Journal of Viral Hepatitis, 2016, 23, 812–819
12545
HBsAg positivity during pregnancy and adverse maternal
outcomes: a retrospective cohort analysis
J. Tan,1,2,3,† X. Liu,4,† X. Mao,1 J. Yu,1 M. Chen,4 Y. Li1 and X. Sun1
Center, West China Hospital, Sichuan University, Chengdu, China; School of Public Health, Sichuan University, Chengdu, China; 3Chengdu
2
University of Traditional Chinese Medicine, Chengdu, China; and 4West China Women’s and Children’s Hospital, Sichuan University, Chengdu,
China
Received January 2016; accepted for publication March 2016
SUMMARY. Hepatitis B virus infection characterized by
HBsAg positivity during pregnancy is a well-recognized
issue in developing countries, but the association between
HBsAg positivity and adverse maternal outcomes remains
uncertain. To examine the association between HBsAg pos-
itivity during pregnancy and adverse maternal outcomes, a
retrospective cohort study was conducted in Sichuan pro-
vince, China. Deliveries were recorded from six hospitals
between 1 January 2009 and 31 December 2010. Pre-
eclampsia, gestational diabetes mellitus (GDM), postpartum
haemorrhage (PPH), intrahepatic cholestasis, Caesarean
section and placenta previa were prespecified adverse
maternal outcomes. We used two multivariate logistic
regression models to assess the association between HBsAg
positivity and adverse maternal outcomes. In total, 948
(4.2%) pregnant women were HBsAg positive from 22 374
deliveries. Pregnant women with positive HBsAg had
INTRODUCTION
Hepatitis B virus (HBV) infection characterized by positivity
for HBsAg during pregnancy is not an issue that can be
neglected. It is common, particularly in developing coun-
tries. The prevalence of HBsAg positivity during pregnancy
is high, for example in China (7.6%) [1] and Thailand
(6.2%) [2].
However, it remains uncertain as to whether HBsAg
positivity during pregnancy is associated with adverse
Abbreviations: aOR, adjusted odds ratio; ART, assisted reproduc-
tive technology; CI, confidence interval; GDM, gestational diabetes
mellitus; HBsAg, hepatitis B surface antigen; HCV, hepatitis C
virus; PPH, postpartum haemorrhage.
Correspondence: Xin Sun, Chinese Evidence-based Medicine Cen-
ter, West China Hospital, Sichuan University, No.37 Guo Xue
Xiang, Chengdu, China.
E-mail: sunx79@hotmail.com
† To examine the association of HBsAg positivity during
Jing Tan and Xing-hui Liu contributed to the work equally and
are co-first authors.
doi:10.1111/jvh.
1
Chinese Evidence-based Medicine
higher risk of GDM (aOR1.41, 95%CI 1.15–1.74), PPH
(1.44, 1.13–1.83), intrahepatic cholestasis (1.74, 1.40–
2.16) and Caesarean section (1.24, 1.06–1.45). No statisti-
cal associations were found between HBsAg positivity and
pre-eclampsia (1.36, 0.94–1.97), and placenta previa
(1.21, 0.87–1.67). HBsAg positivity during pregnancy was
associated with higher risk of multiple adverse maternal
outcomes. Although the causality has yet to be established,
efforts may be warranted in routine care, particularly in
those with high risk for adverse maternal outcomes, given
the volume population infected with HBsAg. Future studies
are needed to establish causality and examine the impact
of HBeAg on the adverse outcomes.
Keywords: caesarean section, gestational diabetes, HBsAg
positivity, intrahepatic cholestasis, postpartum haemor-
rhage.
maternal outcomes. The findings are inconsistent across
studies. For instance, while a couple of studies reported
lower risk of pre-eclampsia in women with HBsAg positiv-
ity [3,4], most showed no association between pre-eclamp-
sia and positive HBsAg [2,5–7]. Some studies suggested
higher risk of gestational diabetes mellitus (GDM) in those
with HBsAg positivity [6–9], but these findings were not
replicated in others [2, 5,10–12].
In addition, previous studies have not examined the
association of HBsAg with other adverse outcomes, includ-
ing postpartum haemorrhage (PPH) that results in about a
quarter of the total maternal mortality globally [13], and
intrahepatic cholestasis which may cause premature deliv-
ery, foetal distress, meconium-stained amniotic fluid and
even stillbirth [14]. Up to now, the published studies exclu-
sively reported data from countries where the prevalence
of HBsAg positivity during pregnancy is low. The findings
may not be applicable to populations at higher risk, such
as those in the region of East Asia.
pregnancy with adverse maternal outcomes in the Chinese
© 2016 John Wiley & Sons Ltd

