Prevalence of extensively drug-resistant tuberculosis in a Chinese multidrug-resistant TB cohort after redefinition

Tuberculosis, caused by Mycobacterium tuberculosis (MTB) complex, continues to be a global public health priority [1, 2]. The World Health Organisation (WHO) estimates that approximately 10.0 million people developed TB and 1.4 million people died from this disease in 2019 [1]. The emergence of multidrug-resistant tuberculosis (MDR-TB; resistant to at least isoniazid and rifampin), with an estimated burden of 78% rifampicin-resistant TB cases worldwide, has jeopardised TB control and subverted the goals of the WHO’s END TB Strategy [1, 3]. Because of inherent resistance to the two most potent anti-TB drugs, MDR-TB treatment requires the use of second-line drugs that are less effective, more toxic, and costlier than first-line regimens [4, 5]; however, the overall rate of treatment success among MDR-TB patients is currently 57% [1]. This unsatisfactory outcome is expected to worsen the epidemic of this severe form of TB. Therefore more efforts are urgently needed for new and effective drugs to improve the chemotherapy of MDR-TB [6].

Over the past decades, several new or re-purposed agents antimicrobial hold promise for MDR-TB treatment, such as bedaquiline (BDQ), linezolid (LZD) and delamanid [6‐9]. Recent clinical trials have demonstrated that higher treatment success rates are achievable by inclusion of these new or re-purposed agents in the regimens [10]. Similarly, A study in South Africa demonstrates that bedaquiline (BDQ) is a safe drug and is associated with the high success rate for the MDR-TB and XDR-TB cohort even in high HIV burden areas [11, 12]. Recently, the WHO released the updated guidelines on the MDR-TB treatment on the basis of new experimental and observational evidence, in which the late-generation fluoroquinolones (FQs, i.e., levofloxacin and moxifloxacin), linezolid and bedaquiline are classified as preferred Group A drugs [13]. Subsequently, the definition of extensively drug-resistant TB (XDR-TB) has been revised by the WHO Global TB Programme, aiming to enable access to more effective treatment options for patients afflicted with drug-resistant strains [14]. The updated XDR-TB is defined as infection with an MDR-TB strain that is also resistant to any fluoroquinolone and at least one additional Group A drug. This redefinition points to increasing progression of the severity of the disease, and predicts poor clinical outcomes. The appropriate drug susceptibility testing (DST) methods are required to optimize the use of Group A drugs to improve the treatment of MDR-TB. Unfortunately, the commercial DST methods remain limited to assess in vitro susceptibility of MTB isolates to bedaquiline and linezolid. The national surveillance of anti-TB resistance in MDR-TB is thus essential to assist countries in planning the scale-up of patient management.

Despite a significant achievement in tackling the TB epidemic over the past years, China has a serious MDR-TB burden, with an estimated 74% rifampicin-resistant/MDR-TB cases in 2019 [1, 15]. Findings on temporal surveillance demonstrated that the prevalence of MDR-TB has dramatically increased in China [16]. The spread of MDR- and XDR-TB has been a major threat from both a clinical and a public health perspective. Although recent changes in the WHO recommendations for treatment of RR/MDR-TB patients have replaced injectables with bedaquiline in preferred therapy [17], the treatment of these patients majorly relies on long-course regimens stipulated by the WHO due to limited accessibility to bedaquiline. Only 54% of RR/MDR-TB cases who started treatment achieved a favorable outcome, which was recognized as a threat to TB control efforts in this country. To date, we still lack national data regarding the prevalence of XDR-TB in China. To address this concern, we conducted a study to determine in vitro susceptibilities of MTB isolates to FQs, bedaquiline and linezolid in a Chinese MDR-TB cohort. The putative drug target genes conferring drug resistance were screened by DNA sequencing.

Drug resistance surveillance is of importance to identify and predict the impact of new empirical anti-TB drug prescribing [22]. To our best knowledge, this was the first snapshot of the prevalence of XDR-TB among MDR-TB patients in a high-burden setting. Our data demonstrate that the rate of XDR-TB was noted in 6.8% of patients afflicted with MDR-TB, which was significantly decreased from 33.6% on the basis of previous definition. This low rate of XDR-TB was majorly attributed to relatively late introduction of BDQ and LZD for clinical management of MDR-TB in China, thereby preventing accumulation of drug resistant mutations [18]. Consequently, this holds great promise for MDR-TB treatment with WHO-endorsed regimens containing BDQ-LZD combination.

