Minimal but significant improvement in survival for non–hepatitis C–related adult liver transplant patients beyond the one-year posttransplant mark

Authors

  • Geoffrey W. McCaughan,

    Corresponding author
    1. Australian National Liver Transplant Unit, Royal Prince Alfred Hospital, University of Sydney, Sydney, Australia
    2. Centenary Research Institute, Royal Prince Alfred Hospital, University of Sydney, Sydney, Australia
    • Royal Prince Alfred Hospital, University of Sydney, Sydney, Australia
    Search for more papers by this author
    • Telephone: +61 2 9565-6125, +61 2 9515-8578, and 0418 212 805; FAX: + 61 2 9565 6100

  • Nicholas A. Shackel,

    1. Australian National Liver Transplant Unit, Royal Prince Alfred Hospital, University of Sydney, Sydney, Australia
    2. Centenary Research Institute, Royal Prince Alfred Hospital, University of Sydney, Sydney, Australia
    Search for more papers by this author
  • Simone I. Strasser,

    1. Australian National Liver Transplant Unit, Royal Prince Alfred Hospital, University of Sydney, Sydney, Australia
    Search for more papers by this author
  • Pamela Dilworth,

    1. Australian National Liver Transplant Unit, Royal Prince Alfred Hospital, University of Sydney, Sydney, Australia
    Search for more papers by this author
  • Patrick Tang

    1. Australian National Liver Transplant Unit, Royal Prince Alfred Hospital, University of Sydney, Sydney, Australia
    Search for more papers by this author

Abstract

Although 1-year survival rates following liver transplantation over the last 20 years may have improved, there is doubt about improvement in long-term survival. We examined survival with and without initial 12-month mortality in adult liver transplant recipients over a 20-year period. Patient and allograft survival for 3 different time periods was compared: 1986-1994 (group 1, n = 547), 1995-2000 (group 2, n = 735), and 2000-2005 (group 3, n = 749). After this, all deaths in the first 12 months of each group were removed. Patient and allograft survival was then once again compared across the 3 groups. There was significant improvement in both patient and allograft survival across the 20-year period (P < 0.001). Overall patient and allograft survival improved in non–hepatitis C virus (HCV) patients but not in HCV patients. A similar comparison with deaths in the first year removed, however, showed no difference in patient survival (P = 0.07) and only a marginal improvement in allograft survival (P = 0.048) between the 3 time periods. When patients were divided into HCV-positive and HCV-negative groups with deaths in the first year removed, there was, however, improved patient and allograft survival in the HCV-negative group but not in the HCV-positive group. The causes of death between 1 and 5 years were then compared. There were 48 deaths in period 1, 63 in period 2, and 43 in period 3 (P = not significant). There were more deaths due to cardio/cerebrovascular disease and hepatitis B virus recurrence in the first time period, but there were more deaths due to recurrent HCV and de novo malignancy in later time periods. In conclusion, although overall survival following liver transplantation in adults seems to be improving over time, the long-term results are not, particularly in HCV patients. Liver Transpl 16:130–137, 2010. © 2010 AASLD.

It is generally recognized that there has been improved survival over the past 20 years for patients undergoing liver transplantation.1, 2 This has been attributed to improvements in surgical techniques, immunosuppression, and anti-infective therapies that particularly benefit early survival. However, in renal transplantation, it has been realized that improved 1-year survival may not necessarily translate into improved long-term benefits.3 This has now been recognized to be associated with what is known as chronic allograft nephropathy. Such a process is not recognized after liver transplantation with a less than 5% rate of chronic allograft rejection.4, 5

However, it has always been realized that patients who require long-term immunosuppression are at risk of cardiovascular diseases and malignancy.6, 7 Furthermore, since the emergence in the past 10 years of recurrent hepatitis C virus (HCV) infection in the liver allograft, it has been realized that long-term results following liver transplantation are not as good as originally anticipated.8

Doubts about improvements in long-term survival in liver transplant patients have been raised in a recent publication that examined mortality following liver transplantation over 2 consecutive 5-year periods.9 Although overall improved survival was seen (but not in all patient diagnoses), when first 12-month mortality was censored, there was indeed no improvement in 1- to 5-year survival between the 2 periods, and this indicated no improvement in long-term outcomes. This is the only publication that has so far addressed this issue in such a way, although a brief comment in a UK analysis suggested similar findings when first 6-month mortality was censored.10 Thus, the aim of this study was to perform a survival analysis with and without initial 12-month mortality on a different database [the Australian and New Zealand Liver Transplant Registry (ANZLTR)]. This database has prospectively gathered mortality data over a 20-year period for patients undergoing liver transplantation in Australia and New Zealand.

