The incidence rate of tuberculosis (TB) rose by 3.6% between 2020 and 2021, reversing years of decline for the previous 2 decades according to the 2022 WHO World Tuberculosis Report (1). Unfortunately, the burden of multidrug-resistant TB (MDR-TB) is also estimated to have increased between 2020 and 2021. Globally, the rate of MDR-TB among patients with new TB was about 3.6% and the best estimate for people who have previously been treated for TB is about 18% in 2021. Drug-susceptible TB and drug-resistant TB are spread in the same way (2). However, treatment of MDR-TB infections is much more complex, less effective, more toxic, and more expensive than treating patients infected with susceptible TB strains (3). Reducing the incidence of MDR-TB infection in the community may be difficult due to a lack of knowledge about the risk factors associated with developing MDR-TB.

The mechanisms that lead to the development of drug resistance are complex, and resistance may have multiple causes even within the same individual patient. Previous studies (4–7) have identified risk factors for developing anti-TB drug resistance. Some studies have suggested that drug resistance may be associated with prior exposure to anti-TB drugs in insufficient amounts to achieve complete cure (8, 9). Other studies have suggested that race, malnutrition, and pulmonary TB may be risk factors for developing drug-resistant TB (10). The onset of MDR-TB can be related to various factors, including race, socioeconomic status, culture, lifestyle, epidemiology, medical condition, and detection technology (10). Some comorbidity studies have shown that people with human immunodeficiency virus (HIV) infection or diabetes are more likely to develop MDR-TB (11–13).

However, few studies have identified the comorbidities or medical conditions associated with the development of drug-resistant TB (12, 13). Moreover, with the exception of HIV and diabetes, it is unclear whether the risk factors associated with TB already act as risk factors for developing drug-resistant TB (14, 15). We investigated the risk factors for the occurrence of MDR-TB in patients with pulmonary TB using a national population-based database.

This study investigated the risk factors for the occurrence of MDR-TB in patients with pulmonary TB. As a result of multivariate analysis, younger adults and males were significantly associated with MDR-TB. While patients with Psoriasis and Rheumatoid arthritis showed low risk of MDR-TB, diabetes mellitus and solid organ transplantation showed no overt associations with MDR-TB.

We found that a prior history of treatment with anti-TNF agents was a significant risk factor for MDR-TB. Treatment with TNF inhibitors is a well-known risk factor for the reactivation or remote acquisition of TB (16).

TNF-α is known to be associated with the function of the immune system of TB patients. Some studies demonstrated that the TNF-α level in peripheral blood cells could differentiate active TB from non-TB patients and strong TNF-a responses at TB diagnosis were significantly decreased after treatment (17, 18). Interestingly, after TB treatment, the levels of TNF-α remained elevated in the sputum-positive population and in extensively drug resistant TB patients contrary to sputum- negative group and sensitive TB patients (19, 20).

TNF-α has a dual effect in TB and acts as host resistance and susceptibility factor (21, 22). Physiological levels of TNF-α have a defensive role against mycobacterial infection. TNF-α is important for phagosome activation to kill mycobacteria and granuloma formation to prevent dissemination (23). Elevated TNF-α levels in some subgroups after treatment may be indicative of the host’s immune status to control the pathogen.

Kviatcovsky et al. reported that MDR strain M inhibited TNF-α secretion influencing on neutrophil activation and ROS production, as part of evasion mechanism to persist in the host (24). Several studies found the distinct difference in immune responses and the immune cellular subpopulation between MDR TB and sensitive TB (25, 26). Drug resistant TB can occur as primary drug resistance or acquired drug resistance happened during failed treatment of drug-susceptible TB (27). In the latter cases, bacterial subpopulation known as persisters can be phenotypically tolerant to anti-TB medication. Evidence suggests drug resistant M. TB strains interact with phagocytes and immune cells in the lung microenvironment and modify to adapt and persist in the host (23).

Therefore, it is reasonable to assume that TNF-α inhibitors suppress cytokine expression and alter the immune microenvironment and would develop resistant TB or advanced forms of TB. Our results, in which the occurrence of MDR-TB was more frequent in the case of longer use of infliximab, support the hypothesis that the risk of developing resistant tuberculosis might increase when TNF-α is continuously suppressed.

