Liver Unit, Queen Elizabeth Hospital, Birmingham, United Kingdom
National Institute for Health Research Biomedical Research Unit and Centre for Liver Research, University of Birmingham, Birmingham, United Kingdom
Clinical Lecturer in Hepatology, Centre for Liver Research, NIHR Biomedical Research Unit, Institute of Biomedical Research (5th floor) University of Birmingham, Edgbaston, Birmingham, United Kingdom B15 2TT
Although late hepatic artery thrombosis (HAT) is an infrequent complication of liver transplantation, it is a major cause of patient morbidity and mortality.1-5 Ischemic cholangiopathy may manifest as jaundice, cholangitis, bilomas, or bile leaks, and hepatic necrosis may result in abscess formation.3, 5-7 Short-term survival without intervention for symptomatic patients has been reported to be as low as 30%.5 Late HAT can occur at any time after transplantation. Consequently, as the number of long-term survivors increases, liver units are encountering a growing number of patients with late HAT and the clinical challenges that accompany them.
Asymptomatic patients with late HAT are generally managed conservatively. Intra-arterial thrombolysis has been described in small case series, yet it is not currently recommended for patients with severe graft dysfunction or sepsis.8 Retransplantation for late HAT is considered the treatment of choice in select patients with ischemic complications. However, despite being performed since the early 1980s, outcomes after retransplantation for late HAT specifically have not been described. Furthermore, variables associated with survival to retransplantation are not known. In an era of organ shortage, risk stratification of patients awaiting liver transplantation and the identification of patients who will not benefit from regrafting are paramount.
The aims of this study were (1) to identify determinants of mortality in patients listed for retransplantation for late HAT, (2) to describe survival after retransplantation for late HAT, and (3) to determine variables associated with mortality after retransplantation for late HAT.
PATIENTS AND METHODS
This was a single-center, retrospective study of consecutive patients diagnosed with late HAT and listed for retransplantation between January 1994 and May 2010. The study protocol was approved by the local clinical audit committee. Patients were identified from a prospectively collected database. Late HAT was defined as the occurrence of HAT more than 4 weeks after transplantation.1, 2, 3, 8-12 The diagnosis of late HAT was made by 1 of 3 liver radiologists on computed tomography or formal angiography, and was confirmed in most cases by examination of the explant or postmortem.
Data were collected for the following patient variables both at the time of listing for retransplantation and at the time of retransplantation: age, sex, ethnicity, indication for the primary liver graft, serum bilirubin, serum creatinine, serum sodium, international normalized ratio (INR) and albumin, smoking status, and comorbidities. The estimated glomerular filtration rate (eGFR) was calculated with the 4-variable Modification of Diet in Renal Disease study equation13:
The Model for End-Stage Liver Disease (MELD) score was determined.14 Complications of late HAT were recorded: bile duct abnormalities, intrahepatic collections (bilomas or abscesses), parahepatic collections, and ascites on imaging and positive blood and drain cultures. Bacteria were described as multidrug-resistant (MDR) according to recently the published guidelines.15 Continued antibiotic therapy was defined as antibiotic therapy up to and including the time of retransplantation.
Surgical technique in our unit has been previously described.1 Routine thromboprophylaxis was not standard practice, but thromboprophylaxis was administered to select patients when there was more than 1 arterial anastomosis or a small-diameter artery or when an aortic conduit was used for reconstruction. The choice of thromboprophylaxis was surgeon-dependent and also was not standardized during the follow-up time period.
After the diagnosis of late HAT, nonsurgical drainage of an intrahepatic collection (percutaneous) or the biliary tree (percutaneous or endoscopic) was considered on an individual basis and was performed when agreed by the multidisciplinary team. Drain fluid was sent for culture when the drain was first placed, and this was repeated only if there was clinical suspicion of active infection. Similarly, blood cultures were taken when clinically indicated. Antibiotics were given to clinically infected patients; the choice and duration were determined by clinical response, bacterial sensitivity testing, and the advice of a liver microbiologist. Immunosuppression until retransplantation was minimized, with many patients receiving a calcineurin inhibitor–free regimen. Thromboprophylaxis was not administered before retransplantation, and given to select patients after transplantation who were perceived to be at high risk for HAT recurrence. In this study, thromboprophylaxis was defined as any antiplatelet or anticoagulant treatment initiated in the immediate postoperative period and continued over the long term after transplantation.
