Krause TG, Hviid A, Koch A, Friborg J, Hjuler T, Wohlfahrt J, Olsen OR, Kristensen B, Melbye M. BCG Vaccination and Risk of Atopy. JAMA. 2003;289(8):1012-1015. doi:10.1001/jama.289.8.1012
Author Affiliations: Department of Epidemiology Research, Danish Epidemiology Science Centre, Statens Serum Institut, Copenhagen, Denmark (Drs Krause, Koch, Friborg, Hjuler, and Melbye, and Mssrs Hviid and Wohlfahrt); Sisimiut Hospital, Sisimiut, Greenland (Dr Olsen); and Pharmacia, Copenhagen, Denmark (Mr Kristensen).
Context It has been suggested that BCG vaccination may protect against development
of allergic diseases, particularly when given just after birth. BCG vaccination
was given routinely to all infants in Greenland until 1990, when it was withdrawn
from the vaccination program. Whether this resulted in an increased prevalence
of atopy in children born after the stop of BCG vaccination is unknown.
Objective To determine whether BCG vaccination and age at BCG vaccination are
associated with development of atopy.
Design, Setting, and Participants Cross-sectional study among schoolchildren aged 8 to 16 years in 4 towns
on the northwest coast of Greenland. Participants had a blood sample drawn
and information on BCG vaccination was obtained during 2 periods, November
1998 and November 2001. A total of 1686 children (79% of available children)
participated, 1575 of whom had complete information on vaccination status.
Atopy was defined as a positive test result in an assay that tests for IgE
specific against the most common inhalant allergens in serum.
Main Outcome Measures Odds ratio (OR) of atopy in BCG-vaccinated compared with unvaccinated
children and OR according to age at vaccination.
Results The risk of atopy was the same in BCG-vaccinated compared with unvaccinated
children after adjustment for confounders (OR, 1.03; 95% confidence interval,
0.72-1.48). The risk of atopy in BCG-vaccinated children was not associated
with age at vaccination (P = .17).
Conclusions BCG vaccination administered to infants is not associated with reduced
risk of development of atopy.
It has been proposed that childhood exposure to certain infections and
vaccinations that induce TH1-type immune responses may protect
against atopic diseases, which are characterized by a TH2-type
cytokine expression.1,2 In particular,
BCG vaccination has been suggested to have such an effect based on its ability
to elicit a strong TH1-type immune response in humans.3 In support of this theory, routine BCG vaccination
has been stopped in most western countries where an increase in allergy has
been observed.4 Results from the few epidemiologic
studies that have addressed the hypothesized association between BCG vaccination
and allergic diseases are conflicting5- 9 and
possibly distorted by selection bias, as they have been carried out in countries
where BCG vaccination is either a part of the routine vaccination program
or given only to high-risk subjects.
In Greenland, BCG vaccination was given routinely to all infants until
1990, when it was abruptedly stopped.10 This
provided a unique opportunity to compare the risk of atopy in unselected groups
of children who were either all vaccinated with BCG or all unvaccinated. Furthermore,
we obtained information on exact date of BCG vaccination, enabling estimation
of the risk of atopy according to age at vaccination.
The study was conducted in the towns of Sisimiut, Ilulissat, Aasiaat,
and Maniitsoq, located on the northwest coast of Greenland. These towns are
similar in the number of inhabitants (3500-5200), climate, infrastructure,
and living conditions. In November 1998, all children aged 8 through 16 years
in Sisimiut alone were invited to participate. In November 2001, all children
aged 9 through 13 years in all 4 towns were invited. The sample was limited
to children born in Greenland. The children answered a self-administered questionnaire
together with their parents regarding sociodemographic variables and had a
venous blood sample drawn. Parental place of birth was used as an indicator
of ethnicity. Thus, if both the child and his or her parents were born in
Greenland, 96% of grandparents were also born in Greenland, making it highly
likely that the child was of Inuit origin.11 Children
living in Sisimiut who were examined twice were only included at the time
of their first blood sample. Ethics committee approval was obtained from the
Commission for Scientific Research in Greenland. Written informed consent
was obtained from all participating children and their parents.
