Factors that identify survival after liver retransplantation for allograft failure caused by recurrent hepatitis C infection



Hepatitis C virus (HCV) is becoming the most common indication for liver retransplantation (ReLTx). This study was a retrospective review of the medical records of liver transplant patients at our institution to determine factors that would identify the best candidates for ReLTx resulting from allograft failure because of HCV recurrence. The patients were divided into 2 groups on the basis of indication for initial liver transplant. Group 1 included ReLTx patients whose initial indication for LTx was HCV. Group 2 included patients who received ReLTx who did not have a history of HCV. We defined chronic allograft dysfunction (AD) as patients with persistent jaundice (> 30 days) beginning 6 months after primary liver transplant in the absence of other reasons. HCV was the primary indication for initial orthotopic liver transplantation (OLT) in 491/1114 patients (44%) from July 1996 to February 2004. The number of patients with AD undergoing ReLTx in Groups 1 and 2 was 22 and 12, respectively. The overall patient and allograft survival at 1 year was 50% and 75% in Groups 1 and 2, respectively (P = .04). The rates of primary nonfunction and technical problems after ReLTx were not different between the groups. However, the incidence of recurrent AD was higher in Group 1 at 32% versus 17% in Group 2 (P = .04). Important factors that predicted a successful ReLTx included physical condition at the time of ReLTx (P = .002) and Child-Turcotte-Pugh score (P = .008). In conclusion, HCV is associated with an increased incidence of chronic graft destruction with a negative effect on long-term results after ReLTx. The optimum candidate for ReLTx is a patient who can maintain normal physical activity. As the allograft shortage continues, the optimal use of cadaveric livers continues to be of primary importance. The use of deceased donor livers in patients with allograft failure caused by HCV remains a highly controversial issue. (Liver Transpl 2004;10:1497–1503.)

The growing number of candidates for liver transplantation has not been followed by a corresponding increase in the number of deceased donors. Even more problematic is the increase in the numbers of transplant recipients requiring retransplantation (ReLTx) for chronic allograft dysfunction (AD). These facts have led to the exploration of alternatives such as split liver transplantation, the use of live donors, and to a more liberal use of extended criteria donors.1–7 Nonetheless, the donor–recipient numeric disparity still prevails. ReLTx, the sole treatment for irreversible liver graft failure after liver transplantation, is now aggravating the scarcity of donor resources.8–10

Hepatitis C virus (HCV) recurrence after orthotopic liver transplantation (OLT) is universal, but the natural history of recurrent HCV in the allograft is quite variable and difficult to predict.11–13 Although patient and graft survival vary by transplant center, there is no dispute that fibrosis develops at a faster rate in transplanted patients with HCV. As many as 20 to 30% will progress to cirrhosis within 5 years of the initial transplant.1, 14–16

Unfortunately, ReLTx in HCV-infected patients has demonstrated marginal results, and reports suggest an accelerated rate of allograft failure in the ReLTx patient caused by HCV recurrence.17 This situation is frequently complicated by marked muscle wasting. However, the course after ReLTx may not be dismal in all patients.

Parameters may be available to better identify the optimum candidate for ReLTx.18 To this end, we examined several pre-ReLTx parameters—including evaluating physical condition, HCV tolerance to antiviral therapy, serum bilirubin, and renal function—with the aim of identifying factors associated with improved post-ReLTx prognosis and higher survival rates.


HCV, Hepatitis C virus; ReLTx, liver retransplantation; OLT, orthotopic liver transplantation; AD, allograft dysfunction; CTP, Child-Turcotte-Pugh score; MELD, model for end-stage liver disease; HCV-ND, HCV nondetectable.


