Liver transplantation (LT) in human immunodeficiency virus (HIV)-positive individuals is considered to be an experimental therapy with limited reported worldwide experience, and little long-term survival data. Published data suggest that the short-term outcome is encouraging in selected patients. Here, we report our experience in 14 HIV-infected liver allograft recipients, and compare outcomes between those coinfected with hepatitis C virus (HCV) and the non-HCV group. A total of 14 HIV-infected patients (12 male, 2 female, age range 26-59 years) underwent LT between January 1995 and April 2003. Indications for LT were HCV (n = 7), hepatitis B virus (HBV; n = 4), alcohol-induced liver disease (n = 2), and seronegative hepatitis (n = 1); 3 patients presented with acute liver failure. At LT, CD4 cell counts (T-helper cells that are targets for HIV) ranged from 124 to 500 cells/μL (mean 264), and HIV viral loads from <50 to 197,000 copies/mL. Nine of 12 patients were exposed to highly active antiretroviral therapy (HAART) before LT. In the non-HCV group (n = 7), all patients are alive, all surviving more than 365 days (range 668-2,661 days). No patient has experienced HBV recurrence, and graft function is normal in all 7 patients. However, 5 of 7 HCV-infected patients died after LT at 95-784 days (median 161 days). A total of 4 patients died of complications due to recurrent HCV infection and sepsis, despite antiviral therapy in 3 of them. A total of 3 patients experienced complications relating to HAART therapy. In conclusion, outcome of LT in HIV-infected patients with HBV or other causes of chronic liver disease indicates that LT is an acceptable therapeutic option in selected patients. However, longer follow-up in larger series is required before a conclusive directive can be provided for HCV / HIV coinfected patients requiring LT. (Liver Transpl 2004;10:1271–1278.)
In the developed world, human immunodeficiency virus (HIV)-related morbidity and mortality have been considerably reduced since the introduction of highly active antiretroviral therapy (HAART).1, 2 The increased survival associated with HAART has exposed the role that liver disease plays in causing significant morbidity in HIV-infected persons,3, 4 to the extent that end-stage liver disease is now the leading cause of death in hospitalized HIV-infected persons.5, 6 Because HIV, hepatitis B virus (HBV), and hepatitis C virus (HCV) infections share some common transmission pathways, coinfection with HCV or HBV is common in HIV-infected persons, particularly in patients with hemophilia7, 8 and injection drug use.9 It is estimated that 30% of HIV-infected persons are coinfected with HCV; that estimate rises to 50 to 80% of HIV-positive intravenous drug users.10, 11 Simultaneous coinfection with HIV and HCV results in more rapid progression to cirrhosis,3, 9, 12 liver failure, and hepatocellular carcinoma.13 HCV-related liver disease is now the leading non–acquired immunodeficiency syndrome cause of death in HIV-infected persons in the developed world. It is therefore likely that HIV-infected persons will develop end-stage liver disease before they experience life-threatening complications related to HIV infection. As a result, the cohort of HIV-infected patients meeting current criteria for liver transplantation will increase rapidly over the next decade.
Liver transplantation (LT) is a proven therapy for a wide range of acute and chronic liver diseases. Concerns regarding the cumulative effect of viral and pharmacologic immunosuppression, and doubts about the appropriateness of utilizing a scarce resource in an area of unproven clinical efficacy, have severely limited the access of HIV-infected individuals to transplant programs.14 The limited data available from pilot studies suggest that HIV infection does not adversely affect the success of LT.15–19 Recent literature have provided ethical arguments in favor of viewing LT in HIV-infected persons as analogous to LT in persons with other chronic illnesses, and have proposed that the major liver transplant centers in the United States not consider LT in HIV-infected persons as experimental.20–22 It is therefore timely to report our experience of liver transplantation in HIV-positive patients and highlight the clinical challenges that have been encountered to date in this cohort.
