Sirolimus-based immunosuppression is associated with increased survival after liver transplantation for hepatocellular carcinoma

Authors

  • Christian Toso,

    Corresponding author
    1. Section of Hepatobiliary, Pancreatic and Transplant Surgery, University of Alberta, Edmonton, Canada
    2. Transplantation Unit, Department of Surgery, University of Geneva Hospitals, Geneva, Switzerland
    • Department of Surgery, Rue Gabrielle-Perret-Gentil, 1211 Geneva, Switzerland
    Search for more papers by this author
    • fax: 41-22-3727755.

  • Shaheed Merani,

    1. Section of Hepatobiliary, Pancreatic and Transplant Surgery, University of Alberta, Edmonton, Canada
    Search for more papers by this author
  • David L. Bigam,

    1. Section of Hepatobiliary, Pancreatic and Transplant Surgery, University of Alberta, Edmonton, Canada
    Search for more papers by this author
  • A.M. James Shapiro,

    1. Section of Hepatobiliary, Pancreatic and Transplant Surgery, University of Alberta, Edmonton, Canada
    Search for more papers by this author
  • Norman M. Kneteman

    1. Section of Hepatobiliary, Pancreatic and Transplant Surgery, University of Alberta, Edmonton, Canada
    Search for more papers by this author
    • N.M.K. was a Senior Scholar of the AHFMR.


  • Potential conflict of interest: Nothing to report.

Abstract

Liver transplantation is an important treatment option for selected patients with nonresectable hepatocellular carcinoma (HCC). Several reports have suggested a lower risk of posttransplant tumor recurrence with the use of sirolimus and a higher one with calcineurin inhibitors, but the selection of an ideal immunosuppression protocol is still a matter of debate. The aim of this study was to define the immunosuppression associated with the best survival after liver transplantation for HCC. It was based on the Scientific Registry of Transplant Recipients and included 2,491 adult recipients of isolated liver transplantation for HCC and 12,167 for non-HCC diagnoses between March 2002 and March 2009. All patients remained on stable maintenance immunosuppression protocols for at least 6 months posttransplant. In a multivariate analysis, only anti-CD25 antibody induction and sirolimus-based maintenance therapy were associated with improved survivals after transplantation for HCC (hazard ratio [HR] 0.64, 95% confidence interval [CI]: 0.45–0.9, P ≤ 0.01; HR 0.53, 95% CI: 0.31–0.92, P ≤ 0.05, respectively). The other studied drugs, including calcineurin inhibitors, did not demonstrate a significant impact. In an effort to understand whether the observed effects were due to a direct impact of the drug on tumor or more on liver transplant in general, we conducted a similar analysis on non-HCC patients. Although anti-CD25 induction was again associated with a trend toward improved survival, sirolimus showed a trend toward lower rates of survival in non-HCC recipients, confirming the specificity of its beneficial impact to cancer patients. Conclusion: According to these data, sirolimus-based immunosuppression has unique posttransplant effects on HCC patients that lead to improved survival. (HEPATOLOGY 2010.)

Liver transplantation is the treatment of choice for selected patients with nonresectable hepatocellular carcinoma (HCC). Although the surgical procedure is well established, the definition of the most appropriate immunosuppression combination, allowing decreased risk of tumor recurrence and improved survival, is still a matter of debate. To date, no single protocol has gained broad acceptance. In recent years this lack of consensus has become more acute, given the increasing number of patients undergoing transplantation for HCC, currently the second commonest indication for liver transplantation in the USA, after hepatitis C virus (HCV) disease (www.ustransplant.org/annual_reports). We can also expect the number of transplantation for HCC to further increase, with several recent studies showing that selected patients beyond Milan criteria can be safely considered for transplantation.1–5

