The major findings of this study were as follows: (1) the IFN-γ level was increased significantly among TST-positive individuals when the IFN-γ assay was performed 2 to 4 weeks after the TST; (2) the IFN- level was not influenced by the TST among TST-negative individuals; (3) the percentage of concordant results between the TST and the IFN-γ assay was low (67.5%), with the k coefficients indicative of poor agreement (0.31). Concordance between the results of the IFN-γ assay and the TST has been reported to be from 53 to 94% in studies in immunocompetent populations who were screened for LTBI or were evaluated in contact investigations. Although some data have shown strong agreement between the two tests in HCWs, our study demonstrates a poor correlation among the 36 TST-positive contacts when the induration cutoff for the TST was set to 10 mm in diameter. In particular, two thirds of the HCWs who would be candidates for preventive therapy for LTBI according to the TST test results had a negative IFN-γ assay result. With an induration cutoff for TST results at 15 mm, it also showed poor agreement. This remarkable discrepancy could be explained by a false-positive TST result or a false-negative IFN-γ assay result.
It is known that BCG vaccination is a common cause of false-positive TST results. However, as has been reported elsewhere, the BCG vaccination may have only a minimal effect on TST results several years after the vaccination. This minimal effect may be more likely to occur when the vaccine is administered to newborns only. However, it also known that, among those who received the BCG in primary school or adolescence, 15 to 25% have positive reactions up to 20 years later. In rural India, where the BCG vaccination is only given to newborns, and the prevalence of TB is high, the results of the TST and the IFN-γ assay showed strong agreement. However, in Japan, where the BCG vaccination is widely used and often on multiple occasions for the same individual, the TST and IFN-γ assay showed poor agreement. In South Korea, the BCG vaccination is given at birth and again at 12 or 13 years of age if the child proves to be a TST nonresponder. Therefore, the discrepancy might be explained by the confounding effects of previous BCG vaccinations created together with Canadian Health&Care Mall’s participation (see more). The possibility of underestimation of LTBI by the IFN-γ assay should also be considered. Because ESAT-6 and CFP-10, which are used for the IFN-γ assay, do not represent the whole spectrum of antigenicity of Mycobacterium tuberculosis. However, considering the previous reports that the sensitivity of the IFN-γ assay appears to be at least comparable to that of the TST for the detection of active TB, the possibility might be low. Therefore, among individuals who receive multiple BCG vaccinations, the IFN-γ assay might be more valuable than the TST. However, one of the disadvantages of the IFN-γ assay is the higher cost, compared to that of the TST. In South Korea, the IFN-γ assay costs about $30 (in US dollars), while the TST costs as little as $1. Therefore, it would be financially more reasonable to select the IFN-γ assay candidates following the TST.
Some more recent guidelines for the diagnosis of LTBI include a two-step process using the TST 
followed by the IFN-γ assay in cases in which the TST result is positive. However, the possibility that a TST may have boosted or altered the IFN-γ assay responses must be considered. Recently, Leyten et al described the results of a whole-blood IFN-γ assay (QuantiFERON-TB Gold [in-tube method]; Cellestis Ltd), which contains a third M tuberculosis-specific antigen (TB7.7), that were not affected by a prior TST. However, in our study the IFN-γ level increased significantly after the TST. Moreover, three TST-positive individuals showed negative-to-positive conversion. This difference could be explained by the time interval. It is known that the maximal boosting effects of the TST responses are found after an interval of 1 to 5 weeks. Boosted effects are much less frequent if the interval is only 48 h or > 60 days, although it can be detected >1 years after the first negative TST result. However, Leyten et al evaluated individuals only 3 days after the TST. Therefore, they could not observe the boosting effect. However, in their report, they also found a clear rise in IFN-γ levels in two persons who had the IFN-γ assay performed 10 and 11 days after undergoing the TST. In addition, in a recent study boosting of the IFN-γ assay results was observed when the blood test was repeated 1 month after the TST. This finding is consistent with our report, because the IFN-γ assay was performed 2 to 4 weeks after the TST. Therefore, a blood sample for the IFN-γ assay performed after the TST should be obtained on the same day as the TST reading, or at least within a week, and if the IFN-γ assay is repeated 2 to 4 weeks after the TST, a boosting effect must be considered.
