Outcomes of liver transplant recipients with hepatitis C and human immunodeficiency virus coinfection§

Authors


  • Burc Barin and Donald M. Stablein have full access to all data in this study and take responsibility for the integrity of the data and the accuracy of the data analysis. Norah A. Terrault, Michelle E. Roland, Margaret V. Ragni, John Fung, Donald M. Stablein, Lawrence Fox, Jonah Odim, and Peter G. Stock contributed to the study concept and design. Norah A. Terrault, Michelle E. Roland, Thomas Schiano, Lorna Dove, Michael T. Wong, Fred Poordad, Margaret V. Ragni, David Simon, Kim M. Olthoff, Lynt Johnson, Valentina Stosor, Dushyantha Jayaweera, John Fung, Kenneth E. Sherman, Aruna Subramanian, J. Michael Millis, Douglas Slakey, Carl L. Berg, Laurie Carlson, Linda Ferrell, and Peter G. Stock contributed to the acquisition of data. Norah A. Terrault, Michelle E. Roland, Thomas Schiano, Lorna Dove, Michael T. Wong, Fred Poordad, Margaret V. Ragni, Burc Barin, David Simon, Kim M. Olthoff, Lynt Johnson, Valentina Stosor, Dushyantha Jayaweera, John Fung, Kenneth E. Sherman, Aruna Subramanian, J. Michael Millis, Douglas Slakey, Carl L. Berg, Laurie Carlson, Linda Ferrell, Donald M. Stablein, Lawrence Fox, Jonah Odim, and Peter G. Stock contributed to the interpretation of data. Norah A. Terrault, Michelle E. Roland, Burc Barin, Donald M. Stablein, and Peter G. Stock contributed to the drafting of the manuscript. Norah A. Terrault, Michelle E. Roland, Thomas Schiano, Lorna Dove, Michael T. Wong, Fred Poordad, Margaret V. Ragni, Burc Barin, David Simon, Kim M. Olthoff, Lynt Johnson, Valentina Stosor, Dushyantha Jayaweera, John Fung, Kenneth E. Sherman, Aruna Subramanian, J. Michael Millis, Douglas Slakey, Carl L. Berg, Laurie Carlson, Linda Ferrell, Donald M. Stablein, Lawrence Fox, Jonah Odim, and Peter G. Stock contributed to the critical revision of the manuscript for important intellectual content. Norah A. Terrault, Michelle E. Roland, Burc Barin, Donald M. Stablein, and Peter G. Stock contributed to the statistical analysis. Michelle E. Roland and Peter G. Stock obtained funding. Norah A. Terrault, Michelle E. Roland, Laurie Carlson, Jonah Odim, Lawrence Fox, and Peter G. Stock contributed to the supervision of the study.

  • All the authors completed and submitted copyright assignment, authorship responsibility, National Institutes of Health funding, financial disclosure, institutional review board/animal care committee approval, and sponsorship forms, and no conflicts were reported.

  • §

    The Committee on Human Research at the University of California San Francisco approved the study protocol, as did the institutional review board of each participating center. Each participant provided written informed consent.

  • Potential Conflicts of Interest: Dr. Schiano consults for Vertex, Merck, Salix, and Gilead. Dr. Poordad advises and received grants from Abbott, Anadys, Achillion, and Tibotec. He advises, is on the speakers' bureau for, and received grants from Gilead, Genentech, Merck, and Vertex. He is also on the speakers' bureau for Salix and Onyx and received grants from Bristol Meyers Squibb, Pharmasset, and Boehringer Ingleheim. Dr. Jayaweera consults, advises, and is on the speakers' bureau for Gillead and consults, advises, and received grants from Viiv. He consults, advises, is on the speakers' bureau for, and received grants from Bristol Meyers Squibb and Tibotec. He also received grants from Vertex. Dr. Sherman advises Bristol Meyers Squibb, Glaxo Smith Kline, Baxter, Regulus, and Fibrogen, and received grants from Roche, Gilead, Siemens, Anadys, and Pharmasset. He advises and received grants from Merck, Vertex, Boehringer Ingleheim, and SciClone

Abstract

Hepatitis C virus (HCV) is a controversial indication for liver transplantation (LT) in human immunodeficiency virus (HIV)–infected patients because of reportedly poor outcomes. This prospective, multicenter US cohort study compared patient and graft survival for 89 HCV/HIV-coinfected patients and 2 control groups: 235 HCV-monoinfected LT controls and all US transplant recipients who were 65 years old or older. The 3-year patient and graft survival rates were 60% [95% confidence interval (CI) = 47%-71%] and 53% (95% CI = 40%-64%) for the HCV/HIV patients and 79% (95% CI = 72%-84%) and 74% (95% CI = 66%-79%) for the HCV-infected recipients (P < 0.001 for both), and HIV infection was the only factor significantly associated with reduced patient and graft survival. Among the HCV/HIV patients, older donor age [hazard ratio (HR) = 1.3 per decade], combined kidney-liver transplantation (HR = 3.8), an anti-HCV–positive donor (HR = 2.5), and a body mass index < 21 kg/m2 (HR = 3.2) were independent predictors of graft loss. For the patients without the last 3 factors, the patient and graft survival rates were similar to those for US LT recipients. The 3-year incidence of treated acute rejection was 1.6-fold higher for the HCV/HIV patients versus the HCV patients (39% versus 24%, log rank P = 0.02), but the cumulative rates of severe HCV disease at 3 years were not significantly different (29% versus 23%, P = 0.21). In conclusion, patient and graft survival rates are lower for HCV/HIV-coinfected LT patients versus HCV-monoinfected LT patients. Importantly, the rates of treated acute rejection (but not the rates of HCV disease severity) are significantly higher for HCV/HIV-coinfected recipients versus HCV-infected recipients. Our results indicate that HCV per se is not a contraindication to LT in HIV patients, but recipient and donor selection and the management of acute rejection strongly influence outcomes. Liver Transpl 18:716–726, 2012. © 2012 AASLD.