Until now the positive and negative effects of various immunosuppressive drugs on HCC have been demonstrated in animal data and suggested from a few retrospective single-center clinical studies. Calcineurin inhibitors, including both tacrolimus and cyclosporine, have been associated with a dose-dependent increase in the posttransplant risk of HCC recurrence.6 Conversely, sirolimus has shown anticancer properties in in vitro and animal models, both alone or in combination with doxorubicin or sorafenib.7–12 Sirolimus can prevent angiogenesis by interfering with vascular endothelium growth factor (VEGF)-mediated pathways in endothelial cells, thus limiting the growth of tumors,7 and also impacts established tumors, by inducing extensive microthrombi and so inhibiting tumor growth.9, 13 Although these animal data are clear, clinical studies are less convincing. We have demonstrated good outcomes with the use of sirolimus in a noncontrolled trial, and more recently the groups at the University of Colorado in Denver and Fudan University in Shanghai demonstrated better survivals in patients on sirolimus compared to control liver recipients.4, 14–16 Although all show similar trends, these retrospective studies included limited numbers of patients, and possible confounding variables could not be taken into account due to the limited sample size.

The present study is based on a large registry transplant population and evaluates the impact of immunosuppression on survival in an attempt to define the best posttransplant treatment combination for HCC patients.

Abbreviations

AFP, alpha fetoprotein, HCC, hepatocellular carcinoma; HBV, hepatitis B virus; HCV, hepatitis C virus; HR, hazard ratio; HRSA, Health Resources and Services Administration; MELD, Model for End-Stage Liver Disease; OPTN, Organ Procurement and Transplantation Network; SRTR, Scientific Registry of Transplant Recipients; TTV, total tumor volume; UNOS, United Network for Organ Sharing.

Patients and Methods

This study analyzed data from the Scientific Registry of Transplant Recipients (SRTR). The SRTR data system includes data on all donors, wait-listed candidates, and transplant recipients in the United States, submitted by the members of the Organ Procurement and Transplantation Network (OPTN), and has been described elsewhere.17 The Health Resources and Services Administration (HRSA), US Department of Health and Human Services, provides oversight to the activities of the OPTN and SRTR contractors. The study was reviewed and approved by the Health Research Ethics Board at the University of Alberta. The study population included all adult (≥16 years) patients who received an isolated liver transplantation from March 2002 to March 2009. In order to ensure that all subjects had a significant exposure to the drugs, only individuals kept on the same maintenance immunosuppression protocol for at least 6 months posttransplant (or until death) were further selected. Overall, 25,201 out of 39,859 patients receiving a liver transplant during the study period were excluded.

The unique outcome variable of the study was patient survival. The occurrence and the date of death were obtained from data reported to the SRTR by the transplanting centers and were completed by data from the US Social Security Administration and from the OPTN. All deaths were taken into account in the analysis, whether they were associated with HCC or not. Of note, the SRTR data were not of sufficient granularity nor previously validated to allow the use of variables such as cancer recurrence or cancer-associated death. As a consequence, some patients may have been alive with an HCC recurrence and were not considered as an event in the survival analyses.

A first analysis was conducted on patients transplanted for HCC only. Subjects with cholangio-carcinoma, hepatoblastoma, hemangio-endothelioma, and benign liver tumors were excluded. We performed a univariate analysis using the Kaplan Meier technique and comparing groups with log-rank tests. The impact of immunosuppression was analyzed, comparing patients put on a specific drug prior to the original posttransplant discharge and kept on the same drug for at least 6 months, to those who had not been put on that specific drug for at least 6 months posttransplant. The following variables were used: tacrolimus (Prograf), cyclosporin (Sandimmune, Neoral, and generics), sirolimus (Rapamune), mycophenolate mofetil (Cellcept), steroids (methylprednisolone, Solumedrol, and oral prednisone, excluding patients treated with steroids for rejection episode), and induction therapy with an anti-CD25 antibody (daclizumab, Zenapax and basiliximab, Simulect) or with Thymoglobulin.

