Multidrug-resistant tuberculosis (MDR-TB), defined by resistance to rifampicin and isoniazid—the two most effective first-line anti-TB drugs—continues to pose a critical global public health challenge. Effective management of MDR-TB requires prolonged treatment with costly second-line regimens that involve complex administration protocols and rigorous monitoring (1, 2). Despite overall global declines in tuberculosis incidence and death, the prevalence of MDR-TB has been increasing (3). The 2024 Global TB Report by World Health Organization (WHO) estimated 10.8 million new TB cases (representing an incidence of 134 per 100,000 population) and 1.25 million TB-related deaths in 2023. These figures surpass deaths from Coronavirus Disease 2019 (COVID-19) and reestablish tuberculosis as the foremost infectious disease killer (4). Achieving the targets of the WHO End TB goal—an 80% reduction in incidence, a 90% decline in death, and the elimination of catastrophic costs for affected families by 2030—remains a major challenge (5).

Among newly diagnosed TB cases, 400,000 were identified as MDR-TB, with resistance observed in 3.2% of new cases and 16% of re-treatment cases (4). Further, significant geographic disparities in MDR-TB epidemiology persist, especially in high-TB-burden nations such as India, the Philippines, China, Russia, and South Africa, where an increasing incidence of MDR-TB is projected despite improvements in acquired drug resistance management (6). Modeling studies anticipate substantial increases in MDR-TB cases in these countries by 2040, thereby underscoring the urgent need for enhanced containment strategies to interrupt transmission (7).

Although MDR-TB is recognized as a significant threat, systematic descriptions of its long-term (>30 years) global spatiotemporal evolution patterns is still limited. Existing research focuses on specific regions, short time span (such as after 2000) or single index (such as incidence rate), and lacks panoramic and long-term scale analysis integrating incidence rate, death, disability-adjusted life years (DALYs) and their age gender geographical heterogeneity. Song (8) focused on 30-year trends, but their analysis mainly focused on the comparison of overall trends and drug sensitivity TB. Lv (9) provided valuable global burden assessments, but their analysis time frame up to 2019 did not capture potential key turning points before and after the COVID-19 pandemic, and lacked depth in spatial heterogeneity and future predictions. Alene et al. (10) conducted research on northwestern Ethiopia, using spatial autocorrelation (Moran’s I) and local spatial association index (LISA) to identify MDR-TB hotspots, but only covered a single country or region and did not expand globally. Sharma et al.’s (7) prediction was limited to four high burden countries. Although Guo et al. (11) used GBD 2021 data, they only used ARIMA models to predict until 2030, without integrating population structure and cohort effects.

Therefore, building on previous research in this field in the early stage, this study has filled some key analysis gaps. It aims to conduct a comprehensive spatiotemporal epidemiological study of global MDR-TB for the first time over a period of 46 years (from 1990 and predicting until 2035) using the latest released GBD 2021 data. We use Joinpoint regression to accurately identify the turning point of the historical trend over the past 30 years. By analyzing the distribution characteristics of key populations through subgroup analysis, such as specific age groups and genders, describe the country clustering of MDR-TB and its spatio-temporal changes in combination with spatial autocorrelation analysis and identify the continuous high burden hotspots. The Bayesian Age Period Cohort (BAPC) model was applied to integrate historical epidemiological data and the United Nations population projections, accounting for future demographic and age-structural changes across different countries and regions. This model was selected not only for its ability to capture temporal trends in the data, but more importantly, for its incorporation of population characteristics (such as age structure, birth cohort effects, and demographic shifts) across different countries and regions, thereby providing a more nuanced and accurate reflection of future disease burden than traditional time-series forecasting methods—enabling more refined predictions of the global incidence rate and number of MDR-TB cases between 2022 and 2035. The projections provide critical evidence for assessing the feasibility and challenges of achieving the WHO End TB goals (12, 13).

This study provides for the first time, a comprehensive spatiotemporal epidemiological picture of global MDR-TB spanning 46 years (1990–2035), filling the knowledge gap regarding its long-term evolution patterns, population and geographic heterogeneity, and refined long-term predictions. Our analysis confirmed and quantified the non-linear changes in MDR-TB burden over time, significant demographic differences, and persistent geographic clustering. Crucially, the BAPC model predicts that the global burden of MDR-TB will remain high until 2035, especially with population growth and aging, the absolute number of cases is projected to increase. This finding posed a serious challenge to achieving the End TB target.

