Full text of DOI:10.1086/315883

1511
CONCISE COMMUNICATION
Prospective Reevaluation of Risk Factors in Mother-to-Child Transmission
of Hepatitis C Virus: High Virus Load, Vaginal Delivery, and Negative
Anti-NS4 Antibody
Manabu Okamoto,1 Ikuo Nagata,1 Jun Murakami,1
Shunsaku Kaji,1 Toshiyuki Iitsuka,1 Tadataka Hoshika,3
Ryu Matsuda,4 Yusaku Tazawa,1 Kazuo Shiraki1
and Shigeo Hino2
Departments of Pediatrics and Virology, Faculty of Medicine,
1
2
3
Tottori University, Yonago, and Tottori Prefectural Central Hospital,
4
and Matsuda Pediatric Clinic, Kurayoshi, Tottori, Japan
Of 21,791 pregnant women screened in Tottori Prefecture, Japan, 127 (0.58%) were positive
for anti–hepatitis C virus (HCV) antibody and 84 (0.39%) were positive for HCV RNA. Of
84 children followed up for at least 6 months, 7 (8%) were infected. All of them were born
to 26 mothers with a high virus load (HVL; у2.5 ϫ 106 RNA copies/mL [27%]), compared
with 0 of 58 children born to non-HVL mothers (P ! .001). Because all the infected children
were vaginally delivered, the infection rate among 16 vaginally delivered children born to the
HVL mothers was as high as 44%. The prevalence of anti-NS4 antibody in the mothers with
an infectious HVL was significantly lower than that in the mothers with a noninfectious HVL
(P p .048). Analysis of our results suggests that maternal HVL, vaginal delivery, and negative
anti-NS4 antibody are significant risk factors for the mother-to-child transmission of HCV.
Screening of donors and heat treatment of blood products es-
transaminase, and only 2 of 6 recovered from viremia within
the 3-year observation period [9, 10].
sentially ceased the transfusion-associated spread of hepatitis C
virus (HCV) [1]. Several groups have observed the mother-to-child
transmission of HCV in ∼10% of HCV-positive mothers [2, 3],
whereas others did not [4]. The true risk of mother-to-child trans-
mission is still unclear, because carrier mothers were screened by
reverse transcriptase–polymerase chain reaction (RT-PCR) anal-
ysis, with varying sensitivities [5–7]. Nevertheless, a maternal high
virus load (HVL) has been suggested as a major risk factor [2, 3].
Maternal coinfection with human immunodeficiency virus (HIV)
complicated the matter further, because, in this population, HCV
transmission occurs more often [3, 8].
The clinical significance of mother-to-child transmission of
HCV has not been confirmed. Although some reports have
indicated that most infected children developed chronic liver
dysfunction [7], children remained asymptomatic in other stud-
ies [3]. In our experience, 4 of 6 patients, including a patient
with 2-year persistence of HCV, possessed chronic elevation of
The relative importance of mother-to-child transmission in de-
veloped countries is now significantly higher than would be ex-
pected because of the elimination of the blood-borne spread of
HCV. We decided to look for a plausible measure to prevent
mother-to-child transmission, so we prospectively reevaluated
risk factors for the infection of children born to HCV-positive
mothers.
Materials and Methods
Subjects. From June 1992 to December 1998, 21,791 pregnant
women were screened for anti-HCV antibody in Tottori Prefecture,
Japan. None of the mothers had risk factors for HIV infection,
and the officially reported carrier rate in the prefecture is !10Ϫ5
.
Seropositive (Abϩ) samples were subjected to RT-PCR analysis.
The RT-PCR positive samples (RNAϩ) were titrated by a branched
DNA assay.
Serum samples of the children born to Abϩ mothers were tested
for anti-HCV antibody, HCV RNA, and liver function approxi-
mately every 3 months during the first year and biannually there-
after. Minimum follow-up period was 6 months. Breast-feeding was
not discouraged.
Received 22 May 2000; revised 24 July 2000; electronically published 9
October 2000.
Informed consent from each mother was obtained before and after the
screening, and human experimentation guidelines of each participating in-
stitution were followed in the conduct of this research.
Financial support: Research on Children and Families of the Ministry of
Welfare of Japan.
