Background: The protective role of BCG vaccination against tuberculous meningitis (TBM) is still controversial in India, largely due to the presence of other predisposing factors such as age, nutritional and socioeconomic status, and household contact. Very few Indian studies have focused on the interaction between BCG and these factors on the occurrence of TBM.
Methods: We did an unmatched hospital-based case-control study with prospective enrolment. Children with TBM diagnosed on the basis of predefined criteria were enrolled as cases. For each case, two children admitted on the same day and who did not have any neurological symptoms were enrolled as controls. Demographic data and information on predisposing factors for tuberculosis were collected for both cases and controls and the presence of a BCG scar was charted. Cases and controls were compared by univariate followed by multivariate analysis to obtain significant independent predictors for the occurrence of TBM. To assess the interaction between other predisposing factors and protective efficacy of BCG, a stratified analysis was also done.
Results: A total of 91 cases and 182 controls were enrolled over a one-year study period, of which 37 cases and 111 controls had a BCG scar. The crude odds ratio for the occurrence of TBM in the absence of a BCG scar was 2.28 (range: 1.32-3.94). The time elapsed since vaccination was significantly longer in the cases. Also, the proportion with a household contact was significantly higher in the cases, the mean age of the cases was higher than that of the controls, and the mean weight and height for age percentage were significantly lower. The cases had a significantly lower socioeconomic status. On multivariate analysis, the significant independent predictors for the occurrence of TBM were positive household contact with tuberculosis (adjusted OR 4.26; 95% CCI 2.26-8.04), absent BCG scar (adjusted OR 1.98; 95%ClI 1.09-3.57) and rural residence (adjusted OR 2.07; 95% ClI 1.02-4.17).
Conclusion: Vaccination with BCG was found to be protective even after controlling for the effect of other variables. Stratified analysis showed that protection due to BCG failed to reach significance for those > 5 years of age, if the weight was <6 0% of that expected for age, in the presence of a household contact with tuberculosis, and in socioeconomic classes III, IV and V.