Rifampicin (RIF) is an important first-line anti-tuberculosis (TB) drug with excellent bactericidal and sterilizing activities [1]. The addition of RIF to isoniazid (INH)-based therapy reduced the treatment duration by half, from 18 to 9 months, while decreasing the risk of treatment failure and acquired drug resistance [2]. When a new drug is introduced, there is always a concern about acquiring resistance. Multidrug-resistant tuberculosis (MDR-TB), defined as resistance to INH and RIF, emerged as a threat to TB control worldwide in the early 1990s [3]. MDR-TB was the first infectious disease to alert national authorities worldwide regarding the importance of antimicrobial resistance as a future public health challenge. Resolving RIF resistance is crucial to achieving the targets of the World Health Organization (WHO)’s End TB Strategy.

Rapid diagnosis for early containment of transmission and the use of appropriate anti-TB drugs to improve treatment success rates are essential to effectively control MDR-TB. The development of rapid molecular drug susceptibility tests (DSTs) such as the Xpert MTB/RIF assay (Cepheid, Sunnyvale, CA, USA), which identifies genes responsible for RIF resistance, has shortened the time required for diagnosing RIF resistance. The WHO highlights the importance of universal access to DST, which is defined as rapid DST for at least RIF among all people with bacteriologically confirmed TB [4]. Further DST for at least fluoroquinolone and second-line injectable drugs among all TB people with RIF resistance is also recommended to individualize appropriate treatment regimens for MDR-TB. However, their implementation faces global challenges owing to cost constraints in all countries.

A recent meta-analysis revealed that the global pooled prevalence of MDR-TB was 11.6% [5]. However, the clinical profiles of MDR-TB vary depending on the TB epidemiology in each country. Additionally, there is a lack of research on rifampicin-monoresistant tuberculosis (RMR-TB). National surveillance of RIF resistance among TB cases is necessary to understand the clinical characteristics for clinicians and policymakers to implement appropriate public health interventions. The current study aimed to identify the clinical characteristics of people with MDR/rifampicin-resistant tuberculosis (MDR/RR-TB) and to determine the presence of concurrent resistance to other anti-TB drugs.

Between July 2018 and December 2021, 24,447 TB cases were registered in the Korean National Registry (Figure 1). Among the 6,154 TB cases identified using the Xpert MTB/RIF assay, 2.7% (165 cases) were resistant to RIF and 0.5% (33 cases) had indeterminate RIF susceptibility results. Among the 165 people identified with RIF resistance through the Xpert MTB/RIF assay, 30.9% (51 RR-TB cases) either did not undergo further phenotypic or molecular DSTs or had erroneous results, rendering the determination of resistance to other drugs unfeasible. Based on phenotypic and molecular DST results, we identified 159 RMR-TB and 292 MDR-TB cases.

Among the 24,447 TB cases, 43.6% (10,656 cases) were PanS-TB cases and 502 (2.1%) were MDR/RR-TB cases (Figure 2A). When only cases with known DST results were considered, the proportions of PanS-TB and MDR/RR-TB were 70.7% and 3.3%, respectively (Figure 2B). Among the 502 MDR/RR-TB cases, 31.7% (159 cases) were RMR-TB, 58.1% (292 cases) were MDR-TB, and 8.4% (42 cases) were fluoroquinolone-resistant TB (Figure 2C). Additionally, of the 292 MDR-TB cases, 19.2% (56 cases) had unknown susceptibility to fluoroquinolones.

The proportions of MDR/RR-TB among new and previously treated patients were 2.8% and 6.7%, respectively (Table 1). Treatment after loss to follow-up comprised the highest proportion (9.2%). The proportions of MDR/RR-TB among new patients decreased significantly decreased as age increased (<45 years, 4.1%; 45−64 years, 2.9%; ≥65 years, 2.1%; p<0.001). Its proportions among previously treated patients also significantly decreased as age increased (<45 years, 14.5%; 45−64 years, 6.9%; ≥65 years, 4.3%; p<0.001).

