IntroductionOnychomycosis, or nail fungal infections, account for around 50% of nail diseases1. This condition is more prevalent in toenails compared to fingernails, with a mean prevalence in Europe and North America of 4.3% in population-based studies2. Overall, one epidemiological study reported 72.4% of onychomycosis cases as moderate-to-severe3 and, while non-life-threatening, it can lead to serious complications such as nail loss, tissue damage, cellulitis, osteomyelitis, and sepsis, particularly in high risk populations, like diabetes4,5. The diagnosis of onychomycosis may become a source of increased concern, considering the commonly reported esthetic issues and psychosocial factors such as embarrassment, low self-esteem, social withdrawal, the reduction of the quality of life, and the significant economic burden of topical and oral therapies6,7,8.Overall, dermatophytes are identified in 90% of toenail and in 50% of fingernail onychomycosis infections9. Trichophyton rubrum is the most frequent causative agent, whilst Trichophyton mentagrophytes and Epidermophyton floccosum can also cause it. Non-dermatophyte molds can also originate around 10–20% of onychomycosis cases, including Aspergillus spp., Fusarium spp., Acremonium spp., Scytalidium, and Scopulariopsis brevicaulis. Onychomycosis caused by yeasts, such as Candida albicans, is uncommon accounting for less than 10% of cases10,11. Confirmatory laboratory testing should be routinely performed before prescribing antifungal therapy to avoid treatment failures12,13. However, in clinical practice, it has been reported that only 39.3% of onychomycosis patients underwent diagnostic testing13. Within the last decade, sensitive tests (fungal culture testing or PCR) have become useful for identifying the origin of infection, especially in cases caused by non-dermatophyte molds and yeasts12. However, diagnostic methods for onychomycosis have notable limitations. Fungal culture, although considered the gold standard, may yield false-negative results in up to 30–50% of microscopically confirmed cases, as non-viable or non-cultivable organisms may preclude species-level identification 14,15. Dermatophyte test strips, while rapid and easy to use, may lack specificity, and PCR-based methods, although more sensitive, are not always available in routine clinical practice and can produce false positives by detecting non-viable organisms.Topical antifungal nail lacquers, such as ciclopirox, have been formulated to provide efficient delivery to the nail unit16,17. Ciclopirox has exhibited broad-spectrum antifungal activity against dermatophytes, molds and yeasts, including certain azole-resistant Candida species18. However, eradicating the causative organism and restoring the nail to its normal appearance may be challenging due to the high recurrence rates, treatment failures, and limited effectiveness associated with topical antifungals. The increasing incidence of antifungal resistance, particularly to systemic antifungals such as terbinafine and azoles, further complicates this challenge and must be carefully considered when selecting a treatment19,20. A recent review20 indicated that no resistance to ciclopirox has been reported, likely due to its mechanism of action. Ciclopirox acts as a chelating agent, binding to trivalent cations (e.g. Fe3+) and inhibiting metal-dependent enzymes involved in the degradation of toxic metabolites within fungal cells18. Furthermore, many onychomycosis patients prefer non-systemic treatment options9,21. Thus, treatment selection should be individualized according to the clinical characteristics of the disease, the causative organism, the presence of comorbidities and concurrent medications, and the patient’s preferences22. Oral agents including terbinafine, itraconazole, and fluconazole are used for moderate-to-severe onychomycosis, with the advantage of high cure rates and short treatment periods, but the disadvantages of potential drug interactions and risk of adverse events (AE) (e.g. hepatotoxicity)11,22,23. In addition to topical and systemic antifungals, other modalities such as laser therapy and photodynamic therapy have been explored as adjuvant treatments for onychomycosis, particularly in recalcitrant or severe cases24,25. However, evidence supporting their efficacy remains limited and inconsistent, and they are not routinely used as first-line options.Topical ciclopirox 8% hydroxypropyl chitosan (hereinafter referred to as CPX 8% HPCH) has a fungicidal/fungistatic as well as sporicidal activity, with HPCH acting as a film-forming agent, protecting the nail and increasing the penetration of ciclopirox26. Hence, CPX 8% HPCH is indicated for the treatment of mild-to-moderate fungal infections of the nails that are caused by ciclopirox-sensitive fungi, without nail matrix involvement, being generally well tolerated26. In some cases, when there is extensive involvement of one or several finger and toenails, additional treatment with oral antifungals, such as combined therapy of oral terbinafine and topical ciclopirox, may be considered27. However, no clinical evidence has been reported regarding combination of topical ciclopirox with oral itraconazole or fluconazole27. Indeed, it could be possible that evidence reported for clinical studies might not reflect the reality in clinical practice.The purpose of this study was to assess clinical characteristics and outcomes of onychomycosis patients treated with CPX 8% HPCH nail lacquer combined with oral antifungal treatment in the setting of real-world clinical practice in Spain. We used natural language processing (NLP) and machine learning (ML) techniques for extracting real-world data (RWD) from electronic health records (EHRs). Studies using this methodology in the assessment of topical and oral antifungal treatments approach in patients with onychomycosis during routinary clinical practice have not been previously reported.Materials and methodsDesign and study populationThis retrospective and multicenter study utilized RWD extracted from onychomycosis patients’ EHRs across 3 hospitals from the public Spanish National Health