Trends in post–liver transplant survival in patients with hepatitis C between 1991 and 2001 in the united states

Authors

  • Paul J. Thuluvath,

    Corresponding author
    1. Section of Hepatology, Department of Medicine, The Johns Hopkins University School of Medicine, Baltimore, MD
    • Section of Hepatology, The Johns Hopkins University Hospital, 1830 East Monument Street, Suite 428, Baltimore, MD 21205
    Search for more papers by this author
    • Telephone: 410-614-5389; FAX: 410-614-9612

  • Karen L. Krok,

    1. Section of Hepatology, Department of Medicine, The Johns Hopkins University School of Medicine, Baltimore, MD
    Search for more papers by this author
  • Dorry L. Segev,

    1. Section of Hepatology, Department of Medicine, The Johns Hopkins University School of Medicine, Baltimore, MD
    Search for more papers by this author
  • Hwan Y. Yoo

    1. Section of Hepatology, Department of Medicine, The Johns Hopkins University School of Medicine, Baltimore, MD
    Search for more papers by this author

  • See Editorial on page 641

Abstract

It has been suggested that the post–liver transplantation (LT) survival rate of patients with hepatitis C virus infection (HCV) has declined in recent years. To compare the outcome of LT in patients with HCV at various time intervals between 1991 and 2001, we used United Network for Organ Sharing data to compare the post-LT survival of adult patients (age >18 years) with HCV with those without HCV. Of the 37,101 patients who underwent LT during the study period, 28,193 patients (HCV 7,459 and 20,734 non-HCV) were eligible for the study. On the basis of the time of transplantation, patients were divided into 3 groups: 1991-1993 (period 1), 1994-1997 (period 2), and 1998-2001 (period 3). The patient and graft survival rates were adjusted for other known confounding variables that influenced outcomes. The 3-year patient survival rate was lower in HCV patients compared with non-HCV recipients (78.5% vs. 81.4%, hazard ratio 1.14, 95% confidence interval 1.05-1.23, P = 0.001). The graft (72.8%, 71.0%, and 69.8%) and patient (77.4%, 79.6%, and 78.5%) survival of HCV patients remained unchanged during study periods 1-3, respectively. However, the graft and patient survival rates of non-HCV recipients improved markedly during study periods 2 and 3 compared with period 1. The graft and patient survival has remained unchanged between 1991 and 2001 in HCV recipients, but during the same period, there was a great improvement in survival among non-HCV recipients. Liver Transpl 13:719–724, 2007. © 2007 AASLD.

Hepatitis C virus (HCV)-induced liver disease is the most common indication for liver transplantation (LT) in the United States and elsewhere. Previous studies have shown that recipients with HCV have 10-15% lower 5-year graft and patient survival rates compared with non-HCV controls.1, 2 This difference in survival has become more apparent among those who were transplanted more recently, whereas there was an overall trend in improved survival rates in non-HCV recipients, including those transplanted for hepatocellular carcinoma (HCC).2–4 Proposed factors that predict poor survival in HCV patients include viral factors such as high pretransplant virus titers and possibly genotype 1b, and host factors including early and severe recurrence of HCV, >2 rejection episodes that require steroid boluses, use of monoclonal antibody (OKT3) or antithymocyte globulin (ATG), cytomegalovirus, and human immunodeficiency virus infection.2, 4–21 In addition, older deceased donors, living donors (LDLT), and severity of graft steatosis also have a negative impact on the natural history of recurrent of HCV.2, 3, 13–15 Finally, there is preliminary evidence to suggest that immunologic factors such as human leukocyte antigen class II (B14, DRB1*04) and donor tumor necrosis factor alpha promoter genotype may also influence the outcome of recurrent HCV.22, 23

Despite the emerging knowledge of viral, host, donor, and immunologic factors that lead to more severe recurrence, there is no satisfactory explanation for the reported decline in survival of HCV patients transplanted more recently. In this study, we wanted to examine whether previous observations applied to HCV patients transplanted in the United States. In order to have adequate sample size to adjust for confounding factors, we used the United Network for Organ Sharing (UNOS) database to compare the graft and patient survival of HCV and non-HCV patients who were transplanted at various time intervals between 1991 and 2001. Our objective was to determine whether there has been a decline in survival among HCV patients transplanted recently, as previously suggested, after adjusting for other known confounding factors.

Abbreviations

HCV, hepatitis C virus infection; LT, liver transplantation; HCC, hepatocellular carcinoma; UNOS, United Network for Organ Sharing; BMI, body mass index; LDLT, live donor liver transplant; DDLT, deceased donor liver transplant.

