This paper was accepted for presentation at the 41st Annual ICAAC in September 2001, Chicago, IL: abstract no. 462.
Interactions Between Cytomegalovirus, Human Herpesvirus-6, and the Recurrence of Hepatitis C After Liver Transplantation
Article first published online: 23 MAY 2002
American Journal of Transplantation
Volume 2, Issue 5, pages 461–466, May 2002
How to Cite
Humar, A., Kumar, D., Raboud, J., Caliendo, A. M., Moussa, G., Levy, G. and Mazzulli, T. (2002), Interactions Between Cytomegalovirus, Human Herpesvirus-6, and the Recurrence of Hepatitis C After Liver Transplantation. American Journal of Transplantation, 2: 461–466. doi: 10.1034/j.1600-6143.2002.20511.x
- Issue published online: 23 MAY 2002
- Article first published online: 23 MAY 2002
- Received 7 September 2001,revised and accepted for publication 14 January 2002
- hepatitis C;
- human herpesvirus-6
Recurrence of hepatitis C (HCV) following liver transplantation is common. Herpesvirus reactivation following transplant may have an immunomodulatory effect resulting in increased HCV replication. We studied whether cytomegalovirus (CMV) and human herpesvirus-6 (HHV-6) may be associated with HCV recurrence and viral load after transplant. We prospectively followed 66 HCV liver-transplant recipients with serial viral load testing for CMV and HHV-6. Infection and viral load were correlated with the development of biopsy-proven HCV recurrence and HCV viral loads. Histologic recurrence of HCV occurred in 41/66 (62.1%) patients. In the primary analysis, CMV infection and disease, and HHV-6 infection were not associated with HCV recurrence. Peak CMV and HHV-6 viral loads were not significantly different in patients with and without recurrence. No correlation was observed between HCV viral loads at 1 and 3 months post-transplant and peak HHV-6 or CMV viral loads. In a subgroup analysis, HHV-6 infection was associated with the development of more severe recurrence (hepatitis and/or fibrosis score ≥ 2) (p = 0.01). Also, fibrosis scores at last follow up were higher in patients with CMV disease (1.67 vs. 0.56; p = 0.016) and in patients with HHV-6 infection (1.18 vs. 0.55; p = 0.031). In conclusion, HHV-6 and CMV infection and viral load were not associated with increased overall rates of HCV recurrence or HCV viral load after liver transplantation but may be associated with more severe forms of recurrence.
Hepatitis C virus (HCV) related chronic liver disease has become one of the leading indications for liver transplantation. However, persistence of HCV is almost universal after transplantation. Biochemical and histologic evidence of recurrent HCV is common during the first year after transplantation, and increases further with longer follow up (1, 2). Bridging fibrosis or cirrhosis develops in approximately 20–30% of patients with recurrent HCV by 5 years. A better understanding of factors favoring recurrence is important in developing strategies to improve outcomes. One of the factors promoting recurrence of HCV may be the reactivation of one or more of the herpesgroup viruses after transplantation.
Cytomegalovirus (CMV) is the herpesvirus that most commonly results in symptomatic disease after transplant, but is also proposed to have an indirect immunomodulatory effect in transplant recipients. Cytomegalovirus infection has been associated with bacteremia (3), invasive fungal disease (4), Epstein–Barr virus-related post-transplant lymphoproliferative disease (5) as well as acute and chronic allograft injury (6). In addition, CMV replication may favor the recurrence of HCV through similar immunomodulatory effects (7, 8).
Another β-herpesvirus, human herpesvirus-6 (HHV-6), also commonly reactivates following transplantation with HHV-6 infection, as reported in 31–55% of solid-organ transplant recipients (9). Human herpesvirus-6 has been implicated as a cause of febrile viral syndromes, hepatitis, pneumonitis, and encephalitis in this patient population (9). This virus may also have an immunomodulatory role, and in-vitro infection of T cells by HHV-6 leads to a reduction of both IL-2 synthesis and T-cell proliferation (10). Human herpesvirus-6 has been implicated as a risk factor for the development of CMV disease, graft dysfunction and other opportunistic infections (11–13). Given the evidence suggesting an immunomodulatory effect of HHV-6 infection, this virus may also play a role in promoting HCV replication after transplant, although no studies have been published that specifically assess this hypothesis.
