Introduction

Invasive candidiasis (IC) is an important cause of nosocomial infections, related to different risk factors such as previous antimicrobial therapy, immunodeficiency, parenteral nutrition, presence of catheters and intensive care unit permanence. [1, 2]. Although Candida albicans is still the main pathogen, non-albicans Candida species are increasing in IC patients [3]. This change has been attributed to the more frequent use of azole antifungals and invasive procedures [3]. In addition, there are geographical differences in the epidemiology of Candida infection [4]. In fact, Candida glabrata is the species more frequent after C. albicans in North America [5]. On the contrary, Candida parapsilosis or Candida tropicalis is relatively more common in Europe, Australia, Latin America and Asia [6,7,8]. In India, C. tropicalis causes are seen in more cases of candidemia than C. albicans [9]. Furthermore, different species such as Candida guilliermondii and Candida rugosa are diffusing [10, 11].

Compared to other bloodstream infections, IC appears to be associated with a particularly high rate of mortality, due to mainly, in the form of a delay in diagnosis or even failure of the antifungal therapy.

Different papers have shown that the rate of resistance to fluconazole ranges from 2.5 to 9% in Candida spp. isolated from blood [12, 13]. Most Candida species are considered good targets for the three echinocandins (anidulafungin, caspofungin and micafungin) that are used as first-line agents for the treatment of fungemia; however, what has been found is the increasing use of these drugs determines the emergence of resistance in Candida and non-Candida species [14].

An effective infection control is required due to higher incidence of IC, increased mortality and the growing prevalence of resistance to fluconazole.

The aim of this study was to conduct a seven-year retrospective analysis in order to analyze the incidence and species distribution in patients with fungemia in a large hospital located in Brescia, Italy. The rates of antifungal resistance were determined according to the newly revised CLSI clinical breakpoints (CBPs) or, in the absence of CBPs, according to epidemiological cutoff values (ECVs) for nine antifungal agents [15].

Materials and Methods

Collection of Isolates

We conducted this study in a large university hospital, Spedali Civili, in Brescia, North of Italy. We collected isolates from patients with candidemia from January 2009 to December 2015. An episode of candidemia was defined as Candida infection involving at least one blood culture. Only the first episode of fungemia was reported per patient with recurrent or subsequent episodes of infection. Patients whose cultures grew >1 documented species of Candida were excluded from the analysis. The study did not require the approval by the Institutional ethics committee due to the descriptive nature.

Yeast Identification

Yeasts were isolated from patient blood cultures collected during normal routine and processed by the BACTEC (BD Diagnostic Systems, Sparks, MD) system. Identification of all species was performed using VITEK 2 YST cards from bioMérieux (Marcy, l’ Etoile, France) or since 2012, by matrix-assisted laser desorption ionization–time-of-flight mass spectrometry (MALDI-TOF).

Antifungal Susceptibility Testing

The in vitro susceptibility to antifungal drugs was performed by using Sensititre Yeast One (YO-08 from 2009 to 2012 and YO-10 from 2013 to 2015) in accordance with the manufacturer’s instructions. Because the ranges of amphotericin B, flucytosine, fluconazole and itraconazole were different from those of the previous version (YO-08), MIC values of 0.008–0.12 μg/mL for amphotericin B and of 0.03–0.12 µg/mL for fluconazole were reported as ≤0.12 g/mL; MIC values of 0.03–0.06 μg/mL for flucytosine were reported as ≤0.06 g/mL; and MIC values of 0.008–0.015 μg/mL for itraconazole were reported as ≤0.015 μg/mL.

Data Analysis

MIC values determined by the Sensititre Yeast One YO-08 and YO-10 were interpreted according to current CLSI species specific CBPs [15]; if no CBPs were defined, ECVs were used.

The CLSI resistance breakpoint for fluconazole was defined as an MIC of >4 g/mL against C. albicans, C. parapsilosis and C. tropicalis, and of >32 μg/mL against C. glabrata; for voriconazole, as an MIC of >0.5 μg/mL against C. albicans, C. parapsilosis, and C. tropicalis, and of >1 μg/mL against C. krusei. The CLSI resistance breakpoint for echinocandins was indicated as an MIC of >0.5 μg/mL against C. albicans and C. tropicalis and of 4 g/mL against C. parapsilosis; for both anidulafungin and caspofungin, an MIC of >0.25 μg/mL was defined against C. glabrata and an MIC of >0.12 μg/mL for micafungin. The ECV of >0.5 μg/mL was used to identify non-WT isolates of C. glabrata to voriconazole; ECVs of >0.06, >0.25, >0.12, and >2 μg/mL were used to identify non-WT isolates of C. albicans, C. parapsilosis, C. tropicalis and C. glabrata, respectively, to posaconazole [16]. ECVs were also used to identify non-WT isolates of C. albicans, C. parapsilosis, C. tropicalis and C. glabrata to amphotericin B (>2 μg/mL for all) and flucytosine (>0.5 μg/mL for all) [16].

Results and Discussion

A total of 196 distinct episodes of candidemia (192 adults and 4 adolescents <18 years of age) were identified during the study period. Of these, 113 cases (58%) comprised of males and 83 (42%) where females. The median age of patients was 81 years (ranging from 17 to 96 years).

