Hepatitis B and pregnancy: an underestimated issue
Maureen M. Jonas, Children's Hospital Boston, Center for Childhood Liver Disease, Division of Gastroenterology, 300 Longwood Avenue, Boston, MA 02115, USA
Tel: +617 355 5837
Fax: +617 730 0716
Hepatitis B infection during pregnancy presents a unique set of management issues. Aspects of care that must be considered include maternal and fetal effects of hepatitis B, effects of pregnancy itself on the course of hepatitis B infection and its complications, treatment of hepatitis B during pregnancy and prevention of perinatal infection. There are insufficient studies to date regarding these concerns; most are from the Far East, and many have important limitations, but some have yielded valuable data. Pregnant women with acute hepatitis B virus (HBV) infection typically have a course not very different from that in the general adult population, but the risk of transmission of HBV to neonates increases the later in gestation the acute infection occurs. Chronic HBV infection is usually mild in pregnant women, but may flare shortly after delivery. The risk of perinatal transmission is highest in women with high levels of viraemia; this may be a factor in the small but reproducible failure rate of current immunoprophylaxis strategies. Obstetrical policies must be assessed with respect to detection of maternal infection and liver disease, as well as with respect to perinatal transmission risk. In addition to the usual issues of drug efficacy and safety in the affected individuals, effects on the developing fetus must be considered. This paper reviews the current experience in each of these areas, and highlights the need for further investigation into this critical but often underestimated topic.
Of the estimated 350 million individuals chronically infected with hepatitis B virus (HBV) worldwide, it is generally accepted that at least 50% acquired their infections either perinatally or in early childhood, especially in countries where HBV is endemic (1). This is attributed to the high rates of HBeAg-positive infections in women of child-bearing age in these parts of the world, and the efficient transmission of infection from these women to their newborns. It has long been recognized that prevention of perinatal transmission is a high priority in the attempt to decrease the global burden of chronic HBV. Immunoprophylaxis with hepatitis B immune globulin (HBIG) and hepatitis B vaccine is known to be safe and effective, but applied variably in different geographical regions. Even with proper vaccination, 5–10% of infants of HBeAg-positive women become infected, and so there is opportunity for improvement in prevention strategies. In addition, the interaction of HBV infection and pregnancy itself is an area for further study.
The prevalence of chronic HBV infection in pregnant women in urban areas of the USA varies by race and ethnicity (2). As expected, the highest rate (6%) is in Asian women. The rates in black, white and Hispanic women are 1, 0.6 and 0.14% respectively. Similar data are not available for European countries, but are expected to mirror those in the general populations of each nation, especially as determined by the numbers and countries of origin of immigrants (3, 4). In areas of high endemicity, such as China, other Far East countries and Africa, rates are proportionately higher. In the USA, testing for HBsAg is recommended for every pregnant woman, regardless of previous testing or vaccination (5). Women who have not had this testing and those with risk factors for HBV acquisition should be tested at presentation for delivery. In Europe, there is no consistent policy with respect to testing of women for HBV infection during pregnancy, and many countries rely on historical ‘risk factors’ to determine indications for screening. However, a recent report from Denmark indicates that approximately 50% of infected pregnant women would not have been identified using this strategy (6), a proportion that closely resembles that found in a comparable study in an urban US hospital a number of years ago (7). In addition, changing immigration patterns in Europe indicate that HBV prevalence will vary greatly at a regional level within each country, and it has been suggested that there is a need to provide a more general immunization programme to protect the population at large (8). The benefits of detection of infected pregnant women include not only identification of infants who require prophylaxis, but of women who might need treatment, and sexual and household contacts who will benefit from testing, counselling, vaccination or therapy if indicated.
Effect of hepatitis B on pregnancy
Susceptible women who develop acute hepatitis B during pregnancy may have an illness indistinguishable from that in the general population. Acute HBV infection must be differentiated from other acute liver diseases that occur during pregnancy such as intrahepatic cholestasis or acute fatty liver of pregnancy if jaundice is present, or haemolysis, elevated liver enzymes and low platelets syndrome if jaundice is absent. It does not appear that acute HBV infection increases mortality during pregnancy, or that it has teratogenic effects. However, a higher incidence of low birth weight and prematurity has been reported. In addition, acute HBV early in pregnancy is associated with a 10% perinatal transmission rate, and the rate increases substantially with HBV infection in the third trimester.
