Mycobacterium bovis

INTRODUCTION — The Mycobacterium tuberculosis complex includes M. tuberculosis (the cause of most human tuberculosis), M. bovis, M. bovis Bacillus Calmette-Guérin (BCG, the vaccine strain), M. africanum, M. microti, and others [1]. All are gram-positive, aerobic, and acid-fast bacilli.

M. bovis is the main cause of tuberculosis in cattle, deer, and other mammals and shows some features distinct from M. tuberculosis in humans. M. bovis may have transitioned from humans to animals; animal domestication facilitated spread and maintenance in animals [2], although the strains are remarkably related genetically (99.95 percent) [3]. Novel molecular methods such as whole-genome sequencing, mycobacterial interspersed repetitive-unit-variable-number tandem-repeat typing, and spoligotyping are now contributing to improved understanding of evolution and strain features [4].

The epidemiology, transmission, clinical manifestations, diagnosis, treatment, and prognosis of human M. bovis (referred to as "zoonotic tuberculosis") will be reviewed here. Issues related to the BCG vaccine strain are discussed separately, as are issues related to BCG for treatment of bladder cancer. (See "Prevention of tuberculosis: BCG immunization and nutritional supplementation" and "Infectious complications of intravesical BCG immunotherapy".)

EPIDEMIOLOGY AND RISK FACTORS

Epidemiology

Worldwide — The burden of zoonotic tuberculosis is underestimated as a cause of tuberculosis in humans, due to the absence of systematic surveillance in countries where bovine tuberculosis is endemic, lack of awareness and expertise among health care providers and because the usual laboratory procedures used to diagnose human tuberculosis do not reliably differentiate M. bovis from M. tuberculosis [5].

Despite these limitations, the World Health Organization estimates that M. bovis caused 143,000 new cases and 12,300 deaths in 2018 [6]. However, only 16 countries reported within this surveillance effort, which in part explains the broad range for that estimate (71,000 to 240,000). Stated as a proportion, M. bovis caused approximately 1.4 percent of the 10 million incident pulmonary tuberculosis cases globally [6]. Among patients with extrapulmonary tuberculosis, the proportion of M. bovis is higher because human M. bovis infection generally occurs in the setting of consumption of infected cow's milk products, so scrofula (infection of the lymph node[s] in proximity to the mouth and esophagus) and gastrointestinal disease are typical clinical manifestations [7].

In developed countries where bovine tuberculosis is controlled and dairy products are pasteurized routinely, the proportion of M. bovis infection among human tuberculosis cases is often lower than the global estimate. As an example, in the United Kingdom, M. bovis caused approximately 0.5 percent of culture-confirmed human tuberculosis in 2007 [8] and has been decreasing: between 2005 and 2008, the annual incidence of M. bovis decreased from 0.065 to 0.047 per 100,000 population [9]. In a Spanish hospital, M. bovis caused approximately 0.95 percent of tuberculosis cases between 1980 and 2003 [10]. However, the proportion may be higher among certain at-risk populations; as an example, one retrospective study among HIV-infected patients in France noted that M. bovis infection accounted for 1.6 percent of tuberculosis cases [11].

In resource-limited settings, reliable and representative national data regarding the relative frequency of human tuberculosis due to M. bovis are limited [12]:

●In an analysis of more than 1100 human tuberculosis isolates obtained in Mexico City between 2000 and 2014, M. bovis accounted for 26 percent of isolates (from all sites) and 16 percent of pulmonary samples [13].

●In a systematic review from Nigeria including 17 studies and 1693 mycobacteria isolates obtained between 1975 and 2014, 1.3 percent of the M. tuberculosis isolates were identified in cattle, while 8 percent of the M. bovis isolates were isolated from humans [14].

●Among 1599 isolates from acid-fast bacilli-smear positive patients in Ethiopia between 2011 and 2013, molecular typing identified two isolates (0.13 percent) as M. bovis [15].

●In a Chinese study including three years of data from more than 5000 patients in a region with known M. bovis cattle infection, M. bovis accounted for 0.34 percent of human tuberculosis cases [16]. In a subsequent survey among more than 4000 clinical tuberculosis strains isolated from sputum between 2007 and 2009, none were M. bovis [17].

United States — Among 59,273 cases of tuberculosis in the United States between 2006 and 2013, M. bovis accounted for 1.3 to 1.6 percent of cases [18]. In some settings, the proportion is higher and may be increasing.