In this study, we tested the hypothesis that CMV and HHV-6 infection and viral load are associated with the histologic recurrence of HCV and HCV viral load after liver transplant. We employed molecular diagnostic tests to determine viral loads for all three viruses and liver biopsies to assess for recurrence of HCV.
Adult patients undergoing liver transplantation for HCV were enrolled between June 1997 and March 2000. These patients were part of a larger trial assessing the clinical impact of HHV-6 infection following liver transplantation (11). The study was approved by the institutional review board. After informed consent was obtained, blood samples for CMV and HHV-6 viral load were collected at baseline and 1–2 weekly intervals until 12 weeks post-transplant. The only antiviral prophylaxis was in patients who were donor (D) seropositive and recipient (R) seronegative (D+/R–) for CMV. These patients received 12 weeks' ganciclovir, and then samples were collected for an additional 8 weeks. Blood was collected for HCV viral load at 1 and 3 months post-transplant. The treating physician was blinded to the test results, and therefore patients did not receive pre-emptive antiviral therapy based on the results. All laboratory testing was done by technologists blinded to the patients' clinical status.
Human herpesvirus-6 PCR. DNA was extracted from peripheral blood leukocytes using Puregene DNA Isolation kits (Gentra Systems, Minneapolis, MN). For quantitation and standardization, an HHV-6 DNA control (Advanced Biotechnologies Inc., Columbia, MD) with a known copy number was used. Polymerase chain reaction (PCR) was carried out using a HHV-6 Custom Probe/Primer set (Digene Inc., Silver Spring, MD, which uses primers directed against the U1102 region of HHV-6 variant A and the Z29 region of HHV-6 variant B. Primer sequences(5′ to 3′) were GTT CCA GGC GGC ATG AAT TC and Biotin-ACA CGG CCT CTC TAC ATC AC. Amplified DNA was detected and quantitated using the SHARP SignalTM System (Digene Inc., Silver Spring, MD), which is a colorimetric absorbance assay. Quantitation of HHV-6 DNA was calculated by using known dilutions of the HHV-6 DNA control to generate a standard curve, plotting the positive sample signals on this curve and extrapolating the copy number. Results were reported as log10 copies/µg input DNA.
Quantitative CMV and HCV-PCR were performed on plasma samples according to manufacturer's instructions using the Cobas Amplicor Monitor Test (Roche Diagnostic Systems, Inc., Branchburg, NJ). Results were recorded as the number of viral copies/mL plasma. For CMV, the lower limit of detection was 400 copies/mL and for HCV was 1000 copies/mL.
Cytomegalovirus infection and disease. Cytomegalovirus infection was defined as any patient with a positive plasma PCR regardless of symptoms. Active CMV disease was diagnosed based on biopsy evidence of tissue invasion or in patients with fever, a positive plasma PCR or antigenemia assay, plus either leukopenia (WBC < 3.5 × 109/L), thrombocytopenia (platelet count < 100 × 109/L), or new onset arthralgias, myalgias and malaise (CMV viral syndrome) (11).
Human herpesvirus-6 infection
In an attempt to distinguish active from latent viral infections, HHV-6 viral load results were analyzed for samples taken between 0 and 7 days post-transplant in a larger cohort of 200 liver-transplant patients (11). These results are described in detail elsewhere (11), but all patients who had a positive result from day 0–7 had viral loads < 2-log10 copies/µg input DNA (median 1.4 log10). Therefore, active HHV-6 infection was defined as a viral load ≥ 2-log10 copies/µg input DNA. Although somewhat arbitrary, this cut-off was chosen as a conservative estimate of patients likely to have viral reactivation. However, as this may underestimate the rate of HHV-6 infection, any positive PCR result (including < 2-log10 copies) was also analyzed as a continuous variable (i.e. HHV-6 viral load) for its association with HCV recurrence.
