Liver Retransplantation in HIV-Infected Patients: A Prospective Cohort Study


  • Presented in part at the 2011 American Transplant Congress, Philadelphia, PA; Abstracts published at Am J Transpl; 2011, 11 (Suppl. 2): 277. Abstract No. 826.

Jose M. Miro,


Information regarding liver retransplantation in HIV-infected patients is scant. Data from 14 HIV-infected patients retransplanted between 2002 and 2011 in Spain (6% retransplantation rate) were analyzed and compared with those from 157 matched HIV-negative retransplanted patients. In HIV-infected patients, early (≤30 days) retransplantation was more frequently indicated (57% vs. 29%; p = 0.057), and retransplantation for HCV recurrence was less frequently indicated (7% vs. 37%; p = 0.036). Survival probability after retransplantation in HIV-positive patients was lower than in HIV-negative patients, 42% versus 64% at 3 years, although not significantly (p = 0.160). Among HIV-infected patients, those with undetectable HCV RNA at retransplantation and those with late (>30 days) retransplantation showed better 3-year survival probability (80% and 67%, respectively), similar to that in their respective HIV-negative counterparts (72% and 70%). In HIV-infected and HIV-negative patients, 3-year survival probability in those with positive HCV RNA at retransplantation was 22% versus 65% (p = 0.008); in those with early retransplantation, 3-year survival probability was 25% versus 56% (p = 0.282). HIV infection was controlled with antiretroviral therapy after retransplantation. In conclusion, HIV-infected patients taken as a whole have unsatisfactory survival after liver retransplantation, although patients with undetectable HCV RNA at retransplantation or undergoing late retransplantation show a more favorable outcome.


acquired immunodeficiency syndrome


confidence interval


chronic ductopenic rejection




cerebrovascular accident


Foundation for Research and Prevention of AIDS in Spain


Spanish Group for the study of acquired immunodeficiency syndrome


Spanish Group for the study of infections in transplanted patients


highly active antiretroviral treatment


hepatic artery stenosis


hepatitis B virus


hepatocellular carcinoma


hepatitis C virus


human immunodeficiency virus


hazard ratio


intravenous drug user


interquartile range


liver transplantation


model for end-stage liver disease


multiorgan failure


men who have sex with men


nucleoside reverse transcriptase inhibitor


primary graft nonfunction


protease inhibitors


liver retransplantation


Spanish Society of Infectious Diseases and Clinical Microbiology


sustained virological response


Before the advent of highly active antiretroviral treatment (HAART) in 1996, poor outcome following liver transplantation (LT) in patients infected with human immunodeficiency virus type 1 (HIV) (1) led experts to consider HIV infection a formal contraindication for this procedure (2). However, several authors showed that HAART was able to control HIV infection after LT (3–6). Despite this success, some recent single-center reports have described reduced long-term survival after LT in patients coinfected by HIV and hepatitis C virus (HCV), mainly due to recurrence of HCV infection (7,8). Moreover, the recently published results of ongoing nationwide prospective studies conducted in Spain (funded by the Spanish Foundation for AIDS Research and Prevention [FIPSE]) and in the United States (Transplant Study for People with HIV funded by the National Institutes of Health) show that survival after LT in HIV/HCV-coinfected patients is shorter than in HIV-negative matched controls (9,10).

The frequency of liver retransplantation (reLT) in HIV-negative patients has ranged from 6.5% to 13.4% in recent years. The most common reasons for emergency reLT are primary graft nonfunction (PNF) and vascular thrombosis; the most common reasons for elective reLT are disease recurrence and chronic rejection (11–14). Overall survival after reLT is 15–20% lower than that of primary transplant recipients (11,15,16), and the marked disparity between the number of patients awaiting their first LT and the scarcity of available organs gives cause for concern. HIV-infected recipients and HIV-negative recipients can suffer from the same complications after LT (graft dysfunction, vascular thrombosis and recurrence of liver disease), and these could eventually lead to reLT. However, published experience with reLT in HIV-infected patients is scant, and most cases are only mentioned in articles reporting single-center experiences after primary transplantation in HIV-infected patients (5,7,17–22). Consequently, the incidence and outcome of reLT in HIV-infected patients is unknown, and, since the benefit of primary LT in HIV-infected patients remains open to debate, the usefulness of reLT in this population may be even more controversial. Therefore, we describe the incidence, indications, main characteristics and outcome of reLT in HIV-infected patients included in the aforementioned Spanish prospective FIPSE study. In addition, we compare survival of reLT in HIV-infected patients with that of matched HIV-negative controls.


Study design

We performed a prospective, multicenter, cohort study of HIV-infected patients who underwent reLT between January 2002 and March 2011 in 17 centers participating in the Spanish nationwide prospective (FIPSE) study of LT in HIV-infected patients. Patients were followed until January 2012. We compared the survival of HIV-infected reLT patients and HIV-negative reLT patients (controls) selected according to frequency matching for center, period of reLT (2002–2011), and HCV or HBV infection at the first transplant. The study was approved by the Institutional Review Boards of all the participating sites. All patients signed the informed consent form.


