Highlights
In a high-burden South African setting, more than half of sequenced second-line drug-resistant tuberculosis cases were genetically clustered, indicating recent transmission rather than isolated reactivation.
Among clustered participants, epidemiologic links attributed to casual contact were far more common than links identified through conventional close-contact pathways such as named person-to-person exposure or overlapping hospitalizations.
Living within 1 km of another clustered case showed the strongest association with genotypic clustering, underscoring the importance of neighborhood-scale transmission.
The findings challenge the heavy reliance on household contact tracing alone and support expanded community-based tuberculosis case finding and infection-control strategies.
Background
Tuberculosis remains one of the world’s leading infectious causes of death, and drug-resistant tuberculosis poses a particular threat to control efforts because diagnosis is often delayed, treatment is prolonged, and outcomes are worse than for drug-susceptible disease. In high-burden countries, transmission is now recognized as a major engine sustaining both tuberculosis incidence and the growing burden of rifampicin-resistant and more extensively drug-resistant forms. Yet an enduring question has been where transmission actually occurs and how often it happens outside traditional “close contact” settings.
Conventional tuberculosis control programs primarily focus on household contact investigation and, in some settings, workplace or school exposure assessment. These strategies are clinically sensible, but they may miss the broader social and spatial networks through which Mycobacterium tuberculosis spreads. Casual contact in crowded clinics, public transport systems, community gathering places, and densely populated residential areas may account for a substantial proportion of transmission, especially in urban settings with high prevalence of undiagnosed disease.
Molecular epidemiology offers a way to address this gap. Whole genome sequencing (WGS) provides much finer resolution than older genotyping methods and can identify genetically linked isolates that are consistent with recent transmission. When these genomic data are combined with geographic information and exposure histories, they can help distinguish whether transmission is more likely to occur among known close contacts or through less obvious community interactions. The current study by Gandhi and colleagues directly addresses this question in metropolitan Durban, South Africa, a setting with a heavy tuberculosis burden and substantial drug resistance.
Study Design and Methods
Design and setting
This was a prospective molecular epidemiology study conducted in metropolitan Durban, South Africa, from June 2018 through December 2022. The investigators recruited persons diagnosed with second-line drug-resistant tuberculosis, including pre-extensively drug-resistant (pre-XDR) and extensively drug-resistant (XDR) tuberculosis. The central aim was to quantify the relative contribution of close contact versus casual contact to transmission among genetically linked cases.
Population
Of 383 persons diagnosed with second-line drug-resistant tuberculosis during the study period, 305 were enrolled, representing 80% of the eligible population. Tuberculosis isolates were successfully sequenced for 251 participants, or 83% of those enrolled. This is an important strength because universal or near-universal sequencing within a defined population reduces ascertainment bias compared with studies sequencing only selected clusters or outbreak suspects.
Exposure assessment
The investigators collected detailed epidemiologic information before diagnosis, including named contacts and GPS coordinates of homes, clinics, and regularly visited community locations. This allowed the study to move beyond standard contact tracing and formally examine spatial and social overlap outside the household.
Participants were classified as genotypically clustered when their M. tuberculosis isolates differed by 12 or fewer single nucleotide polymorphisms (SNPs), a commonly used threshold suggesting likely recent shared transmission chains. Within these clusters, epidemiologic links were categorized as follows:
Close contact: direct person-to-person links or overlapping hospitalizations.
Casual contact: geographic proximity of homes, proximity of regularly visited community locations, or attendance at the same outpatient clinic.
Outcome and analysis
The primary analytic objective was to quantify what proportion of clustered cases could be attributed to close versus casual contact. The investigators also performed multivariable analysis to identify factors independently associated with genotypic clustering. The abstract reports odds ratios but does not provide confidence intervals in the summary available through PubMed, so interpretation of estimate precision is necessarily limited without the full paper.
Key Findings
High proportion of genomic clustering
Among the 251 participants with sequenced isolates, 141 individuals, or 56%, were genotypically linked. These cases formed 25 clusters ranging in size from 2 to 49 persons per cluster. This degree of clustering strongly suggests that ongoing transmission, rather than remote reactivation alone, is a major contributor to second-line drug-resistant tuberculosis in this setting.
From a public health standpoint, this is a striking observation. Drug-resistant tuberculosis is sometimes assumed to arise predominantly through inadequate treatment and acquired resistance. Although treatment-related amplification remains important, the present data reinforce the now well-established view that direct transmission of resistant strains is a critical driver in high-burden environments.
Casual contact links were much more common than close contact links
Among the 141 clustered participants, 69 individuals, or 49%, were epidemiologically linked through casual contact. By contrast, only 13 individuals, or 9%, were linked through close contact. The implication is not merely that casual contact exists, but that it may explain a far larger share of observed transmission than traditional contact investigations capture.
This finding deserves careful interpretation. It does not mean that household or named contact tracing is unimportant. Rather, it shows that a control strategy focused mainly on close contacts is likely to miss a substantial proportion of transmission events, especially for drug-resistant tuberculosis circulating through community networks and neighborhood spaces.
Spatial proximity was strongly associated with transmission
In multivariable analysis, living within 1 km of another clustered case was associated with genotypic clustering with an odds ratio of 17.9. This was the strongest reported association in the study and suggests that neighborhood-level forces may be highly relevant to tuberculosis spread. Such forces could include repeated low-intensity exposure across shared microenvironments, social mixing patterns, crowding, and the presence of undiagnosed intermediary cases.
Notably, this type of result broadens the concept of contact. Tuberculosis transmission may often occur not between people who recognize one another, but among individuals who share air in overlapping residential and community spaces. That idea aligns with prior social mixing and geospatial studies showing that the infectious risk landscape for tuberculosis extends well beyond the household.
Community locations and outpatient clinics also contributed
Visiting proximate community locations was associated with clustering (OR 1.88), as was attendance at a shared outpatient clinic (OR 1.72). These findings are biologically plausible. Tuberculosis is transmitted through inhalation of airborne droplet nuclei, and prolonged or repeated exposure in poorly ventilated settings can facilitate spread even without direct interaction.
Outpatient clinics are particularly important because they concentrate symptomatic patients, often have overcrowded waiting areas, and may see inadequate implementation of ventilation, masking, or patient flow strategies. For drug-resistant tuberculosis, clinics can become both diagnostic gateways and transmission nodes if infection prevention and control measures are insufficient.
Close contact still mattered, but explained only a minority of links
Direct person-to-person links remained associated with clustering in multivariable analysis (OR 5.38). This confirms that close exposure retains epidemiologic importance. However, the relative frequency of such links was small compared with casual-contact pathways. Clinicians and tuberculosis program leaders should therefore interpret this result as additive rather than substitutive: close-contact tracing should continue, but it cannot be the sole cornerstone of transmission control in high-burden urban settings.
Clinical and Public Health Interpretation
The major contribution of this study is conceptual as much as numerical. It reframes drug-resistant tuberculosis transmission as a community phenomenon that often escapes the narrow lens of named contacts. In practical terms, this means that patients presenting with pre-XDR or XDR tuberculosis may have acquired infection from someone outside the household, potentially through repeated exposure in local clinics, neighborhood venues, or shared urban spaces.
Several implications follow. First, passive case finding is unlikely to be sufficient. If many transmission events occur through casual contact or through undiagnosed intermediate cases, then waiting for symptomatic individuals to self-present will allow resistant strains to circulate. Second, infection prevention must extend beyond hospital wards and household counseling. Clinics, transportation corridors, community gathering spaces, and densely populated residential environments should be viewed as important intervention targets.
Third, tuberculosis programs may need to combine molecular surveillance with geospatial intelligence. Universal or large-scale WGS can identify transmission hotspots and reveal whether resistant strains are being propagated locally. When linked to address and mobility data, these tools could support targeted active case finding, ventilation upgrades, mobile diagnostics, and community-centered screening campaigns.
Fourth, the findings are especially relevant in regions with overlapping HIV burden, even though the abstract does not detail HIV-specific analyses. In southern Africa, HIV-associated immunosuppression increases susceptibility to progression after infection, which may amplify the consequences of community exposure. This further strengthens the rationale for integrated tuberculosis-HIV public health strategies.
Strengths of the Study
The study has several notable strengths. It recruited a large, population-based sample of persons with second-line drug-resistant tuberculosis over more than four years in a real-world, high-burden urban setting. The enrollment proportion was high, and sequencing coverage among enrolled participants was substantial. These features increase confidence that the reported clusters reflect actual transmission patterns rather than highly selected outbreaks.
The integration of WGS with GPS-based location data is a major methodological advance. Older tuberculosis transmission studies often relied only on named contacts or coarse residential data, which underestimates community spread. Here, the investigators were able to evaluate multiple exposure dimensions, including proximity of homes, community locations, and outpatient clinics.
Another strength is the explicit comparison between casual and close contact pathways within genomic clusters. This directly addresses a policy-relevant question: whether existing contact tracing paradigms capture the dominant routes of transmission in high-burden settings.
Limitations and Cautions
Despite its importance, the study should not be interpreted as proving the exact site or moment of transmission for each case. WGS can establish genetic relatedness consistent with transmission, but it cannot by itself determine directionality or identify the precise intermediary host. Some epidemiologic links categorized as casual contact may reflect exposure through unmeasured social networks or through unsampled infectious persons.
The use of a 12-SNP threshold is reasonable and common, but all genomic clustering cutoffs involve tradeoffs. A broader threshold may capture more true transmission chains while also including some more distantly related isolates. The full article may provide sensitivity analyses using narrower thresholds; those would be useful in assessing robustness.
Only 251 of the 305 enrolled participants had sequenced isolates. While this is still strong coverage, incomplete sequencing could miss links and alter apparent cluster structures. Similarly, the study includes only diagnosed second-line drug-resistant tuberculosis cases. Transmission involving undiagnosed, culture-negative, or non-enrolled individuals would not be directly observed.
Exposure assessment also depends on participant recall and the completeness of location reporting. Regularly visited places may be forgotten or underreported, and GPS proximity does not guarantee shared airspace. Nevertheless, these limitations would more likely dilute true associations than create the strong spatial signals observed here.
Finally, generalizability deserves consideration. Durban is a high-burden metropolitan setting with specific healthcare utilization patterns, population density, and tuberculosis epidemiology. The relative importance of casual community transmission may differ in rural areas, low-incidence countries, or settings with different clinic infrastructure and social mixing patterns.
Implications for Practice and Policy
For clinicians, the study is a reminder that absence of a known household exposure does not lower the likelihood that a patient acquired drug-resistant tuberculosis through recent transmission. Clinical histories should routinely include time spent in clinics, congregate settings, and neighborhood hotspots, although individual-level attribution will often remain difficult.
For tuberculosis programs, the findings support broadening interventions in at least four domains: first, active case finding in communities with geospatially clustered disease; second, stronger outpatient infection prevention and control, including ventilation, triage, respiratory protection, and reduced waiting times; third, molecular surveillance to detect transmission clusters early; and fourth, public health strategies that account for neighborhood and mobility patterns rather than focusing exclusively on households.
For health systems, the study argues for investment in clinic redesign and airborne infection control as core tuberculosis interventions, not ancillary quality-improvement measures. Drug-resistant tuberculosis control cannot depend only on treatment scale-up after diagnosis; it must also prevent repeated exposure in high-risk community environments.
Conclusion
This molecular epidemiology study provides compelling evidence that casual community contact plays a major role in the transmission of second-line drug-resistant tuberculosis in metropolitan Durban. More than half of sequenced cases were genomically clustered, and among clustered participants, casual-contact links were identified far more often than conventional close-contact links. The strongest signal came from neighborhood proximity, with additional contributions from shared community locations and outpatient clinics.
The practical message is clear: drug-resistant tuberculosis transmission in high-burden settings cannot be understood or controlled through household contact tracing alone. Effective control will likely require a more expansive strategy that combines WGS-informed surveillance, community-based case detection, and rigorous airborne infection prevention in outpatient and neighborhood settings. The study marks an important step toward a more realistic model of how drug-resistant tuberculosis spreads and where prevention efforts should be focused.
Funding and Trial Registration
Funding details and ClinicalTrials.gov registration information were not provided in the abstract available through PubMed. Readers should consult the full published article for complete funding disclosures and protocol registration details, if applicable.
References
Gandhi NR, Naidoo K, Lutchminarain K, Omar SV, Yoshii H, Willis F, Boodhram R, Gwala T, Campbell A, Coe MM, Evans AN, Tang L, Chitwood MH, Hlathi SR, Mbatha PN, Joseph L, Van Der Meulen H, Mlisana K, Jenness SM, Lurie MN, Kreiswirth BN, Warren JL, Nelson KN, Auld SC, Brust JCM, Cohen T, Mathema B, Shah NS. Role of Casual Contact in Drug-Resistant Tuberculosis Transmission: A Molecular Epidemiology Study. American Journal of Respiratory and Critical Care Medicine. 2026-04-28. PMID: 42089682. URL: https://pubmed.ncbi.nlm.nih.gov/42089682/
World Health Organization. Global Tuberculosis Report 2024. Geneva: World Health Organization.
Shah NS, Auld SC, Brust JCM, Mathema B, Ismail N, Moodley P, Mlisana K, Allana S, Campbell A, Mthiyane T, Morris N, Mpangase PT, van der Meulen H, Omar SV, Brown TS, Narechania A, Shiau S, Kreiswirth B, Gandhi NR. Transmission of Extensively Drug-Resistant Tuberculosis in South Africa. New England Journal of Medicine. 2017;376:243-253.
Martinez L, Shen Y, Mupere E, Kizza A, Hill PC, Whalen CC. Transmission of Mycobacterium tuberculosis in households and the community: a systematic review and meta-analysis. American Journal of Epidemiology. 2017;185(12):1327-1339.
McCreesh N, White RG. An explanation for the low proportion of tuberculosis that results from transmission between household and known social contacts. Scientific Reports. 2018;8:5382.