See Editorial on Page 617

With dramatic improvements in the prognosis of individuals living with human immunodeficiency virus (HIV) on long-term antiretroviral therapy, end-stage liver disease has emerged as an important cause of morbidity and mortality.1, 2 Historically, HIV infection was viewed as a contraindication to liver transplantation (LT) because of concerns about immunosuppression-related opportunistic infections and reduced survival, but this is no longer the case. Prior studies, have demonstrated acceptable short-term survival and no increased risk of HIV-related complications.3 However, the results for HIV-infected LT recipients with hepatitis C virus (HCV) are generally worse than those for recipients without HCV,4-7 and this makes HCV a controversial indication for LT in HIV-infected patients. Natural history studies of HCV-infected patients with HIV have demonstrated an accelerated rate of hepatic fibrosis8-10 and a higher rate of complications of cirrhosis11 in comparison with HCV-infected patients without HIV. After LT, patients with HCV monoinfection also experience an accelerated rate of disease progression, and recurrent HCV cirrhosis is the leading cause of graft loss among LT recipients.12 Thus, HCV/HIV-coinfected transplant recipients may be at especially high risk of recurrent cirrhosis after transplantation. We conducted a prospective, multicenter US cohort study of HCV/HIV-coinfected LT recipients to define the natural history of the disease and predictors of graft survival and HCV disease recurrence. We hypothesized that the outcomes of HCV/HIV-coinfected patients would be similar to the outcomes of other higher risk LT recipients without HIV, such as recipients who are 65 years old or older.

Abbreviations:

AIDS, acquired immune deficiency syndrome; BMI, body mass index; CI, confidence interval; CSA, cyclosporine A; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; HIV, human immunodeficiency virus; HR, hazard ratio; IQR, interquartile range; LT, liver transplantation; MELD, Model for End-Stage Liver Disease; MMF, mycophenolate mofetil; NNRTI, nonnucleoside reverse transcriptase inhibitor; PI, protease inhibitor; SRTR, Scientific Registry of Transplant Recipients; TAC, tacrolimus.

PATIENTS AND METHODS

Study Population

This is a prospective cohort study of 89 HIV-infected patients with chronic HCV infections and complications of end-stage liver disease who underwent transplantation at 17 US centers between October 2003 and February 2010 (ClinicalTrials.gov number NCT00074386). The HIV-specific inclusion criteria for transplantation were previously described.3 No HCV-specific selection criteria were used. All patients with hepatocellular carcinoma (HCC) were within the Milan criteria. The research protocol was approved and monitored by the institutional review boards of all participating centers and the data and safety monitoring board of the National Institute of Allergy and Infectious Diseases. Each patient provided written informed consent to participate in the trial.

Two non-HIV control groups were used for analytical purposes. The first control group consisted of HCV-monoinfected LT recipients (up to 3 recipients per HIV/HCV-coinfected subject) who were matched by the study clinical site, the calendar time, and 2 additional clinical parameters: (1) a single transplant versus a combined kidney-liver transplant and (2) HCC. This control group was used for comparisons of survival, rejection, and HCV recurrence. The second control group, which was used for survival analyses only, included all older LT recipients (≥65 years) from the US Scientific Registry of Transplant Recipients (SRTR) with a data cutoff date of August 23, 2010. This group was selected to represent a patient population potentially similar to HIV patients in that LT is offered selectively because of reduced patient and graft survival.

Posttransplant Follow-Up in the HCV/HIV Cohort

A standardized protocol for posttransplant follow-up was used. Liver biopsy samples were obtained at least annually to assess HCV disease severity, and they were read centrally and scored with the Ludwig-Batts method.13 Additionally, liver biopsy was performed for abnormal liver tests, suspected rejection, or suspected drug hepatotoxicity. The number of central biopsy readings performed in the first, second, and third years after transplantation was 203, 33, and 15, respectively.

Antiretroviral therapy was reinitiated after transplantation when the patient was able to take oral agents and renal function and clinical status were stable. Antiretroviral therapy was determined by the HIV provider in consultation with transplant team physicians who had expertise in HIV management. Most patients continued their pretransplant regimen in the posttransplant period.

