Background Prompt diagnosis of Mycobacterium tuberculosis （MTB） infection is an essential step in tuberculosis control and elimination. However, it is often difficult to accurately diagnose pediatric tuberculosis （TB）. The tuberculin test （TST） may have a low specificity because of cross-reactivity with antigens present in Mycobacterium bovis bacillus Calmette-Guerin （BCG） and other mycobacteria, especially in China with a predominantly BCG-vaccinated population.Early-secreted antigenic target 6-kDa protein （ESAT-6） and culture filtrate protein 10 （CFP-10）, stand out as suitable antigens that induce an interferon-gamma （IFN-γ） secreting, T-cell-mediated immune response to infection. While,considered the higher costs and complexity of the IFN-γ release assay （TSPOT）, we aimed to evaluate the TSPOT and TST test in the clinical diagnosis of pediatric tuberculosis and to establish a diagnostic process suitable for China.Methods The sensitivity and specificity of the assay were evaluated in total seventy four children with active tuberculosis and fifty one nontuberculous children with other disease, and then the results were compared with TST.Logistic regression models were used to identify variables that were associated with positive results for each assay. The independent variables included sex, age, birth place, vaccination history, close contract with an active TB patient.Results The sensitivity of TSPOT was higher than TST in active TB children with or without BCG vaccination, as well as in children with culture-confirmed TB. But the difference was not significant statistically. Combining results of the TSPOT and TST improved the sensitivity to 94.6%. Agreement of the TST and TSPOT was low （77.0%, k=0.203） in active TB patients. The difference in specificity between TSPOT and TST test was statistically significant （94.1% vs.70.6%, P=0.006）. Specificity of the two tests in patients without prior BCG vaccination history was similar （80.0% vs.60.0%）. The concordance betw