PATIENTS AND METHODS

We studied the UNOS data on all adult LT that were performed in the United States between January 1991 and October 2001 (Fig. 1). Out of 37,101 patients, we excluded 8,908 patients (18.6% HCV and 23.3% non-HCV) for a variety of reasons including multiple organ transplantation, retransplantation, inadequate survival information, fulminant hepatic failure, or non-A/non-B viral hepatitis. Of the remaining 28,193 patients, 7,459 patients had HCV (217 HCV patients with HCC were not included in this group), and 20,734 patients had liver diseases unrelated to HCV (non-HCV). A total of 5,708 (72%) HCV to 16,116 (82%) non-HCV patients had adequate data for inclusion in the multivariate analysis to determine the impact of confounding factors on the survival. The study time period was arbitrarily divided into 3 time periods to have a minimum of 3-year posttransplantation follow-up. In period 1 (1991-1993) there were 1,176 HCV and 4,806 non-HCV patients, in period 2 (1994-1997) there were 2,679 HCV and 8,419 non-HCV patients, and in period 3 (1998-2001) there were 3,604 HCV and 7,509 non-HCV patients.

Figure 1.

Number of patients in each group and exclusions for each time period (1991-1993, 1994-1997, and 1998-2001). FHF, fulminant hepatic failure; NANB, non-A and non-B liver disease; UNOS, United Network for Organ Sharing; HCV, hepatitis C virus.

We evaluated the recipient's age, gender, race, body mass index (BMI) at the time of LT (after excluding outliers with BMI <15 or >55 for probable inaccurate data), serum creatinine (mg/dL), ABO blood type, cause of liver disease, UNOS listing status, cold ischemia time, donor age, and patient and graft survival status up to 5 years after LT. UNOS listing status was recorded as status 1, status 2a, and others (before Model for End-Stage Liver Disease scoring). ABO matching was divided into 3 groups: matched, compatible, or mismatched.

To compare the survival rates of patients with LDLT with those who had deceased donor liver transplants (DDLT), a case-control study analysis was performed because of the low number of LDLT performed during this period. For each LDLT, we randomly selected DDLT controls matched for age, gender, race, diagnosis, and year transplanted. There were 207 LDLT and 480 DDLT for patients with HCV.

For statistical analysis, the χ2 test and the Student t test or analysis of variance were used for categorical and continuous variables, respectively. Patient and graft survival was determined by Kaplan-Meier survival analysis and Cox multivariate regression survival analysis. For multivariate analysis, we included recipient factors such as age, gender, race, BMI, and serum creatinine, and donor factors such as cold ischemia time, age, and donor liver histology findings (severity of steatosis, portal infiltration, and fibrosis). Differences in survival by Kaplan-Meier were analyzed by log-rank test. For all analyses, a 2-tailed P value of 0.05 or less was considered statistically significant. The statistical analysis was performed by SPSS software, version 11.5.0 (SPSS, Chicago, IL).

RESULTS

The proportion of adult LTs for HCV increased over the study period and reached >50% in the last 2 years of the study period. Only 16.4% of LT were performed for HCV in 1991, and this increased 54.7% in 2001. The characteristics of patients with HCV at different periods are shown in Table 1. There was a trend for more African Americans and fewer women toward the later part of the study. Cold ischemia time appeared to decrease, and the donor age increased during this same time period. There was an increase in the average donor age over the study period with similar increases in both groups. There was a statistically significant increase in the average donor age (period 1: 32.6 ± 15.1 years; period 2: 35.5 ± 16.7 years; period 3: 38.8 ± 17.4 years; P < 0.01) over the study period for the entire group.

Table 1. Demographics of Liver Transplant Recipients With Hepatitis C Virus–Induced Liver Disease
Characteristic1991-1993 (n = 1,176)1994-1997 (n = 2,679)1998-2001 (n = 3,604)P value
  1. Abbreviations: UNOS, United Network for Organ Sharing; BMI, body mass index.