Patient Population

We conducted a retrospective medical record review of patients who received a liver transplant at the University of Cincinnati and the University of Miami from July 1996 to February 2004. Patients were divided into 2 groups according to reason for undergoing their initial liver transplant. Those patients who were transplanted for cirrhosis because of HCV were assigned to Group 1; Group 2 patients had other reasons for the initial transplant. If graft failure after initial transplant occurred, the groups were then further divided into one of 3 causes: chronic AD, primary graft nonfunction, and surgical or technical difficulties. AD was defined as graft dysfunction with persistent jaundice (> 30 days) beginning at least 6 months after primary transplant in the absence of other causes such as vascular or biliary complications, infection, or medication-associated cholestasis. Primary graft nonfunction was defined as massive aminotransferase elevation, encephalopathy, and coagulopathy within 10 days of liver transplantation. Surgical or technical difficulties were defined as patients with allografts with hepatic artery thrombosis, biliary ischemia, or organ failure within the first 6 months after liver transplantation. Recurrent HCV was defined as histologic changes on liver biopsy consistent with recurrent hepatitis, post–liver transplant detection of HCV RNA by reverse transcriptase polymerase chain reaction in association with elevated serum chemistries. Vascular and biliary injuries were excluded by radiographic evaluation and, when necessary, percutaneous cholangiogram or arteriogram.

We collected data on baseline patient demographics, antiviral therapy and results, patient and graft survival, initial and repeat indications for liver transplantation, patient and allograft survival, Child-Turcotte-Pugh (CTP) scores and the model for end-stage liver disease (MELD) when the international ratio values were available.

Immune Suppression

The institution's standard post–liver transplant immune suppression protocol consisted of a combination of tacrolimus and methylprednisolone. Tacrolimus was given immediately before OLT at a dose of .05 mg/kg orally and continued postoperatively with dose adjustments to maintain whole blood 12-hour trough level at 15 ng/mL during the first postoperative week, 10 ng/mL for the first 3 months, and 8–10 ng/mL thereafter. Methylprednisolone was initiated with 1-g intravenous bolus immediately after the reperfusion of the hepatic graft and then tapered to 200 mg per day (day 1), 160 mg per day (day 2), 120 mg per day (day 3), 80 mg per day (day 4), 40 mg per day (day 5), 20 mg per day (day 6), and then reduced to 0 within 3–6 months after OLT.

Physical Condition

The patient's physical condition was determined by an adaptation of the American Society of Anesthesiologists' Physical Scale classification.19 Specifically, condition was assessed by measuring the daily activity level and assigning a numeric value with 1 the best score and 4 the worst according to the following criteria:

  • Level 1: Patient at home and working full time with no limitation of daily activities with occasional complaints of fatigue as a consequence of muscle wasting and deconditioning.

  • Level 2: Patient at home and able to maintain his or her daily activities but limited by fatigue. Patient was not able to maintain full- or part-time employment because of fatigue.

  • Level 3: Patient predominately limited to bed rest with little or no activity and solely dependent on another individual for ambulation and all daily activities. Patient was not hospitalized before ReLTx for liver failure.

  • Level 4: Patient predominately limited to bed rest with little or no activity and solely dependent on another individual for ambulation and all daily activities. This patient was hospitalized before ReLTx for liver failure accompanied by malnutrition, fatigue, and deconditioning.

The database of all patients for the study within the allotted period of time was collected using Microsoft Excel; this database formed the basis for the final analyses.. Data were analyzed by STATISTICA (StatSoft, Tulsa, OK) and Stata version 8.0 (Stata Corporation, College Station, TX). Values were expressed as means ± standard deviation, unless otherwise noted. This study was approved by the University of Miami and University of Cincinnati institutional review boards.



Liver transplantation for cirrhosis was performed in 1114 adults in the study population from July 1996 to February 2004, with 491 patients (44%) suffering from chronic HCV infection. Sixty-seven of 491 patients (14%) had recurrent HCV and resultant allograft failure, comprising Group 1. Thirty-three percent of these patients 30% (20/67) had allograft failure resulting from PNF, and 37% (25/67) had surgical/technical complications (Fig. 1).

Figure 1.

Indication for primary graft loss between HCV orthotopic liver transplant recipients (Group 1; light box) and non-HCV orthotopic liver transplant recipients (Group 2; dark box). PNF, primary nonfunction; TEC, technical/surgical complications; AD, chronic allograft dysfunction; HCV, hepatitis C virus.

