Role of QuantiFERON-TB gold, interferon gamma inducible protein-10 and tuberculin skin test in active tuberculosis diagnosis

Abstract

Background: The measurement of Interferon gamma or Interferon gamma inducible protein (IP)-10 in antigen stimulated blood samples is suggested as an alternative method for latent tuberculosis (TB) diagnosis. Nonetheless, their role in active TB diagnosis, particularly in TB endemic settings is yet to be defined. In this study, the sensitivities and specificities of Interferon gamma release assay (IGRA), IP-10 assay and tuberculin skin test (TST) in detecting active TB cases were assessed in human immunodeficiency virus (HIV) sero-negative TB patients and healthy controls respectively.

Methods/principal findings: A total of 177 adult TB patients and 100 healthy controls were included for this study. QuantiFERON-TB Gold In-tube (QFT-IT) method was used to analyze the sensitivity and specificity of IGRA. QFT-IT, IP-10 and TST yielded the diagnostic sensitivities of 90.6% (95%CI: 86.3%-94.9%), 92.5% (95%CI: 88.6%-96.4%) and 68.9% (95%CI: 60.6%-77.2%) and specificities of 55% (95% CI: 35.2%-54.8%), 48% (95% CI: 38.2%-57.8%) and 75.5% (95% CI: 66.8%-84.2%), respectively. The extent of pulmonary involvement or presence of diabetes mellitus did not appear to influence the sensitivities of any of these tests. The combination of any of the two tests among QFT-IT, IP-10 and TST showed >98% sensitivity among smear negative cases and particularly the combination of IP-10, TST and smear microscopy showed 100% sensitivity, however, the specificity was decreased to 44.8%.

Conclusions/significance: QFT-IT and IP-10 were highly sensitive in detecting active TB cases. The combination with TST improved the sensitivity of QFT-IT and IP-10 significantly. Although the higher sensitivity of combination of QFT-IT/IP-10 and TST may be useful in active TB diagnosis, they are limited by their poor specificity due to the high prevalence of latent TB in our settings.

Figure 1. The levels of IP-10 in unstimulated, TB antigens stimulated and mitogen stimulated samples of healthy controls (a) and PTB patients (b).

The levels of IP-10 in TB antigen stimulated and mitogen stimulated samples were significantly higher than unstimulated samples. Box and Whisker plots show range, inter-quartile range and median. Nil – unstimulated; TB ag – TB antigens stimulated; mitogen- mitogen stimulated samples. *significant difference p<0.05 by Kruskal- Wallis test.

Figure 2. The levels of IFN-γ in unstimulated, TB antigens stimulated and mitogen stimulated samples of healthy controls (a) and PTB patients (b).

The levels of IFN-γ in TB antigen stimulated and mitogen stimulated samples were significantly higher that unstimulated samples. Box and Whisker plots show range, inter-quartile range and median. Nil – unstimulated; TB ag – TB antigens stimulated; mitogen- mitogen stimulated samples. *significant difference p<0.05 by Kruskal- Wallis test.

Figure 3. ROC curve analysis for TB antigens specific IP-10.

The cut-off value for TB antigens specific IP-10 levels was determined by using QFT-IT and TST negative healthy controls as uninfected and culture confirmed TB patients as diseases subjects. The area under curve was 0.946 (95% CI: 0.905–0.987). TB- tuberculosis; IP-10-Interferon gamma inducible protein-10; QFT-IT – QuantiFERON-TB Gold in-tube; TST – Tuberculin Skin test; CI-Confidence interval.