The widespread adoption of first-line anti-TB drugs (isoniazid and rifampicin) since the 1980s, particularly after rifampicin’s inclusion in the short-course regimen that reduced treatment duration from 18–24 months to 6 months, was initially transformative (29, 30). However, these gains were eroded by escalating drug resistance, driven by suboptimal adherence to prolonged regimens and inadequate drug exposure in settings with weak health systems (31, 32). The steady increase in ASIR between 1990 and 2005 reflects both the expansion of drug-resistant TB transmission and historical diagnostic limitations. Late-twentieth-century improvements in diagnostic methodologies—including the advent of phenotypic techniques (e.g., solid culture drug susceptibility testing) and molecular diagnostics (e.g., Xpert MTB/RIF) (33, 34)—significantly enhanced case detection, thereby revealing previously undiagnosed reservoirs. The subsequent decline in ASIR post-2005 (annual percentage change = −3.15%) is likely associated with the implementation of enhanced TB control strategies, such as the directly observed treatment strategy (DOTS), and improved access to second-line therapies (35, 36). However, the recent plateau and slight resurgence in incidence since 2016 underscore persistent systemic gaps; treatment success rates remain at approximately 50%, thereby perpetuating community transmission of resistant strains. The impact of the COVID-19 pandemic on the global healthcare industry may have halted the downward trend of MDR-TB incidence after 2019. A review covering 17 low- and middle-income countries (LMICs) shows that the number of MDR-TB cases rebounded after the pandemic, especially in areas with weak healthcare infrastructure such as Indonesia and India (37). Concurrent declines in ASDR and DALYs from 2015 to 2021, as reported in global TB studies (38), suggest partial effectiveness of intensified MDR-TB management efforts. Nonetheless, persistent incidence trends necessitate the urgent adoption of novel transmission control strategies, optimized treatment regimens (e.g., incorporating bedaquiline and pretomanid), and the scaling up of quality assurance programs to effectively curb further transmission.

Our research indicated that the global burden of MDR-TB exhibits marked gender and age stratification. Notably, there were pronounced gender disparities, with males exhibiting substantially higher ASIR, ASDR and DALYs than females. This sex-based gradient aligned with established tuberculosis epidemiology and likely reflects male-predominant risk factors such as smoking, alcohol consumption, occupational exposures (e.g., mining or migrant labor), delayed healthcare-seeking behaviors, and sub-optimal living conditions (8). Collectively, these factors heightened resistance risks (39, 40) and underscore the importance of implementing gender-sensitive interventions—particularly community-based active screening targeting male populations—to disrupt transmission asymmetries (41). The incidence number burden was concentrated among middle-aged individuals, primarily those between 30 and 60 years of age (42), a trend that might be related to frequent social interactions, occupational exposures, and histories of nonstandard treatment among young and middle-aged populations (8, 43). In contrast, the death burden was predominantly observed in the older adults, specifically among individuals aged 60 years and above. For instance, a study in South Korea reported that patients over 75 years of age had a death risk 68 times higher than those under 24, with individuals over 65 accounting for 65% of total deaths (44). Additionally, DALYs peaked in the 45–60 age group, reflecting significant labor loss and socioeconomic impact (9). The burden among children under 14 years of age should not be underestimated. It was estimated that annually, between 25,000 and 32,000 children contract MDR-TB—accounting for approximately 3% of children’s tuberculosis cases—yet only 3–4% of these children receive standard treatment, resulting in a 21% death (45). Insufficient diagnostic capacity was a core issue, compounded by limited treatment options: while a 6–9-month short-course regimen was effective in adults, appropriate dosage forms and safety data for children were lacking (46). Strengthening the management of family contacts, developing child-friendly formulations, and incorporating children into pediatric clinical trials are essential strategies for reducing the disease burden in this vulnerable population (47).

