Post–liver transplant survival in hepatitis C patients is improving over time

Authors

  • Jacqueline G. O'Leary,

    Corresponding author
    1. Department of Medicine, Baylor University Medical Center, 3500 Gaston Avenue, Dallas, TX
    2. Baylor Regional Transplant Institute, Dallas, TX
    • Baylor University Medical Center, 4th Floor Roberts, 3500 Gaston Avenue, Dallas, TX 75246
    Search for more papers by this author
    • Telephone: 214-820-8500; FAX: 214-820-8168

  • Henry Randall,

    1. Department of Surgery, Baylor University Medical Center, 3500 Gaston Avenue, Dallas, TX
    2. Baylor Regional Transplant Institute, Dallas, TX
    Search for more papers by this author
  • Nicholas Onaca,

    1. Department of Surgery, Baylor University Medical Center, 3500 Gaston Avenue, Dallas, TX
    2. Baylor Regional Transplant Institute, Dallas, TX
    Search for more papers by this author
  • Linda Jennings,

    1. Baylor Regional Transplant Institute, Dallas, TX
    Search for more papers by this author
  • Göran B. Klintmalm,

    1. Department of Surgery, Baylor University Medical Center, 3500 Gaston Avenue, Dallas, TX
    2. Baylor Regional Transplant Institute, Dallas, TX
    Search for more papers by this author
  • Gary L. Davis

    1. Department of Medicine, Baylor University Medical Center, 3500 Gaston Avenue, Dallas, TX
    2. Baylor Regional Transplant Institute, Dallas, TX
    Search for more papers by this author

Abstract

Outcomes after orthotopic liver transplantation for chronic hepatitis C have been reported to be worsening over the last 2 decades. We analyzed our center's experience over 15 years to identify trends in post–orthotopic liver transplantation survival in patients with and without hepatitis C virus infection. Patient survival and graft survival among adult primary orthotopic liver transplantation recipients who survived more than 90 days from January 1991 to June 2006 at the Baylor Regional Transplant Institute (n = 1901) were evaluated by Kaplan-Meier analysis. Those with or without hepatitis C virus infection were analyzed by era: era 1, 1991-1994 (n = 473); era 2, 1995-1998 (n = 421); era 3, 1999-2002 (n = 498); and era 4, 2003-2006 (n = 512). Differences in eras with disparate survivals were assessed by univariate and multivariable analysis. Overall, patient survival and graft survival were significantly lower among hepatitis C virus infection recipients compared to those without hepatitis C virus infection (P < 0.001). This difference was dependent on the era of transplantation, with progressive improvement in hepatitis C virus patient (P < 0.001) and graft (P < 0.001) survival in sequential eras. Several factors accounted for this improvement, notably better selection of hepatocellular carcinoma patients and fewer late cytomegalovirus infections. Improvement occurred despite an increase in the ages of both donors and recipients. In conclusion, posttransplant survival after orthotopic liver transplantation for chronic hepatitis C has improved significantly over the last 15 years despite demographic changes in patients and grafts that have been previously shown to impair survival. A major reason for this improvement is better selection of patients with concurrent hepatocellular carcinoma and fewer late cytomegalovirus infections, although other factors may play a role as well. Liver Transpl 15:360–368, 2009. © 2009 AASLD.

Hepatitis C virus (HCV)–induced end-stage liver disease is the most common indication for orthotopic liver transplantation (OLT).1 Nonetheless, despite improvements in surgical technique, immunosuppression, and antiviral therapy, outcomes in HCV-infected patients have been reported to be worsening in the last 2 decades.2 This was first observed by Berenguer and colleagues,2 who evaluated serial liver biopsies from 284 HCV-infected transplant recipients in Spain and the United States between 1989 and 1996. They discovered that the rate of fibrosis progression was more rapid in patients transplanted in recent years. Fibrosis progression was used as their endpoint, and survival analysis was not reported. In contrast, Thuluvath and colleagues3 evaluated the survival of all HCV-infected adults who underwent OLT in the United States. After the exclusion of patients with incomplete data, acute liver failure, a diagnosis of non-A/non-B hepatitis, retransplant, and multiorgan transplant, no difference in patient or graft survival was seen over time among 28,193 patients.