Univariable analyses were conducted to describe the clinical profiles of MDR/RR-TB cases and revealed that younger age, low body mass index, diabetes, foreigners, and prior TB history were significantly associated with MDR/RR-TB (Table 2). However, only younger age (<45 years: adjusted odds ratio [aOR]=1.736; 95% confidence interval [CI], 1.357 to 2.220; 45−64 years: aOR=1.253; 95% CI, 1.002 to 1.566), foreigners (aOR=3.151; 95% CI, 2.318 to 4.284), and prior TB history (aOR=2.921; 95% CI, 2.370 to 3.601) remained significant based on multivariable logistic regression analyses. Compared to 2021, the calendar year 2018 (aOR=2.407; 95% CI, 1.839 to 3.149) and 2019 (aOR=1.370; 95% CI, 1.056 to 1.777) were significantly associated with MDR/RR-TB.

After categorizing MDR/RR-TB cases based on INH resistance, we compared the clinical characteristics of MDR-TB, RMR-TB, and RR-TB cases (Table 3). Among all variables examined, only patient nationality showed significant differences. The proportion of foreigners was highest (15.1%) among MDR-TB cases and lowest (6.9%) among RMR-TB cases.

We also compared resistance to other anti-TB drugs between MDR-TB and RMR-TB (Table 4). The proportion of resistance to other first-line drugs such as ethambutol (54.7% vs. 14.8%, p<0.001) and pyrazinamide (30.9% vs. 12.3%, p=0.001) was significantly higher in the MDR-TB group than in the RMR-TB group. Resistance to rifabutin was similar between people with MDR-TB and those with RMR-TB (67.8% vs. 67.9%, p=0.986). MDR-TB was significantly associated with fluoroquine resistance (17.8% vs. 7.4%, p=0.024).

This study confirmed the current status of RIF resistance among notified TB cases based on the most recent nationwide data in Korea. We compared the clinical characteristics of RMR-TB and RR-TB with those of MDR-TB. The proportions of MDR/RR-TB cases among all TB cases and TB cases with known DST results were 2.1% and 3.3%, respectively. Younger age, foreigners, and a prior history of TB were significantly associated with MDR/RR-TB. Among all cases with identified RIF resistance, approximately one-third were RMR-TB and one-tenth were RR-TB. The clinical profiles were similar for MDR-TB, RMR-TB, and RR-TB except for nationality. This study could serve as important evidence for the development of future plans to combat RIF resistance and provide useful insights for recommending the optimal use of rapid molecular DST and the composition of appropriate treatment regimens in Korea.

Korea has experienced a significant reduction in the number of TB cases after achieving enormous economic development and implementing universal health coverage, leading to a decline in the occurrence of drug-resistant TB. The first nationwide TB prevalence survey, with the assistance of the WHO, was conducted in 1965, followed by six additional surveys until 1995 [11,12]. According to these surveys, the drug resistance rate remained relatively stable from 1965 to 1975 at approximately 38%, peaked at 47.5% in 1980, and decreased to 9.9% by 1995. Subsequent to national drug resistance surveys [13], which reported MDR-TB strains occurring in 3.4% of cases in 1994 and 3.8% in 2004, there have been no official nationwide data available on MDR-TB status. Prior to this study, Lee et al. [14] analyzed data from 4,417 cases undergoing phenotypic DST at eight university-affiliated hospitals in Korea from 2015 to 2018 and reported an MDR-TB rate of 4.1%. Our study collected nationwide data from 2018 to 2021 and confirmed an MDR-TB rate of 1.9% according to DST results, indicating a lower rate than in previous years. This decline correlates with the reduction in the number of nationwide notified MDR-TB cases since 2012 [15], reflecting improved patient management based on the public-private partnership in Korea [16] and the treatment success rate, likely attributed to the introduction of novel anti-TB drugs [17].