Although the overall prevalence of XDR-TB was low, the high prevalence of FQ-resistance in MDR-TB patients warrants national attention. In a recent population-based study in China, the prevalence of MFX resistance was markedly increased from 3.0% in 2000 to 7.7% in 2010 [23]. Similar results were observed by surveillance data in Beijing, which demonstrated that a statistically significant increase in LFX resistance over the past decade [20]. This phenomenon is probably the result of overuse of FQs in the treatment of undiagnosed bacterial infections in China in view of their promising efficacy and low occurrence of adverse events. In addition, widespread use of FQs in animal and food industries has resulted in reported excessive accumulation of antibiotics in environmental water samples in China [24]. The unexpected exposure to environmental FQs is another possible explanation for increasing FQ resistance in our MDR-TB cohort. Thus our primary results directly address the concern that essential interventions are required to reduce the misuse of antibiotics in clinical practice, as well as in livestock and food industries.

The acquisition of drug resistance is associated with the presence of genetic mutations conferring resistance [25]. The most frequently observed mutations in BDQ-resistant isolates were reported in Rv0678, a transcriptional repressor of efflux genes participating in the regulation of expression of MmpS5–MmpL5 [25, 26]. Consistent with previous investigation, mutations in Rv0678 were the major mechanism for BDQ resistance in our study. The high prevalence of Rv0678 mutation may be related to the prior exposure to clofazimine among MDR-TB patients considering the cross-resistance between these two drugs. Of note, mutations within the Rv0678 are highly diverse, with 4 unique mutations at 4 different positions. Although the basis for the highly diverse Rv0678 mutations remains unclear, our results reveal that the DNA sequencing is more suitable for identify mutations within Rv0678 locus rather than conventional PCR-based assays, such as real-time PCR and line probe assay. We also found that approximate half of BDQ-resistant isolates harbored no mutations within other known BDQ resistance genes. Similar results were noted in MTB isolates with LZD resistance. In addition to target gene mutations, multiple mechanisms confer drug resistance, such as cell wall permeability and efflux [27]. An experimental study by Velayati and co-workers found that the XDR-TB MTB isolates had thicker cell wall than MDR and susceptible isolates though viewing the ultrastructure of cell wall [28]. Thus we hypothesize that the cell permeability-associated mechanism may play an important role of in these MTB isolates. The poor correlation between genetic mutations and phenotypic resistance forebodes unsatisfactory performance using molecular panel to diagnose resistance to these novel drugs. Further analysis is required to elucidate novel mechanisms conferring BDQ and LZD resistance, thereby boosting the development of rapid molecular diagnostics for drug resistance.

Despite limited data, we observed that a patient with resistance to both LZD and BDQ but susceptible to MFX experienced treatment failure. In a recent BDQ clinical trial, the administration of BDQ-containing regimens provides particular benefit for pre-XDR- and XDR-TB patients [18]. On one hand, the existing evidence from studies confirmed BDQ as a cornerstone in the treatment of MDR-TB patients. Although the revised definitions of XDR-TB have emphasized the importance of BDQ, it excludes the patients at high risk of treatment failure from pre-XDR and XDR-TB group. More clinical data is required to subtly stratify drug-resistant TB patients according to their risk of unfavorable outcomes. On the other hand, there is an urgent need for commercial laboratory tests of the sensitivity of tubercle bacilli to BDQ and LZD, such as phenotypic DST using MGIT, which is essential to guide the choice of chemotherapy for MDR-TB patients.

We also acknowledged several obvious limitations to our study. First, despite the enrollment of MDR-TB isolates in our cohort, the relatively small number of MTB isolates tested may result in potential sampling bias. Second, the patients were non-responsive MDR-TB, possibly leading to overestimation of XDR-TB in Chinese population. Third, the MIC values were not assessed in triplicate. Hence, the lack of repeated experiments may introduce the systematic error by microdilution method, especially for isolates with MICs close to the breakpoint. Finally, whole genome sequencing (WGS) was not conducted to identify the molecular mechanism conferring BDQ and LZD resistance among MTB isolates without known mutations. Specially, a proportion of drug-resistant isolates may be caused by the presence of heteroresistant subpopulation, which were missed by Sanger Sequencing. Further WGS analysis is of importance to elucidate the molecular mechanism underlying resistance to these agents.

In conclusion, our results demonstrate that initial XDR-TB occurs in 6.8% of MDR-TB patients in our Chinese population, whereas the high prevalence of FQ-resistance in MDR-TB patients warrants national attention. Approximately half of BDQ-resistant and LZD-resistant isolates harbor no mutations within known resistance genes. The low prevalence of XDR-TB holds great promise for MDR-TB treatment with WHO-endorsed regimens containing BDQ-LZD combination.