As an example, in a retrospective analysis of tuberculosis cases between 1994 and 2005 in San Diego, California, the incidence of M. bovis increased from 0.65 to 0.93 per 100,000 population. Because the overall incidence of M. tuberculosis declined, the annual proportion of culture-positive tuberculosis cases attributable to M. bovis increased from 5 to 11 percent [19]. This was particularly striking among children <15 years of age in whom M. bovis comprised 54 percent of tuberculosis cases [19]. In a study of 86 patients with HIV and tuberculosis coinfection between 2000 and 2007 in southern California, 35 percent had M. bovis infection [20]. A subsequent review of the California tuberculosis registry showed that the percentage of tuberculosis cases attributable to M. bovis increased between 2003 and 2011, from 3.4 to 5.4 percent [21].

Risk factors — Among patients diagnosed with tuberculosis, the possibility of M. bovis should be considered in patients who may have consumed unpasteurized dairy products or had contact with infected animals or humans [22]. A systematic review of occupational zoonotic tuberculosis found livestock farmers, veterinarians and their assistants, abattoir workers, hunters, and workers in contact with wildlife at risk [23]. M. bovis should also be considered in the setting of extrapulmonary disease [20,21,24,25], immunosuppression [20,21], pyrazinamide monoresistance [26], Hispanic ethnicity [19,21], and children <15 years of age with tuberculosis [19]. Multivariate analysis of the California tuberculosis registry also identified diabetes as independently associated with M. bovis disease [21]. Among children with M. bovis tuberculosis during 2010 to 2011, all had >1 parent/guardian born in Mexico, compared with 38 percent for children with M. tuberculosis.

Among 165 patients in the United States with M. bovis who underwent genotyping from 1995 to 2005, 89 percent were Hispanic individuals, 62 percent were born in Mexico, 19 percent were < 15 years old (compared with only 2 percent in those with M. tuberculosis), 65 percent had extrapulmonary disease, and 26 percent were HIV infected [27]. Among 511 patients with tuberculosis in Italy, 1.76 percent were caused by M. bovis. M. bovis was significantly more common among those with extrapulmonary disease and older adults [28]. In a retrospective study including 35 Mexican patients with M. bovis infection, 51 percent were children, 69 percent had malnutrition, 51 percent consumed unpasteurized milk, and 6 percent had contact with animals [25]. Another retrospective study including 39 patients with confirmed M. bovis in Argentina noted strong risk factors included occupational exposure (65 percent), a history of living in a rural area (31 percent), and consumption of unpasteurized milk (4 percent) [22].

In a retrospective review of 533 Mexican patients with a cultured isolate of M. tuberculosis complex, M. bovis was the cause in 30 percent of cases [29]. Patients with M. bovis were more likely to be younger, using glucocorticoids, and to have extrapulmonary disease. Among the subset of patients with only a pulmonary isolate of M. bovis, smoking was an additional risk factor for M. bovis. There was no difference between M. bovis and M. tuberculosis with regards to the proportion with positive sputum smear or cavitation on chest radiograph.

Transmission — M. bovis may be transmitted from humans to animals, animals to humans, between animals [30], or between humans. People can become infected with M. bovis via oral ingestion, droplet inhalation, or cutaneous penetration [7].

M. bovis has been identified in water and soil samples; it is uncertain whether this may be a source of disease transmission [7].

●Animal to human – M. bovis can be transmitted from animals to humans via ingestion of animal products or via airborne particle inhalation [7].

•Cattle – Humans can become infected with M. bovis via ingestion of infected dairy products (the most common mode of M. bovis infection transmission) [31] or via airborne particle inhalation.

In one surveillance study among more than 300 farm workers in Mexico exposed to cattle with M. bovis infection (many of whom consumed unpasteurized milk), more than half of individuals had evidence of latent tuberculosis infection [31]. A positive tuberculin skin test (TST) was observed in 76 percent, and 58 percent had a positive interferon-gamma release assay; this prevalence is higher than other high-risk populations in Mexico, including individuals in border cities (57 percent), migrant workers (26 percent), and health care workers (up to 64 percent). Two cases of active M. bovis infection were identified, one of which matched the prevalent strain found in the cattle.

