Background & objective: To estimate the efficacy of BCG in preventing tuberculosis over a 15-year period, and also to assess the impact of infection with nontuberculous environmental mycobacteria in a rural community in Chingleput district in Tamil Nadu in south India. We re-analysed the 15-year follow up data of a large randomized trial conducted earlier.
Methods: A double-blind randomized control trial was initiated in 1968, in which over 100,000 uninfected subjects with a normal radiograph were allocated to placebo, BCG in low dose (0.01 mg) or BCG in high dose (1.0 mg); two widely used strains of BCG were employed, each in one half of the vaccinated subjects. Sensitivity to purified protein derivative (PPD-B) was also determined. The study population was followed for 15 yr by radiographic surveys of the total population once every 2.5 yr, selective case finding in suspects once in 10 months, and investigation of those reporting voluntarily with chest symptoms.
Results: Coverage by radiography was of the order of 80 per cent throughout, while coverage by sputum examination of suspects was usually 90 per cent or above. The annual incidence of culture-positive tuberculosis (irrespective of smear) was estimated to be 55 per 100,000, and neither strain of BCG had any effect. The failure to protect was seen in both males and females, and in children and adults. However, in a subset of over 40,000 subjects who were also nonreactors to PPD-B, BCG had a low level of protection, i.e., 32 per cent (95% CI=3-52%), 29 per cent with the Danish strain and 34 per cent with the French strain.
Interpretation & conclusion: Our findings reaffirm that BCG was of little value in preventing sputum-positive cases of pulmonary tuberculosis.