Variables were collected for each HIV-infected patient at each site using a standardized case report form as previously described (9). Variables were recorded pre-LT, post-LT, pre-reLT and post-reLT. Donor variables were also recorded (Tables 1–3). We calculated the 1999 Rosen score, which comprises five recipient variables (age, bilirubin, creatinine, United Network for Organ Sharing status and cause of graft failure) and helps to predict survival in patients undergoing reLT (23). The modified Rosen score for patients who underwent re-LT 15 days or more after their primary LT was calculated according to the new criteria of Rosen et al. (24) published in 2003. This new score is derived from four recipient variables, namely age, bilirubin, creatinine and interval to reLT. Patient information was sent every 6 months to the coordinating center and entered into the FIPSE OLT-HIV-05-GESIDA 45–05 database (available at, as previously described (9).

Table 1.  Baseline characteristics of the 14 patients who underwent liver retransplantation in Spain (2002–2011)
CaseAge at reLTGenderHIV risk factorPrimary liver diseaseHCCHCV genotypeHCV RNA at reLTIndication reLTIndication type
  1. CDR = chronic ductopenic rejection; HBV = hepatitis B virus; HCC = hepatocellular carcinoma; HCV = hepatitis C virus; IDU = intravenous drug user; NA = not applicable; OH = alcohol; PNF = primary graft nonfunction; reLT = liver retransplantation.*Early reLT (≤30 days); **Late reLT (>30 days).

137MaleHeterosexualHCV cirrhosisNo1PositiveVascular thrombosisEmergency*
246MaleFormer IDUHCV cirrhosisYes1PositiveVascular thrombosisEmergency*
344MaleFormer IDUHCV cirrhosis + HBVNo1PositivePNFEmergency*
445MaleFormer IDUHCV cirrhosisNo1PositiveVascular thrombosisEmergency*
545MaleFormer IDUHCV cirrhosisNo3PositivePNFEmergency*
651MaleFormer IDUHBV cirrhosisNoNANAVascular thrombosisEmergency*
744MaleFormer IDUHCV cirrhosisNo1PositivePNFEmergency*
841MaleFormer IDUHCV cirrhosisNo2PositiveVascular thrombosisEmergency*
941MaleFormer IDUHCV cirrhosis + HBVYes1NegativeVascular thrombosisElective**
1039MaleFormer IDUHCV cirrhosisNo1NegativeCDRElective**
1152MaleFormer IDUHCV cirrhosisNo1NegativeCDRElective**
1249FemaleHeterosexualHCV cirrhosisYes1PositiveHCV cirrhosisElective**
1350FemaleFormer IDUHCV cirrhosis + HBVNo4NegativeCDRElective**
1447MaleFormer IDUHCV cirrhosis + OHNo3NegativeCDRElective**
Table 2.  Baseline characteristics and outcome of the 14 patients who underwent liver retransplantation in Spain (2002–2011) and characteristics of their donors
CaseDays from 1st LTMELD score at reLT1999 Rosen score at reLT2003 Rosen score at reLTDonor ageDonor cause of deathFollow-up (months)Post-reLT complicationsOutcome/cause of death
  1. CDR = chronic ductopenic rejection; CMV = cytomegalovirus; CRA = cardiorespiratory arrest; HAS = hepatic artery stenosis; LT = liver transplantation; MELD = Model for End-Stage Liver Disease; MOF = multiorgan failure; NA = not applicable; PNF = primary graft nonfunction; reLT = liver retransplantation; SBP = spontaneous bacterial peritonitis.

11320.12879NA59Stroke9CDR, sepsis, MOFDeath/HCV recurrence
2370.14025NA35Stroke13Not reportedDeath HCC recurrence
33220.41136NA73Trauma1Biliary leakage, perihepatic hematomaDeath/sepsis, MOF
46280.40967NA51Stroke5Fibrosing cholestatic hepatitisDeath/HCV recurrence
510230.66883NA51Stroke51Femoral vein thrombosis HASAlive
712250.61056NA55Stroke3PNF, hemoperitoneum, HCV recurrence, esophageal candidiasis, biliary infectionDeath/stroke
818250.3345614.36929Stroke90Chronic renal failureAlive
9276210.5515214.45154Stroke86Biliary stenosisAlive
10281250.5957715.09740Trauma26Biliary stenosisAlive
11510240.8592717.77355Stroke18Not reportedAlive
121114200.3830412.53646Stroke11Severe HCV recurrence, SBPDeath/HCV recurrence
13135230.8602418.60378Stroke12Esophageal herpes ulcers, CMV viremia, hypertensionAlive
14332220.4889413.07948Trauma1Septic shockDeath/sepsis, MOF
Table 3.  Characteristics of HIV infection in the 14 patients who underwent liver retransplantation in Spain (2002–2011)
CasePre-LTPre- reLTLast check-up
Plasma HIV RNA viral loadCD4+ T cells/mm3HAARTPlasma HIV RNA viral loadCD4+ T cells/mm3HAARTHAART resumed (days)Plasma HIV RNA viral loadCD4+ T cells/mm3HAART
  1. ATV = atazanavir; AZT = zidovudine; DDI = didanosine; D4T = stavudine; EFV = efavirenz; ETV = etravirine; FTC = emtricitabine; LOP/r = lopinavir/ritonavir; LT = liver transplantation; MVC = maraviroc; NA = not available; ND = not determined; NFV = nelfinavir; RAL = raltegravir; reLT = liver retransplantation; RTV = ritonavir; SQV = saquinavir; TDF = tenofovir; T-20 = enfurvitide; 3TC = lamivudine.

