Comparison of diagnostic methods in the evaluation of onychomycosis

Abstract

BACKGROUND Onychomycosis is a common problem seen in clinical practice. Given the differential diagnosis of dystrophic nails, it is helpful to obtain a definitive diagnosis of dermatophyte infection before the initiation of antifungal therapy. Potassium hydroxide (KOH) preparation and fungal culture, which are typically used in the diagnosis of these infections, often yield false-negative results. Recent reports have suggested that nail plate biopsy using periodic acid-Schiff (PAS) (Bx/PAS) stain may be a very sensitive technique for the diagnosis of onychomycosis. OBJECTIVE The purpose of this study was to compare KOH preparation, culture, Bx/PAS stain, and calcofluor white (CW) stain in the diagnosis of onychomycosis and to determine their sensitivity and specificity. METHODS We evaluated 105 patients with suspected onychomycosis using 4 diagnostic methods: KOH preparation, culture, Bx/PAS, and CW stain. CW stain binds to cellulose and chitin, and fluoresces when exposed to UV radiation. It is a highly sensitive and specific technique for the detection of dermatophytes. To determine the clinical usefulness and performance characteristics of each test, CW was chosen as the gold standard for statistical analysis. RESULTS Of the patients, 93 had at least 1 of the 4 diagnostic methods positive for the presence of organisms. The following were calculated for each test: sensitivity; specificity; positive predictive value; and negative predictive value. The sensitivities of each of the techniques were as follows: KOH 80%; Bx/PAS 92%; and culture 59%. Both KOH and Bx/PAS methods were more sensitive than culture (P =.00002). Bx/PAS was also more sensitive than KOH (P =.03). The specificities were as follows: KOH 72%; Bx/PAS 72%; and culture 82%. The positive predictive value calculated for the different techniques were: KOH 88%; Bx/PAS 89.7%; and culture 90%. In terms of negative predictive value, the results were: KOH 58%; Bx/PAS 77%; and culture 43%. CONCLUSION Bx/PAS is the most sensitive method for the diagnosis of onychomycosis. It is also superior to the other methods in its negative predictive value. It is indicated if other methods are negative and clinical suspicion is high, and potentially is the single method of choice for the evaluation of onychomycosis.