The consensus statement on childhood tuberculosis constituted by the Working group on Tuberculosis, IAP 2008(1) claims that "Few studies have reported as high as 33% bacteriological positivity even in primary disease such as hilar lymphadenopathy." This contradicts the concept of primary tuberculosis, which we understand till date as being difficult to diagnose by demonstration of AFB due to its paucibacillary nature, and the fact that Ziehl-Neelson stain can reveal AFB only if the sample contains >10,000 bacilli per mL. In fact, both the references quoted by the working group(2,3); on which the entire algorithm for diagnosis of tuberculosis in children is based, are actually studies done on mixed population of primary, progressive primary and cavitatory tuberculosis. In the study by Somu, et al.(2) of the 50 cases, there were only 6 cases of hilar/mediastinal lymphadenopathy, of which only one was positive for AFB on gastric lavage(2). In their study, the positivity rate was highest in cases with cavitation and consolidation. In the study by Singh, et al.(3) of the 58 children, only 13 cases had primary complex or paratracheal/hilar lymphadenopathy. The study did not separately reveal the positivity of AFB on gastric lavage/BAL in this subgroup of children, but only reported the overall positivity in the study as 34.5%. Thus, generalising the conclusions of these studies in the general population with predominant primary complex seems to be unreasonable. Further studies in children with primary complex need to be done before such guidelines are laid down.

As regards treatment, the present algorithm rightly lays emphasis that there is no role for empirical trial of antitubercular therapy. However, in "probable cases" which includes all symptomatic children/children with history of contact with radiology suggestive of tuberculosis, positive skin test, but with bacteriology negative for AFB, the guidelines of treatment have not been specified. With the AFB positivity rate being actually low in primary complex (as mentioned above), and with not enough Indian data available, this would not be a good suggestion in a community set up in an endemic nation like ours were under-treatment of tuberculosis would be more hazardous than overtreatment.

1. It is true that there is scanty literature in India on bacteriological confirmation of childhood tuberculosis and specifically related to primary complex. Both studies quoted do show bacteriological positivity to an extent of 30% though separate data on primary complex is not available in one of the studies while the other study quoted 15% positivity in primary complex. Thus it was not possible to draw definite conclusion with studies involving small number of children. We intended to give a strong message that we must attempt bacteriological diagnosis in every case of childhood tuberculosis including primary complex irrespective of success, and I am sure more we try more we will find AFB.

2. As regards to "probable" case of childhood tuberculosis, decision of treating would depend upon individual physician’s analysis of probability. In case of doubt, one should consider another opinion and then take a decision. There cannot be structured protocol for such cases. It is not correct to presume that overtreatment is safer than undertreatment. In fact mistakes on both sides are hazardous and that is the reason we hope that our members follow the protocol to minimize both undertreatment and overtreatment. That is also the reason that we have stressed on bacteriological diagnosis.