1<50368T-20, 3TC, NFVNDNDNot restarted240319252AZT, 3TC, T-20
2<50644ATV/rtv, EFVNDNDNot restarted4<50866RAL, ATV
3<50277TDF, 3TC EFVNDNDNot restartedNot restarted<5067NA
4<5090TDF+3TC, LOP/rNDNDNot restarted12<5022RAL, TDF+FTC
5<50172EFV,TDF+ FTC<50409Not restarted23<50193EFV+TDF+ FTC
6<50177TDF+ FTC, SQV/rNDNDNot restartedNot restartedNDNDNA
7223346RAL<501023EFV,TDF+FTC26<5045AZT+3TC, MVC
8104108AZT+3TC, EFV<50NDAZT+3T, EFV3<50157DDI, RAL, FTC
1034700344No Treatment<50288RAL, TDF+FTC, ETVNA<50353RAL, TDF+FTC
12<50208FTC+TDF, EFV<50336FTC+TDF, EFV23<5027FTC+TDF, EFV
13<504903TC, DRV/r<50481ABC+3TC, DRV/r1<5026ABC+3TC, DRV/r
14<20115TDF+FTC, SQV/r<20141TDF, FTC, RALNot restartedNDNDNA

The definitions of infection by HIV, HCV and hepatitis B virus (HBV), as well as PNF, vascular thrombosis, chronic rejection, HCV recurrence and other posttransplant complications, were based on standard criteria (25,26).

Data from HIV-negative reLT recipients were obtained from the Spanish Liver Transplant Registry, as previously described (9). Data were analyzed blind at the coordinating center.

Primary LT and reLT criteria

HIV-infected patients had to fulfill the 2005 GESIDA/GESITRA-SEIMC criteria according to their infection status (27). As for liver disease, the criteria for accepting HIV-infected patients for primary LT were the same as those followed in Spain for HIV-negative patients. Since there are no uniform criteria for reLT in Spain, each participating center followed its local protocol for indicating reLT in both HIV-infected and HIV-negative patients.

Patients who underwent reLT within the first 30 days after primary LT were classified as early reLT recipients. The remaining cases were classified as late reLT recipients.

Post-LT management

Antiretroviral therapy was administered until the day of surgery and restarted once the patient was stable and oral intake was reintroduced. Antiretroviral drugs were administered according to Spanish national guidelines (28). HIV-infected patients received the same immunosuppressive regimens as HIV-negative patients according to local protocols. Posttransplant prophylaxis and prophylaxis against HIV infection were administered according to Spanish national guidelines (29,30), as previously described (9,31). HCV recurrence was treated with pegylated interferon α-2a or α-2b and ribavirin based on the same criteria as for HCV-monoinfected LT recipients according to local protocols.

Statistical analysis

Variables are expressed as mean and standard deviation, median and interquartile range (IQR) or as proportions. The Fisher exact test and Wilcoxon test were used to assess whether the predictors were balanced between HIV-infected patients and controls. Survival analyses were performed with the date of reLT as the start date; death from any cause was treated as failure. Survival time from reLT was estimated using the Kaplan–Meier product-limit method; the curves obtained in the different groups were compared using the generalized log-rank test (univariate Cox model analysis). Statistical significance was defined as a bilateral p-value <0.05. All statistical analyses were carried out using Stata (release 9.2).


During the study period, 237 HIV-infected patients underwent primary LT. Patient disposition is shown in Figure 1. Graft failure was recorded in 61 patients (26%), although only 17 out of the 61 patients (28%) were accepted and listed for reLT. Fourteen of the 17 patients underwent reLT, and the remaining three died while on the waiting list. Therefore, the frequency of reLT was 6%. The causes of graft failure and indications for reLT are shown in Figure 1. Three cases underwent reLT in 2004 and 2009, and two cases underwent reLT in 2007, 2008, 2010 and 2011.

Figure 1.

Progress of HIV-infected patients with graft failure after primary liver transplantation (LT). 
a Percentage for HIV-infected patients with primary LT. 
b Percentage for patients with graft failure. 
c Percentage for patients listed for re-LT.

