Version July 2025
INTRODUCTION —
Mycobacterium tuberculosis is an ancient human pathogen which continues to plague humanity despite the introduction of curative and preventive therapy in the last century [1]. The burden of drug resistance continues to evolve.
Multidrug-resistant tuberculosis (MDR-TB) has emerged in epidemic proportions in the wake of widespread human immunodeficiency virus (HIV) infection, particularly in regions impacted by disruptions in health services. Extensively drug-resistant TB (XDR-TB) was first reported in 2006 and has since been documented on six continents.
These trends are critically important for global health, since drug-resistant TB mortality rates are high, accounting for upward of 13 percent of all antimicrobial resistance related deaths worldwide [2]. Second- and third-line agents for treatment of drug-resistant TB can be less potent and less tolerable than first-line therapies.
This topic will focus on the epidemiology of XDR-TB. Issues related to epidemiology of drug-susceptible and drug-resistant TB are discussed separately. (See "Epidemiology of tuberculosis" and "Epidemiology and molecular mechanisms of drug-resistant tuberculosis".)
TERMINOLOGY —
TB terminology is inconsistent in the literature [3]. Relevant terms are defined in the table (table 1).
DEFINITIONS FOR TB DRUG RESISTANCE —
Definitions for drug-resistant TB are summarized in the table (table 2).
ESTIMATES OF XDR-TB DRUG RESISTANCE
1940s to 1990s — Drug resistance was first noted in the 1940s when streptomycin was formally studied as monotherapy for the treatment of TB [4]. As a result, subsequent therapeutic interventions utilized multidrug regimens to decrease the risk of drug resistance.
Despite this strategic approach, outbreaks of multidrug-resistant TB (MDR-TB) occurred throughout the world in the mid-1990s, including the United States, Spain, Italy, Argentina, and Russia [5-9]. During these years, in regions where HIV infection was prevalent, the epidemiology of M. tuberculosis mirrored the epidemic of HIV infection, since severe immunosuppression greatly increases the risk of active TB [10]. As a result, the number of TB cases in regions such as sub-Saharan Africa increased dramatically in the late 1990s and first decade of the 2000s [11]. These outbreaks hinted at the potential for more extensive drug resistance in regions with the greatest prevalence of TB and HIV coinfection and the least infrastructure to screen and manage such patients.
South Africa 2006 epidemic — International alarm surrounding XDR-TB first surfaced in 2006 with the detection of a large cluster of cases from the rural town of Tugela Ferry in the province of KwaZulu-Natal, South Africa [12]. In a prospective study of enhanced surveillance, 2203 sputum specimens from 1539 individual patients with suspected TB were sent for sputum culture and drug susceptibility testing. Of 542 patients with at least one positive sample for M. tuberculosis, 221 (41 percent) had MDR-TB and 53 patients of this subset were confirmed to have XDR-TB; a remarkably high proportion of complex drug resistance for a single geographic location at the time.
In the first five years of that epidemic, more than 400 cases of XDR-TB were reported from Tugela Ferry [13] and, during the height of the epidemic in 2007 to 2008, there were more reported cases of XDR-TB than MDR-TB [14] and nosocomial transmission was identified as an important source of transmission [15].
Since that time, there has been a significant local decline of notified XDR-TB cases, associated with the application of expanded, comprehensive, and integrated TB and HIV treatment and prevention policies and practices [16].
A study including more than 400 patients in KwaZulu-Natal with a diagnosis of XDR-TB between 2011 and 2014 included genotyping of M. tuberculosis isolates and social network analyses to describe epidemiologic linkages [17]. Findings demonstrated that 280 patients (69 percent) with XDR-TB had not received prior treatment for MDR-TB. In addition, genotypic analysis in 386 participants demonstrated 84 percent belonged to one of many specifically defined clusters. The authors concluded that the majority of cases were likely due to person-to-person transmission, rather than inadequate treatment of MDR-TB. Further follow-up study identified the importance of social mixing in urban settings; this has been associated with a higher degree of genetic linkage of XDR-TB isolates (as a proxy for transmission). Unexpectedly, people with smear-positive disease (classically more symptomatic with a higher bacillary burden) were less likely to have an isolate with genetic linkage [18]. This work demonstrates that focus on interrupting transmission that may be asymptomatic or smear negative in both community and health care facilities is paramount to limiting spread.
