Fungi in dialysis water and dialysate: occurrence, susceptibility to antifungal agents and biofilm production capacity

The aim of this study was to investigate the occurrence of fungi in dialysis water and dialysate, in addition to evaluating the susceptibility to antifungals and the biofilm production capacity of isolated microorganisms. The samples were collected in three hemodialysis units in Bauru (Brazil), every 15 days (July 2017–June 2018) at post-reverse osmosis, reuse, and dialysate points. The fungi were isolated by spread plate on Sabouraud dextrose agar. Filamentous fungi were phenotypically identified and yeasts were subjected to molecular evaluation of the ITS region. Susceptibility test to antifungals was carried out by the broth microdilution method and biofilm production capacity was evaluated in microtiter plates using crystal violet staining. Fungi were isolated in 52/216 (24.1%) samples, with an average count of 16.3 (10–40) CFU/ mL. Overall, 61 microorganisms were identified, with 54 (88.5%) filamentous fungi and 7 (11.5%) yeasts. The main genera included were Penicillium, Cladosporium, Scedosporium, Rhinocladiella, Fusarium, and Emmonsia. Most isolates showed high values of minimum inhibitory concentration for 5-flucytosine and fluconazole and 35/45 (77.8%) isolates were classified as strong producers of biofilm. In order to increase the safety of the dialysis process, the adoption of control measures and monitoring of fungi in hemodialysis fluids is suggested.


INTRODUCTION
Chronic renal failure (CRF) is currently an important public health problem worldwide. It is considered a complex and multifactorial disease and has high rates of morbidity and mortality in several countries (Ashby et al. 2019).
Hemodialysis is the main form of treatment for CRF and, in general, is performed three times a week, in sessions that last about 3-4 hours, depending on the patient's clinical condition. In this therapy, water is used in the production of dialysate and also in the reuse of dialyzers, so that, during the hemodialysis sessions, patients are exposed, through the dialyzer's semipermeable membrane, to approximately 400 liters of water per week. Therefore, the quality of the water used in this process is extremely important to avoid additional risks to patients' health (Totaro et al. 2017).
There are several guidelines that establish water quality standards for hemodialysis around the globe; however, except for some countries, such as Italy, Germany, Austria, and Sweden, most international standards do not include the counting of filamentous fungi and yeasts among microbiological parameters (Ward 2009). Naturally ubiquitous and known as saprophytes, in recent decades, fungi have arisen as emerging opportunistic pathogens, causing different episodes of nosocomial infections, especially in immunocompromised individuals (Durán Graeff et al. 2017).
Currently, the constant failures in the treatment of fungal infections have called attention to the emergence and spread of resistance to available antifungals, providing an increase in the lethality of these infections (Perlin et al. 2017). In addition, the possible production of biofilms by fungi can represent an important virulence factor for these microorganisms, giving them greater protection against adverse environmental conditions and greater resistance to antimicrobial agents (Sardi et al. 2014).
Several studies have already pointed out that the water distribution system seems to be an important reservoir of filamentous fungi and yeasts, so water is one of the main ways of spreading these microorganisms in hospital environments (Mesquita-Rocha et al. 2013;Arroyo et al. 2020). However, there are still few studies about the occurrence of fungi in hemodialysis fluids. Thus, the aim of this study was to investigate the occurrence, distribution, and diversity of filamentous fungi and yeasts in dialysis water and dialysate samples, in addition to evaluating the susceptibility to antifungal agents and the biofilm production capacity of isolated microorganisms.

Samples
Dialysis water and dialysate samples were collected from three hemodialysis units, called A, B, and C, located in the city of Bauru, State of São Paulo, Brazil. The collections were performed according to the recommendations of the Standard Methods for the Examination of Water and Wastwater (APHA 2012), every 15 days, from July 2017 to June 2018, in the following points of the water treatment and distribution systems: (i) right after the treatment (post-reverse osmosis), (ii) in the reprocessing room of the dialyzers (reuse), and (iii) in the dialysis machine (dialysate)a total of 216 samples.

Fungal culture and identification
The isolation of fungi was performed by spread plate method on Sabouraud dextrose agar (SDA) supplemented with chloramphenicol (0.05 g/L), and incubated at 25°C/15 days (APHA 2012). Filamentous fungi were phenotypically identified at the gender level through the study of their macroscopic characteristics (giant colony technique) and microscopic characteristics (microculture test). Yeasts were analyzed for their macroscopic, microscopic, reproductive, and physiological characteristics, and afterwards, they were subjected to molecular evaluation of the ITS (internal transcribed spacer) region of the rDNA (White et al. 1990), for confirmation of species.

