Epidemiological Analysis of Tuberculosis Transmission, Risk Factors, and Subclinical Tuberculosis Management in a High School Outbreak, South Korea

Abstract

Background

Tuberculosis (TB) remains a significant public health concern, particularly in congregate settings such as schools, where adolescents are at increased risk transmission. This study aimed to investigate the epidemiological characteristics, transmission dynamics, and control strategies during a TB outbreak in a South Korean high school.

Methods

A retrospective epidemiological investigation was conducted using data from the Korea Tuberculosis Network and official outbreak reports. A total of 935 individuals—including students, staff, and household contacts—underwent screening through chest X-rays, interferon-gamma release assays (IGRAs), and chest computed tomography (CT). Genotyping of Mycobacterium tuberculosis isolates was performed using spoligotyping and whole-genome sequencing (WGS). Logistic regression identified risk factors associated with TB infection.

Results

Among 935 contacts, 133 (14.2%) tested positive for TB infection. In total, 30 cases of TB disease and 66 cases of latent TB infection were identified among 762 student contacts. Prolonged exposure exceeding 10 hours per week was associated with a significantly increased risk of TB infection (adjusted odds ratio = 5.91, 95% confidence interval: 3.06–11.40, P < .001). Notably, subclinical TB accounted for 74.2% of active TB cases, with most detected via chest CT. WGS and phylogenetic analysis identified a distinct genomic cluster of the Beijing clade, indicating a likely single transmission chain within the school setting.

Conclusions

This outbreak highlights the importance of rapid TB diagnosis, targeted screening for high-risk groups, and advanced diagnostic tools such as IGRA and CT in identifying subclinical cases. Strengthened contact investigations and expanded preventive strategies, including household contacts, are essential for effective outbreak control.

Mycobacterium tuberculosis remains a significant global health challenge. In 2023 alone, ∼10.8 million TB cases and 1.2 million deaths were reported globally according to the World Health Organization (WHO). In South Korea, TB remains a critical public health issue, with 20 000 cases reported in 2023, corresponding to an incidence rate of 38 per 100 000 population, making it the second-highest among Organization for Economic Co-operation and Development countries [1].

Schools and other congregate settings are particularly vulnerable to TB outbreaks due to close and prolonged interpersonal contact. High schools, where adolescents share confined spaces for extended periods, represent an environment with heightened transmission risks. Adolescents are not only susceptible to TB but can serve as effective transmitters of the disease [2]. Furthermore, during adolescence, the risk of developing TB increases, compounded by challenges in timely case detection and effective treatment [3]. Outbreaks in educational institutions have far-reaching consequences, potentially extending beyond the immediate school community to affect families and society while disrupting academic and social activities [4]. In South Korea, the vulnerability of schools to TB outbreaks is evident. Between 2019 and 2023, schools accounted for ∼7%–11% of all TB cases reported in congregate settings [5]. This was further illustrated in 2023 when contact investigations among 6405 individuals aged 10–19 identified 158 latent TB infections (LTBI), highlighting the urgent need for robust preventive measures in school settings [5].

This study aims to examine the dynamics of TB transmission within a high school setting in South Korea, with a specific focus on its epidemiological characteristics, risk factors, and management strategies. Analyzing a recent school-based outbreak, this research seeks to provide actionable insights for strengthening TB prevention and control policies in educational settings.

METHODS

Study Design and Study Population

This retrospective study was conducted using data obtained from the Korea Tuberculosis Network (KTBnet) database, managed by the Korea Disease Control and Prevention Agency (KDCA). The epidemiologic investigation was conducted following TB exposure at a high school. The demographic, epidemiological, and clinical data of individuals exposed—including students, staff, and household contacts—were retrospectively extracted from investigation reports and the KTBnet database. Over a 6-year observation period from 9 February 2018, to 31 December 2023, person-years were calculated based on TB diagnoses in the KTBnet.