The main clinical characteristics and the outcome of the 14 HIV-positive patients who underwent reLT are shown in Tables 1 and 2. The median (IQR) age was 45 (37–52) years and all but two patients were males (86%). At primary LT, 13 patients had HCV-related liver disease (93%) (three were coinfected with HBV) and one patient (7%) had HBV-related liver disease. Eight patients (57%) underwent early reLT, five because of vascular thrombosis and three because of PNF. Six patients (43%) underwent late reLT because of chronic rejection in four cases and hepatic artery stenosis and HCV recurrence in one case each. Among the 13 patients with HCV infection, five were HCV RNA-negative and eight were HCV RNA-positive at the time of reLT. The median (range) Model for End-stage Liver Disease score, 1999 Rosen score and 2003 Rosen score were 23.5 (7–32), 0.52 (0.12–0.86) and 14.45 (12.54–18.60), respectively. None of the patients were high-risk in either of the 2 Rosen scores (23, 24). All liver grafts were obtained from deceased donors. Data from donors are shown in Table 2.

After a median (IQR) follow-up of 12 (3–26) months, eight of the 14 HIV-infected patients with reLT (57%) died. The causes of death were severe recurrence of HCV infection in three cases, infectious complications in two, and stroke, recurrence of HCC and intraoperative death in one case each. Survival was 25% (2/8 patients) after early reLT and 67% (4/6 patients) after late reLT, although this difference was not statistically significant (p = 0.227). Survival in patients who were HCV RNA-negative at reLT (80%[4/5 patients]) was higher than that in patients who were HCV RNA-positive (25%[2/8 patients]), although the difference was not statistically significant (p = 0.103).

Data concerning HIV infection are summarized in Table 3. Twelve patients (86%) were former drug users. The median (range) CD4+ T cell counts before LT, before reLT and at the last check-up before death were 310 (90–644), 324 (141–1023) and 239 (26–550) cells/mm3, respectively. Most patients had an undetectable viral load (<50 copies/mL) at all time points. HAART was restarted after reLT in 11 cases (79%) at a median of 6 days (1–26) after surgery. Five patients received a raltegravir-based regimen, two an efavirenz-based regimen and four other regimens.

Post-reLT survival in HIV-infected and HIV-negative recipients

During the study period, 549 reLT were performed in 6,310 HIV-negative patients with a primary LT at the same participating centers (frequency of reLT, 8.7%[HIV-negative] vs. 6%[HIV-infected]; p = 0.165). A total of 157 reLT were performed in HIV-negative and HCV-positive or HBV-positive recipients. The main clinical characteristics and outcome of the 14 HIV-infected and the 157 HIV-negative patients with reLT are summarized in Table 4. HIV-infected reLT recipients were younger (p<0.001), showed a higher proportion of early reLT (57% vs. 29%, p = 0.057) and, accordingly, had a shorter interval between the first and the second LT (15 vs. 265 days, p = 0.055). The reasons for reLT are shown in Table 4. The rate of HCV recurrence as an indication for LT was significantly lower in HIV-infected recipients (7% vs. 37%, p = 0.036). Donor characteristics were similar in both groups.

Table 4.  Characteristics of liver retransplantation in HIV-infected and HIV-negative recipients (2002–2011)
  1. *Median and interquartile range.

  2. **Biliary complications (seven cases), HBV recurrence (two cases), donor-transmitted tumor (one case), massive hemorrhagic graft necrosis (one case).

  3. Five cases had a negative HCV RNA viral load at the time or reLT.

  4. ††Primary graft nonfunction and vascular complications in six and three cases, respectively.

  5. In HIV-infected patients: stroke and recurrence of HCC. In HIV-negative patients: kidney failure in two cases and rejection, myocardial infarction, massive hemorrhagic graft necrosis, tumor, stroke and pulmonary embolism in one case each.

No. of cases14157 
Recipient variables
 Age, y*45 (41, 49)55.5 (47.5, 62)<0.001
 Male gender12 (86%)111 (71%)0.354
Primary LT
 HCV infection13 (93%)142 (90%)0.856
 HBV infection1 (7%)15 (10%)0.992
 Hepatocellular carcinoma3 (21%)54 (34%)0.490
Primary LT donor:
 Donor age, y*53 (49, 64)50 (36, 65)0.297
 Donor age ≥65, y3 (21%)41 (26%)1.000
 Donor brain death by trauma3 (21%)38 (24%)0.696
Days from primary LT*15 (6, 281)256 (12,1049)0.055
 Early reLT (≤30 days)8 (57%)45 (29%)0.057
 Late reLT (>30 days)6 (43%)112 (71%) 
Reasons for reLT
 Primary graft nonfunction (PNF)3 (21%)31 (20%)0.488
 Vascular complications6 (43%)33 (21%)0.310
 Rejection4 (29%)18 (11%)0.086
 HCV recurrence1 (7%)58 (37%)0.036
 Other**11 (8%)0.602
 Not available6 (4%)1.000
ReLT donor:
 Donor age, y*53 (48, 59)52 (34.5, 62.5)0.594
 Donor age ≥65, y2 (14%)37 (24%)0.526
 Donor brain death by trauma3 (21%)38 (24%)0.310
Length of follow-up (months)
Overall12 (3, 26)17 (1.5, 45)0.791
 According to the interval from primary LT
  Early reLT (≤30 days)7 (2, 32)12 (1, 39) 
  Late reLT (>30 days)15 (12, 26)18 (2, 49) 
 According to HCV RNA at reLT
 Negative17 (12, 26)13 (3, 41) 
 Positive9 (3, 13)21 (3, 49) 
Third transplantation10 (6%)
Mortality8 (57%)55 (35%)0.176
 According to the interval from primary LT
 Early reLT (≤30 days)6/8 (75%)21/45 (47%)0.250
 Late reLT (>30 days)2/6 (33%)34/112 (30%)0.764
 According to HCV RNA at reLT
 Negative1/5 (20%)5/23 (22%)1.000
 Positive6/8 (75%)32/100 (32%)0.039
Cause of death
 Intra-operative death1 (7%)3 (2%)0.750
 Technical complications††9 (6%)0.767
 Infections2 (14%)25 (16%)0.825
 HCV recurrence3 (21%)8 (5%)0.069
 Miscellaneous2 (14%)8 (5%)0.418
 Not available3 (2%)