Abbreviations: AIH, autoimmune hepatitis; ALD, alcoholic liver disease; ANLTU, Australian National Liver Transplant Unit–Sydney; ANZLTR, Australian and New Zealand Liver Transplant Registry; FHF, fulminant hepatic failure; HAT, hepatic artery thrombosis; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; NZLTU, New Zealand Liver Transplant Unit; PBC, primary biliary cirrhosis; PSC, primary sclerosing cholangitis; QLTU, Queensland Liver Transplant Unit; SALTU, South Australian Liver Transplant Unit; UNOS, United Network for Organ Sharing; VLTU, Victorian Liver Transplant Unit; WALTU, Western Australian Liver Transplant Unit.

PATIENTS AND METHODS

Patient Population

The ANZLTR has prospectively received mortality data from all transplant units in Australia and New Zealand since 1986. Causes of death and underlying disease indications have also been obtained. The analysis reported here includes only adult patients (>16 years) and occurred in several steps. First, overall patient and allograft survival for 3 different time periods was compared: 1986-1994 (group1), 1995-2000 (group 2), and 2000-2005 (group 3). The longer period in group 1 was chosen to allow similar numbers of transplants to be compared. After this, all deaths in the first 12 months of each group were censored (removed). Patient and allograft survival was then once again compared across the 3 groups. Follow-up was in December 2007. Patients were also separated on the basis of the original disease into HCV infection and non-HCV groups. The diagnosis of HCV infection was based on HCV antibody testing. Causes of death were also compared between 1 and 5 years post–liver transplantation across groups.

Statistical Methods

All statistical analyses were performed with SPSS 15.0.1. Univariate and multivariate analyses of covariates and Kaplan-Meier and Cox proportional hazard models on survival were conducted. The chi-square test was used to compare differences of categorical variables, and the t test was used to compare continuous variables. P < 0.05 was considered to be significant.

RESULTS

From 1986 to 2005, 2031 adults underwent transplantation. There were 123 second transplants and 7 third transplants, and this resulted in 2161 allografts. Between 1986 and 1994 (group 1), there were 547 patients undergoing primary transplantation; between 1995 and 2000 (group 2), there were 735; and between 2001 and 2005 (group 3), there were 749. Primary diagnoses across the 3 eras are shown in Table 1. The major shifts were an increase in HCV and hepatocellular carcinoma (HCC) and a reduction in primary biliary cirrhosis, autoimmune hepatitis, primary sclerosing cholangitis, and fulminant hepatic failure between groups 1 and 2. There was an increase in recipient and donor age across the 3 eras.

Table 1. Indications for Liver Transplantation Across the Three Eras
 Group 1: 1985-1994 (n = 547)Group 2: 1995-2000 (n = 735)Group 3: 2001-2005 (n = 749)P Value
Group 1 Versus Group 2Group 1 Versus Group 3Group 2 Versus Group 3
  • *

    This reflects primary diagnoses (n = 10, 12, and 23) and secondary diagnoses.

Primary diagnosis      
 AIH5231270.00020.00010.5931
 PSC9178750.00200.00050.7331
 PBC7356340.00100.00010.0163
 HBV3675580.02680.44950.1023
 HCV52167251<0.0001<0.0001<0.0001
 HCC*25*95*181*<0.0001<0.0001<0.0001
 FHF5868460.44850.00500.0252
 ALD4794990.01890.00980.7576
 Genetic4685580.07630.67970.0137
 Cryptogenic4644450.09820.09961
 Other3625330.01120.10290.3498
Age      
 Recipient43.5 ± 12.7 (15-67)46.9 ± 11.8 (15-68)49.8 ± 10.5 (16-73)<0.0001<0.0001<0.0001
 Donor (first graft)31.6 ± 13.7 (4-67)37.8 ± 15.8 (6-77)40.6 ± 16.8 (7-81)<0.0001<0.0001<0.0001

Figure 1 shows patient and graft survival in the 3 time periods. There was significant improvement in both patient and allograft survival across the 20-year period (P < 0.001). This was a stepwise improvement in survival between groups 1 and 2 and subsequently between groups 2 and 3.

Figure 1.

Kaplan-Meier curves showing (A) overall patient survival and (B) overall allograft survival over the 3 time periods. There was increased patient and allograft survival (P < 0.001 for both).

Figure 2 shows a similar comparison with deaths in the first year removed. These data show an insignificant improvement in patient survival (P = 0.072) and a marginal improvement in allograft survival (P = 0.048).

Figure 2.

Kaplan-Meier curves showing (A) patient and (B) allograft survival after the removal of deaths in the first year.

Figure 3 separates out HCV and non-HCV patients. Overall patient and allograft survival improved in non-HCV patients but not in HCV patients, although there seemed to be worse allograft survival in the group 2 period for HCV patients (P = 0.042).

