Risk of active tuberculosis development in contacts exposed to infectious tuberculosis in congregate settings in Korea

Contact investigation is an important and effective active case-finding strategy, but there is a lack of research on congregate settings in countries with an intermediate incidence. This study determined the incidence of and risk factors for tuberculosis (TB) development after exposure in congregate settings. This retrospective cohort study included 116,742 contacts identified during the investigation of 2,609 TB cases diagnosed from January to December 2015. We searched the Korean National Tuberculosis Surveillance System TB registry to identify contacts that developed active TB during follow-up until May 2018. During the mean observation period of 2.9 years, 499 of 116,742 contacts (0.4%) developed new active TB. From these contacts, 404 (81.0%) developed TB within 2 years after exposure. The 2-year Kaplan-Meier cumulative risk for TB was the highest in contacts aged ≥65 years [1%; 95% confidence interval (CI), 0.8–1.3]. Contacts with LTBI who completed chemoprophylaxis exhibited a lower risk of active TB development than those without chemoprophylaxis (adjusted hazard ratio, 0.16; 95% CI, 0.08–0.29). Aggressive contact investigation is effective for the early detection and prevention of TB in congregate settings. The risk of progression to active TB among contacts with LTBI can be minimised by the completion of chemoprophylaxis.


Discussion
In this study, we analysed the TB incidence in individuals who had contact with active TB cases in congregate settings in the ROK and sought to confirm the risk factors that influenced the development of TB in these cases. When 116,742 contacts of TB cases reported in 2015 were followed up for an average of 2.9 years, until 31 May, 2018, 499 (0.4%) had newly developed active TB. Of 499 TB cases, 421 (84.4%) were pulmonary tuberculosis and 78 (15.6%) were extra-pulmonary tuberculosis (see supplementary information). The incidence of TB in contacts was 427 per 100,000 person-years, with an incidence of 1,406 per 100,000 person-years in those who tested positive for LTBI. The overall incidence in contacts was almost seven times higher than that in the general population of the ROK in 2015 (63 per 100,000 persons) 3 , and the incidence in contacts who tested positive for LTBI was 22 times higher than that in the general population.
In contrast to our findings, studies conducted in the US and the Netherlands, which have a low TB incidence and burden, reported a higher TB incidence of 1.2%-3.5% in contacts [12][13][14] . Studies conducted in Taiwan 9 and Hong Kong 10 , which have an intermediate TB burden, similar to that in the ROK, reported incidences of 0.7% and 1.7%, respectively. The relatively lower incidence found in the present study could be attributed to the fact that these previous studies 9,10,12-14 included only close contacts of TB cases or household contacts.
Moreover, the definition of contacts in the ROK is more comprehensive than that used in other countries. Whereas the mean number of contacts per index case was 45 in this study, it was six in studies from the US 12,14 , 15 in a study from the Netherlands 13 , and three in a study from Hong Kong 10 . According to the 2015 ROK national guidelines, the entire school population was included as contacts when two or more TB cases were confirmed within 6 months or when three or more TB cases were confirmed among students in the same year. Although such aggressive contact investigations may be less efficient than investigations of targeted groups of contacts, the prevention of new active TB by finding and treating LTBI cases could be an important TB management strategy in countries with a relatively high TB incidence, such as the ROK 15 . Indeed, the TB incidence has been decreasing each year by 5% since 2011, with the number of TB cases among individuals aged <20 years exhibiting a marked decrease. The number of new TB cases among individuals aged <20 years was 1,501 in 2013, accounting for 4.2% of all cases. However, this number had decreased to 508 in 2018, accounting for 1.9% of all cases. Thus, the percentage of younger patients, who are often exposed to congregate settings, has been decreasing in particular 3 , possibly because of intervention involving contacts in congregate settings.
Of the 499 contacts who developed active TB in this study, 81.0% were diagnosed within 2 years. Of the 163 contacts with LTBI who eventually developed active TB, 84.7% were diagnosed within 2 years. A Canadian study that followed up contacts in congregate settings and household contacts for an average of 6 years reported that 86% contacts who developed active TB were diagnosed within 2 years, similar to our findings 16 . Although differing in terms of the follow-up duration and participants, recent studies, including the present one 13,14,[16][17][18] , have reported that the 2-year incidence of TB in contacts was 63%-94.9%, which is higher than that reported in earlier studies 5,19 . According to the ROK guidelines on follow-up of contacts, follow-up chest radiography was conducted at 3 months and 9 months after exposure in contacts who were negative for LTBI. When contacts who were positive for LTBI were not treated, they underwent follow-up chest radiography every 3 months for 2 years 20 . Considering that 60.5% and 81.0% contacts developed TB within 1 and 2 years, respectively, it would be advisable to develop stratified strategies for prolonging the follow-up duration for each contact category beyond that required by the current ROK guidelines.