There were many different findings from previous studies identifying risk factors for MDR-TB (4, 5, 13, 29). In the present study, the differences in MDR-TB risk between the types of anti-TNF agents were not clear, in contrast to previous studies showing that the risk for TB is higher in patients receiving anti-TNF monoclonal antibody therapy than those receiving soluble TNF receptor therapy (28, 29). Wallis et al. (30) suggested a time-varying risk for TB in patients with anti-TNF therapy with a higher early increased risk for infliximab than for etanercept. Similarly, in our study, the average duration of drug therapy before TB diagnosis was higher for etanercept than infliximab and adalimumab.

Previous studies (4, 5, 13, 31) have reported that MDR-TB risk is mainly related to previously diagnosed TB and anti-TB therapy rather than sociodemographic characteristics, such as age or sex, comorbidities such as HIV or chronic obstructive pulmonary disease (COPD), lifestyle factors such as alcohol abuse, and other TB-related characteristics, such as directly observed therapy. The results of meta-analyzes have been from pooled studies that differed in design, patient demographic characteristics, and definitions of MDR-TB, resulting in a large degree of heterogeneity (4).

Age and sex may be related to TB drug adherence. Interestingly, in our results showed that more men and more ICS users are included in non-naïve MDR-TB patients than naïve MDR -TB patients. ICS use appears to be a risk factor for MDR-TB in terms of the decreased immune response, similar to the pathogenesis of pneumonia or non-MDR-TB (4). Further research is needed to determine the correlation between ICS exposure and emergence of drug resistance in TB.

Our data showed that while rheumatoid arthritis and psoriasis were associated with lower risk for MDR-TB, the percentage of patients who had been treated with etanercept for rheumatoid arthritis RA was higher in MDR TB patients than those without MDR-TB. Due to data limitations, it was not possible to ascertain the individual’s immune status, latent TB and latent TB prophylaxis. Given the results of the United Kingdom study showing that incidence of tuberculosis varies with the severity of psoriasis, immune status including TNF-α and IL-17 -induced inflammation may affect the MDR-TB risk (32). Further research taking into account these variables may clarify the association with MDR-TB risk.

Our findings could help improve the rate of successful TB treatment and help prepare for immediate radical treatment by actively observing pulmonary TB patients with early suspicion of MDR-TB.

This study had several limitations. First, the study used the claims data from the NHI in Korea. As the administrative claims data were designed for reimbursement purposes, they do not include sensitive information for privacy reasons, such as socioeconomic status, occupation, nutritional status, or medication adherence. Second, retreatment and recurrence were excluded as index TB infections after confirmation by observation from 1–2 years to 7 years, but recurrence after infection decades ago was not confirmed due to the limitation of claims data. Third, it was also difficult to determine the number of patients with latent TB infection before using TNF-α inhibitors. In this study, the transition rate from latent TB to active infection could not be confirmed. Fourth, this claim dataset was unable to differentiate whether the development of MDR TB is because of the anti-TNF alpha agents or the immune status of the patients using anti-TNF alpha agents. Regardless of these limitations, our results could help understand anti-TNF agents’ effects as a driver of MDR-TB in patients with pulmonary TB.

This study identified an increased risk for MDR-TB in patients who received TNF-α inhibitors. Therefore, successful treatment of TB requires close monitoring of patients who have received drugs that may affect TNF-related immune dysfunction.

The data analyzed in this study is subject to the following licenses/restrictions: We cannot share the administrative data of the Korean government. Requests to access these datasets should be directed to mdhong@hallym.or.kr.

The studies involving human participants were reviewed and approved by Kangwon National University Hospital Institutional Review Board (KNUH-2022–02–011). Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

YH and JP contributed to the design concept, acquisition, analysis, and interpretation of data, and drafting of and revising of the manuscript critically for important intellectual content. JH and JP contributed to the design concept, interpretation of data, and revising of the manuscript critically for intellectual content. All authors contributed to the article and approved the submitted version.

This study was supported by the National Research Foundation of Korea grant (NRF2020R1C1C1009091 and NRF2020R1A2C1011455) and the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by Korean Government (HI21C1074).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.