population, we conducted a retrospective cohort study
enrolling patients from six hospitals in the Sichuan pro-
vince of China.
MATERIAL AND METHODS
Data sources
We selected six hospitals from four cities based geograph-
ically in Sichuan province, China. Of those, three were
located in Chengdu (the largest city in Sichuan, account-
ing for about one-fifth of the population in the province),
one in Panzhihua (south-west of the province), one in
Zigong (south of the province) and one in Suining (mid-
dle of the province). One is an academic teaching hospi-
tal (West China Women’s and Children’s Hospital,
Sichuan University), three women and children hospitals
and two regional general hospitals. This study was
approved by the Medical Ethics Committee of West China
Second Hospital, Sichuan University (Approval number:
2014–2020 and date of approval: September 18th
2014).
The West China Women’s and Children’s Hospital is the
referral centre for the five other hospitals and provides
instructional support to the five hospitals. The care of preg-
nant women was standardized in these hospitals so that
each woman received a health record card with a unique
number at the first visit for pregnancy. Corresponding to
that number, all the data – including physical examina-
tion, laboratory tests, medical and pharmacy orders at all
visits – were recorded at the hospital. All pregnant women
were required to conduct follow-up visits at those hospitals
until delivery.
We used predefined and pilot-tested case report forms to
collect data from medical record databases from those six
hospitals. Prior to the beginning of the study, we trained
research personnel and clinical staff responsible for the
obstetric departments to ensure adequate understanding of
the procedures and case report forms. The trained research
personnel then collected data from medical charts or elec-
tronic medical records and entered the data into the
research database. The accuracy of data collection was
ensured through double entry and subsequent consistency
checks.
Patient selection and ascertainment of exposure
The retrospective cohort consisted of pregnant women
with 20 gestational weeks or older during 1 January
2009 to 31 December 2010. All pregnant women were
screened for HBsAg at their first antenatal visit during
the 11–12+6 gestational weeks, using enzyme-linked
immunosorbent assay (ELISA). The patient status regard-
ing HBsAg positivity was recorded in the patient medical
records.
© 2016 John Wiley & Sons Ltd
HBsAg positivity and adverse maternal outcomes 813
Maternal outcomes
Maternal adverse outcomes included pre-eclampsia, gesta-
tional diabetes mellitus (GDM), postpartum haemorrhage
(PPH), intrahepatic cholestasis, Caesarean section and pla-
centa previa. These outcomes were extracted from the indi-
vidual medical record.
Collection of covariates
For each individual pregnant woman, we collected the fol-
lowing information: demographic characteristics, gesta-
tional characteristics, physical examination at admission,
laboratory examination, gestational comorbidities and
medical interventions. All the information was extracted
from the medical records.
Demographic information included maternal age, resi-
dence location and marriage status. Gestational character-
istics included gestational week, parity, history of
Caesarean section and multiple gestations. Medical inter-
ventions included application of assisted reproductive tech-
nology (ART), use of one or more uterotonics during the
third stage of labour, induction of labour and delivery
mode.
The physical examination included height, weight, blood
pressure, pulse, fundal height, abdominal girth and inter-
tuberous diameter. The laboratory tests included the num-
ber of erythrocytes, thrombocyte, leucocyte, the
haemoglobin concentration, 75 g or 100 g oral glucose
tolerance test, total bilirubin, total bile acid, total protein,
albumin. The findings of both physical examination and
laboratory tests were used for diagnosis of gestational
comorbidities.
Gestational comorbidities included iron deficiency anae-
mia, hepatitis C virus (HCV), hypertension, cardiac dis-
eases, gynaecological diseases (hysteromyoma, ovarian cyst
and pelvic inflammation), respiratory diseases (bronchial
asthma and pulmonary tuberculosis), thyroid diseases (hy-
perthyroidism and hypothyroidism), GDM, sexually trans-
mitted diseases, psychological and neural diseases
(epilepsy, depression or anxiety). Among those, GDM was
the maternal outcome we set previously, and GDM was
deemed as a gestational comorbidity considering its poten-
tial effects on other maternal outcomes.
Hypertension included gestational hypertension and
chronic hypertension. Cardiac diseases included one or
more types of heart disease identified through aetiological
diagnosis or pathological diagnosis, including congenital
heart diseases, rheumatic heart disease, coronary disease,
hypertensive heart disease, perinatal myocardial dis-
ease, dilated cardiomyopathy, myocarditis, mitral stenosis,
interventricular septal defect, myocardial infarction, nodal
tachycardia, atrial fibrillation and sinus irregularity. Sex-
ually transmitted diseases included syphilis, gonorrhoea,
AIDS, genital herpes and Condyloma acuminatum.
814 J. Tan et al.
All gestational comorbidities were self-reported or clini-
cally diagnosed at any time of antenatal care by
clinicians.
Data analysis
We descriptively reported the distribution of demographic