The prophylaxis for opportunistic infections was previously described.3 Immunosuppression was not standardized, but the study protocol strongly recommended specific immunosuppressive drugs, and adherence to those recommendations was high. Maintenance immunosuppression included a calcineurin inhibitor [cyclosporine A (CSA) or tacrolimus (TAC)], mycophenolate mofetil (MMF), and prednisone. Because of potential anti-HIV effects of CSA, there was more use of CSA in HCV/HIV patients versus HCV-monoinfected controls. Sirolimus was used at the discretion of the investigator and was typically part of a renal-sparing drug strategy. Target trough immunosuppressive levels were site-specified. Induction with an interleukin-2 receptor inhibitor (anti-CD25 antibody) or thymoglobulin was not recommended, and alemtuzumab use was prohibited because of the risk of infectious complications and severe HCV recurrence.14, 15 Health care providers followed calcineurin inhibitor dosing guidelines for patients on protease inhibitors (PIs) or on combined regimens based on PIs and nonnucleoside reverse transcriptase inhibitors (NNRTIs).16 Calcineurin inhibitor and sirolimus doses were adjusted to obtain therapeutic trough serum levels. Steroid induction, tapering, and maintenance were performed according to local site practice. The treatment of acute rejection was dictated by local site protocols, and all treated rejection episodes were biopsy-proven. Acute rejection was treated with the amplification of the baseline immunosuppression (eg, a change in the calcineurin inhibitor or the addition of MMF) or the administration of corticosteroids. The use of OKT3 and polyclonal anti-lymphocyte preparations was restricted to patients with severe rejection.

Follow-Up in the HCV-Monoinfected Control Group

Controls were identified retrospectively within strata as the next HCV-infected LT recipients in calendar time after HCV/HIV-coinfected LT recipients. Donor information was obtained from SRTR; recipient information and posttransplant complications, including HCV disease severity by liver biopsy, the treatment of HCV and acute cellular rejection, and the causes of graft loss and patient death, were obtained by medical chart review. Biopsy samples were not read centrally but were scored according to a histopathological review by the local pathologist for the presence or absence of severe HCV disease, which was defined as bridging fibrosis or cirrhosis (Ludwig-Batts stage F3-F4) or histological features of cholestatic hepatitis.17 The number of biopsy readings performed in the first, second, and third years after transplantation was 382, 136, and 60, respectively.

Statistical Methods

The primary study endpoints were patient and graft survival. Secondary endpoints included the rate of severe HCV disease (defined by the presence of cholestatic hepatitis or bridging fibrosis/cirrhosis or by graft loss due to HCV) and the rate of treated acute rejection. Other outcomes that were limited to the HCV/HIV group included the incidence of HIV-associated opportunistic complications and changes in CD4+ T cell counts.

Descriptive statistics were presented as medians, interquartile ranges (IQRs), and ranges as appropriate. A comparison of the baseline characteristics was conducted with Fisher's exact test (categorical variables) or the Wilcoxon rank-sum test (continuous variables). Changes in CD4+ T cell counts from the baseline were analyzed with the Wilcoxon signed-rank test. Estimated rates of patient survival, graft survival, graft rejection, and severe HCV recurrence were calculated with the Kaplan-Meier method, and 95% confidence intervals (CIs) were estimated with Greenwood's formula. Survival distributions were compared with the log-rank test. Predictors of outcomes were analyzed in Cox proportional hazards models; all variables from a univariate analysis with P < 0.10 were included in an initial multivariate model. Subsequently, variables with P ≥ 0.1 were excluded, the model was refitted, and all interactions were examined. Posttransplant characteristics were analyzed as time-dependent covariates. Select antiretrovirals previously linked to liver-related outcomes for nontransplant patients (PIs, nevirapine, didanosine, and stavudine) were evaluated as predictors of graft loss in HCV/HIV patients. Models were developed for the overall cohort and for the HCV/HIV group alone. Models for the overall cohort were stratified by the case-matched group, and the predictor variable of interest was HIV/HCV coinfection. A 2-sided P value < 0.05 was considered to indicate statistical significance. Statistical analyses were performed with SAS 9.2 (SAS Institute, Cary, NC).

RESULTS

Baseline Characteristics

In all, 89 HCV/HIV-coinfected patients and 235 HCV-monoinfected controls were included. Survivors were followed for a median of 2.7 years (IQR = 1.4-3.7 years) in the HCV/HIV-coinfected group and for a median of 2.4 years (IQR = 1.3-3.5 years) in the HCV control group (P = 0.45). When all subjects were included, the median follow-up was 2.0 years (IQR = 1.1-3.4 years) for HIV-negative subjects with HCV and 1.8 years (IQR = 0.7-3.4 years) for HIV-positive subjects with HCV (P = 0.51). Recipient, donor, and transplant-related characteristics are shown in Table 1. Compared to HCV controls, HCV/HIV-coinfected transplant recipients were younger, had a lower body mass index (BMI) at listing, more hepatitis B virus (HBV) coinfections, more donor allografts from non–heart-beating donors, longer median warm ischemia times, and less initial use of TAC-based immunosuppression (versus CSA-based immunosuppression). In the HIV/HCV group, the median CD4+ T cell count at LT was 283 cells/mm3 (IQR = 187-408 cells/mm3), and plasma HIV RNA was undetectable in 88% at the time of transplantation. The majority (97%) of coinfected patients were on antiretrovirals up to the time of LT, and 80% resumed antiretrovirals within the first postoperative week. Among the 8 HCV/HIV-coinfected recipients of combined kidney-liver transplants, the causes of renal dysfunction were diabetic nephropathy (n = 3), hypertensive nephrosclerosis (n = 2), HIV nephropathy (n = 1), acute renal failure (n = 1), and hepatorenal syndrome (n = 1). Two patients were on long-term dialysis with compensated cirrhosis; the remainder had decompensated cirrhosis with acute or acute-on-chronic kidney disease.

Table 1. Characteristics of HCV/HIV-Coinfected Transplant Recipients and HCV-Monoinfected Transplant Recipients
 HCV/HIV (n = 89)HCV (n = 235)P Value
  • *

    The data are presented as medians and IQRs.