Care System between January 1, 2014, and March 31, 2023. The participating hospitals were all located in Madrid (Hospital Universitario Fundación Alcorcón, Hospital Universitario Infanta Leonor, and Hospital Universitario Puerta de Hierro). Inclusion criteria were determined to define the study population, encompassing subjects aged 18 years or older, attended in the dermatology departments, diagnosed with onychomycosis, and treated with CPX 8% HPCH nail lacquer (Ony-Tec®, Almirall S.A., Barcelona, Spain. Other registered (R) brands: Ciclopoli, Fulcare, Kitonail, Myconail, Niogermos, Niogermox, Onytec, Polinail, Privex, Rejuvenail) in combination with oral antifungals. No formal exclusion criteria were applied and only patients not meeting inclusion criteria were not considered part of the study population. The index date was defined as the earliest time point within the study period at which CPX 8% HPCH appeared in the unstructured text of the patient’s EHRs. The follow-up period was defined as the time elapsed from the index date to the last report available in the EHR. A window of 6 months pre- and post-index date was considered to define clinical manifestations and the concomitant use of oral antifungal treatment. A window of 12 months pre- and 6 months post-index date was considered to assess diagnostic methods and etiological pathogens. Specific analyses excluded patients with missing information. Thus, for time-to-event analyses, only patients with a record of the event (positive response, discontinuation or censoring) occurring after the index date were included.Data source and study variablesThe source of information was free text (i.e. unstructured) and structured information, including outpatient clinic reports, discharge reports, emergency reports, prescriptions, and other medical reports of patients with onychomycosis in the participating hospitals. Clinical information was extracted using the EHRead® technology developed by Medsavana S.L. (Madrid, Spain)28a powerful engine based on NLP and ML techniques for extracting text from EHRs29. To assess the quality of the information gathered from EHRs, the total number of screened records and patients were analyzed per site according to the main data sources and hospital departments (dermatology, emergency, and others). These data were later aggregated in the study database.Study variables included demographics and clinical characteristics, diagnostic methods, etiological agents, treatment-related variables, including clinical response and time to positive response, treatment synchronicity, type of oral antifungal agent (terbinafine, itraconazole, fluconazole), treatment persistence, treatment discontinuation, and safety. Diagnostic methods such as fungal culture were assumed to be performed according to the routine diagnostic procedures of each participating hospital. As this was a real-world study based on EHR data, specific methodological details were not uniformly available. However, all sites were tertiary hospitals with dermatology departments, where culture testing is typically performed under standard clinical microbiology protocols. Clinical response measured by standardized outcome definitions such as “complete cure” (defined as the combination of clinical and mycological resolution) could not be systematically applied due to the study nature based on hospital real world data. Therefore, an NLP-based classification was used: positive response, presumed positive response, partial response, and negative response. Positive response comprised explicit mentions in the EHR of clinical improvement or resolution. Conversely, presumed positive response entailed indirect or inferred improvement without clear documentation of clinical resolution such as referral to a primary care physician due to treatment effectiveness or free-text detection of terms associated to healing. Then, clinical response categories were always based on detection of related terms in free text, except for presumed positive response where inferences were also allowed. Treatment synchronicity related to the index date defined three study subgroups: (i) concomitant start: CPX 8% HPCH and oral antifungal mentions registered simultaneously (i.e. index date); (ii) initial oral antifungal: patients with the earliest mention of oral antifungal before index date; and (iii) initial CPX 8% HPCH: patients with the earliest oral antifungal mention after index date. Treatment persistence was calculated for the whole study population considering the time from the first mention of the treatment in the EHRs (combination therapy or each treatment component) until its discontinuation. Discontinuation was defined using data from the following categories in this order: time to cure, time to switch, time to severe potential AE, and time to treatment last mention (i.e. proxy for treatment stop). For combination treatment, time to discontinuation was calculated from the date in which the last treatment component was mentioned (either antifungal at index or later, or CPX 8% HPCH at index) until the earliest recorded cessation of any of them. For oral antifungal treatment, it was calculated as the time from the earliest mention of oral antifungal (either previous, at index, or posterior to index) until its discontinuation. And, for the CPX 8% HPCH treatment, it was calculated from the index date (i.e. CPX 8% HPCH earliest mention) until its discontinuation. Additional topical treatment during the follow-up period was also registered.Data extraction from electronic health recordsAll study variables and those terms required to construct them were predefined, reviewed, and approved by medical research experts, dermatologists, and NLP experts as well as by the required scientific and ethics committees. The EHRead® technology was implemented, using NLP and ML techniques for extracting and translating clinical terms, acronyms, abbreviations, and other clinical expressions used in routine clinical practice from EHRs into a study database. For this, conceptual definitions for all previously approved study variables were prespecified and aligned with clinical entities