Figure 3.

Kaplan-Meier curves showing overall patient survival in (A) HCV-negative and (C) HCV-positive patients and allograft survival in (B) HCV-negative and (D) HCV-positive patients. Improved survival was seen only in HCV-negative patients.

Figure 4 repeats the same analysis on HCV-positive and HCV-negative patients with deaths in the first year removed. This analysis showed improved patient and allograft survival in HCV-negative patients but not in HCV patients. It should be recognized that a significant number of patients with HCV (n = 470) also had a diagnosis of hepatocellular cancer (n = 132). There was no difference in patient or allograft survival across the 20-year period when HCV-associated HCC was removed (Fig. 5A,B). Hence, we did not dissect out further any confounding effect of HCV-related HCC in our analysis.

Figure 4.

Kaplan-Meier curves showing (A,C) patient and (B,D) allograft survival after the removal of deaths in the first year: (A,B) HCV-negative and (C,D) HCV-positive patients. There was improved patient and allograft survival in HCV-negative patients. Abbreviation: HCV, hepatitis C virus.

Figure 5.

Kaplan-Meier curves showing (A) overall patient survival and (B) overall allograft survival in HCV-positive patients without HCC.

It is well known the donor age may influence patient and allograft survival and that donor age increased across the time periods under study. Thus, we repeated the patient survival analysis beyond the 1-year mark with donor and recipient ages measured as confounders with Cox proportion hazard regression. The results indicate survival curves similar to those shown in Figs. 2 and 4 despite the changing variables of donor and recipient age (Fig. 6).

Figure 6.

Patient survival, after the removal of deaths in the first year, adjusted for recipient and donor age in all patients, HCV patients, and non-HCV patients. The data are results from Cox proportional hazard regression analysis.

The causes of death between 1 and 5 years were then compared between all groups. This period was chosen because of the different lengths of follow-up between each group. This meant that groups 1 and 2 had 5-year follow-up on all patients, whereas the average follow-up for group 3 patients who were alive at 12 months was 4.2 years, with 232 patients followed up to 5 years. There were 48/436 deaths in period 1, 63/641 in period 2, and 43/679 in group 3. When these were then converted into deaths per patient year of follow-up, there were no statistical differences in the number of deaths between 1 and 5 years post-transplant in the 3 groups.

Tables 2 to 4 show different causes of mortality and allograft loss in the 3 time periods, with more deaths due to cardiovascular disease and recurrent malignancy in period 1 and with more de novo malignancy in period 2. With respect to causes of allograft loss in years 1 to 5, in period 1, there was more allograft loss due to hepatitis B virus (HBV), and there was less allograft loss due to HCV. There was no particular difference in the pattern of de novo malignancy across the 3 periods (Table 4).

Table 2. Causes of Death (1-5 Years of Follow-Up)
Cause of DeathGroup 1: 1985-1994 (n = 436)Group 2: 1995-2000 (n = 641)Group 3: 2001-2005 (n = 679)P Value
Group 1 Versus Group 2Group 1 Versus Group 3Group 2 Versus Group 3
Recurrent malignancy181650.21350.00710.0896
Graft failure1019150.28590.16190.6739
Cardiovascular/cerebral8240.01900.36330.2202
Sepsis5450.496810.4805
De novo malignancy41490.06840.13271
Respiratory2110.577411
Other1740.13460.18491
Total4863430.54150.00690.02
Table 3. Cause of Death: Graft Failure (1-5 Years of Follow-Up)
Cause of Death: Graft FailureGroup 1: 1985-1994Group 2: 1995-2000Group 3: 2001-2005P Value
Group 1 Versus Group 2Group 1 Versus Group 3Group 2 Versus Group 3
Rejection (chronic)4420.39040.17530.6722
HBV4120.03570.17530.5714
HCV11070.04360.08751
Biliary1000.34480.41
HAT0200.53210.492
Recurrent disease01210.50.5714
Other011111
Rejection (acute)001110.4412
Total1019150.28590.16190.67
Table 4. Cause of Death: De Novo Malignancy (1-5 Years of Follow-Up)
Cause of Death: De Novo MalignancyGroup 1: 1986-1994Group 2: 1995-2000Group 3: 2001-2005P Value
Group 1 Versus Group 2Group 1 Versus Group 3Group 2 Versus Group 3
Colorectal143111
Lymphoma132111
Lung1110.405211
Skin021111
Esophagus011111
Melanoma1100.40520.30771
Larynx001110.3913
Ovarian010111
Pancreas010111
Total4149   