Reprints or correspondence: Dr. Shigeo Hino, Dept. of Virology, Faculty
of Medicine, Tottori University, 86 Nishi, Yonago 683-8503, Japan (hino
@grape.med.tottori-u.ac.jp).
Anti-HCV antibody. A second generation anti-HCV antibody
assay kit based on passive hemagglutination (Dainabot, Tokyo, Ja-
pan) was used for screening. Positive samples were confirmed by a
self-prepared Western blot that used a recombinant core antigen [9].
A third-generation enzyme immunoassay kit was used for epitope-
specific antibodies (Sympep HCV-EIA II; Kyokuto Med, Tokyo,
Japan).
The Journal of Infectious Diseases 2000;182:1511–4
᭧ 2000 by the Infectious Diseases Society of America. All rights reserved.
0022-1899/2000/18205-0029$02.00
Nested RT-PCR analysis. Methods for the nested RT-PCR

1512
Okamoto et al.
JID 2000;182 (November)
analysis targeted on the 50 untranslated region were described else-
where [9]. In brief, RNA was extracted from 50 mL of the serum
sample mixed with 20 mg of glycogen (Boehringer-Mannheim,
Mannheim, Germany) by the guanidinium method. After reverse
transcription with 20 U of Moloney murine leukemia virus RT
(Superscript; Gibco BRL, Gaithersburg, MD) at 45ЊC for 60 min,
the whole product was used for the first PCR for 35 cycles at 94ЊC
for 1 min (3 min for the first cycle), 55ЊC for 1 min, and 72ЊC for
2 min (7 min for the last cycle) for each cycle. A 2-mL aliquot was
used for the second PCR, with the inner primer pair under the
same conditions. Reproducibility of each positive signal was con-
firmed by repeating the test on another day.
The minimum detection level of the RT-PCR was 10 HCV RNA
copies per reaction, or 200 copies/mL of serum calibrated by in vitro
synthesized HCV RNA from a deletion plasmid [9]. Two tubes con-
taining 20 copies per reaction of the deletion plasmid DNA and a
tube without a template were added to each run of the RT-PCR as
the positive and negative controls, respectively. Data were discarded
if both of the positive control tubes in a run were negative, and none
of the negative controls became positive in any run.
Quantitative HCV-RNA assay and genotype. Serum HCV RNA
was quantified by the branched DNA (bDNA) kit (Quantiplex ver-
sion 1.0; Chiron, Emeryville, CA); the minimum detection level of
HCV RNA was 0.5 ϫ 106 copies/mL. The HCV genotype was de-
termined by a nested RT-PCR kit (Sumai Test; Tokusyu Men’eki,
Tokyo, Japan).
Potential maternal and perinatal risk factors. The following
parameters were analyzed as potential risk factors for mother-to-
child transmission: maternal HCV RNA level, mode of delivery,
history of blood transfusion, history of hepatitis, HCV genotype,
gestation period, intrapartum period, body weight at birth, pla-
cental weight, bleeding volume during delivery, feeding method,
and epitope-specific maternal antibodies.
were found to be positive by RT-PCR analysis at titers р103
copies/mL. In contrast, only 1 of 26 cord blood samples of
uninfected children was found to be positive by RT-PCR anal-
ysis. The genotype of HCV in each mother-child pair was iden-
tical: genotype 1b in 4 pairs, including a pair of siblings, ge-
notype 2a in one, and genotype 2b in a pair of siblings.
Maternal HCV titer and transmission. Of 77 Abϩ mothers
without infected children (i.e., mothers who were noninfec-
tious), 53 (69%) were RNAϩ, and 41 (53%) were bDNAϩ. Geo-
metric average titer of the bDNAϩ mothers was 2.7 ϫ 106 cop-
ies/mL (95% CI, 1.9 ϫ 106 to 3.8 ϫ 106 copies/mL). In contrast,
all the mothers with infected children (i.e., mothers who were
infectious) were RNAϩ, with a geometric average titer of
8.0 ϫ 106 copies/mL (95% CI, 3.8 ϫ 106 to 16.7 ϫ 106 copies/
mL; P p .016). It must be noted that 43 (51%) of 84 Abϩ
mothers were bDNAϪ and therefore were not included in the
comparison. By use of the average titer of noninfectious
bDNAϩ mothers, 26 mothers with a virus load у2.5 ϫ 106 cop-
ies/mL were classified as HVL mothers. They consisted of 31%,
44%, and 54% of the Abϩ, RNAϩ, and bDNAϩ mothers, re-
spectively. All the mothers who were infectious were classified
with the HVL mothers. Relative risk of mother-to-child trans-
mission for the HVL mothers versus the Abϩ mothers was 3.5
(95% CI, 1.1–10.9; P p .041), based on mothers, and 3.2 (CI,
1.3–8.4; P p .020), based on children.
