A 35-year-old male laboratory worker was referred to a physician with recurrent fever, headache, arthralgia, weakness, and constipation, which had begun 1 month before. The patient worked in a laboratory that produced antigens for diagnostic use. Three weeks before symptoms began, he had been handling a dense culture of live B. canis M- and was using no personal protection; the procedures were not performed in a biological safety cabinet. Moreover, the patient had attempted resuspension by repeated pipetting with his mouth. The clinical examination disclosed cervical adenomegaly, and laboratory tests indicated a mild increase of hepatic enzymes (Aspartate aminotransferase 46 μ/L and alanine aminotransferase 65 μ/L) and neutropenia. The patient reported not having close contact with dogs or other animals. Taking into account the unprotected exposure to B. canis M-, brucellosis was suspected, and blood samples were drawn for culture and serologic studies. Two weeks later blood cultures indicated a Brucella species that was later typed as B. canis. Conventional tests for antibodies to smooth brucellae (agglutination, complement fixation) yielded negative results. In contrast, slide agglutination for B. canis was strongly positive with undiluted serum and was also positive at 1:10 dilution. Serologic tests for hepatotropic viruses and Toxoplasma gondii were negative.

After diagnosis, a course with oral doxycycline, 100 mg twice a day for 42 days, plus parenteral gentamicin, 180 mg once a day for 10 days, was started. The patient clinically recovered, but on the last day gentamicin was administered, symptoms of the left VIII cranial nerve occurred, which resolved with flunarizine and vitamin B12 administration.

Blood cultures performed 2 weeks after antimicrobial therapy ended were negative for B. canis. During follow-up, the patient remained asymptomatic, his cervical adenitis resolved, and serum levels of hepatic enzymes returned to normal. On his last visit, 4 years after infection, the patient was asymptomatic. While he continues to handle B. canis M-, cultures are now performed under strict biological safety measures (biological safety cabinet, personal protection including goggles, gloves and mask, and autoclaving of contaminated material).

The main finding of our study is that the M- strain of B. canis can produce human disease, which was unexpected in view of the reported avirulent phenotype of this strain in dogs. To the best of our knowledge, this case is the first of human B. canis M- infection ever reported. The M- strain has been widely used for the diagnosis of canine brucellosis because it is less prone to autoagglutination than its wild-type counterpart (called m+). Based on the low virulence of the M- strain in dogs (4), the production protocol of the laboratory where this case occurred did not include bacterial inactivation or personnel protection during initial handling of cultures, which led to a prolonged exposure to a high number of viable bacteria. As the patient was the only person involved in the production of this strain, his co-workers were not tested for B. canis M-. Similar illness in the production plant was not reported. Because Brucella spp. is not usually transmitted from patients to healthy persons, the patient’s family members were not tested for B. canis M- infection.

The clinical manifestations in our patient were similar to those reported for human, wild-type B. canis infections (e.g., fever, headache, anorexia, asthenia, and adenitis). Previous studies in dogs experimentally infected with the M- strain showed that this strain does not revert to the m(+) phenotype in vivo (4). Our case may be analogous to cases of human illness by attenuated strains of Brucella species used for animal vaccination, mainly B. melitensis Rev-1 (9) and B. abortus S19 (10). Altogether, these human infections indicate that attenuation for animals does not necessarily mean immunity for humans and that biological safety measures must be followed in each case.

Figure

Figure. Serological follow-up of a human infection by Brucella canis M-. CP, cytoplasmic proteins; BLS, Brucella lumazine synthase; HS, B. canis hot-saline extract. The enzyme-linked immunosorbent assay titer was calculated as the...

To assess the humoral immune response to the infection with the M- strain, antibodies against outer membrane antigens (HS) and to internal antigens (CP and Brucella lumazine synthase) were measured. Overall, low titers of antibodies were found by all tests, which is similar to those found for M- infections in dogs (4). Low antibody titers also could be related to early administration of antimicrobial therapy, as has been shown in patients infected with smooth brucellae (11). Antibodies to both external and internal antigens declined after antimicrobial therapy was begun and were undetectable 6 months after diagnosis (Figure). This decline, with longer persistence of antibodies to external antigens, is in agreement with our previous findings in human infections by smooth Brucella species (12).

An early and strong cellular Th1-type response to Brucella internal antigens developed in this patient, in agreement with our previous observations in acute human brucellosis (8). The reasons for the decline of this response during follow-up are unknown, but conceivably, bacteria levels were substantially diminished by the early antimicrobial therapy, thus eliminating the internal antigens needed to develop a long-lasting cellular immune response.

In summary, this case shows that, in spite of its reduced virulence in dogs, B. canis M- can produce human disease with a clinical picture similar to that produced by the infection with wild-type strains of B. canis. Therapeutic and immunologic parameters seem to be very similar to those observed in infections by smooth brucellae.