Geospatial clustering analyses identified hyperendemic MDR-TB zones in sub-Saharan Africa, Central Asia, India, and Russia. In these regions, syndemic interactions with HIV/AIDS—reflected in co-infection rates of 20% in Africa—fragile health systems, poverty, and armed conflicts perpetuate transmission (48, 49). In South Africa, one-third of diagnosed TB cases discontinue treatment, thereby fueling the propagation of MDR-TB (50, 51). Moreover, in conflict zones such as Somalia, the collapse of healthcare infrastructure has created persistent hotspots through recurrent treatment interruptions (52). In contrast, European cold spots were associated with robust socioeconomic development and universal healthcare access. Through free treatment, transportation subsidies, and occupational protection policies in Europe and other regions, TB patients do not need to interrupt treatment due to economic pressure, effectively blocking the transmission chain of drug-resistant bacteria in the community (53). To mitigate the occurrence of MDR-TB, it was imperative to implement social protection and poverty reduction strategies in low-income regions with lagging socioeconomic development (54), alongside increased allocation of medical resources to local populations. Phlegm culture, genetic testing (such as Xpert MTB/RIF), and second-line drug costs should be covered, encompassing the entire process from diagnosis to treatment. We strongly urged the international community to make targeted, strengthened, and sustainable investments in the region, such as enhancing diagnostic capabilities, promoting short-term programs, addressing HIV comorbidities, and providing social support.

The BAPC model projected a marginal 1.62% decline in global MDR-TB incidence by 2035, despite an anticipated increase to 480,000 cases annually. These findings, which align with Sharma’s earlier projections (7), suggested that high-burden nations such as India and Russia will continue to experience rising incidence until 2040, driven by demographic expansion and the accumulation of drug-resistant reservoirs. The core engine of future global population growth is sub Saharan Africa (contributing over 50% of the new population), followed by South Asian countries such as India and Pakistan (55). This trend is driven by ultra-high fertility rates, a young population structure, and a lagging transition in fertility. The proportion of older adults aged 65 and above will continue to rise, expected to reach 16% by 2050 (56). This underscores the need to prioritize absolute disease burden metrics in MDR-TB surveillance, especially as the gap with the WHO’s “End TB” targets widens due to multi-factorial drivers including demographic growth, imbalanced healthcare resource distribution, and cross-border transmission of resistant strains (9, 57). Although a six-month short treatment regimen (BPaL/M) has demonstrated improved outcomes, its prohibitive cost has limited accessibility for MDR-TB patients in low-income settings (58). To address these challenges, it was recommended that the WHO and national governments develop targeted strategies for low-resource regions, consider implementing initiatives such as “government subsidies + healthcare” to alleviate catastrophic treatment costs and ensure comprehensive patient care and support (59). A study shows that the local government of the Republic of Congo has taken measures to strengthen molecular monitoring and control to manage drug-resistant TB (60). A study in Brazil shows that measures such as reducing population density and controlling HIV can reduce the incidence rate of drug-resistant TB (61).

This study has limitations. First, we relied on GBD estimated data, which may under- or overestimate the true MDR-TB burden due to variable data quality across countries. Second, the BAPC model assumes future trends mirror historical patterns, overlooking disruptive events like pandemics, conflicts, or medical breakthroughs. Third, lacking global socioeconomic data restricts our analysis of factors driving disease burden disparities. Future work should integrate real-time surveillance with dynamic models and assess socioeconomic determinants of MDR-TB.

This first comprehensive spatiotemporal analysis spanning 46 years reveals that the MDR-TB epidemic presents a far more formidable challenge to the End TB goals than previously appreciated. It is characterized by persistent geographical hotspots (notably in sub-Saharan Africa), a disproportionate burden among males, the older adults, and children, and a projected rise in absolute cases to 480,000 by 2035—despite a marginal decline in ASIR—driven by sustained population growth and aging, presents a far more formidable and enduring challenge to the End TB goals than previously appreciated from fragmented or shorter-term studies. Effectively mitigating this trend demands a paradigm shift: (1) Hyper-targeted resource allocation informed by spatial hotspot mapping to sub-Saharan Africa and other high-burden clusters; (2) Development and implementation of demographically-tailored interventions addressing the specific barriers faced by men (e.g., strengthen management and intervene in behavior), the older adults (e.g., active case finding in high-risk settings, integrated care), and children (e.g., improved diagnostics, child-friendly formulations, contact investigation); (3) Sustained investment in novel tools and strategies (shorter regimens, new drugs, vaccines); (4) Strengthen government commitments, provide patients with more economic support, implement “government subsidies + healthcare” initiatives to reduce the burden of diagnosis and treatment; (5) Carry out international regional cooperation to prevent and control MDR-TB through assistance programs. Resource allocation and effective healthcare interventions in the face of this persistent threat would galvanize more urgent, focused, and equitable global action.