The effects of chronic HBV infection on pregnancy outcomes have not been clearly defined. One large study demonstrated no differences in gestational age at delivery, birth weight, incidence of prematurity, neonatal jaundice, congenital anomalies or perinatal mortality comparing HBsAg-positive women with controls (9). However, a relatively recent study described an association of maternal HBV infection (HBsAg positive) with gestational diabetes mellitus and antepartum haemorrhage (10). There was a suggested association with preterm delivery.
Consideration of active HBV infection during pregnancy raises the question of whether amniocentesis is contraindicated in this setting. In one series of 21 mother–infant pairs, in which the mothers were HBsAg positive (but only one HBeAg positive) and underwent amniocentesis for accepted indications at a mean of 19.5 weeks gestation, none of the infants were HBsAg positive at 1 or 12 months of age (11); they had received the HBIG and HBV vaccine as recommended. In another study, a prospective longitudinal analysis of outcomes in 47 HBsAg-positive women who presented for amniocentesis (12), all the amniotic fluid samples and cord blood samples from the infants were analysed for HBsAg and HBV DNA. In this cohort, 32% of the amniotic fluid samples were HBsAg positive, but HBV DNA was undetectable in all. Although cord blood from 27% of the infants contained HBsAg, none contained HBV DNA. As a control, cord blood was sampled from 72 infants delivered from HBsAg-positive women who did not undergo amniocentesis. Of these, 18% contained HBsAg and 4% contained HBV DNA. The authors of both studies concluded that the risk of HBV transmission by amniocentesis is low.
Effect of pregnancy on hepatitis B
In general, women with chronic hepatitis B do well during pregnancy. However, pregnancy is associated with high levels of adrenal corticosteroids, which might be expected to increase levels of viraemia, and oestrogen, which has been demonstrated in laboratory animals to decrease HBV replication. In one study, no significant differences in HBV viraemia were noted during pregnancy, although alanine aminotransferase (ALT) levels tended to increase late in pregnancy and in the post-partum period (13). It has been known for some time that a proportion of women have hepatitis flares with or without HBeAg seroconversion within the first months after delivery (14). Seroconversion rates of 12.5% (15) to 17% (14) have been described. It has been postulated that the rapid decrease in cortisol levels characteristic of the post-partum state is analogous to the steroid withdrawal therapy that has been used to elicit seroconversion (15, 16). Although usually this is well tolerated, cases of exacerbation of hepatitis (16) and even fulminant hepatic failure (17) have been described in the peripartum period. Exacerbation of hepatitis was not prevented by administration of lamivudine in the third trimester (16). Factors that do not appear to be associated with likelihood of postpartum HBeAg clearance include maternal age, parity and presence of precore or basal core promoter mutations (15). In one report, a low maternal HBeAg level was strongly associated with postpartum HBeAg clearance (15). It appears prudent to monitor HBV-infected women closely for several months after delivery for hepatitis flares and seroconversion.
Hepatitis B virus and human immunodeficiency virus co-infection during pregnancy
There are few reports of HBV co-infection with the human immunodeficiency virus (HIV) in pregnant women. In sub-Saharan Africa, where HBV is endemic, 13% of HIV-infected pregnant women also have HBV. In the only American series, 1.5% of 455 HIV-infected obstetrical patients followed in Texas over 11 years were HBV co-infected (18). Of note, these women had lower CD4 counts when compared with women with both HIV and HCV or those with HIV alone. In addition, women with HBV were compared with those who had serological evidence of previous infection and natural immunity. The women with chronic HBV had lower median CD4 counts than those who had cleared previous HBV infection. The authors raised the question of whether HBV co-infection conferred additional immune suppression in this group.