found in the SNOMED Clinical Terms (a comprehensive, computational processable collection of medical terms utilized in clinical documentation) using the SNOMED CT browser. All study variables were defined by mapping them to clinical terms and clinical expressions to be extracted from EHRs. However, the lack of these terms and expressions did not always represent actual absence of these clinical events since in routine clinical practice not all of them are homogenously registered.Statistical analysisThe sample size was calculated assuming that the total reference population covered by the three included sites was approximately 1 million patients and considering a prevalence of onychomycosis of 4.3% in the general population2 and a conservative estimate of all-age frequency of moderate-to-severe disease of 10%3. Considering that 100% of these patients with moderate-to-severe onychomycosis attended a specialist for treatment in the hospital setting, a total of 4,300 patients with onychomycosis diagnosis (95% confidence interval [CI] 1,900 to 6,800) were estimated to be detected from the participating hospitals. Therefore, we expected to include in the study population around 516 patients (95% CI 228 to 816) treated with CPX 8% HPCH plus an oral antifungal agent if the rate of use of the combination therapy exceeded 12% of the hospital population with moderate-to-severe onychomycosis. Categorical variables were expressed as frequencies and percentages, and continuous variables as mean and standard deviation (SD) or median and quartiles (Q1, Q3). The frequency of available or missing values was also reported. Time to positive response and time to discontinuation were analyzed as time-to-event analysis using the Kaplan-Meier method. A sensitivity analysis of patients having the mention of culture as diagnostic method was conducted ad hoc. Data analysis was carried out using “R” software (v4.0.2) (The R Foundation for Statistical Computing, Vienna, Austria).Ethical considerations and study approvalThe study obtained ethical approval from national and local Ethics committees, adhering to regional regulations. It complied with local laws, the latest Declaration of Helsinki, and Good Pharmacoepidemiology Practices. Due to the retrospective nature of the study, the ‘Comité de Ética de la Investigación con medicamentos del Hospital Universitario Fundación Alcorcón’ (‘Ethics Committee for Drug Research at the Alcorcón Foundation University Hospital’, Madrid, Spain), waived the need of obtaining informed consent (Approval code: M-22212-41, acta no. 5/22. Approval date: 25 May 2022). Being a retrospective study and using RWD, collection and variable evaluation were conducted blindly.ResultsOverall and clinical data of the study populationA total of 85,000,009 EHRs from 1,745,399 patients treated at the participating hospitals during the study period were processed. Within these, 3,541 adult patients were attended at the dermatology departments and registered an onychomycosis diagnosis. After applying other eligibility criteria and filters for antifungal treatment, there were 661 patients treated with CPX 8% HPCH, and 61.7% (n = 408) were concomitantly treated with an oral antifungal agent. Therefore, a study population of 408 patients treated with CPX 8% HPCH and oral antifungals was included in the analysis, which falls within the expected range based on the initial sample size estimation. The most common antifungal agent was terbinafine in 67.6% of patients (n = 276) followed by itraconazole in 20.8% (n = 85), and fluconazole in 11.5% (n = 47). Figure 1 shows the flow chart of the study population and details of the subgroups according to the type of oral antifungal agent and treatment synchronicity.Fig. 1Flow chart of the study population and subgroups by antifungal agent and treatment.synchronicity.Full size imageThe study population comprised 186 men and 222 women (45.6%, and 54.4%, respectively), with a mean (SD) age of 51.1 (14.8) years. The mean body mass index (BMI) was 27.9 (5.3) kg/m2. Dermatological comorbidities included tinea pedis interdigitalis in 12.3% of patients, psoriasis in 5.4%, nail psoriasis in 0.7%, and vitiligo in 0.5%. Hypertension and dyslipidemia were also frequent comorbid conditions, which were recorded in 17.4% and 14.5% of patients, respectively. No comorbidities were detected in the EHRs of more than half of the patients (51.7%). Repeated nail trauma and chemotherapy were the most frequently registered risk factors (7.1% and 2.0%, respectively), although no risk factors for onychomycosis were registered for 89.7% of the patients. Salient features at the index date are summarized in Table 1.Table 1 Sociodemographic characteristics, comorbidities, and risk factors at the index date.Full size tableSubungual hyperkeratosis was the most frequent clinical symptom detected (26.2%), followed by onycholysis (18.6%). The most common diagnostic methods were fungal culture (56.6%) and dermatophyte test strip (13.2%). Among the etiological agents, dermatophytes were the most common, especially Trichophyton rubrum (33.8%), and 10 patients (2.5%) had more than one of the listed causative pathogens. Clinical, diagnostic, and etiological data are shown in Table 2.Table 2 Clinical characteristics, diagnostic tests, and causative pathogens at the index date.Full size tableClinical response and time to responseThe clinical response according to treatment synchronicity is shown in Table 3. The positive response was 15.7% in the study population, with the highest percentage (17.0%) registered in the subgroup of patients who started CPX 8% HPCH before oral antifungal therapy (initial CPX 8% HPCH subgroup). The sum of both positive and presumed positive responses was 75.5% for the entire study population with the higher rate (84.2%) in the subgroup of patients who started oral antifungal therapy before topical use of CPX 8% HPCH nail lacquer (initial oral antifungal subgroup) compared to the subgroups of concomitant start of combined therapy (70.9%) and initial CPX 8% HPCH (77.4%).Table 3 Clinical response to treatment in the study population according to