Introduction

Onychomycosis is the most common cause of abnormal toenails, accounting for over half of all nail-related diseases.1 2 Toenail onychomycosis is primarily caused by dermatophytes, most frequently Trichophyton rubrum and to a lesser extent by yeasts or non-dermatophyte moulds.3–5 With an estimated prevalence of 4.3% in the general population, increasing to over 20% in patients aged ≥60, toenail onychomycosis is common in Europe and North America and often seen in general practice.1 6–8 Clinical signs and symptoms may vary significantly, with some patients unaware of their onychomycosis, while others suffer from progressive thickening, discolouration, separation from the nail bed or even nail plate destruction.3 7 9 10 Onychomycosis can also cause pain and decrease quality of life, especially in the severely affected,11 12 as well as complications such as recurring fungal and bacterial skin infections.13 14

In patients experiencing few or no symptoms, treatment might not be strictly warranted.7 12 15 For patients requiring treatment, oral terbinafine is the recommended and most effective treatment option.12 15 16 However, oral treatment can have serious adverse effects such as severe cutaneous reactions or cause hepatotoxicity ranging from mild transaminitis to fulminant liver failure.17–20 Hence, oral treatment is disadvised for patients with liver disease, hypersensitivity or at risk of drug–drug interactions. Topical treatment would be a welcome alternative, not only for those with contraindications for systemic treatment but also in general, given its low-risk profile.21 Previous studies investigating the efficacy of topical antifungals were carried out in outpatient settings in specialised clinics not necessarily representative of primary care.21 Furthermore, studies have shown that severe cases, for example, involving the matrix, do not respond well to topical treatment.16 21 22 Based on previous results and general availability, miconazole and amorolfine are potentially suitable for treating mild to moderately severe onychomycosis in primary care.22–27 Although effective at treating dermatomycosis, miconazole is not well studied in the treatment of onychomycosis, and the evidence available on the efficacy of amorolfine is of low quality.21–27

Our aim was to investigate the efficacy of these commonly used topical antifungals in the primary care treatment of mild to moderately severe onychomycosis. In this randomised, double-blind placebo-controlled trial, we compared topical miconazole or topical amorolfine with placebo.

MethodsTrial design

The Onycho Trial was a double-blind, placebo-controlled trial designed to investigate the efficacy of topical antifungal treatment in Dutch general practice (ICTRP NL8193 https://trialsearch.who.int/). The study consisted of three parallel arms: two topical antifungals, miconazole and amorolfine, and a placebo group, with randomisation at the individual patient level and an allocation ratio of 1:1:1.

This trial was reported in accordance with the Consolidated Standards of Reporting Trials statement and guidelines.28 29

Participants and setting

The Onycho Trial focused on primary care patients with mild to moderately severe onychomycosis, defined as 10%–75% involvement of the index toenail without matrix involvement or spikes. The index toenail was defined as the largest, most affected toenail. Initially, the minimum involvement was set at 25% for ease of interpretation. One month into the trial and after including six participants, this was lowered to 10% to increase the inclusion of milder cases and to accommodate the Onychomycosis Severity Index (OSI) as a measure of severity as approved by the ethics committee.30 Additional criteria included ages 18–70, no more than three affected toenails per foot, no oral antifungal treatment within 6 months prior to enrolment and no clinically apparent tinea pedis. Patients using contraindicated medication (metformin, phenytoin and coumarins) were excluded, as well as patients with contraindicated conditions (pregnant or lactating women), or conditions potentially affecting nail growth (stages III–IV peripheral arterial occlusive disease or cancers treated with immunotherapy or chemotherapy). There were no patients receiving topical antifungal treatment at the time of enrolment.

Patients suspected of having onychomycosis were referred by participating general practitioners (GPs) or recruited from the general public through social media, with prescreening via email or telephone. Eligible patients were invited for face-to-face screening at the department or the participant’s home. Screening and follow-up visits at 3 and 6 months after the start of treatment were performed by the department’s research staff under supervision of observers RMW (GP in training), TNB and JAHE (practising GPs). At each visit, protocolised clinical evaluations were performed, including the administration of validated questionnaires. Using a small portable photobooth, standardised photographs were taken and independently evaluated by the observers for patient selection and response to treatment. Finally, fungal infection of the nail sample collected during the screening was confirmed by the LUMC’s departments of Dermatology and Medical Microbiology, using a standardised workup of potassium hydroxide preparation, PCR and culture. This was repeated at the final visit after 6 months of treatment.