Susceptibility testing for antifungal agents
To determine the minimum inhibitory concentration (MIC), the broth microdilution method was used according to the recommendations of the European Committee on Antibiotic Susceptibility Testing (EUCAST) (Arendrup et al. 2014(Arendrup et al. , 2015. The

Biofilm production test
The biofilm production capacity was evaluated in microtiter plates using the crystal violet staining method, according to Stepanovićet al. (2007), with modifications for yeasts and filamentous fungi.
The filamentous fungi were grown on potato dextrose agar, at 25°C. Then, for each isolate, a suspension in sterile saline with conidia or sporangiospores was prepared, adjusted in a spectrophotometer (Epoch-Biotek, Winooski, VT, USA) at a concentration of 1 Â 10 4 CFU/mL. Then, 20 μL of the fungal suspension and 180 μL of tryptic soy broth (TSB), supplemented with 1% glucose, were incubated in plates at 25°C/48 h. After incubation, the following steps were performed: washing with distilled water, fixing with methanol (150 μL/20 min), staining with violet crystal 2% (150 μL/15 min), washing with distilled water, resuspension with acetic acid 33% (150 μL/30 min), and absorbance reading in a spectrophotometer at 590 nm.
The yeasts were grown on Sabouraud dextrose agar at 36°C/24-48 h and, based on the growth obtained, a suspension was prepared in 3 mL of sterile saline, according to the McFarland standard 4. Twenty μL of the suspension and 180 μL of Sabouraud broth were used, supplemented with 8% glucose, and the plate was incubated at 35°C/24 h. After incubation, the following steps were performed: washing with distilled water, fixing with methanol (110 μL/15 min), staining with violet crystal 0.4% (110 μL/45 min), washing with distilled water, resuspension with acetic acid 33% (200 μL/30 sec), and absorbance reading in a spectrophotometer at 595 nm.
The tests were performed in quadruplicate and a white well (without inoculum, with culture medium only) was used to correct the absorbance value. After correctionconsidering the average optical density of the sample (OD) subtracted from the average of the optical density of the white (ODc)the isolates were classified as strong producers (OD . 4 Â ODc), moderate producers (2 Â ODc , OD 4 Â ODc), weak producers (ODc , OD 2 Â ODc) or non-biofilm producers (OD ODc). As a positive control, the reference strain C. parapsilosis (ATCC 22,019) was used.

Statistical analysis
Statistical analysis was performed with the aid of the software Statistica 10.0 (Statsoft, Tulsa, OK, USA, 2011), considering a significant difference when p , 0.05.

RESULTS
Fungi were isolated in 52/216 (24.1%) samples, with similar percentages of positivity in all units investigated, however, with significant predominance in samples from the reuse room and dialysate. Among the positive samples, the average (minimummaximum) fungal count was 16.3 (10-40) CFU/mL. The highest count values were noted at unit C and at the post-reverse osmosis point, but without significant statistical differences (Table 1). Among the 52 positive samples for fungi, 45 (86.5%) showed growth of filamentous fungi, 4 (7.7%) of yeasts, and 3 (5.8%) of filamentous fungi and yeasts concomitantly. Overall, 61 microorganisms were identified, with 54 (88.5%) filamentous fungi and 7 (11.5%) yeasts. The main genera included were Penicillium, Cladosporium, Scedosporium, Rhinocladiella, Fusarium, and Emmonsia (Figure 1). Table 2 shows the complete distribution of isolated microorganisms in relation to the hemodialysis units and the collection points investigated.
The susceptibility test to antifungal agents revealed several MIC values for all the analyzed drugs (Table 3). Most filamentous isolates showed high MIC values for 5-flucytosine and fluconazole. Some Penicillium isolates also showed low susceptibility to echinocandins and voriconazole. High MICs for caspofungin and miconazole were noted in microorganisms of the genera Scedosporium and Emmonsia, respectively. Fusarium spp. and Rhinocladiella spp. isolates showed low susceptibility to all tested drugs. Even for amphotericin B, some filamentous isolates exposed high MICs, with values ranging between 0.125 and .16 μg/mL. Among the yeasts, Candida orthopsilosis was sensitive to the tested antifungals and Cryptococcus laurentii showed high MIC values for echinocandins, 5-flucytosine, and fluconazole.
The analysis of the biofilm production capacity included 45 filamentous fungi and 3 yeasts (Table 4). All the filamentous fungi produced biofilm, so 35/45 (77.8%) isolates were classified as strong producers and 10/45 (22.2%) as moderate producers, with no statistically significant differences between the analyzed genera. There were also no significant differences between the filamentous isolates in terms of hemodialysis units and the collection points investigated (Table 5). Among the yeasts, 2/3 (66.7%) isolates were not able to produce biofilm and 1/3 (33.3%) showed a weakly producing profile.
The investigation of a possible correlation between susceptibility to antifungal agents and the biofilm production capacity of the filamentous isolates showed a positive correlation (despite the weak intensity: 0.3 , r , 0.5) between the MIC values for micafungin, amphotericin B, and 5-flucytosine and the absorbance values of the biofilm production test (Table 6).