The Korean War has often been considered the cause of the high TB incidence in the ROK, despite the developed economic status in this country 21 . It is estimated that many Koreans were infected with TB in the poor post-war environments in the 1950s and 60 s. The LTBI rate in the ROK was 64.2% in 1960, 59.3% in 1975, 44.4% in 1990, and 33.2% in 2016 22 . The number of new TB cases aged ≥65 years has increased annually, accounting for 45.5% of all TB patients in 2018 3 . In the present study, the LTBI rate was 5.9% in individuals aged 0-18 years, 11.2% in those aged [19][20][21][22][23][24][25][26][27][28][29][30][31][32][33][34][35]37.4% in those aged 36-64, and 44.5% in those aged ≥65 years (see supplementary information). The risk of incident TB was higher in those aged ≥65 years than in those aged 0-18 years, possibly because the cases were previously infected with TB and developed active TB as they aged and their immune functions decreased. In contrast, among individuals with LBTI, the risks of incident TB were lower in those aged 36-64 years than in those aged 0-18 years, opposite to the trend observed for the LTBI rate. Here, individuals aged 0-18 years were likely to have been infected with Mycobacterium tuberculosis relatively more recently and were more likely to develop active TB within 2 years.
The risks of co-prevalent and incident TB were higher in healthcare facilities and social welfare facilities than in schools. This may be because most contacts belonging to medical and social welfare facilities have acute or chronic conditions relative to those in other settings. In the present study, the mean age of contacts in these settings was 52.1 ± 18.3 years (see supplementary information), an age group with a relatively high prevalence of diabetes and chronic renal failure [23][24][25] , which are known TB risk factors. However, because we could not confirm the presence of underlying diseases in contacts, follow-up research is required.  26,27 . In the present study, contacts of TB cases with cavitary lung lesions on chest radiographs exhibited greater risks of incident TB. Although the contact investigation undertaken for this study indicated that the presence of cavitary lung lesions on chest radiographs predicted the length of infectivity of index cases, radiographic examination is not currently included among the criteria for conducting investigations. Therefore, we suggest that it would be advisable to include radiography in future criteria.
In is generally accepted that 10% individuals who acquire LTBI will develop active TB in the absence of preventive therapy, with 50% developing the disease within 2 years after exposure 4,16 .
A Dutch study 28 reported that the incidence of TB per 100,000 person-years was 187 in individuals treated for LTBI, 436 in those who did not complete treatment, and 355 in those who were not treated, similar to our findings. In the present study, among individuals aged ≥36 years, the TB incidence per 100,000 person-years was 159 in those who completed chemoprophylaxis and 333 in those who were not treated. Although the incidence was lower in those who completed chemoprophylaxis, the difference was not significant. Because the 2015 ROK national guidelines on TB management only recommend treatment for LTBI in contacts aged <36 years, testing and treatment for LTBI were not as active in contacts aged ≥36 years, which may have influenced the results. More detailed analyses on the effects of treatment for LTBI are required.  The present study had the following limitations. First, we lacked information on underlying diseases, HIV infection, diabetes, nutritional status, body mass index, and smoking, which are recognised TB risk factors, for the contacts. Second, the death of participants during the follow-up period could not be confirmed. Third, because molecular epidemiological tests for index cases and contacts were conducted only in some epidemic cases, the results were not included in the analysis. This limited analysis of the exact route of transmission. Finally, the mean follow-up duration was 2.92 years, which may be too short for the evaluation of active TB development.