Worldwide — XDR-TB has been reported in >120 countries from all geographic regions of the world (figure 1) [12,19-24]. The largest documented numbers of cases have occurred in Eastern Europe, countries of the former Soviet Union, India, and China. Countries and regions of high XDR-TB overlap with those of other drug-resistant TB frequency. The proportion of M. tuberculosis isolates identified as MDR-TB that also carry additional pre-XDR or XDR-TB resistance patterns can vary widely within regions depending on the timing and location of surveillance.
In a meta-analysis extracted including 64 studies involving 12,711 patients with MDR-TB who were treated from 2003 to 2020, the pooled proportion of pre-XDR-TB among MDR-TB cases was 26 percent (95% CI 22-31) [25]. A surveillance study from China had the highest proportion of pre-XDR-TB (66 percent) [26] and a study from Ethiopia the lowest (3 percent) [27]. The estimated pooled proportion of XDR-TB among patients with MDR-TB was 9 percent (95% CI 7-11 percent) with the highest proportion of XDR-TB reported in a surveillance study from India (77 percent) [28] and the lowest in Ethiopia [29] and Cameron (1 percent) [30].
In the World Health Organization (WHO) report covering country level TB data from 2022, 410,000 people developed RR-TB or MDR-TB, accounting for 3.9 percent of the 10.6 million estimated people with new TB disease during 2022 [31].
Globally in 2022, 73 percent of people (2.9/4.0 million) diagnosed with bacteriologically confirmed pulmonary TB were tested for rifampicin resistance, up from 69 percent (2.4/3.5 million) in 2021. Among the 149,511 people diagnosed with RR-TB, 27,075 people with pre-XDR-TB or XDR-TB were detected. Previously, a large proportional drop in people diagnosed with drug-resistant TB from 2019 to 2020 (157,903 with drug-resistant TB, including 25,681 cases of pre-XDR-TB or XDR-TB in 2020 dropping from 201,997 with drug-resistant TB in 2019) mirrored the similar proportional drop in the number of new people diagnosed with any form of TB.
These reductions were not estimated to be due to a true drop in disease burden, but rather due coronavirus disease 2019 (COVID-19)-related health systems disruptions to support systems typically in place for people to access TB diagnostic services [32]. The COVID-19 pandemic also reduced previously steady gains in the number of people with drug-resistant TB enrolled in treatment that continue to recover (figure 2). Given the anticipated scale up of all-oral and shorter course drug-resistant TB regimens, these trends are projected to improve. (See "Treatment of drug-resistant pulmonary tuberculosis in adults".)
As the WHO surveillance indicates, the actual reported cases of XDR-TB remain a significant under-representation of the true burden of disease [33]. In some regions, presumed TB is still treated empirically without microbiologic confirmation or the diagnosis is based on diagnostic tests without additional drug-susceptibility testing. In settings where drug-susceptibility testing is available but universal drug-susceptibility testing is not practiced, testing has been highly variable even when standardized patients with clear risk factors for drug-resistance (such as prior TB treatment) are utilized to study testing patterns [34]. In high-burden settings, laboratory diagnosis of MDR-TB does not always trigger investigation for XDR-TB, which requires additional drug susceptibility tests, often using prolonged culture-based techniques or next generation sequencing techniques that may be inaccessible [35]. In addition, referral bias reduces the estimate of drug resistance, since patients may die before seeking medical attention or before drug-resistant TB is suspected.
United States — In a review of United States TB cases reported from 1993 to 2007, a total of 83 people with XDR-TB were reported [36]. The majority of people were United States born (n = 45). The annual rates of XDR-TB declined from 18 people in 1993 (0.07 percent of 25,107 total TB cases) to 2 people in 2007 (0.02 percent of 13,293 total TB cases). In 2014, for example, there were only two cases of XDR-TB reported in the United States, and since 2009 the majority have been among people born in countries other than the United States [37].
The decline of reported cases coincided with the introduction of potent antiretroviral therapy (ART) in the United States and improved public health control of TB and HIV infections. Of 40 cases reported from 1993 to 1997, 25 (63 percent) were known to be HIV infected; from 1998 to 2007, only 6 (14 percent) were known to be HIV infected. In a study of bedaquiline use for MDR-TB, pre-XDR-TB and XDR-TB in the United States from 2020 describing patients diagnosed from 2012 to 2016, of 14 patients, 7 (50 percent) had MDR-TB, 4 (29 percent) had pre-XDR-TB, and 3 (21 percent) had XDR-TB [38]. All were people born in countries other than the United States, 5 (36 percent) had diabetes mellitus, but only one patient (7 percent) was living with HIV.