All patients signed informed consent at the screening visit. All clinical data were recorded using Castor’s Electronic Data Capture System (Castor).

Patient and public involvement

There was no formal patient or public involvement in the design of this study.

Interventions

Topical treatment of affected toenails consisted of once-daily application of miconazole 20 mg/g (Daktarin), or once-weekly amorolfine hydrochloride 5% (Loceryl) nail lacquer according to regular user instructions for a period of 6 months. Placebo consisted of daily application of denatonium benzoate solution (Byte-X, used to prevent nail biting) chosen for its resemblance in look, smell and consistency to the other antifungal lacquers. A thorough literature search was performed which yielded no studies suggesting antifungal properties of denatonium benzoate.31–35

All medications were prepared, packaged and distributed by the LUMC’s trial pharmacy. Trial medication consisted of two vials, one for Monday to Saturday and another for Sunday, identical for all participants regardless of treatment allocation. For participants in the miconazole group, both vials contained miconazole. For the amorolfine group, the Monday to Saturday vial contained denatonium benzoate solution while the Sunday vial contained amorolfine. For the placebo group, both vials contained denatonium benzoate solution. All participants received identical written instructions that included once daily application of one layer of nail lacquer solution after cleaning the nail with acetone provided. For Sundays, participants were additional instructed to file the nail surface before applying treatment, all according to user instructions.

Outcomes

Our primary outcome measure was the proportion of patients achieving complete cure, consisting of both clinical and mycological cure of the index toenail at 6 months.

Secondary outcome measures were clinical improvement (defined as either ≤10% involvement of the index toenail or as ≥40% reduction), symptom burden as expressed by the ONYCHO (Onychomycosis Quality of Life) questionnaire score,15 quality of life based on the Short Form-12 survey,36 37 adverse effects, therapy compliance, patient-perceived improvement and treatment acceptability. Adverse effects included local reactions, as well as serious adverse events (SAEs) and suspected unexpected serious adverse reactions (SUSARs). Therapy compliance was defined as having missed no more than 2 weeks of treatment, that is, <10% of the treatment period. In addition, the effects of treatment on OSI score (range 0–35) and percentage of the affected area of the index toenail were measured over time.

Sample size

Based on previous studies, the expected difference in complete cure between antifungal treatment and placebo was set at 40%.21 22 With a power of 90% and an alpha of 2.5%, applying the Bonferroni correction given the trial’s three study arms, the required sample size was 29 patients per group to demonstrate a significant difference (p<0.05). With an anticipated loss to follow-up of 20%, we calculated that 36 patients would be required for each group, resulting in a minimum inclusion of 108 patients. No interim analyses were performed.

Randomisation

Treatment allocation of included participants was computer generated in Castor, using non-stratified variable block randomisation with block sizes of 6, 9 and 12, and an allocation ratio of 1:1:1 to ensure equally sized groups. The allocation sequence was concealed for observers and research staff involved in screening, enrolment and follow-up. This was not the case for the data manager and pharmacy staff who required access to this information and were not involved in screening, follow-up, analyses of results or study tasks involving direct contact with participants. After confirmation of successful inclusion, the allocation sequence was initiated by the data manager in Castor and automatically sent to the responsible pharmacy staff.

Blinding

All participants, data collection staff and observers responsible for patient selection and outcome assessment (RMW, TNB and JAHE) were blinded for treatment allocation and outcome measurements recorded by the other observers. Regardless of treatment allocation, all dispensed trial medications were distributed in identical vials, including identical user instructions to ensure concealment of allocation.

Statistical methods

Descriptive statistics were used to evaluate baseline characteristics, presenting results per treatment group in absolute numbers with corresponding percentages or as means and SD. To test for statistical differences between the groups, Pearson’s χ2 test or one-way analysis of variance (ANOVA) were used for categorical or numerical data, respectively, with significance defined as a p<0.05.

To evaluate treatment effect after 6 months, primary comparisons were miconazole with placebo and amorolfine with placebo. In addition, both antifungals combined were compared with placebo and miconazole was compared with amorolfine. The primary outcome of complete cure, together with underlying clinical and mycological cure as binary outcomes, were evaluated using Pearson’s χ2 or Fisher’s exact test, the latter used if ≥1 cells had an expected count <5.

