Preliminary reports suggested that hepatitis C virus (HCV) infection has a more aggressive course following living donor liver transplantation (LDLT) compared to cadaveric liver transplantation (CLT). The aim of this prospective study was to establish if HCV disease recurrence differs between LDLT and CLT. A cohort of 116 consecutive HCV-infected patients undergoing 117 LTs in a single center from March 2000 to August 2003 were followed-up, including systematic liver biopsies. Severe recurrence (SR) was defined as biopsy-proven cirrhosis and/or the occurrence of clinical decompensation. After a median follow-up of 22 months (2.6–44 months), 26 (22%) patients developed SR (decompensation in 12), involving 17 (18%) of 95 patients undergoing CLT and 9 (41%) of 22 undergoing LDLT. The 2-year probability of presenting SR was significantly higher in LDLT compared to CLT (45% vs. 22%, P = .019). By univariate analysis LDLT (P = .019) and an ALT higher than 80 IU/L 3 months after LT (P = .022) were predictors of SR. In 93 patients from whom a liver biopsy was available 3 months after LT, a lobular necroinflammatory score >1 (P < .01), LDLT (P < .01), and biliary complications (P = .046) were associated with SR. However, the only variables independently associated with SR were LDLT (odds ratio [OR], = 2.8; 95% CI,1.19-6.6; P = .024) and a lobular necroinflammatory score >1 (OR, 3.1; 95% CI, 1.2-8; P = .013). In conclusion, HCV recurrence is more severe in LDLT compared to CLT. Although our results were based on a single-center experience, they should be considered in the decision-making process of transplant programs, since severe HCV recurrence may ultimately compromise graft and patient survival. (HEPATOLOGY 2004; 40:699–707.)
Hepatitis C virus (HCV)–related cirrhosis is the leading indication for liver transplantation (LT) in the United States and Europe.1, 2 More than half of the patients on the waiting list are infected with HCV. Regretfully, HCV recurrence is universal after LT3 and leads to chronic hepatitis and liver cirrhosis in a significant proportion of patients.4, 5 Although initial reports failed to demonstrate an impact of HCV infection on survival, Forman et al. have recently shown that graft and patient survival after LT were significantly lower in HCV-infected patients compared to noninfected individuals.6
Several variables have been associated with a more severe HCV disease recurrence after LT, such as a high pretransplantation viral load, old donor age, the presence of significant graft steatosis, and the administration of steroid boluses.4, 7–9 It is important to state, however, that most studies are retrospective and that there is a lack of homogeneity in the definition of severe HCV recurrence after LT.
With the limited pool of cadaveric donors, living donor liver transplantation (LDLT) has become the most feasible alternative to cadaveric liver transplantation (CLT) for patients with end-stage liver disease or hepatocellular carcinoma (HCC). Nowadays, more than 3,000 LDLTs have been performed worldwide using the right hepatic lobe. Despite this high number of procedures, the enthusiasm for LDLT is tempered by the need for a highly skilled group of senior liver surgeons, the elevated surgical-related morbidity, and the rare but potential donor mortality.10–12 In addition, the applicability of LDLT is low, and only one fourth or less of the potential recipients undergo the procedure.13, 14
The scientific community assumed that outcomes after LDLT and CLT were comparable, and this assumption provided the rationale to propose LDLT in patients awaiting CLT. Consequently, cost-effectiveness analyses were run with those assumptions.15–17 Although there are studies reporting similar outcomes for both groups,18–20 others suggest that HCV disease recurrence has an earlier and more severe course in LDLT compared to CLT.21, 22 We started a LDLT program in March 2000 in patients on the waiting list for CLT. This program was expanded to patients with HCC exceeding the conventional criteria in 2001.23 The present prospective study was aimed at assessing if the outcome of HCV infection differed between CLT and LDLT. For this purpose, a cohort of 116 consecutive HCV-infected patients undergoing LT in a single institution between March 2000 and August 2003 were followed, including the performance of systematic liver biopsies.