There were 623/1114 adults (56%) transplanted for non-HCV causes. This group contained 60/623 patients (10%) that suffered graft failure from a variety of non-HCV related causes, comprising Group 2. Twenty percent (12/60) were ReLTx for AD, 40% (24/60) for primary non-function, and an additional 40% (24/60) for surgical/technical complications (Fig. 1). The proportion of each group retransplanted for AD trended toward a significant difference—33% in Group 1 vs. 20% in Group 2 (p = 0.10, chi-square test).

CTP and MELD Levels

In the patients with AD, the mean time period from initial OLT to ReLTx in Groups 1 and 2 were 642 and 879 days respectively (P = NS,Table 1). Regarding renal function, serum creatinine greater than 1.5 mg/dL was found in 6/22 (27%) in Group 1 and in 4/12 (33%) in Group 2. No significant difference was seen in CTP (Group 1 = 9.5, Group 2 = 10.0) or MELD scores (Group 1 = 21.7, Group 2 = 20.9) at time of ReLTx.

Table 1. Comparison of Survival in Days Between the 2 Groups for Primary Transplantation and Retransplantation
Time PeriodGroup 1 HCV+Group 2 HCV−P Value
  1. Abbreviations: ReLTx, retransplantation; HCV, hepatitis C virus.

Primary to ReLTx213–2481172–2158.13
ReLTx to present10–24291–2858.12

Activity Levels

The mean activity level at the time of ReLTx differed numerically between the 2 groups. Level of activity for Group 1 was 2.8 (range 1–4), and for Group 2 was 1.6 (range 1–4). Although not statistically significant, it appeared that the patients in Group 2 had greater activity in daily living, despite a similar CTP score.

Transplant recipients suffering from chronic allograft dysfunction had graft survival rates of 50% and 75% at 1 year in Groups 1 and 2, respectively (P = .04). The rates of primary graft nonfunction and surgical or technical difficulties after ReLTx were not different between Groups 1 and 2; however, the incidence of AD was significantly higher in Group 1 (32% versus 17%, P = .04). Recurrent AD after the first ReLTx resulted in 73% and 36% graft loss in Groups 1 and 2, respectively (P = .04).

HCV Viral Load and Antiviral Therapy

Serum HCV testing at time of ReLTx revealed that 8 patients in Group 1 were HCV nondetectable (HCV-ND), 5 patients had fewer than 2 million copies, and 9 patients had levels greater than 2 million copies by reverse transcriptase polymerase chain reaction. Antiviral therapy with interferon monotherapy or with ribavirin was administered to 6/22 (27%) patients at the time of ReLTx. Viral recurrence post-ReLTx occurred in 21/22 (95%) of patients, 8 of whom received pegylated interferon and ribavirin therapy within 4 months of ReLTx. Three of these 8 patients had a virologic response on therapy with a nondetectable HCV RNA by reverse transcriptase polymerase chain reaction. The mean serum alanine aminotransferase was 88 IU/dL but was less than 40 IU/dL in 4 patients (3 patients during antiviral therapy with HCV-ND and 1 patient after treatment with HCV-ND). Biopsies were performed at 12 months of therapy in 3 of 7 patients and demonstrated persistent virus with a mean fibrosis score of stage 1.

Predictors of Survival Post-Retransplantation

To date, 10/22 (45%) patients in Group 1 remain alive. Three of these patients are currently HCV- ND, and all had maintained an activity level of 1–2 before ReLTx. The activity level in all patients who died was at least 3. In this group of retransplanted patients, serum bilirubin and creatinine were not significant predictors of mortality. One-year mortality was higher in Group 1 (55% vs. 33% in Group 2) and was related to a variety of causes, including sepsis (n = 6), pancreatitis (n = 1), cardiac failure (n = 3), and allograft failure related to HCV recurrence (n = 2) (Table 2). Survival time was also shorter in Group 1, although this trend did not achieve statistical significance (Fig. 2). Multivariate regression analysis determined that physical condition was the only significant factor predicting post-ReLTx survival (Table 3).

