Chang M, Shau W, Chen C, Wu T, Kong M, Liang D, Hsu H, Chen H, Hsu H, Chen D, for the Taiwan Childhood Hepatoma Study Group . Hepatitis B Vaccination and Hepatocellular Carcinoma Rates in Boys and Girls. JAMA. 2000;284(23):3040-3042. doi:10.1001/jama.284.23.3040
Author Affiliations: Department of Pediatrics (Drs Chang, H. Y. Hsu, and H. L. Chen) and Hepatitis Research Center (Dr D. S. Chen), National Taiwan University Hospital, Taipei; Graduate Institute of Clinical Medicine, College of Medicine (Dr Shau) and Graduate Institute of Epidemiology, College of Public Health (Dr C. J. Chen), National Taiwan University, Taipei, Department of Pediatrics, Veteran General Hospital, Taipei (Dr Wu); Chang-Gung Children's Hospital, Lin-Kou (Mr Kong); Department of Pediatrics, Mackay Memorial Hospital, Taipei (Dr Liang); and Center for Disease Control, Department of Health, Executive Yuan (Dr H. M. Hsu), Taipei, Taiwan, Republic of China.
Context Hepatocellular carcinoma (HCC) has a male predominance and is closely
related to hepatitis B virus (HBV) infection. Hepatitis B virus vaccination
was launched in 1984 in Taiwan for neonates of mothers carrying hepatitis
B e antigen, resulting in a decreased incidence of HCC in children. The effect
on boys vs girls is not known.
Objective To evaluate the association between a HBV vaccination program with incidence
of childhood HCC by sex.
Design and Setting Analysis of data collected from Taiwan's National Cancer Registry System
and the Taiwan Childhood Hepatoma Study Group between 1981 and 1996.
Participants Children aged 6 to 14 years who were diagnosed as having HCC (201 boys
and 70 girls).
Main Outcome Measure Incidence of HCC in boys and girls before and after implementation of
the vaccination program.
Results The boy-girl incidence ratio decreased steadily from 4.5 in 1981-1984
(before the program's introduction) to 1.9 in 1990-1996 (6-12 years after
the vaccination program was launched). The incidence of HCC in boys born after
1984 was significantly reduced in comparison with those born before 1978 (relative
risk [RR], 0.72; P = .002). No significant decrease
in HCC incidence was observed in girls born in the same periods (RR, 0.77; P = .20). The incidence of HCC in boys remained stable
with increasing age, while an increase of HCC incidence with age in girls
was observed. These age and sex effects remained the same regardless of birth
before or after the vaccination program.
Conclusion Our results suggest that boys may benefit more from HBV vaccination
than girls in the prevention of HCC.
Hepatocellular carcinoma (HCC) occurs mainly in adults 40 to 60 years
of age.1 However, in areas hyperendemic for
hepatitis B virus (HBV) infection, HCC may develop in children.2
We have found a nearly 100% hepatitis B surface antigen seropositivity rate
in Taiwanese children with HCC, indicating an intimate relationship between
HBV and childhood HCC.2 In 1984, a hepatitis
B vaccination program was launched and has effectively reduced the prevalence
of HBV infection, chronic HBV infection rate,3
and incidence of HCC in children in Taiwan.4
A male predominance of HCC has long been observed.5,6
The mechanism is unknown, but a tumorigenic effect of androgens has been suggested.7 We found a male-female ratio of 3-4:1 in children
with HCC,8 which is similar to that in adults.
Since the influence of hormones in children is much less than that in adults,
investigating HCC in children may facilitate understanding of the mechanism
of HCC. We therefore studied children with HCC by sex before and after the
implementation of the HBV vaccination program.
A vaccination program was implemented in Taiwan in July 1984.9 Hepatitis B immunoglobulin was given to neonates of
highly infectious mothers carrying hepatitis B e antigen. All infants received
3 or 4 doses of HBV vaccine.
According to our previous observation, HCC in children was diagnosed
mainly in those older than 6 years,2 while
hepatoblastoma was diagnosed in younger children.10
In this study, we included children 6 to 14 years of age with liver cancer
to preclude the inclusion of hepatoblastoma.
Two independent childhood hepatoma registry systems were used in this
study to ensure accuracy. Data from the following 2 systems, including the
name, identification number, birth date, sex, address, etc, were checked and
merged, and any repetition was deleted. The case information was confirmed
by the reporting hospitals. The capture-recapture method was used to estimate
the total number of cases of childhood HCC (Epi Info, version 6.04; Centers
for Disease Control and Prevention and the World Health Organization). The
number of cases identified by systems 1 and 2 was estimated to be 86% (95%
confidence interval [CI], 80%-92%) of the actual total number of children
Cases of hepatoma diagnosed between July 1981 and June 1996 were analyzed
from the data bank of the National Cancer Registry System at the National
Department of Health. This registry was established in 1979. Cases are reported
by the department of medical records in each of the 167 hospitals with more
than 50 beds in Taiwan.
