Infection with hepatitis E virus (HEV) usually results in acute, self-limiting hepatitis in immunocompetent individuals.1 Acute hepatitis E may rarely progress to fulminant hepatic failure, which more often occurs in pregnant women2 and patients with alcoholism and chronic liver diseases.3 HEV is a single-stranded, nonenveloped RNA virus4 that is endemic in Southern Asia and Africa. At least 5 different HEV genotypes have been described, with 4 of them being able to infect humans. HEV genotype 3 has frequently been associated with zoonotic infections,5–9 whereas HEV genotypes 1 and 2 appear to primarily infect human beings. HEV is most often transmitted by a fecal-oral route; however, there have been reports on HEV transmission by blood transfusion in Asia10–12 and Europe.13, 14
Recently, cases of chronic evolution of HEV infections have been described in French and Dutch immunocompromised patients.15–18 Kamar et al.15 reported 14 cases of acute hepatitis E in kidney and liver transplant recipients, of whom 8 took a chronic course leading to persistently elevated alanine aminotransferase (ALT) levels and significant histological activity and fibrosis. Similarly, Haagsma and colleagues17 described 2 cases of chronic hepatitis E after liver transplantation. Both patients rapidly developed cirrhosis, and retransplantation was needed after 5 and 7 years, respectively. However, the frequency of HEV infection as a possible cause of graft hepatitis has never been systematically investigated specifically in liver transplant recipients in a low endemic area. Moreover, the diagnosis of HEV infection can be difficult in immunosuppressed patients as HEV viremia may be detected long before seroconversion to anti-HEV. We therefore screened more than 220 patients who underwent liver transplantation at a single major transplant center in Northern Germany for both HEV RNA and anti-HEV. We identified 2 cases of chronic hepatitis E in whom seroconversion occurred up to 8 months after the first detection of HEV genotype 3. Retrospectively, the source of infection could not be determined; however, we did prove that the HEV strain of 1 patient could infect pigs, and this supported the hypothesis of reverse zoonotic transmission. We believe that these findings are of importance not only for the management of liver transplant recipients but also for other immunosuppressed patients with unclear elevations of liver enzymes. This conclusion is further supported by a recent report on a case of reactivation of HEV infection in a patient undergoing stem cell transplantation.19
ALT, alanine aminotransferase; AST, aspartate aminotransferase; HAV, hepatitis A virus; HBsAg, hepatitis B surface antigen; HBV, hepatitis B virus; HCV, hepatitis C virus; HE, hematoxylin and eosin; HEV, hepatitis E virus; IgG, immunoglobulin G; IgM, immunoglobulin M; NA, not applicable; PAS, periodic acid Schiff; PCR, polymerase chain reaction; RT-PCR, real-time polymerase chain reaction; SD, standard deviation.
PATIENTS AND METHODS
Stored sera of 464 individuals, including 108 healthy controls, 129 nontransplanted patients with chronic liver disease, and 226 liver transplant recipients, were studied. All patients underwent transplantation at Hannover Medical School. Liver transplant recipients were studied by cross-sectional testing of consecutive samples between February 2008 and August 2008 without any specified selection criteria. In addition, stored sera of liver transplant patients with elevated liver enzymes and unclear graft hepatitis, collected during 2006 and 2007, were studied. Additional testing of stored samples was performed for all anti-HEV–positive or HEV RNA–positive liver transplant patients with samples available to determine the time point of anti-HEV acquisition and the duration of HEV viremia.
All individuals were tested for the presence of HEV immunoglobulin G (IgG) antibodies. Liver transplant recipients were further categorized into 2 groups: patients without current clinical evidence of graft hepatitis (group A, n = 156) and patients with elevated ALT levels and unclear graft hepatitis (group B, n = 70). For this second group, no other obvious cause of graft hepatitis was evident on admission. ALT, aspartate aminotransferase (AST), gamma glutamyl transferase, or bilirubin levels were required to be at least 2 times the upper limit of normal; this led to liver biopsy for further diagnostic workup. All patients of this group were also tested for various herpes viruses (cytomegalovirus, Epstein-Barr virus, and human herpesvirus 6) by serological and molecular methods. Patients with viral hepatitis were screened for hepatitis B virus (HBV) and hepatitis C virus (HCV) reinfections. Biliary complications were investigated by ultrasound and endoscopic retrograde cholangiopancreatography, if indicated. Subsequently, 33 of the 70 patients were diagnosed with rejection; 7 patients had graft hepatitis C; 8 patients had cytomegalovirus, Epstein-Barr virus, or human herpesvirus 6 infection as the most likely cause of graft hepatitis; and 5 patients had more than 1 of the aforementioned causes of hepatitis. In 27 patients, no definite cause of graft hepatitis could be defined.
