Telephone: 902-473-2898; FAX: 902-473-2783
Recurrent hepatitis C posttransplant: Early preservation injury may predict poor outcome
Article first published online: 28 DEC 2005
Copyright © 2005 American Association for the Study of Liver Diseases
Volume 12, Issue 1, pages 134–139, January 2006
How to Cite
Watt, K. D.S., Lyden, E. R., Gulizia, J. M. and McCashland, T. M. (2006), Recurrent hepatitis C posttransplant: Early preservation injury may predict poor outcome. Liver Transpl, 12: 134–139. doi: 10.1002/lt.20583
- Issue published online: 28 DEC 2005
- Article first published online: 28 DEC 2005
- Manuscript Accepted: 5 AUG 2005
- Manuscript Received: 8 MAR 2005
Organ cold/warm ischemia is thought to be a risk factor for increased severity of recurrence of hepatitis C (HCV) post liver transplantation. We had noted some HCV patients with preservation injury (PI) to have particularly poor outcomes. Our goal was to determine if PI on biopsy in HCV patients is associated with earlier, more rapidly progressive recurrence or graft and patient survival. Sixty-nine patients from the University of Nebraska transplant database were included: 23 HCV patients with PI (group = 1), 23 non-HCV patients with PI (group = 2), and 23 HCV patients without PI (group = 3). Patient groups were matched for gender, age, immunosuppression, and time of transplantation for analysis. No difference in time to recurrence was noted between HCV groups (256 vs. 316 days posttransplant). More patients in group 1 had progression to stage 3 or 4 fibrosis, compared to group 3 (43 vs. 9%, P = 0.02). One-year survival for groups 1, 2, and 3 was 78, 82, and 100% respectively, whereas 3-yr survival was 59, 82, and 88% (group 1 vs. group 2 or 3 respectively, P = 0.0055). There was no difference in survival between groups 2 and 3. Patients in group 1 that received antiviral treatment had improved survival, compared to those who did not (P = 0.012). Risk factors for poor survival on univariate analysis included severity of PI (Relative Risk = 2.78, P < 0.001) and donor age of >55 (P = 0.014). Multivariate analysis shows HCV is the most important factor. In conclusion, HCV transplant patients with evidence of early PI on biopsy have poorer survival outcomes than non-HCV transplant patients with PI or HCV transplant patients without PI. Consideration for antiviral therapy early in the posttransplant course may be warranted in this subset of patients. Liver Transpl12:134–139, 2006. © 2005 AASLD.
Preservation injury (PI), otherwise known as transport injury, harvesting injury, or reperfusion injury, occurs not infrequently in liver transplantation. It has been reported to occur in approximately 17% of cases, but rates as high as 50% have been reported.1, 2 It usually is suspected with abnormal liver enzymes in the early posttransplant period, but requires histological confirmation with a liver biopsy.2 Histological findings of neutrophilic infiltration, microvesicular steatosis, and hepatocyte cytoaggregation occur early and progress to centrilobular necrosis, while hepatocyte swelling and cholestasis occur later in the process. Mild cases resolve spontaneously, but more severe cases may have residual damage.2
Cold ischemia and, more importantly, warm ischemia time during implantation are thought to be risk factors for recurrent hepatitis C (HCV) posttransplantation.3–5 Baron et al.5 had shown specifically that duration of ischemic rewarming was significantly associated with severity of recurrent HCV.
We have noted a few cases of rapidly progressive recurrent HCV developing in patients with early (6-15 days) posttransplant biopsies exhibiting changes consistent with moderate to severe preservation or transport injury. We sought to determine if early preservation or transport injury in HCV patients was associated with earlier recurrence, more rapidly progressive recurrence, or graft and patient survival.
We reviewed the University of Nebraska pathology department database for all patients with PI posttransplantation during the years 1996 to 2003. The University of Nebraska transplant database was then reviewed to find the underlying disease etiologies, demographic details, and outcomes of patients identified.