  • Most recent pretransplant value within 16 weeks of transplantation.

  • Most recent pretransplant therapy. The PI, NNRTI, PI and NNRTI, and nucleoside-only regimens included at least 2 nucleoside analogues (except for 7 regimens that included a single nucleoside analogue). Eighteen of the 22 NNRTI-based regimens included efavirenz, 3 included nevirapine, and 1 included etravirine. Two PI-based regimens included raltegravir, and 1 PI-based regimen included enfuvirtide as well.

Recipient characteristics
 Age (years)*49 (44-53)54 (51-60)<0.001
 Male sex [n (%)]67 (75)166 (71)0.49
 Caucasian race [n (%)]58 (65)126 (54)0.08
 BMI at listing (kg/m2)*25 (23-28)28 (25-32)<0.001
 MELD score at LT*20 (15-27)20 (15-28)0.80
 HCC [n (%)]30 (34)73 (31)0.69
 HCV genotype: 1/4/other [n (%)]71 (80)192 (82)0.85
 HBV coinfection [n (%)]5 (6)2 (1)0.02
Donor/transplant characteristics
 Donor age (years)*37 (24-49)42 (30-53)0.07
 Black donor race [n (%)]10 (11)36 (15)0.38
 Living donor [n (%)]1 (1)15 (6)0.08
 Donor risk index*1.31 (1.15-1.62)1.42 (1.21-1.77)0.08
 Cold ischemia time (hours)*7.6 (5.8-9.9)8.0 (6.2-10.1)0.23
 Non–heart-beating donor [n (%)]15 (17)8 (4)<0.001
 Warm ischemia time (minutes)*41 (36-55)21 (13-26)0.001
 Anti-HCV–positive donor [n (%)]12 (13)26 (11)0.14
 HBV core antibody–positive donor [n (%)]12 (13)24 (10)0.16
 Combined kidney-liver transplant [n (%)]8 (9)17 (7)0.64
 Split/partial transplant [n (%)]1 (1)2 (1)>0.99
Posttransplant characteristics
 Initial calcineurin inhibitor use [n (%)]  <0.001
  CSA31 (35)23 (10) 
  TAC52 (58)188 (80) 
  None6 (7)9 (4) 
  Not reported0 (0)15 (6) 
 HCV treatment [n (%)]37 (42)57 (24)0.002
 Follow-up after LT (years)*1.8 (0.7-3.4)2.0 (1.1-3.4)0.51
HIV-specific characteristics
 CD4+ T cells (cells/mm3)*283 (187-408)  
 Undetectable HIV RNA [n (%)]78 (88)  
 On antiretrovirals within first week after LT [(n (%)]71 (80)  
 Antiretroviral therapy [n (%)]   
  PI45 (51)  
  NNRTI22 (25)  
  PI and NNRTI4 (4)  
  Nucleoside only15 (17)  
  None3 (3)  

Patient and Graft Survival

The 1- and 3-year patient survival rates were 76% (95% CI = 66%-84%) and 60% (95% CI = 47%-71%) for HCV/HIV patients and 92% (95% CI = 87%-95%) and 79% (95% CI = 72%-84%) for HCV patients (P < 0.001; Fig. 1A). Similarly, the graft loss rate was significantly higher for HCV/HIV patients versus HCV patients (P < 0.001; Fig. 1B). Patient/graft loss due to sepsis or multiorgan failure was more frequent for HCV/HIV-coinfected recipients versus HCV-infected recipients, and malignancy-related patient/graft loss was less frequent (Table 2). There were no deaths due to HIV-associated infections or malignancies in the HCV/HIV group. A pretransplant history of acquired immune deficiency syndrome (AIDS)–related opportunistic infections or neoplasms did not significantly affect posttransplant survival [hazard ratio (HR) = 1.0 (95% CI = 0.4-3.0), P = 0.96]. In the multivariate analysis, HIV infection was the only baseline factor significantly associated with an increased risk of death (HR = 2.3, P = 0.002) and graft loss (HR = 1.9, P = 0.01; Table 3). Among HCV/HIV-coinfected patients, the receipt of a combined kidney-liver transplant (HR = 3.8, P = 0.003), a BMI < 21 kg/m2 at enrollment (HR = 3.2, P = 0.01), the receipt of an anti-HCV–positive donor (HR = 2.5, P = 0.03), and older donor age (HR = 1.3 per decade, P = 0.04) were significant predictors of reduced graft survival (Table 4). A cytomegalovirus infection treated after transplantation (HR = 9.5, P < 0.001) was also significantly associated with an increased risk of graft loss in the univariate analysis, but it was not included in the multivariate model because there were only 5 cases in this group. Seven of 8 combined kidney-liver transplant recipients (88%) died; 6 of these 8 patients had a Model for End-Stage Liver Disease (MELD) score greater than 25 at LT. Seven of 10 LT recipients (70%) with a BMI < 21 kg/m2 at enrollment and 8 of 12 LT recipients (67%) with an anti-HCV–positive donor experienced graft loss.

Figure 1.