We further conducted a stepwise multivariate Cox regression analysis. The previously described immunosuppression variables were all entered in this analysis and results were corrected for the following covariates: Model for End-Stage Liver Disease (MELD) score, year of transplant, age at transplant, primary underlying liver disease, total tumor volume (TTV), alpha-fetoprotein (AFP) and pretransplant tumor treatment (yes versus no). TTV was calculated as previously reported by adding the volume of each HCC ((4/3)πr3) based on the maximum radiological radius of each tumor.5, 18, 19 Of note, only TTV and AFP were used as HCC factors, as they have been previously reported to be the main variables impacting patient survival.5, 18 Data obtained on the date closest to transplant were used.

In an effort to understand whether the observed results were due to specific impacts of the drugs on HCC or more generally on liver transplantation overall, we further conducted the same univariate and multivariate analyses independently on patients transplanted during the same time period for non-HCC diagnoses. Similar variables and covariates were used, excluding those directly applicable to HCC patients: TTV, AFP, and pretransplant tumor treatment.

The study design, which only allowed the inclusion of patients on stable immunosuppression for at least 6 months posttransplant, resulted in data on immunosuppression being available for all patients. Covariates were missing in less than 35% (the most frequently missing was AFP) in the HCC group and were replaced by the mean. Covariates were complete in the non-HCC group.

Other statistical tests included the use of Student's t and chi-square tests to compare the demographic variables between groups. Results were provided as mean ± standard deviation. A standard alpha level of 0.05 indicated statistical significance. Analyses were conducted using SPSS 15.0 (Chicago, IL).

Results

During the study period, 2,491 adult patients received an isolated liver transplant for HCC and 12,167 for non-HCC diagnoses (Table 1). All analyzed patients remained on the same maintenance immunosuppressive drugs for at least 6 months posttransplant. HCC patients included more males (female/male ratio: 1/3.9 versus 1/1.8, P ≤ 0.001) and were older (56 ± 8 versus 51 ± 11 years on average, P ≤ 0.001). The incidence of HCV- and hepatitis B virus (HBV)-induced liver disease was also higher among HCC patients (P ≤ 0.001). Finally, calculated MELD scores, not adjusted for tumor exception points, were lower in the HCC group (14 ± 6 versus 21 ± 8, P ≤ 0.001).

Table 1. Patient and Tumor Characteristics
 HCC patientsnon-HCC patientsp
  1. HCV: hepatitis C virus infection, HBV: hepatitis B virus infection, NA: non applicable

  2. MELD: Model for End-Stage Liver Disease

Patients (number)249112167 
Mean age (years ± SD)56 ± 851 ± 11≤0.001
Gender (ratio)female:504/male: 1987female:4364/male:7803≤0.001
Cause of liver disease (%)   
 HCV (±alcohol, ±HBV)1343 (54)4225 (35)≤0.001
 Alcohol227 (9)1642 (13)≤0.001
 HBV199 (8)329 (3)≤0.001
 Cryptogenic90 (3.5)1033 (8)≤0.001
 NASH47 (2)343 (3)0.008
 Autoimmune27 (1)462 (4)≤0.001
 Primary biliary cirrhosis19 (0.5)588 (5)≤0.001
 Primary sclerosing cholangitis13 (0.5)857 (7)≤0.001
 Hemochromatosis20 (0.5)65 (0.5)0.1
 Alpha1-antitrypsin deficiency7 (0.5)137 (1)≤0.001
 Metabolic disorder7 (0.5)75 (0.5)0.04
 Acute liver necrosis86 (3)844 (7)≤0.001
 Other433 (17)1567 (13)≤0.001
MELD score at transplantation (±SD)14 ± 621 ± 9≤0.001
Mean Total Tumor Volume (cm3 ±SD)17 ± 39NA 
Total Tumor Volume >115 cm3 (%)5 (0.2)NA 
Mean serum alpha fetoprotein level (ng/ml ±SD)291 ± 1384NA 
Serum alpha fetoprotein level >400 ng/nl (%)157 (6)NA 