GENERAL PRINCIPLES OF DRUG RESISTANCE —
Within any given person infected with M. tuberculosis, there is a varied population of organisms, each with naturally occurring resistance to a given medication, which preexists prior to treatment. This observation mandates multidrug therapy for TB disease given the higher burden of organisms and in order to decrease the risk of drug resistance.
The mechanisms for developing drug resistance are complex and may be multifactorial in origin even within an individual person. While it remains difficult to study whether certain lineages of M. tuberculosis are more virulent independent of their drug-resistance mutations, evidence has suggested that strains of lineage 2 (East Asian and Beijing sublineage) acquire resistance more rapidly than strains of Euro-American lineage due to a higher in vitro mutation rate [13].
Expansion of other drug-resistant sub-lineages with high reproductive fitness have been described in Eastern Europe and Central Asia [39]. In an investigation of MDR-TB strains from Moldova, the M. tuberculosis Ural lineage 4.2 and Beijing lineage 2.2.1 strains were responsible for most circulating MDR-TB, yet the fitness of MDR Ural strains was high relative to non-MDR strains of the same lineage, and even higher than strains of the Beijing sub-lineage commonly associated with successful transmission [40]. The rapid expansion of the large MDR clade of Ural lineage 4.2 strains in Moldova occurred over a relatively short epidemiological period of 10 to 15 years. Even with introduction of novel regimens for drug-resistant TB, selection for resistance to new antibiotics may be difficult to avoid, especially if infected by strains with high reproductive fitness (See "Epidemiology and molecular mechanisms of drug-resistant tuberculosis".)
Acquired versus primary drug resistance — An important distinction is made between acquired and primary drug resistance. Patients with "acquired" or "amplified" drug resistance have had prior treatment with anti-tuberculosis medications, while patients with "primary" or "transmitted" drug resistance may have contracted a previously drug-resistant organism from an infectious contact.
The factors that may lead to acquired drug resistance are discussed below (see 'Risk factors for drug resistance' below). Primary drug resistance can result from:
●Initial infection with drug-resistant TB
●Exogenous reinfection in a patient who had previous eradication of TB
●Exogenous superinfection with a drug-resistant isolate in a patient with a concurrent and more drug-susceptible isolate of TB
Exogenous reinfection with TB is more common in people infected with HIV with advanced immunosuppression than in patients with relatively intact cellular immunity [41]. A study from Tugela Ferry demonstrated exogenous reinfection of either multidrug-resistant TB (MDR-TB) or XDR-TB in a cohort of 23 patients who had previously been successfully treated for drug-susceptible TB [42]. Of the 15 patients who underwent testing for HIV, all were HIV infected. Reinfection was demonstrated using spoligotyping, a genetic assay that helps determine whether a patient may have a relapse of their initial infection versus acquisition of a new strain. Another report from Uzbekistan employed spoligotyping to confirm exogenous superinfection with XDR-TB occurring in four patients being treated for MDR-TB [43].
In a large study from nine sites in Eastern Europe, the Russian Federation, South Africa, and Southeast Asia before the advent of shorter course therapies and introduction of novel agents for drug-resistant TB such as bedaquiline and pretomanid, acquired resistance while on therapy for MDR-TB was substantial [44]. Of 832 patients without baseline resistance to second-line drugs, 68 (8.9 percent) developed new resistance to both a fluoroquinolone and an injectable agent, thereby characterized as XDR-TB. Acquired or amplified drug resistance appears preventable with a robust public health infrastructure and coordinated international effort. Furthermore, the nine-site study found the risk of acquiring drug resistance increased as the number of other susceptible drugs decreased, emphasizing the importance of early and rapid drug susceptibility testing.
Even when treating MDR-TB with drugs defined as susceptible by conventional testing, individual pharmacokinetic variability may predispose to acquired resistance [45], given the need to treat for such extended durations and the co-morbidities predisposing to pharmacokinetic variability that are more common in TB endemic settings (such as HIV and malnutrition from enteric enteropathy) [46].
In a prospective cohort of 290 adults treated for RR/MDR-TB from Bangladesh, Tanzania and the Russian Federation, the serum area under the concentration-time curve over 24 hours (AUC0-24 or total drug exposure) was calculated for each drug in patients' multidrug regimens, and quantitative susceptibility of the M. tuberculosis isolate was measured by minimum inhibitory concentrations (MIC) [47]. Treatment failure was defined as death, culture positivity at week 24 or beyond, acquired drug resistance (pre-specified as a 4-fold or greater increase in MIC of any drug in the patient's regimen), or a lack of improvement of the major TB-related symptom at week 24 or beyond (clinical failure). 62 (21 percent) experienced treatment failure, including 30 deaths. Moxifloxacin AUC0-24/MIC target of 58 was associated with favorable treatment outcome (OR, 3.75; 95% CI, 1.21-11.56; p = .022); levofloxacin AUC0-24/MIC of 118.3, clofazimine AUC0-24/MIC of 50.5, and pyrazinamide AUC0-24 of 379 mg × h/L were associated with faster culture conversion (HR >1.0, P < .05). Clustering by AUC0-24/MIC demonstrated that those with the lowest multidrug exposures had the slowest clearance of M. tuberculosis from the sputum.