synchronicity of combined treatment.Full size tableThe percentage of patients with positive and presumed positive responses were 85.1%, 81.2%, and 72.1% in those treated with fluconazole, itraconazole, and terbinafine, respectively (Table 4).Table 4 Clinical response to treatment according to the type of oral antifungal.Full size tableThe median time to positive response was 5.08 months for the combination of CPX 8% HPCH and oral antifungals in patients who registered a positive response (including clinical and/or mycological cure) within the study period (Fig. 2).Fig. 2Time to positive response in patients who registered positive response within the study period for the combination of oral antifungal agents and CPX 8% HPCH nail lacquer. Note that due to curve stabilization, X-axis was truncated at 30 months. Only patients with an event (positive response) occurring after the index date were included. Accordingly, 13 patients were excluded. Patients were censored at CPX 8% HPCH discontinuation or end of follow-up. Dotted line marks 5.08 months which is the median time to response for patients included in this figure.Full size imageTreatment persistenceTreatment persistence for CPX 8% HPCH, oral antifungal agents, and combined treatment is shown in Table 5. The median (Q1, Q3) treatment persistence for the treatment combination, for oral antifungal agents, and for CPX 8% HPCH was 2.72 (0.95, 4.72), 2.93 (1.38, 6.07), and 4.98 (1.48, 6.77) months, respectively.Table 5 Treatment persistence of combination therapy and of each treatment component (oral antifungal and CPX 8% HPCH) in the entire study population.Full size tableTime to discontinuation (in months) of onychomycosis combination treatment (CPX 8% HPCH and oral antifungal) in all study population is shown in Fig. 3 and in patients with registered positive response in Fig. 4. The probability of discontinuation reached 0.13 at the 24-month point for the study population (Fig. 3). The median time to discontinuation in patients with registered positive response was 3.21 months (Fig. 4).Fig. 3Time to CPX 8% HPCH and oral antifungal combination discontinuation in the entire study population. X-axis truncated at 24 months due to curve stabilization and Y-axis truncated at 0.13 given no events appeared beyond this point. Only patients with a discontinuation or censoring occurring after the index date were included. Accordingly, 26 patients were excluded. Patients were censored at the end of follow-up.Full size imageFig. 4Time to CPX 8% HPCH and oral antifungal combination discontinuation in patients with registered positive response. Note that due to curve stabilization, X-axis was truncated at 18 months. Only patients with a discontinuation or censoring occurring after the index date were included. Accordingly, 9 patients were excluded. Patients were censored at the end of the follow-up. Dotted line marks 3.21 months which is the median time to discontinuation for patients included in this figure.Full size imageAdverse eventsPotential AEs reported in the EHRs during the study period in the entire study population and in the subgroups of patients according to the oral antifungal agent combined with CPX 8% HPCH is shown in Table 6. The most frequent AEs were erythema (5.6%), diarrhea (4.9%), and fever (4.2%). The occurrence of AEs was similar across the different subgroups of oral antifungal agents.Table 6 Adverse events recorded during the study in the entire study population and in the subgroups of CPX 8% HPCH combined with each oral antifungal agent.Full size tableAdditional topical treatmentsA small percentage of patients received additional topical treatments during the study period, including topical amorolfine in 9.1% (n = 37), laser therapy in 4.9% (n = 20), topical miconazole in 4.4% (n = 18), topical ketoconazole in 1% (n = 4), topical clotrimazole in 0.7% (n = 3), and photodynamic therapy in 0.2% (n = 1).Sensitivity analysisPatients with a registered mention of culture as diagnostic method (56.6%, n = 231) showed similar result patterns compared with the entire study population (Supplementary Tables 1 to 6 and Supplementary Figs. 1 to 3).DiscussionThe present study provides real-world evidence of the management of patients diagnosed with onychomycosis in daily practice conditions. A key finding of the study was that treatment combinations of CPX 8% HPCH with terbinafine, itraconazole, and fluconazole were commonly used in our clinical practice setting, which is consistent with data of other studies indicating that the association of antifungal agents to topical therapy may be useful to accelerate the clinical and microbiological healing of superficial dermatophytes infections30.When contextualizing our results within the framework of previously described literature findings, it is important to consider the methodological differences between our study and the previously published randomized controlled trials. First, it is crucial to emphasize distinctions between the research product utilized in the current study (CPX 8% HPCH nail lacquer) and the ciclopirox employed in other studies. In this regard, the addition of HPCH, a water-soluble biopolymer, has been shown to improve the efficacy of ciclopirox nail lacquer as shown in a multicenter, randomized, three-arm, placebo-controlled study of 467 patients31. While brief evidence exists regarding the combination of topical ciclopirox with terbinafine32,33,34no previous studies have evaluated the combinations of ciclopirox 8% with other oral antifungals. In a clinical study of 68 patients randomly assigned to oral terbinafine 250 mg/day for 16 weeks or a combination of oral terbinafine 250 mg/day for 16 weeks and topical ciclopirox nail lacquer once daily for 9 months, the mycological cure rates were 64.7% for terbinafine monotherapy vs. 88.2% for the combined therapy32. In a multicenter randomized pilot study of 73 patients with moderate to severe toenail onychomycosis, mycological cure, at week 48, was observed in 70.4% of patients treated with the combination of ciclopirox 8% nail lacquer topical solution and terbinafine 250 mg/day and in 56% of patients treated