The index case in this study was a second-grade male high school student from Chuncheon City, Gangwon Province, South Korea, who had experienced symptoms, including a persistent cough and weight loss, for ∼10 weeks. A chest X-ray (CXR) showed findings indicative of active tuberculosis (TB) with cavitation, and sputum tests confirmed both smear-positive and TB-PCR-positive, with the case officially reported as TB on 15 September 2017. Drug susceptibility testing of the index case confirmed pan-susceptibility to all first-line anti-TB drugs, including isoniazid, rifampin, pyrazinamide, and ethambutol. Based on this result, the patient received the standard 6-month regimen, consisting of 2 months of isoniazid, rifampin, pyrazinamide, and ethambutol during the intensive phase, followed by 4 months of isoniazid and rifampin in the maintenance phase. The study population at risk comprised 935 individuals, including students (n = 762) and staff members (n = 88) from the high school where the index case was identified, as well as family members of patients with TB (n = 85).

Ethics Approval

This study was reviewed by the KDCA Institutional Review Board and was exempt from review (KDCA-2024-07-02_PE_01). Written informed consent was waived because the patients’ records and information were anonymized and de-identified before the analysis. Data were accessible from 24 July 2024, to 30 August 2024, for research purposes.

Outbreak Setting and Contact Investigation

On 19 September 2017, the TB Specialized Epidemiological Investigation Team in KDCA conducted a field epidemiological investigation at the high school in collaboration with the local public health center. The school is a 4-story, closed rectangular building surrounding a central courtyard, with classrooms that are ∼66 m² in size. Each classroom has 3 windows facing the exterior and 2 facing the hallway. The windows facing outside provided sufficient natural light; however, natural ventilation was restricted from July to September due to the air conditioning.

Students participate in designated administrative classes, receiving cultural and health education and career management guidance while attending elective courses for volunteer work or aptitude- and interest-based activities conducted in mixed-grade groups of grades 1–3. In their daily routine, students begin with elective courses for 20–30 minutes before transitioning to assigned subject classrooms for required courses related to college entrance examinations. Later in the day, they return to elective classes for ∼1 hour.

Contacts were classified based on the risk assessment for TB transmission, and testing was conducted according to KDCA guidelines [6]. Students who attended the same elective or administrative classes as the index, rode the same school bus, or spent ≥10 hours per week were categorized as close contacts and underwent onsite CXR and interferon-gamma release assay (IGRA) testing. On the other hand, the remaining second-grade students and teachers were considered casual contacts and initially underwent only CXR. After identifying 2 additional TB cases, screening was extended to include all first and third-grade students, teachers, and staff members (Figure 1). All study populations underwent a series of CXR and IGRA. Comprehensive testing was conducted for all household members if a family member was identified as having TB. TB was diagnosed using CXR, chest computed tomography (CT), and sputum examinations, with chest CTs performed on IGRA-positive students in compliance with KDCA's CT testing standards [6].

Figure 1.

Flowchart for investigating TB contacts associated with a high school TB outbreak, South Korea (n = 762). Group A includes students who were in the same administrative or elective class, rode the same bus, or shared 10 or more hours per week with the index case. Group B comprises all second-grade students, and group C includes first- and third-grade students. Abbreviations: CT, computed tomography; CXR, chest X-ray; IGRA, interferon-gamma release assay; LTBI, latent tuberculosis infection; neg, negative; pos, positive; TB, tuberculosis.

Genotype of M. tuberculosis Strains

To investigate the molecular epidemiological linkages among the 16 culture-positive M. tuberculosis isolates, spoligotyping and whole-genome sequencing (WGS) were performed. Spoligotyping was performed using a commercial kit (Ocimum Biosolutions, Hyderabad, India), according to the manufacturer's protocol, and the resulting patterns were analyzed using the SITVIT2 database. WGS was performed using either the Illumina or Ion Torrent platform. For sequencing with Illumina (Illumina, San Diego, California, USA), libraries were prepared using the Nextera XT Library Prep Kit, and 250 bp paired-end reads were generated with the MiSeq Reagent Kit v2 [7]. Alternatively, the Ion Torrent libraries were constructed using the Ion Xpress™ Plus Library Kit on the AB Library Builder™ System, and sequencing was carried out on the Ion Gene Studio S5 (Thermo Fisher Scientific, Madison, Wisconsin, USA) [8]. All analyses of WGS data were performed by using BioNumerics 7.6.3 (Applied Maths, Sint-Martens-Latem, Belgium). Reference-guided assemblies were created with the Bionumerics 7.6 Reference Mapper v 1.2.3 using the M. tuberculosis H37Rv reference genome (NC00962.3). The single-nucleotide polymorphism (SNP) analysis allowed the construction of the dendrogram using the complete linkage method. A threshold of ≤10 SNPs difference between isolates was used to categorize genomic clusters [9].