Sixty-nine patients were included in the analysis. All patients with both PI and underlying HCV were identified (n = 23 patients, group 1). Two control groups each consisted of 23 patients with PI and disease etiologies other than viral hepatitis (group 2), as well as 23 patients with HCV and no evidence of early PI (group 3). The HCV control patients were matched for age (within 5 yr), year of transplant (within 1 yr), and then gender. The non-HCV with PI patients were also matched for age and year of transplant, and attempts were made to match for gender (difficult due to the small population).
Data retrieved included recipient age and gender, date of transplant, disease etiology, donor age and gender, donor HCV status, and posttransplant immunosuppression for all patients. Donor steatosis percentage was available on 22 of 23 patients in group 1 and all patients in groups 2 and 3. Days to histological PI, severity of PI (see definition below), episodes of acute cellular rejection, days to virologic and histological recurrence of disease, days to stages 2, 3, and 4 fibrosis documented on biopsy, days until death, and cause of death were also determined. Pretransplant HCV genotype and viral load were only documented for 14 and 12 patients in groups 1 and 3, respectively; therefore, these data were not analyzed. Posttransplant viral load was documented in 18 of 23 in group 1 patients and 16 of 23 in group 3 patients. Additional biopsy data (after initial biopsy of PI) were available for 21 of 23 patients in group 1 and 14 of 23 patients in group 3. A total of 2 patients in group 1 and 6 patients in group 3 did not have documented viral or histologic recurrence of disease. This is not to imply these patients were HCV ribonucleic acid negative, but simply did not have indication for biopsy or did not have an HCV RNA entered into the database. HCV RNA levels were only performed during this time frame when suspicion was high for recurrent disease. It was the practice at this institution to biopsy for histologic recurrence at the time of abnormal liver chemistry (>1.5-2 times normal). Additional biopsies were obtained when clinically indicated (e.g., further abnormal liver chemistry >1.5-2 times previous levels). Protocol biopsies were not performed.
At present, there is no clearly defined grading system for severity of PI. We divided histologic findings into 2 grades, mild and moderate to severe. Mild PI is defined as neutrophilic infiltrate, hepatocyte necrosis, cytoaggregation, and pyknosis, as well as microvesicular steatosis involving <30% of hepatocytes. Progression to more severe injury was defined as centrilobular or zonal necrosis, hepatocyte ballooning, sinusoidal dilation, cholestasis, and fibrosis.1, 2 Severity was only analyzed in the multivariate analysis. All other results discussed are based on the presence or absence of PI on early histologic exam.
The Kaplan-Meier method was used to estimate survival distributions for overall survival, and the log-rank test was used to compare distributions by group. The Kaplan-Meier method and the log-rank were used to assess the associations of risk factors with outcomes of overall survival. Cox regression was used to fit a multivariate model that included patient group and PI severity.
Patient and donor characteristics are displayed in Table 1. There was no statistical difference between the groups, except for gender in group 2, as discussed in the methodology. Histological recurrence of HCV was documented in 21 of 23 (91%) patients in group 1 and 14 of 23 (61%) patients in group 3 (an additional 3 patients in group 3 had documented virologic recurrence of HCV, total 74%). There was no statistical difference in mean time to histological recurrence of HCV between groups 1 and 3 (247.6 vs. 316.8 days, P = 0.10). The median time to histological recurrence was 103 vs. 213 days in groups 1 and 3, respectively. These were not protocol biopsies; thus, the rate of and time to histological recurrence are not truly known.
