We report the first multicenter European study assessing physicians’ knowledge about current recommendations on diagnosis and treatment of IFI. We could demonstrate that even frequent prescribers have a significant need of continuous education. Most common mistakes lead to antifungal over consumption, since many physicians erroneously treat fungal colonization, use unnecessary high doses of L-AmB and administered combined antifungal therapy with no supporting scientific evidence.

Our survey corroborates gaps in both diagnosis and management of IFI in 5 institutions of 4 European countries. Physicians have problems for differentiating colonization from infection when Candida spp. is isolated in urine or in a tracheal aspirate, which could lead to an over prescription of antifungals. In a prospective study conducted in a Thai tertiary care setting, Sutepvarnon et al. demonstrated a positive correlation between isolation of Candida species from urine and unnecessary treatment [5]. Other common scenario of avoidable over treatment is the isolation of Candida in respiratory secretions which should not be the only indication for starting antifungal therapy [7,10].

Another aspect deserving attention is the lack of knowledge of the current indications of antifungal prophylaxis and empirical treatment for invasive candidiasis. Almost 50% of physicians would initiate Candida prophylaxis in every ICU patient colonized by Candida spp. or in the clinical scenario of an ICU patient having an urinary indwelling catheter, central venous catheter and recent surgery, without taking into consideration other risk factors or clinical manifestations. A probable reason for a such inadequate management could be related to the results of at least four prospective studies of ICU and surgical patients suggesting a reduction in Candida infection and mortality under fluconazole prophylaxis [11-14]. However, since the universal administration of antifungal prophylaxis remains an inefficient strategy that may increase subsequent azole-resistance or non-albicans candidemia [15-17], it is currently warranted only in selected ICU patients at highest risk (>10%) of invasive candidiasis [7,18].

On the other hand, we observed a partial impact on medical practice of the articles demonstrating the importance of an early initiation of targeted antifungal treatment when there is a clinical suspicion of invasive candidiasis [19-21]. Unfortunately, about 20% of physicians delay the start of antifungal treatment after being informed of positive blood cultures, whereas the critical window of opportunity for antifungal initiation appears to be 12–24 hours following the first positive blood cultures were drawn [20,22]. Moreover, we reported that 60% of physicians (staff more frequently than residents and intensivists more frequently than medical physicians) failed to identify the indications of antifungal treatment in patients with suspicion of catheter related infection [23].

However, the clinical outcome in IFI related sepsis is not only related to adequacy and timeliness of antifungal administration, but also on appropriate dosing, all these factors associated with length of hospital stay, health care costs, morbidity and mortality [20,22,24]. We found that another critical point regarding knowledge of Candida infections management is the appropriate fluconazole dosage, that was correctly known by only 62% of the physicians. In a retrospective cohort study performed by Labelle et al., inadequate initial fluconazole dose was prescribed in about half of the critical patients with invasive candidiasis and it was associated with an increased risk of in-hospital mortality [25]. Indeed, failure to achieve pharmacodynamic targets for fluconazole has been associated with worse outcomes [26-29]. For this reason, to obtain clinical success, dose of fluconazole should be tailored to achieve AUC/MIC ratios of at least 25 (using Clinical and Laboratory Standards Institute MIC methodology) [30] or 100 (using European Committee on Antimicrobial Susceptibility Testing MIC methodology). These targets usually require 6 mg/kg/day (following 12 mg/kg loading) for susceptible isolates or 12 mg/kg/day for C. glabrata or other isolates with MICs of 16-32 mg/L [7].

Most physicians were not aware of their local incidence of azoles resistance in Candida, which could enhance the prescription of broad-spectrum antifungals. Although non-albicans strains have clearly increased, in many European centers the rate of fluconazole resistance is still less than 5% [31-34].

More surprisingly to us was to find that 12% and 9.2% of physicians would select L-AmB and voriconazole, respectively, instead of a candin to treat these fluconazole-resistant Candida infections. L-AmB should be restricted to selected cases of intra-abdominal candidiasis [35,36] or intolerance to other antifungal agents due to its potential toxicity and higher cost, and voriconazole should be limited to step-down oral therapy for selected cases of candidiasis due to C. krusei or voriconazole-susceptible C. glabrata [7]. Finally, we would like to stress that there are differences among published guidelines regarding some diagnostic and therapeutic aspects of candidiasis that need to be clarified for the sake of a more clear understanding by the prescribing physicians [37].

Regarding IA, problems to differentiate colonization from infection were also evident. As previously demonstrated only 22.3% of Aspergillus isolates from respiratory tract corresponds to probable/proven IA episodes. For this reason, Bouza et al. proposed a prediction score taking into account the procedure used to obtain the sample, the presence of leukaemia, neutropenia or the use of corticoids to help clinicians in the interpretation of Aspergillus cultures [38].

Another problem was that a high percentage of the prescribing physicians still consider L-AmB as the first line therapy for IA, whereas the largest prospective, randomized trial for the treatment of IA demonstrated that voriconazole was superior to D-AmB [8,39]. Furthermore, 3.5% of the physicians continue to believe that high doses of L-AmB are necessary to treat IA, ignoring the results of the AmBiLoad Trial [40,41]. The efficacy of voriconazole was further demonstrated in paediatric and adult patients receiving voriconazole for treatment of IA who were refractory or intolerant to conventional antifungal therapy [8,39,42,43]. The role of combination therapy as primary or salvage therapy is uncertain and it is not actually recommended in international guidelines [8].

The majority of physicians ignored that current guidelines recommend that treatment of IA should be continued for a minimum of 6–12 weeks and that voriconazole plasma levels should be monitored during treatment to avoid toxicity and therapeutic failure [44-47].

Finally, we found differences in knowledge between experimented physicians and residents. As we expected, the haematologists and infectious disease specialists (medical departments), are more proficient in the use of antifungal therapy. However, ICU physicians that are directly responsible for many pre-emptive and empirical antifungal prescriptions did not score so well comparatively. Interestingly, Apisarnthanarak et al. reported that 44% and 31% of inappropriate antifungal use was detected in medical departments and ICUs respectively, and 15% in surgical departments [4].

The primary limitations of our study include the selection of physicians that more frequently prescribe antifungal agents and the hospitals for the questionnaire. This means that the results could be worse if all the physicians of all institutions had been offered to participate. We also included a heterogeneous group of physicians who are mainly involved in different fields of IFI infection and only addressed knowledge in Candida and Aspergillus invasive infections. Finally, although the response rate to different questions was high (85-100%), potential bias could result due to non-response to specific questions.

Given the findings of our study assessing the basal knowledge of antifungal prescribers in 5 European institutions, operative attempts to ameliorate the inappropriate use of medications should be performed. We believe that these European interventions should be based on: 1) Educational programs focusing on correct drugs and dose administration of antifungals; 2) Improvement in appropriate clinical interpretation of fungal isolates; 3) Establishment of antifungal management programs that incorporate infectious diseases specialists that as previously documented [5,48] are important in order to improve quality of care while optimizing hospital costs and antifungal use.