Statistical Analysis

All statistical analyses were performed using IBM SPSS Statistics (Version 23.0). Categorical variables were described using frequencies and percentages and compared using χ² or Fisher's exact test. Continuous variables were performed using a median with an interquartile range. Logistic regression analysis was used to identify risk factors associated with TB infection. We included factors with a P-value <.1 from univariable analysis in multivariable logistic regression.

RESULTS

Latent and Active TB Among Contacts of the Index Case

Among the 762 student contacts, 146 were classified as close contacts (group A), with 1 case of TB disease identified through CXR and 32 students testing positive for IGRA. Of these, 9 cases were subsequently confirmed as TB disease through chest CT. Additionally, among 165 casual contacts (group B), 1 case of TB disease was identified. In November 2017, expanded screening, including IGRA testing for all students (group C) and further evaluation of group B, identified 46 IGRA-positive cases. Subsequent CT confirmed 13 additional cases of TB disease. In repeated tests conducted between November 2017 and February 2018, 13 IGRA-positive cases and 4 cases of TB disease were identified among contacts who had initially tested IGRA negative (Figure 1). All students had received BCG vaccinations during childhood and had no other acute or chronic diseases.

Of the at-risk students, the prevalence of TB disease was 3.9% (n = 30), while 8.8% (n = 66) were diagnosed with LTBI, resulting in an overall infection rate of 12.6%. The second grade exhibited the highest rates of TB disease (7.5%) and LTBI (16.0%), leading to an overall infection rate of 23.3%. In contrast, the first and third grades showed lower overall infection rates of 7.2% and 6.5%, respectively. Among the 88 school staff contacts, no cases of TB disease were identified. The LTBI rate was 10.3% (n = 9), with a slightly higher prevalence observed among other staff (20.0%) compared with teachers (3.8%). The 85 household contacts tested exhibited the highest infection rates, with 1.2% (n = 1) diagnosed with TB disease and 33.8% (n = 27) with LTBI, resulting in an overall infection rate of 32.9%. During the 6-year observation period, 2 more active TB cases were detected through KTBnet (Table 1, Figure 2).

Table 1.

Result of Latent and Active Tuberculosis Among Screened Contacts of the Index Case With TB Disease in a High School Outbreak, South Korea

Group	No. at Risk	No. of Contacts Evaluated With IGRA	TB Disease		LTBIa		Infectedb
			n	%	n	%	n	%
Student
1st grade	235	233	4	1.7	13	5.6	17	7.2
2nd grade	266	262	20	7.5	42	16.0	62	23.3
3rd grade	261	255	6	2.3	11	4.3	17	6.5
Subtotal	762	750	30	3.9	66	8.8	96	12.6
School Staff
Teacher	53	52	0	0.0	2	3.8	2	3.8
Other staff	35	35	0	0.0	7	20.0	7	20.0
Subtotal	88	87	0	0.0	9	10.3	9	10.2
Household contacts
Family member	83	78	1	1.2	26	33.3	27	32.5
Othersc	2	2	0	0.0	1	50.0	1	50.0
Subtotal	85	80	1	1.2	27	33.8	28	32.9
Total contacts	935	917	31	3.3	102	11.1	133	14.2

Abbreviations: IGRA, interferon-gamma release assay; LTBI, latent tuberculosis infection; TB, tuberculosis.

^aThe denominator refers to the number of contacts evaluated using IGRA.

^bThose who have TB disease or LTBI; the denominator is the number of contacts fully evaluated.