Liver biopsies were performed when clinically indicated. Those patients with histologic evidence of hepatitis in the absence of rejection or serologic evidence of another hepatitis virus infection were diagnosed as having recurrent HCV (14). Biopsies were interpreted by a pathologist blinded to the results of CMV, HHV-6, and HCV viral load testing. A standardized semiquantitative score for grading the severity of hepatitis based on the extent of portal inflammation, lobular degeneration and necrosis and periportal necrosis (0 = none, 1 = mild, 2 = moderate, 3 = severe) was used (METAVIR scoring system) (14). In this system mild and moderate hepatitis is differentiated based on the degree of piece-meal necrosis and lobular necrosis. For example, a biopsy with either moderate piece-meal necrosis or moderate lobular necrosis would be classified at least as moderate activity; Bedossa et al. fully describe this activity score (14). A separate score was given for the fibrosis stage (0 = none, 1 = fibrous portal expansion, 2 = portal and lobule fibrosis without bridging, 3 = bridging fibrosis, 4 = cirrhosis) (15).
All statistical analysis was performed using SPSS (Chicago, IL, USA) version 8.0. The association of various factors with the recurrence of post-transplantation HCV was determined using the χ2-test or Fisher's exact test for categorical variables and the Mann–Whitney U-test for continuous variables. Cox's proportional hazard models were used to determine the degree of association between a number of covariates and the time to HCV recurrence.
Sixty-six patients undergoing liver transplant for HCV infection were followed prospectively (47 male, 19 female). Maintenance immunosuppression consisted of cyclosporin or tacrolimus and prednisone (n = 30), cyclosporin or tacrolimus plus prednisone and either mycophenolate mofetil (MMF) (n = 27) or immuran (n = 5), or other regimens (n = 4). Antilymphocyte globulin was used in 16/66 (24.2%) patients for induction therapy. Patients were followed for a minimum of 1 year or until death. Median follow up was 602.5 days (range 38–1361). Ten patients died during the study period of which two patients died of recurrent HCV.
Recurrent hepatitis C
Biopsy-proven recurrent HCV occurred in 41/66 patients (62.1%). Median time to recurrence was 152 days (mean 205; range 35–796). At the time of recurrence, the hepatitis score was 1 in 22 patients (33.3%), 2 in 16 patients (24.2%) and 3 in three patients (4.5%). Grade 1 fibrosis was present in eight patients, grade 2 in three patients, and grade 3 in three patients. Recurrent HCV progressed to chronic active hepatitis or fibrosis ≥ grade 2 in 29/66 (43.9%) patients, a median of 281 days post-transplant (mean 317; range 55–812). Median HCV viral load at 1 month post-transplant was 5.72 log10 copies/mL (range undetectable-7.64 log10 copies/mL) and at 3 months post-transplant was 6.16 log copies/mL (range undetectable-7.76 log10 copies/mL). Virological recurrence (i.e. a positive plasma HCV RNA) occurred in 63/66 (95.5%) patients by 3 months post-transplant and was universal by 6 months post-transplant. The HCV viral load at 1 month was predictive of biopsy-proven recurrence with a relative risk of a 1.22 per log10 copies/mL increase in viral load (95% CI 1.03–1.44; p = 0.02). However, HCV viral load at 3 months was not predictive of recurrent disease (RR = 1.08; 95% CI 0.86–1.44; p = 0.49). The use of antilymphocyte globulin was not associated with a higher recurrence rate in this cohort (Tables 1 and 2). Recurrence was not associated with the type of immunosuppression, and specifically the use of MMF (Tables 1 and 2). The mean number of total grams of solumedrol was 1.30 in patients with recurrence vs. 1.28 in patients without recurrence (p = NS). Mean age was 55.5 in patients without recurrence vs. 53.0 in patients with recurrence (p = NS). Recipient ethnic background (nonwhite vs. white) had no influence on recurrence (p = NS).
