Introduction
In France, primary BCG immunisation is mandatory for children before they
can enter daycare centres or the care of childminders, and must be given
by the age of six years at the latest, when school entry is compulsory.
BCG is also recommended in the first month of life for high risk children.
Until June 2004, a tuberculin test was required between 3 and 12 months
after vaccination and at 11-13 years of age, followed in both cases by
revaccination, if negative. Following an evaluation initiated by the
publication of a report by the national institute for public health surveillance,
the Institut de Veille Sanitaire (InVS), routine tuberculin testing in
children and revaccination were discontinued in July 2004, as was revaccination
for exposed professionals [1]. Tuberculin testing remains indicated as
a diagnosis tool for tuberculosis infection or disease and for the follow
up of health or social workers for whom BCG vaccination is still mandatory.
The InVS report also questioned the need for universal BCG immunisation
of children. For the period 2000-2002, the incidence of positive sputum
smear tuberculosis cases was 4.6 per 100 000 (5.7 when correcting for the
lack of exhaustiveness of the notification). For the 1998-2002 period,
the incidence of meningitis in children less than five years old was 0.4
per 10 millions. This epidemiological situation was thus very close to
the threshold values proposed by the International Union Against Tuberculosis
and Lung Diseases for possible discontinuation of BCG vaccination [2].
An overview of tuberculosis control strategies, and an epidemiological
assessment of the consequences of reducing BCG vaccination activities was
conducted in the context of a multi-disciplinary evaluation by the national
institute for health and medical research (Inserm), at the request of the
health authorities.
In this context, InVS assessed the epidemiological impact of discontinuing
universal children BCG immunisation. Several arguments were in favour of
also studying the impact of a strategy targeted on children at risk. The
data from mandatory notification of tuberculosis cases in France show that
the risk of tuberculosis is highly heterogenous, according to nationality
or country of birth. In 2003, the incidence of the disease was 10.2 per
100 000 but was tenfold higher in non-French nationals than in nationals
(respectively 72.1 versus 8.1 per 100 000, all ages together, and 18.7
versus 1.8 per 100 000, in children under 15 years old) [3]. In 2003, eight
out of the then 15 European Union (EU) countries had chosen to target children
at high risk of tuberculosis for immunisation [4]. By 2005, nine of the
25 member states of the EU had applied such targeted strategies [5].
Two scenarios for changes in BCG immunisation programme in France were
therefore assessed: the total discontinuation of any vaccination and targeted
vaccination for children living in a risk environment.
Methods
Assessment of the impact of total discontinuation of immunisation
The number of excess tuberculosis cases that would be observed if immunisation
were completely discontinued is equivalent to the number of cases of
tuberculosis avoided each year by the current immunisation programme.
This figure was estimated from BCG effectiveness estimates, immunisation
coverage and tuberculosis notification data [6]. Based on published data,
we considered the hypothesis of a protection provided by BCG lasting
until the age of 15 years, and concerning only vaccinated persons (no
indirect protection for unvaccinated subjects due to the absence of reduction
of the circulation of the tuberculosis bacillus, as childhood tuberculosis
is rarely contagious). Two hypothesis on vaccine effectiveness were considered.
In the basic hypothesis, BCG effectiveness was considered to be 75% for
tuberculous meningitis and miliary, the most severe localisations of
the disease, and 50% for other sites, mainly pulmonary. In the hypothesis
most favourable to immunisation, considered in order to avoid underestimating
the number of additional tuberculosis cases that would follow the reduction
in BCG immunisation activities, the effectiveness of BCG was considered
85% on tuberculous meningitis and miliary and 75% on other sites.
The numbers of observed cases were estimated from mandatory notification
data between 1997 and 2002, corrected for lack of exhaustiveness, on
the basis of a notification rate of 75% for childhood tuberculosis [6].
The data on immunisation coverage are available through the analysis
at national level of health certificates completed at 24 months of age
for each child and from a national survey carried out in schools in 1997,
in children 5 to 6 years old [7,8].
Assessment of the impact of immunisation targeted on children
at risk
The definition used was based on the Swedish experience and matched
the French epidemiology of tuberculosis. It included children meeting
at least one of the following criteria:
• A child coming from a country with high tuberculosis prevalence;
• A child born into a family coming from a country with high tuberculosis
prevalence;
• A child of any origin with a history of tuberculosis in his/her family.
Africa, Asia (except Japan), Central and South America, the Russian Federation
and Baltic countries were considered as areas or countries with high
tuberculosis prevalence.
Among total childhood TB cases, the proportion of those occurring in
at-risk children was estimated at 75%, based on a study carried out in
1997 in the Parisian area [9].
Two levels of immunisation coverage of at-risk children were considered:
95% and 50%. Interrupting the universal immunisation of children could
lead to a decrease in the current immunisation coverage among targeted
populations, as such a decision would de facto imply the discontinuation
of mandatory vaccination.
Based on data from a survey carried out by the National Institute for
Demographic Studies (INED) [10], the number of children at risk was estimated
at 14% of each yearly birth cohort (that is about 100 000 children out
of a birth cohort of about 750 000).