After a median (IQR) follow-up of 12 (3–26) and 17 (1.5–45) months, respectively, eight (57%) HIV-infected patients and 55 (35%) HIV-negative patients died (p = 0.176). The distribution of the causes of death was similar in both groups (Table 4), although recurrence of HCV showed a nonsignificant trend toward being more frequent in HIV-infected recipients (21%[3/14] vs. 5%[8/157], p = 0.069). Mortality rates in both groups classified according to the interval to reLT (early vs. late) and to HCV infection status (HCV RNA-positive vs. HCV RNA–negative) are shown in Table 4. The only significant difference was observed in HCV RNA-positive patients, with a higher mortality in HIV-infected patients (6/8 [75%] vs. 32/100 [32%]; p = 0.039).

The survival probability of reLT recipients according to HIV status is shown in Figure 2A. Survival at 1 and 3 years for HIV-infected and HIV-negative patients was 50% versus 72% and 42% versus 64%, respectively (p = 0.160).

Figure 2.

Probability of survival in the whole cohort (A), in patients with negative HCV RNA (B) and in positive HCV RNA (C) at reLT, and in patients with early reLT (≤30 days) (D) or late reLT (>30 days) (E). Solid and dashed lines represent HIV-infected recipients and HIV-negative recipients, respectively.

Survival curves of reLT recipients according to their HCV RNA status at reLT are shown in Figure 2B and C. The survival probability at 1 and 3 years for HIV-infected and HIV-negative patients with a negative HCV RNA viral load at the time of reLT was 80% versus 82% and 80% versus 72.5%, respectively (p = 0.977). Conversely, the survival probability at 1 and 3 years for the same groups of patients with a positive HCV RNA viral load at the time of reLT was 33% versus 74% and 22% versus 65%, respectively (p = 0.008).

Survival curves after early and late reLT are shown in Figure 2D and E. Survival at 1 and 3 years for HIV-infected versus HIV-negative patients who had undergone early reLT was 37.5% versus 59% and 25% versus 56%, respectively (p = 0.282). Conversely, survival at 1 and 3 years for patients who had undergone late reLT was 67% versus 77% and 67% versus 67%, respectively (p = 0.868).


Information on reLT in HIV-infected patients is scarce. In fact, only a few cases of liver regrafting after primary LT have been reported in this population (Table 5) (5,7,17–22). Most cases are described in case series reporting outcome after primary LT in HIV-positive patients and, therefore, have not been studied in depth. Only two studies provide an in-depth analysis of reLT in HIV-infected patients (17,20). We present detailed prospective data on clinical characteristics and outcome in 14 HIV-infected reLT recipients, making ours the most numerous series to date. We also compare these patients with 157 matched HIV-negative reLT recipients.

Table 5.  Retransplantation in HIV-infected patients: literature findings
Author, year (Reference)Gender/ageIndication for primary LTIndication for reLTInterval to reLTOutcome
  1. *HIV infection acquired after primary LT.

  2. reLT = liver retransplantation; HCV = hepatitis C virus; HAART = highly active antiretroviral therapy; FHF = fulminant hepatic failure; HBV = hepatitis B virus; PNF = primary graft nonfunction; MOF = multiorgan failure; LT = liver transplantation; LDLT = living donor liver transplantation; NA = not available.