Vaginal delivery as a risk factor. Of 84 deliveries, 28 babies
were delivered by cesarean section (CS): 9 for precedent CS, 2
for breech presentation, 4 for cephalopelvic disproportion, 2
for twin pregnancy, 1 for gestosis, and 10 for unknown reasons
(table 1). None were directly related to the HCV infection. All
7 infected children were delivered vaginally, and none of 18 CS
children born to RNAϩ mothers were infected. Among the
children born to mothers with HVL, the vaginally delivered
children had a significantly higher risk of infection than did
the CS children (P p .023; table 1). Geometric average of the
HCV RNA level in the 5 infectious mothers, 7.0 ϫ 106 copies/
mL, was significantly higher than that of the 31 noninfectious
Statistical analysis. Unpaired Student’s t test with Welch’s cor-
rection was applied for numeric data after logarithmic conversion,
and Fisher’s exact test was used for dichotomized data.
Results
Sampling of mother-child pairs. Of the 21,791 pregnant
women screened for anti-HCV antibody, 127 (0.58%) were Abϩ
and 84 (0.39%) were RNAϩ; the latter represented 66% of Abϩ
mothers. Of 43 dropouts, 41 were not referred to pediatricians
who participated in this study. Those enrolled in this study
consisted of 73 Abϩ and 50 RNAϩ mothers, and 84 and 59
children born to these mothers, respectively. These mothers in-
cluded 7 with 2 successive children, 1 with 3 successive children,
and 1 with a pair of twins.
Children infected by mother-to-child transmission. Seven
children born to 5 mothers were found to be positive for HCV
RNA by RT-PCR analysis. The other children never tested
positive within the follow-up period. Each possessed high titers
of HCV RNA within 3 months of the age at the geometric
average titer of 19.5 ϫ 106 copies/mL (95% confidence interval
[CI], 9.5 ϫ 106 to 42.5 ϫ 106 copies/mL). All 7 cord blood sam-
ples were negative by the bDNA assay (bDNAϪ), but 3 of 6
RNAϩ mothers, which was
6 copies/mL (
; table
1.5 ϫ 10
P ! .001
2). Interestingly, an HVL mother had 2 successive children de-
livered vaginally who were infected and then bore a third CS
child who was free from infection.
Effect of other maternal and perinatal factors on HCV trans-
mission. We surveyed the possible role of maternal parameters
in the mother-to-child transmission of HCV, including preceding
history of blood transfusion, history of clinical hepatitis, and
HCV genotype. None of these maternal parameters was signif-
icantly different between the mothers who were infectious and
those who were not.
None of the perinatal parameters tested were significant be-
tween the infected and uninfected children, including gestation
period, body weight at birth, placental weight, and bleeding
volume during labor. The average intrapartum period was
longer for the infected children, but the difference disappeared

JID 2000;182 (November)
Mother-to-Child Transmission of HCV
1513
study, у2.5 ϫ 106 copies/mL, might be lower if the sample size
were larger or for mothers coinfected with HIV [3, 11].
Table 1. Mother-to-child transmission of hepatitis C virus (HCV)
and the route of delivery.
Status, delivery
No. enrolleda
No. positive
Positive, %
Pb
All the infected children were vaginally delivered, and the
risk of infection in vaginally delivered children born to HVL
mothers was 44%. This high risk might be explained by intra-
partum microtransfusion from mother to fetus, because it is
well known that a prearranged CS can minimize the micro-
transfusion [12, 13]. These results, as well as the minimum
amount of HCV RNA in the cord blood and the rapid increase
of virus load within the first 1–3 months of life, are strong
evidences for perinatal transmission. A previous study failed
to find HCV infections in siblings of the index patients [14].
However, in this study, 2 of 3 HVL mothers who vaginally
delivered 11 child each had 2 infected children, which suggests
the presence of additional maternal risk factors.