DISCUSSION
Water is essential in dialysis therapy and its physical-chemical and microbiological quality must be guaranteed (Totaro et al. 2017). In the present study, fungi were detected in 24.1% of the dialysis water and dialysate samples from three hemodialysis units in Bauru (Brazil), with an average (minimum-maximum) count of 16.3 (10-40) CFU/mL. A similar study undertaken in Greece found filamentous fungi and yeasts in 95.3% of dialysis water samples and in 90.6% of dialysate samples (Arvanitidou et al. 2000). In Iran, Mahmoudabadi et al. (2011) reported fungi in 27.4% of the analyzed samples, and in Italy, Schiavano et al. (2014) isolated these microorganisms in 12.1% of the samples from eight hemodialysis units, with counts that varied between 1 and 420 CFU/100 mL. In Brazil, other studies described fungal contamination in 66% (Figel et al. 2013) and 89% of the samples, of which, some showed .100 CFU/mL (Pires-Gonçalves et al. 2008). However, it is important to note that due to the diversity of methodologies used, with different isolation techniques, applied volumes, culture medium, and incubation periods, it is not always possible to make direct comparisons between studies.
Some authors have described a higher frequency and concentration of microorganisms in dialysate samples, suggesting that contamination and proliferation occur mainly in dialysis machines (Heidarieh et al. 2016). However, these findings were not so clear in our research, which showed a higher frequency of fungi in the reuse room and in the dialysate, without significant differences related to their concentrations in the different collection points, thus we reinforce the need for intensification of the maintenance, cleaning, and disinfection measures throughout the water treatment and distribution system.
Filamentous fungi were the main fungal contaminants isolated in this study. Similarly, other studies have also shown the predominance of these microorganisms, with emphasis on the genera Cladosporium, Penicillium, Aspergillus, Fusarium, Trichoderma, among others (Mahmoudabadi et al. 2011;Figel et al. 2013;Schiavano et al. 2014). Yeasts, including Candida  The most frequent isolates in our research were members of the genus Penicillium, which have already been described in several cases of invasive fungal infections, with pulmonary, pericardial, hepatosplenic involvement, and others (Geltner et al.  Ramírez et al. 2018). Cladosporium spp. are among the most largely dispersed fungi in the environment and are commonly associated with allergic rhinitis and localized skin and subcutaneous lesions, but may additionally cause deeper and more widespread infections (Sandoval-Denis et al. 2015). The genus Scedosporium has at least ten species that are possibly  pathogenic to humans, capable of causing a wide spectrum of diseases (Chen et al. 2016). Frequently, invasive infections caused by Scedosporium spp. are difficult to treat and are associated with high rates of morbidity and mortality (Bernhardt et al. 2018). Relating to the yeasts, they are certainly the main cause of nosocomial fungal infections. Invasive candidiasis is the most prevalent fungal disease among hospitalized patients and affects approximately 250,000 people, annually, throughout the world, with more than 50,000 reported deaths (Kullberg & Arendrup 2015). Although C. albicans remains the most involved species in pathogenic conditions, in the last decades there has been a significant change in the epidemiological profile of these infections, with an increase in cases caused by non-albicans Candida, including C. orthopsilosis (Pappas et al. 2016;Arastehfar et al. 2019), which was isolated in this study. Cryptococcus laurentii has also expanded as a causative agent of opportunistic infections in humans, with reports of pulmonary infections, meningitis, peritonitis, and fungemias (Smith et al. 2017).
Until now, little has been known about the clinical complications caused by fungal contamination of hemodialysis fluids, because there are few studies in the literature. Some episodes of fungemia by Phialemonium curvatum and C. parapsilosis have been described by some authors, related to inadequate reprocessing of dialyzers and reflux of residual fluids (Rao et al. 2009;Oyong et al. 2014).
In the last years, antifungal resistance has become an eminent global problem (Houšt' et al. 2020). Corroborating with our findings, several studies have pointed out low activity of 5-flucytosine and fluconazole against filamentous fungi (CLSI 2008a;Peng et al. 2019). Commonly, Scedosporium spp. show in vitro resistance to several antifungal agents, including echinocandins, and Fusarium spp. are among the most drug-resistant fungi available (Tortorano et al. 2014). In fact, the emergence of several Fusarium species with a profile of resistance to multiple drugs has become a public health problem, with an unfavorable prognosis and high mortality rates (Taj-Aldeen 2017). Some special attention should be paid to Rhinocladiella isolates which have shown little susceptibility to the antifungal agents used, unlike some studies that reported low MIC values for amphotericin B and azoles in general for these microorganisms (Nucci et al. 2001;Sun et al. 2019). Other studies involving yeasts isolated from hemodialysis fluids have also not shown resistance to antifungals among microorganisms of the C. parapsilosis complex (Pires et al. 2011); however, clinical failures in infections caused by C. orthopsilosis have been reported by some authors (Heslop et al. 2015;Charsizadeh et al. 2018). Cryptococcus laurentii with reduced susceptibility to some drugs has also been described in the literature (Ajesh & Sreejith 2012).
According to some authors, the expansion of antifungal use in hospital environments and in agriculture may be significantly influencing the emergence of more resistant fungi, with elevated MICs (Chowdhary et al. 2013;Jensen 2016).
Currently, the reference documents recommended by EUCAST and the Clinical and Laboratory Standards Institute (CLSI) establish breakpoints for some antifungals for only species of the genera Candida and Aspergillus (CLSI 2008a(CLSI , 2008b(CLSI , 2012 Arendrup et al. 2014Arendrup et al. , 2015. Thus, for most emerging fungal pathogens there are still no defined breakpoints, which makes the interpretation of MIC results quite challenging, because the in vitro reaction does not necessarily represent the clinical response to treatment (Hadrich & Ayadi 2018).
In this study, we showed the capacity of biofilm production by filamentous fungal isolates. Other similar studies have also described in vitro biofilm formations by species of the order Mucorales (Singh et al. 2011) and the genera Scedosporium (Rollin-Pinheiro et al. 2017) and Trichosporon (Montoya et al. 2018). According to our results, Thomaz et al. (2018) reported Candida isolates with low biofilm production capacity. However, Pires et al. (2011), evaluating yeasts from the hydraulic circuit of a hemodialysis unit, found a considerable number of highly producing strains of the C. parapsilosis complex. Similarly, this characteristic has also been demonstrated in other yeasts (Silva et al. 2011;Pannanusorn et al. 2013).
Commonly, the different techniques for in vitro evaluation of the biofilm production capacity of fungal isolates are adapted from methods applied in bacteriological investigations, so they still need to be standardized (Pires et al. 2011;Montoya et al. 2018). Thus, in addition to the origin of the species, differences in experimental conditions (such as pH, culture medium, oxygen availability, among others) may be responsible for different findings in the literature (Pannanusorn et al. 2013). In this study, all parameters were standardized with adequate training for satisfactory fungal growth.
Furthermore, we observed in this study a positive correlation between the MIC values for some antifungals and the absorbance values of the biofilm production test. Likewise, greater antifungal resistance to amphotericin B and fluconazole has been associated with greater capacity for biofilm production by Trichosporon asahii isolates in research conducted by Sun et al. (2012) and Montoya et al. (2018). However, another study did not find this possible correlation (Iturrieta-González et al. 2014), suggesting that such results can be influenced by multiple factors, including the origin of the isolate, fungal species, investigated drug, methodological parameters, and others (Ramage et al. 2012). Therefore, further studies are needed to elucidate these findings, once the treatment/elimination of biofilm is a major clinical and environmental challenge.
Undoubtedly, the production of biofilm can facilitate the proliferation of microorganisms in different points of the water treatment and distribution systems of the hemodialysis units, in addition to providing greater resistance to cleaning and disinfection procedures, because biofilms are less susceptible to antimicrobial agents compared to planktonic cells (Kanamori et al. 2016;Montoya et al. 2018).

CONCLUSIONS
This study showed a wide variety of fungi in dialysis water and dialysate samples from all the investigated hemodialysis units. Considering the pathogenic potential of the isolated fungi, the occurrence of these microorganisms in these samples may represent a risk to the health of patients undergoing hemodialysis, who commonly have a compromised immune system and other comorbidities, which make them more susceptible to opportunistic infections. Although more studies are needed to elucidate the clinical potential and the threshold of fungal contamination in hemodialysis fluids, aiming to increase the safety of the dialysis process, our findings suggest the adoption of control measures and regular monitoring of fungi in dialysis water and dialysate.

DATA AVAILABILITY STATEMENT
All relevant data are included in the paper or its Supplementary Information.