Hepatocellular carcinoma and pregnancy
There are rare recorded cases of hepatocellular carcinoma (HCC) in pregnancy. In several reports, fetal outcome was often satisfactory although some intra-uterine deaths were recorded. Maternal mortality was high, suggesting an adverse effect of pregnancy on the outcome of this malignancy. Twenty of the 33 reported women in a combined series died within a few days of the initial presentation (19, 20) and most others succumbed within months. It has been suggested that oestrogen may accelerate the evolution of HCC as it does for other liver tumours. In addition, gestational immune suppression may be an enabling factor in tumour progression (21).
Vaccination against hepatitis B virus during pregnancy
Vaccination against HBV is both safe and efficacious during pregnancy (22, 23). In addition, passive transfer of maternal antibody to newborns has been demonstrated, although without the addition of active vaccination, titres in the infants were noted to wane over time (22), as would be expected.
Treatment for hepatitis B virus during pregnancy
There are two principal indications for administration of antiviral agents to HBV-infected pregnant women: treatment of chronic hepatitis in the mothers and prevention of perinatal HBV transmission to the newborns.
Most women with chronic HBV infection have mild liver disease during pregnancy, although hepatitis may flare after delivery, as described above. In addition, interferon, lamivudine, adefovir and entecavir are classified by the Food and Drug Administration as Class C, and telbivudine and tenofovir as Class B. In most cases, this is because there are insufficient data in humans to demonstrate teratogenic or embryotoxic effects. For these reasons, in most instances, it is reasonable to defer therapy until after delivery, to avoid fetal exposure to the therapeutic agents. After delivery, standard therapy indications, as expressed in the several available HBV guidelines, will apply. However, if maternal liver disease requires treatment, or if a pregnancy occurs in a woman already receiving a medication for HBV, decisions must be made about treatment course.
There is a long history of use of lamivudine during pregnancy, both for women with HIV infection and for those with chronic HBV. Data from the Antiretroviral Pregnancy Registry 2006 (24) indicate that the rate of birth defects among women exposed to lamivudine was similar to that in the general population. In one cohort of 38 HBV-infected women who became pregnant while taking lamivudine and elected to continue the treatment throughout the pregnancy, there were no pregnancy complications, no instances of fetal injury and no cases of perinatal HBV transmission (25). This compared favourably to historical rates of HBV transmission from the same population in which active and passive immunization was used routinely. In addition, 35 of the 38 women were no longer viraemic with HBV, and 10 (26.3%) had HBeAg seroconversion. Two women who elected to discontinue lamivudine during their pregnancies developed active hepatitis (abnormal ALT) within 6 months (25). This study was small, and the authors concede that more data are needed, but there is some support for the safety of lamivudine in this group. There are no comparable studies of other antiviral agents for HBV.
At this point, there are no standards regarding managing HBV in women who become pregnant while receiving antiviral therapy. One option is discontinuation of treatment as soon as the pregnancy is recognized. This is an option only for those with mild hepatitis, with a low risk of serious flare or disease progression. Other possibilities include continued careful monitoring or change of therapy to lamivudine, either temporarily or permanently, acknowledging the risk of development of resistance.
Perinatal hepatitis B virus transmission
Perinatal transmission of HBV results in a high frequency of chronic infection, up to 90% in infants born to HBeAg-positive women. It is widely accepted that most perinatal transmission occurs at or near the time of birth, because neonatal vaccination prevents newborn infection in about 80–95% of cases. Theoretical risks for HBV transmission at delivery include exposure to cervical secretions and maternal blood. Transplacental (intra-uterine) transmission is presumed to cause the minority of infections not prevented by prompt immunization. Risk factors for transplacental transmission of HBV include maternal HBeAg positivity, HBsAg titre and HBV DNA level (26). In one study, a maternal HBV DNA level of ≥108 copies/ml was associated with increased likelihood of intra-uterine transmission (27). HBV is found in the villous capillary endothelial cells (26) and the trophoblasts (28) of the placenta, supporting the hypothesis that breach of the placental barrier is a mechanism for intra-uterine infection. Threatened preterm labour or spontaneous abortion, with the possible mixing of maternal and fetal blood, appears to increase the risk of HBV transmission (29). Recently, polymorphisms in some cytokine genes, such as those encoding for interferon-γ and tumour necrosis factor-α, have been correlated with risk of intra-uterine infection with HBV (30, 31). Prevention of perinatal transmission is considered critical in the attempt to decrease individual and population morbidity from chronic hepatitis B infection as well as the global burden of hepatitis B.