(A) Kaplan-Meier estimates of patient survival for HCV/HIV-coinfected recipients and HCV-monoinfected LT recipients. The 1- and 3-year patient survival rates were 76% (95% CI = 66%-84%) and 60% (95% CI = 47%-71%) for the HCV/HIV patients and 92% (95% CI = 87%-95%) and 79% (95% CI = 72%-84%) for the HCV patients (P < 0.001). (B) Kaplan-Meier estimates of graft survival for HCV/HIV-coinfected recipients and HCV-monoinfected LT recipients. The 1- and 3-year graft survival rates were 72% (95% CI = 61%-80%) and 53% (95% CI = 40%-64%) for the HCV/HIV patients and 88% (95% CI = 83%-91%) and 74% (95% CI = 66%-79%) for the HCV patients (P < 0.001). The rates of graft survival were calculated on the basis of graft failure for any cause. (C,D) Kaplan-Meier estimates of patient and graft survival for HCV/HIV-coinfected LT recipients by risk groups. The HCV/HIV-coinfected LT recipients were divided into 2 risk groups: a high-risk group, which included 25 subjects determined to have an increased risk of graft loss (enrollment BMI < 21 kg/m2, combined kidney-liver transplantation, or anti-HCV–positive donor), and a non–high-risk group, which included the remaining 64 subjects. The 1- and 3-year patient survival rates were 86% (95% CI = 75%-93%) and 72% (95% CI = 56%-83%) for the non–high-risk HCV/HIV patients and 51% (95% CI = 30%-68%) and 29% (95% CI = 9%-52%) for the high-risk HCV/HIV patients (P < 0.001). The 1- and 3-year graft survival rates were 84% (95% CI = 72%-91%) and 65% (95% CI = 49%-77%) for the non–high-risk HCV/HIV patients and 42% (95% CI = 23%-61%) and 20% (95% CI = 5%-43%) for the high-risk HCV/HIV patients (P < 0.001). However, the rates of patient survival and graft survival for the non–high-risk HCV/HIV group were similar to the SRTR-reported rates for older LT recipients (≥65 years, all indications for LT) and for all LT recipients in the United States during a similar time frame. The rates of graft survival were calculated on the basis of graft failure from any cause.

Table 2. Causes of Graft Loss for HCV/HIV-Coinfected Transplant Recipients and HCV-Monoinfected Transplant Recipients
 HCV/HIV (n = 38)HCV (n = 55)
  1. NOTE: There were no significant differences in causes of graft loss between HCV/HIV-coinfected transplant recipients and HCV-monoinfected transplant recipients (P = 0.25).

Recurrent HCV [n (%)]9 (24)12 (22)
Multiorgan failure/ sepsis [n (%)]12 (32)8 (15)
Postsurgical complications [n (%)]6 (16)11 (20)
Rejection [n (%)]3 (8)4 (7)
Malignancy [n (%)]1 (3)8 (15)
Other/unknown [n (%)]7 (18)12 (22)
Table 3. Predictors of Death and Graft Loss in HIV/HCV-Coinfected Transplant Recipients and HCV-Monoinfected Transplant Recipients
 DeathGraft Loss
HR (95% CI)P ValueHR (95% CI)P Value
  • NOTE: Adjustments were made for clustering.

  • *

    Time-varying covariate

Univariate Predictor    
 HIV coinfection2.3 (1.3-3.8)0.0022.1 (1.3-3.4)0.002
 Recipient age (years)1.0 (0.9-1.0)0.141.0 (0.9-1.0)0.08
 Recipient female sex1.0 (0.5-1.9)0.980.8 (0.5-1.5)0.52
 Recipient Caucasian race1.2 (0.6-2.2)0.621.1 (0.6-2.0)0.65
 HCV genotype: 1/4/other0.6 (0.3-1.5)0.290.6 (0.3-1.3)0.16
 BMI at enrollment < 21 kg/m22.4 (0.8-6.9)0.112.0 (0.8-5.3)0.16
 MELD score at LT > 251.0 (0.5-2.0)0.981.0 (0.5-1.9)0.98
 TAC versus CSA as initial immunosuppression0.8 (0.3-2.1)0.601.1 (0.4-2.7)0.88
 Deceased donor3.1 (0.7-14.8)0.151.8 (0.5-6.8)0.39
 Donor risk index1.2 (0.6-2.4)0.661.1 (0.6-2.2)0.75
 Donor age (by decade)1.0 (0.9-1.2)0.591.1 (0.9-1.2)0.46
 Non–heart-beating donor1.4 (0.4-4.3)0.612.1 (0.7-5.7)0.16
 Cold ischemia time (hours)1.1 (0.9-1.2)0.231.1 (1.0-1.2)0.23
 Donor black race0.7 (0.3-1.6)0.350.7 (0.3-1.6)0.37
 Anti-HCV–positive donor1.6 (0.6-4.3)0.321.1 (0.5-2.7)0.79
 Treated acute rejection*1.6 (0.8-3.3)0.232.4 (1.2-4.6)0.01
Multivariate Predictor    
 HIV coinfection2.3 (1.3-3.8)0.0021.9 (1.2-3.1)0.01
 Treated acute rejection*  2.0 (1.0-4.0)0.06
Table 4. Predictors of Graft Loss in HIV/HCV-Coinfected Transplant Recipients
 HR (95% CI)P Value
  • *

    Time-varying covariate.