Neff, 2003 (17)F / 48HCV cirrhosis*Chronic rejection28 monthsAlive 7 months
Polard, 2005 (19)M / 34HCV cirrhosisHAART induced FHF22 weeksAlive 4 years
Vogel, 2005 (18)M / 34HVB FHFPNFNAAlive
De Vera, 2006 (7)F / 50HCV cirrhosisPNFNADeath 7 days MOF
 M / 52HCV cirrhosisCholestatic HCV recurrence7 monthsAlive 10 months
Roland, 2008 (5)NAN.A.Small for size after LDLT49 daysDeath 13 months, HCV recurrence
Jao, 2010 (20)F / 57Nevirapine-induced FHFPNF7 daysAlive 2 years
Di Benedetto, 2011 (21)NANAPNF4 daysAlive
 NANAPNF5 daysDeath 34 days, sepsis
Cherian, 2011 (22)M/9HAART-induced FHFPNF29 daysDeath 26 months, chronic rejection
 NANAVascular complications11 monthsNA
 NANAVascular complications13 monthsNA

The 14 cases represent a 6% frequency of reLT in HIV-infected patients, which is not substantially different from the 8.7% frequency in the HIV-negative population (p = 0.165). The indications for reLT were similar in both groups, with the exception of HCV recurrence, which was significantly less common in HIV-infected recipients (7% vs. 37%; p = 0.036) (Table 4). This finding is consistent with the very low number of HIV-infected patients with primary LT and graft failure due to recurrence of HCV who were accepted for reLT in our series (1 of 36 patients [3%]) (Figure 1). Transplant teams may be reluctant to indicate reLT in HIV-positive patients with recurrence of severe HCV infection for the following reasons: poorer outcome after primary LT in HIV/HCV-coinfected patients, the lack of knowledge about factors affecting progression of fibrosis in this population, the relatively unsuccessful therapies available to treat HCV recurrence at present and the current shortage of organs (7,8,15). Furthermore, reduced survival after reLT in HIV-negative–HCV-positive patients may have reduced the degree of acceptance of reLT among our HIV/HCV-coinfected patients (15,16). A similar attitude was expressed by many US transplant centers in a survey conducted by Burton et al. (32) to characterize current practice in reLT for allograft failure caused by HCV recurrence in HIV-negative patients. However, the forthcoming introduction of new and highly effective anti-HCV drugs could make it easier to eradicate HCV in HIV/HCV-coinfected patients after primary LT, with the result that more patients could be considered candidates for reLT.

In contrast, reLT for PNF or vascular complications was indicated in 67% of HIV-infected patients (10 out of 15 cases) (Figure 1), probably reflecting the reticence of physicians to deny emergency reLT when the indication is related to technical problems (PNF, vascular complications or small-for-size syndrome). In agreement with our results, most reLT in HIV-infected patients reported in the literature were also indicated because of these technical complications (Table 5). Nonetheless, LT teams from the centers participating in the current study were cautious when making decisions on early reLT in HIV-infected patients, as suggested by the fact that no patients had very high MELD or Rosen scores, which are associated with a high risk of post-reLT mortality.

In our study, overall patient survival after reLT in HIV-infected patients was relatively poor and lower than that recorded in HIV-negative reLT recipients (50% vs. 72% at 1 year and 42% vs. 64%, at 3 years, respectively) (Figure 2A). This difference did not reach statistical significance (p = 0.160), probably because of the low number of HIV-infected cases. We also tried to identify subpopulations with different outcomes. Therefore, we stratified survival according to HCV RNA status (positive vs. negative) at the time of reLT and according to the interval between primary LT and reLT (early vs. late reLT) and compared the results with those of the respective HIV-negative counterparts. Survival of HIV-infected patients with a negative HCV RNA viral load at the time of reLT was almost identical to that of HIV-negative–HCV-RNA negative reLT recipients (80% vs. 72.5% at 3 years, respectively) (Figure 2B). Conversely, among patients with a positive HCV RNA viral load at the time of reLT, survival of HIV-infected patients was significantly lower than that of HIV-negative reLT recipients (22% vs. 65% at 3 years, respectively; p = 0.008) (Figure 2C). Survival of HIV-infected patients with a positive HCV RNA viral load at the time of reLT was poor regardless of the fact that most reLT indications were not directly related to HCV recurrence. These results, if confirmed in larger series, could prove helpful when making decisions on reLT in HIV-infected patients: reLT could reasonably be offered to patients with a negative HCV RNA viral load, whereas it might be inadequate in patients with a positive HCV RNA viral load. However, the introduction of more effective therapies against HCV could modify this approach.

In our series, the 3-year probability of survival of HIV-infected patients who underwent early reLT was only 25% (Figure 2D). Although this rate was lower than that of HIV-negative patients with early reLT (56% at 3 years) and lower than that of HIV-infected patients with late reLT (67% at 3 years), the differences did not reach statistical significance, probably owing to the small numbers of HIV-infected patients in this study. Larger series are needed before robust conclusions can be drawn. Confirmation of these results would cast serious doubts on the indication of early reLT in HIV-infected patients, even in those cases with graft failure due to technical complications directly related to the primary LT procedure. Nonetheless, most deaths after early reLT in our HIV-infected patients were caused by mid-term complications that were unrelated to the procedure, mainly HCV recurrence (Table 2). In contrast, survival after late reLT in HIV-infected patients was acceptable (67% at 3 years) and identical to that seen in HIV-negative patients receiving late reLT (67% at 3 years), thus suggesting that late reLT may reasonably be indicated in HIV-infected patients. However, it remains noteworthy that all but one patient with late reLT had a negative HCV RNA viral load at the time of the procedure (Table 1), thus precluding any conclusion on whether the acceptable survival in this subpopulation was the result of favorable HCV RNA status, the late indication of reLT, or both. A multivariate analysis could have helped to identify which variables were associated with this acceptable outcome; however, we were unable to perform an analysis of this type owing to the small number of cases included.