Contacts Without Contacts With Unadjusted Adjusted
In conclusion, the present study utilised complete enumeration data obtained from contact investigations conducted in the ROK in 2015 to confirm the development of TB in recent contacts within 2 years and evaluated the effects of the current contact investigations conducted in the ROK. The findings suggest that the completion of chemoprophylaxis for LTBI can lower the risk of TB development in contacts. In future, efforts to increase the contact investigation rates and improve the treatment rates for LTBI are necessary. Furthermore, long-term assessment of the effects of contact investigations is also necessary.   www.nature.com/scientificreports www.nature.com/scientificreports/ contact investigation. According to the Tuberculosis Prevention Act in the ROK, physicians who diagnose TB in private healthcare institutions or public health centres should report this to the public health authority through the KNTSS. TB management staff at public health centres and TB management nurses at private healthcare institutions then conduct case investigations to confirm whether the patients are associated with congregate settings and conduct contact investigation for such cases. KTNESS operates on a web -basis and consists of a patient reporting database, database for TB epidemiological investigation in congregate settings, and households contact investigation database In accordance with the 2015 national guidelines for TB management, contact investigation was conducted when respiratory TB cases with a positive acid-fast bacilli smear test or culture test were reported to be associated with congregate settings, or when more than two TB cases were reported in the same congregate setting within 6 months, regardless of the findings in respiratory specimens 20 . On the basis of on-site investigations, contacts were classified as close or casual contacts; chest radiography and LTBI screening were prioritised for close contacts. Close contacts were those who used the same closed indoor spaces and had direct contact with the index cases for prolonged periods of times. LTBI screening involved a tuberculin skin test (TST) or an interferon-gamma release assay (IGRA). A positive TST result is defined as an induration of ≥10 mm (≥5 mm in newborns that had not yet received the BCG vaccine). The IGRA was performed using the QuantiFERON-TB Gold In-tube (QTF) test, and a value of 0.35 international units or more was deemed positive 20 .
Definitions and risk factors. 'Index patient' was defined as the first person with confirmed TB within the congregate settings. When contacts were reported to have developed TB within 90 days after the initial report of TB in the index case, the disease was considered to be 'co-prevalent' . When contacts were reported to have developed TB after 90 days of the report of active TB in the index case, the disease was considered 'incident' 29 . Risk factors related to the index cases included sex, age, sputum smear microscopy results, and chest radiograph results, while those for contacts included sex, age, the congregate setting, the type of contact, the presence of LTBI, and treatment for LTBI. Congregate settings were categorized into school, workplace, healthcare facility, social welfare facility, and others. Study population. Overall, 133,423 contacts of 2,763 index cases were extracted from the KNTSS.
The following participants were excluded: 15,271 contacts of 152 index cases whose contact investigations initiated before 2015 and continued into 2015, 1,178 contact cases previously diagnosed with TB, 232 household contacts confirmed during the study period, 2 index cases and 2 contacts with duplicate registration. The final sample included 116, 742 contacts of 2,609 index cases (Fig. 1). Of the index cases, 2,594 (99.4%) were pulmonary tuberculosis. ethics approval and consent to participate. The need for written informed consent from participants was waived, based on the Korean Infectious Disease Control and Prevention Act (No.4). This study was conducted in accordance with Korean Infectious Disease Control and Prevention Act and Tuberculosis Prevention Act with permission of KCDC. The study design was approved by the Institutional Review Board (IRB) of KCDC. As this is a retrospective study on the existing data on subjects and based on the Korean Infectious Disease Control and Prevention Act (No.4), we received confirmation of the written consent exemption from IRB. Statistical analyses. The Kaplan-Meier method and log-rank tests were used to compare TB incidences according to the contacts' age. Demographic, laboratory, and clinical determinants (both index cases and contact-related cases) of co-prevalent TB were identified using logistic regression, while those of incident TB were analysed using Cox proportional hazards regression. Multivariate analyses with backward elimination were conducted for the sex and age of index cases and contacts as well as variables with a p-value of <0.05 in univariate analyses. Statistical significance was identified with a 95% confidence interval and a P-value < 0.05. All statistical analyses were performed using SAS version 9.4 (SAS Institute Inc) and Epi-infotm (CDC, Atlanta, GA, USA).

Data availability
All data extracted in this study are included in this article.  Table 6. Incidence rate of tuberculosis among contacts treated by preventive treatment for LTBI in congregate settings. PT, preventive therapy; LTBI, latent tuberculosis infection; TB, tuberculosis; CI, confidence interval. a Incident TB = TB diagnosed within 90 days after diagnosis of the index patient. b Incidence rate in cases per 100,000 person-years observation.