|Group 1 HCV/PI (N = 23)||Group 2 Non-HCV/PI (N = 23)||Group 3 HCV/No PI (N = 23)||P value|
|Gender: N (%)|
|Male||6 (26%)||14 (61%)||6 (26%)||0.03|
|Female||17 (74%)||9 (39%)||17 (74%)|
|Donor HCV: N (%)|
|Negative||20 (87%)||23 (100%)||20 (91%)||0.27|
|Positive||3 (13%)||0||2 (9%)|
|Immunosuppression: N (%)|
|FK||12 (52%)||16 (70%)||15 (65%)||0.55|
|CYA||11 (48%)||7 (30%)||8 (35%)|
|PI severity: N (%)|
|Mild||18 (78%)||22 (96%)||NA||0.22|
|Moderate/Severe||5 (22%)||1 (4%)|
|Donor fat: N (%)|
|0||7 (37%)||17 (77%)||14 (61%)|
|<5%||5 (26%)||4 (18%)||4 (17%)||0.17|
|<10%||5 (26%)||1 (5%)||4 (17%)|
|>20%||1 (5%)||0||1 (4%)|
Significantly more patients (10/23, 43%) in group 1 had biopsy documented progression to stage 3 or 4 fibrosis than in group 3 (2/23, 9%, P = 0.02) (Table 2). Progression occurred within 69-994 days in group 1 and 360-907 days in group 3. It should be noted that an additional 3 patients in group 1 did not have biopsy documented progression to stage 3 or 4 fibrosis, but did have end-stage liver disease (ESLD) documented at the time of death; thus, the difference is likely even more significant.
|Group 1 HCV/PI (N = 23)||Group 2 Non-HCV/PI (N = 23)||Group 3 HCV/No PI (N = 23)||P value|
|Histological Recurrence (days)||N = 21||N = 2||N = 14||0.10 (Group 1 vs. Group 3)|
|Stage 3 or 4: N (%)|
|Absent||13 (57%)||NA||21 (91%)||0.02|
|Present||10 (43%)||2 (9%)|
|ACR, N (%)|
|0||11 (48%)||10 (43%)||18 (78%)||0.10*|
|1||7 (30%)||10 (43%)||3 (13%)|
|2+||5 (22%)||3 (13%)||2 (9%)|
Survival distributions were examined for the 3 groups of patients and are displayed in Figure 1. There is a statistically significant difference in survival for the 3 patient groups (P = 0.0055). Patients with HCV and PI (group 1) had significantly higher mortality. Respective 1-yr survivals for groups 1, 2, and 3 were 78% (confidence interval (CI) = 61-95%), 82% (CI = 67-98%), and 100%, whereas 3-yr survivals were 59% (CI = 39-80%), 82% (CI = 67-98%), and 88% (CI = 73-100%). There is no statistically significant difference in survival between groups 2 and 3. Of the 12 patients in group 1 that died, 10 had ESLD at the time of death, 1 had a myocardial infarction, and 1 had sepsis in addition to ESLD at the time of death. The remaining 2 patients died of heart failure and pneumonia. Causes of death for group 2 included posttransplant lymphoproliferative disorder (1), sepsis (2), and ESLD (1). Group 3 patients died from ESLD and hepatocellular carcinoma (1) and suicide (1).
There was only 1 patient in the HCV with PI group listed for retransplant at the time of the data analysis; therefore, graft survival data were not significantly different than patient survival data (data not shown). This patient was alive at the time of data analysis, but has subsequently died.
Patients in group 1 that received treatment for the recurrent HCV (n = 9) demonstrated significantly increased survival compared to those who did not receive treatment (n = 14) (Fig. 2). This is despite the fact that 7 of 9 patients were nonresponders to treatment (positive HCV RNA at any point up to 6 months posttreatment). Four of these treated patients had stage 3 or 4 fibrosis at the initiation of treatment. Of the 5 patients with stage 0-2 fibrosis, 2 progressed to stage 3 or 4 fibrosis despite treatment (or even sustained virologic response in 1 case). One patient with sustained virologic response and 1 nonresponder died with ESLD (808 and 613 days, respectively, after treatment initiation). Six patients that received treatment were alive at the time of analysis (1,040-2,400 days post-OLT); 1 additional patient was still undergoing treatment at the time of analysis.
