The data were collected during an international multicenter trial of LTBI entitled, “A randomized clinical trial comparing 4RIF vs. 9INH for treatment of latent tuberculosis infection: effectiveness” (registration number; NCT00931736). The study protocol and inclusion and exclusion criteria of participants are described in the website (https://clinicaltrials.gov/ct2/show/study/NCT00931736). According to study protocol, the inclusion criterion for treatment was a tuberculin skin test (TST) reaction of ≥ 5 mm for close contacts of active pulmonary TB patients.

This current study was conducted at a single regional tertiary hospital in Korea, which was part of the international multicenter trial. This study was approved by the Institutional Review Board of Gyeongsang National University Hospital, Korea (IRB No. GNUHIRB-2009-042). Informed consent was obtained from the enrolled participants.

From October 2009 to August 2013, we recruited individuals (aged > 18 years), who were considered to be close contacts of active pulmonary TB patients. These individuals were asked to be volunteers for the LTBI test at the Gyeongsang National University Hospital, Korea. A “close contact” was defined as a person who had contact, for at least 4 hours/week, with a patient(s) with active pulmonary TB during an infectious period, and was confirmed by an acid-fast bacilli (AFB) smear and/or AFB culture test. We excluded health care workers, those diagnosed with active TB after chest radiography, and those who were in contact with patients with drug-resistant TB as indicated by the drug susceptibility profile. Participants underwent a TST and/or QuantiFERON-TB Gold (QFT-G, Cellestis Ltd., Carnegie, Australia) test after signing informed consent. If both tests were conducted during the same day, the TST was performed first. We assessed risk factors for contracting LTBI, including age, sex, history of bacilli-Calmette-Guerin (BCG) vaccination, history of pulmonary TB, cohabiting with a TB patient in the same house or room, presence of cough, AFB smear status of the index case, and cavitary lesions on the chest radiographs of the index case.

The TST was performed according to the Mantoux method (Statens Serum Institute, Copenhagen, Denmark). The participants were injected intradermally with 0.1 mL of tuberculin purified protein derivative RT 23 SSI (2TU) by a trained research nurse, and the transverse diameter (mm) of the induration at the injection site (inner forearm) was measured 48 to 72 hours after injection [10,11]. The QFT-G tests were performed by a specialized laboratory company (Samkwang Medical Laboratories, Seoul, Korea) according to the manufacturer’s instructions. We incubated whole blood at 37ºC for 16 to 24 hours with TB-specific antigens containing an early-secreted antigenic target 6 protein and culture filtrate protein 10, and then measured the concentration of the released interferon gamma in the supernatant solution using an enzyme-linked-immunosorbent serologic assay [12,13].

In this study, LTBI was defined as a positive result for either TST (≥ 10 mm) or QFT-G, and participants who tested positive for LTBI were randomly allocated to receive isoniazid for 9 months (9INH; 9 months of 5 mg/kg isoniazid up to 300 mg/day) or rifampin for 4 months (4RIF; 4 months of 10 mg/kg rifampin up to 600 mg/day). Participants who showed resistance to isoniazid or rifampin during the drug sensitivity test were excluded from the study. Participants were allowed to withdraw from treatment if they choose to do so. We also compared LTBI patients who accepted and refused treatment.

Descriptive data were presented as numbers and percentages; continuous data were indicated by mean ± standard deviation or median (range). Categorical variables were compared using the chi-square test or Fisher exact test, as appropriate. Continuous variables were compared using the parametric t test or the Mann-Whitney U test. Differences were considered statistically significant if p < 0.05. Statistical analyses were performed using SPSS version 18.0 (SPSS Inc., Chicago, IL, USA).