Univariate Predictor  
 Recipient age (years)1.0 (1.0-1.1)0.63
 Recipient female sex0.8 (0.3-1.7)0.52
 Recipient Caucasian race0.9 (0.4-1.7)0.68
 Combined kidney-liver transplant3.7 (1.6-8.7)0.002
 BMI at enrollment < 21 kg/m23.2 (1.4-7.5)0.01
 Transplant MELD score > 251.6 (0.8-3.4)0.17
 Detectable HIV RNA at transplant2.6 (1.2-5.7)0.02
 History of splenectomy2.7 (0.8-8.8)0.10
 HCC0.7 (0.3-1.4)0.31
 TAC versus CSA as initial immunosuppression2.8 (1.3-6.2)0.01
 Donor risk index1.9 (0.9-4.1)0.08
 Donor age (by decade)1.3 (1.0-1.5)0.04
 Anti-HCV–positive donor2.7 (1.2-6.1)0.01
 Treated acute rejection*1.4 (0.7-2.9)0.36
 Treated cytomegalovirus infection*9.5 (3.1-29.4)<0.001
 PI use in previous 3 months*0.9 (0.5-1.7)0.70
 Stavudine/didanosine use in previous 3 months*0.3 (0.0-2.0)0.19
 Nevirapine use in previous 3 months*0.6 (0.1-4.2)0.58
 Pretransplant history of AIDS-related opportunistic   infections or neoplasms1.0 (0.4-3.0)0.96
Multivariate Predictor  
 BMI at enrollment < 21 kg/m23.2 (1.3-7.7)0.01
 Combined kidney-liver transplant3.8 (1.6-9.1)0.003
 Anti-HCV–positive donor2.5 (1.1-5.6)0.03
 Donor age (by decade)1.3 (1.0-1.6)0.04

The patient and graft survival rates for the high-risk HCV/HIV patients, who underwent combined kidney-liver transplantation, had a low BMI at enrollment (<21 kg/m2), or had an anti-HCV–positive donor, were significantly lower than the rates for the non–high-risk HCV/HIV patients without those 3 risk factors (P < 0.001; Fig. 1C,D). In contrast, the patient and graft survival rates of the non–high-risk HCV/HIV patients were not significantly different from the rates reported in the national SRTR database for older LT recipients (≥65 years) during a similar time frame (P = 0.67 and P = 0.79, respectively; Fig. 1C,D).

There were no significant changes in survival among the HIV/HCV transplant recipients over time. When the first half of the coinfected cohort was compared with the second half, no significant differences were observed in the patient survival curves (log-rank P > 0.99) or graft survival curves (log-rank P = 0.91). For the first half of the coinfected cohort, the 1-year patient and graft survival rates were 75% (95% CI = 60%-85%) and 71% (95% CI = 55%-82%), respectively, whereas for the second half of the coinfected cohort, the rates were 77% (95% CI = 61%-88%) and 73% (95% CI = 57%-84%), respectively.

HCV Disease Severity

Among recipients with functioning grafts, the median histological follow-up was 0.9 years (IQR = 0.2-1.8 years) for the HCV/HIV group and 1.1 years (IQR = 0.3-2.3 years) for the HCV group (P = 0.20). Sixty-two percent of the HCV/HIV-coinfected recipients and 64% of the HCV-infected recipients had 2 or more biopsy samples for review. The 1-, 2-, and 3-year cumulative rates of severe HCV disease were 17% (95% CI = 10%-28%), 27% (95% CI = 17%-40%), and 29% (95% CI = 19%-43%) for the HCV/HIV patients and 6% (95% CI = 3%-9%), 14% (95% CI = 9%-20%), and 23% (95% CI = 16%-31%) for the HCV patients (P = 0.21). HIV was not significantly associated with severe HCV disease (HR = 1.4, P = 0.41; Supporting Table 1). Recipient female sex (HR = 3.5, P = 0.01) and treated acute rejection (HR = 3.3, P = 0.01) were significantly associated with severe HCV disease, and older recipient age was protective (HR = 0.9, P = 0.046). The proportions of graft losses attributed to recurrent HCV disease were similar in the HCV/HIV and HCV groups (Table 2).

Significantly more HCV/HIV subjects (42%) than HCV/non-HIV subjects (24%) received HCV therapy after transplantation (P = 0.002), and approximately 85% of the HCV/HIV subjects had a fibrosis stage of 0 or 1 before the initiation of HCV therapy. A lack of detailed biopsy grading information before treatment for the HCV-monoinfected patients precluded a comparison of the stages of disease at treatment initiation. The receipt of HCV treatment was not included in the multivariate model of severe HCV disease (Supporting Table 1) because it is a surrogate for the severity of disease. However, when it was included to adjust for the impact of different treatment rates in the 2 groups, HIV coinfection was still not significantly associated with severe HCV disease in the multivariate model [HR = 1.2 (95% CI = 0.6-2.7), P = 0.61]. The rates of sustained virological response were similar for the 2 groups: 13.5% (5/37) for the HCV/HIV-coinfected group and 15.8% (9/57) for the HCV-monoinfected group.

Among HCV/HIV-coinfected transplant recipients, the only factor independently associated with severe HCV disease was treated acute rejection (HR = 2.7, P = 0.04). Treated cytomegalovirus infection (HR = 15.6, P < 0.001), another univariate predictor, was not included in the multivariate model because of the small number of cytomegalovirus infection cases. Three HCV/HIV-coinfected recipients (3%) spontaneously cleared their HCV infection after LT.

