One of the pregnancy complications most frequently examined in HBV carriers is GDM, probably because GDM itself is among the most commonly encountered pregnancy complications. An association between maternal HBV carriers with GDM was first observed when HBV status was examined as a possible confounding factor in the relationship between maternal third trimester ferritin level at the time of the oral glucose tolerance test (OGTT) with the diagnosis of GDM [7]. This association was subsequently confirmed in a case-control study [8] and a cohort study in the same population [9]. This relationship was validated more recently in a repeat case-control study in the same population in which the OGTT was applied to both the study and control groups regardless of the presence or otherwise of established risk factors, when the HBV carriers exhibited a 77% increased incidence of GDM [10]. This association has since been supported by a number of reports from different populations [11–16], and the positive association is unlikely to be influenced by the choice of the diagnostic criteria for the OGTT, since the 19.6% prevalence in the HBV carriers found in the case-control study referred to before [10] was similar to the 19.0% reported in the earlier case-control study [8], whereas the 12.3% prevalence in the control group was similar to the 10.2% in the non-carrier obstetric population [9] and 11.2% in the matched high-risk non-carrier mothers [8]. In the latest report from China, increased GDM was confirmed with the aOR of 1.13 [16]. A meta-analysis on this issue has found an overall 47% increased risk of GDM among HBV carriers [17]. This association is unrelated to HBV markers such as HBeAg status, viral load, or presence of circulating HBV DNA [10, 15], although a Thai study found 43.4% increase in GDM when HBeAg was seropositive [13].

Nevertheless, a number of other studies [18–28] did not demonstrate such an association. The reasons for the conflicting findings have not been elucidated, but several factors could be involved. The most important factor is probably related to the method of screening and diagnosis of GDM. In the studies conducted in Hong Kong, there was either universal screening [7–9] or that OGTT was applied to everyone in both the study and control groups irrespective of the presence of risk factors [10]. It has been shown that between 34% [29] and 54% [30] of the women with GDM diagnosed through universal screening were missed by risk factor-based screening. Thus, if the studies reporting negative findings did not apply universal screening, and that HBsAg seropositivity was not adopted as a risk factor for GDM, it was likely that there was an under-diagnosis of GDM among the HBV carriers. Another reason is that systemic inflammatory response is involved in the development of GDM [31, 32], and early pregnancy elevation in hepatic alanine aminotransferase (ALT) is associated with subsequent GDM [33, 34]. If the presence of low grade inflammation, especially hepatic inflammation, had not been examined in these studies, HBV carriers at risk of GDM would not have been screened and therefore the diagnosis could be missed. On the basis of the aforementioned studies, it would be prudent to include HBsAg seropositivity as a risk factor for GDM, especially in the populations with high prevalence of both chronic HBV infection and GDM, so that the occurrence of GDM would not be overlooked in case universal screening is not adopted as a routine antenatal test.

PE belongs to the spectrum of pregnancy hypertensive disorders, which also included gestational hypertension (GH, also described as pregnancy-induced hypertension or PIH in some reports) [35]. Among this spectrum of disorders, the best defined is PE which refers to the onset of hypertension with proteinuria arising de novo, or superimposed on pre-existing hypertension. The diagnosis of GH/PIH and PE should be mutually exclusive, although some studies labelled both GH and PE together as PIH. There are many established risk factors for PE, including nulliparity status, advanced age (≥ 35 years), high body mass index (BMI) at booking (> 25kg/m²), and presence of significant medical history that included cardiovascular disease, renal and autoimmune disorders, and pre-GDM [36, 37]. Hence in the assessment of the relationship between maternal HBsAg seropositivity with PE, these confounding factors should also be taken into account.

