Correspondence Dr Udayakumar Navaneethan, MD, Department of Internal Medicine, University of Cincinnati Medical Center, Cincinnati, OH, USA Tel: +513-558-6110 Fax: +513-558-1744 e-mail: firstname.lastname@example.org
Hepatitis E virus (HEV) is a single-stranded RNA virus that causes large-scale epidemics of acute viral hepatitis, particularly in developing countries. In men and non-pregnant women, the disease is usually self-limited and has a case-fatality rate of less than <0.1%. However, in pregnant women, particularly from certain geographical areas in India, HEV infection is more severe, often leading to fulminant hepatic failure and death in a significant proportion of patients. In contrast, reports from Egypt, Europe and the USA have shown that the course and severity of viral hepatitis during pregnancy is not different from that in non-pregnant women. The reasons for this geographical difference are not clear. The high mortality rate in pregnancy has been thought to be secondary to the associated hormonal (oestrogen and progesterone) changes during pregnancy and consequent immunological changes. These immunological changes include downregulation of the p65 component of nuclear factor (NF-κB) with a predominant T-helper type 2 (Th2) bias in the T-cell response along with host susceptibility factors, mediated by human leucocyte antigen expression. Thus far, researchers were unable to explain the high HEV morbidity in pregnancy, why it is different from other hepatitis viruses such as hepatitis A with similar epidemiological features and the reason behind the difference in HEV morbidity in pregnant women in different geographical regions. The recent developments in understanding the immune response to HEV have encouraged us to review the possible mechanisms for these differences. Further research in the immunology of HEV and pregnancy is required to conquer this disease in the near future.
Enterically transmitted hepatitis E virus (HEV) infection is the most frequent cause of acute viral hepatitis (AVH) in developing countries (1). The disease was first recognized in the Indian subcontinent in the 1950s. Initially thought of as hepatitis A infection (2), it took almost 30 years to recognize it as a different virus when the sera from persons during two water-borne epidemics in India were negative for hepatitis A and B (3, 4). As research progressed in HEV infection, the HEV genome was isolated and the enzyme-linked immunosorbent (ELISA) and polymerase chain reaction (PCR) assays to the HEV became available over the last decade (5).
In industrialized countries, hepatitis E is considered as an emerging disease of global importance and it has been reported in a number of developed countries. A recent survey of acute hepatitis E cases in France highlighted that hepatitis E is clearly an emerging disease in France as in other developed countries (6). Also, a recent study from France (7) has shown that post-transplant patients who are immunocompromised progressed to chronic HEV infection.
In addition to this peculiar trait of progressing to chronic hepatitis E in immunocompromised patients, HEV has an interesting course in pregnant women in certain geographical regions of the world. Studies from various developing countries have shown that the incidence of HEV infection in pregnancy is high and a significant proportion of pregnant women can progress to fulminant hepatitis, with a mortality rate varying from 30 to 100% (8–16). The mechanism of severe liver injury in pregnant women with hepatitis E remains a mystery. In this review article, we have initially reviewed the normal hormonal and immunological changes occurring during gestation and have later discussed in depth the interaction of hepatitis E and pregnancy and the potential mechanisms responsible for its high mortality.
We searched MEDLINE from 1980 to the present using the Medical Subject Headings terms Hepatitis E, Immunological changes and pregnancy, Hormones and pregnancy, Hepatitis E and pregnancy, Hepatitis E in pregnancy and mortality. Important clinical guidelines, large case series from centres of excellence, consensus conference and our own research developments form the basis of this review article.
Before going into the interaction of HEV and pregnancy, we have discussed some unique features of hepatitis E infection that may play a role in pathogenesis.
Molecular virology of hepatitis E virus
Hepatitis E virus, a single-stranded RNA virus, was first described in 1983 as spherical, 27–30 nm virus-like particles. Analysis of its RNA helicase and RNA-dependent RNA polymerase regions shows that HEV forms a phylogenetically distinct group that was recently placed into a separate genus: Hepevirus (1).
Characterization of HEV genomes from geographically distinct locations has identified at least four major genotypes that may differ up to 20% at the nucleotide level (17). While these are diverse in all three open reading frame regions, they are serologically indistinguishable and cross-reactive.
Genotype 1 includes isolates from Asia, the Middle East and North Africa. Genotype 2 has been found in Mexico and Nigeria. Genotype 3 was recovered from swine in North America, Europe, Egypt, Asia and New Zealand and from humans in North and South America, Europe, Japan and China. Genotype 4 was found in humans and swine in Asia (18).