LT; liver transplantation; HCV; hepatitis C virus; HBV; hepatitis B virus; HIV; human immunodeficiency virus; CD4, T-helper cells that are targets for HIV; HAART; highly active antiretroviral therapy; 3TC, lamivudine; EFV, efavirenz; TEN, tenofovir; d4T, stavudine; AZT, zidovudine; dDI, didanosine; NEV, nevirapine; KAL, kaletra.
Materials and Methods
A total of 14 HIV-infected patients (12 males, 2 females, age range 26-59 years) have received liver allografts in the King's College Hospital program between 1995 and April 2003, representing 1.1% of the activity over that period. The etiology of liver disease was HCV-related cirrhosis (n = 7), HBV-related cirrhosis (n = 2), acute liver failure due to HBV (n = 2), seronegative hepatitis (n = 1), and alcohol-induced liver disease (n = 2). Ten of 11 patients with cirrhosis had refractory ascites, spontaneous bacterial peritonitis, and / or encephalopathy, and had impaired synthetic function (Child-Turcotte-Pugh scores 9-12). The remaining patient with Child's A cirrhosis underwent liver transplantation for hepatopulmonary syndrome. All hepatitis surface antigen–positive candidates were HBV-deoxyribonucleic acid–negative prior to LT; 2 were hepatitis S e antigen–positive. In 1 patient with nonviral acute liver failure, serologic evidence of chronic HBV carriage was documented pretransplantation (i.e., hepatitis surface antigen–positive, hepatitis S e antigen–negative, hepatitis B e antibody–positive). Risk factors for HCV acquisition were previous injection drug use in 2 patients and blood product transfusion in the other 3 patients, who were also hemophilic. All 7 recipients had genotype 1 HCV infection.
HIV infection was diagnosed immediately prior to LT in the 3 patients who presented with acute liver failure; HIV infection had been established for 3 to 18 years before LT in the remaining 9 patients. Prior to transplantation assessment, 2 patients had experienced opportunistic infection with documented pneumocystis carini pneumonia at 12 and 15 months before LT, but responded to high dose co-trimoxazole. CD4 cell counts (T-helper cells that are targets for HIV) ranged from 124 to 500 cells/μL (mean 264, median 258), and HIV viral loads from <50 to 197,000 copies/mL before LT. Eleven of 14 patients received HAART before LT; drug regimens and patient demographics are listed in Table 1. All patients received HAART following LT (Table 2).
Table 1. Demographic Details of HIV-Infected Liver Transplant Recipients
Immunosuppression consisted of tacrolimus-based dual therapy with prednisolone in 13 patients, 1 patient received cyclosporin and azathioprine for 7 months but was subsequently changed to tacrolimus because of rejection. Doses were adjusted to maintain desired blood levels (tacrolimus 5-15 ng/ml, cyclosporin 100-250 ng/ml; trough samples taken 12 hours postdose) for the first year. Prednisolone was commenced at a daily dose of 20 mg reducing by 5 mg every week after the second week, and withdrawn completely at a median of 4 months. Only 1 patient received azathioprine, which was commenced at a dose of 1 mg/kg. Patients with moderate or severe cellular rejection received a 3-day course of intravenous methylprednisolone 1 gm daily. OKT3, ALG, and ATG were not used during the study period in this cohort. All patients received anti-fungal, anti–pneumocystis carini pneumonia, and anti-cytomegalovirus prophylaxis after transplantation. All recipients received a whole graft.
The protocol for prevention of HBV recurrence after LT was as follows: intraoperative administration of 10,000 IU of hepatitis B immunoglobulin intravenously during the anhepatic phase followed by 5,000 IU intravenously for the first 7 postoperative days and thereafter to maintain hepatitis B surface antibody titer above 300 IU during the first postoperative month. Immunoprophylaxis was continued indefinitely with monthly administration of 2,000 to 3,000 IU of hepatitis B immunoglobulin by intramuscular injection to maintain hepatitis B surface antibody titer greater than 100 IU. From 1998, all patients transplanted for HBV infection received dual immunoprophylaxis with hepatitis B immunoglobulin and lamivudine.