There are conflicting findings in the literature. An increase in superimposed PE (1.1% versus 0.67%) [12], as well as no association with PE [38], has both been reported, but the majority of studies reported a variable reduction in PE, such as 0.73% versus 1.10% in one study [24], 0.22% versus 0.79% in another study [22], 29% reduction in a third study [39], and a 20% reduction in a fourth study [27]. A meta-analysis which included 11 studies and 11,566 PE patients confirmed a 23% reduction of PE [40]. It is proposed that the altered immune response in chronic HBV infection [41–43] is compatible with the maternal immune tolerance of the conceptus [39], thus reducing the risk of PE. Indeed, in mothers who develop PE, their immune responses include a shift in favor of T helper 1 (Th1) cell activity, activation of CD4⁺ and CD8⁺ lymphocytes and deficiency of T regulatory cells, lower interleukin-10 (IL-10) production in peripheral blood mononuclear cells, and increased systemic inflammation with production of pro-inflammatory cytokines such as tumour necrosis factor-α (TNF-α) [44–46], which resemble the immune clearance phase of chronic HBV infection. It is therefore possible that one important explanation of the conflicting findings is the phase of chronic HBV infection of the pregnant women in the study population, those who were trying to clear the virus could also manifest features of rejection of the conceptus hence predisposing to PE.

As explained in the preceding section, the condition described as PIH in some studies could be a mixture of different conditions. This may explain the reported increase in hypertensive disorders (15% versus 8%) in one study [47], 13% versus 2.89% in another study [12], and a 2.2-fold increase in a third study [48]. However, the majority of studies found no difference in incidence [8, 11, 19, 23, 38, 49]; or reduced GH (0.97% versus 1.5%) [24] and 17% reduction [27], and reduced PIH (2.01% versus 3.58%) [22] and a 21% reduction [39]. Bearing in mind the aforementioned confounding risk factors for hypertension in pregnancy [36, 37], as well as the relationship between chronic HBV infection with PE, it is very likely that maternal chronic HBV infection per se has no adverse effect on PIH.

This is a rare yet potentially very serious obstetric complication, but despite HBV being a hepatotropic virus, only a handful of studies conducted in China have examined this issue. In an earlier cohort study of 3,329 women, a 70% increase in ICP was found among the 346 HBV carriers as compared with the HBV negative women, and within the group of HBV carriers, ICP was more common among the HBeAg positive (aOR 2.96) then the HBeAg negative (aOR 1.52) women [26]. A recent report also supported this observation, with a much greater increase in ICP [odds ratio (OR)]3.4 found among the HBV carriers [27]. HBV infection and ICP together appeared synergistic, both together produced the highest rates for PROM, meconium-stained amniotic fluid, and CD compared with either complication alone [50]. These reports suggest that although a relatively uncommon complication, ICP could have been a hitherto overlooked event which could have resulted directly in, as well as contributing to, some of the adverse perinatal outcomes observed in some earlier studies.

In many reports, no distinction was made between spontaneous versus medically indicated PTB. Although threatened preterm labour was increased in HBV carrier mothers, it was not translated into increased PTB [8]. A later cohort study [51], and a meta-analysis of 6,781 women [52], also found no association between HBV carriers with preterm labour.

On the other hand, a majority of studies found increased PTB [12, 13, 19, 23, 53–55], which ranged from 1.3% in a study which also showed HBeAg seropositivity to increase PTB by 25% [13], to 5.2-fold in spontaneous PTB which was unrelated to the level of HBsAg or HBV DNA, or viraemia [53]. It is possible that viral activity plays a role here, as a Chinese study found no difference in overall PTB (5.4% versus 5.0%) between the 1,728 HBV carriers and 1,497 controls, but there was 18% increased PTB for each log₁₀ copy/mL increase in second trimester HBV DNA [48]. A larger Chinese study with 489,965 women found seropositivity for HBsAg alone had 26% and 18% higher risk of PTB and early PTB (< 34 weeks gestation) respectively, while seropositivity for both HBsAg and HBeAg had 20% and 34% higher risk respectively [55]. Interestingly, a recent study found no increase in PTB overall, but when analyzed by mode of delivery, HBV carrier women who had vaginal delivery had a 14% increase in PTB [27]. Increased PTB was confirmed in several meta-analyses which reported 2.36-fold increase [56], 26% increase [51], and 16% increase with HBV seropositivity alone and 21% with both HBsAg and HBeAg seropositivity [57]. While the relationship between chronic HBV infection with increased PTB appears established, the extent of the effect could be quite variable due to differences in underlying or associated factors which have been explored in only a few studies.