We identified 348 adults who had close contact with 128 active pulmonary TB patients who were willing to test for LTBI with the TST and/or QFT-G. We excluded health care workers (n = 61) and those aged < 18 years (n = 7). Individuals were also excluded if the TB patient with whom they had close contact with had the following: (1) no Mycobacterium tuberculosis (MTB) growth (n = 31) on their sputum or bronchial washing fluid samples; (2) positive test for mono- or multi-drug resistance (n = 21); and (3) no documented results of a drug sensitivity test for MTB (n = 40) (Fig. 1). After these exclusions, we enrolled 189 participants who were close contacts of 107 active TB patients. The mean number of tested participants per active TB patient was 1.8 ± 1.2 (range, 1 to 7). The number of participants who underwent the TST, QFT-G, and both TST and QFT-G tests were 183, 118, and 112 participants, respectively. Seventy-six participants were tested with TST and QFT-G on the same day, and the mean time interval for undergoing both tests was 3.1 ± 2.2 days (range, 1 to 14). Two participants were tested with the QFT-G 7 and 14 days after the TST.

The mean age of participants was 48.7 ± 15.7 years; 130 (68.8%) were women and 140 were household contacts (74.1%). Most of the participants (n = 170, 89.9%) had a scar from BCG vaccination (90.7% and 88.1% in the TST and QFT-G groups, respectively). A history of pulmonary TB was noted in 8.2% and 7.6% of the participants in the TST and QFT-G groups, respectively (Table 1). Comorbid diseases, including hypertension, diabetes, cardiovascular, chronic liver, and thyroid disease, were reported in 28.9% and 23.7% of participants in the TST and QFT-G groups, respectively. More than half of the participants in both the TST and QFT-G groups were household close contacts (73.8% and 72.9% in the TST and QFT-G groups, respectively) and resided in the same room (41.5% and 43.2% in the TST and QFT-G groups, respectively). In index cases, the rate of positive AFB smears obtained from the sputum or bronchial washing fluid in active pulmonary TB patients was 79.8% and 77.9% in the TST and QFT-G groups, respectively. Coughing as a respiratory symptom was observed in about 60% of active pulmonary TB patients. About 40% of active pulmonary TB patients were observed to have cavitary lesions in their chest radiographs.

Seventy-five of the 183 participants who underwent TST (≥ 10 mm, 40.9%) and 45 of the 118 participants who underwent QFT-G testing (41.0%) had positive results. There were no other significant differences in the clinical characteristics of TST-positive and TST-negative participants, including the mean age, proportion of participants with visible BCG vaccination scars, history of TB, existence of comorbid illnesses, proportion of participants who had never smoked, household contacts (or residing in the same room), AFB smear status, presence of cough, or chest radiograph-confirmed cavitary lesion(s) in the index cases with LTBI. Similar findings were also observed in QFT-G-positive and QFT-G-negative participants (Table 2).

Among the 88 participants with TST or QFT-G positive results, 43 agreed to treatment. Forty-two of the 45 participants did not consent to treatment, two had signs of active TB on their chest radiograph, and one did not visit the hospital after TST/QFT-G testing. There were no differences in age, smoking status, history of TB, presence of cough in index cases, and household contacts (or residing in the same room) between those who agreed to and refused treatment. However, a higher percentage of women (76.7% vs. 60.0%, respectively) and those with cavitary lesions on chest radiograph in index cases (55.8% vs. 24.4%, respectively) accepted rather than refused treatment. Among the 63 participants who underwent testing for both TST and QFT-G, a positive result for both tests was more common in those who accepted treatment compared to those who refused treatment (62.5% vs. 36.4%, respectively) (Table 3). Twenty of the 43 participants who agreed to treatment had positive results for both TST and QFT-G, while eight and six participants tested positive for only TST and QFT-G, respectively. Those who consented to treatment were randomly assigned to receive 9INH (n = 22) or 4RIF (n = 21). The median duration of treatment for participants who underwent 9INH and 4RIF was 6 months (range, 1 to 9) and 4 months (range, 1 to 4), respectively. Nine of the 22 patients in the 9INH group and 16 of the 21 participants in the 4RIF group completed treatment (Fig. 2).