|(n = 41)||(n = 25)||p-value|
|HHV-6 infection||12 (29%)||5 (20%)||NS|
|CMV infection||15 (37%)||11 (44%)||NS|
|CMV disease||7 (17%)||2 (8%)||NS|
|Peak CMV VL1||0 (0–5.1)||0 (0–4.75)||NS|
|Peak HHV-6 VL1||0.85 (0–4.0)||1.18 (0–4.5)||NS|
|HCV VL at 1 month1||5.72 (0–7.64)||5.56 (0–7.24)||0.12|
|HCV VL at 3 months1||6.27 (0–7.76)||5.91 (0–7.68)||NS|
|Antilymphocyte therapy||10 (24.4%)||6 (24.0%)||NS|
|MMF therapy||18 (43.9%)||13 (52.0%)||NS|
|Factor||Severe HCV||No severe HCV|
|(n = 29)||(n = 37)||p-value|
|HHV-6 infection||12 (41%)||5 (14%)||0.01|
|CMV infection||14 (48%)||12 (32%)||NS|
|CMV disease||6 (21%)||3 (8%)||0.14|
|Peak CMV VL1||0 (0–4.82)||0 (0–4.07)||NS|
|Peak HHV-6 VL1||1.20 (0–4.00)||0 (0–4.50)||0.11|
|HCV VL at 1 month1||6.04 (0–7.64)||5.54 (0–7.24)||0.014|
|HCV VL at 3 months1||6.47 (0–7.76)||5.97 (0–7.68)||NS|
|Antilymphocyte therapy||8 (27.6%)||8 (21.6%)||NS|
|MMF use||14 (48.3%)||17 (45.9%)||NS|
Hepatitis C virus and cytomegalovirus
Cytomegalovirus donor (D) and recipient serostatus was as follows: D+/R+, n = 30; D–/R+, n = 27; D+/R–, n = 5; and D–R–, n = 4. Cytomegalovirus prophylaxis consisting of oral ganciclovir was used in only 5/66 patients (D+/R–)(7.6%). Cytomegalovirus infection occurred in 26/66 patients (39.4%). Symptomatic CMV disease occurred in 9/66 patients (13.6%) and was manifest as CMV viral syndrome (n = 5), CMV colitis (n = 3), and CMV hepatitis (n = 1). Peak CMV viral load in patients with CMV infection occurred at a median of 46.5 days post-transplant (range 8–142 days). Median peak CMV viral load was not different in patients with and without recurrence (median = undetectable in both groups; mean 3.87 vs. 3.82 log10 copies/mL; p = NS). The mean number of biopsies in patients with and without CMV disease and infection was the same (2.8 in patients with CMV disease vs. 2.8 in those without, and 3.0 vs. 2.6 in patients with CMV infection vs. those without; p = NS for both comparisons). The mean length of time between transplant and last biopsy was the same in patients with (377 days) and without CMV (423 days) (p = NS). Cytomegalovirus infection, disease and viral load were compared in patients with and without histologic HCV recurrence (Table 1, Figure 1). Cytomegalovirus infection occurred in 44% of patients without recurrence vs. 37% of patients with HCV recurrence (p = NS). Cytomegalovirus disease occurred 8% of patients without recurrence vs. 17% of patients with HCV recurrence (p = NS). The relative risk of recurrent HCV did not change with increasing CMV viral loads (RR = 0.96 for every 1-log10 copies/mL increase in peak CMV viral load; 95% CI 0.82–1.14; p = 0.66). Cytomegalovirus viral load, infection and disease were not associated with increased HCV viral loads at 1 and 3 months (Table 3).
|HCV viral load at 1 month Median (range)||HCV viral load at 3 months Median (range)||p-value|
|HHV-6 infection||6.04 (0–7.64)||5.91 (0–7.67)|
|No HHV-6 infection||5.67 (0–7.43)||6.19 (0–7.76)||NS|
|CMV infection||5.84 (0–7.64)||6.11 (0–7.68)|
|No CMV infection||5.72 (0–7.43)||6.18 (0–7.76)||NS|
|CMV disease||4.32 (0–7.51)||6.11 (0–7.67)|
|No CMV disease||5.72 (0–7.64)||6.18 (0–7.76)||NS|
Hepatitis C virus and human herpesvirus-6
Human herpesvirus-6 PCR was positive in 36/66 (54.5%) patients at some time during their post-transplant course. No patient developed symptoms that were directly attributable solely to infection with HHV-6. Peak viral load occurred a median of 30 days post-transplant. Median peak HHV-6 viral load was 0.85 log10 copies/µg input DNA in patients with recurrence vs. 1.18 log10 copies in patients with no recurrence (p = NS) (Table 1, Figure 1). The relative risk of recurrent HCV was 1.03 per log10 copies/µg input DNA increase in peak HHV-6 viral load (95% CI 0.81–1.33; p = 0.79). Human herpesvirus-6 infection defined as a viral load ≥ 2 log10 copies/µg input DNA occurred in 17/66 (25.8%) patients. Human herpesvirus-6 infection occurred in 29% of patients with HCV recurrence vs. 20% of patients with no recurrence (p = NS) (Table 1). Human herpesvirus-6 viral load and infection were not associated with increasing HCV viral loads at 1 and 3 months (Table 3). The mean number of biopsies in patients with and without HHV-6 was the same (2.8 vs. 2.7 biopsies/patient; p = NS). The mean length of time between transplant and last biopsy was the same in patients with (409 days) and without CMV (382 days) (p = NS).