features, gestational characteristics, medical interventions
and gestational comorbidities in the HBsAg-positive and
HBsAg-negative groups. We checked the missing data for
all variables and compared the distribution difference
between those with complete data and those without. We
included all the patients with complete data in our
analyses.
We also compared those characteristics between the
groups using the Pearson’s chi-square or Fisher’s exact test
for categorical variables, and t-test or rank-sum (Mann–
Whitney) test for continuous variables. We categorized the
maternal age as ≥35 years or <35 years. We combined
the variables of parity and history of Caesarean section,
which was categorized as nulliparity, multiparity and non-
Caesarean history, or multiparity and Caesarean history
for avoiding multicollinearity in the regression model. We
calculated BMI by the height divided by the square of
weight.
For each of the prespecified adverse maternal outcomes,
we conducted a univariate analysis of the association with
HBsAg positivity. We applied the Pearson’s chi-square test
or Fisher’s exact test to the univariate analyses. These
analyses provided a preliminary estimation of the
associations.
We further used the multivariate logistic regression to
assess the association between HBsAg positivity and
adverse maternal outcomes and reported adjusted odds
ratio (aOR) and 95% confidence intervals (CIs). In the
adjusted analyses, we included demographic and gesta-
tional characteristics, use of medical interventions and
gestational comorbidities, given their statistical signifi-
cance in the univariate analyses, biological rationale and
previous findings. The gestational comorbidities were
selected as covariates if the p-value of the univariate
analysis was less than 0.1. We carefully considered the
covariates in the adjusted analysis. For each of the
adverse maternal outcomes, covariates could differ in the
multivariate logistic regression models. Use of uterotonics
during the third stage of labour, induction of labour and
delivery mode were considered appropriate for the out-
come of PPH only.
To distinguish the effects of different covariates, we
developed 2 regression models for each adverse outcome.
Model 1 adjusted for the effects of maternal demographic
characteristics, gestational characteristics and medical
interventions. In model 2, gestational comorbidities were
added to model 1. We used STATA12.0 (StataCorp College
Station, Texas USA) for the statistical analysis.
RESULTS
The database consisted of 22 374 deliveries collected from
1/Jan/2009 to 31/Dec/2010. The number of deliveries
during the 2 years ranged from 358 to 9368 across the 6
hospitals. The median maternal age was 27 (interquartile
range 23–31) years, and the median gestational age was
39 (38–39) weeks. HBsAg positivity was present in 948
(4.2%) pregnant women. Compared with the HBsAg-nega-
tive group, those with positive HBsAg were more likely to
be aged over 35 years (14.2% vs 11.9%, P = 0.03) and
have a history of Caesarean section (10.2% vs 7.7%,
P < 0.001). No statistical differences were present in mar-
riage status (97.7% vs 97.4%, P = 0.63), nulliparity
(75.2% vs 77.0%, P = 0.22), multiple gestation (2.6% vs
1.9%, P = 0.09) and mean BMI (26.8 vs 26.6, P = 0.07)
between HBsAg-positive and HBsAg-negative women. With
respect to the use of medical interventions, those with
HBsAg positivity were more likely to use ART (3.2% vs
2.2%, P = 0.05) and Caesarean delivery (73.9% vs 68.5%,
P < 0.001), but fewer induced labour (3.6% vs 5.5%,
P = 0.009). No significant difference was observed in the
uterotonics use between the groups (99.3% vs 98.9%,
P = 0.26) (Table 1).
Table 2 describes the distribution of gestational comor-
bidities in the HBsAg-positive group and HBsAg-negative
groups. The proportion of GDM in the HBsAg-positive
group was higher than the negative (11.8% vs 8.2%,
P < 0.001). We found 26 cases of HCV in our cohort; the
proportion of HCV in the HBsAg-positive group was
slightly higher than the HBsAg-negative group, but not
significant (0.32% vs 0.11%, P = 0.06). Conversely, the
incidence of gynaecological diseases in the HBsAg-positive
group was 5.4%, lower than 6.8% in the HBsAg-negative
group (P = 0.07). The other gestational comorbidities had
no statistical difference between the HBsAg-positive and
HBsAg-negative groups.
Table 3 shows the unadjusted association and results of
the multivariate logistic regression analyses between
HBsAg positivity and adverse maternal outcomes. Com-
pared with those with negative HBsAg, pregnant women
with positive HBsAg appeared to have higher risk of pre-
eclampsia (3.5% vs 2.4%,P = 0.05). After adjusting for the
gestational and demographic characteristics, and use of
medical interventions, HBsAg positivity was not statisti-
cally associated with pre-eclampsia (aOR 1.34, 95% CI
0.94–1.94); further adjustment for gestational comorbidi-
ties yielded a similar estimate (aOR 1.36, 95% CI 0.94–
1.97). The incidence of GDM in the HBsAg-positive group
was statistically higher than that in the negative group in
the univariate analysis (11.8% vs 8.2%, P < 0.001). In
Model 1 of adjusting for the demographic characteristics
and the use of medical interventions, there was continued
statistical association between HBsAg positivity and
increased risk of GDM (aOR 1.42, 95% CI 1.15–1.75);
© 2016 John Wiley & Sons Ltd
 HBsAg positivity and adverse maternal outcomes 815