with oral terbinafine alone33. However, in an open randomized comparative study of 96 patients, a small difference in mycological cure rates was found between oral terbinafine pulse therapy in combination with topical ciclopirox olamine 8% (cure rate 83.3%) and terbinafine pulse therapy as monotherapy (cure rate 82.6%)34. Although all these results showed higher mycological cure rates in patients treated with the combination therapy, results are conflicting due to methodological differences among the three studies27. Second, our study is entirely based on RWD from clinical records, unlike randomized trials, which usually apply strict inclusion and exclusion criteria that often make it difficult to generalize to real-life settings. This is important as it fills a gap in the current literature by providing information about the characteristics, management, and outcomes of these patients in routine clinical practice, which may translate into easier decision-making by doctors during the implementation of antifungal therapy.As far as we are aware, this is the first study describing the use of ciclopirox 8% HPCH combined with three different oral antifungals in a large study population of 408 patients with onychomycosis. In over half of the patients, the causative pathogen was not recorded in the EHRs. This reflects the real-world clinical context where treatment decisions may be based on clinical presentation or rapid diagnostic tests that suggest fungal infection but do not identify the specific organism. As the study relied on retrospective EHR data, we were unable to assess the exact rationale for antifungal selection in each case. The present results show that the response rates as reported by clinicians in EHRs were similar across the combinations of CPX 8% HPCH with terbinafine, itraconazole and fluconazole, with an overall positive response of 15.7%. We also found that 59.8% of patients had a presumed positive response that was non-confirmed but registered in their clinical records. When positive and presumed positive responses were considered, the overall response rate was 75.5%. This percentage is in line with the high variability of response rates reported by Falotico et al. in a recently published systematic review27. The authors reported that cure percentages for the combination of oral terbinafine plus ciclopirox (non-HPCH) range from 33 to 83.3% depending on the treatment protocol, time to event, and type of endpoints assessed, such as mycological cure, clinical cure, or clinical improvement, among others27. In our study, only two patients had negative response to treatment, meaning that they had treatment failure during the study period. This finding was expected since the efficacy of the combination therapy is not 100% for the onychomycosis population. The median time to positive response was 5.08 months. Interestingly, a similar trend of response rate was observed when stratified by treatment synchronicity (concomitant start, initial oral antifungal, and initial CPX 8% HPCH), and by type of antifungal agent (terbinafine, itraconazole, and fluconazole). This observation facilitates flexibility in the use of combined treatment with CPX 8% HPCH and systemic antifungals in daily practice.In this study, treatment persistence was calculated for the combination therapy and for each treatment component, reporting median times of 2.72, 2.93, and 4.98 months for the combination therapy, oral antifungal, and CPX 8% HPCH, respectively. In parallel, we evaluated the time to discontinuation as reporting this endpoint in real-world studies is relevant to complement information about treatment effectiveness. For the entire study population, the probability of discontinuation reached 0.13 at the 24-month point, indicating that the median time of discontinuation was not achieved during the study period. For patients with a registered positive response, the analysis of time to discontinuation of the combination (CPX 8% HPCH and oral antifungal) showed that the median time to discontinuation was 3.21 months and that almost all patients discontinued the treatment at the time point of 18 months. These results are expected in routine clinical practice since the registered times are under the clinical guidelines recommendations.Another interesting finding was the very small percentage of patients in which additional topical treatment was recorded during the follow-up period. The low frequency of additional topical treatments such as laser therapy and photodynamic therapy observed in our cohort may reflect their limited availability or perceived effectiveness in routine hospital practice. While some studies suggest potential benefits as adjunctive therapies, their role remains controversial and lacks robust supporting evidence, especially in comparison to standard antifungal regimens. This finding may indirectly support the effectiveness of CPX 8% HPCH nail lacquer associated with oral antifungal treatment for the management of onychomycosis. Finally, patients’ demographic characteristics, risk factors and comorbidities were in line with previous studies35 showing a median age over 50 years, sex balanced distribution, tinea pedis as common risk factor, as well as hypertension, dyslipidemia, obesity and diabetes mellitus as comorbidities. Our results are consistent with previous findings on the relationship between diabetes and onychomycosis. A recent study conducted in Spain also highlights the prevalence of fungal pathogens in diabetic patients, supporting the importance of considering metabolic comorbidities when managing onychomycosis in routine clinical practice5.The treatment schedule of CPX 8% HPCH combined with oral antifungals was well tolerated and safe, with erythema (5.6%), diarrhea (4.9%), and fever (4.2%) as the most frequently registered potential AEs. The interpretation of AEs is limited by the fact that symptoms caused by CPX 8% HPCH from those attributable to treatment with oral antifungals cannot be differentiated. Retrospective analysis of AEs with topical onychomycosis medications reported to the United States Food and Drug Administration showed that