^cInclude 2 home tutors.

Figure 2.

The epidemic curve representing 32 patients with TB associated with a high school outbreak, South Korea. Case no. 32 was the mother of case no. 2.

Characteristics of the Patients With TB Diseases

The characteristics of the patients with TB were as follows: of the 32 cases, 31 were students, while 1 involved a family member, the mother of case 2. The majority of patients with TB, 29 (93.5%), were diagnosed with pulmonary TB, while 2 patients (6.5%) had extrapulmonary TB. Most patients were asymptomatic, except for the index case, patients 2, 11, and 30 who presented with TB-related symptoms. Notably, 23 patients (74.2%) were diagnosed based on chest CT findings without reported symptoms, except 1 patient who had sputum production. Among the 29 patients who underwent IGRA testing, 25 (86.2%) tested positive. The highest number of cases, 21 (67.7%), occurred among second-grade students, with 7 cases associated with participation in the elective class B of the English Newspaper Study. One patient (case 29) was initially diagnosed with LTBI based on a positive IGRA result, but did not commence treatment. Subsequently, active TB was identified on a follow-up CXR (Table 2). Two patients (cases 30 and 31) initially tested negative on IGRA but were later diagnosed with TB disease during the follow-up period, with both cases showing identical genotypes (Figures 2 and 3). One patient (case 12) developed persistent pleural effusion requiring surgical intervention and was subsequently retreated for TB; no deaths, complications, or post-TB sequelae were identified among the remaining patients. The demographic, microbiological, and radiological findings of patients with CXR-diagnosed TB, CT-diagnosed TB, and LTBI were summarized and compared (Supplementary Table 1).

Table 2.

Characteristics of TB Diseases Identified During a High School Outbreak, South Korea (n = 32)