Venous blood samples were analyzed for total IgE with the Unicap total
IgE test (Pharmacia, Copenhagen, Denmark) and for specific IgE with the Phadiatop
test (Pharmacia), which is a qualitative (yes/no) assay testing for the 8
most common inhalant allergens (grass, birch, mugwort, dog, cat, horse, Cladosporium herbarum, house dust mite) in 1 pool. Children
with a positive specific IgE test result were considered atopic.
Information on child's place of birth, parents' place of birth, mother's
age at first birth, and birth order was obtained for all children from the
Danish Civil Registration System. All inhabitants of Greenland are registered
in the system with a unique identification number. Information on birth weight
was obtained from birth records.
Until 1990, infants in Greenland received the BCG vaccination once intracutaneously
in the deltoid region; all infants received the same dose within a few days
after birth. The BCG vaccine administered from 1980 to 1990 (Copenhagen strain
No. 1331) has been widely used in many countries around the world. Due to
a reduction in incidence of tuberculous meningitis and pulmonary tuberculosis,
BCG vaccination was withdrawn from the vaccination program at the beginning
of 199010 and stopped completely at all hospitals
in Greenland during the following year. Due to local outbreaks of tuberculosis,
BCG vaccination was introduced again in 1997 to newborns in the routine vaccination
program. As an exception, Maniitsoq children who were born from 1990 to 1997
but not vaccinated were all offered vaccination in 1997.
In most towns in Greenland, exact date of BCG vaccination is registered
both in vaccination protocols kept at the health center where the child is
born and in hospital records kept at the health center in the town where the
child is living. Information on BCG vaccination was primarily obtained from
vaccination protocols (77%) available in Sisimiut, Aasiaat, and Maniitsoq.
The remaining information was obtained from hospital records (23%).
The odds ratio (OR) for the prevalence of atopy in vaccinated compared
with unvaccinated children and the OR for prevalence according to age at vaccination
were estimated using logistic regression. To identify variables confounding
the association between BCG and atopy, we used the change-in-estimate method,12 including the background variables presented in Table 1 with the categories as presented
if a change in the BCG OR of more than 10% was present. Adjustment for age
was made using quadratic splines with knots at ages 8, 12, and 16 years. Analyses
were carried out using SAS v8.2 (SAS Institute Inc, Cary, NC); P = .05 was used to determine statistical significance.
In November 1998, 820 children aged 8 to 16 years (85% of available
children) were enrolled and had a blood sample drawn, of whom 789 were born
in Greenland. In November 2001, 1139 children aged 8 to 13 years (74% of available
children) were enrolled and had a blood sample drawn, of whom 1102 were born
in Greenland. A total of 205 children living in Sisimiut had been examined
twice and were only included in the study with their first blood sample, resulting
in a study group of 1686 children aged 8 to 16 years. Information on BCG vaccination
was obtained for 1575 children (93%). Of these 1065 (68%) had been vaccinated,
but 3 children had missing information on exact dates of BCG vaccination.
A total of 81 children were vaccinated when they were older than 1 year. Of
these, 62 (76.5%) lived in Maniitsoq and had been vaccinated as part of the
vaccination campaign performed in that town in 1997. The BCG vaccination coverage
before age 1 year was 94% among children born before January 1, 1990, 53%
among children born in 1990, and only 2% among children born after January
Children receiving the BCG vaccine were older than unvaccinated children
(Table 1). Furthermore, a higher
proportion of the BCG-vaccinated children were examined in 1998 compared with
2001. The number of persons per room in the home differed between BCG-vaccinated
and unvaccinated children in univariate analysis but was similar after adjusting
for age (P = .46). Otherwise, the distribution of
background variables was similar in BCG-vaccinated and unvaccinated children.
The adjusted OR for prevalence of atopy according to BCG vaccination
and age at BCG vaccination was examined in 2 logistic regression analyses
(Table 2). In the first, only
age was a confounder, whereas in the second, age, year at examination, and
birth weight were confounders. The adjusted risk of atopy was the same in
BCG-vaccinated compared with unvaccinated children (OR, 1.03; 95% confidence
interval [CI], 0.72-1.48). There was no effect modification by geographical
site (P = .23) or by source of BCG vaccination information
(vaccination protocols vs hospital records) (P =
.39). Furthermore, if children with repeated measurements of atopy (n = 205)
and children living in Maniitsoq (n = 176) were excluded, the overall risk
of atopy remained unchanged (OR, 0.94; 95% CI, 0.61-1.47 and OR, 1.00; 95%
CI, 0.68-1.47, respectively). There was no effect of age at BCG vaccination
on risk of atopy (P = .17 for heterogeneity). The
data could not be described by a trend.
