Combined IFN-γ and TNF-α release assay for differentiating active tuberculosis from latent tuberculosis infection

The IFN-γ-release assay (IGRA) cannot differentiate active tuberculosis (TB) from latent TB infection (LTBI). We hypothesized that the TNF-α-release assay (TARA) combined with IGRA might discriminate active TB from not active TB without LTBI.Adult patients with suspected TB, and with unrelated diseases such as herpes zoster as controls, were enrolled in an intermediate TB-burden country. Patients with confirmed or probable TB were regarded as active TB, and patients with not active TB were further classified as those having not active TB with and without LTBI based on IGRA results. The IGRA and TARA by using ELISPOT assays were performed on peripheral mononuclear cells.Thirty six patients with active TB and 53 patients including 18 not active TB with LTBI and 35 not active TB without LTBI were finally included. The sensitivity and specificity of the IGRA for those patients found to have active TB were 94% (CI, 80-99) and 66% (CI 52-78), respectively. Combining the IGRA and the TARA substantially increased the specificity for active TB (93%, CI, 82-98; P = 0.001) compared with the IGRA only, without compromising sensitivity (89%, CI, 73-96; P = 0.67).Combining the IGRA and TARA appears to be useful for diagnosing active TB.