Control of HIV disease was adequate before and after reLT. All patients initially underwent surgery while fulfilling strict criteria for HIV disease control (27), except for one patient with a CD4 T cell count of 90 cells/mm3 in the last test before LT. Due to the short interval, CD4 T cell count and HIV viral load were not recorded before emergency reLT; however, those with available data had an HIV viral load below detection limits and CD4 T cell counts over 100 cells/mm3 while on the waiting list for reLT. All patients who underwent elective reLT had suppressed plasma HIV RNA viral load and stable CD4 T cell counts before reLT. These parameters persisted after regrafting, except at the last check-up in those who died.

The main limitation of our study is its small sample size and the short follow-up of patients, which could have influenced the statistical analysis and did not allow us to investigate subpopulations of interest (e.g. subsets obtained after combining variables such as reLT indication, interval from primary LT, or HCV RNA status), thus preventing us from drawing firm conclusions. Nevertheless, our study is prospective, constitutes the most numerous series of reLT in HIV-infected patients and compares these patients with matched HIV-negative reLT recipients. Given the current low rate of reLT in most transplant centers, we believe that recording cases in a worldwide registry might be the only way to answer questions concerning this procedure in HIV-infected patients. Meanwhile, arguments for and against reLT in HIV-infected patients will necessarily be hypothetical. In this context, the main reasons for reLT are that it is the only therapeutic option for patients with allograft failure and that transplant teams find it difficult to refuse early reLT for graft failure due to technical complications. A powerful argument against reLT in HIV-infected patients is that survival after primary LT is lower in HIV/HCV-coinfected patients (7–10), thus making it difficult to justify reLT when primary LT in HIV-infected patients is still under debate. In addition, offering reLT to HIV-infected patients may be unacceptable when there are so many HIV-negative patients with better expectable outcomes waiting for their first transplant. Unfortunately, current data from both our study and the literature review are not yet sufficiently robust to resolve these doubts.

We conclude that overall survival after reLT in HIV-infected patients is unsatisfactory. Nevertheless, HIV-infected patients with a negative HCV RNA viral load at the time of reLT and patients undergoing late reLT had a more acceptable survival rate. Remarkably, HIV infection is controlled with HAART after reLT. However, since these conclusions are based on a very small sample size, a multinational registry of reLT in HIV-infected patients is necessary.


We are indebted to the study participants and to the staff of the liver transplant units at the centers for retrieving detailed data on donors and recipients. We are also grateful to the following for their constant support from the beginning of the project: `Fundación para la Investigación y Prevención del Sida en España (FIPSE)’, Madrid, Spain; the National AIDS Plan Secretariat and the National Transplant Organization (ONT) of the Spanish Ministry of Health, Madrid, Spain; the Spanish Society of Liver Transplantation (SETH), Madrid, Spain; the Spanish Association for the Study of the Liver (AEEH), Spain; and the HIV/AIDS Study Group (GESIDA) and Infections in Transplants (GESITRA) Working Groups and the SEIMC/GESIDA Foundation (FSG) of the Spanish Society of Infectious Diseases and Clinical Microbiology (SEIMC), Madrid, Spain. CIBEREHD supported by Instituto de Salud Carlos III, Spain.

Funding sources: This study was supported by the Spanish Foundation for AIDS Research and Prevention (FIPSE, Madrid, Spain), grants TOH-VIH/05 and TOH-VIH/08. JMM had in the year 2011 an INT10/219 Research Intensification Grant (I3SNS and PRICS programs) from the Instituto de Salud Carlos III, Madrid (Spain) and the Departament de Salut de la Generalitat de Catalunya, Barcelona (Spain).


The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.


