The animal study was conducted beginning in February 2009, and the human study was conducted beginning in June 2009. R. conorii, R. akari, R. rhipicephali, R. africae, R. parkeri, and O. tsutsugamushi were grown in L929 cell monolayers, purified, and titrated as reported (12). A suspension of each rickettsial species (200 μL containing 1 × 10⁵ rickettsia) was injected intradermally into 8 shaved areas on the backs of 6 Hartley guinea pigs (1 species/guinea pig) by using aseptic procedures (12). A negative control guinea pig was infected with 200 μL (1 × 10⁶ cells/mL) of an L929 cell suspension. Infection sites were inspected daily for skin lesions. Animals were handled according to the regulations of Décret No. 887–848 du 10/19/1987, Paris. The experimental protocol was reviewed and approved by the Institutional Animal Care Committee, Université de la Méditerranée, Marseille.

Infection with each rickettsial species caused an eschar at the infection site (12). Eschars were observed at day 3 postinfection. A sterile cotton swab (Copan Italia S. p. A., Brescia, Italy) was rotated against the eschar (3 circular motions) and stored at 4°C for 24 h. Swabs were then placed in 400 μL of phosphate-buffered saline, and DNA was extracted by using the QIAamp DNA Mini Kit (QIAGEN, Hilden, Germany). Lesions were swabbed daily until the animal showed clinical recovery (day 20 postinfection for those infected with R. akari, R. conorii, and R. rhipicephali and day 13 postinfection for those infected with R. africae, R. parkeri, and O. tsutsugamushi).

Maximum number of DNA copies for R. rhipicephali, R. akari, and R. conorii was detected on day 4 postinfection (2.27 × 10⁷, 2.96 × 10⁷, and 9.28 × 10⁷copies/5 μL of swab DNA extracts, respectively (Figure 1, panel A). Maximum number of DNA copies for R. parkeri was detected on day 3 postinfection (2.66 × 10⁵ copies/5 μL), for R. africae on day 6 postinfection (6.73 × 10⁵ copies/5 μL), and for O. tsutsugamushi on day 10 postinfection (2.7 × 10⁷ copies/5 μL) (Figure 1, panel B).

Effects of temperature and storage time of cotton swabs on bacterial DNA were evaluated in 3 guinea pigs infected with R. conorii. Twelve swabs per animal were obtained daily for 5 days and stored in groups of 3 at 22°C, 4°C, −20°C, or −80°C. DNA was extracted after 1, 2, or 3 days of storage. Eschars appeared by day 3 postinfection and reached their maximum size by day 7. Storage at 4°C was the optimal temperature condition for isolation of DNA (7.53 × 10⁶ copies/5 μL vs. 1.03 × 10⁶, 3.77 × 10⁶, or 4.49 × 10⁶ copies/5 μL for swab storage at 22°C, −20°C, or −80°C respectively; p = 0.0001) (Figure 2, panel A). Storage time (24 h, 48 h, and 72 h) had no effect on DNA yield (Figure 2, panel B). Temperature had a significant effect (p<0.05) on DNA yield and for the same extraction (Figure 2, panels C–E).

To demonstrate the usefulness of skin lesion swabs for detection of rickettsial infection, we used this technique with eschars from patients with suspected rickettsioses. Nine patients were included in this experiment after informed consent was obtained. This experiment was reviewed and approved by the local ethics committee (reference 09–016). DNA was extracted from swabs or skin biopsy specimens and tested by quantitative PCRs (qPCRs) (13) specific for a fragment of the citrate synthase A gene, which is conserved among spotted fever group rickettsiae, or the gene coding periplasmic serine protease of O. tsutsugamushi; β-actin gene was used as a control (14).

When rickettsial DNA was amplified in samples, specific qPCR was performed by using specific primers and probes and on the basis of clinical and epidemiologic data (Technical Appendix Table 1 ) (4). If specific rickettsial DNA was not detected, PCR amplification and sequencing were performed to identify the causative agent (4,15). R. montanensis DNA was used as a positive control, and DNA from sterile biopsy samples and sterile water were used as a negative control.

The qPCR for the β-actin gene showed cycle threshold (C_t) values of 19–23 for skin biopsy samples and 22–37 for swab samples (Table). Spotted fever group rickettsial DNA was detected in biopsy samples from 5/5 patients and swab samples from 8/9 patients (Technical Appendix Table 2). Specific qPCR showed a diagnostic result in 3/7 swabs samples and 4/5 skin biopsy samples.

We amplified R. conorii DNA from patients 1 and 2, R. africae DNA from patients 4 and 5, and R. australis DNA from patient 9. Rickettsial DNA from patients 3 and 7 showed 100% homology with the R. sibirica mongolitimonae citrate synthase A gene (GenBank accession nos. DQ097081 and DQ423370, respectively). Rickettsial DNA from patient 6 showed 99.1% homology with DNA from R. slovaca. Patient 9 was a technician who was accidentally infected by the aerosol route when handling R. australis. Only 2/11 swabs obtained from vesicular lesions of patient 9 were positive for rickettsial DNA and R. australis DNA after reamplification of primary PCR products. These samples showed 98% homology with R. australis 23S rRNA gene (GenBank accession no. AJ133711) (Technical Appendix Table 2).

Our study showed the efficacy and reliability of skin lesion swabs for molecular detection of 6 Rickettsia species (Figure 1). Rickettsial DNAs were detected by using this technique as long as eschars persisted (<19 days). For short-term storage of swabs, 4°C was the optimal temperature. Using swabs of eschars, we made a diagnosis of rickettsiosis for 8/9 patients. For patients 6, 7, and 8, for whom biopsy samples were not available, we confirmed the diagnosis by using swab samples. We also showed that for patient 9, who had a rickettsiosis but no eschar, swabbing of vesicular lesions may be useful for diagnosis, although these lesions were less sensitive than eschars.

Our results indicate that swabs of eschars can be used for molecular detection of rickettsial infections when biopsy samples are not available or biopsies are difficult to perform. We recommend that swabs be used for DNA extraction immediately after sampling or stored at 4°C until needed.