The incidence of late HAT was calculated as the number of transplanted livers with late HAT divided by the total number of transplanted livers. Normally distributed continuous variables and nonparametric continuous variables were compared with the Student t test and the Mann-Whitney test, respectively. Survival was estimated using Kaplan-Meier plots with log-rank tests for differences. Patient survival was defined as the time from placement on the waiting list for liver transplantation for late HAT to patient death, last known follow-up, or retransplantation and as the time from transplantation for late HAT to patient death or last known follow-up. Graft survival was defined as the time from transplantation for late HAT to retransplantation, death, or last known follow-up. Cox proportional hazards analysis was used to identify variables predictive of mortality on the waiting list for liver transplantation for late HAT and after retransplantation for late HAT. In the multivariate models, clinically significant variables were included simultaneously. The number of included variables was limited by collinearity and patient numbers. When observing the association of thromboprophylaxis with survival the 3 patients who died within 7 days of retransplantation were excluded from all analyses. P < 0.05 was considered statistically significant at all times.
Data were analyzed with the SPSS 18 package. All values are expressed as means and standard deviations or as medians and interquartile ranges (IQRs) as appropriate.
Two thousand and forty-seven liver transplants were performed during the observation period. Late HAT affected 80 grafts in 78 patients (2 patients experienced the recurrence of late HAT after regrafting). Thus, the incidence of late HAT was 3.9%.
Of the 74 patients with adequate documentation to be included in the study, 49 patients (66.2%) were listed for re-transplantation. The reasons documented for not relisting the remaining 25 patients were as follows: too unwell (considered to be unlikely to survive retransplantation; 11 patients), too well (considered to have no survival benefit with retransplantation; 6 patients), comorbidities (6 patients), active alcohol abuse (1 patient), and living abroad (1 patient). The estimated 1-year survival rate was 0% for non-relisted patients who were too unwell, 100% for patients who were too well, and 62.5% for patients who had other reasons for not relisting (Supporting Fig. 1).
Table 1 lists the baseline characteristics of the patients (n = 49) at the time of listing for retransplantation for late HAT. The median time from transplantation to the diagnosis of late HAT was 216 days (IQR = 53-832 days), and the median time from the diagnosis of late HAT to relisting was 14 days (IQR = 3-55 days).
Table 1. Clinical Characteristics of the Patients at the Time of Listing for Retransplantation (n = 49)
The data are presented as means and standard deviations.
Complications of Late HAT at the Time of Relisting
The frequencies of complications of late HAT at listing are outlined in Table 2. Eleven patients (23.9%) had positive blood cultures, of which 63.6% were MDR organisms (MDR Escherichia coli, 2 patients; MDR Serratia species, 1 patient; MDR Klebsiella pneumonia, 1 patient; MDR Enterobacter cloacae, 1 patient; vancomycin-resistant enterococci, 1 patient; and methicillin-resistant Staphylococcus aureus, 1 patient). Twenty-one patients (42.9%) had undergone nonsurgical drainage of the biliary tree or a biloma (Table 3). The rate of positive drain cultures in this group was 47.4%; 44.4% of these cultures showed MDR organisms [MDR E. coli, 1 patient; MDR Serratia species, 1 patient; MDR K. pneumonia (extended-spectrum beta-lactamase), 1 patient; MDR E. cloacae, 1 patient; and vancomycin-resistant enterococci, 1 patient].
Table 2. Frequency of Complications of Late HAT at the Time of Listing for Retransplantation
NOTE: The data are presented as numbers and percentages. Percentages are based on the number of patients listed for retransplantation (n = 49) except where noted otherwise.
Bile duct abnormality on imaging
Positive blood cultures
MDR bacteria–positive blood cultures
Positive bile cultures
MDR bacteria–positive bile cultures
Table 3. Univariate and Multivariate Cox Regression Analyses of Variables Associated With Wait-List Mortality in Patients With Late HAT
Patients [n (%)]
HR (95% CI)
HR (95% CI)
NOTE: The reference group (relative risk = 1.00) was characterized as follows: no diabetes mellitus, MELD scores ≤ 12, no bile duct abnormalities on imaging, no intrahepatic collections, no biliary tree/biloma drainage, and no MDR bacteria–positive cultures/culture negative. Bolded values are significant. Percentages are based on the number of patients listed for retransplantation (n = 49) except where noted otherwise.