HCV, hepatitis C virus; LT, liver transplantation; LDLT, living donor liver transplantation; CLT, cadaveric liver transplantation ; HCC, hepatocellular carcinoma; HIV, human immunodeficiency virus; HBV, hepatitis B virus; SR, severe recurrence.
Patients and Methods
HCV-infected patients who underwent LT for end-stage cirrhosis or HCC between March 2000 and August 2003 were included in the study. Exclusion criteria were (1) double kidney and liver transplantation, (2) coinfection with the human immunodeficiency virus (HIV) or hepatitis B virus (HBV), (3) recipients of a nonbeating-heart donor, (4) undetectable HCV-RNA before transplantation, and (5) survival shorter than 3 months following transplantation. Patients fulfilling these criteria entered the study, which was approved by the Investigation and Ethics Committee of the hospital.
Data from 10 LDLT recipients not infected with HCV who underwent transplantation during the same period of time were recorded following the study protocol described below, except for protocol liver biopsies.
During hospital admission, patients were managed according to a previously published schedule.3, 24 In brief, induction immunosuppression was cyclosporine A or tacrolimus, and prednisone. Ten patients were treated with tacrolimus and anti–interleukin-2 receptor antibodies. Mycophenolate mofetil was given to patients who required cyclosporine or tacrolimus dose reduction or discontinuation. Immunosuppression therapy was recorded throughout the study. Acute rejection episodes were documented by liver histology25, 26 and treated with steroid boluses if moderate or severe. After discharge, patients were visited at the outpatient clinic, monthly for the first 3 months, with complete record of clinical and analytical variables (including viral load), and every 2 months thereafter. Patients underwent protocol liver biopsies 3 months after LT and yearly thereafter, as well as when clinically indicated. Liver biopsies were evaluated by a single pathologist (M.B.) with wide experience in the histopathology of LT. Necroinflammatory activity and fibrosis stage were assessed according to Scheuer's classification.27
Definition of Severe Recurrence
Severe HCV recurrence was defined as the presence of liver cirrhosis in a liver biopsy and/or the development of clinical decompensation secondary to liver disease with portal hypertension (ascites, variceal bleeding, hepatic encephalopathy).
Prognostic Factors of Severe HCV Recurrence
A total of 29 variables potentially associated with severe HCV disease recurrence were prospectively recorded. Pretransplantation variables included recipient age and gender; Child-Pugh and model for end-stage liver disease scores; presence of HCC; HCV genotype; and pretransplantation viral load. In case of antiviral treatment before LT, the duration and doses of interferon and ribavirin were recorded. Variables related to the donor included age, graft steatosis, and the type of donor (cadaveric or living). The presence of graft steatosis was evaluated in a postreperfusion liver biopsy and assessed by a single pathologist (M.B.). Graft steatosis was classified as absent, mild (<25% of hepatocytes), moderate (25%-50% of hepatocytes), and severe (>50% of hepatocytes).
Recorded peritransplant variables were cold ischemia time and transfusion requirements.
Post-transplantation variables included the doses and levels of immunosuppressive drugs; rejection episodes; administration of corticosteroid boluses; cytomegalovirus infection or disease, antiviral treatment after LT; posttransplantation viral load; alanine aminotransferase levels; and vascular and biliary complications. Biliary complications were defined as any leak or stenosis documented by transhepatic or endoscopic retrograde cholangiography, requiring either surgery or interventional radiology/endoscopy.
The graft weight/recipient body-weight ratio was analyzed in patients undergoing LDLT. In addition, we calculated the increase in graft volume using magnetic resonance imaging volumetry measured prior to transplantation and 1 month after the procedure.28
HCV-RNA Quantification and Genotyping
Blood samples were collected before transplantation and at weeks 1, 4,12, 24, and 48 following the procedure. HCV viral load was determined using a commercially available assay (Amplicor Monitor v2.0, Roche Diagnostics, Branchburg, NJ). For negative samples serum was retested using a more sensitive qualitative test (Amplicor HCV v2.0, Roche Diagnostics). HCV genotype was determined by restriction fragment length polymorphism (3).