Prevention and management — The degree to which primary or acquired drug resistance drives MDR- and XDR-TB rates has significant implications for designing effective prevention and management strategies:
●In areas with a high proportion of primary drug resistance, strategies aimed at airborne infection control to reduce transmission of TB in hospitals, clinics, and community congregate sites are essential.
●In contrast, variable adherence or inadequate infrastructure may lead to acquired drug resistance; in this situation, strengthening and improving the performance of TB treatment programs, including access to rapid drug-susceptibility testing, regular treatment response monitoring, and consistent and effective patient support, is critically important.
Geographic distribution — The global burden of TB drug resistance is a consequence of both primary and acquired disease. The 2024 World Health Organization report on TB drug resistance provides MDR-TB rates in newly diagnosed patients compared with those who were previously treated (figure 3) [24].
Previously, primary drug resistance was thought to occur in isolated outbreaks; however, large-scale national survey data has found otherwise [48,49]. In a survey of 3037 new and 892 previously treated cases from China, MDR-TB occurred in 5.7 percent (95% CI 4.5-7.0) and 25.6 percent (95% CI 21.5-29.8), respectively, and approximately 8 percent of the patients with MDR-TB had XDR-TB [48].
RISK FACTORS FOR DRUG RESISTANCE
Adherence and other factors — Factors contributing to drug resistance include social determinants (such as inadequate income, housing, food, alcohol and other substance use and social support) as well as health system limitations (such as prescribing errors or suboptimal dosing, treatment interruptions from medication stock shortages, or poor treatment completion due to overstressed directly observed therapy programs).
In addition, biologic factors play a role in acquired resistance when the decreased bioavailability of anti-tuberculosis medication occurs secondary to malabsorption and drug-drug interactions that affect metabolism and excretion. This point is particularly pertinent to people living with HIV. (See 'Effect of HIV infection' below.)
Clinical trial simulations have demonstrated that approximately 1 percent of TB patients with perfect adherence would still develop multidrug-resistant TB (MDR-TB) as a result of pharmacokinetic variability alone [50], a rate that may be even higher when considering the development of XDR-TB in a patient on a standardized MDR-TB regimen. Even in directly observed settings, adherence to second-line drug regimens for MDR- or XDR-TB can be adversely affected by perceived social stigma, isolation, and psychological intolerance of medication, which has even led to selective adherence to HIV medications over MDR- or XDR-TB medications [51]. Integration of HIV and TB services may overcome such barriers.
Historical clues — Certain historical factors that should raise the possibility of XDR-TB have been suggested in observational studies and expert opinion but have not been formally evaluated [52-56].
●Living in or originating from an area with a known high MDR- or XDR-TB prevalence
●Previous TB therapy without documented success despite adequate adherence
●Hospitalization within the last two years, particularly in areas of high MDR or XDR prevalence
●HIV/acquired immunodeficiency syndrome (AIDS)
●Incarceration
Other comorbidities that may predispose to drug resistance include a history of malabsorption or taking other medications with known pharmacokinetic interactions with TB therapy.
Health care workers — Health care workers in endemic areas are at increased risk of TB, including MDR-TB [57,58].
A retrospective chart review was conducted of patients who were hospitalized for initiation of treatment for MDR-TB and XDR-TB in KwaZulu-Natal province, South Africa [59]. All patients were asked whether they were health care workers as part of a solicited occupational history. From 2003 to 2008, 4941 patients were hospitalized and 213 health care workers were diagnosed with either MDR-TB or XDR-TB. When compared with the general population, the average incidence rate ratio for TB infection was higher for health care workers for both MDR- and XDR-TB (fivefold and sixfold higher, respectively), while HIV infection rates were comparable between groups. Health care workers were less likely to report prior treatment for TB compared with other patients, suggesting nosocomial acquisition in the majority of cases. Another study found a direct relationship between the amount of time a health care worker spent in hospital rooms with a patient with documented TB and risk of subsequent TB infection [58].