Mode of delivery has been examined as a potential risk factor for HBV transmission. In a report from China in 1988, of 447 infants born to HBsAg-positive women, 24.9% (96/385) of newborns delivered vaginally were HBV infected at birth, compared with <10% (6/62) delivered by caesarean section (32). Both groups received HBV vaccine. These authors advised caesarean section delivery for mothers with high levels of viraemia. However, a later study compared outcomes among three groups: 144 infants born by spontaneous vaginal delivery, 40 by forceps or vacuum extraction and 117 by caesarean section (33). All infants received the HBIG and HBV vaccine at the recommended schedule. Chronic HBV infection was detected in the infants in 7.3, 7.7 and 6.8%, respectively, and response rates to immunization were similar in all groups. The authors concluded that mode of delivery does not influence the likelihood of HBV transmission. At this point, most obstetrical algorithms do not include change in the planned mode of delivery for HBsAg-positive women regardless of HBeAg status or level of viraemia.
In the USA, all pregnant women are supposed to be tested for HBsAg, regardless of assessed risk and previous testing. Neonates born to HBsAg-positive women should receive HBIG and vaccine before discharge (5) and be followed to determine the adequacy of immune response and the vaccine failures. All infants, regardless of maternal HBsAg status, should receive HBV vaccine in the first months of life. In Taiwan, all infants have been receiving the HBV vaccine for almost 20 years, with a significant impact on perinatal transmission and childhood and adolescent infection and its complications (34, 35). Whether universal immunization will be adopted in all European countries, as the World Health Organization has recommended, depends on many factors, such as perceived prevalence and risk, changing immigration patterns, cost–benefit analyses and budgetary priorities.
Immunoprophylaxis provided to newborns clearly reduces the incidence of perinatal HBV transmission. In a recent meta-analysis of clinical trials (36), the relative risk of neonatal HBV infection in those who received HBV vaccine (plasma-derived or recombinant) was 0.28 [95% confidence interval (CI) 0.2–0.4] compared with those who received placebo or no intervention. Compared with vaccine alone, the addition of HBIG to the regimen further reduced the relative risk (0.54, 95% CI 0.41–0.73) when compared with active prophylaxis only. Nonetheless, there are clearly a substantial number of newborn infections, even with prompt administration of active and passive vaccination. The estimates vary, and depend on maternal HBeAg status, but most studies demonstrate anywhere from 1% (37) to 10% (36) chronic HBV infection in infants who were appropriately immunized. Clearly, with millions of at-risk pregnancies each year throughout the world, significant numbers of perinatally acquired chronic HBV infection are still occurring.
The major target for neutralizing anti-HBs is the a determinant of the surface antigen protein. Mutations in the S gene of HBV causing conformational changes in the a determinant have been detected in humans infected with HBV, and concern has been expressed that these variants might replicate in the presence of vaccine-induced anti-HBs or anti-HBs contained in HBIG (38, 39). At this point, no evidence suggests that S gene immunization escape mutants pose a threat to programmes using hepatitis B vaccines (40), but perhaps enhanced surveillance to detect the emergence of these variants will be necessary for monitoring the effectiveness of current vaccination strategies.
One approach to prevention of perinatal HBV transmission is provision of HBIG during pregnancy. Several reports have documented the results of this intervention, demonstrating varying efficacy (27, 41–43). Unfortunately, the studies are quite heterogeneous, using different doses and routes of HBIG administration, and utilizing different outcomes to determine neonatal infection, such as HBV DNA in cord blood, or HBsAg in the infants at 6 months of age. In some studies, only HBeAg-positive mothers were included, and in others HBeAg status was not specified. Three of the four reports, however, documented a beneficial effect (27, 41, 43), while in one no obvious difference was noted (42). One of the studies included examination of the effect of maternal HBIG receipt on the results of newborn vaccination, and found that maternal treatment with HBIG was associated with higher seroprotection (development of antibody to HBsAg, anti-HBs) rates in their offspring than observed in those whose mothers were not treated (41). Although a decrease in the perinatal HBV transmission rate was documented only in HBeAg-positive women, the beneficial effect on development of anti-HBs was seen in infants of both HBeAg-positive and HBeAg-negative women (41).