Hepatitis E virus genotypes are also important as they correlate with the severity of infection (19, 20). Accumulating evidence suggests that genotypes 3 and 4 are less pathogenic in humans, while genotype 1 has been shown to be more pathogenic. This explains the high severity of infection in India, where genotype 1 is the most common subtype, in comparison with US, where genotype 3 is the most common type seen.
Hepatitis E virus is classically transmitted feco-orally, although person-to-person transmission has also been reported (21). HEV has been occasionally linked to nosocomial spread (22). Vertical transmission from mother to infant is also known to occur (23). It is infrequently transmitted by transfusion of blood or blood products (24, 25).
Hepatitis E can occur either in large epidemics (2, 4, 21) or in the form of sporadic cases. Although hepatitis E infection is endemic in Southeast and Central Asian countries, outbreaks have also been reported from several parts of the Middle East, Africa and Mexico.
The outbreaks of hepatitis E are large and the overall rates range from 1 to 15%, varying from 3–30% in adults to 0.2–10% in children (4, 21, 26, 27). Children have a high rate of subclinical infection. In the US and Western Europe, <1% of patients of AVH have hepatitis E as the aetiology of their infection and it was thought to be associated with their travel to HEV-endemic regions (28, 29). However, a recent paper from France reporting that 90% of acute hepatitis E patients acquired it through the indigenous route by contaminated water supplies and uncooked shellfish may change this current perception of hepatitis E epidemiology (6).
Clinical presentation of hepatitis E in pregnancy
The relationship between hepatitis E and pregnancy is quite interesting. Hepatitis E has both a high incidence and a severe course in pregnant women in some geographical regions of HEV-endemic countries, such as Northern India (12, 13), while in other HEV-endemic countries, such as Egypt, it has been shown to have a benign course with little or no morbidity (30). In a recent large prospective study from Northern India on the maternal and fetal outcomes of hepatitis E infection, close to 60% of viral hepatitis in pregnant women was attributed to hepatitis E infection. Fulminant hepatic failure (FHF) was more common among HEV-infected women (55%), who were at a 2.7 times higher risk than non-HEV-infected women (20%); maternal mortality was also higher secondary to FHF in the HEV-infected group (41%) vs. 7% in the non-HEV group (13).
Sporadic hepatitis E infection is also associated with increased incidence and severity in pregnant women as reported by a study from India. Hepatitis E alone contributed to approximately 50% of cases of AVH. Fulminant liver failure was significantly higher in pregnant women with HEV infection as opposed to other causes of AVH (69.2 vs. 10%, P<0.001). Also, the prevalence and severity of HEV infection in pregnant women did not differ significantly in various stages of gestation (16).
In contrast, in Egypt, where the prevalence of anti-HEV in rural communities is very high, severe HEV-caused AVH in pregnant women has not been reported. In one study of 2428 pregnant women, the anti-HEV prevalence was 84.3%. No patients with AVH were reported (30).
The reasons for the differences in the outcome of HEV in different geographical areas remain unclear but could be the result of early childhood HEV exposures, producing long-lasting immunity and/or modifying subsequent responses to exposure to the virus. Alternatively, the predominant HEV genotype(s) in Egypt could be less virulent than those in Asia (30, 31). The high risk of vertical transmission of HEV infection from mother to infant was investigated in a study of 469 pregnant women and a mother-to-infant transmission of 100% was reported, although there may have been a selection bias. Nonetheless, the high transmission rate indicates the importance of vertical transmission of HEV infection. A small percentage of the babies born to mothers with active disease were either preterm or had anicteric hepatitis. Two of the babies died within 48 h, while the remaining 24 alive infants showed full recovery (32).
Studies in animals to better understand the pathogenesis have also been non-contributory. In an animal study that investigated the changes induced by HEV in pregnant and non-pregnant primates (33), the course of liver injury was similar in both groups. However, this is not surprising, considering that HEV infection in primates leads only to a milder form of liver injury. The severe liver injury because of HEV infection during pregnancy may be related to one of several possible host factors, such as differences in immune and hormonal factors occurring during pregnancy and genetic and environmental factors, with its occurrence in certain developing countries. We have explored all these issues in this article.