The primary end-point was severe HCV recurrence. Baseline characteristics of the patients are expressed as median (range). Differences between qualitative variables were assessed by the Chi-square or the Fisher exact test; differences between quantitative variables were analyzed by a nonparametric test (Mann-Whitney). Cumulative probability curves of severe HCV disease recurrence according to the Kaplan-Meier method were compared by the Cox-Mantel test. Stepwise forward Cox regression analysis of severe recurrence (SR) was used to evaluate baseline and postoperative variables found to be significant (P < .05) or near significant (P < .1) in the univariate analysis. The cutoff level chosen for quantitative variables was the median value, unless stated. Follow-up was maintained until death or retransplantation or was censored at the last visit before November 2003. The software used for statistical analysis was SPSS 10.0 (SPSS Inc., Chicago, IL).
A total of 283 liver transplantations were performed in our center from March 2000 until August 2003, 151 (53.4%) in 140 HCV-infected patients. Ten retransplantations performed within 3 months after implantation of the first graft were excluded from the analysis (Fig. 1). Twenty-four patients were excluded due to double liver-renal transplantation (3), HIV coinfection (1), nonbeating donor (2), anti–HCV positive but HCV-RNA negative (4), and survival less than 3 months after LT (cause of death not related to HCV recurrence) (14). Therefore, a total of 117 consecutive liver transplantations performed in 116 HCV-infected recipients were included in the study, 95 (81%) corresponding to CLT and 22 (19%) to LDLT.
Characteristics of the entire cohort and of patients undergoing CTL and LDLT are summarized in Table 1. Baseline features of CLT and LDLT recipients were similar, except for some variables inherently linked to LDLT, such as donor age and graft steatosis (Table 1). Regarding posttransplantation variables, biliary complications were significantly more frequent in LDLT than in CLT (73% vs. 22%, P < .01). Importantly, 15 (68%) of 22 LDLT recipients underwent double or multiple biliary anastomoses. Patients undergoing LDLT received tacrolimus more frequently than recipients of a cadaveric graft (86% vs. 41%, P < .01). Blood levels of tacrolimus, however, were similar between both groups at week 1 (9.8 ng/mL vs. 8.1 ng/mL, P = .1), month 1 (13.9 ng/mL vs. 13.3 ng/mL, P = .6), month 3 (10.9 ng/mL vs. 8.3 ng/mL, P = .2), month 6 (8.9 ng/mL vs. 8.2 ng/mL, P = .4), and month 12 (8.6 ng/mL vs. 7.2 ng/mL, P = .3), respectively. Similarly, the length and cumulative doses of steroid therapy were comparable between both groups. Acute rejection was diagnosed (2 weeks after LT; range, 1-52) in 29 (30.9%) and 9 (42.9%) CLT and LDLT recipients, respectively. Only in 2 cases acute rejection occurred after the third month (both in CLT recipients). Follow-up after transplantation was identical in both groups.
Table 1. Baseline and Posttransplantation Characteristics of Patients Undergoing LT According to the Type of Transplantation
All (n = 117)
CLT (n = 95)
LDLT (n = 22)
Abbreviations: MELD, model for end-stage liver disease; CyA, cyclosporine A; MMF, mycophenolate mofetil.
Quantitative variables expressed as median (range).
Difference between mild (<25%) vs. moderate or severe (>25%) steatosis.
In 6 cases, the indication was using expanded criteria.
Median doses of prednisone (mg) at 1, 3, 6, 9, and 12 months were 20, 15, 10, 5, and 0, respectively, in both CLT and LDLT recipients.
Three patients had minor leakages that solved after single endoscopic papillotomy (2) or percutaneous drainage (1).
After a median follow-up of 22 months (2.6-44 months), 26 (22.2%) patients developed SR, involving 17 (18%) of 95 patients receiving a CLT and 9 (41%) of 22 undergoing LDLT. The cumulative probability of being free of SR was 71% at 2 years and 67% at 3 years (Fig. 2). Diagnosis of SR relied on histology in 14 patients and on clinical decompensation in 12 (ascites in 11 patients, and variceal bleeding in 1). At the time of decompensation, 10 of the 12 patients underwent a liver hemodynamic study; hepatic venous pressure gradient was >10 mm Hg in all cases.
