Decreased annual risk of tuberculosis infection in South Korean healthcare workers using interferon-gamma release assay between 1986 and 2005

Background Tuberculosis (TB) has been a major public health problem in South Korea. Although TB notification rate in Korea is gradually decreasing, still highest among the member countries of the Organization for Economic Cooperation and Development. To effectively control TB, understanding the TB epidemiology such as prevalence of latent tuberculosis infection (LTBI) and annual risk of TB infection (ARI) are important. This study aimed to identify the prevalence of LTBI and ARI among South Korean health care workers (HCWs) based on their interferon-gamma release assays (IGRA). Methods This was single center, cross-sectional retrospective study in a tertiary hospital in South Korea. We performed IGRA in HCWs between May 2017 and March 2018. We estimated ARI based on IGRA results. Logistic regression model was used to identify factors affecting IGRA positivity. Results A total of 3233 HCWs were analyzed. Median age of participants was 38.0 and female was predominant (72.6%). Overall positive rate of IGRA was 24.1% and IGRA positive rates age-group wise were 6.6%, 14.4%, 34.3%, and around 50% in the age groups 20s, 30s, 40s, and 50s and 60s, respectively. The ARIs was 0.26–1.35% between 1986 and 2005; rate of TB infection has gradually decreased in the last two decades. Multivariable analysis indicated that older age, healed TB lesion in x-ray, and male gender were risk factors for IGRA positivity, whereas working in high-risk TB departments was not. Conclusions Results showed that ARI in South Korean HCWs gradually decreased over two decades, although LTBI remained prevalent. Our results suggest that the LTBI test result of HCWs might be greatly affected by age, rather than occupational exposure, in intermediate TB burden countries. Thus, careful interpretation considering the age structure is required. Supplementary Information The online version contains supplementary material available at 10.1186/s12879-021-06855-5.


Background
South Korea is an intermediate tuberculosis (TB) burden country; in 2019, the annual incidence of TB was reported to be 59 per 100,000 [1]. TB is still an important public health concern in South Korea; for effective TB control, understanding its epidemiology, such as the prevalence of latent tuberculosis infection (LTBI) and annual risk of TB infection (ARI) is important. ARI represents the proportion of newly infected or reinfected people over a 1-year period calculated based on the prevalence of LTBI. LTBI prevalence has steadily decreased in South Korea [2,3]; however, about one third of the general population in South Korea is still thought to have LTBI [4,5].
To date, LTBI prevalence studies and analysis of ARI have been conducted mainly using the tuberculin skin test (TST) in South Korea [2,6,7]. In previous studies, the annual risk of TB infection in South Korea was estimated using TST results in ages 5-9 years, and the ARI was 0.5-5.3% based on a 10-mm induration between 1958 and 1988 [6]. However, studies on ARI using interferon-gamma release assay (IGRA) results in South Korea are lacking.
In this study, we aimed to identify the prevalence of LTBI and estimate ARI in health care settings based on IGRA results of HCWs. We also aimed at identifying the risk factors of LTBI for HCWs with a particular focus on their age in an intermediate TB burden country.

Study population
This study was conducted at Severance Hospital, a 2500-bed tertiary referral hospital in South Korea. A total of 3283 HCWs underwent the LTBI test between May 2017 and March 2018 as part of a national TB elimination program. The LTBI test used was QuantiF-ERON ® TB-Gold In Tube test (QFT-GIT). High-risk departments such as respiratory department and outpatient clinic, medical intensive care unit, emergency department, microbiology laboratory, and radiology department were defined as TB-related departments. We collected data related to participants' age, sex, occupation, hospital department, and years of employment; other electronic medical records were retrospectively reviewed. Previously healed TB lesions on chest radiographs were defined as calcified nodules or lymph nodes, discrete fibrotic scars with volume loss, or pleural thickening or calcification [8,9] Keywords: Latent tuberculosis infection, Prevalence, Annual risk of infection, Interferon-gamma release assays, Health care workers Fig. 1 Flowchart of the HCWs in the study. HCWs health care workers; IGRA interferon-gamma release assay; LTBI latent tuberculosis infection

IGRA
The IGRA was performed using a QuantiFERON ® TB-Gold In Tube test (QIAGEN, Hilden, Germany) following the manufacturer's instructions. A positive IGRA result was defined as an interferon-gamma response to the TB antigen minus that of the nil tube of ≥ 0.35 IU/mL. Chest radiographs were evaluated to exclude active TB at the same period. The HCWs were diagnosed with LTBI if the IGRA results were positive.

Statistical analysis
Continuous variables were presented as means ± standard deviation (SD), or median and interquartile range (IQR), and were analyzed using the Student t-test or the Mann-Whitney test. Categorical variables were reported as number and percentage, and were analyzed using the chisquare test or Fisher's exact test. To determine if IGRA positivity differed according to age in this study, we obtained the probability of a positive result in each subject using logistic regression and expressed it in the form of a figure.
Risk factors associated with IGRA positivity were investigated using logistic regression models, and the results were reported as odds ratios (ORs) and 95% confidence intervals (CIs). All tests of significance were 2-tailed, and a p value of less than 0.05 was considered statistically significant. All analyses were performed using SPSS software, version 23.0 (IBM, Armonk, NY, USA) and SAS 9.4 (SAS Institute, Cary NC).