Among 129 cases of human M. bovis infection in the United Kingdom between 2005 and 2008, most patients were older than 65 years, suggesting that reactivation may have occurred following latent infection acquired prior to widespread pasteurization of milk [9]. In another study including 70 livestock traders in Nigeria, 10 percent had sputum culture positive for the M. tuberculosis complex; of these, 28 percent were further identified as M. bovis. Work in the livestock trade for at least three years was a risk factor for infection [32]. A genetic characterization of M. bovis strains in Burkina Faso suggests that M. bovis transmission occurs mainly between cattle within and beyond Burkina Faso, but also occasionally from cattle to humans and potentially between humans [33].

•Other animals – Other animal species can transmit M. bovis to humans, although this is less common than transmission from cattle. One report in 1990 noted epizootic M. bovis infection from domesticated elk (Cervus elaphus) in Alberta, Canada. Among 394 humans with animal contact, 22 percent were noted to have positive TSTs, and most were known to have had contact with culture-positive animals [34]. In addition, transmission of M. bovis from deer to hunters in Michigan (where M. bovis is endemic in deer) has been described [35,36]. Rare cat-to-human transmission has also been suggested by a United Kingdom study [37].

●Human to human – Human-to-human transmission of M. bovis infection occurs rarely. Case reports include descriptions of transmission among homeless shelter residents, within households, at a church, and between siblings on a farm in the United Kingdom, and from a Dutch health care worker [38-42]. In one cluster of six cases in the United Kingdom, the source case had prior contact with cattle and consumed unpasteurized dairy products [40]; all subsequent cases were visitors to bars with favorable conditions for transmission (eg, prolonged close contact in a confined space with poor ventilation).

Nosocomial transmission of multidrug-resistant M. bovis has also been described [43-46].

●Human to animal – M. bovis can be transmitted from humans to animals presumably through airborne particle inhalation. One such example is described in a report of M. bovis transmission between a human and two indoor-housed pet cats [47].

Drug resistance — M. bovis is intrinsically resistant to PZA; apart from PZA resistance, primary and acquired drug resistance has been described [26,43-46]. In a report from Mexico City between 2000 and 2014, isoniazid, rifampin, and streptomycin resistance in previously untreated patients was higher among M. bovis than M. tuberculosis isolates (10.9 versus 3.4 percent) and appeared to be increasing (from 3.4 percent [2000 to 2004] to 7.6 percent [2010 to 2014]). Resistance rates to at least isoniazid and rifampin in patients who were previously treated for M. bovis and M. tuberculosis were 38.5 percent and 34.4 percent, respectively [13].

Data regarding genetic mechanisms for drug resistance in M. bovis are emerging [43,48,49]. An analysis including 2074 M. bovis isolates recovered from non-human hosts employed whole genome sequencing to identify the most frequent mutations associated with tuberculosis drug resistance; M. bovis isolates were found to harbor the same mutations that confer resistance to both first- and second-line antimycobacterial antibiotics [50].

CLINICAL MANIFESTATIONS — Tuberculosis due to M. bovis is clinically and radiographically indistinguishable from tuberculosis due to M. tuberculosis. Like M. tuberculosis, M. bovis can manifest with primary and reactivation forms; involvement may be pulmonary, extrapulmonary, or disseminated. When M. bovis is acquired via ingestion of contaminated dairy products, extrapulmonary disease is more likely than pulmonary tuberculosis [20,24,25]. Extrapulmonary sites of M. bovis include lymph nodes (scrofula) [51], pleural space [10], joints [52,53], eye [54], the central nervous system [55,56], and abdominal organs [57,58].

The clinical manifestations, physical examination findings, and radiographic abnormalities of tuberculosis are discussed in detail separately. (See "Pulmonary tuberculosis: Clinical manifestations and complications".)

DIAGNOSIS — The initial diagnostic approach for M. bovis is identical to the diagnostic approach for M. tuberculosis, including acid-fast bacilli staining and mycobacterial culture of relevant specimens, and molecular testing for mycobacterial DNA (such as GeneXpert MTB/RIF) [59]. Both the tuberculin skin test and interferon-gamma release assays can detect M. bovis infection [60]. (See "Tuberculosis infection (latent tuberculosis) in adults: Approach to diagnosis (screening)".)

The United States Centers for Disease Control and Prevention the National Tuberculosis Genotyping Service which allows differentiation of M. bovis from M. tuberculosis [61]. However, distinguishing M. bovis from other M. tuberculosis complex (MTBC) strains is not done in most laboratories outside the United States. Additional laboratory investigation should be requested when there is epidemiologic risk for M. bovis, failure to respond clinically to initial empiric therapy, or susceptibility testing that demonstrates pyrazinamide (PZA) resistance.