Predictors of Severe HCV Recurrence
Entire Cohort (N = 117).
We analyzed the prognostic value of pretransplantation and posttransplantation variables on the development of SR. LDLT was the only baseline variable predictive of SR by univariate analysis (P = .019), whereas there was a nonsignificant trend in patients with pretransplantation viral load above 5.43 log10 IU/mL (P = .08) (Table 2). Donor age, graft steatosis, recipient age, pretransplantation Child-Pugh and model for end-state liver disease scores, HCV genotype, and the administration of antiviral therapy before transplantation did not show any value as predictors of severe HCV recurrence (Table 2). Regarding posttransplantation variables, an alanine aminotransferase value 2 times the upper limit of normal (80 IU/L) 3 months following LT was predictive of SR (P = .022), whereas biliary complications (P = .09) and infection with cytomegalovirus (P = .072) showed a nonsignificant trend. None of the variables related to the type and intensity of immunosuppression influenced the severity of HCV disease recurrence (Table 2), even when the analysis was restricted to the LDLT group (data not shown). Patients who developed severe HCV recurrence received antiviral treatment more frequently than patients who did not (61.5% vs. 18.7%, P < .01). There was no rejection episode related to antiviral therapy. Follow-up after LT was comparable between patients with and without SR (Table 2).
Table 2. Prognostic Factors Associated With the Development of SR Following Liver Transplantation. Univariate Analysis
Non–severe HCV Recurrence (n = 91)
Severe HCV Recurrence (n = 26)
P Value (log rank)
Abbreviations: MELD, model for end-stage liver disease; ALT, alanine aminotransferase; CyA, cyclosporine A; MMF, mycophenolate mofetil; CMV, cytomegalovirus.
Necroinflammatory index >1 in a liver biopsy performed 3 months after LT. Available in 93 patients.
Blood levels of tacrolimus and cyclosporine were comparable between both groups at week 1, and month 1, 3, 6, and 12 after LT.
Median doses of prednisone (mg) at month 1, 3, 6, 9, and 12 were 20, 15, 10, 5, and 0 in patients without SR and 20, 10, 10, 5, and 2.5 in patients with SR (nonsignificant at all points).
In patients undergoing cadaveric liver transplantation, SR occurred in 4 (19%) of 21 with biliary complications and in 13 (18%) of 74 without biliary complications (log rank = 0.5).
Seven patients received antiviral treatment before and 9 after the diagnosis of SR.
The type of transplantation was the only independent predictor of SR (odds ratio, 2.5; 95% CI, 1.13-5.68; P = .025) (Table 3). Multivariate analysis including the variable biliary complications provided identical results. Therefore, the 2-year probability of presenting SR was significantly higher in LDLT compared to CLT (45% vs. 22%, P = .019) (Fig. 3).
Table 3. Prognostic Factors Associated With the Development of SR Following LT in the Entire Cohort (n = 117) and in Patients Who Underwent Liver Biopsy 3 Months After LT (n = 93). Cox Regression Analysis
Entire Cohort (n = 117)
Abbreviation: ALT, alanine aminotransferase.
Type of transplantation (LDLT vs. CLT)
ALT >80 IU/mL
Liver Biopsy Available 3 Months After LT (n = 93)
Type of transplantation (LDLT vs. CLT)
Lobular necroinflammation >1
Patients With Liver Biopsy Available 3 Months After LT (n = 93).
Ninety-three patients underwent a protocol liver biopsy 3 months after transplantation. In the remaining 24 patients it was not performed due to a nonprotocol biopsy obtained between the 1st and 2nd month after LT for clinical indication (13 cases), biliary complications (8 cases), and patient denial (3 cases). In this subgroup of 93 patients, the univariate analysis disclosed that LDLT (P = .004), biliary complications (P = .046), and a lobular necroinflammatory score >1 in the third-month liver biopsy (P = .006) were predictors of SR (Fig. 3). By Cox regression analysis, only the type of transplantation and the presence of necroinflammatory changes showed independent predictive value for SR (Table 3).