As binary measures, secondary outcomes for clinical improvement (≤10% or ≥40% reduction of affected area), adverse effects, therapy compliance, patient-perceived improvement and treatment acceptability, were also evaluated using χ2 or Fisher’s exact tests. To evaluate the effect of treatment on symptom burden and quality of life, as numerical values based on corresponding questionnaire scores, one-way ANOVA applying Tukey-Kramer’s post hoc test was used to compare mean differences between the groups. The corresponding effect sizes are reported as ORs with 95% CIs and p values for the binary outcomes, or as mean differences with p values for the numerical outcomes, respectively.

In addition, repeated measures ANOVA was used to evaluate the effect of treatment on clinical severity as expressed by the percentages of index toenail affected and OSI scores over time.

For the purpose of intention-to-treat analysis, the last-observation-carried-forward method was applied for both primary and secondary outcome measures, replacing missing numerical values for those of the last observation; for binary outcomes, missing values were replaced by their negative option, that is, not having reached the outcome of interest.

Finally, a per-protocol analysis was done to evaluate treatment effect in compliant patients defined as those that had missed no more than 2 weeks of treatment, as well as a subgroup analysis to evaluate the influence of the affected area of the index toenail at baseline on primary outcome and clinical improvement.

P values of <0.05 were considered statistically significant. Statistical analyses were performed using IBM SPSS Statistics (V.28).

ResultsParticipant recruitment

Participants were recruited from February 2020 to August 2022, when the required number of participants was reached. The date of first enrolment was 10 February 2020. Between March and November 2020, recruitment was suspended due to the COVID-19 pandemic. The first subsequent screening visit was on 9 December 2020. The last follow-up visit took place on 17 March 2023.

Participant flow

Figure 1 illustrates participant flow. Of 953 respondents, 760 were non-eligible, consisting of 735 who did not meet or were unable to verify criteria, while 25 declined, were unwilling or unable to participate (eg, for practical reasons). The remaining 193 respondents were eligible and invited for screening and intake.

Figure 1

Participant flow. ITT, intention to treat; LTFU, lost to follow-up; N, number of patients; PP, per protocol.

82 patients were subsequently excluded through screening: 52 of whom did not meet criteria due to extensive involvement (n=36), less than 10% involvement (n=15) or use of contraindicated medication (n=1); for the remaining 30 patients, fungal infection could not be confirmed.

Of the 111 patients included and randomised to one of the treatment arms, 99 (89.2%) completed the full treatment period, of whom 86 (77.5%) were also compliant. 12 patients (10.8%) were lost to follow-up: 6 in the miconazole group (due to no effect (n=3), discolouration (n=1), no given reason (n=1) and one never started); 4 in the amorolfine group (due to no effect (n=2), moving (n=1) and pregnancy (n=1)); as well as 2 patients in the placebo group (due to no effect (n=1) and unspecified private circumstances (n=1)).

Baseline characteristics

Table 1 shows the baseline characteristics per treatment group. The overall study population had a mean age at baseline of 51 years (SD 13.1), a slight female predominance (51.4%) and showed moderately severe onychomycosis (mean affliction 41% (SD 15.5), mean OSI 13.0 (SD 5.5)) primarily caused by dermatophyte infection (93.7%). Specific organisms identified are presented in online supplemental table S1. Most participants reported a moderate symptom burden, with an average ONYCHO questionnaire score of 58.9 (SD 13.8). Regarding risk factors and relevant history, tinea pedis was most commonly reported (56.8%), and a majority had previously treated their nail infection (82.9%), including the use of topical antifungals. A minority had previously received oral treatment (10.8%). No statistically significant differences were found between groups at baseline.

Table 1

Participant characteristics at baseline

Numbers analysed

All randomised patients (N=111) were included in the intention-to-treat analyses of primary and secondary outcome measures. For the per-protocol analysis, only compliant cases were selected (N=86).

Primary and secondary outcomes

Table 2 summarises results for each treatment group at 6 months. No patient in the miconazole and amorolfine group reached a complete cure, compared with two in the placebo group. In the amorolfine group, one patient reached clinical cure, compared with two in the placebo group. In each antifungal group, 3 patients reached mycological cure, compared with 10 in the placebo group.