One such factor is liver function. In a Chinese study, abnormal liver function test, half of which was related to non-alcoholic fatty liver disease, has turned out to be the independent risk factor for PTB rather than HBsAg seropositivity [58]. Furthermore, it is long established that preterm labour and PTB are largely attributed to the presence of genital tract and/or intra-amniotic infection [59], and it is possible that maternal chronic HBV infection predisposes to increased genital tract/intra-amniotic infection due to altered maternal immune response. As well, PTB is almost an inevitable outcome following preterm prelabour rupture of the membranes (PPROM), which was shown in a study that examined this complication specifically [12]. On the other hand, in countries with high endemicity for HBV infection such as China, PPROM is generally attributed to genital tract pathogens [60], and as genital tract or intrauterine infections were not specifically examined or excluded in the aforementioned studies, it is not possible to establish the independent role of chronic HBV infection in the occurrence of PTB with or without preceding PPROM. Thus, HBeAg status, hepatic inflammation, and possibly predisposition towards genital tract infection due to the immune modulation from chronic HBV infection, probably interact in different and varying ways to influence the occurrence of PTB.

The occurrence of PROM often occurs at term, but only a handful of studies have examined PROM specifically. A Chinese study found PROM to be increased by 38% in HBsAg seropositive mothers [26], and an Israeli study found a 30% increase [23], but it was not clarified if the cases of PROM all occurred at term or included those occurring preterm. In an Iranian study, although HBsAg seropositive mothers had higher incidence of PROM (6.89% versus 2.44%), this failed to reach statistical significance, probably due to the relatively small number of subjects, even though there was a significant 3.2-fold increase for PPROM [12]. Finally in a recent report, a 9% reduction in PROM was found [27]. However, in the Chinese study comparing 346 HBsAg carriers with 2,983 non-carriers, PPROM was neither translated into increased PTB (19.8% versus 16.6%) nor was it influenced by HBeAg status [26]. In these studies, other reported factors for PROM, including genital tract infections, were not examined or reported, and a definitive role of chronic HBV infection in the occurrence of PROM remains to be established.

The two major types of APH include PA and PP. In general, APH was increased by 2.18-fold with HBV alone [8], and 2.82-fold with HCV coinfection [19]. Where the type of APH was specified, HBsAg seropositivity was shown to increase PA by 4-fold [8] and 2-fold [23]. In a recent study which had also analyzed pregnancy outcome by mode of delivery, a 44% increased risk of PA was found in women with vaginal delivery [27]. However, there are a number of other confounding factors which influence the risk of PA, such as maternal smoking, and these were not examined or reported in the aforementioned reports. In the latest meta-analysis, no association with PA could be demonstrated [61].

Although there may be an association between HBsAg seropositivity with PA, no associated increase with PP was reported in the studies which examined this issue [8, 11, 21]. No association with PP was found in the latest meta-analysis [61]. Nevertheless, the preference of CD in some countries for the purported reason of reducing MTCT would inevitably increase the incidence of multiparous women with a scarred uterus, so that their risk of PP in subsequent pregnancies would likely be increased on account of their uterine scars [62].

CD includes those performed as scheduled (elective) or as an emergency, but studies often did not distinguish between these two categories in their reports. In a US study comparing 814 HBV carriers with 296,218 non-carriers, the former group had 31% reduction in CD [19]. On the other hand, a Hong Kong study on 9,526 women found more elective CD (13.8% versus 12.9%) and less emergency CD (11.6% versus 15.5%), but the indications for CD were not provided [24]. An Israeli study on 749 HBV (78.6%) and/or HCV carriers found a 50% increase in overall CD [23]. Studies in China similarly documented increased CD by 70% in one study [48] and 12% in a more recent report [16]. It should be mentioned that in China, reducing or preventing MTCT is often taken as the indication for CD, both on maternal request and obstetrician’s recommendation [63]. But CD has been shown not to further reduce MTCT with timely neonatal immunoprophylaxis [64], so that elective CD is not indicated for this reason. Nevertheless, any associated pregnancy complications could conceivably increase the incidence of CD, and future studies should adjust for pregnancy complications as confounders.