Immunological changes during pregnancy
During pregnancy, the maternal immune system is clearly altered to tolerate a genetically different fetus (34, 35). The outer layer of the placenta is made of trophoblasts, which forms the interface between the maternal and the fetal circulations. Trophoblasts do not express major histocompatibility complex (MHC) class proteins and are hence resistant to T-cell-mediated injury, which is a protective phenomenon to sustain the fetus. However, the natural killer (NK) cells do not require MHC proteins and the trophoblasts are protected against the NK cells as they express a unique human leucocyte antigen (HLA) molecule called HLA-G, which binds to NK receptors CD 16 and CD 56 and inactivates them (36). The placenta also expresses an enzyme called indoleamine 2, 3-dioxygenase, which inactivates and depletes tryptophan, an amino acid essential to T-cell function, and hence suppresses cell-mediated immunity at the fetus–placental interface (37).
Cytokines also contribute to immunological tolerance as both the placenta and the trophoblasts secrete cytokines, including tumour growth factor β, interleukin (IL)-4 and IL-10, which inhibit cell-mediated immunity. In an attempt to understand the immunological changes in pregnancy, Orsi et al. (38) measured serum cytokines in mouse models during different stages of pregnancy; levels of tumour necrosis factor α, IL-1β, IL-2, IL-6, IL-10, IL-12 (p40), IL-12 (p70) and IL-17 (P<0.05) were low during the initial phase of pregnancy and increased markedly in late pregnancy and post-partum. Dudley et al. (39), in their murine model, also observed this, which reinforced the hypothesis that cytokine production during pregnancy favours antibody production over cytotoxic T-cell responses.
T-cells are markedly reduced during early pregnancy up to the 20th week of gestation, leading to a reduced level of immunity (40–42). This modulation of cell-mediated immunity occurs to allow fetal allograft retention, but it also alters the immune response mounted against infections (43). The decrease in T-cell activity has been suggested to increase susceptibility to viral infections such as hepatitis, rubella, herpes and human papilloma virus and also to infections like malaria during pregnancy (44), but it also explains why cell-mediated immune diseases like rheumatoid arthritis improve during gestation.
While some studies have argued that there is no alteration in the number of total T-lymphocytes or in CD4+ lymphocytes in pregnancy (45, 46), others have suggested an initial decrease until 20 weeks to sustain the fetus during the implantation phase and then an increase or normalization later during pregnancy or the post-partum period (47–53). The apparent response to pregnancy of CD8 lymphocytes, although less well studied, is either a slight decrease or stability throughout gestation (46, 50–52, 54, 55).
To summarize, the immunological changes during pregnancy promote the maintenance of the antigenic fetus in the maternal environment by suppression of T-cell-mediated immunity. There is a clear shift in the T-helper type 1 (Th1):Th2 cell paradigm during pregnancy, with a definite skew towards Th2 cells. The levels of most cytokines are depressed particularly during the initial 20 weeks of pregnancy, which is an important phase to sustain the fetus. Whether this suppressed immune system translates into an increased risk of infections during pregnancy is still not clear with the available data.
Hormonal factors in pregnancy
Hormonal factors during pregnancy may also play a significant role in altering immune regulation or viral replication (56, 57). Levels of progesterone, oestrogen and human chorionic gonadotropin (HCG) increase with pregnancy. In animal studies, these hormones have been shown to have a clear suppressive effect on the cell-mediated immunity. HCG has been shown to inhibit cell-mediated immunity in guinea-pigs (58), while oestrogen causes shrinkage of thymus and depletes the CD4 and CD8 populations in mice (59, 60) and progesterone produces involution of the thymus and blocks T-cell development and inhibits Th1 cell and promotes Th2 cell development (61). Progesterone has specifically been shown to impair the transition of pro-T-cells to early pre-T-cells in mice models (61). The expression of thymic stromal progesterone receptors is required to produce involution of the thymus and they play a greater role than oestrogen in producing thymic involution (60, 61). Despite these changes, the numbers of peripheral T- and B-cells are unchanged as the half-life of peripheral lymphocytes is higher. Studies have also shown that there is also a decrease in bone marrow B-cell production, mainly pre-B and immature (fractions B–D) bone marrow B-cells, in pregnant mice because of the increase in oestrogen and progesterone levels during pregnancy (62).
In addition, there are evidences indicating that steroid hormones may influence viral replication (63, 64). For example, hormonal enhancement of cytomegalovirus (CMV) replication may be a possible mechanism responsible for the increased incidence of CMV infection observed during human pregnancy (63). There are also reports of increased predisposition to viral infection in certain high-oestrogen states (64).
