Rocky Mountain Spotted Fever in Dogs, Brazil

Clinical illness caused by Rickettsia rickettsii in dogs has been reported solely in the United States. We report 2 natural clinical cases of Rocky Mountain spotted fever in dogs in Brazil. Each case was confirmed by seroconversion and molecular analysis and resolved after doxycycline therapy.


The Cases
On August 23, 2007, a 4-year-old, female, Dogue de Bordeaux (dog 1) was brought to a veterinary clinic in São Paulo because of a high load of ticks noticed 5 days after she had been to a farm in the Itu Municipality (23°15′S, 47°17′W), state of São Paulo. The dog was treated with fi pronil and sent home. Tick taxonomic identifi cation was not performed. The next day, the dog had diarrhea and hematochezia and was taken back to the clinic, where laboratory test results were within reference ranges, except for a slight leukocytosis (18,000 cells/mm 3 ) and elevated al-kaline phosphatase level (278.6 U/L). Metronidazol was prescribed, and the dog was again sent home. Three days later, the dog was febrile (40.5°C), anorexic, and lethargic. Blood was sent to a private laboratory, where a battery of PCR tests failed to detect DNA of Babesia spp., Borrelia spp., Mycoplasma spp., or Ehrlichia spp. The dog was treated with subcutaneous imidocarb and oral doxycycline. The next day, the dog was still febrile (39.4°C) and anorexic, and neurologic signs (ataxia and vestibular syndrome with spontaneous nystagmus) had developed. The animal was hospitalized; doxycycline was switched to the subcutaneous route; and the next day oral prednisone was added. Blood values remained within reference range, except for a slight leukocytosis (17,600 cells/mm 3 ). On August 30, neurologic improvement was noted, and the dog had no fever (38.5°C) and started to eat. Despite slight nystagmus, the dog was discharged the next day. On September 3, (8 days after doxycycline therapy began), the dog showed no clinical abnormality, and a new blood sample was collected for serologic testing. Another blood sample collected on September 10 showed hematologic parameters within reference range except for leukopenia (6,900 cells/mm 3 ).
Serologic evaluation was performed by indirect immunofl uorescence assay (IFA) by using antigens of 6 Rickettsia isolates from Brazil (8). Plasma from the sample collected on August 24 showed an IFA endpoint titer of 128 for R. rickettsii and no reactivity for the remaining rickettsial antigens at a 1:64 dilution. Plasma from the sample collected on September 3 showed the following endpoint titers for rickettsial antigens: R. rickettsii 2,048, R. parkeri 512, R. amblyommii 512, R. felis 512, R. rhipicephali, and 512; R. bellii 256.
DNA was extracted from the blood samples collected on August 24 and September 3 (before and after antimicrobial drug therapy) by using the DNeasy Tissue Kit (QIAGEN, Chatsworth, CA, USA). Samples were tested by 2 PCR protocols: one targeting a 147-bp fragment of the rickettsial gltA gene (9), and the other, a heminested PCR, targeting a fragment of the rickettsial ompA gene (10). Extracted DNA from the fi rst blood sample yielded expected products by both PCR protocols. No product was obtained from the second blood sample. Sequencing of the ompA product resulted in a 452-bp fragment 100% identical to the corresponding sequence of the Bitterroot strain of R. rickettsii from the United States (GenBank accession no. U43804). Ehrlichia spp. were not detected by PCR (6) in either sample.
The second case was noted on August 28, 2007, when a 10-month-old, female, miniature Schnauzer (dog 2) was examined at the same veterinary clinic for anorexia, lethargy, fever (40.2°C), vomiting, and tick infestation. This dog had visited the same farm at the same time as dog 1 DNA was extracted from the samples collected on August 24 and September 3 (before and after antimicrobial drug therapy) and processed by the PCR protocols cited above. Extracted DNA from the fi rst sample yielded expected product for the gltA-PCR, which was not sequenced. No other PCR product was obtained.

Conclusions
Defi nitive diagnoses of naturally acquired R. rickettsii infection in 2 dogs in Brazil are supported by 1) paired serum samples with >8-fold rise in antibody titer to R. rickettsii antigen; 2) titers to R. rickettsii >4-fold higher than titers to other rickettsial antigens known to occur in Brazil; 3) detection of rickettsial DNA in canine blood, confi rmed to be R. rickettsii in at least 1 of the dogs; 4) compatible clinical signs and laboratory abnormalities (i.e., thrombocytopenia in at least 1 dog); 5) response to doxycycline; and 6) compatible epidemiologic history (i.e., prior contact with ticks in an RMSF-endemic area). This sixth statement is supported by the fact that Itu municipality is an area where RMSF laboratory-confi rmed cases in humans have been reported since 2003 (www.cve.saude.sp.gov.br). Owners of the 2 dogs reported here noted various capybaras (Hydrochoerus hydrochaeris) in the area where their dogs had become infested with ticks (data not shown). Capybaras are one of the main hosts of Amblyomma cajennense ticks, the most common important vector of R. rickettsii in Brazil (9,11).
In a recent study of experimental infection, dogs exposed to a Brazil isolate of R. rickettsii had fever, lethargy, anorexia, anemia, and thrombocytopenia; 1 also had ocular lesions (12). These clinical signs have been reported in the United States for dogs with active R. rickettsii infection (4,5) and were also noted in the present study under natural conditions, except for anemia and ocular lesions. Studies in the United States have shown that neurologic dysfunction occurs in as many as 43% of dogs with RMSF; vestibular dysfunction is possibly the most frequent neurologic abnormality (13). These results suggest that clinical illness caused by R. rickettsii in dogs has similar patterns in Brazil and the United States.
Veterinarians in Brazil should include R. rickettsii infection in their differential diagnoses of CME and other acute nonspecifi c febrile illnesses of dogs, especially because R. rickettsii is highly pathogenic for humans. In the United States, several cases of human infection have been preceded by RMSF in dogs (14,15). Accurate diagnosis of RMSF in dogs should lead to dog owners understanding risk for infection from ticks in their location (14) and provide valuable information for the surveillance of RMSF in humans.