Few studies examined this outcome. A cohort study in Hong Kong has found increased PPH (4.0% versus 2.7%) although the mechanism for this association was not elaborated [24]. In another study from Hong Kong which examined the influence of maternal age on PPH, with maternal HBsAg status included as a potential confounding factor in the analysis, had unexpectedly found a 16% increase in PPH, and again the mechanism of the increase was not further examined [65]. A recent report also demonstrated increased PPH (OR 1.16) [27]. It is possible that some of the increase in PPH could well be related to the associated increase in complications as well as CD, and this association should be clarified in further studies with adjustment for obstetric complications and interventions.

An earlier study reported increased miscarriage among 264 women with viral hepatitis [54]. A later study found 71% increased miscarriage when 513 HBV carriers were compared with 20,491 controls [25]. As well, increased early abortion and overall abortion were also noted in HBV carrier women on in vitro fertilization (IVF) treatment, attributed to HBV infection of some embryos because HBV messenger ribonucleic acid (mRNA) fragments were detected in 13.2% (5/38) and 5.6% (1/18) of the cleavage embryos from the mothers or the fathers having HBV infection [66]. It is therefore possible that in women with HBV infection of their gametes or in their ovaries, there would be increased risk of spontaneous miscarriage, but as miscarriage is usually not included in the analysis of obstetric outcome, information on this issue is extremely limited. More detailed laboratory studies would be warranted to delineate the role of HBV infection of the gamete in miscarriage in these mothers.

Fetal malformations, especially serious and lethal malformation, are collectively a cause for spontaneous miscarriage, and such data is usually not available in the studies on obstetric outcome. The fact that HBV mRNA fragments could be detected in 13.2% and 5.6% of the cleavage embryos from the mothers or the fathers having HBV infection [66] could be the underlying reason for the reported increased congenital malformations in the handful of studies which examined this issue [23, 50, 54]. It is noteworthy that in one study [50], the rate of birth defect was 2% (1/50 cases) for maternal HBV infection, and increased to 16% (8/50 cases) among mothers with both HBV and ICP. However, the data should be interpreted with caution because in this study, there were only 50 cases of HBV infection, and 50 cases of ICP with HBV infection, compared with 50 controls, so that there could be some form of bias. In another study, the aOR for malformation was 1.4 [23]. However, no details on the types of malformations, or the organs/systems with malformation, were provided in these reports, and it remains unclear if there was an increase in overall structural malformations or some specific organs are involved in particular, and whether the increased malformations were mostly major, minor, or a mixture.

Increased SB has been reported in an early study [54]. Subsequently other studies have also reported increased IUFD or SB, and perinatal mortality. A later study found markedly increased IUFD (5.56% versus 0.44%, OR 12.1) and SB (2.89% versus 0.44%, OR 8.8) between HBV carriers and non-carriers, but the distinction between IUFD and SB was not provided [12]. Another study found a significantly increased aOR of 1.8 for perinatal mortality [23]. The explanations for the increased IUFD, SB, and perinatal mortality were not provided in these studies, but factors which include increased congenital malformations, fetal hypoxaemia (see later discussion), and pregnancy complications such as PTB could all have contributed.

An increase in LBW [11, 23, 54, 67] and SGA [11, 67] infants have been reported. As LBW infants are often related to PTB, one of the explanations is likely to be the increased PTB in HBV carriers [51, 55–57]. Although another study found no increased LBW infants, HBeAg seropositivity was associated with 25.8% increase in LBW infants [13]. Nevertheless, the contribution of fetal growth restriction to LBW infants, and the potential mechanisms/explanations for the increased SGA infants, were not available.

