Dose-response relationship between treatment delay of smear-positive tuberculosis patients and intra-household transmission: a cross-sectional study.

In order to document the effect of treatment delay on tuberculosis (TB) latent infection among the household contacts of TB patients, a cross-sectional TB infection prevalence survey was conducted among household contacts in Yunnan Province, southern China. In total, 1360 household contacts of 393 smear-positive pulmonary TB patients were enrolled, together with 308 household contacts of 90 non-TB patients. Using the contacts of non-TB patients as the baseline of TB infection, there was a dose-response relationship between household infection and delay of TB treatment (TB infection prevalence 9.7, 7.8, 19.9, 25.7 and 26.9% for non-TB case, TB case with delay < or =30 d, 30-60 d, 60-90 d and >90 d, respectively). Older age, TB index patient with lung cavitation, and sleeping in the same bedroom with a TB patient were all associated with an increased risk of being tuberculin-skin-test positive. In conclusion, 30 d delay in treatment seems to be the turning point at which a significant increase in risk for TB infection occurs. Apart from conventional indicators, magnitude of treatment delay should be considered as a performance indicator for TB control programmes in high-TB-burden countries. Measures for the detection of early cases should be intensified.