Increase of Non-albicans Candida Species and Their Antifungal Susceptibility in Intensive Care Unit Patients (Mexico)

In Mexico, little is known about candidemia by non-albicans Candida species, or the antifungal susceptibility of such strains without performing antifungal tests; fluconazole is one of the most used treatments in empirical therapy. In the present study, we included patients from the intensive care unit of one hospital in Mexico (2019–2020) with yeast infection and positive cultures. Yeasts obtained from cultivable isolates were identified using an automated identification instrument and by PCR (nrDNA ITS), and their susceptibilities to six antifungals were characterized across a range of concentrations. Yeast cultures from 105 patients which were suspected etiological agents of primary diagnosis were recovered and identified as mainly non-albicans Candida species (57.2%). The most prevalent was C. glabrata (41.9%), followed by C. albicans, C. krusei, C. parapsilosis, C. tropicalis, and Cryptococcus neoformans. The most common infection site was urine (56%), followed by the bronchial aspirate (30%). Most isolated fungi were susceptible to 5-flucytosine (98%) and amphotericin B. However, C. glabrata, C. krusei, and C. tropicalis demonstrated to be resistant to itraconazole, miconazole, and fluconazole. The present investigation contributes to the knowledge of non-albicans Candida species infections in patients and opens the possibility for a better understanding and management in antifungal empirical therapy in Mexico.


Introduction
In Mexico, little is known about nosocomial fungal infections, their comorbidities, treatment, and antifungal susceptibility profiles [1], and an increase in the incidence of fungal diseases has been documented [2]. The lack of sensitivity of diagnostic tools coupled with the high morbidity and mortality caused by these infections represents a public health problem with unique challenges [3,4].
In hospitalized patients, the main etiological agents causing mycoses belong to the genus Candida [5,6]. Mycoses caused by yeasts, mainly candidemia and in a lesser proportion cryptococcosis, are one of the major causes of morbidity and mortality [7,8]. Candida and Cryptococcus infections are associated with high mortality, with rates of 38-75% and 70%, respectively [9,10]. Yeast infections are related to patients in intensive care units (ICUs), prolonged hospitalization, and comorbidities such as immune suppressed status, chronic renal insufficiency, diabetes, and hypertension [11]. Around 15 species of the genus Candida have been documented to cause invasive infections in humans [12]. Of these, about 90% of infections correspond to Candida albicans, C. glabrata, C. parapsilosis, C. tropicalis, and C. krusei, with C. albicans and C. glabrata being the most prevalent species [13].
Candidiasis often involves multifocal colonization in the human body, including cutaneous, pulmonary, gastrointestinal, and urogenital localization, and dissemination in the  [14]. The empirical treatment for yeast infections includes several antifungal drugs such as polyenes, azoles, and echinocandins, mainly amphotericin B and fluconazole [15]. However, the extensive use of antifungals in ICU patients during therapy and for prophylaxis may cause an increase of infections and drug resistance by C. albicans and nonalbicans Candida species [16]. Several isolates of Candida and Cryptococcus have been described as resistant to conventional antifungals such as polyenes (amphotericin B), azoles (fluconazole and itraconazole), and echinocandins (caspofungin) in countries such as Spain and France [7,17]. Concerningly, fluconazole is one of the most used antifungals in empirical therapy in hospitals in Mexico against disseminated and localized infection by Candida or any fungi, regardless of the origin or type of mycosis [2]. The aim of the present study was to screen Candida infections in ICU patients at the Hospital General de Querétaro (HGQ) over a period of 1 year, to identify cultivable isolates, and to characterize their resistance to the six most used antifungals in Mexico.