Acute Rejection

The cumulative incidence of acute rejection by year 3 requiring treatment was significantly higher for HCV/HIV patients (39%) versus HCV patients (24%, HR = 2.1, P = 0.01). More than half (54%) of the initial rejection episodes among the HCV/HIV-coinfected recipients occurred within the first 21 days after LT. Twenty-three of the 28 first-treated acute rejection episodes in coinfected recipients were centrally read, and 16 of those episodes were able to be graded according to the Banff criteria: 2 were indeterminate, 3 were mild (grade I), 10 were moderate (grade II), and 1 was severe (grade III). Twenty-five of the 28 coinfected subjects who experienced rejection were on antiretroviral therapy at the time of first-treated acute rejection. The majority of the HCV/HIV patients with acute rejection (82%) were treated with corticosteroids; only 1 patient received OKT3 for severe acute rejection. In the multivariate analysis, HIV infection was the only factor significantly associated with treated acute rejection (HR = 2.0, P = 0.01; Supporting Table 2). Similar results were obtained when the multivariate analysis was restricted to the first 30 days after transplantation (data not shown). Among HCV/HIV-coinfected recipients, older recipient age (HR = 0.95, P = 0.03) and prednisone use after LT (HR = 0.4, P = 0.04) were associated with treated acute rejection in the multivariate model (Table 5).

Table 5. Predictors of First-Treated Acute Rejection in HIV/HCV-Coinfected Transplant Recipients
 HR (95% CI)P Value
  • *

    Time-varying covariate.

Univariate Predictor  
 Recipient age (years)0.94 (0.90-0.98)0.002
 Recipient Caucasian race1.3 (0.6-2.9)0.56
 TAC versus CSA as initial immunosuppression0.7 (0.3-1.5)0.34
 Donor risk index1.3 (0.5-3.4)0.57
 CD4 count to (50 cells/mm3)*1.0 (0.9-1.1)0.60
 PI use in previous 3 months (any)*1.2 (0.6-2.7)0.58
 Efavirenz without PI in previous 3 months*1.1 (0.4-3.0)0.78
 MMF use*0.8 (0.4-1.7)0.53
 Prednisone use*0.3 (0.1-0.7)0.01
 CSA trough level (ng/mL)*1.00 (0.99-1.00)0.31
 TAC trough level (ng/mL)*0.86 (0.74-0.99)0.04
Multivariate Predictor  
 Recipient age (years)0.95 (0.91-1.00)0.03
 Prednisone use*0.4 (0.2-1.0)0.04

HIV Disease After LT

The median changes in CD4 cell counts 6 months, 1 year, and 3 years after LT versus the baseline values were −83 cells/mm3 (IQR = −164 to 1 cells/mm3, P < 0.001), −11 cells/mm3 (IQR = −106 to 86 cells/mm3, P = 0.82), and −2 cells/mm3 (IQR = −74 to 102 cells/mm3, P = 0.61), respectively. HIV RNA was well controlled after LT. After transplantation, there were 3 cases of esophageal candidiasis, 1 case of candidiasis of bronchi, and 1 case of pneumocystosis. All these complications responded to antimicrobial therapy. One case of cutaneous Kaposi's sarcoma was treated successfully with a change to sirolimus-based immunosuppression.

Ten of the 11 subjects with detectable HIV RNA at the baseline had undetectable levels within 3 months, and 1 subject had undetectable levels 9 months after transplantation. Although the latter case subsequently had an episode of HIV viremia (544-63,409 copies/mL beyond 6 months), plasma HIV RNA was undetectable again at the most recent visit. Fourteen of the 78 individuals (18%) who were nonviremic at the baseline had 2 or more consecutive detectable measurements after transplantation. Among these 14 cases, there were 30 episodes of HIV viremia (median peak HIV RNA level = 7259 copies/mL, IQR = 1450-42,570 copies/mL). HIV RNA was detectable at the most recent follow-up visit in 9 of the cases (5 at the time of graft loss).

DISCUSSION

In this study, HCV-infected LT recipients with HIV were found to have significantly lower graft and patient survival than HCV-infected recipients without HIV. This is consistent with previously published studies.5, 6 Importantly, however, our study has uniquely evaluated predictors of poor outcomes and provides for the first time insights into critical patient selection and management strategies for maximizing good outcomes. The association of graft loss in HCV/HIV-coinfected recipients with a low BMI and the need for combined kidney-liver transplantation suggests that wait-listed patients who are more debilitated do less well, and the findings of higher graft losses among HCV/HIV-coinfected recipients with older or anti-HCV–positive donors highlights the importance of careful recipient and donor selection for achieving optimal outcomes.

In HCV transplant recipients, donor age is strongly associated with an increased risk of recurrent HCV cirrhosis, reduced responsiveness to HCV treatment, and higher rates of graft loss.18-20 Among the HCV/HIV patients in this study, older donor age was also associated with higher rates of graft loss, as was the use of anti-HCV–positive donors. The latter finding is important because the use of anti-HCV–positive organs in HCV-infected transplant recipients without HIV has not been associated with an increased risk of severe HCV disease progression or graft loss.21 Donor biopsy data are not available, so it is unknown whether greater severity of disease in anti-HCV–positive donors could account for this difference between HCV/HIV-coinfected recipients and HCV-monoinfected recipients.