drug ineffectiveness was the most common AE associated with ciclopirox 8% solution36whereas taste disturbance (terbinafine) and drug interactions (itraconazole and fluconazole) were the most frequent AEs associated with systemic onychomycosis medications37.The use of artificial intelligence (AI) related tools for extracting data from EHRs is a novel approach that is very useful for the assessment of the clinical management of medical conditions in the real-world setting. In this study, we described the effectiveness and safety profile of ciclopirox 8% formulated with HPCH in combination with oral antifungals. An additional advantage of ciclopirox 8% HPCH is its unique mechanism of action, which involves disruption of metal-dependent fungal enzymes, and the lack of documented clinical resistance. This pharmacological profile may provide a complementary mechanism to systemic antifungals and could be particularly relevant in the current landscape of increasing antifungal resistance, including the emergence of resistant strains such as Trichophyton indotineae. However, this study presents some constraints that should be considered. As is common in real-world retrospective studies based on EHRs, a key limitation lies in the extent and accuracy of clinical documentation. For example, the outcomes reported do not differentiate between clinical and mycological cure, since this distinction is rarely reflected in routine EHRs. To address this, we developed an NLP-based classification of treatment response (positive, presumed positive, partial, or negative). While we acknowledge that this approach entails a degree of subjectivity and potential misclassification bias, it responds to the high variability in how clinical improvement is documented—often with underreporting of favorable outcomes compared to treatment failure. Therefore, we consider it a valid and pragmatic method for extracting complex outcome variables from unstructured data, while also highlighting the need to standardize clinical documentation to improve the quality of future real-world research. Similarly, details such as antifungal dosage and duration, clinical form of onychomycosis, fingernail or toenail involvement, and antifungal susceptibility testing were not consistently recorded and thus could not be analyzed. This limits our ability to explore whether treatments differed between clinical subgroups or adhered to guidelines. The same applies to podiatric care and nail debridement (important for enhancing topical treatment efficacy), which were rarely documented. These absences underscore the importance of improving the systematic registration of relevant clinical variables in routine practice. Additional limitations relate to missing data on key diagnostic variables. Signs and symptoms, diagnostic methods, and etiological agents were not identified in a substantial proportion of patients, likely due to documentation variability and the retrospective design. In particular, the absence of pathogen identification in over 50% of cases may reflect incomplete reporting or the use of diagnostic methods that do not provide etiological detail. Importantly, this does not necessarily imply lack of diagnostic confirmation, as data may have been misclassified or not extractable. To assess the impact of this limitation, we conducted a sensitivity analysis showing that patients with culture-confirmed diagnoses had clinical patterns like those in the overall cohort. While clinical similarity does not guarantee diagnostic certainty, this suggests a low risk of false positives in our population. Additionally, it is important to note that inclusion in the study population was based not only on diagnostic mentions but also on references to treatment in the EHRs, which helps reduce the likelihood of including false-positive cases, even when diagnostic methods were not explicitly documented. The reliance on routine diagnostic methods—primarily fungal culture—may have led to underreporting of non-dermatophyte molds and mixed infections. As molecular techniques such as PCR were not systematically used or recorded, pathogen identification was limited to what could be captured by standard methods. Fungal identification was often restricted to the genus level, particularly for non-dermatophyte molds, due to documentation practices and NLP extraction constraints. Accordingly, the microbiological analysis was performed at this taxonomic level, which limits the granularity of findings. While 176 patients had at least one identified pathogen and 10 had more than one, it remains uncertain whether rare pathogens were present among the remaining 232 cases, precluding further etiological subgroup analyses. Lastly, since the population consisted of hospital-attended patients, it likely reflects moderate to severe cases, limiting the generalizability of findings to primary care. Nevertheless, the multicenter approach and the application of EHRead® technology strengthen the representativeness and reliability of the patient characterization and clinical management observed in real-world practice.ConclusionsThis study offers a comprehensive overview of onychomycosis patients in a real-world hospital setting, describing effectiveness and safety profile of a product containing ciclopirox 8% formulated with HPCH in combination with oral terbinafine, itraconazole and fluconazole. Our results corroborate that CPX 8% HPCH combination with terbinafine is the most common association prescribed, being either itraconazole or fluconazole combinations similarly used. Singularly, we did not observe clinically relevant differences in response rates according to oral antifungal or treatment synchronicity, being frequency of response rates analogous to those previously published. In addition, no threatening AEs were detected, highlighting the safety profile of the combinations studied. The large study population and the robustness of data extracted from EHRs and analyzed using AI-related tools provide evidence of the benefit of combined treatment. While these results must be interpreted within the constraints of observational real-world research, they provide relevant and robust evidence that complements findings from clinical trials.Data availabilityThe study data are available from the corresponding author upon request.ReferencesGupta, A. K. et al. The prevalence of unsuspected onychomycosis and its causative organisms in a multicentre Canadian sample of 30 000 patients visiting physicians’ offices. J. Eur. Acad. Dermatol. Venereol. 