The rational design of a successful vaccination strategy against tuberculos
is requires certain kinds of information and must take account of several c
onsiderations: (i) the nature of the immune response that protects the larg
e majority of individuals infected by Mycobacterium tuberculosis, designate
d as healthy contacts, must be defined and distinguished from that in tuber
culosis patients, whose immune system must have failed; (ii) the vaccinatio
n strategy must incorporate a way of priming the immune system to guarantee
in all individuals this protective response, normally generated in healthy
contacts, upon natural infection by hi. tuberculosis; (iii) the strategy m
ust incorporate a mechanism for ensuring that the effectiveness of this pri
ming is not abrogated by exposure to environmental mycobacteria; and (iv) t
he strategy must take account of the fact that the vaccinated population is
genetically heterogeneous, and that individuals will therefore respond var
iably to most standard vaccination protocols. We describe a tentative propo
sal for how these interrelated problems might be solved and discuss predict
ions of this tentative vaccination strategy. Critical testing of the neonat
al, low-dose BCG vaccination strategy can only be achieved by a field trial
and we outline the considerations underlying this proposal.