Mechanisms responsible for the high morbidity of hepatitis E in pregnancy
As discussed previously, pregnancy is associated with high levels of steroid hormones. These steroid hormones may promote viral replication. It also directly inhibits hepatic cells, which may predispose to hepatic dysfunction/failure when exposed to infectious pathogens (65). Steroid hormones are immunosuppressive (66) and mediate lymphocyte apoptosis through NF-κB. NF-κB is a eukaryotic dimeric transcription factor that has multiple cellular effects, including liver development and regeneration and their effects on the immune response (67). Animal experiments in mice studying the p65 component of NF-κB have shown their primary role in liver development and regeneration. Mice lacking the p65 component of NF-κB showed evidence of widespread apoptosis, which led to recent attempts to study this phenomenon in humans. Prusty et al. (68) studied the changes in NF-κB activity using electrophoretic assays of the p50 and p65 components in pregnant and non-pregnant patients with FHF because of hepatitis B, C and E. Their results replicated the results in animal experiments and they found that the activity of the p65 component of NF-κB was diminished in both the peripheral blood mononuclear cells and the post-mortem liver biopsy specimens in pregnant patients with fulminant liver failure. There was a higher than normal level of p50 expression, but there was a near-complete absence or a minimal expression of p65. The absence of p65 from the NF–κB complex produced fulminant liver damage (68). Their results established that the absence of p65 was probably responsible for severe liver damage in pregnant FHF patients. This hypothesis has also been further supported in other viruses, where recent studies have shown decreased expression of p65 causing liver fibrosis and liver damage in patients with HCV-induced chronic liver disease (69). The expression of NF-κB, physiologically downregulated during pregnancy, also plays an important role in sustaining the fetus during pregnancy (70).
Jilani et al. found that HEV-infected pregnant women with FHF had lower CD4 counts and higher CD8 counts. They also observed that the levels of oestrogen, progesterone and β-HCG were significantly higher in the above-mentioned group when compared with HEV-negative patients or control healthy pregnant females (71). Although the levels of hormones were physiologically high in the normal control population, patients with HEV infection seemed to have significantly higher levels than controls, which probably explains the direct interaction of HEV with the immune system.
Pal et al. (72) studied the cellular immune response in both pregnant and non-pregnant women with acute hepatitis E and the control population and found that pregnant women with HEV had generalized immune suppression characterized by a decrease in lymphocyte response to phytohaemagglutinin (PHA) with a predominant Th2 bias as compared with non-pregnant women with hepatitis E and normal healthy controls. This challenged the previously existing hypothesis that normal pregnancy is associated with systemic immune suppression with an increased risk of infections (73–76). This study was important from a number of perspectives. The thought that normal pregnancy is an immunosuppressed state is challenged because normal healthy pregnant women did not demonstrate a decreased response to PHA. Also, non-pregnant patients with HEV did not show any defective PHA response, probably indicating that HEV by itself does not produce the immunological changes and needs pregnancy as a physiological state to produce the above-mentioned changes. The Th2 bias observed in the present study was specific to HEV infection during pregnancy. It may be just that the Th2 bias is very much prominent in HEV infection as compared with normal pregnancy. The mechanism by which Th2 bias may lead to a more severe disease course in pregnant women with hepatitis E needs further investigation. With this Th2 bias, it was suggested that decreased cellular-mediated immunity is considered a major cause of death in Asian pregnant women with fulminant hepatitis caused by HEV infection.
If all the hypotheses of immunological and hormonal factors interacting with the genetic susceptibility in Asian women hold true, we should expect a high mortality in pregnancy from all HEV-endemic regions. But this is not always true. Two studies, from Chennai, Southern India, and Egypt, despite indicating the high prevalence of hepatitis E infection in pregnancy, also had very interesting observations. The mortality rate of hepatitis E infection was very low (3.4%) (77) and absent (30), respectively, as against 30–100% reported in various studies in HEV-endemic regions (8–16). Also, most of the HEV-infected pregnant women had normal-term deliveries. These studies may underline the importance of viral genotypes in the pathogenesis and severity of HEV infection. The results of various studies from endemic regions have been summarized in Table 1. We hypothesized that the difference in the genotype or subtypes of the HEV infection could be the answer (77). Genotype 1 is the most common subtype causing HEV infection in India, while genotype 3 predominates in the US. Genotype 1 has been further classified into four subtypes and most of them have been grouped into genotype 1A. A subgenotype shift (78) may have been responsible for the different geographical morbidity in pregnant women in Southern India and Egypt. If this hypothesis holds true, it opens up the intriguing possibility of the exploration of the genotype in pregnancy.