Effect of HIV infection — People living with HIV are more likely to have problems with malabsorption, altered drug metabolism, or drug interactions due to concomitant antiretroviral therapy (ART). This can lead to acquired drug resistance because of inadequate antituberculosis drug levels and increased selection for mutations.
People living with HIV are more susceptible to acquisition of drug-resistant strains, as illustrated by the following data:
●In the original Tugela Ferry, South Africa case series, all XDR-TB patients who underwent testing were HIV infected; subsequent epidemiologic surveys have shown that greater than 90 percent of multidrug-resistant (MDR-) and XDR-TB patients from this high-prevalence HIV community are HIV infected [12,60].
●In an observational analysis of the 5200 TB patients diagnosed from Latvia, the risk of developing drug-resistant TB was more than twofold higher in patients with underlying HIV infection compared with those who were HIV uninfected or without a known HIV status [20].
●Similarly, in a Ukrainian study of 1496 confirmed TB patients, MDR-TB was identified in 24 percent of HIV-uninfected and 32 percent of HIV-infected patients (odds ratio 1.3, 95% CI 1.1-1.5) [61].
As a result of the association between HIV and TB drug resistance, the World Health Organization (WHO) Global Task Force has recommended initial sputum culture and drug susceptibility testing for all people living with HIV and with TB; use of Xpert MTB/RIF assay should be used where available [62]. As an example, modeling suggests that integration of community-based screening and treatment programs for HIV and TB (with Xpert MTB/RIF) in a rural district in South Africa alone would have the potential to avert 69 percent of new XDR-TB cases (95% CI 34-90 percent) over a 10-year span [63].
Significant concern exists for the potential of a swell of new cases of HIV-related MDR- and XDR-TB in the Russian Federation and countries of the former Soviet Union, given the rise in HIV incidence and HIV-related mortality [64,65]. The Russian Federation estimated over 1.1 million people living with HIV in 2017, accounting for 80 percent of new HIV infections in all of Eastern Europe and Central Asia [66]. Reports have highlighted the urgent opportunity for HIV and TB services integration in the Russian Federation including for people who inject drugs [67,68].
COMMON GENETIC MUTATIONS —
The discovery of specific genetic mutations resulting in drug resistance has allowed targets for novel polymerase chain reaction or targeted next generation sequencing-based diagnostics. Some of the best understood mutations are those associated with isoniazid and rifampin resistance but, while helpful for MDR-TB, confirmatory diagnosis of XDR-TB can still require standard culture and drug susceptibility testing, which can take up to three months to complete [69,70]. Mutations in resistance-determining regions of genes in important drugs for the treatment of pre-XDR or XDR-TB, for example, gyrA for fluoroquinolones in a real-time PCR assay [71], or mmpR5 for bedaquiline and clofazimine non-susceptibility in sequencing based assays [72], have been incorporated in the commercially available platforms. (See "Epidemiology and molecular mechanisms of drug-resistant tuberculosis".)
SOCIETY GUIDELINE LINKS —
Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Diagnosis and treatment of tuberculosis".)
SUMMARY AND RECOMMENDATIONS
●Introduction – Extensively drug-resistant tuberculosis (XDR-TB) emerged in epidemic proportions in the wake of widespread HIV infection in the world's poorest populations, including sub-Saharan Africa, and in areas with and/or disrupted programmatic infrastructure, such as countries of the former Soviet Union. The Russian Federation is experiencing a rise in HIV-related drug-resistant TB and drug-resistant strains. (See 'Introduction' above.)
●Definitions – XDR-TB is defined as resistance to both isoniazid and rifampin with additional resistance to at least one fluoroquinolone and at least one additional group A drug (as of 2021, these are bedaquiline and linezolid; the group A drugs may change in the future) (table 2). (See 'Definitions for TB drug resistance' above.)
●Estimates of XDR-TB resistance – The largest reported cluster of XDR-TB cases was described from South Africa in 2006 and was associated with HIV infection and a high rate of mortality. Subsequently, both nosocomial and community transmission have been identified as critical epidemiologic drivers. (See 'Estimates of XDR-TB drug resistance' above.)
●Acquired versus primary resistance – Patients with "acquired" drug resistance have had prior treatment with antituberculosis medications, while patients with "primary" or "transmitted" drug resistance have contracted a drug-resistant organism from an infectious contact. (See 'General principles of drug resistance' above.)
●Effect of HIV infection – People living with HIV are more likely to have problems with malabsorption, altered drug metabolism, or drug interactions due to concomitant antiretroviral therapy (ART) that may lead to acquired drug resistance because of altered pharmacokinetics and increased mutation selection. (See 'Effect of HIV infection' above.)