Table 1 Demographic and clinical characteristics of cohort

Variables HBsAg positive (%) HBsAg negative (%) P value

Maternal age (years)

<35 813 (85.8) 18 876 (88.1)
≥35 135 (14.2) 2550 (11.9) 0.03
Residence location
City or town 559 (59.0) 12 824 (60.0)
Rural area 389 (41.0) 8602 (40.1) 0.59
Married
Yes 926 (97.7) 20 874 (97.4)
No 22 (2.3) 552 (2.6) 0.63
Parity
Nulliparity 713 (75.2) 16 489 (77.0)
Multiparity 235 (24.8) 4937 (23.0) 0.22
Caesarean history
Non-Caesarean history 851 (90.0) 19 769 (92.3)
Caesarean history 97 (10.2) 1657 (7.7) <0.001
Multiple gestation
Yes 25 (2.6) 402 (1.9)
No 923 (97.4) 21 024 (98.1) 0.09
BMI
/ 26.8
3.4 26.6
3.2 0.07
ART
Yes 30 (3.2) 473 (2.2)
No 918 (96.8) 20 953 (97.8) 0.05
Uterotonics use
Yes 941 (99.3) 21 183 (98.9)
No 7 (0.7) 243 (1.1) 0.26
Induced labour
Yes 34 (3.6) 1188 (5.5)
No 914 (96.4) 20 238 (94.5) 0.009
Delivery mode
Vaginal 248 (26.2) 6764 (31.6)
Caesarean section 700 (73.9) 14 662 (68.5) <0.001

similar findings (aOR 1.41, 95% CI 1.15–1.74) were
obtained with Model 2 after adding gestational comorbidi-
ties, including presence of iron deficiency anaemia, hyper-
tension, respiratory disease and thyroid disease. The 8.3%
of pregnant women with HBsAg positivity had PPH, and
this was significantly higher than that in these HBsAg neg-
ative (8.3% vs 5.5%, P < 0.001). Considering the effects of
uterotonics use, induced labour and delivery mode, Model
1 also showed a significant association between HBsAg
positivity and PPH (aOR 1.45, 95% CI 1.14–1.85); a simi-
lar effect estimate was observed when gestational comor-
bidities were additionally included in Model 2 (aOR 1.44,
95% CI 1.13–1.83). The risk of intrahepatic cholestasis in
the HBsAg-positive group was higher than the negative
group in univariate analysis (10.7% vs 6.3%, P < 0.001).
Adjusting for demographic characteristics and the use of
medical interventions continued to show statistical associa-
tion of HBsAg positivity with increased risk of intrahepatic
cholestasis (aOR 1.77, 95% CI 1.43–2.20); the addition of
gestational comorbidities showed similar findings (aOR
1.74, 95% CI 1.40–2.16). In univariate analysis, the inci-
dence of Caesarean section in the HBsAg-positive group
was higher than in their negative counterparts (73.8% vs
68.4%, P < 0.001). Similar estimates of increased risk of
Caesarean section in the HBsAg-positive group were
observed in Model 1 (aOR 1.26, 95% CI 1.08–1.48) and
Model 2 (aOR 1.24, 95% CI 1.05–1.45). No significant dif-
ference was observed in risk of HBsAg positivity and pla-
centa previa (4.2% vs 3.3%, P = 0.11). Multivariate
analyses were consistent with the results of both Model 1
(aOR 1.25, 95% CI 0.90–1.74) and Model 2 (aOR 1.21,
95% CI 0.87–1.67).
DISCUSSION
In this study, we found that pregnant women with positive
HBsAg had statistically higher risk of gestational diabetes,
postpartum haemorrhage, intrahepatic cholestasis and