Sample Collection and Identification
The data were collected in a public secondary referral hospital in Mexico. The HGQ covers different medical and surgical specialties and has an ICU; it is an 85-bed public hospital serving a population of 1,405,992 inhabitants in the metropolitan area of Querétaro City, Mexico. A prospective study was aimed on Candida infection in ICU. The authors confirm that the study adhered to the policies of the World Medical Association Declaration of Helsinki and the Ethical and Investigation Committee of Servicios de Salud del Estado de Querétaro which reviewed and approved the implementation of this study (09-28-2019/11110/UNIVER-SIDAD AUTONOMA DE QUERETARO) in the HGQ, and that Official Mexican Standard NOM-024-SSA3-2012 was followed.
Patients admitted to ICU for > 24 h that required intensive post-operative monitoring or underwent therapy for one or more acute organ system failures from January 2019 through January 2020 were included. Clinical and demographic data were recorded prospectively and analyzed (age, gender, comorbidities, and risk factors). Patients with positive cultures obtained from different clinical samples (bloodstream, bronchial aspirate, catheter tip, cerebrospinal fluid, peritoneal fluid, skin and soft tissue, and urine) were studied. All samples were collected and cultured on Sabouraud Dextrose Agar (SDA) (Bioxon®, Mexico) supplemented with chloramphenicol (50 mg/L) (Sigma®) to prevent Bacteria growth, for 24-72 h at 37 °C. Positive cultures were cultured in CHROMagar Candida™ (BD®, US) for 24 h at 37 °C. Macroscopic characteristics of the cultures were observed directly on solid medium. Axenic isolates were identified using the Vitek 2k system (bioMérieux, Lyon, France), based on the manufacturer's instructions [18], and by PCR based on method described by Calvillo-Medina et al. [19]. Briefly, DNA was obtained using cetyl-trimethylammonium bromide (CTAB)-based method (Sigma®), and the fungal barcode nrDNA internal transcribed spacer (ITS) sequence was amplified using primers ITS1f and ITS4r [20] in a GeneAmp PCR System 9700 (Thermo Fisher Scientific). PCR products were sent to Macrogen Inc. (Seoul, Korea) for standard-seq single DNA sequencing (Sanger sequencing). Sequences were compared via BLAST with ITS of Candida species and genera-related sequences from the NCBI database (http:// www. ncbi. nlm. nih. gov/ genba nk/).

Antifungal Susceptibility Assay
A panel of six antifungals (Sigma Aldrich®, USA), commonly used in systemic or localized infections therapy in Mexico, were used for susceptibility testing. Concentration ranges for the assays were for amphotericin B (AMB) 1 to 8 mg/mL, 5-flucytosine (FCZ) 1 to 32 mg/mL, fluconazole (FLC) 8 to 64 mg/mL, itraconazole (ITR) 0.5 to 4 mg/ mL, ketoconazole (KTC) 0.5 to 4 mg/mL, and miconazole (MCN) 0.5 to 8 mg/mL. Each antifungal stock solution was prepared in sterile dimethyl sulfoxide (DMSO) and sterilized by 0.22 μm filtration (Millipore®, USA) and stored at -70 °C until their use [21]. Isolates were screened to determine the minimum inhibitory concentration (MIC) in vitro, defined as the minimum concentration of antifungal that inhibited the growth of isolates after the incubation period, based on broth microdilution M27-A3 Clinical and Laboratory Standards Institute (CLSI) [22,23].
Each isolate grew for 24 h in SDA, inoculum concentration was calculated, adjusted to 1 × 10 3 to 5 × 10 3 CFU/ mL in RPMI 1640 medium (Sigma®, Germany), and 100 μL was placed into each well with the antifungal agent. Antifungal susceptibility assays were done in quadruplicate. The microplates were incubated at 35 °C and read after 48 h. Plates were measured with a microplate spectrophotometer (Multiskan Ascent Thermo Labsystems, USA) at an optical density (OD) of 595 nm and were compared with drug-free controls. Strains from the American Type Culture Collection (ATCC) were used as quality control, Candida krusei (ATCC 6258), and C. parapsilosis (ATCC 22019). MIC breakpoints were interpreted according to the CLSI M27-S3. [22,24].

Statistical Analysis
Differences between continuous variables were evaluated statistically by means of two-tailed Student's t test for differences in means and percentages. Differences between categorical variables were evaluated by chi-square test. Values of p < 0.01(***) were considered statistically significant. Graphs and tests were performed with GraphPad Prism 7.0 (San Diego, CA, USA).

Demographic Data of Patients
The study included 105 patients that showed a positive yeast culture. The age range at which the fungal infections presented the highest incidence was 41-60 years, and the average age was 54.5 years (± 18.13
According to in vitro results and after 48 h of growth, the two most prevalent fungi, C. glabrata and C. albicans, were both susceptible (with no growth) to AMB (MIC < 1 mg/ mL) and FCZ (MIC < 1 mg/mL) and resistant to ITR (MIC > 4 mg/mL), FLC (MIC > 64 mg/mL), and MCN (MIC >8 mg/mL). Isolates of C. krusei and C. parapsilosis showed susceptibility to all antifungals tested, but C. tropicalis isolates were susceptible to FCZ (100%), FLC (78%), AMB (67%), and KTC (67%) and only moderately resistant to ITR > MCN. C. neoformans isolates presented sensitivity to all the antifungals tested, except for FCZ (50%). However, it is important to recall that both cases of cryptococcosis were fatal. Comparison of all isolate controls against antifungal treatments showed a significant difference (p < 0.01) in all experiments ( Table 2).

