The study was undertaken in a sample of 103 villages of 3 sub-divisions of Bangalore district as a part of the 5 year study of epidemiology of tuberculosis between 1961-68. The follow ups were done at 1.5, 3 & 5 years after the first survey. The entire population was offered tuberculin test with 1 TU RT 23, a second test with 20 TU RT 23 to those persons who were having reactions of 0-13 mm to 1 TU. All aged 5 years or more were offered 70mm photofluorograms at each survey. Two specimens of sputum were collected from persons having abnormal X-ray shadows for examination of tubercle bacilli. Procedures were uniform at each survey. The population was divided into three groups on the basis of their tuberculin reactions: (a) reactors to 1 TU (infected with M.tuberculosis), b) non-reactors to 1 TU but reactors to 20 TU (infected with atypical mycobacteria), c) non-reactors to both 1 TU & 20 TU (not infected with either M.tuberculosis or other mycobacteria). Incidence of disease and crude mortality were studied separately among these groups.

The five year incidence of culture positive disease was the highest among 1 TU reactors and the least among reactors to 20 TU. In the younger age group (5-14 years) the five year incidence of culture positive disease among reactors to 20 TU was significantly lower compared with that among 20 TU non-reactors. The reduction of incidence of culture positive cases in the former group over that in the latter was 75% for culture positive cases and 61% for combined culture positive and negative disease. As regards crude mortality, the overall rate was significantly lower among 20 TU reactors compared with non-reactors. Even if the significance of the finding on crude mortality is debatable, it could be concluded that non-specific infection provides some protection against development of tuberculosis, at least in younger age groups.