Table 2

Summary results after 6 months treatment, per treatment group

Tables 3 and 4 show the comparison of primary and secondary outcomes for miconazole versus placebo and for amorolfine versus placebo, respectively. No OR could be calculated for comparisons with zero counts, indicated as ‘not estimable’.

Table 3

Effect size differences at 6 months between miconazole and placebo

Table 4

Effect size differences at 6 months between amorolfine and placebo

As regards complete or clinical cure, there were no statistically significant differences between antifungals and placebo while miconazole and amorolfine were significantly less effective at reaching mycological cure compared with placebo (OR 0.25, 95% CI 0.06 to 0.98, p=0.037 and OR 0.23, 95% CI 0.06 to 0.92, p=0.029, respectively).

Regarding secondary outcomes, the placebo group showed superior clinical improvement, defined as ≤10% involvement after treatment, compared with either miconazole or amorolfine (OR 0.26, 95% CI 0.08 to 0.91, p=0.028 and OR 0.25, 95% CI 0.07 to 0.85, p=0.021, respectively). No statistically significant differences were found for the remaining secondary outcome measures.

The comparison of both antifungals combined versus placebo showed similar results, as presented in online supplemental table S2. The results for the comparison of miconazole and amorolfine are presented in online supplemental table S3, showing no significant differences between both treatments.

The effect of treatment on the percentage of affected area and OSI scores over time is illustrated in figure 2.

Figure 2

Percentage of index toenail affliction and Onychomycosis Severity Index (OSI) scores with corresponding CIs over time. Y-axes represent percentage affliction (left) and OSI scores (right). X-axis represents the time point of measurement at baseline and follow-up.

Regarding the affected area, repeated measures ANOVA showed an overall significant difference in treatment effect over time between all groups (F(3.4–182.1)=4.8, p=0.002). Using pairwise comparisons, we found that this effect was due to a significant mean difference between amorolfine and placebo (10.7%, 95% CI 0.2% to 21.2%, p=0.045) rather than between miconazole and placebo (5.5%, 95% CI −5.2% to 16.2%, p=0.639), with both favouring placebo.

Similarly, we found an overall significant difference in treatment effect on OSI scores over time between all three groups (repeated measures ANOVA F(3.6–189.0)=2.6, p=0.045). Again, this was due to a significant mean difference between amorolfine and placebo (3.2, 95% CI 0.1 to 6.8, p=0.041) rather than between miconazole versus placebo (2.9, 95% CI −0.5 to 6.2, p=0.121), in favour of placebo.

Per-protocol analysis

The overall results of the per-protocol analysis are summarised in online supplemental table S4. The corresponding statistical tests, comparing miconazole with placebo, amorolfine with placebo and miconazole with amorolfine, are shown in online supplemental table S5A–C, respectively. In contrast to the intention-to-treat analysis, we found no significant differences in mycological cure for miconazole versus placebo (OR 0.22, 95% CI 0.04 to 1.15, p=0.082) or for amorolfine versus placebo (OR 0.28, 95% CI 0.07 to 1.19, p=0.073). In addition, neither miconazole nor amorolfine achieved a significant clinical improvement (≤10% affected) when compared with placebo (OR 0.35, 95% CI 0.09 to 1.28, p=0.105 and OR 0.28, 95% CI 0.08 to 1.03, p=0.048, respectively). As for the intention-to-treat analysis, there were no significant differences between amorolfine and miconazole.

Subgroup analysis

The results of the subgroup analysis based on the affected toenail area at baseline are shown in online supplemental table 6A,B, comparing miconazole and amorolfine with placebo, respectively. Miconazole was significantly less effective in reaching clinical improvement defined as ≤10% affected area at 6 months for patients having 11%–25% involvement at baseline (OR 0.35, 95% CI 0.11 to 1.116, p=0.049) compared with placebo. There were no other significant differences, nor between amorolfine and placebo.

Harms

The number of patients reporting adverse effects of miconazole or amorolfine compared with placebo did not differ significantly (OR 0.88, 95% CI 0.28 to 2.73, p=1.000 and OR 0.82, 95% CI 0.26 to 2.54, p=0.729, respectively). Furthermore, no SAEs or SUSARs were reported. The number and specific type of side effects are described in online supplemental table S7.