© 2016 John Wiley & Sons Ltd

816 J. Tan et al.

Table 2 The gestational comorbidities by HBsAg status

Variables HBsAg positive (%) HBsAg negative (%) P value

Iron deficiency anaemia

Yes 45 (4.8) 961 (4.5)
No 903 (95.2) 20 465 (95.5) 0.70
Hepatitis C
Yes 3 (0.32) 23 (0.11)
No 945 (99.68) 21 403 (99.89) 0.06
Cardiac diseases
Yes 14 (1.5) 286 (1.3)
No 934 (98.5) 21 140 (98.7) 0.71
Hypertension
Yes 9 (1.0) 288 (1.3)
No 939 (99.1) 21 138 (98.7) 0.30
Gynaecological diseases
Yes 64 (5.4) 1151 (6.8)
No 884 (94.6) 20 275 (93.2) 0.07
Respiratory disease
Yes 7 (0.7) 124 (0.6)
No 941 (99.3) 21 302 (99.4) 0.53
Thyroid disease
Yes 5 (0.5) 127 (0.6)
No 943 (99.5) 21 299 (99.4) 0.80
GDM
Yes 112 (11.8) 1758 (8.2)
No 836 (88.2) 19 668 (91.8) <0.001
STD
Yes 2 (0.2) 99 (0.5)
No 946 (99.8) 21 327 (99.5) 0.26
Psychological and neurological diseases
Yes 1 (0.1) 37 (0.2)
No 947 (99.9) 21 389 (99.8) 0.60

Caesarean section, and potentially higher risk of pre-
eclampsia. HBsAg positivity was, however, not associated
with placenta previa. Although the current analyses do
not establish that HBsAg positivity during pregnancy is a
causal risk factor for those adverse maternal outcomes,
these findings suggest that HBsAg positivity may represent
an important factor for consideration in the care of preg-
nant women – those with positive HBsAg in the course of
pregnancy may have higher maternal risk, and appropriate
interventions may be warranted.
In our analysis, we found a potential association of posi-
tive HBsAg during pregnancy with higher risk of pre-
eclampsia (aOR 1.36, 95% CI 0.94–1.97). The nonsignifi-
cant association is likely because of insufficient power in
the multivariate regression analyses given the relatively
small number of events in the HBsAg-positive group (33
pre-eclampsia). Biological evidence also suggests that preg-
nant women with positive HBsAg may have problems with
the cardiovascular system. Ishizaka and colleagues found
that the incidence of carotid atherosclerosis significantly
increased in HBV carriers [15]. Another study showed
higher risk of stroke and myocardial infarction in patients
with positive HBsAg [16]. A couple of studies additionally
suggested that pre-eclampsia occurred in pregnancy could
significantly increase women’s risk of ischaemic heart dis-
ease and death in the future [17,18].
The analysis also suggested that pregnant women with
positive HBsAg have significantly higher risk of GDM (aOR
1.41, 95% CI 1.15–1.74). Laboratory evidence suggested
that exposure to HBV infection may cause significant insu-
lin resistance in pregnant women, potentially because
tumour necrosis factor-a could induce insulin resistance
during pregnancy [19,20], and tumour necrosis factor-a
and its receptor may significantly be increased in those
infected with HBV [21,22]. Second, the elevated concentra-
tion of serum ferritin due to the exposure to HBsAg may
also be associated with increased insulin resistance, thus
leading to increased risk of GDM [9,23,24]. Our finding is
consistent with previous studies that showed that HBsAg
positivity was an independent risk factor for GDM [6–9]; a
few other existing studies, however, did not identify statisti-
cal significant association[11,12]. This is likely due to the

© 2016 John Wiley & Sons Ltd

 HBsAg positivity and adverse maternal outcomes 817

Table 3 Univariate analyses and multivariate logistic regression analyses of associations between HBsAg positivity and
adverse maternal outcomes

Univariate analyses Multivariate analyses

Variables HBsAg positive HBsAg negative OR (95% CI) Model 1 (aOR, 95%CI) Model 2 (aOR, 95%CI)