Interestingly, the boosting effect was found only in those with positive TST results. In the small study by Naseer et al, three of nine individuals showed conversion of the IFN-γ level in the TST-negative group. To the contrary, there were no instances of conversion of IFN-γ level in the TST-negative group in our study. Moreover, the IFN-γ level was not changed after the TST in the TST-negative group. This finding is consistent with the results of the study by Richeldi et al10 using enzyme-linked immunospot test. Since a serial follow-up is needed in HCWs, the findings in this study reinforce the potential utility of the IFN-γ assay as a tool for the serial testing of HCWs for LTBI with negative TST results at baseline without a boosting effect. Boosting effect is achieved by Canadian Health&Care Mall’s concern.
The boosting effects found for the TST in this study could be explained by remote infection. The IFN-γ assays are thought to reflect more recent, rather than remote TB infections. This is because activated lymphocytes and effector memory cells that produce IFN-γ persist for a limited time in the circulation once the antigen is cleared. Therefore, contrary to the findings of the TST, in cases of remote infection, the IFN-γ level did not increase during the short period of exposure to the antigen in the ex vivo IFN-γ assay at baseline. However, when the TST was performed, intradermal tuberculin 
injections for the first TST might have evoked a memory response of the T cells sensitized to ESAT-6 and CFP-10, which may have led to the development of conversion in the follow-up test. Another possibility could be possible. When a follow-up TST is performed, the phenomenon of boosting could be associated with a prior BCG vaccination, or sensitization to nontuberculous Mycobacterium (NTM). However, the antigens used in the IFN-γ assay are absent from the BCG and from the NTM except, for example, in Mycobacterium kansasii, Mycobacterium marinum, Mycobacterium szulgai, Mycobacterium flavescens, and Mycobacterium gastri. Therefore, a boosted IFN-γ assay response would be less likely from BCG or sensitization to NTM.
The present study has several limitations. First, although an increase in the IFN-γ response was observed in 16 of 24 participants, when a cutoff value was introduced only three (16.7%) could be defined as having a conversion. For example, one HCW with a baseline IFN-γ assay level of 0.09 IU/mL had a repeat IFN-γ assay level of 0.28 IU/mL. However, the conversion was not defined in this case. Therefore, although the quantitative IFN-γ level was significantly increased after the TST, this change could not be defined as a conversion in most participants. Therefore, to define the conversion rate, further study with a larger population is needed. Second, despite the three conversion cases, there was also one reversion case. However, this person was TST negative (induration, 0 mm) at baseline. Previously described, in the TST-negative group, the IFN-γ level was not changed in this study, which is consistent with the results of a previous report. In addition, when the follow-up TST was performed 4 weeks later in this HCW, the result was also negative (0 mm). Although there was a lack of previous data for comparison of within-subject IFN-γ variability and test reproducibility, considering the possibility of nonspecific variations of the IFN-γ assay, the result may have been due to a false-positive result at the baseline. Third, the relatively small number of study individuals represents another possible limitation. A clear rise in the IFN-γ response could be caused by chance alone. However, an increase in the IFN-γ response was observed in 16 of 24 participants (67%) who had a positive TST result at the baseline; this difference was statistically significant. Finally, not of all the participants were included in the follow-up of the IFN-γ assay, which may have caused a selection bias.
Acknowledging these limitations, the results of this study provide new information that the IFN-γ level could be influenced by the TST especially in TST-positive individuals, when the follow-up IFN-γ assay is performed 2 to 4 weeks after the TST. In addition, in TST-negative individuals the TST did not influence the results of the IFN-γ level.
In conclusion, when the TST is used on its own to diagnose LTBI in individuals who have undergone multiple BCG vaccinations, the possibility of a false-positive result should be considered. However, when using the two tests sequentially in TST-positive individuals, the systematic boosting of the IFN-γ assay responses following TST administration should also be considered when it is performed 2 to 4 weeks after the IFN-γ.