There are several factors that have a negative impact on survival in HCV-infected patients following OLT.4 Recipient factors include older age, female gender, nonwhite race, and severity of illness at the time of transplant.5-13 Viral factors include pretransplant viral load, viral load during the first few weeks post-transplant, and coinfection with cytomegalovirus (CMV).14-18 In addition, older donor age, shorter time to HCV recurrence, and treatment of rejection with steroids or lymphocyte-depleting antibodies have been reported to decrease survival after OLT in HCV-infected recipients.19-25 However, none of these well-known factors can in isolation explain why outcomes in recent years might be worse than in previous years.

Our aims were to determine whether outcomes were really changing over time in patients transplanted for HCV and to examine the factors associated with such changes.

Abbreviations

CMV, cytomegalovirus; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; HR, hazard ratio; MELD, Model for End-Stage Liver Disease; MMF, mycophenolate mofetil; NA, not applicable; NS, not significant; OLT, orthotopic liver transplantation.

PATIENTS AND METHODS

Institutional review board approval was granted prior to the initiation of this retrospective study. We analyzed both patient and graft survival for all adult recipients of primary OLT at the Baylor Regional Transplant Institute who survived longer than 90 days from January 1991 to June 2006 (n = 1901). Because most patients that die within 90 days of transplant die from surgical and bacterial complications, they were excluded from this analysis. Patients transplanted since anti-HCV serologic testing became available were divided into approximate quartiles based on the year of OLT. They are hereafter referred to as follows: era 1, 1991-1994 (n = 473); era 2, 1995-1998 (n = 421); era 3, 1999-2002 (n = 498); and era 4, 2003-2006 (n = 512).

The Baylor Regional Transplant Institute maintains a prospective database called the Liver Transplant Research Database System, which contains data on all OLT patients since the program's inception. A patient's cause of death is obtained from death certificates and hospital records and is reviewed by a physician before it is entered into the database. Graft survival has been defined as absence of death or retransplantation.

CMV prophylaxis is undertaken with intravenous ganciclovir for the first 7 to 10 days post-transplant. The length of treatment is determined by the length of stay. After discharge, no prophylaxis or laboratory monitoring is undertaken. When signs or symptoms of infection arise, patients are tested for CMV antigenemia. All patients with CMV antigenemia receive intravenous ganciclovir adjusted for renal function until the CMV polymerase chain reaction becomes negative. They are then switched to oral valganciclovir for 2 additional weeks of therapy.

Statistical Methods

Patient characteristics were compared with Wilcoxon rank sum, and median values are reported in Tables 1 and 2. Graft survival and patient survival were evaluated by Kaplan-Meier analysis. In addition, Kaplan-Meier analysis was used to perform univariate analysis of predictors of survival post-OLT. All the factors were included in a stepwise Cox proportional hazards multivariable analysis. Only factors that reached statistical significance (P < 0.05) were reported as part of the multivariable model. SAS 9.1 was used for all statistical analyses.

Table 1. Patient Characteristics from 1991 to 2006 for All Post–Orthotopic Liver Transplantation Patients Displayed by Indication for Transplant
  All Patients (n = 1901)P Value
HCV (n = 777)Other (n = 1124)
  1. NOTE: Results presented as medians or percentages, as appropriate.

  2. Abbreviations: CMV, cytomegalovirus; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; MELD, Model for End-Stage Liver Disease; MMF, mycophenolate mofetil.