Statistical analysis of survival was performed with SSPS version 10 (SPSS, Chicago, IL). Differences in survival between subgroups of patients were tested by log-rank test. Comparisons were made between HIV-positive and HIV-negative HBV-infected patients, HIV-positive and HIV-negative HCV-infected patients, and HIV-positive and HIV-negative allograft recipients during the study period.
All patients survived beyond 30 days following LT. The actuarial 1-, 2-, and 5-year survival rates for the HIV cohort were 78.6, 69.8, and 62.1%, respectively (Fig. 1). All HIV-infected recipients with HBV coinfection and those HIV-infected recipients transplanted for nonviral liver disease are alive, the longest surviving patient is now 88 months after LT (P = .0011) (Fig. 2). Consequently, the 1-, 2-, and 5-year actuarial survival for HBV coinfected vs. HIV-negative HBV-infected recipients was 100 and 86.4%, 100 and 82%, and 100 and 80.4%, respectively. However, inferior survival rates were seen for patients with HCV coinfection (n = 7). At 12 months, 4 of 7 (57.1%) of those coinfected with HCV were still alive, but by 25 months a further 2 had died (Fig. 2). The survival rate of HCV / HIV coinfected persons was clearly lower than in the 182 HCV monoinfected candidates who received organs during the same period. The latter had actuarial 1- and 2-year survival rates of 87.5 and 83.9%, respectively. The 2 surviving patients with HCV are the most recent organ recipients and neither has significant hepatitis C–related disease at 1 year posttransplant.
Causes of death in those who died included ruptured cerebral AV malformation in 1 patient on a background of HCV recurrence and allograft dysfunction, while 3 HCV-infected recipients died of complications relating to recurrent HCV infection with associated graft dysfunction, septicemia and multiorgan failure. One patient died at 3 months from septicemia and allograft failure unrelated to HCV recurrence in the allograft. Two patients had received preemptive therapy from 2 weeks (interferon-alfa 3 MU tiw and ribavirin) and 3 weeks (pegylated interferon-alfa 180 μg weekly) after LT, but had histological evidence of recurrent HCV infection at 176 and 130 days, respectively. A further patient with histologic evidence of recurrent HCV infection at 179 days received interferon and ribavirin therapy at standard doses but therapy was withdrawn due to laboratory abnormalities. The fourth patient had HCV recurrence documented in the allograft at 63 days and died 3 weeks later from sepsis. None of the 5 HCV-infected individuals who died had pre-LT CD4 cell counts below 150 × 106/L or acquired immunodeficiency syndrome–defining illnesses, and had stable HIV infection on HAART. In contrast, no patients with HBV coinfection have experienced HBV recurrence, and graft function is normal in 6 of 7 HIV-infected recipients without HCV infection at 12 to 88 months after LT.
Five of 14 patients (35.7%; 2 HCV-coinfected patients: patient 10 and 12, Table 2) experienced at least 1 episode of rejection 5 to 34 days after LT, comparable to immunocompetent patients transplanted during the same period. In 3 cases, histologic analysis confirmed mild rejection (rejection activity index = 3)23 within the first 8 days of surgery, and the patients responded to a standard course of methylprednisolone therapy (1 gm daily for 3 days). Two patients transplanted for alcohol-related liver disease failed to respond to standard therapy (rejection activity index = 5) and had histologic evidence of ongoing cellular rejection (rejection activity index = 7), which required a further course of methylprednisolone and augmentation of baseline immunosuppression with mycophenolate before graft function returned to normal. No patient has had histologic evidence of chronic rejection. Infection complications included respiratory tract infections (n = 5), wound infection (n = 1), line-related infection (n = 2), and cytomegalovirus viremia in 1 patient; all infective episodes responded to standard antimicrobial treatment. There were no fungal infections.