Pre-eclampsia 33 (3.5%) 523 (2.4%) 1.44 (1.01–2.06) 1.35 (0.94–1.94) 1.36 (0.94–1.97)
GDM 112 (11.8%) 1758 (8.2%) 1.50 (1.22–1.84) 1.42 (1.15–1.75) 1.41 (1.15–1.74)
PPH 79 (8.3%) 1181 (5.5%) 1.56 (1.23–1.97) 1.45 (1.14–1.85) 1.44 (1.13–1.83)
Cholestasis 101 (10.7%) 1352 (6.3%) 1.77 (1.43–2.19) 1.77 (1.43–2.20) 1.74 (1.40–2.16)
Caesarean section 700 (73.8%) 14 662 (68.4%) 1.30 (1.12–1.51) 1.26 (1.08–1.48) 1.24 (1.05–1.45)
Placenta previa 40 (4.2%) 698 (3.3%) 1.31 (0.94–1.81) 1.25 (0.90–1.74) 1.21 (0.87–1.67)

Model 1 included sociodemographical variables, gestational characteristics, and medical interventions (components of medi-
cal interventions may vary across the outcomes).
Model 2 included covariates used in Model 1 plus gestational comorbidities.
Covariates used for each of the outcome measures are listed below. The selection of the covariates was based on statistical
significance in the univariate analyses, biological rationale and previous findings.
The categorization of the covariates was identical at each of the models. The details of the categorization of covariates are
shown in the first outcome (i.e. pre-eclampsia).
Pre-eclampsia: maternal age (less than 35 vs 35 or older), residence location (city or town vs rural), marriage status (mar-
ried vs others), parity and Caesarean history (nulliparity vs multiparity and non-Caesarean history vs multiparity and Cae-
sarean history), multiple gestation (yes vs no), BMI, use of ART (yes vs no), presence of cardiac diseases (yes vs no),
hypertension (yes vs no), presence of respiratory disease (yes vs no), gynaecological diseases and GDM (yes vs no).
GDM: maternal age, residence location, marriage status, parity and Caesarean history, multiple gestation, BMI, use of ART,
presence of iron deficiency anaemia, hypertension, presence of respiratory disease and thyroid disease.
PPH: maternal age, residence location, marriage status, parity and Caesarean history, multiple gestation, BMI, use of ART,
uterotonics use (yes vs no), induced labour (yes vs no), delivery mode (vaginal delivery vs Caesarean section), presence of
iron deficiency anaemia, HCV, cardiac diseases, hypertension, presence of gynaecological diseases and GDM.
Cholestasis: maternal age, residence location, marriage status, parity and Caesarean history, multiple gestation, BMI, use of
ART, presence of iron deficiency anaemia, HCV, gynaecological diseases and GDM.
Caesarean section: maternal age, residence location, marriage status, parity and Caesarean history, multiple gestation, BMI,
use of ART, presence of iron deficiency anaemia, cardiac diseases, hypertension, presence of respiratory disease, gynaecolog-
ical diseases, presence of thyroid disease and GDM.
Placenta previa: maternal age, residence location, marriage status, parity and Caesarean history, multiple gestation, BMI,
use of ART, presence of hypertension, gynaecological diseases and GDM.

lower prevalence of HBsAg positivity in those negative
studies. On the other hand, in this study, GDM was estab-
lished using the 75 g or 100 g oral glucose tolerance test
at 24–28 weeks of gestation according to the American
Diabetes Association diagnosis criteria of 2003 [25], but
other studies barely reported on the specific screening
methods used. The varied diagnosis of GDM among coun-
tries and hospitals probably has contributed to the differ-
ence in results.
Our study showed that HBsAg positivity was associated
with higher risk of PPH (aOR 1.44, 95% CI 1.13–1.83)
and intrahepatic cholestasis (aOR 1.75, 95% CI 1.41–
2.17), findings that were rarely investigated before.
Although not conclusive, these findings are intriguing and
may have important clinical implications, given the seri-
ousness of consequences towards the mothers and their
infants. One of the challenges to address is the difficulty
in measuring PPH. We have, in our study, implemented a
rigorous method to collect and measure blood loss
through a parallel study that we conducted during 2009–
2010, the period that we used the data for the current
analyses. The underlying mechanism about the increased
risk of PPH is possibly due to coagulation dysfunction
resulting from HBsAg positivity. The possible explanations
for the higher risk of intrahepatic cholestasis included pro-
motion of hepatocellular systemic inflammatory responses
resulting from HBsAg positivity, hypohepatia and accelera-
tion of virus replication, and the interactive effect of the
two that leads to further deterioration of maternal hepatic
function [26].
Findings regarding the HBsAg positivity and subsequent
risk of Caesarean section from previous studies were not
consistent with previous studies [2,5,11]. Our study
showed statistically higher risk of Caesarean section (aOR
1.24, 95% CI 1.05–1.45). Although active and passive
immunization after neonatal birth is effective in preventing
mother-to-child transmission of HBV, the worries of possi-
ble infection of the infant during labour in pregnant