Among HVL mothers, the prevalence of anti-NS4 in infec-
tious mothers was marginally lower than that in noninfectious
mothers (P p .048), which is consistent with findings in a pre-
vious report [15]. The finding may be related to a more recent
acquisition of HCV infection among the mothers who were
infectious, which explains the higher levels of viremia.
As maternal HVL and vaginal delivery are the major risk
factors for the mother-to-child transmission of HCV, HVL at
the time of delivery would be the critical factor for transmission.
Approximately 40% of vaginally delivered children born to
HVL mothers were infected with HCV, and more than half the
infected children developed the chronic liver disease. Near-term
titration of HCV would be recommended for Abϩ pregnant
women and CS for those with HVL. The natural history of
infected children is still uncertain. However, if it is strongly
associated with chronic active hepatitis, liver cirrhosis, or even
hepatocellular carcinoma, more intensive measures might be
necessary. For example, interferon therapy might be considered
to lower the HCV RNA titer at the time of delivery, if the
treatment is proved safe and tolerable in the third trimester. A
further study on a larger scale is required to substantiate this
suggestion.
Abϩ mothers
Mothers
Vaginal
Cesarean
Children
Vaginal
Cesarean
RNAϩ mothers
Mothers
Vaginal
Cesarean
Children
Vaginal
Cesarean
HVL mothers
Mothers
Vaginal
Cesarean
Children
Vaginal
45
23
5
0
11
0
.159
.045
50
28
7
0
14
0
36
14
5
0
14
0
.304
.089
41
18
7
0
17
0
13
8
5
0
38
0
.111
.023
16
10
7
0
44
0
Cesarean
NOTE. Abϩ, antibody positive, as assessed by second-generation assay and
confirmed by Western blot test; HVL, high virus load у2.5 ϫ 106 copies/mL, as
determined by branched DNA assay; RNAϩ, HCV RNA positive by reverse
transcriptase–polymerase chain reaction (у200 copies/mL). All the mothers who
were infectious had HVL.
a
Six deliveries without record of the delivery method were deleted.
Fisher’s exact test.
b
when compared with the vaginally delivered children. The sam-
ple size was too small to test the effect of breast-feeding.
Epitope-specific antibodies and mother-to-child transmission.
All Abϩ mothers were positive for anti-core antibody. Anti-
NS3 and -NS5 antibodies were positive in ∼70% and ∼40%,
respectively, of mothers who were infectious and mothers who
were noninfectious. Anti-NS4 antibody was positive in 12 (44%)
of 27 RNAϩ but non-HVL mothers and in 13 of 21 HVL
mothers (62%; not significant). The positive rate for anti-NS4
antibody among the infectious mothers, 1 (20%) of 5, was sig-
nificantly lower than that of the noninfectious HVL mothers,
12 (75%) of 16 (P p .048). However, the significance was lost
when compared with the mothers who delivered vaginally, al-
though a trend remained (5 [71%] of 7; P p .242).
Table 2. Maternal hepatitis C virus (HCV) RNA titers of vaginally
delivered children born to RNAϩ mothers.
HCV RNA, ϫ106 copies/mLa
No.
RNAϩ subjects
tested
Geometric average
95% CI
Pb
Mothersc
Infectious
Noninfectious
Children
Infected
Uninfected
Discussion
5
31
7.0
1.5
2.4–20.0
0.9–2.3
!.001
The most well documented risk factor for mother-to-child
transmission of HCV is maternal HVL [2, 9, 11]. The risks of
mother-to-child transmission were 8%, 12%, and 27% in children
born to Abϩ, RNAϩ, and HVL mothers, respectively. Analysis
of our data suggests that the mother-to-child transmission of
HCV among non-HIV mothers is, in fact, more frequent than
generally thought, if the mothers were carefully selected. Im-
portantly, no mothers who were infectious were found in the
non-HVL group. However, the threshold of HVL used in this
7
34
8.0
1.4
3.8–16.7
0.9–2.2
!.001
NOTE. CI, confidence interval; RNAϩ, HCV RNA positive by reverse tran-
scriptase–polymerase chain reaction (у200 copies/mL).
a
HCV RNA titer determined by branched DNA assay. For the negative sam-
ple in branched DNA assay, 0.2 ϫ 106 copies/mL was assigned.
b
Unpaired Student’s
t test with Welch’s correction after logarithmic
conversion.
c
For computation on 10 mothers with 11 child, data obtained from the first-
order children were used.