Case No.	Sex	Age (y)	Administrative Class	Elective Classa	Date Of Diagnosis	TB Site	Symptom	Radiologic Examination		IGRA Result	Sputum Examination			Culture
								Diagnosed by	Presence of Cavity		Smear	PCR	Xpert
1 (index)	M	16	2nd/1	A	2017 Sep 19	Pulmonary	Cough, sputum, weight loss	CXR	pos	NA	2+	1+	NA	pos
2	M	17	2nd/2	B	2017 Sep 30	Pulmonary	cough, sputum	CXR	neg	pos	neg	1+	NA	pos
3	M	17	2nd/8	C	2017 Sep 30	Pulmonary	asx	CXR	neg	NA	neg	neg	NA	neg
4	M	17	2nd/1	D	2017 Nov 06	Pulmonary	asx	Chest CT	neg	pos	neg	neg	neg	neg
5	M	17	2nd/2	B	2017 Nov 06	Pulmonary	asx	Chest CT	neg	pos	neg	neg	NA	neg
6	M	17	2nd/1	A	2017 Nov 06	Pulmonary	asx	Chest CT	neg	pos	neg	neg	NA	pos
7	M	17	2nd/5	E	2017 Nov 07	Pulmonary	asx	Chest CT	neg	pos	neg	neg	NA	neg
8	M	17	2nd/3	F	2017 Nov 07	Pulmonary	asx	Chest CT	neg	pos	neg	neg	NA	pos
9	M	16	1st/2	B	2017 Nov 07	Pulmonary	asx	Chest CT	neg	pos	neg	neg	NA	neg
10	M	17	2nd/2	G	2017 Nov 07	Pulmonary	asx	Chest CT	neg	pos	neg	neg	neg	pos
11	M	16	2nd/2	B	2017 Nov 08	Pulmonary	sputum	Chest CT	neg	pos	neg	neg	NA	pos
12	M	17	2nd/2	G	2017 Nov 08	Pulmonary and extrapulmonary (pleura)	asx	Chest CT	neg	pos	neg	NA	NA	posb
13	M	18	3rd/3	H	2017 Nov 17	Pulmonary	asx	Chest CT	pos	pos	neg	neg	NA	neg
14	M	17	2nd/4	I	2017 Nov 21	Pulmonary	asx	Chest CT	neg	pos	neg	neg	NA	neg
15	M	17	2nd/8	J	2017 Nov 21	Pulmonary	asx	Chest CT	neg	pos	neg	neg	NA	pos
16	M	17	3rd/8	C	2017 Nov 21	Pulmonary	asx	Chest CT	neg	pos	neg	neg	NA	neg
17	M	17	2nd/6	K	2017 Nov 23	Pulmonary	asx	Chest CT	neg	pos	neg	neg	NA	neg
18	M	17	2nd/5	L	2017 Nov 23	Pulmonary	asx	Chest CT	neg	pos	neg	neg	NA	neg
19	M	16	1st/5	M	2017 Nov 23	Pulmonary	asx	Chest CT	neg	pos	neg	neg	NA	neg
20	M	17	2nd/5	E	2017 Nov 23	Pulmonary	asx	Chest CT	neg	pos	neg	neg	NA	neg
21	M	16	2nd/3	N	2017 Nov 23	Pulmonary	asx	Chest CT	neg	pos	neg	neg	NA	neg
22	M	15	1st/2	B	2017 Nov 24	Pulmonary	asx	Chest CT	neg	pos	neg	neg	NA	pos
23	M	16	1st/4	O	2017 Nov 24	Pulmonary	asx	Chest CT	neg	pos	neg	neg	NA	pos
24	M	18	3rd/4	I	2017 Nov 26	Pulmonary	asx	Chest CT	neg	pos	1+	neg	NA	neg
25	M	18	3rd/2	P	2017 Nov 28	Pulmonary	asx	Chest CT	neg	pos	neg	neg	NA	pos
26	M	17	2nd/4	Q	2017 Dec 04	Pulmonary	asx	CXR	neg	neg	neg	neg	NA	pos
27	M	17	2nd/4	O	2017 Dec 04	Pulmonary	asx	CXR	neg	neg	neg	neg	NA	neg
28	M	17	2nd/3	R	2017 Dec 06	Pulmonary	asx	Chest CT	neg	pos	neg	neg	NA	pos
29	M	18	3rd/2	G	2018 Feb 08	Pulmonary	asx	CXR	neg	pos	neg	neg	NA	neg
30c	M	18	3rd/2	B	2018 Apr 24	Pulmonary	cough	CXR	neg	neg	1+	neg	NA	pos
31c	M	17	2nd/2	B	2022 Nov 01	Extrapulmonary (bone)	asx	NAd	neg	neg	neg	NA	NA	pos
32e	F	46	NA	NA	2017 Oct 20	Pulmonary	asx	CXR	neg	pos	neg	neg	NA	pos

Abbreviations: AFB, acid-fast bacilli; asx, asymptomatic; CT, computed tomography; CXR, chest X-ray; IGRA, interferon-gamma release assay; NA, not applicable; neg, negative; pos, positive; TB, tuberculosis; Xpert, Xpert MTB/RIF (Cepheid, California, USA).

^aAlphabet, from A to R, denotes a different elective class that a student attended.

^bPleural fluid.

^cProgression to active TB during the observation period.

^dCXR was unremarkable; diagnosis of TB was confirmed by histopathological examination.

^eMother of case no. 2.

Figure 3.

Molecular analysis of 16 culture-positive Mycobacterium tuberculosis strains. All isolates were classified within the Beijing clade and formed a single genetic cluster with 0–4 SNP differences.

Genotypic Analysis of M. tuberculosis Strains Among 16 Patients With TB Diseases

All 16 culture-positive M. tuberculosis isolates, including the isolate obtained from the family member of case 2, were classified within the Beijing clade based on spoligotyping analysis. The results of wgSNPs analysis showed that all isolates from patients differed by 4 or fewer SNPs, forming a distinct genomic cluster (Figure 3). WGS and phylogenetic analysis, together with strong epidemiologic links, suggested that the cases were likely part of a single transmission chain.