In the HCV control group, 17 of 23 patients had documented HCV recurrence (14 histological), with 11 of 17 receiving antiviral treatment. Three patients had documented sustained virologic response. Ten treated patients were alive at the time of analysis (range = 280-2,224 days posttransplant).
The following risk factors were not associated with survival: patient gender (P = 0.50) or patient age (≤55 vs. >55, P = 0.26), donor gender (P = 0.26), donor HCV status (P = 0.51), and immunosuppression (P = 0.24). There was no significant difference in frequency of acute cellular rejection between the 3 groups (Table 2). Acute cellular rejection was not a risk factor for death (P = 0.62), nor associated with progression to stage 3 or 4 disease (P = 0.79).
Risk factors (besides group) associated with decreased survival on univariate analysis included PI severity at initial biopsy (P = 0.001) and donor age of >55 (P = 0.014). Donor steatosis percentage was not a significant predictor of survival, but donor livers with >10% steatosis were few. On multivariate analysis between groups 1 and 2 (both with PI), patients with HCV and PI had marginally significant (P = 0.08) poorer survival relative to non-HCV patients with PI after controlling for PI severity and donor age (Table 3). Multivariate analysis between groups 1 and 3 (both with HCV) showed a significantly increased risk of death (P = 0.02) in patients with PI vs. no PI after controlling for donor age.
|Variable||Parameter estimate||Hazard ratio||95% CI||P value|
|Group 1 vs. 2|
|Group 1 vs. 3|
PI is associated with hepatocellular death, followed by cellular proliferation. In the setting of HCV, it is thought that the virus may be able to incorporate into these proliferating cells.6 In the study by Baron et al.,5 graft rewarming duration, but not cold preservation duration, was correlated with increased risk of significant recurrent HCV. Rewarming duration during implantation of <10 minutes was associated with minimal recurrence, whereas a duration of >70 minutes was associated with moderate to severe recurrent hepatitis. This study did not use protocol biopsies, and there were no data on survival statistics. This finding was supported by Velidedeoglu et al.'s7 United Network for Organ Sharing database study showing that a warm ischemia time of >90 minutes was associated with decreased graft survival in HCV-positive individuals. Prolonged ischemic time is a risk factor for PI.2 PI in the early posttransplant course has been thought to be associated with increased risk for subsequent rejection and biliary complications.1, 8, 9 In addition, 1 study2 has suggested decreased overall survival of patients diagnosed with PI. Our study suggests that PI also increases the risk of poor outcomes and death in the HCV posttransplant population.
PI may be a result of (or a surrogate marker for) a combination of several perioperative factors, including prolonged ischemic time ± donor factors, such as age and steatosis. All of these factors have been described as potential risk factors for poor outcomes in recurrent HCV.3, 4 None of these individual risk factors can identify an individual at highest risk for poor outcomes posttransplant. This study suggests that PI injury may be an easily identified surrogate marker of a combination of known risk factors for poor outcomes in HCV patients. Thus, PI found early posttransplantation may identify a cohort of HCV patients at highest risk for poor outcomes.
This study was unable to confirm earlier recurrence of HCV in the setting of early PI, due to the retrospective nature of the study and lack of protocol biopsies. We do show, however, that histological progression to stage 3 or 4 fibrosis occurred more frequently in the HCV patients with early PI on biopsy. As most of the deaths occurring in group 1 were ESLD related (or at least had documented ESLD at the time of death in 2 cases), it appears that the recurrent disease is more rapidly progressive in this group of patients.
