Abstract:

Objective To explore factors influencing T-SPOT.TB results in order to improve the accuracy of tuberculosis diagnosis.Methods Clinical data from 1537 inpatients was collected between May 2014 and April 2015 in Shanghai Pulmonary Hospital. Patients were grouped into a clinical TB group (1159 cases) and a clinical non-TB group (378 cases), according to the preliminary clinical diagnosis. Patients were then regrouped into a TB group and non-TB group according to the pathogenic diagnosis. The TB group (1103 cases) included 229 cases confirmed by acid-fast bacillus (AFB) staining and mycobacterium bacillus culture, and 874 cases confirmed by final clinical diagnosis, and the non-TB group (395 cases) included 94 cases with a previous history of tuberculosis or old tuberculosis lesions demonstrated by radiology, and 301 patients with no past history of tuberculosis. Thirty-nine cases, whose final diagnosis was uncertain, were excluded from the study. The 1103 cases in the TB group consisted of 557 cases of pulmonary tuberculosis, 107 cases of tuberculous pleurisy, 16 cases of lymphoid tuberculosis, 51 cases of osteoarticular tuberculosis, 7 cases of cerebral tuberculosis, 363 cases of multiple organ tuberculosis and 2 cases of other extrapulmonary tuberculosis. There were 93 cases of nontuberculous mycobacterial (NTM) infection among the 395 cases in the non-TB group. We retrospectively analyzed the sensitivity and specificity of T-SPOT.TB in the different groups and discuss factors influencing the results.Results (1) The diagnostic sensitivity of T-SPOT.TB in the clinical TB group and the TB group were 81.97% (950/1159) and 83.77% (924/1103), respectively, and the diagnostic specificity of T-SPOT.TB in these two groups were 53.44% (202/378) and 54.43% (215/395), respectively. The positive diagnostic ratio of T-SPOT.TB was 90.73% (235/259) for TB cases with positive AFB results and 92.58% (212/229) for those with positive culture results. (2) There was a significant difference in the positive rate of T-SPOT.TB between non-TB cases with a previous history of tuberculosis or with old tuberculosis lesions demonstrated by radiology (69.15%, 65/94), and non-TB cases with no previous history of TB (38.21%, 115/301; χ ²=27.65, P=0.000). (3) The TB specific antigen A and B spot counts in the TB-group (69.81% (652/934) and 69.67% (627/900) for spot counts ≥1 and spot counts ≥11) were significantly higher than in the non-TB group (46.83% (96/205) and 46.63% (83/178) for spot counts ≥1 and spot counts ≥11; Z=-14.20, -14.63, respectively; P<0.01). (4) The positive ratios of T-SPOT.TB for tuberculosis varied according to the type of organ lesion. The positive diagnostic ratio was highest for lymph node tuberculosis (87.50%, 14/16), followed by multiple organ tuberculosis (87.33%, 317/363), pulmonary tuberculosis (86.54%, 482/557), tuberculous pleurisy (74.77%, 80/107). The positive diagnostic ratio was low for cerebral tuberculosis (57.14%, 4/7) and osteoarticular tuberculosis (50.98%, 26/51). (5) The average age of NTM cases with positive T-SPOT.TB results (53.61±18.43) was higher than that of TB cases with positive T-SPOT.TB results (44.98±18.88) (t=-2.63, P=0.009). Conclusion Diagnostic methods, sputum bacteria counts, previous history of tuberculosis or old tuberculosis lesions, non-tubercular mycobacterium infection secondary to tuberculosis, and the type of tuberculosis can influence T-SPOT.TB results and should not be ignored in T-SPOT.TB-based TB diagnosis.

Key words: Tuberculosis, Diagnosis-related groups, Mycobacterium infections,atypical, Enzyme-linked immunospot assay, Diagnostic techniques and procedures, Factor analysis,statistical, Evaluation studies