Assessment of adverse effects of BCG immunisation
Frequency of clinically significant adverse effects was studied in the
evaluation carried out by Inserm [4]. BCG immunisation was estimated
to result each year in about 300 lymphadenitis (corresponding to a
rate of about 40 per 100 000 vaccinated children) and about 12 disseminated
BCG infections (corresponding to a rate of about 1.6 per 100 000 vaccinated
children), the latter occurring in children with severe immunodeficiency.
These data have been used to calculate the decrease of the expected
number of side effects for the different options of reduction of immunisation
activities.
Results
The epidemiological consequences of the different immunisation options
that were considered, for children under 15 years old, are summarised
in the table. If immunisation is restricted to children at risk, there
might be 80 to 200 additional cases per year in the non vaccinated low
risk population, corresponding to an incidence rate between 0.9 and 2.3
per 100 000 non vaccinated children. In case of a decreased coverage
in the targeted population, additional cases would also occur in children
at risk. For a vaccination coverage of 50%, annual additional cases could
range from around 200 to almost 500, corresponding to an incidence rate
in non vaccinated children between 2,1 and 5,2 per 100 000. If vaccination
were discontinued, 320 to 800 additional cases might occur every year,
depending on the hypothesis for vaccine effectiveness, corresponding
to an estimated incidence of additional tuberculosis cases between 3,2
and 8,0 per 100 000.
The projected increase to the current incidence in children 0-14 years,
and the number of additional cases and incidence in non-vaccinated children,
broken down for pre-school (0-5 years) and school-age children (6 to
14 years) are also presented in the table.
As about 15% of the children can be considered to be at risk, at least
85% of side effects due to BCG would be avoided through targeting immunisation
to those children.
Discussion
The first option analysed the total discontinuation of BCG vaccination.
Our study shows that such a choice could lead to several hundreds of
additional tuberculosis cases in children each year. These results
are in accordance with observations from other European countries,
particularly Sweden, when immunisation was first completely discontinued
[11], and are in favour of maintaining the current programme. In another
hand, such an option would induce several disseminated BCGitis cases
each year.
The alternative option analysed was the targeting of BCG vaccination
on children living in a risk environment. Based on our assessment, this
programme would avoid approximately three quarters of tuberculosis cases
that are currently avoided by the universal vaccination, while requiring
the vaccination of only about 15% of children. The real impact of this
option will depend on the ability in maintaining a high vaccine coverage
among children at risk. Such an option would also facilitate the interpretation
of tuberculin tests performed in non-vaccinated children exposed to tuberculosis
cases.
To analyse the epidemiological impact of immunisation, we relied on hypotheses
or point estimates, particularly those concerning the exhaustiveness
of tuberculosis notifications in France, the vaccine effectiveness, the
proportion of tuberculosis cases occurring among children with a risk
factor, and the size of this latter population. The resulting estimates
should be considered as orders of magnitude of the current or future
impact of different immunisation options. Nevertheless, the conclusions
on the relevance of the different options can be considered as fairly
reliable. It is worth mentioning that in this analysis, we did not consider
the increase of incidence of non-tuberculosis mycobacterial diseases
that would follow the decrease in BCG immunisation coverage, and that
only mainland France was considered.
These results were presented in 2005 to the Comité Technique des
Vaccinations (CTV, national advisory board on vaccination) and to the
Conseil Supérieur d’Hygiène Publique de France (CSHPF,
national high committee of public hygiene). Both recommended that the
Ministry should adopt the BCG vaccination strategy targeted at children
at high
risk of tuberculosis. They also recommend, as a prerequisite before actually
switching to such a strategy, the strengthening of other tuberculosis
control measures aimed at reducing the risk of infection for children
(such as early identification of cases, tracing of secondary cases and
of the source of contamination, and supervision of treatment of cases),
in the context of a national tuberculosis control plan. Following this
recommendation, the Ministry of Health has, in early 2006, set up an
ad hoc national committee that has began to formulate such a document.
The schedule for implementation of a targeted strategy will have to take
into account a new situation that could influence the BCG vaccination
coverage and the acceptability of maintaining the current programme.
Since early 2006, the multipuncture device used in France for more than
90% of BCG primary vaccinations, is no more available. Administrating
the vaccine intradermally in young infants is a difficult technique for
untrained vaccinators. Currently, more than 80% of children are vaccinated
in their first year of life. A survey of over 800 general practitioners
carried out in 2005 found that fewer than 30% felt ready to routinely
immunise very young infants intradermally [12]. To avoid the problems
that would arise from mandatory vaccination before entering school or
daycare centre, when only intradermal inoculation would be possible,
a strategy targeting the most exposed children, who are mainly those
from families of foreign origin, should perhaps be considered rapidly.
The feasibility and social acceptability of this option would have to
be ascertained beforehand.
Acknowledgements
We wish to thank the members of the Inserm collective expertise for their
contribution in this analysis, as well as Didier Che and Bénédicte
Decludt† from InVS, for supplying epidemiological data on tuberculosis.
|