The average incidence of candidemia was 0.365 per 1000 admissions, which is comparable to that reported for centers in Denmark (0.41 case per 1000 admissions) [17], but lower than that of Israel [18], China [19], Brazil [20], Portugal [21] and even from that of different Italian regions [22, 23]. These differences in candidemia rates may depend on various factors such as differences in demographic characteristics, variations in health care practice, long antibacterial therapies and on the local resistance epidemiology. During the years analyzed, this incidence rate varied, increasing from 0.23 cases per 1000 admissions in the year 2009 to 0.55 cases per 1000 admissions in 2015 (Fig. 1). This might be due to an increased exposure to different risk factors such as prolonged antibiotic therapy, the use of urinary catheters, parenteral nutrition and central venous catheters.

Fig. 1
figure 1

Patients with Candida bloodstream infections (bars) and incidence rate (line) observed during a seven years period

Full size image

The four most prevalent species were Candida albicans (60%), Candida parapsilosis (15%), Candida glabrata (12%), and Candida tropicalis (6%), while other species were relatively rare: Candida famata (1.5%), Candida lusitaniae (1.5%), Candida lipolytica (1%), Candida ciferrii (0.5%), Candida dubliniensis (0.5%) and Candida guilliermondii (0.5%). The most affected age group was 61-80 years and these where mostly isolated from intensive care units, medical and surgical wards (Table 1). This result is not surprising, due to the fact that most of the patients admitted to these wards are in immunosuppressive conditions, and are often submitted to antibiotic therapies, predisposing them to an increasing rate of fungal infections.

Table 1 Distribution and characteristics of patients with bloodstream yeast infections
Full size table

Tables 2 and 3 show all the data for susceptibility testing using CLSI guidelines.

Table 2 Activities of nine antifungal agents against the yeast isolates by Sensititre Yeast One method
Full size table
Table 3 Distribution of MIC values according to the Clinical and Laboratory Standards Institute (CLSI) guidelines
Full size table

All isolates studied showed a WT phenotype to amphotericin B. The MIC50 and MIC90 values of amphotericin B for all the species were either 0.25 or 0.5 mg/L (Table 2). All isolates were inhibited at drug concentrations of 0.5 mg/L.

This full susceptibility to amphotericin B is important because, although this agent is not a first-line drug for the treatment of invasive candidiasis and candidemia in most clinical approaches [14], it may be used as alternative therapeutic option or where the isolates exhibit a resistance against azoles or echinocandins.

When we looked at the susceptibility to flucytosine, 10 isolates across C. albicans (2/116 isolates; 1.7%), C. parapsilosis (3/28 isolates; 10.7%), C. glabrata (2/22 isolates; 9%) and C. tropicalis (2/12 isolates; 25%) were found to be not WT to flucytosine.

The most frequently used antifungals systematically and locally are the azoles. Of the azoles used systematically, fluconazole is the most frequently used one in the yeasts.

Among the group of azoles, C. parapsilosis species was susceptible to all four azoles tested. With regard to C. albicans, 3 isolates were resistant to fluconazole and posaconazole (2.8 and 2.6%, respectively). Three isolates (2.5%) were resistant to voriconazole, while two isolates (1.7%) were resistant to itraconazole. With regard to C. glabrata, one isolate (4.5%) was resistant to fluconazole, 2 isolates (8.3%) were resistant to itraconazole, and four isolates (16.6%) were resistant to voriconazole. All isolates were of the WT phenotype for posaconazole. With regard to C. tropicalis, 1 isolate (10%) was resistant to fluconazole and 2 isolates (18%) were resistant to posaconazole. All isolates were of the WT phenotype for itraconazole.

Overall, the frequency of fluconazole resistance in our isolates was 3.5%. The MIC values for fluconazole were between 0.25 and 32 mg/L for the 175 Candida strains analyzed.

Our findings show a low resistance to fluconazole by C. glabrata isolates (4.5%), a percentage comparable to that reported in other Italian studies [23, 24]. Furthermore, we found some isolates of C. glabrata susceptible to fluconazole and resistant to voriconazole, a phenotype already described [25].

Echinocandins are in general active against various Candida and Aspergillus spp. and this explains their large use in clinical practice. In fact, recent European guidelines are recommending echinocandins as the first-line treatment in severe or neutropenic patients, or when there is prior use of azoles or suspected resistance to azoles [26].

In our study, all Candida–non-albicans species were full susceptible to echinocandins. Echinocandin resistance in Candida albicans was low (1/54, 1.8%) for all the three echinocandins and was similar to that reported in Spain [27], however, different from other countries such as the USA where echinocandin resistance is emerging [28].

Table 4 shows the trend in the resistance rate over the period of the study. We observed a stable trend for azole resistance during the study period with the exception of the year 2014 where there was a decline in the rate of resistance.

Table 4 Trends in resistance according to CLSI breakpoints to the antifungal agents studied for Candida and non-Candida species during the study period (2009–2015)
Full size table

This study has several limitations, i.e., the data associated with underlying diseases, risk factors, mortality and previous antifungal therapy were unknown, and therefore, this information could not be analyzed. It is also important to highlight the relatively small size of our samples and underline that more isolates (by extending the period of study) would give more reliable and substantial importance to our results.

Finally, since we included only isolates from a single institution, we may not able to extrapolate them to other hospitals.

However, the rates of fluconazole and echinocandin resistance that we have recorded were similar to those reported from other countries [25,26,27,28,29,30].

Antifungal susceptibility studies performed locally should be a priority in order to monitor continuously the emergence of Candida or non-Candida species with intrinsically reduced susceptibility or resistance.