Discussion
There is a lack of data about Candida infections in Mexico despite the frequency and severity of the infections that they cause. Most information regarding yeast infections is from Mexico City, and a few investigations have been conducted for other parts of the country [2,25]. Our study addresses this knowledge gap, by characterizing 105 fungal isolates from patients in HGQ in Querétaro, Mexico. These samples were isolated from different human sites across the span of 1 year, and each was tested for antifungal resistance. Here, the results showed a correlation between risk factors (diabetes mellitus, hypertension, renal failure, and obesity) and mycoses. These comorbidities predispose patients to fungal infection specifically candidiasis and cryptococcosis [2,25]. The mortality rate found in HGQ was 25.8%, lower than the Brazilian rate which ranges from 54 to 72.2%, and similar to that found in China (23.3%) [26].
We described the distribution of the most frequent species of the genus Candida: C. albicans, C. glabrata, C. tropicalis, C. parapsilosis, and C. krusei [1,12] and the presence of C. neoformans (1.9%). Our report highlights the predominance of non-albicans Candida species (57.2%), with C. glabrata being the most frequent which may be a cause of concern in the management of candidiasis. According to de la Torres-Saldaña et al. [27] and Reyes-Montes et al. [2], C. albicans is the most frequent etiological agent isolated from Mexican patients. However, González et al. [1] found a prevalence of non-albicans Candida species in hospitals in Monterrey, Mexico, similar to that described here. In tropical countries such as Brazil and India, common causes of nosocomial candidemia are C. parapsilosis and C. tropicalis, respectively  [12,14]. According to Manzano-Gayosso et al. [28], the etiology of Candida infections in Mexico has changed in the last 20 years, with non-albicans Candida species being more prevalent.
Regardless of Candida infection etiology, the antifungal most used in public hospitals in Mexico is FLC [2] which has been broadly used in empirical therapy against disseminated mycoses [29]. However, indiscriminate use of FLC (for example prophylaxis) can produce a great economic and ecological impact [2,18]. In addition to generating resistant yeast, excessive use of FLC has caused its efficacy rate to decrease. FLC offers only variable protection against non-albicans Candida species [12]. Based on differences in susceptibility to FLC, Jenks et al. [30] recommended the implementation of newer triazoles with antifungal activity such as voriconazole [31]. Therefore, to minimize the further development of resistance by Candida and Cryptococcus, appropriate antifungal treatments should be carefully selected, started as early as possible, and only used in patients with proven infection [12,32].
General resistance to azoles, and specific resistance to FLC, have been reported in Candida and Cryptococcus [2,18]. Moreover, the data generated from this study indicate that ITR, FLC, and MCN were the less effective antifungals, which contrasts with the use of FLC in therapy. The findings for C. glabrata resistance to azoles shown in this study are similar to those previously published [1,33]. In contrast to Corzo-Leon et al. [18] and González et al. [1], we found high resistance of C. tropicalis and C. albicans to azoles and AMB. Additionally, C. parapsilosis was susceptible to all antifungal agents tested. According to González et al. [1], Pfaller and Diekema [30], and Corzo-Leon et al. [18], and based on our results, the use of FLC should not be continued for prophylaxis or in patients with recurrent yeast infections.
Our study had several limitations. First, some epidemiological data concerning (previous infection, transfers between hospitals, or previous exposure to antifungals) were not available. Second, the isolates were only collected and analyzed for 1 year. The duration of the study could be extended to provide a more comprehensive analysis of the state of fungal infections and treatments. Third, we did not collect samples from the hospital environment or health care workers, which could provide information to better understand the source and spread of infections. Finally, the decision on which treatments were administered was made by physicians at HGQ and was not based on the results of our antifungal assays. In conclusion, it is important to improve the identification and antifungal test (identification by PCR or susceptibility to echinocandins) of non-albicans Candida species in ICU patients in Mexico to ensure specific identification, diagnosis, and treatment. Selection of appropriate therapeutic strategies should be based on antifungal susceptibility patterns and control measures for risk factors. Nevertheless, the use of standardized technology in diagnostics and research remains a major global challenge, especially in countries such as Mexico. Improving identification and antifungal testing strategies will allow a better understanding of the epidemiology and susceptibility Table 2 In vitro susceptibilities of Candida albicans and non-albicans Candida species against six antifungal agents from clinic isolates In bold are the most effective antifungals (susceptible culture) and underlined the less effective (resistant yeast) used in vitro experiments for each yeast. At different concentration at 37 °C for 48 h. The concentrations of each antifungal in susceptible column are AMB < 1 mg/mL; FCZ < 1 mg/mL, FLC < 8 mg/mL, ITR < 0.5 mg/mL, KTC < 0.5 mg/mL and MCN < 0.5 mg/mL. In resistance column are AMB > 8 mg/ mL; FCZ > 32 mg/mL, FLC > 64 mg/mL, ITR > 4 mg/mL, KTC 4 > mg/mL and MCN 8 > mg/mL