Appendix I

List of Investigators of the Spanish OLT in HIV-Infected Patients Working Group: J.D. Pedreira, M.A. Castro, S. López, F. Suárez, P. Vázquez, Hospital Universitario de A Coruña, A Coruña; J.M. Miro, F. Agüero, J. Blanch, M. Brunet, C. Cervera, E. de Lazzari, C. Fondevila, A. Forner, J. Fuster, N. Freixa, J. C. García-Valdecasas A. Gil, J.M. Gatell, M. Laguno, M. Larrousse, J. Mallolas, C. Manzardo, M. Monrás, A. Moreno, J. Murillas, D. Paredes, I. Pérez, F. Torres, C. Tural, M. Tuset, A. Rimola, Hospital Clínic-IDIBAPS, University of Barcelona, Barcelona; A. Antela, E. Losada, E. Molina, E. Otero, E. Varo, Hospital Clínico Universitario, Santiago de Compostela, A Coruña; R. Lozano, J.J. Araiz, E. Barrao, J. Larraga, S. Letona, P. Luque, A. Navarro, I. Sanjoaquín, T. Serrano, E. Tejero, Hospital Clínico Universitario Lozano Blesa, Zaragoza; M. Salcedo, R. Bañares, J. Berenguer, G. Clemente, J. Cosín, J.P. Ferreiroa, J.L. García-Sabrido, I. Gutiérrez, J.C. López, P. Miralles, M. Ramírez, D. Rincón, M. Sánchez, Hospital General Universitario Gregorio Marañón, Madrid; M. Jiménez, J. de la Cruz, J.L. Fernández, J.M. Lozano, J. Santoyo, J.M. Rodrigo, M.A. Suárez, Hospital Regional Universitario Carlos Haya, Málaga; M. Rodríguez, M.P. Alonso, V. Asensi, M.L. González-Diéguez, I. González-Pinto, Hospital Universitario Central de Asturias, Oviedo; A. Rafecas, C. Baliellas, J. Carratalá, J. Fabregat, N. Fernández, R. Jorba, L. Lladó, X. Xiol, Hospital de Bellvitge-IDIBELL, University of Barcelona, Hospitalet de Llobregat, Barcelona; M. Montejo, J. Bustamante, J.R. Fernández, M. Gastaca, J. González, E. Montejo, J. Ortiz de Urbina, P. Ruiz, M.J. Suárez M. Testillano, A. Valdivieso, A. Ventoso, Hospital Universitario de Cruces, University of the Basque Country, Bilbao, Vizcaya; M. Abradelo, J. Calvo, J.R. Costa, A. García-Sesma, C. Jiménez, A. Manrique, J.C. Meneu, E. Moreno, V. Moreno, S.P. Olivares, F. Pulido, R. Rubio, Hospital Universitario Doce de Octubre, Madrid; M. Blanes, V. Aguilera, M. Berenguer, J. López, R. López, M. Prieto, F. San Juan, Hospital Universitari La Fe, Valencia; M.C. Fariñas, F. Casafont, S. Echevarria, E. Fábrega, J.D. García-Palomo†, M. Gomez-Fleitas, M. Gutiérrez-Cuadra, J.L Herrera-Noreña, Hospital Universitario Marqués de Valdecilla, Santander; S. Moreno, R. Barcena, S. del Campo, J. Fortún, A.M. Moreno, P. Martín-Dávila, Hospital Universitario Ramón y Cajal-IRYCIS, Madrid; J. Torre-Cisneros, P. Barrera, J. Briceño, J.J. Caston, G. Costan, M. de la Mata, R. Lara, P. López-Cillero, J.L. Montero, A. Rivero, S. Rufian, Hospital Universitario Reina Sofía-IMIBIC, Córdoba; Ll. Castells, I. Bilbao, I. Campos-Varela, R. Charco, J.I. Esteban, J. Gavaldá, O. Len, A. Pahissa, E. Ribera, V. Vargas, Hospital Vall d’Hebrón, Univesitat Autonoma de Barcelona, Barcelona; J.A. Pons, Hospital Universitario Virgen de la Arrixaca, El Palmar, Murcia; E. Cordero, C. Bernal, J.M. Cisneros, M.A. Gómez, J.M. Pascasio, M.J. Rodríguez, M. Sayago, J.M. Sousa, G. Suárez, Hospital Universitario Virgen del Rocío, Sevilla; J. González-García, Hospital Universitario La Paz-IdiPAZ, Madrid, E. Aznar, E. Barquilla, H. Esteban and B. Moyano, SEIMC-GESIDA Foundation, Madrid.†Deceased.

Steering Committee: J.M. Miro (Chair), L. Castells, G. de la Rosa, J. Torre-Cisneros, J. Fortún, J. González-García, L. Guerra (FIPSE), F. Lozano, M. Manzanera (FIPSE), P. Miralles, A. Moreno, A. Rafecas, A. Rimola (Vice-Chair), and A. Valdivieso. P. Stock (University of San Francisco, San Francisco, California), M. Roland (University of San Francisco, San Francisco, California) and D. Samuel (Hôpital Paul Brousse, Paris, France) were the external advisors of the committee.

Follow-up Committee: J.M. Miro (Chair), S. del Campo, H. Esteban, J. González-García, C. Manzardo, E. Montejo, B. Moyano and M. Manzanera (FIPSE).

Coordinating Center Staff: E. Aznar, E. Barquilla, H. Esteban, J. González-García, and B. Moyano from the SEIMC-GESIDA Foundation, Madrid.

Methodological Committee: J.M. Miro, I. Perez, C. Manzardo, A. Moreno and A. Rimola from Hospital Clínic-IDIBAPS, Universitat de Barcelona, Barcelona.