As the SRTR registry is based in the US, where HCC patient selection is performed according to Milan criteria,1 only 0.2% of HCC subjects had a TTV higher than 115 cm3. Six percent had an AFP >400 ng/mL. As a result, the included HCCs were relatively homogenous and with similar expected outcomes.5

The use of immunosuppressive drugs was similar between HCC and non-HCC patients. An induction therapy was used in a minority of recipients (anti-CD25 antibody: 12% and 10.8%, Thymoglobulin 6.3% and 7.3%). The most frequently used maintenance therapies were tacrolimus (90.6% and 92.5%), steroids (82.9% and 85.7%), and mycophenolate mofetil (57.6% and 59.5%).

We first performed a univariate analysis based on the HCC group only. Patients receiving induction with anti-CD25 antibodies and those treated with a sirolimus-based maintenance protocol demonstrated significantly higher survivals that reached 6% and 14.4% advantages by 5 years (P ≤ 0.01 and P ≤ 0.05, respectively; Table 2, Fig. 1). On multivariate analysis, corrected for MELD score, year of transplant, age at transplant, primary underlying liver disease, TTV, AFP, and pretransplant tumor treatment, both anti-CD25 antibodies and sirolimus remained significant predictors of patient survival (hazard ratio [HR] 0.64, 95% confidence interval [CI]: 0.45–0.9, P ≤ 0.01; HR 0.53, 95% CI: 0.31–0.92, P ≤ 0.05).

Table 2. Univariate Analysis of Factors Impacting Patient Survival
HCC patientsNon-HCC patients
 3-year patient survival (%)5-year patient survival (%)p* 3-year patient survival (%)5-year patient survival (%)p*
  • *

    Log-rank tests

Tacrolimus       
  on (n = 2256)79.969.60.87on (n = 11259)87.379≤0.001
  off (n = 235)77.568.2 off (n = 908)81.872.9 
Cyclosporin       
  on (n = 207)76.669.20.67on (n = 735)81.872.9≤0.001
  off (n = 2284)79.969.4 off (n = 11432)87.279 
Sirolimus       
  on (n = 109)85.683.1≤0.05on (n = 430)84.2740.14
  off (n = 2382)79.268.7 off (n = 11737)8778.7 
Mycophenolate mofetil       
  on (n = 1434)78.368.50.32on (n = 7244)8778.80.62
  off (n = 1057)8170 off (n = 4923)86.878.2 
Steroids       
  on (n = 2066)7968.50.07on (n = 10429)86.878.10.11
  off (n = 425)82.773.8 off (n = 1738)87.580.6 
Induction: anti-CD25 antibody       
  on (n = 299)87.974≤0.01on (n = 1320)8878.70.6
  off (n = 2192)78.568 off (n = 10847)86.778.5 
Induction: thymoglobulin       
  on (n = 158)75.666.30.44on (n = 894)88.379.90.5
  off (n = 2333)79.869.6 off (n = 11273)86.878.4 
Figure 1.

Kaplan-Meier analysis of survival after liver transplantation for HCC. Both sirolimus (A, P ≤ 0.05) and anti-CD25 antibody induction (B, P ≤ 0.01) demonstrated significantly improved survivals.

Of note, the protective effect of sirolimus did not appear to be linked to a selection bias, as patients on sirolimus demonstrated higher MELD scores than those sirolimus-free (15 ± 7 versus 14 ± 1, P = 0.02). In addition, the other studied characteristics were either similar between both groups, or are without known impact on HCC-free posttransplant survival (Table 3).