According to the multivariable regression analysis in the current study, the proportion of MDR-TB was significantly higher in 2018 and 2019 than in 2021, but there was no statistical difference in 2020. This suggests that there was no reduction in MDR-TB incidence between 2020 and 2021. Although the total number of notified MDR-TB cases in Korea decreased from 580 in 2019 to 399 in 2020, it has remained relatively stable, with 371 notified cases in 2021 and 368 in 2023, indicating no significant decrease in recent years. This may be ascribed to the COVID-19 pandemic, which significantly disrupted healthcare systems worldwide [18]. Korea has also observed that the decline in the TB incidence rate has slowed since 2022 despite the relatively well-maintained healthcare system during the pandemic. It is challenging to maintain sustained MDR-TB support and care because of the long-term management from initial diagnosis to treatment completion. To comprehensively manage people with MDR-TB, the MDR-TB consortium was established in Korea in 2020 [7]. In addition, it is critical to prevent acquired RIF resistance through the comprehensive management of drug-susceptible TB. In 2023, the Third National Strategic Plan for TB Control was announced, which emphasized meticulous patient management for both drug-resistant and drug-susceptible TB. Securing financial support and attracting public attention are essential to ensure the successful implementation of these policies in the near future.

Our study revealed that younger individuals were significantly associated with MDR/RR-TB compared to the elderly population and the proportions of MDR/RR-TB were high at a younger generation in both new and previously treated case, which were similar to the prior Korean study [14]. It indicates possible ongoing direct person-to-person MDR strain transmission and acquired RIF resistance. Although the WHO has recommended TB preventive therapy for contacts with MDR/RRTB since 2017, its implementation has been poor due to a lack of evidence. Based on recent findings from clinical trials [10], TB preventive therapy using 6 months of levofloxacin for identified contacts of RIF-resistant cases can be considered to prevent person-to-person transmission. Considering high RIF resistance rate in relapse and treatment after loss-to-follow-up patients in this study, active patient management should also be implemented in young people during initial anti-TB treatment to prevent acquired RIF resistance. Another explanation is that RIF began to be used in Korea in the late 1970s and became widely used in the mid-1980s. Elderly people who were infected with RIF-susceptible M. tuberculosis complex before this period might develop TB later in their life, which result in lower rates of RIF resistance compared to younger people.

Foreign-born nationality was significantly associated with MDR/RR-TB in the present study. In Korea, the proportion of foreigners among all notified TB cases was 5.3% in 2022 [19], showing no change in the past 5 years, but it increased to 5.7% in 2023. This increase can be attributed to the influx of foreigners following the lifting of the COVID-19 response measures. Although foreign population accounted for only 3.5% of the total population in Korea in 2022, immigration from the high TB burden country, such as China, Viet Nam, Uzbekistan, and Philippines, is expanding to solve the labor shortages caused by the aging population. Strengthening TB management for foreign entrants was a key task of the Third National Strategic Plan for TB Control in Korea. Therefore, it is necessary to enhance TB screening and expand treatment support for foreigners [20].

In this study, we diagnosed RIF resistance using the Xpert MTB/RIF assay and phenotypic and molecular DSTs. The total number of MDR/RR-TB cases in this study was 502, with 317 cases diagnosed using phenotypic DST. Therefore, 185 MDR/RR-TB cases were identified using the molecular DST and Xpert MTB/RIF assays. Phenotypic DSTs are only feasible for culture-positive TB cases, rendering them impractical for culture-negative cases. Moreover, there are instances where phenotypic DST cannot be conducted, or where contamination prevents accurate results. Approximately 10% of cases in this study were bacteriologically confirmed to have unknown drug susceptibility. Considering clinically diagnosed cases, this implies that approximately 40% of all TB cases do not show drug susceptibility results. Approximately 3% of these cases could potentially be RR-TB, underscoring the importance of monitoring and assessing anti-TB treatment outcomes in these patients. Active and vigorous efforts to detect RIF resistance through currently available DSTs are crucial, especially for the high-risk populations identified in our study, such as younger generations, foreign-born individuals, and those with a history of anti-TB treatment.