To ensure the accuracy of the data from the National Cancer Registry,
we formed a multicenter Childhood Hepatoma Study Group to register children
with hepatoma during the same study period. Pediatric gastroenterologists
or oncologists from 17 major hospitals, including all 12 tertiary referral
centers in Taiwan, participated.
The study population was stratified both by age at diagnosis and the
year of birth. Children with HCC who were older than 6 years on July 1, 1984,
when the HBV vaccination program was launched, were born before July 1978.
Children born before 1978 and after 1984 were the respective cohorts without
and with the effect of HBV vaccination. Children born between 1978 and 1984
were born during the transition to full implementation of the HBV vaccination
program. They might have received HBV vaccination beyond infancy.
Age-specific and birth-year–specific incidences of HCC were calculated
for boys and girls. Relative incidences of HCC among children divided into
groups by age, birth cohort, and sex were analyzed using Poisson regression.11 The modification of age effect on HCC incidence by
sex was statistically tested by cross production of age and sex variables
(the interaction term) and expressed in separate models when the age trends
were significantly different between female and male.
A consistent predominance of HCC in boys was found throughout the observation
period. The incidence of childhood HCC declined gradually in boys during 1981-1996,
while the incidence in girls remained stable (Table 1). Although the trend of the predominance in boys remained,
the boy/girl ratio of the incidence of HCC declined gradually with time from
4.5 for years of diagnosis 1981-1984 to 1.9 for years of diagnosis 1990-1996.
The relative risk (RR) of HCC in boys born between 1978 and 1984 declined
significantly in comparison with those born before 1978 (RR, 0.83; P = .02); the trend of decrease in the RR of HCC was even more evident
in those born after 1984 when compared with those born before 1978 (RR, 0.72; P = .002) (Table 2).
However, the trend of decline in girls was not significant for the birth cohort
born between 1978 and 1983 vs those born before 1978 (RR, 1.02; P = .90) or for those born after 1984 vs those born before 1978 (RR,
0.77; P = .20).
The age trend of HCC risk was significantly modified by sex. The risk
of HCC in boys remained constant from age 6 to 14 years, while the risk of
HCC in girls, though lower than in boys, increased significantly with age.
In separate models, there was no significant age trend for boys (RR, 0.97;
95% CI, 0.92-1.03; P = .33), while the incidence
of HCC in girls increased significantly by 1.15 times for each year increment
of age (RR, 1.15; 95% CI, 1.04-1.28; P = .007). The
age effect in boys and girls was the same before and after vaccination (Figure 1).
In the present study, we observed a predominance of HCC in boys both
before and after the HBV vaccination program. This predominance cannot be
explained by the effect of sex hormones, as in adults. Tumor suppression gene
regulation, the metabolism of carcinogens, or genetic alterations have been
proposed to differ between men and women and need further study.5
This predominance decreased after the vaccination program because the
incidence of HCC decreased significantly in boys but not in girls. Why the
vaccination program seems to have had more of an effect on boys remains unclear.
The low incidence of HCC in girls may render the statistical comparison of
the incidences difficult. It is possible that HCC in girls is less intimately
related to HBV infection than in boys, but seems unlikely given our previous
observations2 and evidence of HBV infection
in girls with HCC born after implementation of the program (unpublished data
by authors). The possibility that intrauterine infection with HBV, which would
not be affected by vaccination, occurs more frequently in female infants also
is unlikely, as there was no female predominance in infants who were seropositive
for the hepatitis B surface antigen at birth.12
Additionally, there was no difference in the vaccination coverage rate between
male and female infants in Taiwan. (National Taiwan University Hospital's
coverage rate is 100% for all mature neonates. The number of delivery of neonates
is approximately 3000 per year. The national coverage rate for neonates was
between 84% and 94% for 1986 to 1994 [M. H. Chang, unpublished data]). It
also seems unlikely that case finding for such a serious disease would differ
between boys and girls or change over time.
Seroepidemiologic studies in Taipei conducted in both 1984 and 1994
in children showed no or a slight predominance in boys in the incidence of
HBV infection.3,13 In contrast,
the remarkable predominance of HCC in boys suggests that factor(s) in addition
to chronic HBV infection may contribute to hepatocarcinogenesis in males,
particularly the early occurrence in prepubertal males.