Baseline characteristics of the study population
Overall, 3233 HCWs were analyzed after excluding 50 HCWs who had undergone previous anti-tuberculosis treatment or LTBI treatment (Fig. 1). We also excluded participants with presumptive active TB by chest x-ray and respiratory symptoms. There was no active TB case among our study participants. Baseline characteristics of HCWs are shown in Table 1. Of the total 3233 HCWs, 2348 (72.6%) were female, and the median age was 38 (IQR 29-46). Occupationally, the proportion of heath aids, physicians, and nurses was similar. The working period was divided into groups ranging from less than 1 year to more than 30 years with the percentage of employees working for 10-20 years being the highest (30.1%). Forty-two HCWs (1.3%) reported healed TB lesion on chest x-ray. Overall, 780 of 3233 HCWs were IGRA positive (24.1%). Prevalence of LTBI and predicted probability for IGRA positivity according to age Figure 2A shows the IGRA positive rate by age. Based on IGRA results, the prevalence of LTBI were 6.6% in the 20s, 14.4% in the 30s, 34.3% in the 40s, and around 50% in the 50s and 60s, respectively. As age increased, the likelihood of being positive on IGRA also increased, and the slope was 0.0929, which confirmed statistical significance (p < 0.001) (Fig. 2B). At age 40, the predicted probability of IGRA positivity was around 25% (Fig. 2B), and the results were confirmed to be similar to actual IGRA positive rates ( Fig. 2A) (AUC = 0.749, diagnostic performance was good) [11] (Additional file 1: Fig. 1).  Table 3 shows the risk factors associated with positive IGRA results. According to univariable analysis, older age group, male gender, healed TB on chest x-ray, and longer working duration were risk factors for IGRA positivity. On the other hand, currently working in high-risk (TB related) departments was not a risk factor for IGRA positivity. Even though doctors and nurses face patients directly, IGRA positivity was lower than that among administrative staff, which is explained by the fact that the average age of doctors and nurses was younger than that of administrative staff (Additional file 2: Table 1). In multivariable analysis, older age, healed TB lesion on chest x-ray, and male gender were risk factors for IGRA positivity, whereas currently working in high-risk TB department was not.

Discussion
In this study, using 3233 IGRA results of HCWs between 20 and 68 years of age, the overall prevalence of LTBI was 24.1%, and age was the strongest predictor of LTBI, rather than working in TB-related departments. The annual risk of TB infection ranged from 0.26 to 1.35% between 1986 and 2005, and its risk has gradually decreased over the last two decades.  ARI with Mycobacterium tuberculosis is known to be calculated from an observed prevalence of infection, approximating the incidence of infection. ARI for TB is one of the most important indicators for accessing its epidemiology in a population and is potentially informative about the transmission of TB within a community [12]. The main cause for the decline in ARI was a decreased prevalence of active TB in South Korea over 30 years, estimated from 443 to 167 cases per 100,000 population between 1985 and 2005, owing to economic growth and governmental effort to control TB via Bacillus Calmette-Guérin (BCG) vaccination, a TB notification system, and appropriate treatment through public health systems [3,13]. Additionally, improvement on infection control at medical institutions is believed to have contributed to reducing ARI among HCWs.
To date, trends for TB prevalence and ARI have been documented in several countries, and most research has been based on national TST surveys [14][15][16]. In South Korea, Kim et al. reported that the prevalence of LTBI using TST in persons < 30 years of age decreased from 55.9% in 1965 to 30.8% in 1995. In conjunction with this change, the ARI of TB in persons < 30 years of age decreased from 5.3% in 1965 to 0.5% in 1995 [2]. The use of TST however, has a poor specificity due to its crossreaction with BCG vaccine strain and nontuberculous mycobacteria. According to a study conducted in South Korea comparing TST and IGRA results according to TB infection risk [4], the agreement between TST and IGRA was not high, and IGRA was a better indicator of the risk of TB infection than TST in a BCG-vaccinated population. There have been attempts to estimate ARI using IGRA [17,18]. Nishimura et al. used IGRA data from HCWs in a university hospital in Japan and reported estimated ARIs for TB of 0.156% in 1986 and 0.049% in 2004 [19]. In a recent prospective study of a population of 13,580 individuals in China, ARI was estimated by conversion of IGRA and TST [20]. The annual TB infection rate was suggested to be 3.1% based on IGRA conversion and 1.5% based on persistent positive results after IGRA conversion. Meanwhile, however, ARI based on TST conversion was 14.5%, suggesting of limitation of ARI estimation based on TST results. The study concluded that the use of IGRA was more accurate than TST when estimating ARI [20]. However, studies on ARI using IGRA results are lacking in South Korea. Among recent largescale surveys, a survey of 2051 sample of the 2016 Korea National Health and Nutrition Examination Survey (KNHANES) showed that the overall prevalence of LTBI was 33.2% using TST results [3,21], which was higher than our study results. Considering the false positivity of TST, we believed that our data using IGRA results would reflect more accurate LTBI prevalence and ARI in South Korea where BCG vaccination is mandatory.