The bond between humans and animals has been recognized for many years, and pet ownership has been associated with both emotional and health benefits (1-4). However, pet ownership may also pose health risks through the zoonotic transmission of infectious diseases, especially in the immunocompromised (5). Animal-associated pathogens of concern to immunocompromised persons include Toxoplasma gondii, Cryptosporidium spp., Salmonella spp., Campylobacter spp., Giardia lamblia, Rhodococcus equi, Bartonella spp., Mycobacterium marinum, Bordetella bronchiseptica, Chlamydia psittaci, and zoophilic dermatophytes (2,6). However, with the exception of Bartonella henselae and zoophilic dermatophytes, infections in humans are more commonly acquired from sources other than pets, and the infectious disease risk from owning pets is considered low (2,7). Nonetheless, HIV-infected persons may still be advised not to own pets (8).

Since human medicine often does not delve deeply into the role of animals in the transmission of zoonotic agents (7,9) and veterinary medicine does not cover the clinical aspects of human disease, zoonotic disease control requires involvement of both physicians and veterinarians. We examined how frequently physicians and veterinarians encounter zoonotic diseases, what role physicians think veterinarians should play in zoonotic disease prevention, how often physicians and veterinarians communicate about zoonoses issues, and what physicians and veterinarians perceive as the disease risk of immunocompromised persons from pets.

Our sample populations were drawn from membership lists of the Wisconsin Veterinary Medical Association (WVMA) and the State Medical Society of Wisconsin (excluding retired practitioners). Veterinarians (n = 526) were chosen by a systematic sampling of every third name on the WVMA membership list. Since veterinarians in all types of practice may encounter zoonotic problems, sampling was not stratified by specialty. Physicians (n=698) were chosen by specialty most likely to involve both zoonotic diseases and immunocompromised patients (all physicians who listed infectious disease [n = 38] or hematology/oncology [n = 103] as specialties), as well as randomly selected cohorts of pediatricians (n = 100), and general internal medicine physicians (n = 500). Duplicate names were removed.

Each participant was mailed a cover letter; a number-coded, postage-paid return envelope; and a physician- or veterinarian-specific survey. Nonresponders received a second survey 3 weeks after the first. For questions with a response scale of 1 to 5, the sample size was large enough for a 2-sample Z-test to statistically compare mean responses between physicians and veterinarians. Additionally, responses were analyzed by veterinary practice type and physician specialty. Statistically significant differences in responses by specialty are noted in the text. A 2-sample Z-test employing the standard large sample approximation to binomial data was used to compare the proportions of responses by physicians and veterinarians to the question about zoonotic pathogens of concern to immunocompromised patients.

Surveys were completed by 327 veterinarians and 322 physicians (overall response rate of 53%). Responses from veterinarians were as follows: 142 (43%) small-animal practice, 65 (20%) large-animal practice, 98 (30%) mixed-animal practice, and 22 (7%) exotic-animal practice. The distribution of respondents by practice type is very similar to the distribution by practice type across the State of Wisconsin (46% small animal, 19% large animal, 35% mixed, and 0.3% exotic [data courtesy of M. Mardock, Wisconsin Veterinary Medical Association]), except for an overrepresentation of exotic-animal practitioners. Nationally, 8% of veterinarians are in large animal practice exclusively, 29% in mixed practice, 58% in small animal practice and 5% "other" (data from the Veterinary Economics Statistics Brochure of the AVMA, September 1998). The distribution of responses from physicians included 24 (7%) in infectious diseases, 48 (15%) in hematology/oncology, 53 (16%) in pediatrics and 197 (61%) in general internal medicine (including 16 who specifically categorized themselves as pulmonologists and 11 as rheumatologists). Among our random selection of pediatricians and general internists, the respondent ratio of 3.7 internists for every 1 pediatrician is slightly higher than the statewide ratio of 2.5 (Wisconsin physician data courtesy of M. O'Brien, State Medical Society of Wisconsin) and the national ratio of 2.0 (10).

The survey results indicate that veterinarians (Table 1) encounter zoonotic diseases in their practices or discuss them with their clients more frequently (p < 0.00001) than physicians (Table 2). Among veterinarians, small-animal practitioners encounter zoonoses more frequently than veterinarians as a whole (mean = 2.80, p = 0.05), and large-animal veterinarians less frequently (mean = 3.41, p = 0.001). Among physicians, infectious disease specialists encounter zoonoses more frequently than the overall population of physician respondents (mean = 3.44, p = 0.001), but these specialists still encounter zoonoses problems less frequently than veterinarians (p = 0.05).