Nonsurgical biliary drainage
Any positive cultures
Time since listing (years)
Variables Associated With Wait-List Mortality in Patients With Late HAT
After listing for retransplantation for late HAT, 36 patients underwent retransplantation, 9 patients died on the list, and 4 patients remained active at the time of data analysis. The cause of death was sepsis in 8 patients and graft failure in 1 patient. The median time from listing to transplantation was 34 days (IQR = 12-147 days), and the median time from listing to death was 168 days (IQR = 72-266 days, P = 0.006). The estimated 1-year survival rate for patients with late HAT after listing for retransplantation was 53.7%.
Variables associated with wait-list mortality in the univariate analyses are outlined in Table 3. Only the presence of MDR bacteria–positive blood or drain cultures at the time of listing was predictive of death (P = 0.005). There was no clear association between the MDR subtype or site and mortality (MDR E. coli–positive blood cultures, 1 patient; MDR E. cloacae–positive blood and drain cultures and vancomycin-resistant enterococci–positive drain cultures, 1 patient; and vancomycin-resistant enterococci–positive blood cultures, 1 patient). The estimated 3-month and 1-year survival rates were 53.3% and 0%, respectively, for patients with MDR bacteria–positive cultures and 90.5% and 61.7%, respectively, for patients without MDR bacteria–positive cultures (cultures were negative or cultures were positive but MDR bacteria–negative; log-rank P < 0.001; Fig. 1). In a multivariate model adjusting for other clinically relevant variables, MDR bacteria–positive cultures remained predictive of wait-list mortality (P = 0.01).
Interventions for Late HAT by the Time of Retransplantation
Thirty-six patients listed for late HAT underwent retransplantation. By the time of retransplantation, 16 patients (44.4%) had undergone nonsurgical drainage of the biliary tree/biloma: 15 patients before listing and 1 patient on the list. Ten of these 16 patients had positive drain cultures (62.5%), and 5 had MDR organisms (31.3%). The frequency of positive blood cultures was 25.0% (9/36), of which 66.7% were MDR bacteria. In all, 8 patients had MDR bacteria–positive cultures by the time of retransplantation, including 3 patients who acquired MDR after listing [MDR E. coli, 1 patient; MDR Serratia species, 1 patient; MDR K. pneumonia, 1 patient; MDR K. pneumonia (extended-spectrum beta-lactamase), 1 patient; vancomycin-resistant enterococci, 4 patients; and methicillin-resistant S. aureus, 1 patient].
Twenty-nine of the 36 patients (80.6%) were hospitalized with pyrexia, and 11 patients (30.6%) had been an inpatient for longer than 2 weeks. Twenty-one patients (58.3%) were on continued antibiotic therapy at the time of retransplantation: 13 patients were on oral antibiotics and 8 patients were on intravenous antibiotics. Patients receiving antibiotics were more likely to have demonstrated positive blood or drain cultures (57.1% on antibiotics and 20.0% on no antibiotics, P = 0.03). Of patients with MDR positive cultures 87.5% were on antibiotics compared to 50.0% of patients without MDR positive cultures (P = 0.06).
Variables Associated With Mortality After Retransplantation for Late HAT
Fifteen patients died after retransplantation for late HAT. The overall estimated 1- and 5-year patient survival rates after retransplantation were 71.9% and 62.5%, respectively. Two patients underwent further transplantation for recurrence of late HAT; the first patient died of primary nonfunction, and the second died 30 days after regrafting. Consequently, the estimated 1- and 5-year graft survival rates were also 71.9% and 62.5%. The causes of death and indications for retransplantation are documented in Table 4.
Table 4. Causes of Death and Indications for Retransplantation After Retransplantation for Late HAT in All Patients, Patients With MDR Bacteria in Blood or Drain Cultures Before Retransplantation, Patients on Continued Antibiotic Therapy at the Time of Retransplantation, and Patients Who Received Long-Term Thromboprophylaxis After Retransplantation
All Patients (n = 36)
MDR-Positive Patients (n = 8)
Patients on Antibiotics (n = 21)
Patients on Thromboprophylaxis (n = 18)
NOTE: The data are listed as numbers of patients. Some patients are listed in more than 1 group.
