Human-to-Human Transmission of Tuberculosis Caused by Mycobacterium bovis in Immunocompetent Patients

Abstract

Human-to-human transmission of Mycobacterium bovis in two immunocompetent patients from the same family was confirmed by spoligotyping (pattern F35, which was only observed in cattle from the same area in France). A single allelic difference between animal and human isolates was observed with mycobacterial interspersed repetitive units containing variable-number tandem repeats, suggesting a jump across the species barrier.

CASE REPORTS

Case 1.

In April 2004, a 50-year-old man, born in France, was admitted to the Respiratory Disease Department of the Limoges Hospital for pulmonary tuberculosis. Since 1987, he had been working in a slaughterhouse and handled offal. Clinical manifestations were cough and asthenia. The chest radiograph showed a cavitation of the left upper lobe. The tuberculin skin test (TST) reaction was considered positive. Gastric aspirations were positive for acid-fast bacilli (AFB), and cultures were secondarily positive for M. bovis. M. bovis was identified by phenotyping methods, and identification was then confirmed by molecular analysis (9). The human immunodeficiency virus (HIV) antibody test was negative. Therapy consisted of an association of isoniazid (INH), rifampin (RMP), and ethambutol (EMB). Sample cultures became negative after 1 month of treatment. EMB was discontinued after 3 months; INH and RMP were continued for a total duration of 12 months. Contact tracing was performed. Only his daughter had a positive TST, but she refused any prophylactic treatment.

Case 2.

In May 2007, the daughter of case 1, a 20-year-old woman, born in France, was admitted to the Tours Hospital for suspicion of pulmonary tuberculosis. She had had productive cough with night sweats, thoracic pain, and asthenia for 1 month before hospitalization. The chest radiograph showed a cavitation of the left upper lobe with mediastinal adenopathy confirmed by computed tomography. Two sputum specimens and one gastric aspiration were positive for AFB. Cultures grew M. bovis. The HIV antibody test was negative. Therapy was initiated with an association of INH, RMP, and EMB and continued for 3 months. Sputum smears and culture were negative after 1 month of treatment. EMB was discontinued after 3 months of treatment, and INH and RMP were continued for a total duration of 9 months. Contact tracing was conducted and identified one contact, the patient's grandmother, whose TST reaction was positive at 20 mm. Thus, she received a regimen of INH and RMP for 3 months to treat her latent tuberculosis infection.

Case 1 worked in the Limoges slaughterhouse from 1987 to 2004 and was certainly occupationally exposed during this period. On the other hand, case 2 was born after hygienic measures were applied and she had never traveled to countries where they were not applied. Moreover, she had no contact with cattle herds. Therefore, it is most likely that she was infected by her father.

We compared the M. bovis strains from cases 1 and 2 using spoligotyping and mycobacterial interspersed repetitive units containing variable-number tandem repeats (MIRU-VNTR). Our results confirmed the genetic link between the two strains: no difference was found by spoligotyping and on 10 MIRU-VNTR loci considered to have a high discriminatory power for M. bovis species (ETR-A, -B, -C, and -D; QUB 11a and 11b; QUB 26; MIRU 23; VNTR 3232; and VNTR 3336) (1) (Fig. 1).

FIG. 1.

Unweighted-pair group method with arithmetic average) tree based on combined data, spoligotypes, and MIRU-VNTR of the nine samples under study (human and cattle isolates).

In order to test the hypothesis of a probable occupational exposure to M. bovis for case 1, spoligotypes of cattle strains of M. bovis originating from the Limoges region and other regions in France during the period from 1987 to 2005 were analyzed. Some of the animal strains collected in Limoges between 1997 and 2005 belonged to a peculiar and original spoligotype. This pattern, called F35, was not present in France except in the Limoges area (5). Based on spoligotyping results, the M. bovis isolates from our two cases were identical to the F35 type. These results strongly suggest that case 1 was probably contaminated during the period from 1997 to 1998. Results of the MIRU-VNTR analysis performed on F35 cattle strains were compared with the father's and daughter's isolates (Fig. 1). Except for QUB 26, the rest of the markers tested were identical. QUB 26 was the only marker able to distinguish F35 cattle strains from the intrafamilial strains.

Human tuberculosis caused by Mycobacterium bovis is infrequent in industrialized countries (approximately 0.5 to 1.5% of all culture-positive cases of human tuberculosis) (3, 6, 8). In historical descriptions in Western Europe, acquisition of M. bovis by humans was mainly related to ingestion of unpasteurized milk from infected cows (3). In France, the widespread pasteurization of milk and the systematic slaughtering of tuberculin test-positive cattle have been general practice since 1955 and 1963, respectively. Consequently, most human cases of zoonotic tuberculosis today occur in people born before the introduction of these measures or originating from countries where they are not applied (2), as recently observed in the Hispanic population in the United States (6). Respiratory transmission is possible, mainly in people regularly handling carcasses or offal contaminated with M. bovis, such as slaughterhouse workers (3).

We have reported here on intrafamial transmission of human tuberculosis caused by M. bovis in two immunocompetent patients.

In a nationwide U.S. epidemiologic study, it has recently been shown that patients of Hispanic origin, aged less than 15 years, HIV positive, and with extrapulmonary disease were more likely to be infected by M. bovis than by M. tuberculosis (6). This suggests M. bovis has two different routes of transmission: airborne transmission in older patients with pulmonary disease and food-borne exposure in younger patients with extrapulmonary disease, via contaminated unpasteurized dairy products (6). Person-to-person transmission of M. bovis has been reported as nosocomial in HIV-infected adults (10) and community acquired in immunocompromised patients (4). Two other clusters with interhuman transmission have been suggested, but contact with infected cattle (13) or consumption of unpasteurized milk in Mexico (7) cannot be ruled out as possible sources of M. bovis transmission.

Molecular methods, today the best tools available in epidemiological studies of M. tuberculosis and M. bovis, were employed to test the hypothesis of genetic link between the two intrafamilial cases and to compare human strains to cattle strains. The three most frequently used techniques are (11): restriction fragment length polymorphism (RFLP) typing using IS6110 as the hybridization target (IS6110 RFLP), spoligotyping, and DNA fingerprinting by characterization of MIRU-VNTR. IS6110 RFLP provides limited discrimination among M. bovis isolates (1). Indeed, most strains of this species of the M. tuberculosis complex usually possess only one or a few copies of IS6110, which are not enough to obtain discriminant results with this technique. In recent studies, a combination of spoligotyping and MIRU-VNTR has proven to be a practical and discriminatory method in epidemiological research (1).

The difference between human and animal strains observed only on locus QUB 26 could be due to the absence of the real causative strain in humans among the collection of characterized strains isolated from animals. Either the infected bovine was not found, or the animal was coinfected with at least two variants of the strain. In the latter hypothesis, the strain found in case 1 would not have been isolated in the animal. Nevertheless, bacterial subpopulations with slight variations in fingerprints due to evolutionary events may arise (12). This variant strain defined by a minor allelic difference may have appeared in cattle. It may also have appeared in case 1 when the cattle strain was in contact with a new host. The crossing of the species barrier may be at the origin of the genetic evolution.

In conclusion, the genetic identity between these two strains based on spoligotyping and MIRU-VNTR demonstrated human-to-human transmission from the father to his daughter. This type of transmission has been reported in immunodeficient subjects (4, 10) and in immunocompetent subjects with persisting doubt regarding a zoonotic source of infection (7, 13). An important feature of our case report is transmission of M. bovis between two immunocompetent persons.