Severe hepatitis C virus recurrence
Mean fibrosis score at last follow up was 1.67 vs. 0.56 in patients with CMV disease vs. those without (p = 0.016) and 1.03 vs. 0.50 in patients with CMV infection vs. those without (p = 0.063) (Table 4). The hepatitis score at last biopsy was not significantly different. Fibrosis score was 1.18 vs. 0.55 in patients with HHV-6 infection vs. those without (p = 0.031) (Table 4). Evidence of a more severe recurrence of HCV (hepatitis or fibrosis score ≥ 2) occurred in 29/66 (43.9%) patients. Factors were analyzed for their association with severe recurrence (Table 2). There was a nonsignificant trend towards CMV disease and the development of more severe recurrence (21% vs. 8%; p = 0.14). Human herpesvirus-6 infection was associated with the development of severe recurrence (41% vs. 14%; p = 0.01). In a proportional hazards model, the relative risk of severe HCV recurrence was 2.10 in patients with HHV-6 infection (95% CI 1.0–4.41; p = 0.05). No significant association was observed between peak HHV-6 viral load (RR 1.23/log10 increase in HHV-6 copy number; 95% CI 0.93–1.62; p = 0.14) and severe HCV recurrence or peak CMV viral load and severe HCV recurrence (RR 1.06/log10 copies/mL increase in CMV viral load; 95% CI 0.88–1.28; p = 0.55). Hepatitis C virus viral load at 1 month was associated with severe recurrence (RR 1.42/log10 copies/mL increase in HCV viral load; 95% CI 1.07–1.87; p = 0.01) but HCV viral load at 3 months was not. The mean number of total grams of solumedrol was 1.32 in patients with recurrence vs. 1.27 in patients without recurrence (p = NS). Risk factors for the most severe fibrosis (score ≥ 3) were also analyzed separately. Cytomegalovirus disease was associated with severe fibrosis (44% of patients with CMV disease had ≥ grade 3 fibrosis vs. 7% in patients without CMV disease; p = 0.009). It should be noted however, that only eight patients fell into the category of fibrosis score ≥ 3, representing a very small sample size. No significant association with HHV-6 and fibrosis score ≥ 3 was observed.
|Hepatitis C fibrosis score at last biopsy1||p-value|
|HHV-6 infection||1.18 ± 1.28|
|No HHV-6 infection||0.55 ± 1.02||0.031|
|CMV infection||1.03 ± 1.28|
|No CMV infection||0.50 ± 0.96||0.063|
|CMV disease||1.67 ± 1.41|
|No CMV disease||0.56 ± 1.00||0.016|
This is the first study to prospectively assess quantitative CMV and HHV-6 viral loads and their relationship with both histologic HCV recurrence and HCV viral loads post-transplant. We found no correlation between CMV and HCV viral loads and between HHV-6 and HCV viral loads. In addition, when viral loads were analyzed as a continuous variable, peak CMV and HHV-6 viral load were not associated with biopsy-proven HCV recurrence. In the primary analysis, CMV infection, disease and HHV-6 infection were also not associated with HCV recurrence. However, the fibrosis score at last biopsy was significantly higher in patients with CMV disease and with HHV-6 infection. In addition, the development of severe recurrence (fibrosis or hepatitis score ≥ 2) was significantly more common in patients with HHV-6 infection. This difference was not directly explainable by observable differences in immunosuppression therapy because no significant difference in the type or amount of immunosuppressive therapy was seen in the different groups.