When physicians were asked (on a scale of 1 to 5 with 1 = very comfortable and 5 = not comfortable) how comfortable they felt about advising patients on the role of animals in the transmission of zoonotic agents and associated risks, with the exception of infectious disease specialists (whose mean comfort level = 1.92 was significantly [p £ 0.0001] better than that of the overall population of physicians), they responded that they were not very comfortable in this role (mean = 3.69, Table 2); moreover, physicians indicated that veterinarians should play an equal or greater role in advising patients about zoonotic diseases (Table 2). In particular, they suggested that veterinarians should be involved not only in controlling zoonotic disease pathogens in animals, but also in providing information for patients and physicians (Table 2). However, the survey demonstrated a nearly complete lack of communication between physicians and veterinarians about zoonotic disease issues (Tables 1,2). In addition, patients themselves do not appear to view veterinarians as a source of zoonotic disease information. Of 310 veterinarians, 96 indicated that they offer special consultation about additional steps for zoonotic disease prevention if they are aware of the fact that a client is immunocompromised; however, for 205 of 310 respondents, the client's health was never discussed (Table 1).

In the second portion of the survey, we examined the views of physicians and veterinarians on the possible disease risks (from specific animals or pathogens) to immunocompromised persons. Various animals were ranked on a risk scale of 1 to 5 (1 = highest risk to 5 = lowest risk, with an option to respond "unsure"). Veterinarians assigned a higher risk than physicians to reptiles (p = 0.004) and puppies (p = 0.01); physicians assigned a higher risk than veterinarians to cats (p ≤ 0.00001) and kittens (p = 0.001) (Tables 1,2). Physicians and veterinarians were also asked to list the two zoonotic pathogens of greatest concern for immunocompromised persons (Table 3). The two most frequently named pathogens were Salmonella spp. and Toxoplasma gondii. Within this ranking, Salmonella spp. were listed more frequently (p = 0.001) by veterinarians than physicians, and this concern may explain why veterinarians thought that reptiles pose the greatest risk to the immunocompromised (Table 1). (Because of the high prevalence in reptiles of Salmonella infection, immunocompromised persons are advised not to own or handle reptiles [2,11]). In contrast, T. gondii was listed as a potential disease risk more often by physicians (p = 0.001), which is consistent with physician's concern about immunocompromised persons owning kittens and cats as pets (Table 2).

In summary, our survey results indicate that physicians and veterinarians hold very different views about the disease risks from certain animals and infectious agents and communicate very little about zoonotic disease prevention. The perceived risks posed by specific pathogens raise some questions. First, for both Salmonella spp. and T. gondii, contact with pets is not the only, or even the most important, source of infection for humans. Contaminated foods are the most common vehicle of Salmonella spp. infection (12); undercooked meat is also a common vehicle of T. gondii infection. Up to 25% of lamb and pork samples contain Toxoplasma tissue cysts (13). Therefore, although cats are the definitive hosts for T. gondii, cat ownership is not associated with an increase in Toxoplasma seroconversion among HIV-infected persons (14). Secondly, a number of the infectious disease agents (e.g., Borrelia burgdorferi, Histoplasma capsulatum, Blastomyces dermatitidis, and Pneumocystis carinii) listed as zoonotic disease risks by both physicians and veterinarians are not truly zoonotic, but rather shared infections. Both animals and humans are infected, but animals are not the direct vehicles of infection for humans. In addition, cytomegaloviruses of humans and animals are not infectious across species. Finally, it is surprising that infection with B. henselae, the causative agent of cat scratch disease, which also causes bacillary angiomatosis, peliosis hepatis, and other conditions in immunocompromised persons, was listed relatively infrequently by both physicians and veterinarians (Table 3). Exposure to kittens has been clearly implicated as a significant risk factor in the epidemiology of B. henselae (15). Both physicians and veterinarians need to recognize the role of this pathogen in the zoonotic infection of immunocompromised persons and the role of cats in its transmission.

Our finding that 205 of 310 veterinarians never knew a client's immunocompromised condition is consistent with a previous study in which only 21% of HIV patients felt most comfortable in asking their veterinarian about the health risks of pet ownership (16). Through approaches such as small signs in exam rooms, zoonotic disease brochures in reception areas, comments in practice newsletters, and affiliation with support groups in the community, veterinarians can encourage immunocompromised persons to avail themselves of the diagnostic and preventive measures that can be provided for zoonotic agents.

Our results suggest that communication between physicians and veterinarians about zoonotic diseases is largely absent. Enhancing such communication could help prevent transmission of zoonotic agents. In addition to directly contacting veterinary practitioners in their community, physicians can also contact their state health departments for information, since some health departments have public health veterinarians on staff. Links between the professions on a broader scale (e.g., through combined veterinary/medical student training and continuing education) to foster a broader consensus about zoonotic disease risks and prevention should also be encouraged.

Top