Age 49510.005
MELD at transplant (calculated) 15170.001
Cold ischemia (hours) 8.18.60.002
Male sex 69%52%<0.001
HCC 21%8%<0.001
Donor age 3835.50.047
RaceWhite76%75%0.5
 Black7%8% 
 Other17%17% 
Prior alcohol use 25%23%0.3
Calcineurin inhibitorCyclosporine52%46%0.04
 Tacrolimus46%52% 
AntimetaboliteMMF44%30%<0.001
 Azathioprine28%27% 
Rejection 57%58%0.6
Steroid-resistant rejection 14%20%<0.001
CMV in <3 months 18%18%0.9
CMV in >3 months 8%7%0.6
Reoperation <3 months 26%23%0.3
Biliary surgery 6%6%0.8
Table 2. Patient Characteristics from 1991 to 2006 Displayed for HCV-Infected Patients Post–Orthotopic Liver Transplantation by Era
  HCV Patients [n = 777 (41% of All Patients)]P Value
Era 1: 1991-1994 (n = 139)Era 2: 1995-1998 (n = 162)Era 3: 1999-2002 (n = 231)Era 4: 2002-2006 (n = 245)
  1. NOTE: Results presented as medians or percentages, as appropriate.

  2. Abbreviations: CMV, cytomegalovirus; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; MELD, Model for End-Stage Liver Disease; MMF, mycophenolate mofetil.

Age 474749500.003
MELD (calculated)All HCV patients161415150.18
 HCV and HCC17.51514120.002
 HCV without HCC151416160.002
Cold ischemia (hours) 10.58.87.27.5<0.001
Male sex 68%60%69%75%0.015
Donor age 333738420.005
RaceWhite80%80%72%73% 
 Black4%5%7%10% 
 Other16%15%21%17%0.12
Prior alcohol use 19%19%30%28%0.02
Calcineurin inhibitorCyclosporine91%86%30%27% 
 Tacrolimus9%13%65%71%<0.001
AntimetaboliteMMF3%4%68%72% 
 Azathioprine59%65%11%1%<0.001
Rejection 73%59%63%42%<0.001
Steroids for rejection (at least 1 g) 70%56%58%34%<0.001
Steroid-resistant rejection 31%17%12%4%<0.001
CMV in <3 months 22%15%18%18%0.58
CMV in >3 months 12%8%9%5%0.07
Reoperation <3 months 27%30%20%28%0.14
Biliary surgery 5%4%3%10%0.01
Number of patients with HCC 18 (13%)17 (11%)55 (24%)70 (29%)<0.001
Median size of largest HCC (cm) 4.43.52.22.5<0.001
Median number of HCCs 12110.003
Median waiting time (days) 10821128830<0.001

RESULTS

Table 1 compares the characteristics of patients transplanted with HCV infection and those transplanted without HCV infection between January 1991 and June 2006 who survived greater than 90 days post-transplant. Patients transplanted for HCV were younger, more likely to be male, and more likely to have hepatocellular carcinoma (HCC) than those transplanted for indications other than HCV. In addition, HCV patients had shorter cold ischemia times, older donors, and a lower calculated Model for End-Stage Liver Disease (MELD) score at transplant. Mycophenolate mofetil and cyclosporine were used more frequently in patients with HCV. Rejection was seen in similar numbers in the 2 groups; however, steroid-resistant rejection was less common in HCV-infected patients. Recipient race and prior alcohol use were similar in the 2 groups. Patients received intravenous ganciclovir for CMV prophylaxis only during their initial hospital stay or during high-dose steroid or antibody treatment with OKT3 or anti-thymocyte globulin for rejection. Therefore, CMV infection was seen frequently but in a similar percentage of both groups. Although the number of HCC patients has increased over time, the size of lesions, number of lesions, and waiting time all decreased in successive eras.

Both patient survival (P < 0.001) and graft survival (P < 0.001) were significantly lower among HCV-infected recipients than uninfected recipients (Fig. 1). As a large percentage of patients infected with HCV were transplanted for HCC, an evaluation was also performed comparing HCV-infected patients without HCC to both all uninfected patients and uninfected patients without HCC. The results remained unchanged for both analyses (data not shown).