Characteristics of the liver transplant recipients are shown in Table 1. All patients received standard immunosuppressive regimens consisting of a calcineurin inhibitor, steroids, and/or mycophenolate.
Table 1. Characteristics of the Liver Transplant Recipients
Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; HBV, hepatitis B virus; HCV, hepatitis C virus; NA, not applicable; SD, standard deviation.
Number of patients
Underlying liver disease
Primary sclerosing cholangitis
Acute liver failure
Primary biliary cirrhosis
Cystic liver disease
Age [years; mean ± SD (range)]
58 ± 13 (25–82)
50 ± 13 (16–79)
Time since transplant [years; mean ± SD (range)]
7.7 ± 3.6 (0.1–20.6)
2.6 ±4.4 years (0.1–25.0)
ALT levels [IU/L; mean ± SD (range)]
24 ± 20 (6–435)
176 ± 253 (4–1910)
AST levels [IU/L; mean ± SD (range)]
31 ± 50 (12–224)
122 ± 131 (13–722)
Bilirubin [μmol/L; mean ± SD (range)]
21 ± 55 (3–516)
69 ± 110 (4–653)
Causes of graft hepatitis
More than 1 cause
All healthy controls were employees of Hannover Medical School. Nontransplanted patients with chronic liver disease, including cases of cryptogenic liver disease, that were admitted between 1998 and 2008 were studied as controls. Of the 129 nontransplanted patients with chronic liver disease, 103 had cryptogenic cirrhosis, pregnancy-associated elevated liver enzymes, or steatosis. Ten patients were infected with HCV (anti-HCV–positive and HCV RNA–positive), 7 were infected with HBV [hepatitis B surface antigen (HBsAg)–positive and HBV DNA–positive], 2 had a coinfection with HBV and HCV (HBsAg-positive and HCV RNA–positive), and 7 suffered from autoimmune hepatitis. The viral load for the HCV-infected patients ranged from 1250 to 2.01 × 106 IU/mL.
Serological Testing for Hepatitis Virus Infections
Antibody to hepatitis A virus [anti-HAV; IgG antibodies and, if suspected, HAV infection followed by anti-HAV immunoglobulin M (IgM)], antibody to HBsAg, antibody to hepatitis B core antigen, HBsAg, and anti-HCV were tested with the Abott Architect System (Abbott Diagnostics, Wiesbaden, Germany).
Anti-HEV IgM and IgG were investigated with a commercial enzyme-linked immunosorbent assay kit (Abbott Laboratories, Abbott Park, IL) according to the manufacturer's instructions.20
For the 2 patients with chronic HEV infection, sera were also tested with the HEV recomBlot assay according to the manufacturer's instructions (Microgen, Martinsried, Germany).
Polymerase Chain Reaction (PCR)
All samples of liver transplant patients (n = 226) and all patients with chronic liver disease (n = 129), regardless of the HEV IgG antibody status, were tested for HEV RNA. As for healthy controls, only samples positive for HEV IgG antibodies were tested. Nested real-time polymerase chain reaction (RT-PCR) was performed with primers derived from the open reading frame 2 region (Table 2) as described by Meng et al.21 Positive samples of the nested RT-PCR were cross-checked by a second independent RT-PCR as described by Mansuy et al.22 Primer and probe sequences are shown in Table 2.
Table 2. Primers and Probe Used for HEV RNA PCR
Abbreviations: HEV, hepatitis E virus; PCR, polymerase chain reaction.
Primers used for nested PCR
Sense primer: 5′-GTT GTC TCA GCC AAT GGC GAG CC-3′
HEV probe: 5′-(6-Fam)-GTYGTCTCRGCCAATGGCGAGCTT-Tamra-3′
For patients positive for anti-HCV antibodies, HCV RNA was determined with the Roche Cobas TaqMan system (Roche Diagnostics, Mannheim, Germany).
Experimental Infection of Pigs with Sera from a Chronically Infected Patient
Five pigs were infected by the intravenous injection of 2 mL of the HEV RNA–positive serum of patient 1. Blood was taken twice a week after infection, and stool was collected daily. All pigs were necropsied to determine the status of the HEV load in different organs such as the liver, muscles, and lymph nodes. These organ samples were histopathologically examined and tested by RT-PCR for the presence of HEV RNA.
To generate the phylogenetic tree, nucleotide sequences were aligned and analyzed with Geneious Basic 4.5.4 software using the neighbor-joining method and pairwise genetic distances. The phylogenetic tree is based on the partial nucleotide sequence of the open reading frame 2 region (159 base pairs) of the HEV genome obtained from 2 chronic patients and swine in comparison with previously identified genotypes available in GenBank. An avian HEV isolate was used as an outgroup. The branch length is a measure of the amount of divergence between 2 nodes in the tree. Bootstrap values are indicated for the major nodes as percentages of the data obtained from 1000 resamplings. The bar (0.05) indicates substitutions per site.
Data for the different patient groups are presented as means and standard deviations. A comparison of continuous and categorical data between groups was performed with the chi-square test. A P value < 0.05 was considered significant.
HEV Infection in Immunocompetent Patients
Only 1 of 108 healthy controls tested positive for HEV IgG antibodies (0.9%). This individual had no biochemical evidence of hepatitis and was HEV RNA–negative in serum, as tested by 2 independent PCR assays. In the group of nontransplanted patients with chronic liver disease, 4 patients (3.1%) were found to be HEV IgG antibody–positive (Table 3). One of these 4 patients was an HCV-infected male with an HCV viral load of 480,000 IU/mL, 1 male patient was coinfected with HCV and HBV (HBsAg-positive and HCV RNA–positive), 1 woman had autoimmune hepatitis, and 1 woman was pregnant. HEV RNA was undetectable in all 4 anti-HEV IgG–positive immunocompetent patients, but we were able to detect HEV RNA in 1 of the patients with chronic liver disease. Only the pregnant woman also tested positive for HEV IgM antibodies. She had a prolonged course of hepatitis of at least 12 weeks, which became evident during the third trimester of pregnancy without any severe complications such as liver failure. In addition, 1 patient was identified who tested negative for anti-HEV but positive for HEV RNA. This 57-year-old individual was admitted to our hospital because of cryptogenic hepatitis with an ALT level of 1297 U/L and an AST level of 651 U/l. Biochemical parameters of liver disease normalized spontaneously within 2 weeks, and no further follow-up was performed.
Table 3. Prevalence of HEV Antibodies and HEV RNA in Control Patients and Liver Transplant Recipients
Number of Patients
Abbreviations: ALT, alanine aminotransferase; HEV, hepatitis E virus; IgG, immunoglobulin G; IgM, immunoglobulin M.
Immunocompetent healthy control individuals were tested for IgM and HEV RNA only if they tested positive for IgG.
Nontransplanted patients with chronic liver disease
Liver transplant recipients with no graft hepatitis (group A)
Liver transplant recipients with elevated ALT levels (group B)
HEV Infection in Liver Transplant Patients
The primary aim of this study was to investigate the frequency of acute and chronic hepatitis E in patients who underwent liver transplantation at Hannover Medical School. Hannover is located in Northern Germany, and Hannover Medical School is the major tertiary referral center for this region, performing between 130 and 150 liver transplants each year.23 Overall, 226 liver transplant recipients (n = 226) were tested for both HEV IgG antibodies and HEV RNA in serum. One hundred fifty-six patients had no evidence of graft hepatitis (group A), and 70 individuals underwent a diagnostic workup for graft hepatitis (group B). Seven of the 156 patients in group A were found to be HEV IgG antibody–positive (4.5%). Two of them underwent transplantation for cystic liver disease, 1 had alcoholic cirrhosis, 1 had bile duct atresia, 1 had hemangiomatosis, 1 had a neuroendocrine carcinoma, and 1 underwent transplantation because of Caroli syndrome. All HEV IgG–positive patients from group A tested negative for HEV RNA and HEV IgM antibodies (Table 3). Five of these 7 patients were already anti-HEV IgG–positive before liver transplantation, whereas 2 of them tested HEV IgG–negative before liver transplantation. In 1 of the HEV IgG–positive group A patients, we could prove by studying stored serum samples that this female patient with cystic liver disease had acquired her HEV infection after liver transplantation. She was anti-HEV–negative 6 months after liver transplantation and tested positive in a sample collected 12 years later. She was HEV RNA–positive 6 months after transplantation but became HEV RNA–negative during further follow-up.
In the second group of liver transplant recipients with elevated liver enzymes (group B), 3 patients were HEV IgG antibody–positive (4.3%). One patient with cryptogenic liver cirrhosis before transplantation was HEV IgG–positive but was negative for both HEV RNA and IgM antibodies. Importantly, this patient tested anti-HEV IgG–positive in a serum sample obtained 8 days before liver transplantation; thus, it is likely that he had already been infected with HEV before transplantation.
Two other patients of group B were found to be anti-HEV IgG–positive and HEV RNA–positive in serum. Both patients were anti-HEV–negative before transplantation and became infected with HEV genotype 3 (Fig. 1). None of the patients had traveled to HEV-endemic areas within the preceding 12 months before the first detection of HEV RNA or had contact with potentially HEV-infected animals such as pigs or cats. The possibility of HEV transmission by the transplanted liver or blood transfusion13, 24 was excluded by the testing of stored serum samples of the deceased liver donors and all available sera of donors of blood (n = 57) for anti-HEV and HEV RNA. Both patients tested negative for HBsAg, HBV DNA, anti-HCV, and HCV RNA, and this excluded the possibility of de novo HCV or HBV infection. The course and outcome of HEV infection in these 2 individuals are described in more detail next.
Patient 1 was a 34-year-old male who underwent transplantation in December 2006 because of glycogenosis. He received a triple immunosuppressive medication regimen consisting of tacrolimus, steroids, and mycophenolate. The patient tested negative 3 times for HEV RNA and HEV IgM and IgG before transplantation. He became HEV RNA–positive 44 days after transplantation but remained anti-HEV–negative for at least 4 months. Seroconversion to anti-HEV IgG was first detected 9 months after transplantation; however, no sera were available between May 2007 and September 2007. He repeatedly tested negative for anti-HEV IgM until August 2008, but it became detectable later during the course of chronic hepatitis E. HEV IgG antibodies were confirmed by a second independent immunoblot assay (recomBlot) as indicated in the Patients and Methods section. The patient remained HEV RNA–positive at all time points studied until the most recent follow-up in February 2009. HEV RNA was also detected repeatedly in stool samples.
The course of ALT levels over time is shown in Fig. 2A. Patient 1 had an episode of acute hepatitis when HEV RNA became first detectable and showed persistently elevated ALT levels during further follow-up. Although the liver histology was normal early after transplantation, interface hepatitis without fibrosis was detected after 8 months (Fig. 2B). The most recent liver histology sample obtained in October 2008 showed mixed lymphocyte infiltration with bridging fibrosis (Fig. 2C).
Patient 2 was a 40-year-old male with primary sclerosing cholangitis who underwent transplantation in January 2007. He received a triple immunosuppressive medication regimen consisting of cyclosporine, steroids, and mycophenolate. The patient tested negative for anti-HEV and HEV RNA 3 months before transplantation. He became HEV RNA–positive within the first 5 months after transplantation and remained HEV RNA–positive until January 2009. Patient 2 became anti-HEV IgG–positive between 5 and 19 months after transplantation, and this was confirmed by anti-HEV IgG with the recomBlot assay. Unfortunately, no serum samples were available to determine the exact time point of seroconversion. Similarly to patient 1, an episode of acute hepatitis was detected early during HEV infection; however, liver enzymes normalized thereafter, despite the persistence of HEV RNA in serum and stool. A liver biopsy sample was taken only during the early phase of acute hepatitis, showing cellular infiltrates and signs of rejection. No further biopsy procedures were performed during follow-up.
Interestingly, neither the wife nor the daughter of patient 2 developed anti-HEV IgG until February 2009, although they lived in close family contact without any special precautions to prevent HEV infection.
Experimental Infection of Pigs
The experimental infection of pigs with the serum of patient 1 was performed with serum taken 20 months after transplantation. Five pigs were infected with serum intravenously (2 mL for each) and necropsied at different time points (days 6-34 post-infection). All 5 animals showed a lymphohistiocytic infiltrate in the liver (Fig. 3), and HEV RNA was detected in the liver in 2 animals (Fig. 1). Moreover, HEV RNA was also detected in the muscle of 2 animals.
We here show that HEV infection has to be considered in the differential diagnosis of graft hepatitis in liver transplant recipients, even in a low endemic country such as Germany. We confirmed recent reports15–18 that HEV infection may indeed take a chronic course in immunosuppressed individuals, and this suggests that adaptive immune responses are of major importance for controlling HEV. Importantly, chronic hepatitis E can be associated with significant graft fibrosis. One of our patients developed bridging fibrosis within a rather short period of just 21 months after transplantation, and this is in line with the observations by Haagsma17 and Kamar et al.15, 16 Thus, hepatitis E has to be taken seriously in immunosuppressed patients, and all efforts should be made to prevent HEV infection.
The initial screening for viral infections is usually based on serological testing for IgG and/or IgM antibodies. However, the diagnostic window between the first detection of viral nucleic acids and the detection of antibodies can last several weeks and may even be longer in immunocompromised individuals who have an impaired ability to produce antigen-specific antibodies. However, even in immunocompetent patients, antibody screening may miss the diagnosis of acute HEV infection, as highlighted by the retrospectively identified case of HEV viremia with highly elevated liver enzymes in an apparently immunocompetent individual with cryptogenic hepatitis.
We have recently shown that antibodies to HBsAg rapidly decline to undetectable levels after the introduction of standard immunosuppressive medications in liver transplant recipients.25 No data are available that show to what extent the same scenario might also hold true for HEV antibodies. The present study suggests that the detection of both anti-HEV IgG and IgM may be very much delayed in individuals being treated with a triple immunosuppression regimen of calcineurin inhibitors, steroids, and mycophenolate. Gerolami et al.18 found that a kidney transplant recipient became positive for anti-HEV IgM together with HEV RNA, and this was not the case in our patients. In fact, patient 1 was anti-HEV IgM–negative for more than 1.5 years before IgM antibodies became detectable. Thus, the role of anti-HEV IgM testing requires further investigation. In general, immunological assays to screen for HEV antibodies are not that well standardized as is the case for other hepatitis viruses. We therefore confirmed in this study the positive anti-HEV IgG results obtained in immunosuppressed patients by a second independent blot assay (recomBlot, Microgen). Christensen and colleagues9 found an extremely high prevalence of HEV antibodies of 20% to 30% in blood donors from an apparently low endemic country (Denmark) with a different noncommercial assay developed at the National Institutes of Health. It might be interesting to apply this assay to immunosuppressed individuals and to determine if this rather long diagnostic window until seroconversion could be shortened. Similarly, future studies also need to address if there is any correlation between anti-HEV titers and disease activity. We found much higher HEV IgG titers in patient 1 with active persisting hepatitis than in patient 2 who had normal liver enzymes after the initial phase of transient hepatitis (data not shown). In any case, it is obvious that the diagnosis of both acute and chronic HEV infection must be based on the detection of HEV RNA.
HEV genotype 3 infection has been suggested to be a zoonosis by several investigators. Potentially exposed subjects such as farmers display a much higher prevalence of anti-HEV, which can exceed 50%.8, 9, 26 In line with these earlier data, a recent German case control study confirmed that autochthonous HEV infection was mainly due to genotype 3 infection, whereas travel-associated hepatitis E was caused by HEV genotype 1.27 Importantly, this study also confirmed that consumption of offal and wild boar meat was independently associated with an increased risk of HEV infection. Thus, these data and the recent evidence of chronic evolution of HEV in immunosuppressed patients strongly suggest that immunocompromised individuals should avoid any consumption of raw meat from animals that are potential carriers of HEV. Indeed, HEV RNA has been detected in raw meat in different countries, including Germany,10, 27–31 and this was confirmed by our experimental infection of pigs with serum from patient 1.
We tried to exclude the possibility of nosocomial HEV infection, especially as both cases of HEV infection occurred within a few months in our hospital. Both organ donors had no evidence for HEV infection (they were tested for HEV RNA and anti-HEV; the data are not shown); similarly, none of the hospital staff in contact with the 2 patients had evidence for acute or chronic hepatitis E. Moreover, all blood donors were anti-HEV–negative. However, it still may be possible that one of the patients had infected the other patient.
Our findings have significant implications for the management of graft hepatitis in liver transplant recipients. First, we suggest that all patients with elevated liver enzymes should be tested for HEV RNA unless other obvious reasons explain the hepatitis. Whether universal screening for HEV should be implemented obviously depends on the overall frequency of HEV infections in the respective regions. We detected anti-HEV in less than 5% of our Northern German patients, whereas Kamar and colleagues15 found antibodies against HEV in more than 10% of patients undergoing liver transplantation in Southern France. Second, therapeutic consequences have to be considered for patients with chronic hepatitis E. Once chronic infection is established, it is likely that in the case of retransplantation, the new organ will also be infected, as shown by Haagsma and colleagues.17 The question to be addressed in the future is to what extent antiviral therapies might be effective against HEV. We are not aware of any data supporting a role for ribavirin to treat HEV infection. In contrast, type I interferons are well established in the treatment of hepatitis B32 and C.33 It is tempting to speculate that patients with chronic hepatitis E may also benefit from interferon alpha treatment. However, the potential risks and benefits of interferon alpha have to be weighed carefully as the drug may induce graft rejection in addition to other frequent and sometimes severe side effects. Third, chronic HEV carriers are potentially infectious, and this has to be considered both during inpatient and outpatient management.
Unfortunately there is no vaccine yet available to protect relatives or medical health professionals from HEV infection. Recently, successful phase II vaccine trials have been presented.34, 35 However, to our knowledge, there is no vaccine program close to approval, and thus it is unlikely that there will be a vaccine available within the next few years. However, once a vaccine is available, we recommend that patients on a transplant waiting list be vaccinated against HEV. Until then, strict hygienic precautions remain the only option to prevent HEV infection. Importantly, both family members of patient 2 remained anti-HEV–negative after 2 years of infection, although they were not aware of the risk of infection for about 18 months. We also screened more than 50 nurses and doctors who potentially had been in contact with patient 1 and patient 2, and none of them had evidence of recent HEV infection (data not shown).
The limitations of our study need to be considered. First and most importantly, we did not prove a causal relationship between HEV infection and liver disease. Graft hepatitis after liver transplantation frequently has a multicausative nature, and several factors may contribute to fibrosis progression. A possible relationship between chronic HEV infection and cirrhosis development is only suggested by cumulative evidence from different cases presented by various groups in Europe.16–18 Second, we did not identify the true source of infection in our cases and only showed a reversed zoonotic transmission. Third, the usefulness of primary HEV RNA testing in immunocompromised individuals needs to be confirmed in a prospective study. Moreover, more studies are needed to determine the incidence and prevalence of HEV infection in other cohorts of immunocompromised patients to investigate if the phenomenon of chronic hepatitis is more often found in certain types of organ transplantation or if all immunosuppressed patients display the same risk for chronic evolution of HEV infection. In this context, it is interesting to note that Spanish investigators have recently reported a lack of chronic hepatitis E in human immunodeficiency virus–infected individuals.36
In conclusion, the prevalence of HEV infection in Central European liver transplant recipients is low; however, chronic hepatitis E may occur and needs to be considered in the differential diagnosis of graft hepatitis. The diagnosis of HEV infections should be based on HEV RNA determination in immunosuppressed patients; however, this needs to be confirmed in a prospective study. We suggest that immunocompromised individuals should avoid eating uncooked meat to avoid zoonotic infection with HEV genotype 3.