Table 2. Demographic Information Including Retransplant Status With Patient Survival Information for Patients With HCV
PatientGenderAge (yr)Time from Primary OLT (days)HCV PCR ReLTx (copies/mL)Interferon Therapy at Time of ReLTxCr (mg/dL)Total Bil (mg/dL)MELDCTPFunction Level (1–4)Cause of DeathSurvival (days)
  1. Abbreviations: CR, chronic rejection; HCV, hepatitis C virus; ND, nondetectable; Int/riba; interferon and ribavirin; PEG, pegylated interferon and ribavirin; bil, bilirubin; MI, myocardial infarction; ReLTx, retransplantation; PCR, polymerase chain reaction; MELD, model for end-stage liver disease; CTP, Child-Turcotte-Pugh.

1F53840RNA > 37None1.029.620102Alive2320
2M49878RNA > 73Int/riba1.60.81592Alive1866
3F25453RNA > 100Yes1.013.51573Alive842
4M501416RNA 347 K IUInt/riba1.210.419101Alive1515
5F71286RNA < 1None0.60.91683Alive2436
6M48306RNA > 120None2.3483693Alive838
7F49509RNA < 2Int/riba1.59.62093Alive2460
8M48716HCV RNA > 11None1.11.816104Sepsis762
9M41601HCV RNA > 56None2.0282694Pancreatitis627
10M492446HCV RNA > 11None1.82.818103Sepsis2687
11M47209RNA < 200 KPEG0.81518114Sepsis223
12M53848RNA < 200 KPEG1.2241883Sepsis1726
13F66525RNA NDPEG2.04930104Sepsis619
14M41431RNA NDNone1.13725114Sepsis601
15M48440RNA > 10None0.61212104Recurrent disease789
16M55331RNA > 8None2.53420104MI362
17M501131RNA NDNone1.94118103MI1226
18F33759RNA > 120None1.0243093Recurrent disease874
19F58828RNA NDNone1.215.315124MI861
20M50604RNA NDNone1.5352591Alive75
21M48180RNA > 5None1.5151891Alive163
22F521753RNA > 10None1.4252091Alive138
Figure 2.

Kaplan–Meier survival curves for HCV+ retransplanted patients (Group 1) and HCV– retransplanted patients (Group 2). The mean days of survival for Group 1 = 1498 (SE = 192); mean days of survival for Group 2 = 1931 (SE = 242). However, the largest observed survival time for Group 2 is censored, so the mean is underestimated. The difference in survival functions trends toward statistical significance (P = .15, log-rank test). HCV, hepatitis C virus; CI, confidence inteval; SE, standard error.

Table 3. Statistical Analysis of Various Retransplantation Parameters for Survival
Univariate Results
ParameterHazard RatioP Value
  1. Abbreviations: MELD, model for end-stage liver disease; CTP, Child-Turcotte-Pugh.

Physical Condition7.56.002
Multivariate Results
ParameterHazard RatioP Value
Physical Condition8.19.02

Surgical Variables

There were no statistically significant differences between the 2 groups for donor age (29.8 years versus 35.8 years) or cold ischemia times.


The literature regarding post-ReLTx outcomes for recurrent HCV-related allograft failure demonstrates conflicting results.11, 16, 20, 21 Some centers have suggested that the survival difference between HCV-infected grafts and non-HCV-infected grafts is negligible, but others are less optimistic.11, 15–18, 22–25 Overall, however, the consensus is that transplantation for HCV results in an increased incidence of allograft dysfunction and has a negative effect on the results of ReLTx,26 even with interferon-based therapy.27 When ReLTx is necessary because of allograft dysfunction there also appears to be an accelerated destruction of the ReLTx graft observed in HCV-infected patients.

Limited information from transplant centers predicts survival after ReLTx. Rosen et al. showed that candidates with total bilirubin greater than or equal to 5 mg/dL and an elevated serum creatinine should not be considered.18 McCashland recently showed that MELD score could predict survival in ReLTx; specifically, a MELD score greater than 25 was predictive of a poor outcome, most likely attributed to combined renal failure and cholestasis.20 The donor organ scarcity also has forced us to search for parameters that will identify the candidate with the best chance at survival when considering ReLTx for patients with recurrent HCV.

