Objective An interferon-gamma release assay (IGRA) is used to screen for latent tuberculosis infection (LTBI). Among IGRAs, the QuantiFERON-TB Gold In-Tube (QFT-GIT) results are highly variable, so the borderline zone has been proposed to reduce unnecessary LTBI treatment. The aim of this study was to examine the proportion of the borderline zone of QFT-GIT in healthcare workers' (HCWs) serial IGRA and to retrospectively identify the utility of predicting tuberculosis (TB) in a moderate TB incidence setting. Methods The participants were HCWs who had undergone serial LTBI screening between June 2013 and June 2018. IGRA-positive HCWs underwent examinations that included low-dose computed tomography (LDCT) and TB culture, if necessary. Applying the borderline zone (0.2-<0.7 IU/mL), the results were classified as definite negative, borderline negative, borderline positive and definite positive. Results Through the follow-up of 477 HCWs, 441 (92.5%) invariant, 30 (6.3%) conversion, 2 (0.4%) reversion and 5 (1.0%) indeterminate results were observed with the manufacturer's cutoff. Applying the borderline zone, 419 (87.8%) invariant, 22 (4.6%) conversion, 1 (0.2%) reversion and 36 (7.5%) decision pending, including 5 (1.0%) indeterminate results, were observed. At the time of screening, five TB cases were identified. Chest X-ray (CXR) identified one TB case, and LDCT identified four additional TB cases. After one year, two TB cases were diagnosed, and their screening QFT-GIT results were definite positive and borderline negative. In the Cochran-Armitage trend test, the greater the maximum difference in