Background: Scabies is a contagious skin infestation caused by the human mite Sarcoptes scabiei. The usual reference method for definitive diagnosis is ex vivo identification of the mite with microscopic examination of skin scrapings. We compared diagnostic accuracy of in vivo dermoscopic (DS) mite identification using a pocket handheld low-magnification DS with the reference method.
Methods: We conducted a prospective, nonrandomized, evaluator-blinded, noninferiority study to compare sensitivities (main outcome) and other diagnostic properties of DS and microscopic examination of skin scrapings. Among 756 patients with a presumptive diagnosis of scabies consulting in one center, 238 were sequentially submitted to the two diagnostic procedures. Three dermoscopists (one expert, two inexperienced) were involved. Diagnostic strategies using clinical skills only, DS results, and a combination of both were compared.
Results: Sensitivities were 91% (95% confidence interval: 86-96) for DS and 90% (95% confidence interval: 85-96) for microscopic examination of skin scrapings (P = .005 for noninferiority). Specificities were 86% (95% confidence interval: 80-92) for DS and 100% (by definition) for microscopic examination of skin scrapings. DS sensitivities were similar for the expert and inexperienced dermoscopists, whereas differences were observed in specificities. However, diagnostic accuracy of inexperienced dermoscopists steadily increased during the study. Compared with clinical-based, DS-based treatment decision rule minimized the number of false-positive and false-negative findings, whereas a treatment decision rule based on combination of clinical presumption and DS result drastically reduced the number of patients with scabies left untreated.
Limitations: There is no definitive standard for ruling out the diagnosis of scabies.
Conclusions: Standard DS with a handheld DS is a useful tool for diagnosing scabies, with high sensitivity, even in inexperienced hands. It greatly enhances clinical skills for making treatment decisions.