To the Editor: Rickettsiosis due to Rickettsia felis is an emerging disease that has been reported worldwide (1). Fever, headache, myalgia, and macular rash have been attributed to R. felis infection in humans (1). In South America, R. felis infection in fleas (mostly Ctenocephalides spp.) has been reported only in Brazil, Peru, and Uruguay (2–3). Although a growing number of articles have reported that R. felis is transmitted by fleas, the acquisition mechanism of R. felis by vertebrates or uninfected fleas in nature remains unknown (4).

Cats experimentally exposed to R. felis–infected fleas have been shown to become seropositive (5). However, neither serologic nor molecular evidence of R. felis infection has been reported in cats under natural conditions, despite the fact that most C. felis fleas are infected by R. felis (6,7).

In November 2006, we investigated the presence of rickettsial DNA in 30 C. felis fleas randomly collected from 22 domestic cats privately owned and housed indoors in a single household in Santiago, Chile. To detect rickettsial DNA in each individual flea, PCRs were performed that targeted a 398-nt fragment of the rickettsial gltA gene and an 856-nt fragment of the rickettsial ompB gene (7,8).

A total of 21 individual fleas (70%) yielded expected PCR products for both gltA and ompB genes. PCR gltA products from the 21 fleas and ompB products from 5 fleas were subjected to DNA sequencing as described (7). The gltA partial sequences obtained from 21 fleas were identical, as were the ompB partial sequences from 5 fleas. These sequences were 100% identical to corresponding sequences in the R. felis genome (GenBank accession no. CP000053).

Blood serum samples were collected from the 22 cats and tested by indirect immunofluorescence assay (IFA) with crude antigens derived from 6 Rickettsia isolates from Brazil: R. bellii, R. amblyommii, R. rhipicephali, R. rickettsii, R. parkeri, and R. felis (7,9). Serum was considered to contain antibodies against rickettsiae if it displayed a reaction at 1:64 dilution. End-point titers against each Rickettsia species were determined by testing serial 2-fold serum dilutions. Reactive serum specimens were tested in 2 or 3 replications by 2 readers before the end-point titer was determined. Serum showing a Rickettsia species titer at least 4-fold higher than those observed for the other Ricketttsia species was considered homologous to the first Rickettsia species or to a very closely related genotype (7,9). In each slide, a nonreactive cat serum specimen (negative control) and a known reactive cat serum specimen (positive control) were tested at the 1:64 dilution (7).

IFA detected antibodies reactive with R. felis (titer >64) in 16 (72.7%) of 22 cats. Among those, 5 (22.7%) also reacted with R. rhipicephali, 4 (18.2%) with R. bellii, 3 (13.6%) with R. parkeri, 2 (9.1%) with R. rickettsii, and 1 (4.5%) with R. amblyommii. No serum reacted with any other Rickettsia species without reacting with R. felis (Table). Four cat serum specimens (cats 1, 3, 8, and 11) showed titers to R. felis at least 4-fold higher than those to any of the other 5 antigens. The antibody titers in these 4 animals were considered to have been stimulated by R. felis infection. For the remaining 12 seropositive cats, we could not discern whether R. felis had been the infection agent because the results displayed a single titer of 64 for R. felis or showed similar titers for other Rickettsia species.

We report 70% R. felis–infected fleas in this study on the basis of the concordant results of 2 PCR amplifications (gltA and ompB) and DNA sequencing. This infection rate is within the range (13.5%–90%) that has been reported for R. felis infecting Ctenocephalides fleas in Brazil and Uruguay (2,3,7). Sixteen (72.7%) cats contained R. felis–reactive antibodies; 4 of them showed titers to R. felis at least 4-fold higher than those to the other 5 rickettsial strains, findings that enabled us to technically conclude that these cats were exposed to R. felis or a closely related organism (1,7,9). Our finding of 70% R. felis infection in fleas infesting the cats indicates that cats acquired the infection through infected fleas. However, the mechanism of R. felis transmission by fleas is yet to be demonstrated under experimental conditions.

To our knowledge, the presence of R. felis, or a spotted fever group Rickettsia species, has not been reported in Chile. Recent investigations have provided clinical and serologic evidence of canine (10) and human (K. Abarca and J. Lopez, unpub. data) infection by spotted fever rickettsia in Chile, confirmed by IFA that used R. conorii commercial antigen. Since substantial serologic cross-reaction occurs between R. conorii and R. felis antigens (1), R. felis could be causing infection in dogs or humans in Chile.