Predictors of Discordant Tuberculin Skin Test and QuantiFERON-TB Gold In-tube Results in Eastern China: A Population-based, Cohort Study

Background. Discordance between the QuantiFERON-TB Gold In-tube (QFT) and tuberculin skin test (TST) is not well understood. We aimed to identify the factors that determine discordance between the TST and QFT when compared to either TST(+)QFT(+) or TST(-)QFT(-) results in a medium tuberculosis (TB) burden setting. Methods. We conducted a population-based study in Eastern China and administered TSTs and QFTs to participants. We calculated kappa values while constructing multivariable logistic regression models to evaluate predictors of test discordance. We analyzed the predictive value of discordant and concordant test results for progression to TB over 6 years of follow-up. Results. Overall, 5405 participants were enrolled; 2043 (37.8%) and 1104 (20.4%) were TST and QFT positive, respectively. There was fair agreement between the TST and the QFT (kappa values between 0.30-0.39 at different TST cutoffs). Agreement was lower among participants vaccinated with Bacillus Calmette-Guerin (BCG; kappa, 0.17 versus 0.47 in nonvaccinated participants). TST(+)QFT(-) results were associated with decreasing age, smoking, undiagnosed diabetes, and BCG vaccination (adjusted odds ratio, 1.45; 95% confidence interval [CI], 1.11-1.90). TST(-)QFT(+) results were associated with increasing age, male sex, smoking, and diagnosed diabetes. Compared to participants with TST(-)QFT(-) results, QFT(+) and TST(+)QFT(+) participants were 6.3 (95% CI, 1.9-20.4) and 7.5 (95%CI, 2.3-25.1) times more likely to progress to TB, respectively. Conclusions. In this population-based study of over 5000 participants from a medium TB burden region, the test agreement between QFT and TST was fair overall and we found multiple novel predictors of discordant QFT/TST results. QFT provides a substantial improvement to the TST among these populations and was multi-fold better at predicting progression to TB.