Identification of PZA monoresistance in an MTBC strain should prompt consideration of M. bovis; however, dependence solely on PZA resistance testing to identify M. bovis is not appropriate because PZA resistance testing is technically difficult and is not performed routinely or reliably in many laboratories. In a United States surveillance study performed between 1995 and 2005, the sensitivity of PZA susceptibility for identification of M. bovis was about 80 percent [27]. A subsequent report including 500 M. bovis isolates noted that 7 percent were incorrectly reported as PZA susceptible [26].

For conclusive identification of M. bovis, the isolate should be sent to a public health laboratory or a mycobacterial reference laboratory. Mechanisms for differentiation of M. bovis from M. tuberculosis include biochemical assays (niacin test, heat-sensitive catalase, nitrate reduction), susceptibility to an isoniazid analogue (thiophene-2-carboxylic acid hydrazide; TCH), and genomic analysis [1].

DIFFERENTIAL DIAGNOSIS — The differential diagnosis for M. bovis pulmonary disease includes:

●M. tuberculosis disease – Tuberculosis caused by M. tuberculosis is clinically and radiographically indistinguishable from tuberculosis caused by M. bovis. The diagnosis is established via culture, with additional testing for M. bovis as described above. (See "Diagnosis of pulmonary tuberculosis in adults".)

●Nontuberculous mycobacterial infection (NTM) – Symptoms of NTM include fatigue, dyspnea, and occasional hemoptysis; fever and weight loss occur less frequently than in patients with tuberculosis. NTM is distinguished from tuberculosis by culture results and/or molecular diagnostic testing. (See "Overview of nontuberculous mycobacterial infections".)

●Fungal infection – Fungal pneumonia can present with a range of manifestations including pneumonia, pulmonary nodule, and cavitary lung disease. It is distinguished from tuberculosis by epidemiologic exposure and culture results. (See "Diagnosis and treatment of pulmonary histoplasmosis".)

●Sarcoidosis – Sarcoidosis most commonly presents with diffuse interstitial lung disease. It rarely forms cavities and is distinguished from tuberculosis by histopathologic detection of noncaseating granulomas. (See "Clinical manifestations and diagnosis of sarcoidosis".)

●Lung cancer – Lung cancer most commonly presents with cough, hemoptysis, chest pain, and dyspnea. It is distinguished from tuberculosis by histopathology. (See "Clinical manifestations of lung cancer".)

●Lymphoma – Lymphoma typically presents with a rapidly growing mass together with fever, night sweats, and weight loss. It is distinguished from tuberculosis by histopathology. (See related topics.)

TREATMENT — M. bovis is intrinsically resistant to PZA; therefore, the approach to treatment of disease due to M. bovis is extrapolated from experience with the treatment of pyrazinamide (PZA)-resistant M. tuberculosis.

Drug-susceptible disease — For patients with disease due to drug-susceptible M. bovis, treatment consists of two months of isoniazid, rifampin, and ethambutol administered daily, followed by seven months of isoniazid and rifampin [62,63]. The duration of therapy for pulmonary and most extrapulmonary disease should be nine months, although the optimal duration for meningitis has not been established; some experts recommend at least 12 months for such cases [55].

Data to guide treatment of M. bovis disease are limited:

●A systematic review identified studies reporting two cohorts with 156 patients who received isoniazid and rifampin for 6 to 9 months and two cohorts with 113 patients who received isoniazid, rifampin, and ethambutol for 6 to 12 months. Analysis of outcomes suggests that the efficacy of both regimens is adequate (99 percent and 93 percent with and without EMB, respectively), and that there were inadequate data to support a treatment duration of <9 months [64].

●In one study including more than 290 patients with M. bovis (sputum culture-positive; sensitive to isoniazid and rifampin) and more than 30,800 patients with M. tuberculosis (sputum culture-positive; sensitive to PZA, isoniazid, and rifampin) who were treated with initial four-drug therapy between 2006 and 2013, the rate of sputum culture conversion by three months of therapy was similar (85 versus 83 percent); however, more M. bovis patients died during treatment (9 versus 5 percent), suggesting that early culture conversion is not necessarily associated with a favorable clinical outcome [65].