Donor age of >55 and PI severity were risk factors associated with increased mortality. Donor age may reflect increased PI risk; thus, the 2 risks are not exclusive. Three patients with moderate to severe PI early posttransplant developed rapidly progressive graft failure and eventually died from recurrent HCV. All 3 received organs from donors who were >55 yr old. Two of these patients died within 100 days of transplant. Two patients in group 1 had moderate PI, with 1 dying of ESLD. All other patients had only mild PI described on biopsy. It is thought that mild PI spontaneously resolves, with minimal clinical impact.2 Our results suggest that this is not the case in the setting of HCV. Even if we analyze survival in all groups with the moderate to severe PI patients excluded (n = 5 in group 1 and n = 1 in group 2), there is a borderline significant (P = 0.06, data not shown) difference, despite the even smaller sample size. These deaths appear to be related to rapid progression of recurrent HCV, as most patients had documented ESLD at the time of death. The reason for this finding is not clear. It is possibly related to hepatocyte proliferation after PI, but most of our patients experienced mild injury, and significant proliferation would not be expected.
In the multivariate analysis, once we adjust for HCV status, severity is no longer a significant predictor of survival among patients with PI; HCV status is a more important predictor of outcome. Thus, in patients with PI, those with underlying HCV had marginally significant poorer survival (Hazard Ratio = 2.78, P = 0.08), relative to non-HCV patients after controlling for PI severity and donor age. Given the small sample size, this may be suggesting a clinically relevant finding. When the 2 HCV groups are analyzed, the presence of PI is a significant risk for death with a hazard ratio of 5.66 (P = 0.02) after adjusting for donor age. Thus, it appears that patients with HCV and any degree of PI are at increased risk of death, and those with moderate to severe PI are at highest risk.
Patients with HCV and PI had significantly improved survival if antiviral therapy was initiated, regardless of the response to treatment. Given the retrospective nature of the study, this may be a reflection of the patients' clinic status to tolerate treatment (i.e., those with significant liver dysfunction were less likely to receive treatment). However, the survival benefit does not appear to be related to viral eradication (1 of 2 patients that died had sustained viral response, defined as negative HCV RNA 6 months posttreatment) and may reflect a direct treatment effect. Further prospective data are needed to know the true effect.
Fibrosing cholestatic HCV can occur in 5 to 7% of patients transplanted for HCV-related cirrhosis.10–12 This entity is characterized by periportal fibrosis, prominent cholestasis, and hepatocyte swelling progressing to centrilobular ballooning degeneration and bridging fibrosis/cirrhosis, with only minimal to mild inflammation.10–12 Three patients in group 1 with HCV/PI developed rapidly progressive liver failure within weeks of transplantation. Their initial biopsies within 5-15 days of transplant were consistent with moderate to severe PI, whereas subsequent biopsies initially read as PI eventually met the criteria of fibrosing cholestatic HCV. This entity is generally seen after the first few months posttransplantation10 and has not been reported to occur within days of transplant. As fibrosing cholestatic HCV and severe PI have overlapping histologic features, it is not clear if they demonstrate a progression from PI to fibrosing cholestatic HCV or if they are, in fact, the same process. One study suggesting the presence of a late or prolonged PI cholestasis effect (up to 7-yr posttransplant) certainly makes the differentiation of the 2 processes difficult. A clear characterization of the histologic progression of both entities is needed.
Cold ischemic time and/or intraoperative rewarming duration has been associated with HCV recurrence posttransplant in previous studies.3–5 This study supports these findings and helps to identify PI as a significant risk factor for poor outcomes in HCV transplant patients. This is the first study to our knowledge to show an association between histological evidence of early PI and decreased survival posttransplant in patients with HCV. Minimizing ischemic times, and perhaps donor age limitations, may reduce the likelihood of PI, which may impact patient survival after transplantation for HCV-related liver disease.
Many risk factors for poor outcomes posttransplant in the HCV patients have been identified. None of these risk factors individually identify a cohort of patients at high risk for progressive disease or poor survival. Our data suggest that assessment for PI posttransplant may be warranted in patients with HCV. The presence of PI on early posttransplant biopsies in HCV patients may identify a cohort of patients at high risk for poor outcomes. These patients should be considered for antiviral medication as soon as can be tolerated.
- 12Hyperbilirubinemia and cholestatic liver injury in hepatitis C-infected liver transplant recipients. Am J Gastroenterol 2000; 95: 753–759., , , , , ,Direct Link: