Zoonotic tuberculosis: When diagnostic and surveillance systems are not designed to detect it

Abstract

Zoonotic tuberculosis (zTB) remains a blind spot in TB control, sustained less by biological rarity than by diagnostic and surveillance design. Brazil illustrates this gap: despite endemic bovine TB and sustained zoonotic transmission risks, zTB is rarely identified because it is not actively targeted. We show that only two human zTB cases were notified between 2013 and 2024 following introduction of reporting, and six cases are described in the literature. Further detection of mono-pyrazinamide resistance may represent undiagnosed zTB, but diagnostic gaps prevent estimation. By outlining current diagnostic paths, we argue that the human-centric diagnostics obscures occupational, foodborne, and animal-linked infection. Emerging subnational One Health initiatives demonstrate that integrated animal–human surveillance is feasible. Therefore, we propose a risk-based, One Health diagnostic protocol for zTB within the Brazilian public health system, and argue that without structural and policy adaptation, human-TB elimination strategies will continue to exclude zTB by design, not only in Brazil.

1. Introduction

More than 10.8 million cases and 1.25 million deaths occur worldwide due to tuberculosis (TB) every year, making it the deadliest disease in humans [1]. While most of these cases are caused by Mycobacterium tuberculosis sensu stricto (Mtb), around 147,000 cases and 12,500 deaths per year are attributed to zoonotic infection with other members of the Mycobacterium tuberculosis complex (MTBC), mainly Mycobacterium bovis [2] (Mbo, aka Mycobacterium tuberculosis variant bovis, according to recent phylogenomic proposal [3]). Unfortunately, the World Health Organization (WHO)’s Global Tuberculosis Reports do not address zoonotic TB (zTB); hence, the precise number of zoonotic cases and deaths is unavailable. The main reason for this lack of data is that routine surveillance is absent in 90% of WHO signatory countries [4]. The failure to recognize and diagnose Mbo infections undermines TB control efforts, particularly in rural and livestock-dense regions.

Zoonotic and human TB are clinically indistinguishable, and most countries do not perform differential laboratory diagnosis [5]. Common rapid diagnostic tests for TB, such as sputum smear microscopy and rapid molecular testing (e.g. Xpert MTB/RIF Ultra and TrueNat MTB), do not differentiate among MTBC species. A positive culture is also not species-specific, and Mbo and other MTBC strains have poor growth in media commonly used to isolate Mtb, such as Lowenstein Jensen and BD BACTEC™ MGIT™960. While assays such as GenoType MTBC (Hain LifeScience), PCR- and LAMP-based methods, as well as MALDI-TOF and Fourier transform infrared spectroscopy (FT-IRS), can differentiate among MTBC species [6], [7], [8], [9], [10], [11], available evidence suggests that their use in routine diagnostics is limited, particularly in low- and middle-income countries, as reflected by the ongoing lack of robust epidemiological data on zTB. Whole-genome sequencing (WGS) offers species-level resolution but is typically reserved for drug resistance and transmission studies due to its cost and complexity [12].

Brazil is classified as a high-burden country for TB [1] and hosts the world's largest commercial cattle herd, where bovine tuberculosis (bTB), caused by Mbo, is endemic [13]. The implementation of bTB control measures and zTB diagnostics remains insufficient because most existing directives function as non-binding guidelines, undermining the ability to operationalize effective control strategies at the interface of human and animal health. Despite this, at least six cases of zTB in Brazil have been reported in the literature since the 1990's [6], [14], [15]. Other exploratory studies found no evidence of zTB among TB patients in certain populations and regions of the country [16], [17]. It is likely that zTB accounts for a relatively small fraction of the total TB burden. However, zTB cases may carry significant clinical and epidemiological consequences [5], [18] and identifying these infections within the current diagnostic workflow remains exceptionally challenging, akin to searching for a needle in a haystack. In this context, the present study examines current zTB notification data and its diagnostic and surveillance barriers in Brazil and proposes strategies to strengthen zTB identification within a One Health framework.

1.1. Why is it important to identify cases of zoonotic TB?

Identifying cases of zTB is crucial for both clinical management and public health. Human and bovine TB have significant geographic overlap, as both diseases are highly prevalent in low- and middle-income countries where socioeconomic and agricultural conditions favor their persistence [18], [19]. Over 80% of cases and deaths of human TB occur in these settings [1], where endemic bovine TB increases the likelihood of zoonotic transmission through direct contact with infected animals or the consumption of unpasteurized dairy products.

At the individual level, accurate identification of Mbo infection is essential to ensure appropriate treatment. Since most Mbo strains are intrinsically resistant to pyrazinamide (PZA), patients misclassified as having Mtb infection may receive suboptimal therapy, leading to prolonged disease and poorer outcomes [2], [20]. They also receive PZA unnecessarily, increasing the risk of adverse events without providing therapeutic benefit. Moreover, zTB often presents as extrapulmonary disease compared with TB caused by Mtb, which contributes to delayed diagnosis and complicates clinical management [21].

To date, no clinical trials have been conducted to establish evidence-based treatment regimens for zTB. In practice, pulmonary disease is generally managed by extending the rifampicin–isoniazid–ethambutol regimen for an additional three months in the absence of pyrazinamide [22], yet data on treatment efficacy, including relapse rates, remain unavailable. The knowledge gap is even greater for extrapulmonary presentations and new schemes for drug-resistant TB. Although current recommendations implicitly assume that Mbo responds similarly to Mtb strains with PZA resistance, this assumption has not been validated by empirical data.

From an epidemiological perspective, identifying zTB cases enables a more accurate estimation of the true TB burden, but requires tracing zoonotic transmission routes between humans, livestock, and wildlife. This is particularly relevant for populations in close contact with animals, such as farmers, livestock and abattoir workers, and veterinarians [23]. Above all, the epidemiology of zTB diverges markedly from that of human TB: the populations at risk are distinct, shaped by occupational and environmental exposures rather than the classical socio-demographic risk factors associated with Mtb transmission. These groups face different vulnerabilities, including repeated exposure to infected livestock or wildlife, limited access to veterinary–public health interfaces, and diagnostic pathways that rarely consider zoonotic infection [24], [25]. As a result, existing TB control strategies, designed around the dynamics of human-to-human transmission, are poorly equipped to detect and manage zTB, requiring programs to adapt to an entirely different epidemiological scenario. There is also limited interest of the medical community to distinguish between human TB and zTB. Thus, zTB represents a critical blind spot in TB programs, underscoring the need for a One Health approach that integrates veterinary and public health system, to fill surveillance and diagnostic gaps.

1.2. Notified cases of zoonotic tuberculosis (zTB) from 2013 to 2024 in Brazil

In Brazil, all human TB cases must be reported in the Information System for Notifiable Diseases (SINAN). If a patient requires a treatment regimen different from the standard combination of the first line-drugs, the case is closed in SINAN and reported in the Information System for Special Tuberculosis Treatments (SITE-TB) [26]. Therefore, SITE-TB will account for patients with drug-resistant TB, patients with non-tuberculous mycobacteria (NTM) disease identified as differential diagnoses of TB, and patients with TB requiring alternative treatment regimens due to comorbidities or adverse events associated with standard treatment [27], [28]. Since most strains of Mbo are intrinsically resistant to PZA [20], its diagnosis implies an alternative treatment regimen, and such cases should be notified in SITE-TB. Notification in the SINAN-TB and SITE-TB systems is mandatory for drug dispensing to be authorized by the public health system.

In 2023, the Brazilian Ministry of Health issued a national guideline providing federal-level recommendations on the clinical management, surveillance, diagnosis, and treatment of zTB [22]. Concurrently, a new variable, called “M. bovis diagnosis”, was introduced in SITE-TB to improve zTB case reporting. Prior to this update, if a TB case was identified as being caused by Mbo, it was reported as PZA-resistant, solely to authorize the dispensing of a non-standard treatment regimen. With the new classification, notifications can explicitly reflect Mbo infection, ensuring that the appropriate drug regimen is authorized and enabling more accurate surveillance of zTB within the public health system.

To evaluate the possible number among zTB cases we retrieved all cases notified in SITE-TB between January 1, 2013, and August 30, 2024. This study was approved by the ethics committee of the Institute of Biomedical Sciences of the University of São Paulo (CAAE 79941524.0.0000.5467). Of the 37,291 validated cases reported in this period, 30,714 (82.36%) were caused by MTBC strains and 6577 (17.64%) were NTM diseases. Among the TB cases, only two (0.006%) were attributed to Mbo (i.e. “M. bovis diagnosis”). Both were notified in 2024, each from a different city and state. The patients were a 68-year-old female with pulmonary disease and an 82-year-old male with extrapulmonary genitourinary involvement. There is no information (i.e. the data cells are blank) regarding culture and drug-susceptibility testing by either phenotypic drug susceptibility testing (pDST), line probe assay (LPA), or Xpert MTB/RIF, limiting our ability to describe how the Mbo diagnosis was performed. Drugs used to treat these patients included rifampicin, isoniazid and ethambutol, for a total extended period of 9 months, which is the recommended regimen for PZA-resistant cases. Both patients had neoplasms/cancer (type not specified) listed as comorbidities, which raises the question of whether these were rare cases of Bacillus Calmette Guerin (BCG) infection (BCGosis) following intravesical BCG therapy for bladder cancer and not true zoonotic cases. At the time of data retrieval, treatment outcomes for both cases were still pending.

Mono-PZA resistance has historically served as a proxy for diagnosing Mbo infection. Since no other cases were defined as “M. bovis diagnosis”, we identified potential cases by searching for reported PZA resistance. Of the 30,714 TB cases, only 2395 (7.8%) had pDST results for PZA at the time of notification, among which 720 were resistant. Of these, sixteen cases were phenotypically susceptible to rifampicin, isoniazid, and ethambutol, i.e., resistant only to PZA (Table 1).

Table 1.

Diagnostic results of patients with tuberculosis resistant to pyrazinamide (PZA) and susceptible to rifampicin, isoniazid and ethambutol by phenotypic drug-susceptibility testing (pDST) at the time of notification in the Brazilian Information System for Special Tuberculosis Treatments (SITE-TB), Brazil, from January 1, 2013 to August 30, 2024.

Patient ID	pDST				LPA*	Xpert MTB/RIF	Other drug resistances	Final drug-resistance profile	Year of notification	Clinical presentation
	RIF	INH	PZA	ETH
1	S	S	R	S	N/A	MTB+ RIF+	pDST: AMK, CAP	Poly-resistance	2016	Pulmonary
2	S	S	R	S	N/A	N/A	–	Mono-PZA	2021	Lymphadenitis
3	S	S	R	S	N/A	MTB+ RIF-	–	Mono-PZA	2021	Pulmonary
4	S	S	R	S	rpoB	MTB+ RIF+	–	Poly-resistance	2022	Pulmonary
5	S	S	R	S	N/A	MTB+ RIF-	–	Mono-PZA	2022	Pulmonary
6	S	S	R	S	N/A	MTB+ RIFindet	–	Mono-PZA	2022	Pulmonary
7	S	S	R	S	rpob	MTB+ RIF+	–	Poly-resistance	2023	Pulmonary
8	S	S	R	S	negative	MTB+ RIF-	–	Mono-PZA	2023	Pulmonary
9	S	S	R	S	N/A	N/A	–	Mono-PZA	2023	Pulmonary
10	S	S	R	S	N/A	N/A	pDST: STR	Mono-PZA?	2023	Genitourinary
11	S	S	R	S	N/A	N/A	–	Mono-PZA	2023	Pulmonary
12	S	S	R	S	negative	MTB+ RIF-	pDST: STR	Mono-PZA?	2023	Pulmonary
13	S	S	R	S	N/A	MTB+ RIF-	–	Mono-PZA	2024	Pulmonary
14	S	S	R	S	N/A	MTB+ RIF-	–	Mono-PZA	2024	Pulmonary
15	S	S	R	S	N/A	MTB+ RIF-	–	Mono-PZA	2024	Pulmonary
16**	S	S	R	S	negative	MTB+ RIF-	–	Mono-PZA	2024	Pulmonary + Meningoencephalitis

pDST: phenotypic drug-susceptibility, RIF: rifampicin, INH: isoniazid, PZA: pyrazinamide, ETH: ethambutol, AMK: amikacin, CAP: capreomycin, STR: streptomycin, R: drug-resistant, S: drug-susceptible, LPA: line probe assay for first- and second-line drugs, N/A: results not available (blank). *list genes in which mutation(s) was(were) detected in the LPA-1 or LPA-2; negative means no mutations were detected in the sample for any of the targeted genes. **Patient #16 only had LPA for first line drugs conducted.

Among these sixteen cases, five had available LPA results for first- and/or second-line drugs, and twelve had Xpert MTB/RIF results. Two cases carried rpoB mutations associated with rifampicin resistance as detected by LPA and were also reported as rifampicin-resistant by Xpert MTB/RIF, while one additional case was identified as rifampicin-resistant only according to Xpert MTB/RIF (Table 1). Borderline rpoB mutations are relatively frequent in certain regions of Brazil [29], which may explain these discrepancies between pDST and molecular assays. Therefore, at least 13 cases reported in Brazil between 2013 and 2024 likely represent true mono-PZA resistance.

It is important to highlight that the completeness of pDST data in SITE-TB is limited (see Table S1). Among first-line drugs, PZA shows the highest proportion of missing data, at 86.18%, compared to 27.08–45.85% for other drugs (Table S1). This likely reflects that the test was not performed rather than a failure to report results. Importantly, there is no other diagnostic test listed in SITE-TB capable of assessing PZA resistance; pDST remains the only available method.

An explanation for this data gap is the well-documented technical limitations of pDST for PZA [30]. These include the requirement for acidic pH (≈5.5) for drug activity, which inhibits bacterial growth and leads to false resistance; poor reproducibility across laboratories; the fact that very few laboratories routinely perform the assay; and complex culture conditions that are not routinely available in many TB reference centers. Additionally, there is no universally accepted standardized method for PZA pDST, and results often show poor concordance with genotypic testing targeting the pncA gene [31], [32], [33]. These factors have led to the progressive discontinuation of phenotypic PZA testing in routine practice in the country. For these reasons, the potential identification of zTB through PZA resistance detected by pDST should be seen with caution and not be interpreted as representative of the true national burden.

1.3. A brief overview of animal TB in Brazil

Brazil is considered endemic for bTB caused by Mbo [13]. Other host-adapted members of the MTBC, such as Mycobacterium caprae and Mycobacterium orygis, have not been reported in the country to date. Current bTB control strategies are defined by the National Program for the Control and Eradication of Brucellosis and Tuberculosis (PNCEBT). Following PNCEBT guidelines, bTB diagnosis is mainly performed through intradermal tuberculin tests and post-mortem inspection in slaughterhouses.

A recent comprehensive review by Ferreira Neto (2024) [13] detailed the epidemiological status of bTB in Brazil. Accordingly, prior to 2001, when the Ministry of Agriculture and Livestock established the PNCEBT, the disease circulated unchecked and there were no systematic epidemiological surveys. Between 2003 and 2020, efforts were then initiated to estimate state-level prevalence of bTB in both individual animals and herds through comparative cervical tuberculin testing [13]. Prevalence surveys were conducted in 15 out of the 26 Brazilian states, as well as in the Federal District. The highest prevalences are seen in dairy herds [13]. Reported animal-level prevalence ranged from 0.009% (95% CI: 0.001–0.063%) in Tocantins [34] to 0.91% (95% CI: 0.68–1.0%) in Pará [35]. Herd-level prevalence estimates ranged from 0.36% (95% CI: 0–2.0%) in the Federal District [36] to 9.0% (95% CI: 7.8–10.5%) in São Paulo [37] [13]. These disparities reflect a combination of factors; northern São Paulo, together with bordering Minas Gerais, is a region of intensive dairy production with high animal density, and differences in official state-level veterinary services, which are responsible for enforcing bTB control strategies, likely contribute to variability in reported prevalence among states.

Only the State of Paraná conducted a second prevalence survey, comparing data from 2005 to 2007 [38] with that from 2018 to 2019 [39]. No significant difference in prevalence was observed between these periods [13], suggesting limited impact of the PNCEBT on reducing bTB prevalence over time. According to Ferreira Neto (2024) [13], a reflection of this is that Brazil currently has a total of 5472 farms accredited as free of bTB, an accreditation with specific rules defined by the PNCEBT. Of these, approximately 97% are concentrated in the two southernmost states of Brazil, Santa Catarina and Rio Grande do Sul. The remaining 197 bTB-free farms are sparsely distributed across 11 other states, highlighting the markedly uneven national distribution of disease-free certifications.

The PNCEBT establishes guidelines for the surveillance, control, and eventual eradication of bTB through measures such as testing, culling of positive animals, and movement restrictions [40]. The program is voluntary, but it sets a standardized framework for disease control across the country. While PNCEBT provides general guidelines, official veterinary services at the state level may adopt additional tools and stricter measures depending on their specific strategies and commitment to eradicating the disease, mostly in collaboration with the dairy and/or beef industry. The Brazilian Department of Inspection of Products of Animal Origin also conducts passive vigilance of carcasses in slaughterhouses [41]. Nevertheless, data remain scarce to assess whether these programs have had any significant impact over time.

Another important aspect to consider is the presence of TB caused by Mbo in wild animals kept in captivity. In Brazil, despite the rare Mbo detection in free-ranging wildlife, [42] numerous reports describe TB caused by both Mbo and Mtb in captive animals [43], [44], [45], [46], [47], [48], [49]. These findings raise serious concerns about cross-species transmission and the potential for spillover between wildlife, livestock, and humans. Since the PNCEBT does not make it mandatory to report these cases in wildlife, the actual number and impact of these outbreaks are unknown.

1.4. National and state-level guidelines on zTB

At the present, Brazil counts two official guidelines addressing zTB. The first was issued by the government of the state of Santa Catarina in 2020 [50] and addresses epidemiology, transmission, pathogenesis, diagnosis, treatment, and control of zTB. Mostly, it establishes a formally implemented intersectoral workflow for the active surveillance of individuals exposed to outbreaks of bTB, constituting an operational One Health policy. Under this protocol, notification by the official veterinary to the Municipal Health Secretariat is mandatory when an outbreak of bTB is detected in a herd. This notification serves as a trigger for a coordinated response by the public health sector. Measures include the seizure of animal-origin products originating from the affected herd and the implementation of active epidemiological surveillance of potentially exposed individuals, such as farm workers, producers, and household contacts. Surveillance actions may include collection of respiratory samples from suspected human cases, health education and awareness initiatives, and referral for clinical evaluation by the local health network. This integrated workflow represents a significant advancement by establishing a formal communication and response mechanism between previously siloed sectors, namely, agriculture/livestock and human health, within a One Health framework. It operationalizes cross-sectoral collaboration at the local level, enabling early identification of zoonotic transmission and ensuring a timely public health response.

Three years later, the Brazilian Ministry of Health issued a federal guideline about zTB [51]. It delineates a set of priority occupational risk groups for targeted surveillance and intervention, including individuals involved in agricultural and livestock production systems; animal handlers; supervisors of farming and livestock operations; slaughterhouse personnel; and artisanal workers engaged in milk pasteurization, dairy product manufacturing, and traditional food preservation. Despite the clear occupational risk associated with exposure to Mbo, the guideline does not explicitly include veterinarians or veterinary technicians among the defined high-risk populations. Additionally, it does not address other vulnerable occupational groups, such as zoological institution personnel, including zookeepers, biologists, and veterinarians, who have frequent and direct contact with captive wildlife. The exclusion of these groups from targeted surveillance strategies represents a gap in the current national zTB response.

This federal guideline proposes that targeted surveillance be operationalized through exposure-based case finding, including referral of individuals exposed to infected animals, active screening in high-risk workplaces, and periodic monitoring of occupational groups. On laboratory diagnostics, it states that routine species-level identification is not recommended for all TB cases, but in suspected zTB (with documented animal exposure) differential protocols should be used. These include culture on media suitable for Mbo, followed by species identification using biochemical and/or PCR-based methods when available. The guideline adopts a flexible approach, reflecting current limitations in laboratory capacity for routine species differentiation. Laboratories are asked to record suspected zTB cases in the Laboratory Environment Management System (GAL) under “M. bovis” and specify the clinical suspicion in the comments field to enable proper processing and notification. Nevertheless, these procedures rely entirely on clinician suspicion and accurate clinical history reporting. In most cases, occupational or exposure data are not systematically collected at the time of TB notification, resulting in missed opportunities for zTB diagnosis.

Unfortunately, this federal document remains a guideline rather than a binding public health policy; it lacks mandatory enforcement and depends on voluntary adoption by local health authorities. As a result, surveillance remains fragmented, with veterinary and human health services operating in parallel rather than in concert. The translation of technical guidance into enforceable practice is uncertain, and survey evidence suggests that such One Health coordination remains weak. Together, these gaps highlight the need for a unified, proactive national strategy that links human and animal health surveillance, expands occupational risk assessment, and integrates zTB detection into routine TB diagnostic workflows.

1.5. Unpasteurized milk consumption and artisanal dairy: emerging challenges for zTB surveillance

The consumption of raw milk is a recognized risk factor for zTB [52], and although the commercialization of raw and unpasteurized milk for human consumption is prohibited in Brazil, its informal sale persists in certain regions [53], [54]. This dietary habit is strongly linked to socioeconomic determinants, as low-income families are the ones who most frequently consume unpasteurized raw milk. A recent study involving 55,700 households showed that participation in Bolsa Família, a conditional cash transfer program aimed at improving education, health, and nutrition among impoverished families, increased the likelihood of purchasing unpasteurized milk [55]. This likely reflects that increased purchasing power raises milk consumption, which in vulnerable settings is often met through informal markets offering unpasteurized products. Another study conducted in the state of Paraná also demonstrated that the predominant profile of consumers consisted of low-income households, particularly those with a family income between one and four minimum wages [54]. In parallel, epidemiological investigations in other countries reinforce the correlation between the occurrence of zTB and socially vulnerable groups, highlighting that factors such as poverty, limited access to health services, and traditional dietary practices enhance exposure to Mbo [54], [56], [57].

Beyond the consumption of raw milk itself, Brazil has a strong tradition of artisanal cheese production [58]. Many of these cheeses are made from unpasteurized milk and produced in small-scale, low-volume facilities, such as Minas Artesanal, Colonial, and Canastra cheeses [59], [60]. Historically produced and consumed mainly in rural and informal contexts, these products have recently undergone commercial revalorization, driven by their territorial and cultural identity [61]. Their production process has been incorporated into specific regulatory frameworks at state and federal levels, which include the need to use milk coming from a herd that is free of brucellosis and TB according to PNCEBT guidelines [62], [63]. This revalorization is not merely economic, such cheeses were transformed into high value-added goods, positioned as exclusive items with territorial authenticity, expanding their presence in urban markets [64].

As many of these products did not fit within the traditional federal inspection system, and given their rising market demand, the federal government instituted, in 2018, the Selo ARTE (Law N°. 13,680/2018), which authorizes interstate commercialization of animal-origin foods classified as artisanal, provided that producers demonstrate adherence to good agricultural and manufacturing practices. This measure represents an important step toward sector formalization and sanitary traceability, but it also introduces new surveillance challenges, particularly when raw milk is used in production. In practice, the cultural and commercial expansion of the artisanal cheese market, if not accompanied by robust laboratory monitoring, may sustain the potential for human exposure to Mbo and maintain the circulation of zTB.

1.6. zTB laboratory diagnostic challenges

In a high burden TB country such as Brazil, culture of Mtb is not universal and is usually conducted with the objective of performing phenotypic or genotypic DST [65]. This limitation is largely due to financial constraints, restricted human resources, and laboratory infrastructure challenges that hinder universal culture implementation. The prioritization of specific groups for DST has led to a situation in which these populations are the primary recipients of culture testing, reaching >80% of the priority population, whereas among non-priority groups, culture coverage is around 50% (unpublished data). In the case of the state of São Paulo, for example, populations known to be at risk for zTB are not among those prioritized for DST (Table 2). In addition, there is no communication to the diagnostic laboratories regarding suspected Mbo infection or specific epidemiological risk factors, which prevents laboratories from directing samples to appropriate culture conditions or species-level investigation when warranted. Consequently, if routine diagnostic isolates are screened for Mbo without specifically targeting high-risk populations, the likelihood of detecting positive cases is extremely low, limiting the usefulness of blind surveillance approaches.

Table 2.

List of populations at risk for zoonotic tuberculosis (zTB) according to the Manual of Clinical and Surveillance Guidelines for Zoonotic Tuberculosis, and populations prioritized for drug-susceptibility testing according to the Program of Tuberculosis Control of the State of São Paulo.

Populations at risk of zTB

Populations prioritized for DST

•Workers and producers in agricultural and livestock operations •Animal caretakers •Supervisors of agricultural and livestock operations •Butchers (slaughterhouse professionals) •Artisanal workers in milk pasteurization and dairy production •Workers in milk pasteurization and dairy production •Artisanal workers in food preservation

•Re-treatment cases •Population deprived of liberty •Population experiencing homelessness •Patients with HIV co-infection •Patients with other immunosuppressive conditions •Healthcare professionals, correction workers, and other professionals working in enclosed institutions (e.g. nursing homes, shelters, mental health hospitals etc) •Contact individuals of patients diagnosed with drug-resistance •Patients with drug and/or alcohol addiction •Immigrants •Patients from enclosed institutions •Patients with a Xpert® MTB/RIF Ultra positive test for rifampicin resistance •Patients with positive culture or smear microscopy after the second month of treatment

The use of culture as a diagnostic tool also depends on the local availability in laboratories. The main culture methods currently used for TB diagnosis in Brazil are egg-based solid media, such as Ogawa-Kudoh (OK) or Lowenstein-Jensen (LJ), and BACTEC^TM MGIT^TM 960 which are supplemented with glycerol but not sodium pyruvate, a key component for the optimal growth of Mbo strains [66]. When zTB is suspected, laboratory guidelines recommend adding pyruvate to LJ or OK media to support growth [65]. However, the presence of glycerol can still hinder or even inhibit the growth of certain Mbo strains [67], [68]. As a result, zTB cases may produce false-negative cultures or show prolonged time to detection [18]. In reference centers located in major cities, automated methods such as the BACTEC™ MGIT™ 960 are applied and considered the gold standard for culture [65]. Liquid media generally recover Mbo more effectively than solid media, even without sodium pyruvate [69], but may also require longer incubation periods.

Current sample processing introduces additional challenges. The standard decontamination (liquefaction procedure with NaOH) significantly reduces Mbo viability, lowering the number of recoverable bacilli and contributing to false-negative results; hence; new methods have been proposed [70]. Moreover, human zTB often has extrapulmonary presentations [21], which restricts or prevents bacillus detection in sputum, the most used sample for routine TB diagnosis [21]. Alternative specimens, such as biopsy, are difficult to obtain and even more challenging to culture [71]. Together, these factors illustrate how the current diagnostic workflow in Brazil is poorly suited to detect zTB.

1.7. Proposal for inclusion of zoonotic TB risk variables in national surveillance systems and triggers for specific diagnostics

To strengthen the detection and surveillance of zTB in Brazil, we propose the integration of a targeted set of epidemiological risk variables into the national TB notification and surveillance platforms, namely SINAN and TBWeb (TB Notification System of the State of São Paulo). The incorporation of these variables would allow the health system to better identify individuals at higher risk of infection with Mbo, and to guide laboratory and clinical workflows more effectively. These variables are related to occupational, exposure and behavioral risk data at the time of TB notification and are listed in Table 3.

Table 3.

Proposed variables to be included in tuberculosis notification systems.

	Proposed variables	Categories
1	zTB related occupation	•Agricultural and livestock producers/farmers, workers, and supervisors •Wildlife professionals •Animal health, food safety, and agricultural inspection personnel •Hunters and game meat processors •Butchers and slaughterhouse workers •Artisanal dairy production and processing workers •Industrial dairy production and processing workers •Meat processing and preservation workers •Veterinarians •Veterinary technicians and animal science professionals •Household contacts of the above individuals
2	zTB related exposure	•Direct contact with livestock or wildlife •Direct contact with livestock or wildlife diagnosed with tuberculosis
3	zTB food consumption habits	•Regular intake of unpasteurized milk or dairy products •Regular intake animal-origin products from informal or unregulated sources

Integrating these variables as structured fields withing SINAN and TBWeb can enhance the epidemiological sensitivity of the system for detecting potential zoonotic transmission routes. When one or more of these risk factors are present, they should automatically trigger a diagnostic workflow that includes the request for culture on selective media suitable for Mbo. We also propose that these variables can be included in the laboratory culture request. Even if these variables are collected only at the time of notification (when the disease was already diagnosed), their systematic inclusion in SINAN/TBWeb can raise awareness among healthcare workers and improve the likelihood of identifying zTB cases.

Another important trigger for a Mbo-specific diagnostic workflow arises from molecular–culture discordant cases, where samples yield a strong positive result on Xpert MTB/RIF or Ultra (e.g., low Ct values) but fail to grow in culture [72]. This pattern may suggest infection by Mbo, which exhibits more fastidious growth requirements than Mtb [70]. In such cases, laboratory professionals should consider repeating the culture in liquid media using automated systems like BACTEC™ MGIT™ 960 with added pyruvate [73] or using Mbo-specific solid media, such as Stonebrink. Species identification can then be performed with species-specific PCR, GenoType MTBC, FT-IRS, WGS, or equivalent molecular methods. When local capacity is limited, specimens should be referred to reference laboratories equipped for culture, species identification, and DST.

Finally, risk-based surveillance should be considered, particularly for individuals with known exposure to infected animals or herds. The protocol implemented in state of Santa Catarina, which mandates active surveillance of farmers, producers and household contacts linked to confirmed bovine TB outbreaks [50], provides a model that could be scaled up nationally, and expanded to institutions that handle captive wildlife [zoological gardens, wildlife assessment/rehabilitation centers (CETAS), aquariums, wildlife breeding facilities etc]. The integration of this same workflow at the national level, supported by shared notification between the agricultural, wildlife and health sectors, would operationalize a One Health response, ensuring early detection of zoonotic transmission events.

We suggest a new protocol for detection of zTB (Fig. 1) considering three triggers for a Mbo-specific diagnostic workflow: epidemiological risk factors, molecular–culture discordance, and active surveillance of groups exposed to animal TB outbreaks. This protocol will align the Brazilian public health system with international recommendations for integrated human-animal health monitoring systems. Furthermore, it would enhance the country capacity to generate robust data on the burden and characteristics of zTB, supporting more effective prevention, diagnostic, and treatment strategies.

Fig. 1.

Proposed protocol for the identification of zoonotic tuberculosis (zTB) cases within the Brazilian public health system. Three triggers for a Mycobacterium bovis specific diagnostic workflow are shown: suspected TB cases with epidemiological risk factors (occupation, food consumption habits and exposure); TB cases that show strong positive in Xpert MTB/RIF Ultra and a negative culture; human contacts of animal TB outbreaks in livestock and/or wildlife. SINAN: Information System for Notifiable Diseases. SITE-TB: Information System for Special Tuberculosis Treatments. FT-IRS: Fourier transform infrared spectroscopy. IGRA: Interferon gamma release assay.

2. Conclusion

zTB remains a neglected public health challenge, yet it is of significant relevance in Brazil. The persistent coexistence of endemic bovine TB, social vulnerability, and consumption of unpasteurized dairy products creates conditions conducive to sustained zoonotic transmission, which is rarely detected, reported, or adequately treated. To improve detection of zTB in Brazil, we propose integrating risk-based criteria into sample selection for culture, ensuring that individuals with relevant exposures are prioritized for testing. Short-term priority actions to address the main challenges associated with zTB include developing and implementing official policies and guidelines with mandatory compliance rather than merely advisory; adopting effective active surveillance strategies and routine case assessments to ensure continuous monitoring of groups with occupational and exposure risks; and expanding the use of specific diagnostic tools to generate more accurate and representative data on the incidence of Mbo infections in humans, particularly in regions where the pathogen is endemic. In parallel, ongoing education of at-risk populations, including rural producers, veterinarians, and food handlers, is essential. Without these measures, zTB will continue to be a hidden component of the TB burden in Brazil, undermining control strategies and perpetuating health disparities. Importantly, the absence of systematic detection cannot be interpreted as evidence of absence: without targeted surveillance, the true burden of zTB will remain unknown.

CRediT authorship contribution statement

Ana Marcia Sá Guimarães: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Kevim Bordignon Guterres: Writing – review & editing, Writing – original draft, Visualization, Validation, Methodology, Investigation, Formal analysis, Data curation. Maria Clara da Silva Galrão Cunha: Writing – review & editing, Writing – original draft, Visualization, Validation, Methodology, Investigation, Formal analysis, Data curation. Erica Chimara: Writing – review & editing, Visualization, Validation, Methodology, Investigation, Formal analysis. Maria de Lourdes Viude de Oliveira: Writing – review & editing, Visualization, Validation, Methodology, Investigation, Formal analysis. Flábio Araújo: Writing – review & editing, Visualization, Validation, Methodology, Investigation, Formal analysis.

Funding source

The funders have no role in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; and in the decision to submit the paper for publication.

Funding

This study was funded by a grant from the National Council for Scientific and Technological Development (CNPq, atendimento@cnpq.br), grant 445745/2023-1 (Chamada N° 29/2023—Pesquisas em tuberculose para o fortalecimento da vigilância e controle da doença) and by São Paulo Research Foundation (FAPESP), grant 13388-9/2023. KBG's fellowship is funded by FAPESP grant 2023/07582-7. AMSG is a fellow of CNPq (311543/2023-5). Coordination for the Improvement of Higher Education Personnel (CAPES), grant 001, provided graduate studies support.

Declaration of competing interest

We declare no competing interests.

Appendix A. Supplementary data

Supplementary material: Table S1.

Data availability

Data will be made available on request.