Hepatitis E Virus
DAVID A. ANDERSON AND ISWAR L. SHRESTHA
Hepatitis E Virus
DAVID A. ANDERSON AND ISWAR L. SHRESTHA
Hepatitis E Virus (HEV) is the causative agent of Hepatitis E, known as enterically transmitted non-A, non-B hepatitis prior to the molecular cloning of HEV by Reyes et al. in 1990 (57). Hepatitis E is an acute and generally self- limiting infection of the liver but is unique among the hepatitis viruses in causing a high mortality rate during pregnancy. Clinical hepatitis E is of greatest importance in developing countries, where it is a major disease burden with an urgent need for practical diagnostic assays and effective vaccines. It is well established that animal strains of HEV are common throughout the world, with at least some potential for zoonotic infection and a low level of locally acquired disease in developed countries, in addition to that seen in travelers returned from regions of endemic infection. Improved diagnostic assays are required to determine the true incidence of HEV infection and disease in developed countries.
HEV was tentatively assigned to the Caliciviridae family for some years on the basis of its particle structure and overall genome organization; however, detailed analysis of the viral genome (41), together with more extensive studies of many human and animal caliciviruses, provided the impetus for the reassignment of HEV to “unclassified” in the most recent taxonomic report (23).
There is insufficient evidence to support the division of HEV isolates into serotypes; however, antigenic variation has important implications for the serological detection of acute and prevalent HEV infection. The type specificity of many epitopes was first recognized by Yarbough and colleagues (81) in comparisons of the Burmese (57, 64) and Mexican (29) strains of HEV. In addition, some viral antigens appear to elicit highly variable responses in different individuals. For example, many experimentally infected animals (and some patients) fail to develop antibodies to the ORF3 protein (44, 45). This variable reactivity contributes to the poor sensitivity and concordance of HEV diagnostic tests based on such antigens (49).
Conversely, all isolates of HEV share some important cross-reactive antigens. Immunization of macaques with recombinant PORF2 proteins of 55 kDa (soluble, 112 to 607 amino acids [aa]) or 62 kDa (virus-like particles [VLP], 112 to 636 aa) expressed using a baculovirus system confers immunity to both homologous and heterologous virus challenge (70, 71, 80). This suggests that major protective epitopes are common among HEV genotypes. One immunodominant conformational epitope has also been identified in the HEV capsid (58).
While this ORF2.1 epitope is highly conserved between HEV strains and represents as much as 60% of the total convalescent-phase antibody repertoire (58), it is not yet known whether it is associated with protection.
Initial studies of the prototype Burmese strain of HEV (57, 64) and a highly divergent Mexican strain (29) demonstrated wide genetic variation, and the recent reports of many equally divergent HEV strains provide a basis for classification of strains into genotypes, with around 75% nucleotide identities between genotypes. Wang et al. (76) have suggested the following classifications: the Burmese (57, 64) and related strains (including most Chinese strains) as genotype 1, the Mexican strain (29) as genotype 2, the swine HEV strain discovered in the United States (51) and closely related strains isolated from patients infected in the United States (60) as genotype 3, and distinct isolates from patients in China (T1 strain) (76) and both patients and swine in Taiwan (27, 28) as genotype 4. The strong relationship between swine and human HEV isolates in the Unites States (51, 60) and Taiwan (27, 28) demonstrates the importance of viral genotypes in helping to understand the epidemiology of HEV infection, especially in countries where infection is not endemic.
Virion Morphology, Structure, and Size
Virions of HEV isolated from the
bile or feces are nonenveloped, icosahedral particles of around 32 nm in
diameter, without any obvious distinguishing morphology (Fig. 1).
Since HEV is refractory to growth in continuous cell cultures and is not present in large amounts in clinical material, there has been very little characterization of authentic viral particles. However, as noted above, the expression of truncated PORF2 (aa 112 to 660) in the baculovirus system leads to the formationof HEV VLPs (46, 80, 81). At around 27 nm, these VLPs are smaller than the intact virus particle, but Xing et al. have used cryoelectron microscopy to give the first indications of HEV structure (78). Their analysis suggests that HEV VLPs are assembled as a T=1 icosahedral particle containing 30 dimeric subunits of 50-kDa PORF2, with the potential to form an intact virion of the correct size with a T=3 arrangement of 90 dimeric subunits (78). PORF2 dimerization appears to be due to noncovalent interactions in the C-terminal part of the protein and may contribute to the assembly of the immunodominant ORF2.1 epitope (3). It is not known whether the PORF2 protein is truncated in viral particles (as it is in VLPs), and further characterization of infections viral particles is required to enhance our understanding of HEV structure.
HEV has a single-stranded RNA genome of positive polarity. “Full-length” HEV sequences deposited in Genbank range in size from 7,138 to 7,277 nucleotides (nt), but some of this variation may be due to incomplete 5’ and 3’ ends in reported sequences. Based on alignments of 5’ and 3’ ends for a selection of sequences, the 7,277-nt genome of the HEV
US-2 strain – including a 26-nt poly (A) tail – is estimated to be complete (3). On the basis of these alignments, the revised genome lengths for all reported complete HEV genomes would be 7,202 to 7,251 nt (3), plus a poly (A) tail which is most probably of variable length.
Recent studies have shown that the 5’ end of the genome has a 7-methylguanosine cap (36). The viral RNA has short, highly conserved 5’ and 3’ untranslated regions of 35 and 68 to 75 nt, respectively, which are likely to play roles in RNA replication and encapsidation. The HEV genome contains three open reading frames (ORFs), organized as 5’-ORF1-ORF3-ORF2-3’ (Fig.2), with ORF3 and ORF2 largely overlapping.
The HEV genome encodes three proteins, PORF1 (replicative polyprotein), PORF3 (unknown function but a possible regulatory protein), and PORF2 (capsid protein).
Translation ofORF1 is expected to yield a polyprotein (PORF1) of approximately 186 kDa containing sequence motifs of methyltransferase, papain-like protease, RNA helicase, and RNA-dependent RNA polymerase (RDRP) activities (64). Expression of PORF1 alone in HepG2 cells or in an in vitro translation system failed to demonstrate any proteolytic processing into mature products (4), suggesting that other cofactors may be required for correct processing.
PORF3 is a very basic protein (pI » 12.5) with a molecular mass of around 13 kDa and is the most variable protein among HEV strains. Zafrullah et al. have shown that PORF3 is associated with the cytoskeleton and is phosphorylated at serine residue 80 by mitogen-activated protein kinase (82), which suggests that PORF3 may have regulatory functions in virus replication or assembly, but its precise role remains unclear.
The capsid protein (PORF2) is translated as a 660-aa protein, and when it is expressed in mammalian cells, a large proportion of the nascent protein is modified by N glycosylation (35); however, this glycosylated form of the protein is highly unstable (69), and it is not clear whether the authentic viral particle contains glycosylated capsid proteins. When PORF2 is expressed in insect cells, it is cleaved at a predominant site between aa 111 and 112 and at various sites within the C terminus of the protein. At least some of these truncated forms of PORF2 have the ability to self-assemble into VLPs of subviral particles (46, 50, 59, 72, 83), and
cryoelectron microscopy has revealed structural details of such particles (78). However, studies of native virus particles from patients or animals are required to confirm the biological relevance of these self-assembled particles, in particular whether the native capsid protein is full-length or truncated.
Our knowledge of HEV replication is poor, due largely to the lack of practicable cell culture systems for the virus. Propagation of HEV has been demonstrated in primary macaque hepatocytes (65, 66), but the level of replication is very low. Two groups have reported the detection of HEV replication in continuous cell cultures (18, 30-32), but there have been no independent confirmations of these reports. A major shortcoming of these systems has been the very low level of viral products, preventing any detailed study or viral replication.
The lack of any related, cultivable viruses further hampers HEV research, although it may be hoped that some animal strains of HEV may prove less refractory to cell culture. However, the general outline of HEV replication can be inferred (Fig. 3). Following interaction with a specific viral receptor and uncoating of the virus, the input viral RNA serves as mRNA (at least for PORF1), after which the PORF1 polyprotein is cleaved by viral (and perhaps cellular) proteases to yield the mature replicative proteins. The RDRP then copies the input viral genome to yield
minus-strand RNA, which in turn serves as a template for the transcription of further plus-strand RNA molecules, including new genomes.
HEV proteins are encoded in three separate ORFs, which suggests that multiple mRNA species would be synthesized by the RDRP, in addition to the full-length genomic and antigenomic strands. Multiple RNA species have indeed been detected in the livers of HEV-
Infected primates (64); however, the origin and functions of these RNAs are unknown. For example, we do not know whether subgenomic m RNAs are transcribed from a full-length minus-length or, conversely, from multiple sugenomic minus-strands.
The site of HEV assembly in the cell, and the role of proteins other than the capsid protein PORF2 in assembly, are also unknown. In particular, it is unclear whether PORF2 is membrane associated and glycosylated, as observed following expression in cell culture, which would suggest that this nonenveloped virus is assembled with the involvement of cell membranes.
A small amount of HEV is found in plasma during infection, consistent with the release of progeny virus through the basolateral domins of hepatocytes, leading to spread through the liver; however, most of the virus appears to be excreted through the biliary system to complete the replication cycle, consistent with release of virus through the apical domain of hepatocytes.
Very recently, Panda et al. have reported the construction of an infectious cDNA clone of HEV (56). This work holds great promise for future molecular characterization of HEV replication and pathogenesis, but further confirmation of this important breakthrough is required.
Human strains of HEV can readily infect many nonhuman primates, with cynomolgus macaques ( Macaca fascicularis) and rhesus macaques ( Macaca mulatta) being widely used for experimental purposes. HEV has also been detected in free-roaming pigs in Nepal, where infection is endemic (12), but the wider significance of swine HEV was first demonstrated with the detection of a unique HEV strain infecting pigs in the United States (51). Notably, most human isolates of HEV detected in countries without endemic infection are more closely related to swine HEV strains, suggesting that transmission from swine to humans occurs, and Meng et al. have clearly demonstrated the transmission of swine HEV to macaques and of human HEV to swine (54). However, in countries without endemic infection, the true rate of human HEV infection, presumably acquired from animal sources, remains unclear.
HEV has proven largely refractory to cell culture. As noted above, propagation of HEV has been demonstrated in macaque hepatocytes (65, 66), but this system is not widely available or practicable. Preliminary reports of HEV replication in continuous cell cultures have yet to be confirmed independently (18, 30-32), and these systems have not yet contributed to our knowledge of HEV.
Inactivation by Physical and
Chemical Agents
There is no specific information on the appropriate conditions for inactivation of HEV in water or food. Since the virus is transmitted via the fecal-oral route, it must with-stand exposure to bile salts during excretion and low PH during ingestion, but it is generally considered to be more labile than hepatitis A virus. In areas of endemic infection boiling is the most reliable treatment for water, since chlorination may be ineffective in the presence of large amounts of organic matter (for example, in water contaminated with feces).
In considering the epidemiology of HEV infection, it must first be noted that there is a wide divergence in the sensitivity, specificity, and concordance of serological assays in use around the world (49) and hence in the resulting estimates of the prevalence of HEV infection. In general, there is broad agreement about the epidemiology of HEV infection in developing countries, where clinical HEV infection is common; however, the situation in developed countries is less clear.
Contamination of water sources with human feces is the most common risk factor for epidemic HEV infection, as for hepatitis A. As such, the major disease burden of HEV is in developing countries, with the Indian subcontinent, Egypt, and parts of China being recognized as areas of strongly endemic infection. HEV is the most common cause of acute hepatitis in these countries. HEV infection appears to be less common in the developing countries of South America, although this conclusion may be biased by the small numbers of studies undertaken there. In many developing countries the incidence and prevalence of HEV infection has not been examined, and it should be assumed that countries with poor sanitation have a high risk for endemic HEV infection and disease.
Clinical HEV infection in developed countries is frequently associated with recent travel to areas of endemic infection. However, swine HEV strains have a worldwide distribution and infection in pigs from commercial farms is almost ubiquitous (10, 12, 27, 51, 52). In view of the demonstrated potential of swine HEV to cross species barriers (54) and the close relationship between swine HEV and strains isolated from some humans (60), HEV infection should be considered a possibility in acute hepatitis patients who do not have a relevant travel history or markers of other hepatitis viruses.
The true rate of HEV infection in developed countries remains controversial. Serological assays based on truncated PORF2 protein expressed using the baculovirus system have yielded seroprevalence rates of over 20% in blood donors from Baltimore (67), whereas rates lower than 2% were found in Australian blood donors when using he ORF2.1 protein expressed in
Escherichia coli (2); however, the two assays appear to be equally sensitive when applied to samples from countries where infection is endemic. Surprisingly, Thomas et al. found no association between risk factors for enterically transmitted viruses and the rate of HEV seroprevalence when using the baculovirus-based assay (67), whereas the ORF2.1-based assay demonstrated a large difference in HEV exposure rates between presumed high-risk and low –risk populations in Malaysia (63). These results suggest that differences in assay specificity (favoring the ORF2.1-based assays) are a major problem in the detection of HEV prevalence, but differences in assay sensitivity (favoring the baculovirus-based assays) cannot be excluded.
In countries where endemic clinical HEV infection is common, the majority of HEV infections still appear to be subclinical, with a ratio of clinical to subclinical infection of between 1:2.6 and 1:7 reported (48). In countries where infection is not endemic, the rate of clinical HEV is clearly much lower, but it is impossible to estimate the subclinical attack rate because of the limited use of specific diagnostic tests in patients with possible acute hepatitis E and the very wide divergence in seroprevalence rates obtained using different assays (see above). Factors which may influence the clinical attack rate in developed countries include the possibility that swine strains of HEV may be attenuated for humans (24, 53, 54) and the tendency of low-dose infections to produce mild or no disease, as observed in experimentally infected animals. Such low-dose exposure might be expected in developed countries, where massive fecal contamination of water supplies is rare.
In countries where infection is endemic, sporadic HEV infection is often the most common form of viral hepatitis on an annual basis, accounting for around 70% of case in Kathmandu, Nepal (1, 14) (see Fig. 4),but above this background major epidemics also occur with a period of around 7 to 10 years. For example, HEV was shown retrospectively to be responsible for 16 of 17 epidemics of enterically transmitted hepatitis in India (5), but HEV is responsible for at least 25% of sporadic hepatitis infections between epidemics (37). In general, epidemics are associated with the wet season (summer in most countries with endemic infection).
Epidemics of HEV have not been reported in developed countries.
Clinical HEV infection in countries where infection is endemic is most common in adolescents and young adults, but it also occurs to a lesser extent from childhood (34) to old age. These appear to be considerable differences in exposure rates among countries where HEV is considered endemic: in Egypt around 60% of children were exposed to HEV by the age of 10, and this rate did not increase further with age (17, 21), whereas in Nepal only 16% of 12-year-old children had evidence of exposure to HEV, and the incidence peaked at 31% later in life (14). It is likely that the risk of clinical disease increases with age, as with hepatitis A, but while clinical HAV infection is uncommon in many developing countries because of high exposure rates in children (for example, 100% HAV and a large proportion of the population remains susceptible.
For residents of developed countries, travel to countries where infection is endemic remains the greatest risk factor for clinical HEV infections, but sporadic cases do occur in the absence of travel. Risk factors for sporadic cases have not been clearly identified in these settings, but they may be different from those for hepatitis A since the prevalence of anti-HEV does not correspond with risk groups for HAV exposure (67). There is some evidence for increased exposure in groups with occupational exposure to swine (38) but HEV is common in many countries where swine are not kept and no risk has been identified from consumption of pork products. In developing countries, exposure to water contaminated with human waste is a clear risk factor and close contact with domestic animals may also be a risk. Importantly, boiling of drinking water appears to be protective (26), but while this will not be practical for all residents in many developing countries, it is advisable for pregnant women because of the high risk of fulminant hepatitis E in this population.
Levels of HEV-specific antibody decline rapidly in the first year after infection, but they probably remain stable for many years thereafter. The presence of antibody to HEV appears to be protective for individuals in epidemics (9), but antibody may not persist at protective levels for life. Although it is not clear whether seroposìtive individuals have sterilizing immunity, clinical HEV infection appears to be limited to individuals who are seronegative at the time of exposure (9)
Peak rates of HEV infection occur in the wet season in some countries where infection is endemic (for example Nepal, China, and India), associated with fecal contamination of water supplies. However, in some regions there appears to be no seasonal variation (16), and at least one study has shown an increased rate of infection associated with a period of unusually low rainfall, presumably by concentrating human wastes in a riverine ecology (15).
Analysis of serological test results from patients with sporadic hepatitis in Kathmandu, Nepal, shows an interesting seasonal relationship for both the rate of infection and the level of anti-HEV immunoglobulin M (IgM) reactivity in patients during a nonepidemic year (Fig.4). HEV was responsible for almost 60% of acute hepatitis cases seen at the Siddhi Polyclinic in 1997, based on reactivity in an IgM assay ( modified from the IgG assay which uses ORF2.1 [2]), but the levels of reactivity for infections giving high reactivity ( for example, human HEV strains spread via contaminated water, with a peak in the wet season) and those giving low reactivity ( for example, zoonotic HEV strains spread via close contact at a constant rate through the year). Further studies are required to understand the transmission patterns of HEV.
Hepatitis E is transmitted via the fecal-oral route. Consumption of contaminated water is by far the most common route of transmission for clinical cases of HEV in countries with endemic infection; however, in the United States the lack of association between HEV seropositively and risk factors for HAV suggests that atypical routes may be involved for the putative zoonotic HEV infections observed in countries where infection is endemic.
Exposure to a common source (such as contaminated water) within a home must be considered, but hepatitis E is not commonly transmitted by person-to-person contact (in contrast to hepatitis A), and household contact is therefore not a significant risk factor.
Nosocomial Infection
Nosocomial Infections with HEV have not been reported.
Excretion of HEV in feces is the only significant risk for person-to-person infection. Although detection of virus in feces by reverse transcription- PCR (RT- PCR) is technically difficult, HEV RNA is detectable in 50% of patients 2 weeks after onset (13) and in one case has been detected at 52 days after onset (55). As noted above, person-to-person spread is much less common than for hepatitis A.
The incubation period for hepatitis E is reduced and the likelihood of clinical disease is increased with increasing doses of virus (74). The incubation period for hepatitis E is most commonly 5 to 6 weeks (range, 3 to 8 weeks), but in experimental infections of macaques with intravenous challenge and high does of virus, disease can be evident as early as 2 weeks after exposure (45, 74).
HEV replication has not been observed in tissues other than the liver; however, it remains possible that a low level of replication may occur in enteric epithelia prior to infection of the liver, as for HAV (6).
Low levels of HEV can be detected in serum during the late incubation period and for 2 to 6 weeks after the onset illness. HEV is excreted in feces; it has not been found other excreations.
The generalized course of infection and serological responses to HEV are shown in Fig. 5. Infection is presumed be initiated via cells lining the alimentary tract, although direct evidence is lacking. Virus then spreads to liver, eventually infecting a large proportion of the hepatocyte population but without causing direct cytolytic damage. After an incubation period of 5 to 6 weeks after exposure to the virus, liver damage results and is thought to be mediated by the cellular immune response the virus, with infections in young children generally swing a benign course.
Histopathological studies of liver specimens obtained from involved in epidemics from New Delhi (1955 to 1956), Ghana (1962 to 1963), Kashmir (1978 to 1979), and the Xingjiang province in China (1988) have demonstrated that infection with HEV can produce morphological changes in the liver, comprising both cholestatic and classical acute hepatitis (8, 42, 77, 84), but these features are not unique or diagnostic in hepatitis E. Typical histopathological changes include lobular disarry with enlargement of portal tracts, Kupffer cell proliferation, focal hepatocyte necrosis and bridging necrosis, ballooning of hepatocytes, acidophilic degeneration of hepatocytes, and mononuclear cell infiltration. Within hepatocytes there is dilatation of the cisternae of the endoplasmic reticulum, with an increase in the number and size of lysosomes within the cytoplasm. Condensation of the matrix within mitochondria, together with dilatation of the matrix within mitochondrial membrane, has also been reported. Cholestatic hepatitis of hepatitis E is characterized by bile stasis in canaliculi and gland-like transformation of hepatocytes, degeneration of hepatocytes, and intralobular and portal tract infiltrates of lymphocytes and polymorphonuclear leukocytes. A prominent feature is the presence of cholestasis and glandular transformation of the liver cell plates, with the cholestatic changes persisting until clinical recovery occurs. These histopathological changes gradually resolve over 3 to 6 months.
Patients with fulminant hepatitis E infection demonstrate necrosis of parenchyma with collapse of liver lobules, swelling of hepatocytes (which have a foamy apperamce), arrangement of hepatocytes into an acinar pattern, proliferation of small bile ductules, phlebitis of portal and central veins, and portal inflammation with lymphocytic and polymorphonuclear leukocyte infiltration (7, 25, 42 ).
Nonspecific (Cytokines and Inflammatory Mediators)
Cytokine responses during HEV infection have not been studied.
Specific (Humoral, Cell-Mediated, and Mucosal Immunity)
The typical patterns of IgG and IgM class antibody responses during HEV infection are shown in Fig. 5. However, it should be noted that antibody responses to individual viral antigens are highly variable, due to both strain-specific differences in some epitopes and differences in responses to single antigens between individual patients (see “ Serotypes and Antigenicity” above). For example, PORF3 varies greatly between strains, but even in macaques challenged with a common-source inoculum, as few as one in three individuals will mount a detectable IgG response to PORF3, and this response is transient (45). In contrast, the 55- to 63-kDa antigens expressed with a baculovirus system and the ORF2.1 protein expressed in E. coli detect specific IgG and IgM responses in the great majority of patients. Typically, both IgG and IgM antibodies are detectable at the onset of disease, with IgM declining to undetectable levels over 2 to 6 months, while an approximately 10-fold decline in IgG levels is seen over this period but the levels then stabilize.
IgA responses to HEV (as a correlate of mucosal immunity) have been detected in around 50% of patients (11), but these antibodies rapidly declined to undetectable levels. The role of IgA in immunity to HEV infection is unknown, but since passive immunization with IgG appears to be sufficient for protection, it is likely that IgA is not essential.
Cellular immune responses during HEV infection have not been studied.
Tsarev et al. have clearly shown that antibody is sufficient for protection by successful passive immunization of macaques (71). Successful active immunization with protein-based vaccines, which are unlikely to promote significant cellular immune responses, further reinforced the essential role of antibody to epitopes within the capsid (PORF2) protein in protection from HEV infection (71).
An understanding of antigenic structure will be important in the clinical evaluation of HEV vaccines. The development and implementation of effective vaccines for hepatitis A and B over the past 20 years have been aided by serological assays in which the level of reactivity (dominated by only a few epitopes) is highly predictive of immunity, by virtue of the correlation between the immunodominant epitopes and protection. Such a correlation may not be true for HEV. Within the 660-aa PORF2, the conformational ORF2.1 epitope appears to be highly immunodominant (58), but its role in protection is unclear. Other, less dominant epitopes may play important roles in immunity. One such protective epitope has been identified to date (62), but current serological assays are probably unable to distinguish between protective and non-protective antibody responses.
Resolution of HEV is accompanied by normalization of biochemical markers (serum alanine aminotransterate [ALT] and bilirubin levels) over a period of 6 weeks in most patients, whereas histological changes may persist for up to 6 months without over disease. Small numbers of patients appear to have a protracted course of disease, with resolution taking many months.
Hepatitis E is clinically indistinguishable from other forms of acute viral hepatitis. The clinical presentation of acute viral hepatitis commonly begins with nonspecific, flu-like prodromal symptoms lasting from 1 to 10 days, consisting of fatigue, malaise, anorexia, nausea, vomiting, and some alteration in taste and smell. A low-grade fever between 38 and 390C is common.
The first distinctive sign of hepatitis is often dark urine and pale-clay-colored stools followed by onset of clinical jaundice, the prodromal symptoms usually subside; however, some patients may not show visible signs of jaundice despite experiencing severe symptoms. Abdominal examination may reveal an enlarge and tender liver associated with pain and discomfort in the right upper quadrant. The spleen is enlarged in 10 to 15% of patients. Occasionally patients present with cholestasis, which is more common in pregnant women.
Clinical recovery is usually within 4 to 6 weeks. During this phase all the constitutional symptoms disappear. However, mild hepatomegaly and a slight increase in ALT levels may persist in some patients. Interestingly, Tong et al. noted a marked difference in the clinical progression of HEV between African citizens and French soldiers who acquired their infections in Africa, with 0 of 27 French soldiers having fulminant hepatitis but with a mean time to recovery of over 8 weeks, whereas 7 of 44 adult male Africans had fulminant hepatitis with 6 deaths, yet there was a reduced time to recovery in the other patients (mean of around 3 weeks) (68). Fulminant hepatitis and death was even more common in pregnant females (4 of 17), consistent with reports from other developing countries.
The reasons for the differing clinical course between the French soldiers and male African citizens is unexplained. Similarly, the basis of the high mortality during pregnancy is not understood; studies of pregnant macaques have not demonstrated any increase in disease severity (73). While hormonal factors may contribute to pathogenesis during pregnancy, other factors may also be important, such as the underlying general health status or chronic infection with hepatitis B or C in patients at the time of HEV infection. No specific pathogenic factors have been identified.
The majority of HEV infections appear to be subclinical, with a ratio of between 1:2.6 and 1:7 in countries where the disease is common (48). In countries where the disease is not endemic, the subclinical attack rate may be proportionally higher, since between 2 and 30% of individuals have anti-HEV IgG (as a marker of prior infection) depending on the assay used but acute clinical HEV is rarely detected. However, it should be noted that the rate of clinical HEV may be underestimated in these countries because of very limited testing and the lack of sensitivity of many tests (see below).
Liver function tests are an important adjunct to diagnosis. The serum aminotransferases ALT and aspsrtate aminotransferases (AST) show a variable increase during the prodromal phase. The ALT level peaks at the onset of symptoms before the serum bilirubin level beings to oncrease. Peak levels of ALT vary from 1,000 to 2,000 U/ liter at the onset. ALT progressively diminishes during the recovery phase. The level of ALT, however, does not correlate with the degree of liver cell damage. Some patients presenting with anicteric acute HEV infection have only raised ALT levels, which is helpful in the early diagnosis of clinically suspected cases, along with the presence of anti- HEV IgM in the serum.
Jaundice is visible in the sclera or skin when the serum total bilirubin level exceeds 2.5 mg/ dl, usually following the peak levels of ALT. peak serum total bilirubin levels range from 5 to 25 mg/dl; both conjugated and unconjugated fractions are increased. In cholestatic HEV infection (which occurs in around 10% of patients), serum bilirubin levels may remain elevated for prolonged periods.
The prothrombin time and international normalized ratio may be increased in acute viral hepatitis, especially in fulminant hepatitis, indicating extensive hepatocellular necrosis and a worse prognosis. Similarly, a reduction in the serum albumin level may occur.
Reduced blood glucose levels leading to hypoglycemia may be observed in patients with prolonged nausea and vomiting and inadequate carbohydrate intake.
Neutropenia, lymphopenia, and atypical lymphocytes are occasionally observed during the acute phase of viral hepatitis.
The risk and severity of clinical HEV disease increases with age at exposure, as with hepatitis A. Fever as the presenting sign of hepatitis E may be more common in the young (68). There is evidence that HEV infection contributes to fetal death and other complications when the mother has symptomatic hepatitis E (25, 75), but the situation with sub clinical infections is unclear.
The course of HEV disease in immunocompromised patients has not been studied. Since the mechanism of liver damage is likely to be immune mediated, there may be no increased disease risk in immunocompromised patients. However, there is no adequate explanation for the high case fatality rate in pregnant women, and it is conceivable that this could be related to their immunological status.
The major complications of HEV infection are fulminant hepatitis, observed in fewer than 1% of patients from the general population but up to 30% during the third trimester of pregnancy, with a high death rate in most studies. Cholestatic hepatitis is seen in around 10 to 25% of patients (40), with a prolonged period of cholestasis being observed in some studies.
Clinical Diagnosis Including Differential Diagnosis
Acute hepatitis E has no distinguishing clinical characteristics which allow a differential diagnosis with respect to other forms of acute viral hepatitis. In many cases, the nonspecific symptoms found in the prodromal phase (fatigue, anorexia, abdominal pain, nausea and vomiting, and fever) may not lead to a suspicion of acute viral hepatitis unless the patient becomes jaundiced or has a relevant history, such as travel to an area of endemic infection. Due to their identical clinical presentation and (likely) modes of transmission, hepatitis E may be suspected in the same circumstances as hepatitis A, but it must be noted that hepatitis A infection is far more common than hepatitis E in developed countries, and the sensitivity and specificity of assays for HAV,
HBV, and HCV are far superior to available commercial assays for the diagnosis of acute HEV infection. As such, it will usually be inappropriate to consider HEV without excluding other forms of viral hepatitis for patients in countries where hepatitis E is not endemic. Conversely, in countries with endemic infection, HEV is often the most common cause of acute hepatitis.
Diagnosis of HEV infection in individual patients remains problematic. Various assays which are in use in research laboratories have high sensitivity and specificity, but routine (commercial) diagnostic assays are not available in many countries and those available are of questionable sensitivity and specificity.
Virus isolation is not appropriate for HEV; as for other hepatitis viruses. HEV is largely or completely refractory to routine isolation in cell culture.
Antigen Detection
Detection of viral particles in feces by immunoelectron microscopy provides a specific diagnostic market but is no longer in use due to the technical demands and very poor sensitivity of the technique. Other antigen detection methods have not been adopted.
Detection of HEV RNA in serum by RT-PCR provides the “gold standard” in specificity for diagnosis of acute hepatitis E but is not suitable for routine use. RT-PCR has been very useful in research situations for the detection of divergent HEV strains whose serological responses have not been detected by some assays (28, 61, 76), especially in countries where infection is not endemic.
As mentioned previously a number of research and commercial immunoassays are available in various countries but with major differences in their sensitivity and specificity (49). The appropriate use and interpretation of current serological assays for HEV infection must take into account these differences, the pattern of serological responses (IgG and IgM) to various antigens, and the widely different prevalence of clinical HEV infection worldwide (1).
Diagnosis of HEV
Infection in Areas of Low Prevalence
In areas with a low incidence of clinical HEV infection, such as the United States and Western Europe, assay specificity will have a very large impact on the predictive value of HEV-specific IgG (manufactured by Genelabs Diagnostics, Singapore, and Abbott Diagnostics, Chicago, ILL.) have considerable value for the diagnosis of acute hepatitis in travelers returned from areas of endemic infection, among whom the incidence may be much higher than the background rate of reactivity of around 2% in the healthy population. However, with the recognition that HEV should be considered in the diagnosis of sporadic acute hepatitis without a travel history (43), the need for mare specific tests become evident. For example, if the incidence of HEV infection among acute hepatitis patients in the United States is 0.2% then only 1 in 10 patients reactive in a test for HEV-specific IgG would be true positives, leading to an unacceptable positive predictive value.
The detection of HEV-specific IgM should therefore become the method of choice for diagnosis of acute HEV infection in areas of low prevalence. One such assay, manufactured by Genelads Diagnostics, is commercially available in some countries, but the false-positive rate in United States blood donors with this assays is reported as 26 of 856 (3%) (Genelabs, 1998 technical bulletin), severely limiting its usefulness in areas where infection is not endemic. Published studies have also demonstrated that the antigens used in this assay may fail to detect around 40% of patients with acute HEV infection (79).
Improved HEV diagnostic assays will need to demonstrate very high specificity, high sensitivity, and the ability to detect infections caused by the diverse range of HEV strains (including potentially zoonotic strains) which appears to be present worldwide. Assays for HEV-specific IgM with improved specificity and sensitivity should become available in the near future through the use of improved recombinant antigens which are in use inresearch laboratories (2, 20, 46, 59, 79, 83).