DiscussionSummary

In this primary care study, we investigated whether topical miconazole or amorolfine is effective in the treatment of mild to moderately severe onychomycosis. Neither was effective at reaching a complete or clinical cure. Based on intention-to-treat analyses, both antifungals were actually significantly less effective at reaching mycological cure and clinical improvement compared with placebo. By contrast, per-protocol analysis of therapy-compliant patients found no statistically significant differences between groups in mycological cure or clinical improvement. In addition, adverse effects, perceived improvement, acceptability, disease burden or quality of life were not significantly different between the antifungal and placebo groups.

Comparison with previous literature

The efficacy of miconazole has been investigated previously, but in studies that either lacked placebo control, or compared topical to systemic treatments.25–27 One study reported a high mycological cure rate (60%) at 6 months but suffered from substantial drop-out (40%) in the treatment group due to inefficacy.27 Previous investigations have exclusively used miconazole cream, thus the Onycho Trial was the first to apply miconazole as a nail lacquer, as well as the first to compare topical miconazole with placebo in a primary care setting.21

Compared with our results, previous studies of amorolfine reported noticeably better complete and mycological cures (38%–52% and 60%–75%, respectively), although again those studies were not placebo-controlled or blinded.23 24 A more recent open-label trial (N=1029) reported outcomes more comparable to ours, with complete, clinical and mycological cure rates for amorolfine of 0.96%, 3.8% and 15.7%, respectively.38 Although we did not find a significant treatment effect of miconazole or amorolfine as monotherapy, previous studies showed that amorolfine was effective when combined with oral treatment or when used as a prophylactic to prevent recurrence of onychomycosis.39 40 These methods of application were not included in our study.

In summary, previous studies on miconazole and amorolfine as monotherapy that showed better results were not placebo controlled or blinded. This could explain at least in part the higher cure rates. Our findings support a limited, in our case non-significant effect of amorolfine and contrast with previously reported cure rates for miconazole, possibly explained by performance and observer bias in previous studies.

Strengths and limitations

An important strength of this trial was the predominance of moderately severe cases, generally considered the most appropriate for topical treatment. Furthermore, selection, treatment and follow-up all took place in an outpatient setting representative of daily practice.

Standardised measurement approaches were used for clinical evaluation as well as confirmatory testing, improving the reliability and reproducibility of our findings. Strict blinding of patients and observers for treatment allocation and evaluations further limited potential bias.

Moreover, we applied accepted, frequently used outcome measures for a complete, clinical and mycological cure, as well as for clinical improvement, allowing our results to be easily compared with those of other trials.21 In addition, we anticipated criticism levelled at previous studies by evaluating treatment effect on symptom burden and quality of life.21 22

A limitation inherent to the selection criteria used in our study is the potential underrepresentation of certain patient groups, most notably those with severe onychomycosis, but also patients with Fitzpatrick skin-type >3 or diabetics excluded due to the risk of metformin–miconazole interactions. Our results are, therefore, not generalisable to these groups. In view of our inclusion of primarily milder cases, we expect that the use of miconazole or amorolfine in a broader population would likely worsen general outcomes.

Importantly, we found a stronger effect of placebo on mycological cure and clinical improvement (to <10%) compared with the antifungal treatments. Nevertheless, it is unlikely that this statistically significant difference was due to antifungal properties of the placebo given the lack of response in the per-protocol analysis, although an immune-stimulating effect of denatonium benzoate cannot be ruled out.31–35 Despite the common assumption that onychomycosis does not resolve if untreated, our findings are consistent with some previous trials showing comparable numbers of participants reaching clinical and mycological cure in the placebo groups, supporting the possibility of spontaneous improvement.21 41–43

Another limitation was the relatively short treatment period compared with other trials, in particular for reaching complete cure given the slow nail growth.21 However, miconazole had previously shown significant results at 6 months and amorolfine had been reported to give no further improvement beyond 6 months.22 27 In addition to these factors, we expected that a 6-month period would increase willingness to participate and limit potential drop-out, as well as better represent daily practice.

Conclusion and future directives

This study supports previous results for topical antifungal treatment of onychomycosis, which have been modest at best and often disappointing in their currently commercially available formulations.21 Based on our findings, topical treatment of mild to moderately severe onychomycosis with miconazole or amorolfine should not be advised.

As the antifungal efficacy of topical antimycotics in skin infections is well established, previously described poor penetration into nail tissue seems likely to have played an important role in the present and previous studies.44 45 Further research into topical antifungals may be justified if combined with modalities that enhance antifungal penetration into the nail.21 45 46