Multiorgan failure (cause not documented)
Retransplantation for HAT
Variables associated with mortality after transplantation for late HAT are shown in Table 5. In the univariate analysis, MDR bacteria–positive blood or drain cultures before retransplantation was a predictor of death (P = 0.04). Five of 8 patients with MDR bacteria–positive cultures died (Table 4). There was no clear association between the MDR subtype or site and mortality [MDR E. coli positive blood and drain cultures, 1 patient; MDR K. pneumonia (extended-spectrum beta-lactamase)–positive drain cultures, 1 patient; MDR K. pneumonia–positive blood and drain cultures and vancomycin-resistant enterococci–positive drain cultures, 1 patient; vancomycin-resistant enterococci–positive drain cultures, 1 patient; and methicillin-resistant S. aureus–positive blood cultures, 1 patient]. The estimated 1- and 5-year survival rates were 37.5% and 37.5%, respectively, for patients with MDR bacteria–positive cultures and 81.8% and 69.4%, respectively, for patients without MDR bacteria–positive cultures (log-rank P = 0.03; Fig. 2).
Table 5. Univariate and Multivariate Cox Regression Analyses of Variables Associated With Mortality After Retransplantation for Late HAT
Patients [n (%)]
HR (95% CI)
HR (95% CI)
NOTE: The reference group (relative risk = 1.00) was characterized as follows: no diabetes mellitus, no active smokers, MELD scores ≤ 12, no antibiotics at the time of retransplantation, no positive cultures, and no MDR bacteria–positive cultures. Bolded values are significant. Percentages are based on the number of patients who underwent retransplantation (n = 36) except where noted otherwise.
Log eGFR (mL/minute/1.73 m2)
Continued antibiotic therapy
Any positive cultures
Time since transplantation (years)
Eleven of 21 patients receiving continued antibiotic therapy at the time of retransplantation died (Table 4). The estimated 1- and 5-year survival rates were 56.4% and 44.0%, respectively, for patients on antibiotics and 93.3% and 86.7%, respectively, for patients not on antibiotics (log-rank P = 0.02). When patients were stratified by the MELD score, there was a strong association between continued antibiotic therapy and survival in patients with a MELD score > 12 (estimated 5-year survival: 88.9% with no antibiotics and 13.6% with antibiotics), which was not observed in patients with a MELD score ≤ 12 (estimated 5-year survival: 100.0% with no antibiotics and 80.0% with antibiotics, log-rank P < 0.001; Fig. 3).
In a multivariate model adjusting for clinically relevant variables, MDR bacteria–positive cultures (P = 0.047) and continued antibiotic therapy (P = 0.001) at the time of retransplantation were predictive of posttransplant death (Table 5).
Eighteen of 33 patients (54.5%) surviving longer than 7 days after retransplantation received long-term post retransplantation thromboprophylaxis. The types of thromboprophylaxis were aspirin alone in 10 patients, warfarin alone in 3 patients, aspirin and warfarin combined in 3 patients, aspirin and dipyridamole in 1 patient, and warfarin followed by clopidogrel in 1 patient. Four patients who received long-term thromboprophylaxis died (Table 4). The estimated 1- and 5-year survival rates were 82.6% and 76.7%, respectively, for patients who received thromboprophylaxis and 73.3% and 60.0%, respectively, for patients who did not receive thromboprophylaxis (log-rank P = 0.21). In a multivariate model (adjusting for the relevant clinical variables; age, diabetes mellitus, smoking status, MELD score, antibiotics, MDR bacteria positive cultures; data not shown), long-term thromboprophylaxis from the time of retransplantation for late HAT was associated with a trend toward better patient survival [hazard ratio (HR) = 0.19, 95% confidence interval (CI) = 0.03-1.11, P = 0.07].
In this large single-center study, we have examined for the first time the wait-list mortality and post–liver retransplant outcomes of patients listed for late HAT. Importantly, we have shown that select patients with late HAT–induced liver failure or ischemic complications have satisfactory survival following liver retransplantation. This confirms that liver retransplantation may be an appropriate treatment option for patients with late HAT. Moreover, we have identified patient risk factors for wait-list mortality and death after retransplantation for late HAT. These findings should aid future decision making regarding organ allocation and prioritization for such patients.
HAT is generally subdivided into 2 categories, early and late, according to the time from transplantation to diagnosis. Most publications have defined late HAT arbitrarily as the occurrence of HAT more than 4 weeks after transplantation.1, 2, 3, 8-12 Early HAT frequently presents with profound graft dysfunction and sepsis, and urgent revascularization or retransplantation may be indicated.8, 16 In late HAT, the presentation is less predictable probably reflecting collateral presence and level of immunosuppression.17 The diagnosis may be made incidentally on imaging or after the development of ischemic complications of varying severity.6 Consequently, the management of patients with late HAT is challenging and is not evidence-based. In particular, it remains unclear who will benefit from retransplantation.