Table 1. Studies on hepatitis E infection and pregnancy
Prevalence of hepatitis E virus infection (%)
Prevalence of fulminant liver failure (%)
Mortality rate (%)
Studies with low morbidity and mortality in pregnancy.
In addition to the above-mentioned factors, Khuroo et al. (79) suggested that infection of the fetus with HEV may be responsible for the increased severity of the disease in the mother; Variations in the MHC, which mediate antigen presentation, may also explain some of the differences in the mortality in different geographical areas in women infected with HEV (80, 81). Recently, an editorial commented that HEV-infected pregnant women as a group may be more commonly taking herbal medications, which could explain the high mortality in certain geographical regions (82). However, in our centre back home, where herbal medication use is very common, we have observed that the mortality rate is low (77, 83). At the same time, we observed that the use of herbal medications was an independent predictor for poor prognosis in patients with acute liver failure because of other aetiologies (83).
Also, a recent study of post-transplant patients from France, reporting the increased risk of acute hepatitis E progressing to chronicity, has reiterated the importance of an immune response to protect against infection (7). However, as observed, all these patients were post-transplant and patients who progressed to chronic hepatitis had significantly lower levels of CD3 and CD4 cells, highlighting the importance particularly of T-cell-mediated immunity for pathogen clearance. This difference in presentation is really interesting. The progression to fulminant liver failure in pregnant women could be because of immunological injury, while chronic hepatitis could be mediated by failure to inhibit viral replication, given that these patients were immunosuppressed and immunological injury was absent. Further research is required to understand the implications of immunology and hepatitis E infection.
Thus, there is a complex interaction among viral, host, immunological and hormonal factors, producing a paradigm of severe liver damage in pregnancy, probably immunologically, which is summarized in Figure. 1. The probable hypothesis of pathogenesis of hepatitis E depending on the host immune function is summarized in Figure. 2.
The mechanism of liver injury in hepatitis E is not clear and all the hypotheses put forth have not yet been proved conclusively. In this situation of uncertainty, the management of HEV infection-induced liver failure assumes more importance than ever before. All the studies have shown that pregnant women have a differential immune response that triggers fulminant liver failure. Hence, the logical treatment should be to deliver the fetus as soon as possible. Unfortunately, very few such studies have been undertaken in this field. Therapeutic termination of pregnancy, which has been proven to be beneficial in pregnancy-specific disorders like HELLP syndrome and acute fatty liver of pregnancy (84), has not been fully explored in hepatitis E infection. However, in a retrospective study from India, Banait et al. (85) studied 42 patients with HEV-induced liver failure; there was no difference in maternal mortality in pregnant women who delivered and those who did not, questioning the role of therapeutic termination. The literature at present is not supportive of the fact that delivery of the baby may decrease maternal mortality. However, this was a small retrospective study and must not discourage physicians from pursuing that option, considering that HEV infection also produces immunological changes in the fetus. Studies have also explored the role of preventing HEV in high-risk endemic countries, given the high maternal and fetal morbidity and mortality. Arankalle et al. (86) used an Indian-made preparation of immune serum globulin based on locally prevalent genotypes in preventing hepatitis E among pregnant women during an epidemic. Patients who received immune globulin had a lower frequency of HEV infection than in the control population (18.1 vs. 33.9%). However, that study was limited by the small number of patients and did not have sufficient power to be conclusive. At present, although there is no consensus to treat patients with HEV infection in pregnancy, early delivery of the fetus, if possible, to prevent maternal mortality should be attempted. Randomized studies are required in the future to decide upon the best way of treating patients with HEV infection in pregnancy.
The interaction of hepatitis E and pregnancy is fascinating and has provided new insights into the pathophysiology and understanding of the immunology and host susceptibility factors and their interaction to produce the disease processes. The severe liver injury because of HEV infection during pregnancy may be related to several possible factors, such as differences in immune and hormonal factors occurring during pregnancy, and genetic and environmental factors, with its occurrence in certain developing countries. From a molecular biology aspect, a more interesting target in the future for scientists will be the discovery of probable modifier genes that might explain the differences in the presentation of hepatitis E in different populations and targeting them to alter the immune response. Immunological research in the future might provide us with more information if the management strategies are to be improved and to conquer this disease in the near future.
This article is supported by NIH grant R21 AI067868 to Dr Shata.