Three patients experienced complications attributed to HAART. One patient taking dDI, d4T, efavirenz, and tenofovir experienced mitochondrial toxicity at 5 months after LT, as evidenced by hyperlactatemia (8.3) and 80% steatosis on liver biopsy. dDI was withdrawn without significant improvement in graft function or serum lactate levels; all antiretroviral drugs were subsequently withdrawn. He subsequently experienced a generalized seizure, and computerized tomography revealed a large occipitoparietal hemorrhage, from which he did not recover. A second patient experienced high serum lactate levels in the immediate postoperative period for 7 days. During this time all HAART drugs were withdrawn, and lactate levels returned to within the normal range. Reintroduction of antiretrovirals was uneventful. A third patient experienced zidovudine-related rhabdomyolysis 3 weeks following LT, but recovered fully following withdrawal of zidovudine. Adverse interactions between HAART and calcineurin inhibitors were not encountered, but 2 patients on PIs required markedly reduced tacrolimus doses (1 mg at 2-4 week intervals) to maintain levels in the therapeutic range.
Our data clearly indicate that LT in HIV-infected persons is feasible and effective in selected patients. Patients transplanted for HBV-induced and alcohol-related cirrhosis had a 1-year actuarial survival of 100%, and survival of HBV coinfected individuals was not significantly different to HIV-negative HBV recipients. The longest surviving HIV-infected patient transplanted for HBV-related liver disease is now 88 months post-LT, with stable CD4 cell counts and an undetectable HIV viral load. These observations suggest that HIV infection per se is not a barrier to LT. Further evidence in support of this statement is provided by rejection and infection rates that were no different in the HIV-infected cohort compared to immunocompetent recipients in the follow-up period. In addition, no specific procedural alterations were required during transplant surgery of HIV-positive patients compared to HIV-negative patients. In light of these results, we strongly endorse the argument that HIV infection should not be an absolute contraindication to liver transplantation.
In the pre-HAART era, patients with HIV infection diagnosed prior to or acquired as a consequence of transplantation had poorer survival rates. In 1 French series,24 7 of 9 HIV-positive individuals died at 2 to 55 months post-LT, 2 as a consequence of acquired immunodeficiency syndrome. In a report on LT in 6 HCV-infected HIV-positive hemophiliacs in the United Kingdom,25 survival was significantly worse than in HIV-negative hemophiliacs, and 3 of 6 patients died of acquired immunodeficiency syndrome–related illnesses. In a series of 15 liver recipients transplanted between 1981 and –1988,26 8 had died before publication, while 7 were alive at a mean follow-up of 4.5 years. Our series differs from these earlier reports, principally in that our patients did not die of HIV-related illnesses but mainly as a consequence of HCV recurrence. The clinical pattern was one of rapid progression to cholestatic allograft dysfunction and multiorgan failure, despite preemptive antiviral therapy in 2 individuals. This early experience, with a probable “learning-curve” effect, triggered a reevaluation of the role of liver transplantation in HIV / HCV patients, leading to modifications in protocols for immunosuppression and antiviral therapy. The 2 patients subsequently transplanted are currently doing well but longer follow-up is required and it is premature to comment on the role of protocol modifications. This later experience falls more in line with data from the United States. A recent retrospective review of 16 HIV-infected liver allograft recipients (11 of which who were coinfected with hepatitis C) from the University of Miami and University of Pittsburgh transplantation programs19 has reported a 2-year actuarial survival of 80%; however, in this group 90% of recipients have experienced histologic HCV recurrence and the longer-term outcome for these patients will be of considerable importance when evaluating the size of the challenge.