The OR for atopy in children with missing information on BCG vaccination
compared with BCG-vaccinated children was 0.89 (95% CI, 0.51-1.77). In a subanalysis
we classified children with missing information on BCG vaccination as either
BCG-vaccinated or unvaccinated according to year of birth (children born before
1990 and the first 6 months of 1990 as vaccinated and children born later
as unvaccinated). The overall estimate was unchanged (OR, 1.02; 95% CI, 0.70-1.48).
Total IgE levels did not differ between BCG-vaccinated and unvaccinated
children (mean [SD] total IgE level among BCG-vaccinated and unvaccinated
children: 233  and 217  kU/L, respectively, P = .65).
Based on a large cross-sectional study, we found BCG-vaccinated children
to have the same risk of atopy as unvaccinated children. Furthermore, no significant
differences in the risk of atopy were found according to age at BCG vaccination.
Overall, these results do not support the hypothesis that BCG vaccination
has a protective effect on the development of atopy.
Five previous studies have addressed the hypothesis that BCG vaccination
protects against allergic diseases by comparing the prevalence of atopy or
allergic diseases in BCG-vaccinated and unvaccinated children. Strannegård
et al5 found no protective effect of BCG vaccination
on allergic diseases in Swedish 4- to 9-year-old children, while a nonsignificant
effect was observed among foreigners. Similarly, Alm et al6 found
no association with BCG vaccination and atopy or allergic diseases in 2-to
7-year-old children with atopic heredity. However, a potential protective
effect of BCG vaccination may have been masked by the strong genetic predisposition
to atopy in the children. Furthermore, in Sweden BCG vaccination is only given
to children at high risk of tuberculosis, resulting in a BCG vaccination rate
of 4%,13 a likely source of selection bias.
In a study conducted in Africa, Aaby et al7 reported
that BCG-vaccinated children had a lower prevalence of atopy compared with
unvaccinated children, particularly when the vaccine had been administered
in the first week after birth. Grüber et al8 found
no effect of BCG vaccination given before age 6 months on development of atopy
or allergic manifestations at age 7 years in a cohort study; in this study
13% of the children at high risk of tuberculosis were BCG-vaccinated, and
children born outside of Germany were overrepresented in the BCG-vaccinated
group. In another cross-sectional German study among preschool children a
weak protective effect of BCG vaccination against asthma was observed in German
children, whereas a stronger protective effect on atopic manifestations was
observed in children of non-German ethnicity.9
The strength of our study was its ability to minimize selection bias,
as practically all children within a birth cohort were either vaccinated or
not vaccinated. By examining children of the same age in 1998 and 2001 we
were able to compare the prevalence of atopy in vaccinated and unvaccinated
children within the same age group. The 2 examinations were performed in exactly
the same manner at the same time of year, and we obtained high participation
rates in both years. It could be speculated that an increase in the prevalence
of atopy from 1998 to 2001 would bias our results. However, we found the prevalence
of atopy according to age to be the same in 1998 and 2001, and we adjusted
for year of examination in the analyses. Furthermore, an increase in the prevalence
of atopy in the last 3 years would have tended to bias our estimates in favor
of a protective effect of BCG vaccination. Our study was further strengthened
by measuring atopy objectively in all children, and by obtaining information
on BCG vaccination status and exact date of vaccination independently of the
outcome for 93% of the children. The fact that the majority of the children
were vaccinated just after birth, when the protective effect of BCG on atopy
has been suggested to be strongest, reduced the risk of missing a potential
protective effect of BCG vaccination, as did the fact that our study was not
restricted to children who were genetically predisposed to atopy.
Although our study does not support the hypothesis that vaccination
with BCG is capable of preventing the development of atopy, we cannot rule
out that the effect of BCG vaccination may be different in populations with
other genetic constitutions.14 Furthermore,
our primary focus was on the effect of early BCG vaccination. Thus, we had
limited data regarding children vaccinated at ages outside the general recommendations.