30, 1567–1572. https://doi.org/10.1111/jdv.13677 (2016).Article CAS Google Scholar Sigurgeirsson, B. & Baran, R. The prevalence of onychomycosis in the global population: a literature study. J. Eur. Acad. Dermatol. Venereol. 28, 1480–1491. https://doi.org/10.1111/jdv.12323 (2014).Article CAS Google Scholar Gupta, A. K. et al. Prevalence and epidemiology of onychomycosis in patients visiting physicians’ offices: a multicenter Canadian survey of 15,000 patients. J. Am. Acad. Dermatol. 43, 244–248. https://doi.org/10.1067/mjd.2000.104794 (2000).Article CAS Google Scholar Thomas, J. et al. Toenail onychomycosis: an important global disease burden. J. Clin. Pharm. Ther. 35, 497–519. https://doi.org/10.1111/j.1365-2710.2009.01107.x (2010).Article CAS Google Scholar Navarro-Perez, D. et al. Prevalence and risk factors predicting onychomycosis in patients with and without diabetes mellitus in spain: A Cross-Sectional study. J. Fungi (Basel). 10 https://doi.org/10.3390/jof10110790 (2024).Chacon, A., Franca, K., Fernandez, A. & Nouri, K. Psychosocial impact of onychomycosis: a review. Int. J. Dermatol. 52, 1300–1307. https://doi.org/10.1111/ijd.12122 (2013).Article Google Scholar Gupta, A. K. Pharmacoeconomic analysis of oral antifungal therapies used to treat dermatophyte onychomycosis of the toenails. A US analysis. Pharmacoeconomics 13, 243–256. https://doi.org/10.2165/00019053-199813020-00007 (1998).Article CAS Google Scholar Gupta, A. K. & Mays, R. R. The impact of onychomycosis on quality of life: A systematic review of the available literature. Skin. Appendage Disord. 4, 208–216. https://doi.org/10.1159/000485632 (2018).Article Google Scholar Bodman, M. A., Syed, H. A. & Krishnamurthy, K. in StatPearls (2024).Leung, A. K. C. et al. Onychomycosis: an updated review. Recent. Pat. Inflamm. Allergy Drug Discov. 14, 32–45. https://doi.org/10.2174/1872213X13666191026090713 (2020).Article CAS Google Scholar Maskan Bermudez, N., Rodriguez-Tamez, G., Perez, S., Tosti, A. & Onychomycosis Old and new. J. Fungi (Basel). 9 https://doi.org/10.3390/jof9050559 (2023).Falotico, J. M. & Lipner, S. R. Updated perspectives on the diagnosis and management of onychomycosis. Clin. Cosmet. Investig Dermatol. 15, 1933–1957. https://doi.org/10.2147/CCID.S362635 (2022).Article Google Scholar Geizhals, S., Cooley, V. & Lipner, S. R. Diagnostic testing for onychomycosis: A retrospective study over 17 years. J. Am. Acad. Dermatol. 83, 239–241. https://doi.org/10.1016/j.jaad.2019.12.019 (2020).Article Google Scholar Graser, Y., Czaika, V., Ohst, T. & Diagnostic PCR of dermatophytes–an overview. J. Dtsch. Dermatol. Ges. 10, 721–726. https://doi.org/10.1111/j.1610-0387.2012.07964.x (2012).Article Google Scholar Gupta, A. K., Hall, D. C., Cooper, E. A. & Ghannoum, M. A. Diagnosing Onychomycosis: What’s New? J. Fungi (Basel). 8 https://doi.org/10.3390/jof8050464 (2022).Baran, R. & Kaoukhov, A. Topical antifungal drugs for the treatment of onychomycosis: an overview of current strategies for monotherapy and combination therapy. J. Eur. Acad. Dermatol. Venereol. 19, 21–29. https://doi.org/10.1111/j.1468-3083.2004.00988.x (2005).Article CAS Google Scholar Feng, X., Xiong, X. & Ran, Y. Efficacy and tolerability of Amorolfine 5% nail lacquer in combination with systemic antifungal agents for onychomycosis: A meta-analysis and systematic review. Dermatol. Ther. 30 https://doi.org/10.1111/dth.12457 (2017).Almirall. in almirall.com. (ed. Almirall) 60 (2020).Gupta, A. K., Talukder, M., Carviel, J. L., Cooper, E. A. & Piguet, V. Combatting antifungal resistance: paradigm shift in the diagnosis and management of onychomycosis and dermatomycosis. J. Eur. Acad. Dermatol. Venereol. 37, 1706–1717. https://doi.org/10.1111/jdv.19217 (2023).Article CAS Google Scholar Gupta, A. K. et al. Treatment of onychomycosis in an era of antifungal resistance: role for antifungal stewardship and topical antifungal agents. Mycoses 67, e13683. https://doi.org/10.1111/myc.13683 (2024).Article CAS Google Scholar Mickle, A. T. et al. Toenail onychomycosis with or without diabetes in canada: patient treatment preferences and health state utilities. Patient Prefer Adherence. 18, 475–486. https://doi.org/10.2147/PPA.S450215 (2024).Article Google Scholar Lipner, S. R., Scher, R. K. & Onychomycosis Treatment and prevention of recurrence. J. Am. Acad. Dermatol. 80, 853–867. https://doi.org/10.1016/j.jaad.2018.05.1260 (2019).Article Google Scholar Aggarwal, R., Targhotra, M., Kumar, B., Sahoo, P. K. & Chauhan, M. K. Treatment and management strategies of onychomycosis. J. Mycol. Med. 30, 100949. https://doi.org/10.1016/j.mycmed.2020.100949 (2020).Article CAS Google Scholar Yousefian, F., Smythe, C., Han, H., Elewski, B. E. & Nestor, M. Treatment options for onychomycosis: efficacy, side effects, adherence, financial considerations, and ethics. J. Clin. Aesthet. Dermatol. 17, 24–33 (2024).Google Scholar Alves, R. O. et al. Antimicrobial photodynamic therapy in onychomycosis management: A systematic review of clinical trials. Photodiagnosis Photodyn Ther. 104640 https://doi.org/10.1016/j.pdpdt.2025.104640 (2025).Piraccini, B. M., Iorizzo, M., Lencastre, A., Nenoff, P. & Rigopoulos, D. Ciclopirox hydroxypropyl Chitosan (HPCH) nail lacquer: A review of its use in onychomycosis. Dermatol. Ther. (Heidelb). 10, 917–929. https://doi.org/10.1007/s13555-020-00420-9 (2020).Article Google Scholar Falotico, J. M., Lapides, R. & Lipner, S. R. Combination therapy should be reserved as Second-Line treatment of onychomycosis: A systematic review of onychomycosis clinical trials. J. Fungi (Basel). 