1514
Okamoto et al.
JID 2000;182 (November)
7. Paccagnini S, Principi N, Massironi E, et al. Perinatal transmission and man-
ifestation of hepatitis C virus infection in a high risk population. Pediatr
Infect Dis J 1995;14:195–9.
Acknowledgments
We thank Koichi Irie (Tottori Prefecture Health Promoting Council),
Akio Nagata (Tottori Prefecture Health Promoting Council and Japan
Association of Obstetricians and Gynecologists), Naoki Terakawa (De-
partment of Gynecology and Obstetrics, Faculty of Medicine, Tottori
University) and Yoshito Hosoda (Department of Pediatrics, Faculty
of Medicine, Tottori University) for collecting samples and clinical
data. We also thank Toshio Kamahora, Hironori Miyata, and Sachiko
Ishikura (Department of Virology, Faculty of Medicine, Tottori Uni-
versity) for technical advice and assistance.
8. Weintrub PS, Veereman-Wauters G, Cowan MJ, Thaler MM. Hepatitis C
virus infection in infants whose mothers took street drugs intravenously.
J Pediatr 1991;119:869–74.
9. Okamoto M, Nagata I, Murakami J, Hino S, Shiraki K. Shift in the buoyant
density of HCV particles in infants infected by mother-to-infant trans-
mission. Pediatr Int 1999;41:369–73.
10. Murakami J, Okamoto M, Miyata H, Nagata I, Shiraki K, Hino S. Evolution
in the hypervariable region of hepatitis C virus in infants after vertical
transmission. Pediatr Res 2000;48:450–6.
11. Thomas SL, Newell ML, Peckham CS, Ades AE, Hall AJ. A review of hepatitis
C virus (HCV) vertical transmission: risks of transmission to infants born
to mothers with and without HCV viraemia or human immunodeficiency
virus infection. Int J Epidemiol 1998;27:108–17.
References
1. Japanese Red Cross Non-A Non-B Hepatitis Research Group. Effect of
screening for hepatitis C virus antibody and hepatitis B virus core antibody
on incidence of posttransfusion hepatitis. Lancet 1991;338:1040–1.
2. Ohto H, Terazawa S, Sasaki N, et al. Transmission of hepatitis C virus from
mothers to infants. N Engl J Med 1994;330:744–50.
3. Zanetti AR, Tanzi E, Paccagnini S, et al. Mother-to-infant transmission of
hepatitis C virus. Lombardy Study Group on Vertical HCV Transmission.
Lancet 1995;345:289–91.
4. Lam JP, McOmish F, Burns SM, Yap PL, Mok JY, Simmonds P. Infrequent
vertical transmission of hepatitis C virus. J Infect Dis 1993;167:572–6.
5. Sabatino G, Ramenghi LA, di Marzio M, Pizzigallo E. Vertical transmission
of hepatitis C virus: an epidemiological study on 2980 pregnant women
in Italy. Eur J Epidemiol 1996;12:443–7.
12. Lin HH, Kao JH, Hsu HY, Mizokami M, Hirano K, Chen DS. Least
microtransfusion from mother to fetus in elective cesarean delivery. Obstet
Gynecol 1996;87:244–8.
13. Kaneda T, Shiraki K, Hirano K, Nagata I. Detection of maternofetal trans-
fusion by placental alkaline phosphatase levels investigations on materno-
fetal transfusion by measuring placental alkaline phosphatase. J Pediatr
1997;130:730–5.
14. Moriya T, Sasaki F, Mizui M, et al. Transmission of hepatitis C virus from
mothers to infants: its frequency and risk factors revisited. Biomed Phar-
macother 1995;49:59–64.
15. Mazza C, Ravaggi A, Rodella A, et al. Prospective study of mother-to-infant
transmission of hepatitis C virus (HCV) infection. J Med Virol 1998;54:
12–19.
6. Wejstal R, Widell A, Mansson AS, Hermodsson S, Norkrans G. Mother-to-
infant transmission of hepatitis C virus. Ann Intern Med 1992;117:887–90.