We performed a similar analysis using liver histology (F3-F4) as an end-point. Thirty patients (25.6%) developed stage 3 or stage 4 fibrosis during follow-up: 20 (21%) of 95 patients receiving a CLT and 10 (45.4%) of 22 undergoing LDLT (P = .02). The results of the univariate and multivariate analysis were identical to those presented above, and LDLT and a lobular necroinflammatory score >1 in the third-month liver biopsy were independently related to the development of F3-F4.
Analysis of HCV Disease Recurrence in CLT vs. LDLT
Alanine aminotransferase at 1 and 3 months after transplantation was significantly higher in LDLT compared to CLT recipients (113 IU/L vs. 50 IU/L, P < .01, and 121 IU/L vs. 65 IU/L, P = .016, respectively). Similarly, GGT values were significantly higher in LDLT than in CLT recipients both at 1 and 3 months after transplantation (420 IU/L vs. 128 IU/L, P < .01, and 647 IU/L vs. 119 IU/L, P < .01, respectively). Severe acute cholestatic hepatitis (lobular necroinflammatory score >2, and cholestasis) was confirmed by liver biopsy in 9 of 93 patients; its incidence was significantly higher in LDLT (6 of 17, or 35%) compared to CLT (3 of 76, or 4%) (P < .01). Regarding viral kinetics, HCV viral load at weeks 1, 4, 12, and 24 after transplantation was higher in LDLT compared to CLT recipients, though the differences did not reach statistical significance.
Patients who underwent LDLT received the right lobe of a living donor in all cases. The median weight of the right lobe was 759 g (range, 550-1045), and the right lobe/recipient weight ratio was 1.07 (range, 0.76-1.66). The median increase in liver volume during the first month following transplantation (measured by magnetic resonance imaging volumetry) was 67% (range, 4.5%-161%). SR occurred more frequently in patients with an increase in liver volume above the median value (6 of 11) compared to patients with an increase in liver volume below the median value (3 of 11) (P = .1).
Follow-up of Anti–HCV Negative LDLT Recipients
We analyzed the outcome of 10 HCV-RNA negative patients who underwent LDLT during the same period of time. Indication for LT was alcoholic cirrhosis (5), HBV-related cirrhosis (2), alfa-1-antitrypsin deficiency (1), Caroli's disease (1), and cryptogenic cirrhosis (1). Except for follow-up liver biopsies, anti–HCV negative patients were followed with the same protocol as HCV-infected patients. Baseline and posttransplantation characteristics of LDLT recipients with and without HCV infection were similar, except for the presence of HCC (only in 2 of 10 anti–HCV negative patients). Importantly, the incidence of biliary complications was identical (72% in HCV infected patients, 70% in noninfected patients). Median follow-up was 14 months (3.5-31 months). Liver fibrosis was absent in 1 or more follow-up liver biopsies available in 3 of the 10 anti–HCV negative patients (all 3 with biliary complications). Moreover, none of 10 individuals presented with clinical decompensation or ultrasonographic evidence of ascites during follow-up.
The outcome of HCV disease recurrence after LDLT is still controversial.18, 19, 21, 22, 29 Our data, though limited to a single center, show that LDLT is a strong and independent predictor of severe HCV disease recurrence following transplantation. Accordingly, the 2-year probability of presenting SR was significantly higher in LDLT compared to CLT (45% vs. 22%). This study has some relevant differences from previous reports. First, we designed a prospective study specifically aimed at assessing whether HCV disease recurrence was different between both types of transplantation. Although the numbers of patients (116) and events (26) are small, the strength of the differences after a median follow-up of 22 months prevents us to expand the series with additional cases. More importantly, severe recurrence was defined by the presence of cirrhosis in a follow-up liver biopsy or by the occurrence of clinical decompensation. This definition allows an unbiased classification of patients and represents a relevant event in the natural history of HCV disease recurrence.30 Finally, patients were recruited in a single center, and the same standard of care was established. Additionally, all relevant variables that might influence HCV disease recurrence were included in the analysis, and, except for factors inherently related to living donation (and the type of calcineurin inhibitor), patients receiving the graft of a living donor or a cadaveric donor were comparable.