In our study, we found that the rate of ReLTx for patients with end-stage liver disease caused by HCV was no different than patients with other indications for their initial liver transplant. Neither the CTP score nor the MELD score predicted survival after repeat liver transplant. Use of CTP/MELD scores as ReLTx exclusion criteria would have eliminated 10 patients in our review. Importantly, 3 of these patients are currently alive.

The rates of primary graft nonfunction and after ReLTx were not different between Groups 1 and 2; however, the incidence of AD was higher in Group 1 (32% versus 17%, P = .04). Recurrent AD after the first ReLTx resulted in 73% and 36% graft loss in Groups 1 and 2, respectively (P = .04).

In contrast to Rosen et al., we found no significant differences in the CTP or MELD scores at the time of ReLTx, although it should be noted that our sample size may have been too small to detect potential differences. The most significant predictors were the physical status of the patient at the time of ReLTx (P = .002) and CTP score (P = .008). All 3 patients who are currently alive and HCV-ND maintained an activity level of 1–2 before ReLTx. The activity level in all patients who died was 3 or more, although it should be noted that the physical activity scale adaptation used in this study has not been validated within this context. The results, however, clearly suggest that physical condition before ReLTx is of prognostic importance and that efforts to maintain or improve patient conditioning before transplantation should be implemented when possible.

Reports from the International Liver Transplant Society in 2003 suggest that interferon therapy before ReLTx may help prevent serum HCV recurrence.28–30 It has been our experience that few patients in this clinical setting can tolerate interferon therapy, especially with the addition of ribavirin, unless they are in good physical condition. This is in agreement with earlier reports.27 In fact, Charlton et al. have reported through a retrospective audit that 90% of patients require 12 weeks to recover after liver transplantation.31 This work suggests that allowing patients a full 12 weeks to recover may assist in rendering them better able to tolerate the common, and often significant, drug-related side effects associated with interferon-based treatment. Our results demonstrate that most of the patients listed with AD were in poor physical condition (Table 3) and were unable to tolerate interferon therapy.

Our current approach to evaluating the hypothetically most successful ReLTx candidate with HCV can be found in Fig. 3. This figure includes composite information from both the literature and our study results.18 Once the candidate has been selected for ReLTx, we implement an aggressive therapeutic approach along with antiviral therapy (if tolerated) and a nutritional and physical therapy program. Decisions regarding the initiation of antiviral therapy are based on the candidate's capacity to tolerate interferon therapy, historical response to HCV antiviral therapy, HCV viral load, and genotype. Candidates with genotype 1, those patients unresponsive to interferon because of intolerance or other reasons, or those with a high HCV viral titer (> 800,000 IU/mL) are considered high risk and most likely would not benefit from ReLTx. In should be noted, however, that 5 patients with HCV were indeed negative for HCV RNA at the time of ReLTx, suggesting progression of allograft failure despite apparent viral clearance. This has been observed previously in the context of interferon treatment.27 Conversely, active candidates without muscle loss and known response to HCV antiviral therapy, including stable HCV polymerase chain reaction, tend to be good candidates. This is a speculative model that has yet to be prospectively evaluated for predictive values.

Figure 3.

Retransplantation prognostic grid using information from our review and current literature for candidates with recurrent HCV. HCV status is scaled from best to worst (1–4), depending on the number of positive factors as follows: 1. High HCV viral load (serum HCV RNA > 850,000 copies/mL); 2. Genotype 1; 3. Poor virological response historically to interferon; 4. Intolerance to antiviral therapy. Physical function is scaled from best to worst (Level 1–Level 4); See text for a description of the physical function.

The debate regarding ReLTx in the HCV LTx population hopefully will resolve as safer, more effective HCV therapies emerge. In view of these results, along with the scarce donor resources and limited effect of antiviral therapies against HCV, in-depth multicenter studies are needed to better identify parameters that will predict individual benefit from ReLTx. In conclusion, HCV transplant recipients who require ReLTx from AD have a better opportunity at survival when maintaining an active life style, normal bilirubin, and serum creatinine and are tolerant and responsive to antiviral therapy.