Between May and September 2005, three unrelated intact male Yorkshire terriers with a mean age of 4.3 years from Catania, Sicily, were brought to a local veterinarian; the dogs had the following histories: anorexia and lethargy of 2 days' duration (dog 1); anorexia, lethargy, and intermittent lameness of a few days' duration (dog 2); and intermittent vomiting, anorexia, and lethargy of a few days' duration (dog 3). Despite living mostly indoors, all 3 dogs had a recent history of tick exposure. All dogs had received current vaccination histories and had no history of serious illness. Results of the physical examination and hematologic, biochemical, and serum electrophoresis abnormalities at the time of onset of clinical signs and after 1 month (dogs 2 and 3) and 2 months (dog 1) of follow-up are provided in Table 1. Treatments instituted for all 3 dogs at onset of illness are described in Table 1.

EDTA-blood and serum samples were obtained by the attending veterinarian at the time of clinical assessment (before treatment), then 1 week later and 1 month (dogs 2 and 3) or 2 months later (dog 1). DNA extraction was performed from whole blood samples (5,7). A quantitative PCR (qPCR) for detection of Rickettsia spp., Anaplasma phagocytophilum, Ehrlichia canis, and Leishmania infantum in DNA samples was performed by using a Light Cycler (Roche, Mannheim, Germany). PCR amplification was carried out with Rickettsia (Rr-prim3 5´-GAAACCGAAAGAGAATCTTCCGAT-3´ and Rr-prim4 5´-TCCTAGTGTAGAGGTGAAATTCTTA-3´ [8]), E. canis, A. phagocytophilum (fragment of 16S rRNA gene), and L. infantum LCSet primers and probes following manufacturer's instructions (TIB Molbiol, Centro Biotecnologie Avanzate, Genova, Italy) (5,7). Conventional Babesia genus PCR was performed (9). Borrelia burgdorferi sensu lato qPCR was performed by a commercial laboratory (http://www.scanelis.com). PCR results for all infectious agents listed above, with the exception of Rickettsia, were negative in all dogs.

PCRs for Rickettsia that use the outer membrane protein A (ompA) gene to amplify 632 bp (10) and 212 bp (107F 5´-GCTTTATTCACCACCTCAAC-3´ and 299R 5´-TRATCACCACCGTAAGTAAAT-3´) (7) amplicons were performed. For dog 1, a 632-bp amplicon was cloned by using the TOPO TA Cloning (Invitrogen, Carlsbad, CA, USA) and sequenced (GenBank accession no. DQ518245) (7). For dogs 2 and 3, a 212-bp amplicon was subjected to direct sequencing (accession no. DQ518246, DQ518247) (7). PCR results are summarized in Table 1.

Consensus sequences were aligned [(BIOEDIT version 7.0 (ClustalW)] with known sequences in GenBank using the basic local alignment search tool (BLAST; available from http://www.ncbi.nlm.nih.gov/BLAST/). The sequence obtained from all 3 dogs was 100% homologous to a portion of the complete genome sequence corresponding to the ompA gene from R. conorii (Malish 7, accession no. AE008674).

Immunofluorescent assays to detect antibodies to R. rickettsii, R. conorii, B. burgdorferi sensu stricto, E. canis, Babesia canis, A. phagocytophilum, L. infantum, Bartonella henselae, and B. vinsonii ssp. berkhoffi antigens were performed (3,7). Results are presented in Table 2.

Clinicopathologic abnormalities detected in these dogs at initial examination, including acute onset of fever, lethargy, thrombocytopenia, anemia, mildly increased liver enzyme activities and hypoalbuminemia, were very similar to abnormalities associated with spotted fever group (SFG) rickettsioses in dogs and humans (1). In addition, R. conorii DNA was amplified in all dogs during the acute illness. Further evidence for R. conorii infection as a cause of the associated clinical signs was provided by the subsequent failure to detect DNA in dogs 1 and 2, 1 week after treatment with doxycycline and the rapid resolution of clinical signs 2 days after initiating doxycycline therapy. Clinical signs in dog 3 resolved in 4 days, while the dog was receiving ceftriaxone, which has no known anti-rickettsial efficacy (1). Spontaneous immune clearance of R. conorii likely accounted for the resolution of clinical signs in dog 3 (6).

The 4-fold increase in IgG antibody titers in dogs 2 and 3 supports seroconversion, which is consistent with an acute R. conorii infection (11). Additionally, the initially high IgM titer in dog 1 after the onset of illness compared with a much lower IgM titer after 65 days is also supportive of an acute infection and is consistent with observations of human serologic test results (1). IgM titers rise rapidly and then disappear by day 35 and 80 in dogs experimentally infected with R. conorii and R. rickettsii, respectively (6,11). However, high R. rickettsii IgM titers are detected in dogs that do not seroconvert, based upon IgG antibodies (11). Thus, the presence of IgM supports but does not prove acute SFG infection in dogs.

Coinfection with A. phagocytophilum or B. burgdorferi could have contributed to clinical signs observed in dog 1. This dog had a low serum A. phagocytophilum titer 7 days after initial examination and also seroconverted to B. burgdorferi. A. phagocytophilum causes an acute febrile illness in dogs and humans, similar to the findings described here (12). B. burgdorferi does not cause clinical signs in dogs until 60–150 days after experimental infection (13); therefore, despite seroconversion, the acute clinical signs in dog 1 were not likely to have been related to B. burgdorferi infection. Moreover, PCR amplification of DNA from organisms other than R. conorii was not found in any dog.

All dogs were intact, male, genetically unrelated Yorkshire terriers. Although an increased risk for Rocky Mountain spotted fever has not been reported in Yorkshire terriers, purebred dogs infected with R. rickettsii appear to be more prone to clinical illness (14). Notably, this breed seems to be at increased risk for Babesia canis infection (15). Male dogs and men may be at increased risk for infection and may develop more severe illness with R. rickettsii and R. conorii (1,14), and male dogs are more likely to be R. conorii seroreactive (3). It has been suggested that more severe illness may develop in English springer spaniels with suspected phosphofructokinase deficiency and persons with glucose 6-phosphate dehydrogenase deficiency when infected with R. rickettsii and R. conorii (1,14). Although inherited immunodeficiencies have not been reported in Yorkshire terriers, and all dogs were previously healthy, an inherited metabolic or immunologic defect cannot be ruled out because specific testing was not performed.

Although a metabolic or immunologic defect may be necessary for illness to develop in dogs of various breeds after R. conorii infection, other potential explanations can be made for the discrepancy between high R. conorii seroprevalence rates among healthy dogs and lack of reports of clinical illness. The high R. conorii seroprevalence in healthy dog populations suggests that exposure to SFG rickettsiae is common. However, the acute, nonspecific, and potentially self-limiting nature of R. conorii infection, combined with a low index of suspicion by regional veterinarians and a historical lack of specific diagnostic techniques, may have precluded the prior association of clinical signs with R. conorii infection in dogs. Further evidence should be gathered regarding the extent to which R. conorii causes clinical disease in dogs.