Figure 1.

(A) Patient and (B) graft survival post–liver transplant in HCV–infected recipients versus all other indications for liver transplant (1991-2006). Patients were excluded if they died within 90 days of transplant. HCV–infected patients post–liver transplant show a decrease in patient and graft survival in comparison with non–HCV–infected patients. Abbreviation: HCV, hepatitis C virus.

The demographics of the HCV-infected OLT recipients in the 4 eras are shown in Table 2. Over time, the proportion of patients transplanted with HCC, recipient age, and donor age have all increased. In addition, the percentage of men, those with previous alcohol consumption, and patients with concurrent HCC have all increased. Cold ischemia time, rejection, and steroid-resistant rejection have all decreased over time. Over the last 15 years, there has been a change in immunosuppression from cyclosporine and azathioprine to tacrolimus and mycophenolate mofetil. Recipient race, CMV infection, and reoperation rates within the first 3 months after OLT remain unchanged. Although the MELD of all HCV-infected patients has not appreciably changed over time, the MELD of HCV patients with HCC has decreased, whereas the MELD of HCV patients without HCC has increased. The median size of HCCs at transplant and the median number of lesions have all decreased over time.

Patient survival and graft survival in HCV-infected persons in sequential eras are shown in Fig. 2A,B, respectively. Patient (P < 0.001) and graft (P < 0.001) survival improved significantly over time. When Bonferroni correction was used to make pairwise comparisons between the eras, patient survival in era 3 (1999-2002) was significantly better than patient survival in era 1 (1991-1994; P < 0.001) and era 2 (1995-1998; P < 0.001). Similarly, graft survival was significantly improved from era 1 to era 3 (P < 0.001) and from era 2 to era 3 (P < 0.001). Comparisons with era 4 were limited by shorter follow-up and therefore were not performed.

Figure 2.

(A) Patient and (B) graft survival post–liver transplant by era in recipients with hepatitis C virus infection. Patients were excluded if they died within 90 days of transplant. Patient survival and graft survival have improved over time. Abbreviation: NA, not applicable.

There was no statistically significant difference in patient survival over time for those transplanted for indications other than HCV (P = 0.102), although graft survival (P = 0.028) improved (Fig. 3). When uninfected patients without HCC were evaluated, patient survival was also unchanged over time (data not shown).

Figure 3.

(A) Patient and (B) graft survival post–liver transplant by era in recipients without hepatitis C virus infection. Patients were excluded if they died within 90 days of transplant. Patient survival is unchanged, but graft survival has improved over time. Abbreviation: NA, not applicable.

As HCC is a common indication for transplant among HCV-infected patients and the tumor size limitations have changed over time, we compared the survival of patients transplanted with HCC between the eras (Fig. 4). There was a striking increase in patient survival from era 1 or 2 to era 3 or 4 (P < 0.001). This improvement was temporally associated with publication of the Mazzaferro study in 1996 and its implementation as criteria for transplantation.26, 27 No change was seen in the survival of patients transplanted with HCC without HCV over time; however, this is likely due to the small number of patients transplanted for this indication (shown later in Table 4). Patient survival over the eras was not different in HCV-infected recipients without HCC (Fig. 5A; P = 0.054). In a multivariable analysis, a more recent era of transplant still predicted improved survival. Despite this change in patient survival, graft survival remained significantly better (Fig. 5B; P < 0.001).

Figure 4.

Patient survival post–liver transplant by era in recipients with hepatocellular carcinoma (A) with hepatitis C virus infection and (B) without hepatitis C virus infection. Patients were excluded if they died within 90 days of transplant. Abbreviation: NA, not applicable.

Figure 5.

(A) Patient and (B) graft survival post–liver transplant by era in recipients with hepatitis C virus infection without hepatocellular carcinoma. Patients were excluded if they died within 90 days of transplant. Patient survival is unchanged, but graft survival has improved over time. Abbreviation: NA, not applicable.

