Background Onychomycosis refers to fungal infections of the nail apparatus that may cause pain, discomfort, and disfigurement. This is an update of a Cochrane Review published in 2007; a substantial amount of new research warrants a review exclusively on toenails. Objectives To assess the clinical and mycological effects of topical drugs and device-based therapies for toenail onychomycosis. Search methods We searched the following databases up to May 2019: the Cochrane Skin Group Specialised Register, CENTRAL, MEDLINE, Embase and LILACS. We also searched five trials registers, and checked the reference lists of included and excluded studies for further references to relevant randomised controlled trials. Selection criteria Randomised controlled trials of topical and device-based therapies for onychomycosis in participants with toenail onychomycosis, confirmed by positive cultures, direct microscopy, or histological nail examination. Eligible comparators were placebo, vehicle, no treatment, or an active topical or device-based treatment. Data collection and analysis We used standard methodological procedures expected by Cochrane. Primary outcomes were complete cure rate (normal-looking nail plus fungus elimination, determined with laboratory methods) and number of participants reporting treatment-related adverse events. Main results We included 56 studies (12,501 participants, average age: 27 to 68 years), with mainly mild-to-moderate onychomycosis without matrix involvement (where reported). Participants had more than one toenail affected. Most studies lasted 48 to 52 weeks; 23% reported disease duration (variable). Thirty-five studies specifically examined dermatophyte-caused onychomycosis. Forty-three studies were carried out in outpatient settings. Most studies assessed topical treatments, 9% devices, and 11% both. We rated three studies at low risk of bias across all domains. The most common high-risk domain was performance bias. We present results for key comparisons, where treatment duration was 36 or 48 weeks, and clinical outcomes were measured at 40 to 52 weeks. Based on two studies (460 participants), compared with vehicle, ciclopirox 8% lacquer may be more effective in achieving complete cure (risk ratio (RR) 9.29, 95% confidence interval (CI) 1.72 to 50.14; low-quality evidence) and is probably more effective in achieving mycological cure (RR 3.15, 95% CI 1.93 to 5.12; moderate-quality evidence). Ciclopirox lacquer may lead to increased adverse events, commonly application reactions, rashes, and nail alteration (e.g. colour, shape). However, the 95% CI indicates that ciclopirox lacquer may actually make little or no difference (RR 1.61, 95% CI 0.89 to 2.92; low-quality evidence). Efinaconazole 10% solution is more effective than vehicle in achieving complete cure (RR 3.54, 95% CI 2.24 to 5.60; 3 studies, 1716 participants) and clinical cure (RR 3.07, 95% CI 2.08 to 4.53; 2 studies, 1655 participants) (both high-quality evidence) and is probably more effective in achieving mycological cure (RR 2.31, 95% CI 1.08 to 4.94; 3 studies, 1716 participants; moderate-quality evidence). Risk of adverse events (such as dermatitis and vesicles) was slightly higher with efinaconazole (RR 1.10, 95% CI 1.01 to 1.20; 3 studies, 1701 participants; high-quality evidence). No other key comparison measured clinical cure. Based on two studies, compared with vehicle, tavaborole 5% solution is probably more effective in achieving complete cure (RR 7.40, 95% CI 2.71 to 20.24; 1198 participants), but probably has a higher risk of adverse events (application site reactions were most commonly reported) (RR 3.82, 95% CI 1.65 to 8.85; 1186 participants (both moderate-quality evidence)). Tavaborole improves mycological cure (RR 3.40, 95% CI 2.34 to 4.93; 1198 participants; high-quality evidence). Moderate-quality evidence from two studies (490 participants) indicates that P-3051 (ciclopirox 8% hydrolacquer) is probably more effective than the comparators ciclopirox 8% lacquer or amorolfine 5% in achieving complete cure (RR 2.43, 95% CI 1.32 to4.48), but there is probably little or no difference between the treatments in achieving mycological cure (RR 1.08, 95% CI 0.85 to 1.37). We found no difference in the risk of adverse events (RR 0.60, 95% CI 0.19 to 1.92; 2 studies, 487 participants; low-quality evidence). The most common events were erythema, rash, and burning. Three studies (112 participants) compared 1064-nm Nd:YAG laser to no treatment or sham treatment. We are uncertain if there is a difference in adverse events (very low-quality evidence) (two studies; 85 participants). There may be little or no difference in mycological cure at 52 weeks (RR 1.04, 95% CI 0.59 to 1.85; 2 studies, 85 participants; low-quality evidence). Complete cure was not measured. One study (293 participants) compared luliconazole 5% solution to vehicle. We are uncertain whether luliconazole leads to higher rates of complete cure (very low-quality evidence). Low-quality evidence indicates there may be little or no difference in adverse events (RR 1.02, 95% CI 0.90 to 1.16) and there may be increased mycological cure with luliconazole; however, the 95% CI indicates that luliconazole may make little or no difference to mycological cure (RR 1.39, 95% CI 0.98 to 1.97). Commonly-reported adverse events were dry skin, paronychia, eczema, and hyperkeratosis, which improved or resolved post-treatment. Authors' conclusions Assessing complete cure, high-quality evidence supports the effectiveness of efinaconazole, moderate-quality evidence supports P-3051 (ciclopirox 8% hydrolacquer) and tavaborole, and low-quality evidence supports ciclopirox 8% lacquer. We are uncertain whether luliconazole 5% solution leads to complete cure (very low-quality evidence); this outcome was not measured by the 1064-nm Nd:YAG laser comparison. Although evidence supports topical treatments, complete cure rates with topical treatments are relatively low. We are uncertain if 1064-nm Nd:YAG laser increases adverse events compared with no treatment or sham treatment (very low-quality evidence). Low-quality evidence indicates that there is no difference in adverse events between P-3051 (ciclopirox hydrolacquer), luliconazole 5% solution, and their comparators. Ciclopirox 8% lacquer may increase adverse events (low-quality evidence). High- to moderate-quality evidence suggests increased adverse events with efinaconazole 10% solution or tavaborole 5% solution. We downgraded evidence for heterogeneity, tack of blinding, and small sample sizes. There is uncertainty about the effectiveness of device-based treatments, which were under-represented; 80% of studies assessed topical treatments, but we were unable to evaluate all of the currently relevant topical treatments. Future studies of topical and device-based therapies should be blinded, with patient-centred outcomes and an adequate sample size. They should specify the causative organism and directly compare treatments.