© 2016 John Wiley & Sons Ltd

818 J. Tan et al.

women with HBsAg positivity might force them to choose
selective Caesarean section, given the high prevalence of
selective Caesarean sections in the Chinese population
(46.2%) [27].
This study has several strengths. To the best of our
knowledge, this study is the first multicentre study to
address the association of HBsAg positivity with adverse
maternal outcomes in the population from mainland
China, where HBV infection has been a serious epidemic in
past decades. Second, we have included a relatively large
sample size that allowed us to adjust for the influence of
residual confounders. Indeed, our adjusted analyses have
considered possible gestational comorbidities to ensure ade-
quate assessment of the association. Third, our study
assessed the association between HBsAg positivity and the
risk of PPH and intrahepatic cholestasis, the two adverse
outcomes that received insufficient investigation in previ-
ous studies.
Our study also has a few limitations. As a retrospective
cohort study, we hardly acquired all related and desired
variables in our database. For instance, the variables of
income and educational level about pregnant women could
be considered as covariates but missing just like other pub-
lished studies. Second, we analysed the association of
HBsAg positivity with adverse maternal outcomes only,
but did not examine the impact of other HBV infection
markers (e.g. HBeAg) on the outcomes, mainly because a
large proportion of patients were not tested for HBeAg in
the hospitals. We are planning to address the impact of
HBeAg in our subsequent study.
In summary, our study suggests that pregnant women
with positive HBsAg have a higher risk of GDM, PPH,
intrahepatic cholestasis and Caesarean section, and
potentially higher risk of pre-eclampsia. Increased effort
may be warranted in the routine care, particularly in those
with high risk for adverse maternal outcomes, given the
volume population infected with HBsAg worldwide, and
the increased risk of adverse outcomes important to
those pregnant women. In the meanwhile, further studies
are needed to investigate whether HBsAg positivity repre-
sents a causal risk factor, and to examine the impact of
HBeAg on the adverse maternal outcomes.
ACKNOWLEDGEMENTS
We thank all research personnel and clinical staff from
Chinese Evidence-based Medicine Center, the West China
Women’s and Children’s Hospital, Zigong Maternal and
Child Care Service Centre, Panzhihua Maternal and Child
Care Service Centre, Pengzhou Maternal and Child Care
Service Centre, Chengdu Fifth People’s Hospital, Suining
Central Hospital for project coordination.
DISCLOSURES
Each author declares that he or she has no personal or
financial conflicts of interest to this report.
FUNDING
This study was funded by ‘Science-technology Support Plan’
of Science and Technology Department of Sichuan Province
(project No. 2013SZ0004), Young Investigator Award,
Sichuan University (project No. 2013SCU04A37) and
‘Thousand Youth Talents Plan’ of China (project No.
D1024002) and Sichuan Province.

REFERENCES
1 Shi G, Zhang SX. Meta-analysis on
the positive rate of hepatitis B sur-
face antigen among pregnant
women in China. Chin Prev Med
2013; 14(1): 26–30.
2 Sirilert S, Traisrisilp K, Sirivatanapa
P, Tongsong T. Pregnancy outcomes
among chronic carriers of hepatitis
B virus. Int J Gynaecol Obstet 2014;
126(2): 106–110.
3 To WWK, Cheung W, Mok KM.
Hepatitis B surface antigen carrier
status and its correlation to gesta-
tional hypertension. Aust NZ J Obstet
Gynaecol 2003; 43: 119–122.
4 Lao TT, Sahota DS, Cheng YKY,
Law LW, Leung TY. Maternal hep-
atitis B surface antigen status and
incidence of pre-eclampsia. J Viral
Hepat 2013; 20(5): 343–349.
5 Lobstein S, Faber R, Tillmann HL.
Prevalence of Hepatitis B among
Pregnant Women and Its Impact on
Pregnancy and Newborn Complica-
tions at a Tertiary Hospital in the
Eastern Part of Germany. Digestion
2011; 83: 76–82.
6 Tse KY, Ho LF, Lao T. The impact of
maternal HBsAg carrier status on
pregnancy outcomes: a case–control
study. J Hepatol 2005; 43: 771–
775.
7 Connell LE, Salihu HM, Salemi JL,
August EM, Weldeselasse H, Mbah
AK. Maternal hepatitis B and hepati-
tis C carrier status and perinatal
outcomes. Liver Int 2011; 31(8):
1163–1170.
8 Lao TT, Chan BCP, Leung WC, Ho LF,
Tse KY. Maternal hepatitis B infection
and gestational diabetes mellitus. J
Hepatol 2007; 47(1): 46–50.
9 Lao TT, Tse KY, Chan LY, Tam KF,
Ho LF. HBsAg carrier status and the
association between gestational dia-
betes with increased serum ferritin
concentration in Chinese women. Dia-
betes Care 2003; 26(11): 3011–3016.
10 Wong S, Chan LY, Yu V, Ho L.
Hepatitis B Carrier and perinatal
outcome in singleton pregnancy.
Am J Perinat 1999; 16: 485–488.
11 Safir A, Levy A, Sikuler E, Sheiner
E. Maternal hepatitis B virus or hep-
atitis C virus carrier status as an