Age (yr)50.3 ± 10.049.9 ± 9.250.6 ± 8.2 
Gender (female %)37.9%38.1%31.3%<0.01
Race   <0.01
 White77.1%75.8%75.2% 
 Black5.1%6.1%7.0% 
 Asian3.8%3.1%2.5% 
 Hispanic11.8%12.4%12.8% 
UNOS listing   <0.01
 18.4%10.0%1.1% 
 2A21.8% 
BMI (kg/m2)27.7 ± 5.327.9 ± 5.528.4 ± 5.5 
Creatinine (mg/dL)1.23 ± 0.861.20 ± 0.951.20 ± 0.88 
Cold ischemia time (hours)10.8 ± 4.09.3 ± 3.38.2 ± 3.2<0.01
ABO matching    
 Match92.5%93.8%93.3% 
 Compatible6.4%5.1%5.6% 
 Mismatch0.8%1.1%1.0% 
Donor age (yr)32.6 ± 15.135.5 ± 16.738.8 ± 17.4<0.01

During the study period (1991-2001), the 3-year survival rate of HCV patients was lower than non-HCV patients (78.5% vs. 81.7%, P = 0.001, hazard ratio [HR] 1.14, 95% confidence interval [95% CI] 1.05-1.23) (Fig. 2A). Figure 2B shows the survival, after adjusting for confounding factors by Cox multivariate regression analysis, for the entire cohort (including HCV and non-HCV patients). Three-year patient survival was much better in the later part (periods 2 and 3) of the study. The 3-year survival was lower for period 1 (78.3%, HR 1.2, 95% CI 1.09-1.3, P <0.001) when compared with survival during period 3 (reference period); there was no difference between period 2 (81.8%, HR 0.98, 95% CI 0.89-1.07, P = 0.6) and period 3.

Figure 2.

(A) Patient and graft survival rate after liver transplantation for patients with or without hepatitis C virus infection (HCV), after adjusting for confounders by Cox multivariate regression analysis. (B) Adjusted patient survival rate for the entire group for 3 time periods: 1991-1993 (period 1), 1994-1997 (period 2), and 1998-2001 (period 3).

Kaplan-Meier survival analysis showed that there was no improvement in survival during the study period for patients with HCV (3-year survival for periods 1, 2, and 3 was 76.3%, 78.4%, and 75.3%, respectively, P = 0.12) (Fig. 3A); the graft survival also showed similar results (72.8%, 71.0%, and 69.8%, respectively). The 5-year patient survival rate was available for only periods 1 and 2, and it was similar (69.4% and 71.8% for periods 1 and 2, respectively, P = 0.10). In contrast, there was an improvement in survival for non-HCV patients during the study period (3-year patient survival was 77.5%, 81.4%, and 80.0% for periods 1, 2, and 3, respectively, P < 0.001) (Fig. 3B). When the survival was adjusted for confounding variables by Cox multivariate regression analysis, HCV patients showed similar survival during periods 1, 2, and 3. However, the survival improved for non-HCV patients with lower survival for period 1 (HR 1.27, 95% CI 1.13-1.43, P < 0.001), and similar survival for period 2 (HR 0.99, 95% CI 0.89-1.10, P = 0.85) and period 3 (reference period).

Figure 3.

(A) Kaplan-Meier patient and graft survival analysis for patients with hepatitis C virus (HCV) at 3 different periods. (B) Kaplan-Meier patient and graft survival analysis for non-HCV patients at 3 different periods.

HCV-Positive Donor

For patients transplanted for HCV, there have been approximately 20-40 HCV-positive donor transplants per year since 1994. Because of the lack of data for period 1 and the small number of patients each year, survival of patients with HCV who had LTs that used HCV-positive donors and HCV-negative donors were compared over the 10 years combined and not divided into the 3 time periods. Kaplan-Meier survival analysis showed that there was no difference in the patient survival between those who had LT with HCV-positive donors or HCV-negative donors (3-year survival rate 81.9% vs. 77.0%, respectively, P = 0.38) (Fig. 4A).

Figure 4.

(A) Kaplan-Meier patient survival analysis comparing survival rates for patients transplanted for hepatitis C virus (HCV) who received HCV-negative livers vs. HCV-positive livers. (B) Comparison of survival rate for HCV patients who received live donor (LDLT) or deceased donor (DDLT) transplants by Kaplan-Meier survival analysis.

Living Donor Liver Transplants

We also compared the survival rates of patients with LDLT with those who had DDLT. Comparison of survival for each of the 3 time periods was not possible because of the low numbers of LDLT, but the overall survival rate during 1991-2001 for those transplanted for HCV by LDLT (n = 207) and DDLT (n = 480 matched) was analyzed. By Kaplan-Meier survival analysis, there was no difference in overall patient survival (P = 0.6) (Fig. 4B), but there was a decreased (P < 0.01 by log rank test) graft survival in HCV-positive patients transplanted with a living donor.