Causes of death were compared between HCV-infected patients and uninfected patients (Table 3). Recurrent disease was the most common cause of death in HCV-infected patients. Among uninfected patients, infection, cancer, and other causes were more likely to cause death. When causes of death among HCV-infected individuals were compared over the eras, death from HCC recurrence and rejection decreased dramatically.

Table 3. Causes of Death for HCV Patients Versus Non-HCV Patients, Causes of Death for HCV Patients by Transplant Era, and HCC Recurrence
 All Deaths (n = 620)P Value
HCV (n = 273)Other (n = 347)
  1. NOTE: Four patients remained alive despite HCC recurrence, and 1 patient died of HCV recurrence with incidental HCC recurrence.

  2. Abbreviations: HCC, hepatocellular carcinoma; HCV, hepatitis C virus.

HCC recurrence22130.03
Rejection10140.8
Recurrent disease7733<0.001
Infection32590.06
Cancer21520.004
Other18390.04
Cardiac/pulmonary19380.09
Graft failure12161
Unknown cause6283 
 All HCV Deaths (n = 273)P Value
Era 1: 1991-1994 (n = 97)Era 2: 1995-1998 (n = 85)Era 3: 1999-2002 (n = 50)Era 4: 2002-2006 (n = 41)
HCC recurrence88330.9
Rejection82000.01
Recurrent disease212916110.15
Infection129830.7
Cancer85530.8
Other66330.9
Cardiac/pulmonary113320.2
Graft failure33240.3
Unknown cause20201012 
 HCV with HCCP Value
Era 1: 1991-1994Era 2: 1995-1998Era 3: 1999-2002Era 4: 2002-2006
No recurrence984964 
Recurrence9966<0.001

Factors that appeared to favor improved patient survival among HCV-infected recipients in era 3 are shown in Table 4 and include less CMV infection after the first 3 months, less advanced HCC (post–Milan criteria), and lower MELD. Era 3 remained associated with better survival despite more frequent need for biliary surgery, greater donor age, and higher recipient age. Despite a significant decrease in the rate of rejection and need for steroid boluses in more recent eras, this was not found to predict a change in survival in multivariable analysis.

Table 4. Stepwise Cox Proportional Hazards Multivariable Analysis With Either All HCV-Infected Patients or HCV-Infected Patients Without HCC
Factors Associated with Death in All HCV-Infected PatientsAll HCV (Multivariate)HCV without HCC (Multivariate)
P ValueHRP ValueHR
  1. Abbreviations: CMV, cytomegalovirus; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; HR, hazard ratio; MELD, Model for End-Stage Liver Disease; NA, not applicable; NS, not significant.

CMV infection 3 months post-transplant<0.0012.3<0.0012.8
Era 3<0.0010.480.0030.57
HCC0.0051.6NA 
Donor age > 50 years0.0051.50.0091.5
Recipient age > 60 years0.0121.50.021.6
Higher MELD at transplant (continuous)0.0151.02NS 
Biliary surgery0.041.6NS 

In order to assess whether the observed temporal changes in survival among HCC patients had an impact on overall survival, we repeated the stepwise multivariable Cox proportional hazards modeling using only HCV-infected patients who did not have HCC (Table 4). CMV infection after 3 months of transplant, an earlier era of transplant, donor age greater than 50 years, and recipient age greater than 60 years all remained predictors of decreased patient survival.

DISCUSSION

Because HCV is the leading indication for liver transplantation in the United States, it is essential to optimize survival outcomes post-OLT. Berenguer and colleagues2 previously reported that fibrosis progression in HCV-infected transplant recipients has been occurring more rapidly in recent years than previously. In a more recent study, they demonstrated that fibrosis progression was slower after their program implemented a dual initial immunosuppression regimen (prednisone and a calcineurin inhibitor) with slow tapering of prednisone.28 Although fibrosis progression is often used as a surrogate predictor of survival, actual long-term patient survival and graft survival are most relevant. We evaluated survival post-OLT in a large number of HCV-infected patients observed over a 15-year period.