A number of studies, examining this issue from different angles, suggest that maternal chronic HBV infection probably enhances fetal growth in utero. Some studies have found significant reduction in SGA infants, 20–25% in one study [11], 21% in another [18], and 16% in a third [28]. Interestingly, one study found reduction in SGA infants (OR 0.68) only among women who had CD [27]. As well, increased birthweight was found in infants born to HBV carriers, although this was significant only in Oriental women [47]. Finally, increased macrosomic infants have been reported. In Iran, a 2.5-fold increase in macrosomic infants was found in a case-control study [12]. In Hong Kong, a cohort study confirmed an independent association with 15% and 11% increase in macrosomic and LGA infants respectively, and a 14% reduction in LBW infants [68]. Other Chinese studies similarly found higher mean birthweight [24], 68% increase in macrosomic infants [48], and increased LGA infants (15.1% versus 10.7%) which was confined to HBeAg negative women (aOR 1.91) [26].

The increase in LGA and macrosomic infants is intriguing, as the effect persisted despite controlled for the presence of GDM as a confounding factor [68]. The underlying mechanism could be related to that accounting for the reduction in PE [21, 22, 39, 40], in form of enhanced immune tolerance induced by chronic HBV infection, which becomes extended to the conceptus thus facilitating fetal growth as well. Another factor could be higher maternal ALT level, as non-GDM women delivering LGA infants had higher mean ALT levels and ALT level > 90th percentile (26 IU/L) was associated with a 4-fold increased LGA babies [69]. The enhanced immune tolerance and higher ALT level are most likely related to the phase of chronic HBV infection in the majority of these mothers.

Increased fetal distress, including meconium stained amniotic fluid (MSAF), non-reassuring fetal heart rate (FHR) patterns, and low fifth minute Apgar score (AS), have been reported [22, 23, 48]. In one of these studies [48], maternal HBV infection was shown to be independently associated with fetal distress with aOR of 1.40, but the underlying reasons for fetal distress were not provided. In another study, there was 60% increased non-reassuring FHR patterns but 20% lower incidence of MSAF [23]. A meta-analysis on 7,600 pregnant HBV carriers from 18 studies indicated an 80% increase in birth asphyxia [56]. The mechanisms of increased fetal distress probably reflect increased fetal hypoxaemia, which could help to explain the 80% increased perinatal mortality reported in one study [23], and the 12.1-fold increased IUFD and 8.8-fold increased SB in another study [12]. Irrespective of the mechanism, the increased IUFD, SB and fetal distress would call for proper fetal assessment and monitoring whenever fetal wellbeing is suspected to be jeopardized, and it is prudent to apply continuous intrapartum fetal monitoring for all pregnancies in HBV carriers.

It has been shown that non-pregnant asymptomatic HBV carriers were more likely to have psychiatric disorders than comparison subjects (30.2% versus 11.6%), and had higher depression and anxiety scores and lower Global Assessment of Functioning scores [70]. Overall, up to half of patients with HBV infection have psychiatric morbidity [71]. Thus, it would be logical to anticipate increased psychiatric morbidity in HBV carriers. However, no studies has examined either antenatal and postnatal psychiatric or psychological status or disturbances, especially the risk of postnatal depression, in pregnant HBV carriers. Only one study examined pregnant HBV carriers and found higher levels of anxiety and stress in these mothers compared with healthy pregnant women [72]. However, it is unclear to what extent could such finding be influenced by the fact that these women only discovered their HBV carrier status at routine antenatal screening, as they were asymptomatic and unlike the patients who were aware and possibly symptomatic of their chronic HBV infection. Since postnatal depression could have serious and potential fatal consequences for the mother, prospective studies are warranted to elucidate the psychological status and psychiatric morbidity among mothers identified to have chronic HBV infection for the first time during pregnancy.

Although breast feeding is often avoided to reduce MTCT, breast feeding did not increase MTCT [73], especially when timely combined passive-active neonatal immunoprophylaxis is administered routinely. The safety of breast feeding with regards to MTCT has been confirmed in a meta-analysis of 10 studies with 751 breast-fed and 873 non-breast-fed infants by assessing infant peripheral blood HBsAg or HBV DNA positivity at age 6–12 months, and antibody to hepatitis B surface antigen (anti-HBs) positivity at 6–12 months [74]. Therefore, breast feeding should be allowed with proper neonatal immunoprophylaxis.