Table 3. Characteristics of HCC Patients On or Off Sirolimus
 HCC patients on sirolimusHCC patients sirolimus-freep
  1. HCV: hepatitis C virus infection, HBV: hepatitis B virus infection

  2. MELD: Model for End-Stage Liver Disease

Patients (number)1092382 
Mean age (years ± SD)56 ± 656 ± 20.82
Gender (ratio)female:32/male:77female:472/male:19100.02
Cause of liver disease (%)   
 HCV (±alcohol, ±HBV)52 (57)1291 (54)0.18
 Alcohol7 (6)220 (9)0.34
 HBV4 (4)195 (8)0.09
 Cryptogenic4 (4)86 (3.5)0.97
 NASH1 (1)46 (2)0.45
 Autoimmune2 (2)25 (1)0.44
 Primary biliary cirrhosis1 (1)18 (0.5)0.85
 Primary sclerosing cholangitis013 (0.5)0.44
 Hemochromatosis1 (1)19 (0.5)0.89
 Alpha1-antitrypsin deficiency2 (2)5 (0.5)0.002
 Metabolic disorder1 (1)6 (0.5)0.2
 Acute liver necrosis1 (1)85 (4)0.3
 Other33 (30)373 (16)≤0.001
MELD score at transplantation (±SD)15 ± 714 ± 10.02
Mean Total Tumor Volume (cm3 ±SD)19 ± 2.317 ± 1.10.78
Total Tumor Volume >115 cm3 (%)05 (0.2)0.5
Mean serum alpha fetoprotein level (ng/ml ±SD)147 ± 72297 ± 360.06
Serum alpha fetoprotein level >400 ng/nl (%)3 (1)154 (6)0.18

In an effort to understand whether the positive impact of these drugs on survival was linked to a specific effect on HCC or to a general effect on liver transplantation, we conducted a similar analysis on the non-HCC group. Based on this analysis, a drug having a positive effect on the HCC group and a negative effect on the non-HCC group should be assumed to have a mode of action with beneficial impact mainly targeting HCC. This could be the case for a drug holding anticancer properties. Conversely, a drug associated with improved survivals on both HCC and non-HCC patients should be assumed to act on the outcome of liver transplantation in general.

When looking at the non-HCC group and performing a univariate analysis, tacrolimus-based therapy was associated with improved survivals, whereas cyclosporine-based treatment was linked to decreased survivals (Table 2). On multivariate analysis, only the use of cyclosporine-based maintenance protocols remained associated with decreased outcomes (HR 1.3, 95% CI: 1–1.7, P ≤ 0.05).

In order to better understand the effects of the various studied drugs on patient survival, we plotted the HR (±95% CIs) found in the HCC and non-HCC groups (Fig. 2). Of the significant variables in the multivariate analyses, both anti-CD25 antibodies and cyclosporine demonstrated trends in similar directions in both groups, suggesting that their effect was primarily directed toward liver transplant in general (and not specifically toward HCC). Conversely, sirolimus-based immunosuppression had a trend toward a protective effect in the HCC group and a negative impact in the non-HCC group, thus suggesting that the significant effect of this drug was primarily directed toward HCC. Of all the studied protocols, sirolimus-based immunosuppression was the only one showing such a pattern.

Figure 2.

Hazard ratios (±95% CI) to compare the risk of mortality after liver transplantation using various immunosuppression protocols. Results were corrected for MELD, year of transplant, primary liver disease (non-HCC), age at transplant, and, when applicable, TTV, AFP, and pretransplant HCC treatment. *Significant variables.

As described earlier, the SRTR does not include data on HCC recurrence. In order to approximate these data we performed an assessment of patients dying of malignancy, and have shown that twice as many patients not on sirolimus died from cancer (HCC or other) compared to those on sirolimus (11% versus 5%, respectively, at 5 years). Although this observation did not reach significance (P = 0.15, log-rank), possibly due to the low report rate of this variable, the observed trend further supports the message of the study.

Discussion

According to the present SRTR registry data, the use of sirolimus-based immunosuppression protocols has unique beneficial posttransplant effects on HCC patients, leading to significantly improved survival. Our analysis of the SRTR dataset also suggests that anti-CD25 antibody induction is associated with a longer posttransplant life expectancy (significant for HCC and with a trend for non-HCC).