Risk Factors Associated With TB Infection Among Students

Second-grade students had a higher risk than first-grade students (adjusted odds ratio [OR] = 2.19, 95% confidence interval [CI]: 1.15–4.20, P = .018). Prolonged exposure emerged as the strongest predictor, with students spending ≥10 hours per week with the index case showing nearly 6-fold higher odds of infection (adjusted OR = 5.91, 95% CI: 3.06–11.40, P < .001) (Table 3). Similarly, prolonged exposure of >10 hours per week with the index also significantly increased the risk of TB disease (OR = 5.65, 95% CI: 2.13–15.02, P = .001) and LTBI (OR = 5.94, 95% CI: 3.02–11.69, P < .001; Supplementary Tables 2 and 3). Exposures such as being in the same administrative class, riding the same school bus, or attending the same elective class as the index case were not statistically significant risk factors for TB infection in either univariable or multivariable analyses.

Table 3.

Risk Factors Associated With TB Infection Among Students in Multivariable Logistic Analysis (n = 762)

Variables	Students Infected With TB		χ2	P-Valuea	Univariable			Multivariableb
	No n (%)	Yes n (%)			Odds Ratio	(95% CI)	P-value	Odds ratio	(95% CI)	P-value
Grade
1st grade	218 (92.8)	17 (7.2)	42.63	<.001	Referent			Referent
2nd grade	204 (76.7)	62 (23.3)			3.90	(2.21–6.89)	<.001	2.19	(1.15–4.20)	.018
3rd grade	244 (93.5)	17 (6.5)			0.89	(.45–1.79)	.751	0.87	(.43–1.76)	.708
The same administrative class with the index
No	641 (87.8)	89 (12.2)	2.61	.107	Referent
Yes	25 (78.1)	7 (21.9)			2.02	(.85–4.80)	.113
School bus rider
No	622 (87.5)	89 (12.5)	0.06	.802	Referent
Yes	44 (86.3)	7 (13.7)			1.11	(.49–2.54)	.802
The same elective class with the index
No	652 (87.8)	91 (12.2)	3.33	.079	Referent			Referent
Yes	14 (73.7)	5 (26.3)			2.56	(.90–7.27)	.078	0.96	(.22–4.11)	.957
Exposure time (per wk)
< 1 h	560 (91.2)	54 (8.8)	96.52	<.001	Referent			Referent
1–3 h	42 (93.3)	3 (14.3)			0.74	(.22–2.47)	.625	0.71	(.21–2.38)	.577
4–6 h	17 (89.5)	2 (10.5)			1.22	(.27–5.42)	.794	0.69	(.15–3.19)	.637
7–9 h	13 (81.3)	3 (18.8)			2.39	(.66–8.66)	.184	2.71	(.51–14.45)	.243
≥10 h	34 (50.0)	34 (50.0)			10.37	(5.98–18.00)	<.001	5.91	(3.06–11.40)	<.001

Abbreviations: CI, confidence interval; IGRA, interferon-gamma release assay.

^aχ² test.

^bVariables with P–value <.1 in univariable analysis were included in the multivariable model.

Treatment Results of Contacts Diagnosed With LTBI

Among the 102 contacts diagnosed with LTBI, 98 (96.1%) initiated prophylactic treatment, and 84 (85.7%) completed the regimen. Of the 66 students who began LTBI treatment, 62 received a 3-month regimen of rifampin plus isoniazid, 2 received a 9-month regimen of isoniazid monotherapy, and 2 received a 4-month of rifampin monotherapy. A total of 56 students (84.8%) completed treatment, while 10 discontinued, including 5 due to adverse effects and 3 due to personal choice. Among the 9 staff members, 8 initiated LTBI treatment with a 3-month regimen of rifampin plus isoniazid, and 6 completed the treatment. Two staff members discontinued therapy due to side effects. Of the 27 family members, 24 initiated treatment, with 20 receiving a 3-month regimen of rifampin plus isoniazid, 2 a 9-month isoniazid regimen, and 2 a 4-month rifampin regimen. Among them, 22 family members (91.7%) completed the regimen. Two additional TB cases were reported over the total observation period of 5447 person-years, resulting in an incidence rate of 3.67 per 10 000 person-years (Supplementary Table 4).