The mechanisms that might explain the more aggressive course of HCV recurrence after LDLT are unknown. Theoretically, there are variables specifically linked to LDLT that might prevent from severe HCV disease recurrence, such as the young donor age, the lack of significant steatosis of the graft, and the short ischemia time during surgery.8, 31 On the contrary, other variables might affect negatively HCV disease recurrence, such as an increased HLA donor-recipient matching, the type of immunosuppression, a high incidence of biliary complications following transplantation, and liver regeneration.21, 32, 33
Our initial hypothesis was that either biliary complications or liver regeneration (or both) would accelerate liver fibrosis in patients undergoing LDLT. Biliary complications are frequent in the latter group, and it is well known that persistent cholestasis induces fibrogenesis. Despite the lack of homogeneity in the definition of biliary complications among reported series, the incidence of biliary leaks or stenosis in our cohort was very high. Three reasons might explain this high incidence: first, the prospective nature of data collection (including even minor leakages); second, the high frequency of double and multiple biliary anastomoses in our LDLT series34; and third, the learning curve.35 Despite this center-specific issue, neither the multivariate analysis nor the follow-up of a small cohort of anti–HCV negative LDLT recipients support an independent predictive value of biliary complications on the severity of HCV disease recurrence. However, a synergistic effect of persistent cholestasis on HCV-infected grafts cannot be excluded.
Biochemical markers of hepatitis increased earlier and reached significantly higher levels in patients undergoing LDLT compared to CLT, strongly suggesting that hepatitis C recurrence is a distinct process in those patients. In fact, the occurrence of cholestatic hepatitis was significantly more frequent in LDLT compared to CLT. One of the key differences between LDLT and CLT is liver regeneration.36In vitro, HCV internal ribosome entry site activity and replication were found to be higher in actively dividing cells, and it is possible that viral translation may be enhanced by factors that stimulate the regeneration of hepatocytes.37, 38 Moreover, there are experimental data suggesting that liver regeneration induces LDL receptor expression,39 which might facilitate HCV entrance into the hepatocytes.40 The absence of significant differences in viral load between LDLT and CLT at the analyzed time points does not exclude increased viral production in regenerating cells, as viral load also depends on the number of cells producing viral particles and on viral clearance mechanisms.
Regretfully, this study presents several caveats. First, the limited number of patients who underwent LDLT did not allow a thorough analysis of the variables that predict SR among them. Our data suggest that patients who experienced a greater increase in liver volume during the first weeks following LDLT had a higher probability of developing SR afterward. Though the increase in liver size cannot be use as a direct marker of the degree of regeneration, this increase might support a negative influence of liver regeneration on HCV disease recurrence. These results should be further confirmed in extensive series. Second, the relatively short follow-up of this cohort did not allow a consistent analysis of graft and patient survival. However, it is well established that around 40% of patients with compensated HCV graft cirrhosis will develop clinical decompensation within the first year following the diagnosis. Once decompensation occurs, survival is lower than 50% at 1 year,30 whereas retransplantation leads to disappointing results.41 The expected decrease in graft or even patient survival derived from this scenario might make LDLT a non-cost-effective approach in HCV-infected patients. Possible alternatives would be to restrict LDLT to very long waiting times and/or high dropout rate settings,17 limiting LDLT to non-HCV-infected patients (at least during the learning curve), or to treat HCV infection before LDLT.42–44 Implementation of any of these strategies would require confirmation of our results in other prospective series.
In summary, our data indicate that LDLT is a strong predictor of severe HCV disease recurrence after transplantation. Although the data need to be validated, the more aggressive course of HCV infection in LDLT compared to cadaveric transplantation should be considered in LDLT programs, since it may ultimately compromise graft and patient survival.