Although the anticancer effect of sirolimus has been suggested by previous animal and single-center studies,7–12 the present data are the first to be adequately powered. It highlights a key potential role for this agent in patients undergoing liver transplantation for HCC. It should also encourage both physicians and regulatory agencies to consider removal of the “black box” linked to the use of sirolimus in liver transplantation (www.fda.gov). This warning was put forward by the Food and Drug Administration (FDA) in 2002, based on Phase II trials showing a significantly increased incidence of early posttransplant hepatic artery thrombosis and infection-associate deaths (www.fda.gov). Sirolimus has now passed the test of time in several centers, all reporting no increase in the incidence of these complications.14, 20 Like any potent immunosuppressive drug, sirolimus is linked to a potential for development of numerous side effects, including dyslipidemia, peripheral edema, anemia, leukopenia, delayed wound healing, and a substantially increased risk of incisional hernia.14, 21, 22 In general, however, we believe that these side effects are relatively minor and easy to manage, and that the data revealed by the present study justify a broader use of protocols including sirolimus after liver transplantation for patients with HCC.

It should be clearly emphasized that this study was not designed to look at the effect of specific drugs, but rather reports on protocols containing specific drugs. We had no access to drug doses or trough levels. As such, whereas it sounds logical that the improved survival associated with sirolimus-containing protocols is the result of its anticancer effects, we cannot rule out that lower doses of calcineurin inhibitors (CNIs) were used in these patients, perhaps reinforcing the effect of sirolimus.6

Of all protocols, sirolimus-based immunosuppression was the only one associated with an improved posttransplantation survival specific to HCC patients (and not to non-HCC patients), further reinforcing the clinical evidence of its anticancer properties. The use of anti-CD25 antibodies demonstrated similar trends to improved survival in both HCC and non-HCC patients. These observations, together with previous reports combining anti-CD25 antibody induction and delayed introduction of CNIs, speak in favor of the use of this drug after liver transplantation in general.23 Finally, the present data also support the use of tacrolimus-based rather than cyclosporine-based maintenance immunosuppression after liver transplant.

The registry nature of the study is linked to several limitations. We did not have access to data on HCC recurrence, which would have been useful to better define the anticancer impact of the drugs. However, due to the lack of access to effective treatment, most patients with HCC recurrence posttransplant are expected to die from the disease, making the rate of survival a reasonable marker. In addition, most deaths occurring during the first 5 years after transplantation for HCC are related to tumor recurrence (and not other causes like HCV recurrence).14 In an effort to provide better understanding of the anti-HCC effect of the drugs, we studied both HCC and non-HCC groups separately and subsequently compared them, thus providing a good reading of the effect of sirolimus-based therapies on HCC.

In order to further decrease the risk of bias, the multivariate analyses were corrected for multiple variables that could potentially impact survival. To illustrate, the opposite results linked to sirolimus between the HCC and non-HCC groups could have been linked to differences in the indications for the use of the drug. The use of sirolimus was linked to the presence of HCC in one group, and may have been used to spare the use of CNIs in the other group, especially in patients with renal dysfunction. Although the risk of bias was decreased by the integration of pretransplant MELD in the analysis, we could not completely rule out differences in posttransplant kidney function between groups.

According to the present study, sirolimus-based immunosuppression is associated with improved patient survival after liver transplantation for HCC. Anti-CD25 antibody induction demonstrates a similar effect in patients transplanted for HCC and non-HCC diagnoses. We believe that these data will help in the transplant management of HCC patients, integrating a balanced selection of candidates with expected good outcomes and a posttransplant adjuvant treatment including appropriate and effective immunosuppression with anticancer properties.

Acknowledgements

The data reported here were supplied by the Arbor Research Collaborative for Health (Arbor Research) as the contractor for the Scientific Registry of Transplant Recipients (SRTR). The interpretation and reporting of these data are the responsibility of the authors and in no way should be seen as an official policy of or interpretation by the SRTR or the U.S. Government. C.T. had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Ancillary