DISCUSSION

This study presents a detailed investigation of a high school outbreak, highlighting its strengths: a 6-year follow-up of 935 individuals, integration with the national TB surveillance database, and the application of advanced genotyping methods such as spoligotyping and WGS. Leveraging these tools, the study conducted a rigorous epidemiological investigation, identifying 102 LTBI and 31 additional TB disease cases. The study results also uncovered significant risk factors, such as ≥10 hours of weekly contact with the index case, which was associated with nearly a 6-fold increase in the risk of infection (OR = 5.91, P < .001).

The prevalence of M. tuberculosis infection among students in our study was 12.6%, which falls within the wide range reported in other studies using the IGRA test, with a pooled prevalence of 33.2% (95% CI: 0.0%–73.0%). This substantial variability may be influenced by factors such as study location, characteristics of the index case, school type, and differences in diagnostic methods [10].

A increased risk of transmission due to the following contributing factors resulted in 133 TB infections identified in this outbreak. Firstly, the index patient exhibited persistent cough and significant weight loss for >10 weeks but was misdiagnosed with bacterial pneumonia, resulting in a critical delay in TB identification and isolation. Likewise, in an Italian school, prolonged symptom duration ranging from weeks to months among teachers and students led to 10 confirmed TB cases and an 8.3% prevalence of LTBI among their contacts [11]. Similarly, a high school student with smear-positive TB remained undiagnosed for ∼5 months, culminating in the identification of 14 additional TB cases in China [12]. Secondly, errors in the initial interpretation of CXRs contributed to delayed diagnoses. While all TB-exposed individuals were initially assessed as having normal CXRs, a subsequent review by the Korean Society of Radiology revealed that 3 cases had already progressed to pulmonary TB at the time of the initial screening. These errors highlight the inherent limitations of CXR interpretation, including subjective variability and interobserver inconsistency [13]. In mass CXR screening, as exemplified by the large cohort of students in this study, incorporating AI-driven diagnostic tools and computer-aided detection systems coupled with standardized interpretation protocols may be a complementary tool in enhancing diagnostic accuracy and reliability [14, 15]. Thirdly, the classroom system, which combines teaching and mentoring, allows students to choose their subjects and transition between classes. This arrangement, which involves intermingling first-, second-, and third-grade students, is thought to have enhanced the transmission in school. Lastly, the transmission period overlapped with Korea's hot and humid summer months, during which windows were often kept closed and air circulation was restricted due to air conditioning. This environmental condition may have contributed to transmission and should be considered when assessing ventilation and infection control measures in school settings. Taken together, despite the relatively large physical dimensions of the classroom, the combination of insufficient ventilation, high infectiousness of the index case, and prolonged, repeated exposure likely contributed to extensive airborne transmission.

Multivariable regression analysis identified prolonged exposure as the most significant risk factor for TB infection, highlighting the critical influence of exposure duration and intensity in M. tuberculosis transmission. Students with 10 or more hours of weekly contact with the index case exhibited markedly higher risks of IGRA positivity and disease progression, consistent with prior studies that showed extended periods of close contact, such as in classrooms or dormitories, further elevated transmission risk [11, 12, 16]. Similarly, attending multiple shared classes with an index patient significantly increased the likelihood of LTBI in the United States [17, 18]. Interestingly, shared bus rides did not independently predict TB infection in our study, diverging from findings in a Missouri high school where bus rides were identified as a risk factor [17]. Distinct from other TB outbreak epidemiological studies, our study quantified class overlap with the index patient and carefully measured exposure time, which enabled a more precise assessment of total contact duration for each individual. Of note, in our multivariable analysis, mere co-presence in the same location as the index patient did not emerge as a significant risk factor for TB transmission. These findings underscore the necessity of incorporating detailed exposure metrics, such as contact duration and interaction intensity, into TB control strategies to more accurately identify high-risk individuals and optimize intervention measures in outbreak settings.