© 2016 John Wiley & Sons Ltd


independent risk factor for adverse
perinatal outcome. Liver Int 2010;
30(5): 765–770.
12 Reddick KL, Jhaveri R, Gandhi M,
James AH, Swamy GK. Pregnancy
outcomes associated with viral hep-
atitis. J Viral Hepat 2011; 18:
e394–e398.
13 World Health Organization. WHO
recommendations for the prevention
and treatment of postpartum haemor-
rhage. Geneva: World Health Orga-
nization, 2012.
14 Rook M, Vargas J, Caughey A, Bac-
chetti P, Rosenthal P, Bull L. Fetal out-
comes in pregnancies complicated by
intrahepatic cholestasis of pregnancy
in a Northern California Cohort. PLoS
One 2012; 7(3): e28343.
15 Ishizaka N, Ishizaka Y, Takahashi E
et al. Increased prevalence of carotid
atherosclerosis in hepatitis B virus
carriers. Circulation 2002; 105:
1028–1030.
16 Sung J, Song YM, Choi YH, Ebra-
him S, Smith GD. Hepatitis B virus
seropositivity and the risk of stroke
and myocardial infarction. Stroke
2007; 38: 1436–1441.
© 2016 John Wiley & Sons Ltd
HBsAg positivity and adverse maternal outcomes
17 Irgens HU, Reisaeter L, Irgens LM,
Lie RT. Long term mortality of moth-
ers and fathers after pre-eclampsia:
population based cohort study. BMJ
2001; 323(1213–6): 19.
18 Smith GCS, Pell JP, Walsh D. Preg-
nancy complications and maternal
risk of ischaemic heart disease: a
retrospective cohort study of
129 290 births. Lancet 2001; 357:
2002–2006.
19 Kirwan JP, Mouzon SHD, Lepercq J
et al. TNF-alpha is a predictor of
insulin resistance in human preg-
nancy. Diabetes 2002; 51: 2207–
2213.
20 Winkler G, Cseh K, Baranyi E et al.
Tumour necrosis factor system in
insulin resistance in gestational dia-
betes. Diabetes Res Clin Pract 2002;
56: 93–99.
21 Marinos G, Naoumov NV, Rossol S
et al. Tumour necrosis factor recep-
tors in patients with chronic hepati-
tis B virus infection. Gastroenterology
1995; 108: 1453–1463.
22 Sheron N, Lau J, Daniels H et al.
Increased production of tumour
necrosis factor alpha in chronic
819
hepatitis B virus infection. J Hepatol
1991; 12: 241–245.
23 Tuomainen TP, Korpela H, Nyys-
sonon K et al. Body iron stores are
associated with serum insulin and
blood glucose concentrations: popu-
lation study in 1013 eastern Fin-
nish men. Diabetes Care 1997; 20:
426–428.
24 Fernandez-Real JM, Casamitjana-
Abella R, Ricart-Engel W, Cabrero D,
Arroyo E, Ferna0 ndez-Castan˜er M,
Soler J. Serum ferritin as a component
of the insulin resistance syndrome.
Diabetes Care 1998; 21: 62–68.
25 American Diabetes Association. Ges-
tational Diabetes Mellitus. Diabetes
Care 2003; 26(Suppl1): S103–S105.
26 Medina LJM, Jouregui MRA, Medina
CN, Medina CD. Intrahepatic
cholestasis of pregnancy:review.
Ginecol Obstet Mex 2012; 80(4):
285–294.
27 Lumbiganon P, Laopaiboon M, Gul-
mezoglu AM et al. Method of delivery
and pregnancy outcomes in Asia:the
WHO global survey on maternal and
perinatal health 2007–08. Lancet
2010; 375: 490–499.
Copyright of Journal of Viral Hepatitis is the property of Wiley-Blackwell and its content
may not be copied or emailed to multiple sites or posted to a listserv without the copyright
holder's express written permission. However, users may print, download, or email articles for
individual use.