Diagnosis of HEV Infection in Areas of High Prevalence
Although the titer of HEV-specific IgG tends to decline markedly in the first year after infection, this relationship is not reliable enough to form the basis of differential diagnosis (2, 22). The detection of HEV-specific IgG is therefore of little use for diagnosis of acute infection in developing countries where HEV is endemic and large numbers of patients have antibody from past infections. Detection of HEV-specific IgM must therefore be the method of choice in these areas.
In endemic settings, the assay sensitivity is of primary concern. This includes the robustness of the assays, in order that the results can be obtained and interpreted manually since equipment such as enzyme-linked immunosorbent assay (ELISA) washers and readers are not available. As noted above, even in a research laboratory setting the current Genelabs IgM assay may fail to detect around 40% of patients with acute HEV infection (79) and may therefore be unsuitable for use in developing countries. HEV IgM assays based on recombinant antigens expressed using the baculovirus system (47, 59, 79, 83) or the ORF2.1 antigen expressed in E. coli (fig.4) appear to have improved sensitivity, which may be more appropriate for endemic settings (1).
Sensitive HEV IgM assays should become widely available in the near future.
Since HEV accounts for as much as 70% of the acute sporadic hepatitis in countries with endemic infection, the specificity of assays is less critical in these settings. For example, if the false-positive rate of an assay is 2% only 1 in 35 hepatitis patients in areas of endemic infections might be misdiagnosed as having acute HEV. Higher specificity would of course be advantageous, and at least HEV IgM assay has been reported with a false-positive of only 0.1% (47).
The most significant influence on HEV infection is protection of water supplies from contamination with human fecal waste. Many epidemics in developing countries have occurred due to leakage of sewage pipes into municipal water supply pipes laid in the same or adjacent trenches. A barrier between these two supplies is essential for long-term prevention. Chlorination and filtration systems are generally inadequate if the source water is heavily contaminated.
Travelers to regions of endemic infection must take precautions against the consumption of contaminated water. Only boiled or bottled water should be consumed. Although infection via food appears to be much less common for hepatitis E than for hepatitis A. it is important that travelers maintain vigilance about the risks of contaminated water, ice, and food. This is especially true for individuals who have been vaccinated against hepatitis A and B and may have a false sense of security with respect to all forms of hepatitis. Women should avoid unnecessary travel to areas of endemic infection during pregnancy.
Isolation does not appear to be justified, due to the very low rates of person-to-person transmission; however, infected individuals should not be involved in food preparation or handling until their symptoms have fully resolved.
Passive immunization has the potential to protect against disease (71), but gamma globulin prepared in countries without endemic infection is ineffective, and even pools prepared in countries with endemic infection may offer little or no protective efficacy. Some protective epitopes have now been characterized (62), and it may be possible to identify suitable high- titer plasma in the future.
Prospects for active immunization are very promising. Vaccines based on the PORF2 protein expressed in insect cells are clearly protective in animal studies (70, 71, 80). Clinical trails which are currently planned in countries with endemic infection should establish whether one or more these vaccine candidates is efficacious and safe. If so, it is hoped that these vaccines will be available for use in countries with endemic infection, where the impact of disease is greatest, rather than only as a “boutique” vaccine for the protection of travelers from developed countries.
Outbreaks are most often associated with massive contamination of water supplies, commonly due to weather conditions (flooding or even lack of rain) which cannot be controlled. Chlorination or boiling of water may be the short-term solutions; special efforts should be taken to supply pregnant women with safe water during outbreaks even if this is not possible for the whole population.
Prevention of Perinatal and Congenital Infection
As above, efforts should focus on protection of pregnant women from infection, especially during outbreaks. Vertical transmission of HEV with severe hepatitis in the infant has been demonstrated (39), but it is unclear whether anything can be done to prevent this once the mother is infected. HEV has not been associated with congenital abnormalities.
Symptomatic and Anti- Inflammatory Treatment
No specific symptomatic
or anti-inflammatory treatments can be advised for hepatitis E. Bed rest where
possible and attention to diet (avoiding fatty foods) may minimized symptoms and
speed recovery; alcohol intake should be minimized. Pruritus is a feature of
cholestatic hepatitis A which may justify the cautious use of corticosteroids
(19): this may also be true in hepatitis E, where cholestasis is quite common,
but controlled studies have not been performed.
Role of End- Organ Support in Severe Disease
There in no specific treatment for acute liver failure, and management strategies for severe hepatitis E are not well established, but it appears that intensive supportive medical care, as for acute liver failure due to other causes, may be able to reduce the high mortality of fulminant hepatitis E during pregnancy seen in regions of endemic infection. Two pregnant HEV patients have been described who acquired infection in countries with endemic infection but presented in the United Kingdom (33). One recovered without specific interventions; the other required ventilation and a bolus of mannitol, which may have aided full recovery.
Antiviral are not available for hepatitis E and would probably be of no clinical use since the peak of virus replication precedes clinical presentation.
References