Because the risk of intra-uterine and perinatal transmission of HBV is clearly related to the level of maternal viraemia (26, 44), another strategy to interrupt this process is maternal treatment with a nucleoside analogue late in pregnancy. As stated above, the only HBV drug with a record of safe use in pregnant women is lamivudine. van Zonneveld et al. (45) treated eight pregnant women with high HBV DNA levels with 150 mg lamivudine daily from gestational week 34 until delivery. The infants received both active and passive immunization at birth. The HBV DNA levels declined at least 1 log in five of the eight women, after 6–40 days. Although four of the infants were HBsAg positive at birth, all but one was negative by 12 months of age (12.5% transmission). The rate of chronic HBV infection in a comparable group of 24 historical controls was 28%. The largest study was a randomized, double-blind, placebo-controlled trial in 114 highly viraemic women, 68 of whom received lamivudine 100 mg daily beginning at week 32 (46). Again, all infants received HBIG and vaccine in the standard regimen. Viral load reduction to <1000 mEq/ml was achieved in 98% of the lamivudine-treated mothers and 31% of controls. At 1 year of age, 18% of infants of lamivudine-treated mothers were HBsAg positive, compared with 39% of those whose mothers received placebo. In addition, there was a greater incidence of anti-HBs positivity in infants whose mothers had been treated with lamivudine, 84 vs. 61% in controls. No adverse effects of lamivudine were noted in either the mothers or their infants.
At this point, there is no consensus regarding using HBIG or a nucleoside analogue in pregnant women to prevent perinatal transmission. One proposed algorithm includes consideration of both the level of maternal viraemia and the history of a previous child becoming infected with HBV perinatally (47).
Breastfeeding by hepatitis B virus-infected women
Decades ago, studies from the Far East demonstrated that HBsAg could be detected in breast milk in a large proportion of HBV-infected women (48, 49). Because many infants became infected before the availability of immunization, there was concern about the additional risk that breastfeeding might confer. However, around the same time, Beasley et al. (50) reported 53% HBV infection in breastfed vs 60% in formula-fed infants. More recently, several studies have documented no difference in rates of perinatal infection between breastfed and formula-fed vaccinated infants, which was between 0 and 5% in both groups, although many of the women in these studies were HBeAg negative (51, 52). These data support the recommendation of the American Academy of Pediatrics that HBV infection not be considered a contraindication to breastfeeding of infants who receive the HBIG and HBV vaccine as advised (53). In addition, it appears that breastfeeding does not interfere with the immune response to the HBV vaccine. In one group of 230 infants, the rate of anti-HBs at 1 year of age was 80.9% in breastfed compared with 73.2% in formula-fed infants who received the HBV vaccine alone, and 90.9 vs 90.3% in those who received the HBV vaccine and HBIG (54). It may be prudent, however, to counsel against breastfeeding by women receiving antiviral agents, because the safety of these drugs during lactation has not been demonstrated.
It has long been recognized that perinatal transmission of HBV accounts for the majority of chronic infections worldwide, and strategies to affect HBV burden should incorporate methods to decrease this mode of acquisition. Many individuals chronically infected with HBV are women of child-bearing potential, and while it is the minority who have serious liver disease requiring intervention during pregnancy, care of this special population is understudied and merits consideration. Management of HBV during pregnancy includes recognition of maternal virological status, assessment of liver disease and minimization of risk for perinatal transmission of infection. This may include simple monitoring, changes in obstetrical care or administration of antiviral therapy in late pregnancy or throughout pregnancy. Passive and active immunoprophylaxis and monitoring for infection or immunity in newborns is an integral part of this management. Opportunities exist for case detection and prevention in household and sexual partners as well. The unique aspects of this management, with consequences for both mothers and newborns, and the relative lack of data, make this a critical medical challenge.
Conflicts of interest
The author declares no conflicts of interest.