As previously reported by others, we found that patients transplanted for complications of chronic HCV infection have lower survival than patients transplanted for other indications.3, 4, 6, 8, 14, 29, 30 Our patient and graft survival curves diverge only after the first year of transplant, and the divergence increases over time as a result of HCV recurrence and graft loss.30 Twenty to forty percent of patients with recurrent HCV infection post-OLT develop cirrhosis in 5 years.4 Decompensation occurs in 40% of these patients with cirrhosis within 1 year, and half of these patients die in the following year.2, 4 This inferior survival among HCV-infected patients remains with or without the inclusion of HCC patients (data not shown) and therefore is almost certainly secondary to cirrhosis from HCV recurrence.3

Patient survival post-OLT is affected by HCV recurrence, but concurrent HCC also plays a role, albeit a smaller one in the recent 2 eras.7, 29 This improvement in survival in patients with HCC temporally corresponds to the implementation of the Milan criteria for HCC transplantation.26, 27 Prior to the Milan criteria, many patients with large tumors and even patients with extrahepatic disease were transplanted, and this led to 5-year survival rates lower than 35%.31-35 However, Starzl and colleagues31 realized in 1985 that patients with small tumors incidentally discovered by the pathologist had excellent long-term survival. With improvement in imaging, Mazzaferro et al.26 were able to define HCC criteria using the size and number of tumors that led to excellent long-term survival post-transplant. Indeed, HCV-infected patients with HCC can now expect a 5-year survival rate of approximately 80%, which is equivalent to that for other indications. Although it is most likely that the improvement in outcomes is related to transplantation of less advanced tumors, ablation or chemoembolization of tumors pre-transplant and a shorter time from listing to transplant after the implementation of the MELD exception for HCC could also have had some impact on outcomes. Regardless, our observation of improved HCV survival over time is partly attributable to improved survival among HCV-infected patients with HCC transplanted before and after implementation of these criteria. Although Kaplan-Meier survival analysis of HCV-infected patients without HCC appears to have changed little over the eras, the recent eras had significantly better survival in multivariable analysis. This improvement may be attributable to several factors. Late CMV infections have decreased in the more recent eras. In addition, the rate of rejection, steroid boluses, and death from rejection have all decreased in recent eras. However, the improvement could also be attributable to a greater use of antiviral therapy (not evaluated in this study).36

Thuluvath and colleagues3 previously found no temporal change in survival among US patients transplanted for HCV before 2001. The lack of change in these studies may be surprising given the changes in the demographics of donors and recipients in recent years that are well known to have a negative impact on survival. Specifically, donor and recipient ages have increased. Unlike Thuluvath et al., we were able to examine the impact of some of these factors on survival. CMV infection occurring greater than 3 months post-OLT, recipient age greater than 60 years, donor age greater than 50 years, and post-OLT complications requiring biliary surgery all increased the probability of death after liver transplant. Belli and colleagues7 in Europe also found improved post-OLT survival of HCV-infected patients in recent years. Although they were not able to identify the reason that survival had improved, multivariable analysis revealed that donor age, recurrent HCV with progression to an Ishak fibrosis score greater than 3, and earlier year of transplant were associated with post-OLT mortality. They did not analyze the impact of HCC on survival over time, although its effect on survival was significant in their univariate analysis.

Although successful antiviral therapy has a positive impact on the survival of HCV-infected patients, in our cohort, few patients were treated before era 4, and given the low sustained virologic response seen in posttransplant patients, it is difficult to determine if it altered our patients' survival. Regrettably, we do not have reliable data on antiviral therapy and response in our cohort.

In summary, survival of HCV-infected patients post-OLT has improved over the last 15 years, despite older donor and recipient ages, which are well known to have a negative impact on transplant outcome. Reasons for the improvement in survival in recent years are better selection of HCV-infected patients with HCC for transplant and likely fewer late CMV infections.

Ancillary