The lower infection rate among staff relative to students aligns with findings from TB outbreak studies in Chinese schools [11, 16], suggesting differences in exposure patterns or immune response influenced by age or work-related behaviors. In contrast, high LTBI rates among household contacts highlight the need for enhanced family-based interventions. Given the heightened risk of in-house TB and community transmission [12, 18], robust household contact investigations and preventive measures—such as early screening, LTBI treatment, and infectious control education—are essential to complement school-centered measures.

In our study protocol, high-resolution chest CT scans were performed on all students with positive IGRA results to differentiate TB disease from LTBI. CT identified 23 out of 31 (74.2%) additional TB cases, underscoring its superior sensitivity in detecting early or subtle pulmonary abnormalities compared with conventional CXRs. This finding raises critical questions regarding current TB definitions and diagnostic criteria, particularly as adolescents frequently exhibit atypical or subtle radiologic findings that challenge the traditional binary classification of active versus latent TB. The International Consensus for Early TB (ICE-TB) framework defines subclinical TB as the presence of detectable macroscopic pathology without infectiousness, with any accompanying symptoms either unnoticed or unacknowledged [19]. However, as our study demonstrates, attributing these radiologic abnormalities to viable M. tuberculosis remains challenging when microbiological examinations yield negative results, emphasizing the need for operational tests or biomarkers.

CT's high-resolution imaging facilitates the identification of subtle parenchymal changes—such as nodules, cavitations, or lymphadenopathy—that may be missed on conventional CXRs [20, 21]. Early detection of such changes may contribute to a reduction in the reservoir of undetected TB and, consequently, to the limitation of disease progression and secondary transmission. This is further supported by our observed incidence rate of 3.67 cases per 10 000 person-years, significantly lower than the 42.8 cases reported among household contacts [22] and the 79.0 cases reported among individuals with LTBI during a school outbreak [23]. Nevertheless, the widespread application of chest CT necessitates careful consideration of associated costs, radiation exposure [24], and overall feasibility in large-scale screening, suggesting that CT screening may be best reserved for cases with a high risk of TB transmission or significantly elevated IGRA levels.

In our study protocol, all students diagnosed with subclinical TB were treated with a routine anti-TB regimen. This approach was based on the plausible hypothesis that early treatment might limit further pathological damage, resolve chronic inflammation, and prevent subsequent illness, post-TB sequelae, or death. Moreover, timely detection and treatment may reduce community transmission of M. tuberculosis [25]. However, the clinical relevance of treating subclinical TB remains controversial. Radiological abnormalities detected on CT can sometimes represent changes following a recent primary infection that may not progress to active disease over long-term follow-up [26], thus raising questions about the necessity and impact of treatment in such cases. These uncertainties underscore the need for further research to clarify the individual and public health benefits of treating subclinical TB and optimize treatment regimens, particularly in outbreak settings. Future guidelines should weigh the advantages of improved diagnostic accuracy achieved through advanced imaging against the potential risks and resource implications.

This study has several limitations that warrant consideration. Firstly, as a retrospective analysis, it relied on data from a single high school outbreak, limiting the generalizability of the findings to other settings or populations. The specific environmental, social, and institutional characteristics of the studied school may not reflect those of other schools or congregate settings. Secondly, while chest CT was extensively used and provided valuable diagnostic insights, the reliance on advanced imaging techniques raises concerns about the feasibility and cost-effectiveness of such an approach in routine outbreak management, particularly in resource-limited settings. Finally, the study did not include a detailed assessment of socioeconomic, behavioral, or immunological factors that could influence TB infection and progression. Understanding these factors could have provided deeper insights into the observed patterns and risk factors.

In conclusion, this study identifies prolonged close contact as a significant risk factor for TB transmission in schools, highlighting the need for targeted screening and early detection strategies. The high proportion of subclinical TB cases detected through chest CT underscores the importance of utilizing advanced imaging for high-risk contacts, particularly IGRA-positive individuals. Additionally, given the notable rate of LTBI among household contacts, expanding contact investigations beyond the school setting could enhance outbreak control. Moving forward, optimizing diagnostic protocols, refining definitions of subclinical TB, and tailoring prevention strategies for high-risk groups will be pivotal in enhancing TB control measures and reducing the disease burden in educational settings.

Supplementary Data

Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.