log-log plots and with an examination of the correlation of Schoenfeld residuals with time. Significance was defined at a P value < 0.05 (<0.017 with a Bonferroni adjustment). Statistical analysis was performed with SPSS software, version 17.0 (SPSS, Inc., Chicago, IL).
Between January 2003 and September 2009, OLT was performed 1336 times by the Mayo Clinic Florida transplant program; 160 cases (12.0%) involved DCD donors, and 1176 (88%) involved DBD donors. Eighty-three HCV+ patients received DCD liver grafts. We were unable to match 6 patients according to the matching criteria. Therefore, the 77 remaining patients (HCV+DCD+) were matched to HCV+DBD+ patients. The HCV−DCD+ group included 77 patients who had end-stage liver disease secondary to a non-HCV disease etiology who received DCD liver grafts.
The recipient and donor characteristics and perioperative data are summarized in Table 1. The median follow-up was 52 months. There were no differences among the groups with respect to the recipient age, recipient sex, donor age, DWIT, CIT, WIT, or MELD score. None of the patients in either HCV+ group received grafts from donors with a positive HCV serology. One patient in the HCV+DCD+ group died of cardiac arrest during OLT. The perioperative packed red blood cell (pRBC) transfusions and the lengths of the intensive care unit stay and hospital stay were also similar among the groups. Peak aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels in the first posttransplant week were significantly higher in the HCV+DCD+ and HCV−DCD+ groups versus the HCV+DBD+ group (P = 0.04; Table 1).
Table 1. Recipient and Donor Characteristics and Perioperative Data
HCV+DCD+ (n = 77)
HCV+DBD+ (n = 77)
HCV−DCD+ (n = 77)
NOTE: Categorical variables are presented as numbers and percentages; continuous variables are presented as means and standard deviations (with medians and ranges in parentheses).
n = 76.
The pairwise comparison was significant for HCV+DCD+ patients versus HCV−DCD+ patients.
The pairwise comparison was significant for HCV−DCD+ patients versus HCV+DBD+ patients.
The pairwise comparison was significant for HCV+DCD+ patients versus HCV+DBD+ patients.
Postoperative complications are summarized in Table 2. Four patients in the HCV+DCD+ group had PNF (P = 0.017). In the HCV+DCD+ group, 6 patients (7.8%) developed IC, whereas in the HCV−DCD+ group, 11 patients (14.3%) developed IC (P = 0.2); there was no IC in the HCV+DBD+ group (P = 0.003). Anastomotic biliary strictures requiring reoperation or an endoscopic/percutaneous procedure were found in 13 patients (16.9%) in the HCV+DCD+ group, in 12 patients (15.5%) in the HCV+DBD+ group, and in 13 patients (16.9%) in the HCV−DCD+ group (P = 0.963). HAT occurred in 5 patients (6.5%) in the HCV+DCD+ group, in 3 patients (3.9%) in the HCV+DBD+ group, and in 1 patient (1.3%) in the HCV−DCD+ group (P = 0.241). Moderate to severe acute rejection episodes and CMV infection rates were similar among the groups (P = 0.770 and P = 0.527, respectively). Only 1 patient in the HCV+DCD+ group received antilymphocyte treatment for recurrent moderate to severe rejection 3 months post-transplant; this patient was alive 53 months after OLT without signs of recurrent HCV. Ten patients (13.0%) in the HCV+DCD+ group, 9 patients (11.6%) in the HCV+DBD+ group, and 7 patients (9.1%) in the HCV−DCD+ group underwent retransplantation at the end of the study period (P = 0.359).
Table 2. Postoperative Complications
HCV+DCD+ (n = 77)
HCV+DBD+ (n = 77)
HCV−DCD+ (n = 77)
NOTE: Variables are presented as numbers and percentages.
Portal vein thrombosis
Acute rejection (requiring steroid bolus)
Graft and patient survival rates are summarized in Figs. 1 and 2, respectively. There were no differences in 1-, 3-, and 5-year graft and patient survival rates among the groups. Pairwise graft and patient survival comparisons between the HCV+DCD+ and HCV+DBD+ groups at 1, 3, and 5 years were also similar (data not shown).
A Cox regression model was used to investigate the factors predicting graft loss and fibrosis progression (Tables 3 and 4). On univariate analysis, the MELD score and CMV infection in the first year post-transplant were significantly associated with an increased risk of graft loss. Both variables were included in the multivariate model. To adjust for the possible confounding effect of recipient age, this variable was entered into the multivariate model, as was the graft type, which was the treatment variable. Because there were differences among the groups in terms of local share versus nonlocal share, we adjusted for this factor in the multivariate analysis. With adjustments for recipient age, graft type, and organ sharing, both the MELD score [hazard ratio (HR) = 1.037, 95% confidence interval (CI) = 1.006-1.069, P = 0.018] and CMV infection in the first year post-transplant (HR = 3.367, 95% CI = 1.493-7.593, P = 0.003) remained significant predictors of graft loss (Table 3).