The results of our study suggest that intercurrent infection is the critical determinant of outcome in patients listed and re-transplanted for late HAT. In individuals with focal infection controlled by antibiotic therapy, despite no clinical evidence of sepsis, there was an increased risk of mortality after retransplantation. Continued antibiotic therapy when the MELD score was greater than 12 was associated with estimated 1- and 5-year survival rates of only 36% and 14%, respectively. It is possible that previously contained pathogens may have been released systemically during surgery and may have produced sepsis in these patients. On the other hand, chronic infection may have been a consequence of, or resulted in, greater graft dysfunction and undernutrition and may have impacted on the severity of parahepatic adhesions.18, 19
Similarly, the presence of MDR bacteria–positive drain or blood cultures was associated with greater wait-list mortality and inferior outcomes after regrafting. MDR bacteremia has been shown to be associated with a worse prognosis in infected liver transplant recipients as well as general critical care patients.20, 21 An explanation for this increased risk of death is the causative pathogen itself.18 Alternatively, inappropriate antimicrobial therapy during the course of the illness may have resulted in more extensive infection.18 The acquisition of MDR bacteria may have been a sign of a longer duration of infection or a more complex pre-HAT illness.18, 20
Long-term thromboprophylaxis from the time of retransplantation for late HAT was associated with a trend toward a reduced risk of death. Unfortunately, because of the relatively small patient numbers, we were not able to assess the impact of antiplatelet therapy versus anticoagulation. Patients were more likely to receive thromboprophylaxis in later years (Supporting Table 1). Consequently, it is possible that the perceived favorable outcomes were due to improved surgical techniques and intensive care management. However, aspirin prophylaxis has previously been linked to a reduced incidence of late HAT in liver transplant recipients with no apparent increased rate of adverse events.9 Therefore, we recommend long-term thromboprophylaxis with antiplatelet therapy for patients retransplanted for late HAT.
The study has some additional potential limitations that should be mentioned. First, all patients in our unit undergo a thorough clinical assessment before listing for liver transplantation. Those considered unlikely to survive the surgery are not relisted. Similarly, patients considered to have no survival benefit with retransplantation are managed conservatively. Therefore, our observations concern a highly select group of patients and should not be extrapolated to all patients with a diagnosis of late HAT. Second, we included patients diagnosed with late HAT during a 16-year time period. Because of the inevitable changes in medical and surgical practice the long time duration may have influenced our results. Nevertheless, we were unable to demonstrate any significant time effect on statistical analyses (Tables 3, 5 and Supporting Table 1). Third, the management of patients with late HAT and the use of nonsurgical interventions such as antimicrobials and drainage procedures were not standardized. This was a single-center study, and in our unit, all decisions are made by the multidisciplinary team; this limits variability. Finally, it is always difficult to ensure the accuracy of clinical data in retrospective studies. For example, it is possible that some patients may have been commenced on treatment for infections locally prior to transfer and subsequently falsely labeled as culture-negative. In addition, the incidence of the recurrence of HAT may have been underestimated if patients died of sepsis or multiorgan failure outside the Liver Unit. To the best of our knowledge, the presented data are accurate.
The findings of our study have important implications for patient management. We have demonstrated that liver retransplantation is an appropriate treatment option for select patients with late HAT. However, the presence of active intercurrent infection and MDR bacteria–positive cultures should be considered when decisions regarding organ allocation and prioritization are being made in this setting. Patients with a MELD score greater than 12 who require continued antibiotic therapy at the time of retransplantation may be a particularly high-risk group. We recommend long-term thromboprophylaxis with antiplatelet therapy after retransplantation for late HAT.
In conclusion, in this large single-center study, we have examined for the first time the wait-list mortality and post–liver retransplant outcomes of patients listed for late HAT. We have shown that select patients with late HAT–induced liver failure or ischemic complications have satisfactory survival after liver retransplantation. We have identified patient risk factors for wait-list mortality and death after retransplantation for late HAT. These findings should aid future decision making regarding organ allocation and prioritization.
The authors thank the statistician James Hodson for his helpful advice concerning the analysis of the data.