There are several possible mechanisms by which herpesviruses may interact with HCV. These viruses may exert an immunomodulatory effect resulting in enhanced immunosuppression. For example CMV has been shown to be a risk factor for the development of invasive fungal and bacterial infections (3, 4). Similarly, in-vitro studies have demonstrated that HHV-6 infection of T cells results in the down-regulation of IL-2 mRNA and protein synthesis, and a significant reduction in mitogen-driven proliferative responses resulting in a cell-mediated immune defect (10). Other factors such as cytokine dysregulation induced by herpesvirus reactivation may have effects on HCV recurrence. Both CMV and HHV-6 infection have been associated with the up-regulation of several cytokines (16–18). For example, HHV-6 infection results in the potent induction of tumor necrosis factor-α, which is thought to play important role in the pathogenesis of HCV (16, 19). Finally, it may be that the progression of HCV and the reactivation of herpesviruses may both just be markers of an enhanced net state of immunosuppression. Although differences in immunosuppression regimens were not observed in our study among patients with and without more severe HCV recurrence, there is currently no gold standard for measuring the degree of immunosuppression in a given patient. Therefore, as immunosuppressive therapy was heterogeneous among the patients, this may impact on HCV viral load and histological severity. Also, therapy for acute rejection may affect both herpesvirus and HCV recurrence and represents another potential bias.
Previous studies of herpesviruses and HCV have focused on CMV. Teixeira et al. (7) analyzed a cohort of 39 HCV liver-transplant recipients who had routine CMV-PCR testing, and found that the occurrence of CMV viremia did not influence HCV recurrence. However, as patients received pre-emptive ganciclovir based on positive PCR tests, it is difficult to interpret potential interactions between HCV and CMV in a setting not modified by antiviral therapy. Rosen et al. (8) found that the incidence of cirrhosis was higher in HCV liver-transplant recipients with CMV viremia. However, this study was limited by the small number of patients with CMV viremia (n = 8). There are currently no published studies assessing interactions between HHV-6 and HCV. However, as previously described, HHV-6 has been associated with CMV disease, graft dysfunction, and other opportunistic infections in liver-transplant recipients (11–13).
Other risk factors for HCV recurrence noted in previous studies include high pretransplant HCV viral loads, older age, nonwhite race, coexistent HCC, and a high Child-Pugh score at the time of transplant (20). Our study was not designed or powered to assess these variables in relation to recurrent HCV. The specific purpose of our study was to assess herpesvirus reactivation and its potential association with HCV.
Strengths of our study include the prospective use of quantitative viral load testing to correlate CMV, HHV-6 and HCV. Also, the results of all testing was kept blinded, and patients did not receive pre-emptive ganciclovir based on these tests. However we did use prophylactic ganciclovir in patients who were D+/R– for CMV. As this comprised only 5/66 (7.5%) patients, it is unlikely to have influenced the results substantially. A limitation of our study is that HCV genotype testing was not performed, although studies analyzing genotype importance in predicting recurrence have been conflicting (21, 22). Also, we had a relatively small sample size and the positive findings may have arisen by chance alone because of the use of multiple statistical comparisons. However, our study is the largest study to date to specifically assess CMV and HHV-6 interactions with HCV molecular diagnostic testing. Finally, we did not perform protocol liver biopsies on patients, which may lead to some biases in interpreting the exact onset of HCV recurrence, as biochemical findings may not correlate with histological recurrences. However, the vast majority of patients did have multiple biopsies throughout their post-transplant course. Also, as the median follow up was only approximately 2 years, it is likely that many more patients would have progressed to significant fibrosis if follow up had been further extended.
We found no overall association between CMV, HHV-6 infection or viral loads and HCV recurrence. However, viral reactivation may be associated with more severe forms of HCV-related fibrosis or hepatitis. Further study into the immunomodulatory effects of different viruses in a larger cohort of patients is required to confirm whether such an interaction truly exists. Further studies could also examine if suppression of herpesvirus reactivation with antiviral therapy influences the course of HCV infection post-transplant.
Funded in part by a grant from the physicians' Services Incorporated Foundation of Ontario, grant no. 99–57. Janet Raboud was supported by the Skate the Dream Foundation.
- 11The clinical impact of human herpesvirus-6 infection following liver transplantation. Transplantation; in press., , , .