Table 3. Univariate and Multivariate Analyses of Risk Factors for Graft Loss.
HR (95% CI)
HR (95% CI)
Donor age (years)
Recipient age (years)
Donor risk index
Acute rejection in the first year post-transplant
CMV infection in the first year post-transplant
Table 4. Univariate and Multivariate Analyses of Risk Factors for ≥Stage 2 Fibrosis in HCV+ Recipients
HR (95% CI)
HR (95% CI)
Donor age (years)
Recipient age (years)
Log viral load (IU/mL)
Donor risk index
Moderate to severe rejection in the first year post-transplant
A comparison of the progression of fibrosis based on liver biopsy samples between the HCV+ groups at 4 months and at 1 to 5 years did not show any statistical difference (Fig. 3). Among the risk factors studied in the multivariate analysis, HCV genotype 1 was significantly associated with the development of stage 2 fibrosis (genotype 1 versus non-1 genotypes: HR = 2.739, 95% CI = 1.047-7.143, P = 0.040; Table 4). Recipients who developed moderate to severe acute rejection in the first year post-transplant had an increased risk of developing stage 2 fibrosis in comparison with recipients who did not, and the association neared significance (HR = 1.953, 95% CI = 1.001-3.808, P = 0.050). The type of liver graft (DCD versus DBD) was not a factor for either graft loss or the development of stage 2 fibrosis in HCV+ recipients. It was not possible to examine the association between CMV infection in the first year post-transplant and the risk of developing stage 2 fibrosis because 5 of 7 patients who had CMV viremia experienced graft loss during the first year before the development of stage 2 fibrosis. Because the Kaplan-Meier plots for the development of fibrosis between the 2 study groups seemed to diverge approximately 48 months post-transplant, an extended Cox regression model using Heaviside step function was performed to investigate differences in HRs between time intervals (<48 versus ≥48 months): there was no statistical difference between the time periods (HR = 1.174 and 95% CI = 0.659-2.090 for <48 months versus HR = 3.082 and 95% CI = 0.730-13.011 for ≥48 months, P = 0.587).
Twenty-five patients (32.5%) in the HCV+DCD+ group and 30 patients (39.0%) in the HCV+DBD+ group underwent antiviral treatment during the first year after OLT (P = 0.4). Fourteen of 25 patients (56%) in the HCV+DCD+ group and 18 of 30 patients (60%) in the HCV+DBD+ group had a sustained virological response (P = 0.76).
Six HCV+ patients who received DCD grafts were unable to be matched to any patients in the DBD group (5 males and 1 female, mean age = 46.8 years, mean MELD score = 17.7). Two of these were alive 29 and 43 months after OLT. The mean graft survival and the mean patient survival for 4 patients who died were 24.2 and 33 months, respectively.
HCV is the most common indication for OLT in the United States. Approximately 35% of the recipients in our program have end-stage liver disease due to HCV. The average waiting times on the list for HCV+ patients increased after 2003 when our practice adopted the policy of using liver grafts from donors less than 60 years old for HCV+ recipients.9 With the limited number of organs available from DBD donors, one option for increasing the donor pool is the use of DCD donors. DCD liver grafts have been used at a relatively low frequency because of the poor graft survival rates reported by different institutions and pooled United Network for Organ Sharing data.10-13 We have previously reported long-term graft and patient outcomes for recipients with DCD grafts.1, 14 Apart from the inherent problems related to the DCD graft itself, reports looking at the results of these inferior grafts in HCV+ patients are scarce. It is well known that overall patient and graft survival is inferior for HCV patients versus non-HCV patients.2, 15 Interactions between the primary disease and the quality of the organ (eg, age, ischemic time, reperfusion injury, and graft size) as well as the presence of comorbidities in the recipient play significant roles in graft and patient survival. An approach to reversing inferior outcomes for HCV+ patients is the optimization of donor selection. One previous publication showed that overall graft and patient survival rates at 1 and 5 years were significantly lower when DCD grafts were used in HCV patients.4 However, the number of DCD grafts used in that retrospective review was small, and it was a cohort comparison without direct matching between DBD and DCD graft recipients. A recently published match-controlled, retrospective review of 37 HCV+ patients who received DCD liver grafts demonstrated that there was a trend toward poorer outcomes in DCD graft recipients; however, because of the small number of patients, statistical significance could not be reached. In addition, because of the large percentage of grafts lost during the first year after OLT in the DCD cohort, a firm conclusion regarding the recurrence of HCV was difficult to extract.5 Our study presents the largest number of DCD liver graft recipients to date. Our primary goal was to compare patient and graft survival when DCD grafts were used for HCV+ patients. The match-controlled design of the study was used to eliminate covariables as much as possible and thus provide a clear picture of survival differences.