None of the HBV coinfected patients in our series experienced recurrent HBV infection following transplantation. Since the introduction of intra- and postoperative HBV immunoglobulin as passive immunoprophylaxis in the 1980s,27 and subsequently the introduction of lamivudine,28, 29 HBV recurrence rates have fallen dramatically with an attending increase in both short-term and long-term survival. In the nontransplant setting, the clinical course of chronic HBV infection does not seem to be significantly altered by coinfection with HIV, and studies have indicated that alanine aminotransferase levels and HBV-deoxyribonucleic acid titers are similar in HIV-positive and HIV-negative persons infected with HBV.30, 31 Although the numbers in our series are small, the same seems to hold true for HBV-infected HIV-positive liver recipients over 3 to 6 years follow-up. Concerns regarding HBV recurrence as a consequence of lamivudine resistance should be alleviated by emerging new therapies such as adefovir, dipivoxil, and tenofovir.32
Specific characteristics in the pretransplant period that may have adversely affected outcome in our cohort were not identified: renal function was normal in all HCV coinfected patients at transplant (mean serum creatinine 69 μmol/L), HCV viral load pretransplant was not available for 3 of the 5 patients, and consequently correlation with survival was not possible. Child-Turcotte-Pugh score in these patients (range 9-12) was similar to HIV-negative HCV-positive candidates. Following transplantation, the possibility of HAART-related hepatotoxicity as a contributing factor to graft dysfunction was considered. Considerable data are available that suggest that 10 to 20% of HIV-infected patients receiving HAART experience hepatotoxicity, and that HCV coinfected patients have a higher risk of this complication.33 Recent data from the United States indicate that patients with antiretroviral tolerance after LT have poorer survival; this may occur more readily in the presence of HCV recurrence in the allograft.16 In our current series, histologic examination of allograft biopsies in 4 of 5 HCV coinfected patients did not confirm features of HAART hepatotoxicity or intolerance (e.g., microvesicular steatosis, oxyphilia), and serum lactate levels were less than 2. In 1 patient, HAART-related mitochondrial injury was suspected, and mitochondrial abnormalities were subsequently confirmed by ultrastructural study at 5months after surgery (patient 8, Table 1). This was attributed to dDI. Another patient experienced high lactate levels within 8 days of transplantation associated with poor allograft dysfunction. HAART was withdrawn and reintroduced when liver function tests normalized, with no subsequent elevation in lactate levels. However, it is possible that HAART-induced hepatotoxicity may have contributed to graft dysfunction in our HCV coinfected patients. At least 2 of the 5 HCV coinfected patients who died after LT had received zidovudine, reflecting antiretroviral therapy prescribing practice at the time they underwent transplantation. These patients would now be deemed to have received inadequate antiretroviral therapy. It is possible that inadequate HIV treatment and therefore immunosuppression related to HIV may have led to more aggressive recurrence of HCV infection. However, any effect on outcome is likely to reflect efficacy of the combination and not an effect of an individual drug. As more HIV-infected patients are offered transplantation, HAART-related hepatotoxicity or HAART intolerance may emerge as a more significant problem than reported heretofore. However, 1 recent series recorded this complication in only 1 in 16 (6.2%) LT recipients,19 and recent data have reconfirmed that the benefits of HAART significantly outweigh associated risks of severe hepatotoxicity.34
Pharmacologic interactions between calcineurin inhibitors and protease inhibitor-containing HAART regimens have been documented, and up to 50-fold reductions in calcineurin inhibitor dosage have been reported in HIV-infected recipients of solid organ allografts.35, 36 In contrast, patients prescribed NNRTI-based regimens have not required significant alteration in cyclosporin dosing. Mycophenolate mofetil, is hydrolyzed in vivo to its active metabolite mycophenolic acid, a potent selective noncompetitive and reversible inhibitor of inosine monophosphate dehydrogenase, an enzyme involved in the synthesis of dGTP. By reducing the intracellular concentration of dGTP, mycophenolate mofetil augments the antiretroviral effect of abacavir, tenofovir, and didanosine against HIV-1,37 and consequently, their potential toxicities may be augmented by coadministration of mycophenolate mofetil following LT. In our series, tacrolimus dose reductions were not observed in the first 4 HCV coinfected recipients transplanted before 2001. However, the 2 surviving HCV coinfected recipients are receiving 1 mg tacrolimus every 2 to 4 weeks to achieve therapeutic levels, representing a 65% reduction in tacrolimus dosage. It is possible that inadvertent overimmunosuppression in those transplanted before 2001 may have contributed to HCV recurrence and poor outcome, but the expected nephrotoxicity expected with calcineurin overexposure was not observed.