Graft and patient losses due to HCV did not differ between HCV-infected recipients and HIV/HCV-coinfected recipients. This contrasts with a previous study from France indicating more severe HCV disease in coinfected patients.5 The rates of advanced fibrosis (F3-F4) were 20% (4/20) and 57% (8/14) in the HCV/HIV-coinfected French cohort compared to 17% and 27% in our cohort, 1 and 2 years after transplantation, respectively. In addition to the recognized limitations of the sample size in providing precise estimates of recurrence rates, donor, recipient, and posttransplant factors, including the use of anti-HCV therapy, likely vary from center to center and may account for differences between studies. Additionally, in our study, HCV/HIV-coinfected patients had a higher rate of graft loss due to sepsis and multiorgan failure (although it was not significant), and this may have contributed to a survivor bias in the assessment of fibrosis severity. The higher frequency of sepsis and multiorgan failure in the coinfected recipients may be related to their greater level of debilitation before transplantation, the higher rate of treated acute rejection, or the presence of underlying recurrent HCV disease.

The treatment of acute rejection in HCV-infected LT recipients is a well-recognized risk factor for recurrent cirrhosis and graft loss.12, 22 The finding of a significantly higher rate of acute rejection for HCV/HIV-coinfected recipients versus HCV-monoinfected recipients is a critically important finding of our study because the prevention of acute rejection may offer a means of optimizing outcomes for HIV/HCV-coinfected recipients. This higher rate of acute rejection may be due to a higher rate of misinterpretation of acute rejection (versus recurrent HCV or drug toxicity effects) in HCV/HIV-coinfected recipients, an overly cautious use of immunosuppression due to concerns about exacerbating HIV or HCV-related diseases, or a reflection of the difficulties in achieving adequate immunosuppression due to interactions between antiretroviral drugs and calcineurin inhibitors,16, 23 and it mirrors the experience with kidney transplant recipients.24 Our data suggest that HCV/HIV patients with acute rejection were underimmunosuppressed because a lower TAC trough level was associated with higher rates of treated acute rejection in the univariate model, and prednisone use after LT was protective in the multivariate model. Alternatively, it is also possible that the higher rates of acute rejection in patients with HIV reflect an inherently enhanced immune response and dysregulation of the immune response.25 Previous infections with other viruses (eg, cytomegalovirus) lead to the generation of memory alloreactive T cells as a result of cross-reactivity.26 Other studies have shown that the homeostatic expansion of T cells in HIV-infected patients is often coupled with the acquisition of a memory phenotype, which in turn is associated with an increased responsiveness of T cells and a nonspecific enhancement of alloimmunity.25 Studies investigating T cell responses in coinfected patients are ongoing and are expected to shed light on contributing mechanisms.

Similarly to the kidney transplant experience,24 there was no evidence of accelerated HIV disease progression, as indicated by stable or improved CD4+ T cell counts and the control of HIV viremia. HIV-specific infections/neoplasms did not contribute to morbidity or mortality in coinfected transplant recipients with follow-up periods as long as approximately 3.5 years.

Although this study represents the largest US experience with the transplantation of HIV/HCV-coinfected patients, its limitations include a lack of a uniform immunosuppressive regimen and standard antiretroviral therapy followed by all centers. Additional limitations include a lack of prospective data collection and central biopsy reading for the HCV-monoinfected controls. To prevent potential biases in selection, we carefully matched the HCV-monoinfected patients by center, year of transplantation, and key recipient factors. However, this may not have completely mitigated potentially important differences in the HCV/HIV and HCV cohorts. Finally, to minimize errors related to the assessment of liver disease severity, we focused on advanced fibrosis (bridging fibrosis and cirrhosis) as the critical endpoint. However, liver biopsy adequacy was not assessed, and thus an underestimation of fibrosis in both HCV/HIV-coinfected patients and HCV-monoinfected patients with small biopsy samples may have occurred.27 However, any measurement bias related to the assessment of fibrosis severity would be nondifferential.

In summary, this study highlights that patient and graft survival rates are lower for HCV-infected patients with HIV versus HCV-infected patients without HIV, but acceptable results are achievable in most coinfected patients. Importantly, our results indicate how recipient and donor selection and the management of posttransplantation complications may potentially be improved to maximize patient and graft survival. First, the severity of the recipient's illness, as reflected by a low BMI and the need for kidney transplantation, influences outcomes after LT. Thus, an early referral for the consideration of transplantation and the utilization of donor options (eg, living donor) that shorten the wait-list time are the best means of overcoming this potential barrier to transplantation. Second, donor selection is important. Older donors are already recognized as a risk for HCV-infected transplant recipients, and coinfected patients have higher rates of graft loss with older donors also. Anti-HCV–positive donors should be used cautiously because of the significant association with graft loss in our study. Third, cytomegalovirus infection was strongly associated with graft loss, and we recommend the use of universal prophylaxis to minimize any risk for this complication. Finally, reducing the rates of early acute rejection is highly desirable because the treatment of rejection is associated with graft loss and more severe HCV disease. Vigilance for rejection and a low threshold for performing biopsy to evaluate abnormal liver tests are recommended. Although the use of PIs and efavirenz has not been shown to be associated with rejection, graft loss, or death, we hypothesize that newer antiretroviral regimens avoiding the use of PIs and efavirenz may facilitate conventional immunosuppressive drug dosing with better drug exposure and potentially reduce rejection rates. Thus, although there remain many challenges, our results support the continued select use of this lifesaving treatment in HCV/HIV-coinfected patients.

Acknowledgements

The authors thank Rodney Rogers for his expert coordination of the study consortium. They also thank all the Solid Organ Transplantation in HIV: Multi-Site Study investigators and coordinators for their hard work and dedication to the study and subjects. A complete list can be found in the supporting information for this article.

Ancillary