Polymerase chain reaction for diagnosis of dermatophyte and Scytalidium spp. onychomycosis

SIR, Diagnosis of onychomycosis relies on direct microscopic examination and culture of a nail sample. Optimally, the specimen must be obtained from the nail bed, as close as possible to the cuticle and to the advancing infected edge of the lesion, as this area is most likely to contain viable hyphae. On a sample obtained under the correct conditions, direct microscopic examination can provide rapid evidence of fungal infection. Culture remains the only method for fungal identification; however, culture results are delayed by 2–3 weeks because of the slow growth of dermatophytes. When the sample is not collected under the correct conditions, both direct examination and culture are totally unreliable, with a high rate of false-negative results and of false-positive results due to fungal or bacterial contamination. Therefore, the sampling technique appears to be a major limiting factor for mycological diagnosis. DNA amplification techniques such as polymerase chain reaction (PCR) might represent an alternative technology as DNA is extremely resistant and can persist even in the absence of viable hyphae. We previously showed the excellent sensitivity and specificity of a PCR-restriction fragment length polymorphism (RFLP) method for rapid diagnosis of dermatophytes and Scytalidium spp. from properly sampled clinical specimens. In the present study, we confirm our previous findings and show the performance of this method on nail specimens that had not been sampled as recommended for mycological examination. Samples were collected prospectively from patients suspected of having onychomycosis. In a first series, 48 nail samples were collected according to the optimal technique for mycological examination. These samples are named ‘correct’ samples. Part of each sample was processed by using classical mycological techniques, i.e. direct microscopic examination in black chlorazol solution followed by culture on both Sabouraud’s dextrose agar containing chloramphenicol and Sabouraud’s dextrose agar containing chloramphenicol and cycloheximide (Bio-Rad, Marnes-la-Coquette, France) for specific identification. The other part was tested by PCRRFLP blindly to the results of the mycological examination. In a second series, two samples were collected in 88 patients. A first sample was obtained by clipping the dystrophic nail plate at the hyponychium or at a distance from the active lesion (named ‘incorrect’ sample). The second sample was collected from the same nail under optimal conditions (‘correct’ sample). The ‘correct’ sample was processed by mycological techniques as described above; the ‘incorrect’ sample was used for PCR-RFLP. DNA extraction, PCR and RFLP analysis were performed as previously described. Among the 48 patients with suspected onychomycosis who gave ‘correct’ nail samples, 30 were diagnosed as having onychomycosis both by direct microscopic examination and by culture (28 dermatophytes and two Scytalidium). PCRRFLP was positive in 29 cases (27 dermatophytes and two Scytalidium), resulting in a sensitivity of 93% and a specificity of 94%, compared with mycological examination. Among the 88 patients with suspected onychomycosis who gave both ‘incorrect’ and ‘correct’ samples, 69 were diagnosed as having onychomycosis either by direct microscopic examination (n 1⁄4 69) or by culture on ‘correct’ nail samples (n 1⁄4 61; 56 dermatophytes and five Scytalidium). In comparison, PCRRFLP on ‘incorrect’ samples was positive in 67 cases (64 dermatophytes and three Scytalidium), resulting in a sensitivity of 93% and a specificity of 84% (Table 1). Because fungi are responsible for only about half of all nail dystrophies, the use of appropriate mycological techniques is