8 https://doi.org/10.3390/jof8030279 (2022).Hernández Medrano, I. et al. Savana: Re-using electronic health records with artificial intelligence. IJIMAI J. 4, 5. https://doi.org/10.9781/ijimai.2017.03.001 (2018).Article Google Scholar Espinosa-Anke, L. et al. Savana: A global information extraction and terminology expansion framework in the medical domain. Proces Del. Leng Nat. 57, 23–30 (2016). http://journal.sepln.org/sepln/ojs/ojs/index.php/pln/article/view/5333Google Scholar Brescini, L., Fioriti, S., Morroni, G. & Barchiesi, F. Antifungal combinations in dermatophytes. J. Fungi (Basel). 7 https://doi.org/10.3390/jof7090727 (2021).Baran, R. et al. An innovative water-soluble biopolymer improves efficacy of Ciclopirox nail lacquer in the management of onychomycosis. J. Eur. Acad. Dermatol. Venereol. 23, 773–781. https://doi.org/10.1111/j.1468-3083.2009.03164.x (2009).Article CAS Google Scholar Avner, S., Nir, N. & Henri, T. Combination of oral terbinafine and topical Ciclopirox compared to oral terbinafine for the treatment of onychomycosis. J. Dermatolog Treat. 16, 327–330. https://doi.org/10.1080/09546630500420183 (2005).Article CAS Google Scholar Gupta, A. K. & Study, O. C. T. Ciclopirox topical solution, 8% combined with oral terbinafine to treat onychomycosis: a randomized, evaluator-blinded study. J. Drugs Dermatol. 4, 481–485 (2005).Google Scholar Jaiswal, A., Sharma, R. P. & Garg, A. P. An open randomized comparative study to test the efficacy and safety of oral terbinafine pulse as a monotherapy and in combination with topical Ciclopirox olamine 8% or topical Amorolfine hydrochloride 5% in the treatment of onychomycosis. Indian J. Dermatol. Venereol. Leprol. 73, 393–396. https://doi.org/10.4103/0378-6323.37056 (2007).Article Google Scholar Albucker, S. J., Falotico, J. M., Choo, Z. N., Matushansky, J. T. & Lipner, S. R. Risk factors and treatment trends for onychomycosis: A Case-Control study of onychomycosis patients in the all of Us research program. J. Fungi (Basel). 9 https://doi.org/10.3390/jof9070712 (2023).Wang, Y. & Lipner, S. R. Retrospective analysis of adverse events with topical onychomycosis medications reported to the united States food and drug administration. Arch. Dermatol. Res. 312, 581–586. https://doi.org/10.1007/s00403-020-02044-7 (2020).Article Google Scholar Wang, Y. & Lipner, S. R. Retrospective analysis of adverse events with systemic onychomycosis medications reported to the united States food and drug administration. J. Dermatolog Treat. 32, 783–787. https://doi.org/10.1080/09546634.2019.1708242 (2021).Article Google Scholar Download referencesAcknowledgementsThe authors thank Dr. Alba Martínez-Alcocer Martínez, Dr. Ángel Rosell Díaz, Dr. Ángela García Miñarro, Dr. Elena García Zamora, Dr. Joseph Griffiths, and Dr. María Gamo Guerrero for performing the external evaluation of clinical variables extracted with NLP. The authors thank Dr. Marta Pulido for editing the manuscript and editorial assistance. Savana Research Group members in alphabetical order: Lucía Cabal-Hierro, David Casadevall, Víctor Fanjul, Judith Marín-Corral, Luisa Martínez, Claudia Maté, Sebastian Menke, Natalia Polo, Margarita Posso, Ignacio Salcedo, Daniel Salvador and Miren Taberna.FundingThis research was founded by Polichem S.A.Author informationAuthor notesA comprehensive list of consortium members appears at the end of the paper.Authors and AffiliationsDermatology Department, Hospital Puerta de Hierro Majadahonda, Madrid, SpainGastón RoustanDermatology Department, Hospital Universitario Fundación Alcorcón, Madrid, SpainJosé Luis López EstébaranzDermatology Department, Hospital Universitario Infanta Leonor, Madrid, SpainPablo De La CuevaSavana Research S.L., Madrid, SpainLucía Cabal-Hierro, David Casadevall, Víctor Fanjul, Judith Marín-Corral, Luisa Martínez, Claudia Maté, Sebastian Menke, Natalia Polo, Margarita Posso, Ignacio Salcedo, Daniel Salvador & Miren TabernaAlmirall S.A, Medical Affairs Department, Barcelona, SpainFrancesca PajueloAlmirall S.A., Research and Development Center, Sant Feliu de Llobregat, Barcelona, SpainMaría Luisa TamaritAlmirall S.A, Global Medical Affairs Department, Barcelona, SpainAida Valmaseda & Jordi GalvánAuthorsGastón RoustanView author publicationsSearch author on:PubMed Google ScholarJosé Luis López EstébaranzView author publicationsSearch author on:PubMed Google ScholarPablo De La CuevaView author publicationsSearch author on:PubMed Google ScholarFrancesca PajueloView author publicationsSearch author on:PubMed Google ScholarMaría Luisa TamaritView author publicationsSearch author on:PubMed Google ScholarAida ValmasedaView author publicationsSearch author on:PubMed Google ScholarJordi GalvánView author publicationsSearch author on:PubMed Google ScholarConsortiaSavana Research GroupLucía Cabal-Hierro, David Casadevall, Víctor Fanjul, Judith Marín-Corral, Luisa Martínez, Claudia Maté, Sebastian Menke, Natalia Polo, Margarita Posso, Ignacio Salcedo, Daniel Salvador & Miren TabernaContributionsConceptualization: G. Roustan, J. López-Estebaranz, P. de la Cueva, Savana Research, and J. Galván; methodology, Savana Research; software, Savana Research.; validation, G. Roustan, J. López-Estebaranz, P. de la Cueva, Savana Research, and J. Galván; formal analysis Savana Research; investigation, G. Roustan, J. López-Estebaranz, P. de la Cueva, Savana Research, F. Pajuelo, M. Tamarit, A. Valmaseda, and J. Galván; resources, F. Pajuelo, M. Tamarit, A. Valmaseda, and J. Galván; data curation, Savana Research; writing—original draft preparation, Savana Research; writing—review and editing, G. Roustan, J. López-Estebaranz, P. de la Cueva, Savana Research, and J. Galván. All authors have read and agreed to the published version of the manuscript.Corresponding authorCorrespondence to Jordi Galván.Ethics declarationsCompeting interestsThe authors declare no competing interests.Institutional review board statementApproval by the Ethics Committee of the participating hospitals was obtained.Additional informationPublisher’s noteSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Electronic supplementary materialBelow is the link to the electronic supplementary material.Supplementary Material 1Rights and permissionsOpen Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.Reprints and permissionsAbout this article