Donor Liver Histology

There were only limited data on donor biopsy findings. Most of the data provided came from the last 2-3 years of the study period. The degree of fatty infiltration, fibrosis, and portal infiltrate had no influence on patient or graft survival, but data were not complete enough to draw any firm conclusions.

DISCUSSION

There has been a major shift in the indication for LT between 1991 and 2001, and HCV has become the single most important cause of end-stage liver disease leading to LT. Although only 16% of transplantations were performed for HCV-related cirrhosis in 1991, this had increased to >50% in 2001. There was an overall improvement in survival of transplant patients during 1994-1997 compared with 1991-1993, with no further improvement after 1998. When compared with non-HCV patients, those with HCV did not show an improvement in survival during the study period. The results remained unchanged when adjusted for other confounding factors that influenced survival.

Previous studies from Europe had reported a decreased survival rate in patients transplanted more recently for HCV compared with other causes of end-stage liver disease.2, 3 In a single-center study from Spain, Mutimer et al.2 reported a marked decline in survival of patients with HCV transplanted more recently in their center. Our study is different from their study for many reasons. We used a large data set that included 7,459 HCV patients and 20,734 non-HCV patients, and this allowed us to adjust the survival for many other confounding predictors. Our data included patients transplanted from all centers in the United States, and this reduced the impact of center-specific differences in immunosuppression and treatment bias with regard to rejection episodes. We also excluded patients with HCC from the HCV group because its presence, however carefully selected, is likely to have an impact on patient survival.6 Mutimer et al.2 included 79 patients with HCC in their HCV group. Although they did not find a statistically significant survival difference between patients with and without HCC, HCC is still an important predictor of survival, as we have shown with a large data set.6 It is important to note that a majority of HCC patients in their group were transplanted more recently, and in fact 38% of their patients transplanted in 1999-2000 had HCC.2 When 283 HCV patients are divided into 5 groups with a large proportion of them with HCC, small differences become exaggerated because there will be only a small number of patients at each follow-up time interval.

As one examines their data more carefully, in their HCV group, there was no real trend in 1-year graft survival (81% for 1991-1992, 69% for 1993-1994, 82% for 1995-1996, 77% for 1997-1998, and 58% for 1999-2000) or Kaplan-Meier survival curves except for those who were transplanted in 1999-2000. In our study, there was no decline in survival for HCV group in recent years. Nevertheless, relative to non-HCV group, there was no expected improvement in survival in HCV group in the more recent era. Our observations could not be explained by donor age or other confounding factors because the lower survival persisted even after adjusting for these factors. Donor characteristics, such as the use of HCV-positive donors or LDLT, also could not explain our observations. We cannot comment on the effect of various immunosuppressive regimens or viral characteristics on graft and patient survival because these data were not available for analysis.

The UNOS database that was used in this study had few limitations. There were missing or partial data for recipient and donor liver histology, complications after LT, treatment used for rejection episodes, and maintenance immunosuppressive regimens. The recurrence or aggressiveness of HCV after LT might be affected by degree of immunosuppression, number of rejection episodes, steroid boluses, use of monoclonal antibody such as OKT3, or antilymphocytic globulin.7–10 Immunosuppression in HCV patients is a fine balance between adequate immunosuppression to prevent rejection, yet not too excessive to prevent severe recurrence. Recently, in small study, it was shown that the choice of calcineurin inhibitors might have limited impact on the outcomes. It is more likely that it is the degree of immunosuppression that matters most. Because there are no routine tests to assess the adequacy of immunosuppression in these patients, the choice of drugs (monotherapy or combination therapy; corticosteroid tapering) and degree of immunosuppression (trough levels) is mainly dependent on the treating physician's preference. Moreover, it is often difficult to make a firm distinction between rejection and recurrence when they coexist, leading to over- or undertreatment. Because only a minority of recurrent HCV was treated during the study, treatment of HCV is unlikely to have had had any impact on the outcome in our study. Despite some of its weaknesses, the large UNOS database that we used for this study is robust enough to compare trends in survival. Moreover, this large data set allowed us to adjust for important potential confounders and understand the real trends in survival better than relatively small-single center studies.

In summary, the survival of patients transplanted with HCV is far lower than those without HCV. There has been a statistically significant improvement in patient and graft survival for non-HCV recipients between 1991 and 2001, but for HCV recipients, the survival rate has remained unchanged during the same period without any obvious explanations.

Acknowledgements

We thank UNOS for providing us the database.

Ancillary