In 2014, it was estimated that only 1.1% of individuals globally diagnosed with TB carried RIF resistance without concomitant INH resistance [21]. The prevalence of RMR-TB was estimated to increase from 12% in all MDR/RR-TB cases in 2014 to 22% in 2019 [10]. In Korea, where universal DST is feasible, an accurate diagnosis of RMR-TB is achievable. Our rapid narrative review revealed that RMR-TB in Korea occurred at a steady rate of approximately 1% among cases performed with DST, despite a decrease in MDR-TB during the last decade. Considering the notably higher proportion of retreatment cases among RMR-TB cases than among PanS-TB cases (33.3% vs. 11.7%, p<0.001) in our study, it can be speculated that prior anti-TB treatments may have induced RIF monoresistance. In addition, spontaneous mutations in chromosomal genes at a frequency of 10⁻⁶ to 10⁻⁸ mycobacterial replications [22] may still occur in retreatment cases, leading to genetic resistance to anti-TB drugs, such as RIF. For example, prior studies [23,24] have reported that rpoB gene mutations causing low-level or borderline RIF resistance are more frequently detected in RMR-TB than in MDR-TB. Recent technological advancements may lead to the increased detection of such strains using both phenotypic and molecular DSTs. While RIF monoresistance theoretically implies INH susceptibility, allowing for its inclusion in effective drug regimens, current guidelines classify RMR-TB as MDR-TB, resulting in unnecessary prolongation of treatment duration and exposure to potential avoidable side effects. Understanding the frequency and mechanisms of RMR-TB is crucial for formulating treatment strategies, and further research is required to verify the efficacy of treatment regimens.

Fluoroquinolone is a pivotal drug in MDR-TB treatment and classified as one of three ‘group A’ anti-TB drugs. Fluoroquinolone resistance is associated with poor outcomes in patients with MDR-TB. Therefore, early identification of fluoroquinolone resistance after MDR/RR-TB diagnosis and subsequent adjustment of treatment regimens are crucial. Among the MDR/RRTB cases tested for fluoroquinolone resistance, 18% (95% CI, 16% to 20%) were estimated to be resistant to any fluoroquinolone globally by 2022 [25]. On examining nationwide data from Korea, the proportion of fluoroquinolone-resistant MDR-TB cases decreased from 34.2% in 2011 to 19.3% in 2015 [26]. Other multicenter studies have indicated that 26.2% of 633 MDR-TB cases in 2010-2019 [27] and 20.1% of 179 cases in 2015-2018 [14] were found to be fluoroquinolone-resistant. In our analysis, fluoroquinolone-resistant MDR-TB cases between 2018 and 2021 accounted for 15.1% of all cases, indicating a declining trend in fluoroquinolone resistance. However, there are still people with MDR/RR-TB whose fluoroquinolone susceptibility remains unknown. Therefore, it is important to provide and support strategies for conducting additional tests to confirm resistance to second-line anti-TB drugs, including fluoroquinolones, especially when RIF resistance is confirmed.

This study had several limitations. First, our study population was recruited from PPM project-participating hospitals in Korea, which oversaw 80% of the notified TB cases across the country. Many hospitals participating in the PPM project are affiliated with universities, suggesting that the TB cases analyzed in this study may have a higher likelihood of being difficult cases, such as drug-resistant TB. Therefore, the drug resistance rate may have been overestimated. Second, we utilized molecular DST results to diagnose RIF resistance and were unable to collect data on the specific locations of rpoB mutations. There are instances where discordant results of phenotypic and molecular DST exist, and further genome sequencing is necessary to identify the causes of false positivity and borderline resistance. Future research should focus on comparing various DST methods and identifying and comparing the prevalence of RIF resistance using various DST methods.

In conclusion, the proportions of MDR/RR-TB cases among all TB cases and TB cases with known DST results were 2.1% and 3.3%, respectively. Despite a reduction in the number of MDR-TB cases in recent years, RMR-TB has occurred at a steady rate, resulting in approximately one-third of all cases of RIF resistance. It is important to actively screen for RIF resistance by targeting the high-risk populations identified in our study, such as younger generations, foreign-born individuals, and those with a prior history of anti-TB treatment.