This is a single-center analysis with uniform surgical techniques and strict medical protocols used throughout the study period and with complete long-term follow-up. In this study, the graft and patient survival rates of HCV+ patients were similar regardless of the graft type that they received: even when the inherent problems of DCD grafts were taken into consideration and when we controlled for other recipient- and donor-related factors, there was no difference in overall graft and patient survival or HCV recurrence between DCD and DBD grafts in HCV+ recipients. As expected, the postoperative peak AST and ALT levels were higher in the DCD group, and this was a reflection of ischemic injury related to the DWIT at the time of procurement. Similarly, factor V, which was measured daily after OLT, started at a lower level in the DCD group, and this perhaps was a reflection of somewhat delayed function in the DCD grafts within the first few postoperative days. Similarly to our previous assessment of our overall DCD graft experience, the incidence of complications inherent to DCD grafts (ie, IC, PNF, and HAT) was less than that reported in the literature. In fact, the incidence of HAT, hepatic artery stenosis, bile leakage, and anastomotic biliary strictures in DCD grafts was similar to the incidence in DBD grafts. IC remains the Achilles heel of DCD liver grafts. In most patients, the development of IC requires retransplantation. This is a significant clinical and allocation dilemma because patients with IC experience recurrent episodes of cholangitis that require repeat hospitalizations and procedures, but they retain low MELD scores that preclude early retransplantation.16 In contrast to previously published single-center reports, we achieved lower IC rates in patients with DCD grafts. Although the reasons for the low IC rates in our overall practice are currently not clear to us, we can postulate that they may be a result of short DWITs and CITs and overall surgical experience that we have gained from the large volume of DCD grafts.
In multivariate analysis, we examined donor and recipient factors that would affect graft survival as well as the development of stage 2 fibrosis. Among the studied factors, only the MELD score at the time of OLT and CMV infection within the first postoperative year affected graft survival. Among the factors studied for fibrosis progression, HCV genotype 1 emerged as a significant factor, whereas moderate to severe acute rejection episodes within the first year approached statistical significance in multivariate analysis. Interestingly, factors previously thought to be risks for graft loss or HCV recurrence, such as steatosis in the liver graft and posttransplant diabetes, did not increase rates of either graft loss or fibrosis progression.
The gap between supply and demand for liver grafts is widening worldwide. To close this gap, livers from DCD donors should be considered more frequently by transplant centers. DCD grafts have been used more extensively since 1997 when the Institute of Medicine determined that these organs are medically effective and ethically acceptable. DCD grafts are considered to be less than optimal for transplantation because of the damaging effect of a variable warm ischemia time before cold preservation. Grafts of reduced quality may show increased sensitivity toward additional damaging events such as HCV and acute rejection episodes. Extended criteria donor features having an impact on the outcomes of HCV+ recipients, such as donor age, allograft steatosis, and prolonged ischemia times, have been reported previously.3 However, only 1 previous study has specifically assessed the impact of using DCD liver grafts in HCV+ patients.3, 5, 14
Extended criteria donor usage should be justified by adequate recipient and graft survival and acceptable early graft function. As suggested previously, the systematic utilization of extended criteria donors, including DCD donors, maximizes donor use, increases access to OLT, and reduces wait-list mortality by providing satisfactory outcomes for select recipients.17 Our analysis illustrates that patients with HCV had satisfactory outcomes with DCD liver grafts.
As the organ scarcity intensifies, further emphasis on expanding donor criteria will occur. The benefit of earlier access to OLT provided by a DCD graft could outweigh the risks of prolonged waiting for a standard graft. With the utilization of DCD grafts, it is clear that the recipients assume an earlier increased risk related to the graft; however, they avoid the morbidity and mortality associated with waiting. Since the beginning of 2003, our program has not used liver grafts from donors older than 60 years for HCV+ patients. As such, HCV+ patients are limited to younger donors, so their wait times are longer than those of patients with other etiologies of liver disease. As a result, DCD grafts were used more often in HCV+ patients (51%) compared to overall percentage of HCV+ patients in our program (35%). It is clear that we have taken advantage of DCD graft utilization in HCV+ patients. This approach has likely shortened wait times for HCV+ patients and affected the overall wait-list mortality rate.18, 19 Even though this is the largest reported single-center experience, the number of patients is relatively limited for adjustments for all the risk factors for HCV-related graft loss and fibrosis progression. The limited number of patients may have resulted in a type 2 error for detecting differences in disease progression.
The most appropriate use of scarce livers continues to evolve. In conclusion, this match-controlled, retrospective analysis demonstrates that DCD liver graft utilization did not cause untoward effects on disease progression or patient and graft survival in comparison with DBD liver graft utilization in HCV+ patients.