DISCUSSION

This article has led to important observations about the long-term follow-up of liver transplant recipients. These observations have implications for long-term patient care. The data here show that, although there has been highly significant overall improved survival over a 20-year period, this translates into only marginally improved survival if deaths in the first year are removed from such an analysis. These observations, however, have some caveats that need to be discussed. Although overall survival has improved, this was not seen in HCV-positive patients (Fig. 3C,D). In fact, there was a tendency for worse survival in HCV-positive patients, particularly in the 1995-2000 period. It is interesting to compare these overall survival data to those recently published from the United Network for Organ Sharing (UNOS) and European databases.9, 10 In the UNOS study, there was no improved overall survival in patients with HCV, primary sclerosing cholangitis, primary biliary cirrhosis, or acute hepatic necrosis when 2 eras (1992-1996 versus 1997-2002) were compared. Improved survival was seen only in patients with autoimmune hepatitis and alcoholic liver disease. We have not further separated out our non-HCV patients because of the much smaller numbers. In contrast, there has been a steady overall improvement in the European Liver Transplant Registry data.10

Although the aforementioned results are of interest, the main point of this study was to examine survival beyond 1 year as an endpoint to address long-term survival issues. These results show that there was no difference in survival over a 20-year period (Fig. 2A,B). However, when patients were divided into non-HCV and HCV-positive groups, there was definitely a significant improvement in non-HCV patient and allograft survival. This occurred despite an increase in donor and recipient age.

In a similar analysis of the UNOS database, there were in fact no differences in survival when deaths of the first 12 months were removed for any group, including all non-HCV groups.9 In fact, the HCV and primary sclerosing cholangitis groups had worse allograft survival in the most recent era. This was despite shorter follow-up in the most recent group (1997-2002). (It is obvious that some patients in this latter group had less than the full 5 years of follow-up as the study was first received for publication in 2005). As mentioned earlier, a UK study reported similar survival over 4 separate time periods from 1985 to 2003 when deaths in the first 6 months were censored.11 These data are mentioned in passing in the article and are shown in a single figure. The authors used these data to estimate a 22-year median survival time post–liver transplantation in adult patients surviving beyond 6 months. The European Liver Transplant Registry did not formally address this. However, if Figure 4 in that report is examined closely, it seems that overall survival was improved across different time periods by differences in 12-month survival. In contrast, the survival curves after 12 months seem to have very similar slopes.

It is of interest to examine the causes of death and allograft failure between 1 and 5 years in the 3 eras. This shows that, although the death rate was the same, the causes of death were different in the 3 eras, with more cardiovascular system–related deaths and recurrent HBV deaths in the earlier period versus more de novo cancer deaths in the second period. Increased de novo cancer is recognized after liver transplantation,12 but there did not seem to be any particular pattern concerning this in our patients. In age-matched control studies, we can expect to see increases in posttransplant lymphoproliferative disease, skin cancer, colorectal cancer, and oral malignancy in liver transplant recipients. In contrast to HBV, there was a reversed pattern, with HCV playing a significantly lesser role in the earliest period.

In conclusion, the overall picture from these results is that, although overall survival seems to be improving over time, the long-term results are not, particularly in HCV patients. Thus, HCV recurrence is masking any gains from improved immunosuppression, cardiovascular risk management, and control of HBV recurrence that are seen in the non-HCV group. Whether there is in addition more de novo cancer is hard to examine as there did not seem to be any particular cancer that was more common. The results highlight the need to continue to gather long-term data on liver transplant survivors. It is clear that changing indications for transplantation in combination with changing immunosuppression make comparisons over long time periods difficult. Also, different levels of pretransplant disease severity may confound analyses. These issues have not been able to be addressed in this article as immunosuppression data and severity of liver disease were not collected in the ANZLTR. However, despite these limitations, it is heartening to see that cardiovascular deaths seem to be decreasing. However, caution remains with respect to HCV recurrence and long-term malignancy.

Acknowledgements

In addition to the authors, the members of the Australian and New Zealand Liver Transplant Study Group are as follows: Stephen Lynch [Queensland Liver Transplant Unit (QLTU)], Jonathan Fawcett (QLTU), Robert Jones [Victorian Liver Transplant Unit (VLTU)], Peter Angus (VLTU), Michael Fink (VLTU), Paul Gow (VLTU), Robert Padbury [South Australian Liver Transplant Unit (SALTU)], John Chen (SALTU), Gary Jeffrey [Western Australian Liver Transplant Unit (WALTU)], Luc Delrievere (WALTU), Gerry McQuillan (WALTU), Stephen Munn [New Zealand Liver Transplant Unit (NZLTU)], Ed Gane (NZLTU), John McCall (NZLTU), Debbie Verran [Australian National Liver Transplant Unit–Sydney (ANLTU)], Michael Crawford (ANLTU), and James Gallagher (ANLTU). Glenda Balderson is manager of the Australian and New Zealand Liver Transplant Database.

Ancillary