The effect of BCG-vaccination and tuberculosis on the risk of leukaemia.

The information concerning leukaemia in the Cancer Register in Finland was matched with a BCG vaccination index (FVI) constructed in 1946-49. Male birth year cohorts 1926-41 were analyzed by comparing observed/expected ratios. There were 422 cases of leukaemia (A59) in these cohorts during 1949-78 which was the follow-up period. The expected values were counted on the basis of experience of the whole population of included cohorts. There were approximately 533,000 persons in the cohorts studied. About 315,000 participated in the mass BCG vaccination campaign. 115,000 were tuberculin positive, 200,000 were negative and BCG vaccinated. The cohort analysis techniques were applied and the findings presented as Lexis diagrams. A sample of the population register was used in order to check the matching errors in FVI. Obviously, the presented number of participants is not quite reliable. This is understandable because the register had not been in use from 1953-77 and contact with the organizers is no longer possible. With correction factors the errors can be adjusted, at least to some extent. The register is functioning as expected in relation to general mortality and tuberculosis. The results indicate that there is no marked difference between naturally T. B. infected and BCG vaccinated persons. The participants seem to have considerably less cases of leukaemia than the non-participants of which a great deal are not infected or vaccinated. Structural weaknesses in the FVI do not allow firm conclusions to drawn. There is in any case so much evidence by age, by calendar year and by cohort that a hypothesis of the diminishing long term effects on leukaemia is justified. The interrelationships are long standing and were noticeable during the whole follow-up period about 30 years. It seems that there is no difference in the possible effect between different types of leukaemia. The structural weaknesses, like the real number of participants in the register can be corrected. The analysis can be broadened to cover females as well, which will also facilitate analysis by different types of leukaemia. The possible long standing effect of mycobacterial infections are discussed. Some strategies of further research are outlined.