Drug-resistant disease — M. bovis is intrinsically resistant to PZA; apart from PZA resistance, primary and acquired drug resistance has been described. (See 'Drug resistance' above.)

For patients with disease to an M. bovis strain that is resistant to drugs (apart from PZA), or for patients who are intolerant of first line drugs, the treatment regimen should be chosen according to susceptibility testing results with expert consultation, similar to the approach for treatment of multidrug-resistant tuberculosis. (See "Treatment of drug-resistant pulmonary tuberculosis in adults".)

PROGNOSIS — The prognosis for tuberculosis due to M. bovis is worse than for disease due to M. tuberculosis [62,65]. In one study including more than 290 patients in the United States with M. bovis and more than 30,800 patients with M. tuberculosis between 2006 and 2013, the rate of death was higher among patients with M. bovis infection (9 versus 5 percent) [65]. In another study including 35 Mexican patients with M. bovis, 31 percent died [25]. In the Netherlands between 2003 and 2011, mortality due to M. bovis was higher than M. tuberculosis (20 versus 4 percent) [24].

Among patients with HIV infection, the prognosis for M. bovis tuberculosis also appears worse than the prognosis for M. tuberculosis. One cohort including 86 HIV-infected patients with tuberculosis in the San Diego, California, area noted a mortality rate of 10 percent (versus 3.6 percent with M. tuberculosis) [20]. In another report including 19 HIV-infected patients with primary multidrug-resistant M. bovis, all patients died [46].

These observations may be due to delay in recognition of M. bovis infection, insufficient duration of treatment in the setting of pyrazinamide resistance, older patient age (due to reactivation disease following remote cattle exposure), or increased virulence of M. bovis.

PREVENTION — A roadmap to reduce zoonotic tuberculosis was developed by the World Health Organization, the World Organization for Animal Health, and the Food and Agricultural Organization of the United Nations, and the International Union Against Tuberculosis and Lung Disease [6]. The roadmap calls for a multidisciplinary "One Health" approach that includes improved diagnosis and surveillance and reduced transmission between animals and humans.

Immunization with Bacillus Calmette-Guérin and treatment of latent infection are likely to be effective for prevention of M. bovis as M. tuberculosis; further study is needed. Infection control precautions for M. bovis should be followed as for M. tuberculosis [44]. (See "Tuberculosis transmission and control in health care settings".)

SUMMARY AND RECOMMENDATIONS

●Mycobacterium bovis is a member of the Mycobacterium tuberculosis complex (MTBC), which also includes M. tuberculosis. (See 'Introduction' above.)

●In resource-limited settings, minimal data are available regarding the relative frequency of disease due to M. bovis because of limited laboratory facilities for culture and identification. In developed countries, approximately 1 to 2 percent of human tuberculosis cases are attributable to M. bovis; some settings have higher proportions such as the San Diego, California region, where the proportion of culture-positive tuberculosis cases attributed to M. bovis increased from 3 to 11 percent between 1980 and 2005. (See 'Epidemiology and risk factors' above.)

●Risk factors for M. bovis infection include consumption of unpasteurized dairy products and contact with infected animals or humans. Among patients diagnosed with tuberculosis, the possibility of M. bovis infection should be considered in the setting of Hispanic ethnicity, age <15 years, immunosuppression, HIV infection, and extrapulmonary disease. (See 'Risk factors' above and 'Transmission' above.)

●Tuberculosis due to M. bovis is clinically and radiographically indistinguishable from tuberculosis due to M. tuberculosis. Like M. tuberculosis, M. bovis can manifest with primary and reactivation forms; involvement may be pulmonary, extrapulmonary, or disseminated. (See 'Clinical manifestations' above.)

●Establishing a diagnosis of M. bovis is challenging because many laboratories in tuberculosis-endemic settings do not distinguish among members of the MTBC. M. bovis infection should be considered in the setting of an MTBC isolate with pyrazinamide (PZA) monoresistance. State public health laboratories can assist with definitive identification. (See 'Diagnosis' above.)

●For patients with disease due to drug-susceptible M. bovis, we recommend treatment with two months of daily isoniazid, rifampin, and ethambutol, followed by seven months of isoniazid and rifampin (daily or twice weekly) (Grade 1B). Susceptibility testing and expert consultation should guide individualized therapy for polyresistant M. bovis. (See 'Treatment' above.)