These results might be interpreted as providing evidence that LT should not be offered to HCV-infected HIV-positive individuals. However, this needs to be put into context, as recurrent hepatitis C is now regarded as 1 of the most important problems after liver transplant. HCV viremia persists in the majority of recipients and allograft reinfection has been described as early as 4 weeks after surgery.38 Recurrent HCV infection is a significant cause of allograft dysfunction and progression to cirrhosis has been documented in 8 to 28% of reported series to date.38–40 Studies on the use of antiviral therapy to prevent or treat recurrent HCV infection have achieved sustained suppression of viral replication in only 8 to 27% of recipients,41–43 suggesting that the majority of HCV-positive recipients will experience graft dysfunction of varying severity.
Unlike the slow natural course of disease observed in immunocompetent HCV-infected persons, allograft failure as a consequence of HCV recurrence occurs at an accelerated rate, and it has become clear that immunosuppression plays a role. Steroids in particular have been implicated in this process.44, 45 In the nontransplant setting, immunosuppression also potentiates fibrosis progression.11, 46 It could therefore be anticipated that long-term survival in HCV / HIV-infected patients would be doubly compromised. Given the concerns relating to immunosuppression-related HCV recurrence, immunosuppression in HCV coinfected patients has evolved to steroid-free therapy, with primary immunosuppression consisting of tacrolimus and mycophenolate. Whether this will have an advantageous effect on long-term survival is currently unknown. Other factors that may have a negative impact on rate of histologic HCV recurrence and patient survival after LT include advancing donor liver age.47, 48 In our cohort, however, the median donor liver age received by HCV-infected patients was less than non-HCV-infected patients (32 years vs. 43 years, respectively), and was therefore unlikely to have played a role in the reduced survival observed in the former group.
The characteristics of HIV-infected patients who are suitable for transplantation have not been precisely defined. Factors that were associated with successful outcome in our series include stable CD4 cell counts on HAART (≥150 cells/mm3) and no history of acquired immunodeficiency syndrome–defining illnesses. The one exception here may be pneumocystis carini pneumonia infection, as improvements in therapy have downgraded the negative impact of this opportunistic infection. None of our patients had significant antiretroviral drug resistance profiles, allowing greater clinical freedom in substituting anti-HIV therapies in the presence of drug interactions or drug-induced complications. A further factor contributing to successful transplantation beyond the first 30 days after surgery was the lack of significant malnutrition in all candidates prior to the transplant procedure (<10th centile). In addition, the indication for liver transplantation in our series was morbidity related to their liver disease, and not imminent death.
HIV infects approximately 750,000 to 1.5 million individuals in the United States alone, with 40,000 additional cases diagnosed each year. Current figures indicate that 240,000 HIV-infected individuals are coinfected with HCV in the United States.49 If 7 to 25% of HCV coinfected patients progress to cirrhosis within 10 to 15 years, as has been reported,3, 4, 9 we can anticipate that 16,800 to 60,000 HIV-positive persons will become liver transplant candidates over the course of their lives. This would have a major impact on a resource that is already oversubscribed: expansion of waiting lists, prolonged waiting times for LT, and a rise in deaths on the LT list. However, equal access to organs among patients with equal need is a policy to which all practitioners subscribe. If transplantation provided substantially less benefit in terms of survival and quality of life for HIV-positive individuals, or a subgroup of HIV-infected individuals, it would be difficult to argue in favor of continuing to offer organs to this cohort. The wider view of nontransplant physicians, the public, and third party insurers also need to be considered. It seems, therefore, imperative that a register be developed nationally to facilitate the documentation of all relevant information pertaining to liver transplantation in HIV-infected patients. It also behooves us to develop new trials of antiviral therapy in HCV-infected patients.