Background

Candida species are the most common cause of fungal infections, leading to a range of life-threatening invasive diseases such as blood stream candidiasis to non-life-threatening mucocutaneous candidiasis such as genitourinary candidiasis, vulvovaginal candidiasis, and oropharyngeal candidiasis [1]. They are also an important cause of superficial mycosis such as onychomycosis. Among fungal infections, invasive candidiasis is commonly associated with high morbidity and mortality rate. For example, Candida species are among the top ten pathogens causing bloodstream infections [1], resulting in increased death rate, patient hospitalization, and healthcare costs [2]. Mucocutaneous candidiasis is one of the indirect signs for cell-mediated immunodeficiency and estimated to have more than 90% positive predictive value for invasive candidiasis [3].

Until recently, C. albicans was recognized as the commonest species causing most of the cases of candidiasis. However, in the last few decades, several studies reported that there has been a progressive shift from a predominance of C. albicans to non-albicans Candida species (NAC) such as C. tropicalis, C. glabrata and C. krusei [4,5,6]. NAC species have been reported to be a major cause of fungal opportunistic infection [7, 8]. An increase in opportunistic fungal infections is the result of an increase in the number of immune-compromised patients. Excessive use of broad-spectrum antibiotics, metabolic disorders, and the emergence of AIDS are among the various contributing factors for an increase in opportunistic fungal infections [9,10,11,12].

Development of resistance to azoles, the treatment of choice for fungal infections, mainly by NAC species, differences in drug susceptibility profile among yeast isolates, and frequent isolation of emerging yeasts (i.e., NAC species) in clinical samples initiated the use of accurate species identification and in vitro susceptibility testing methods [13]. It has been recognized that many phenotypic automated and molecular yeast identification techniques as well as standardized methods for antifungal susceptibility testing are available in many laboratories. Unfortunately, in Ethiopia, even the simplest yeast diagnostic methods such as chromogenic medium and carbohydrate assimilation as well as simple antifungal susceptibility testing procedures are lacking. Consequently, the spectrum of yeasts and their antifungal susceptibility profile are poorly known and treatment of fungal disease has remained empirical. Against this back ground, the aim of this study is to determine the spectrum and antifungal susceptibility profile of yeasts particularly of Candida species recovered from different clinical samples from patients referred to Arsho Advanced Medical Laboratory for routine tests. Identification and sensitivity testing were performed by using the VITEK 2 compact system. The VITEK 2 compact system (bioMérieux) is a fully automated system that is accurate in the identification of yeasts and evaluating their drug susceptibility profile simultaneously [14, 15].

Methods

Study population and period

A study on the composition of yeast species and antifungal susceptibility profile of yeasts recovered in culture from different clinical samples was carried out from January 2018 to September 2018 at Addis Ababa, Ethiopia.

Sample collection

A total of 776 clinical samples including throat swab (119), nail scrapings (154), sputum (142), corneal scrapings (146), and vaginal swabs (215) were collected from patients referred to Arsho advanced medical laboratory according to standard procedures related to each sample. One clinical sample per patient was collected. All clinical samples were inoculated on to Sabouraud dextrose agar (Oxoid, Basingstoke, UK) to which 50 μg/1 ml gentamicin is incorporated. Inoculated plates were incubated at a temperature of 370 C for at least 72 h aerobically. Then yeast isolates were transferred to Brain heart infusion broth (Oxoid, Basingstoke, UK) containing 10% glycerol and transferred to the Department of Medical Laboratory Technology, College of health Sciences, Addis Ababa University and stored at minus 80o deep freezer until use.

Identification and antifungal susceptibility testing

The viability of yeast cultures that have been stored at minus 80o deep freezer was checked by inoculating them on to Sabouraud dextrose agar (Oxoid, Basingstoke, UK). Identification of yeast cultures and their in vitro antifungal susceptibility profile were determined by the automated VITEK 2 compact system (bioMérieux, France) using YST-21343 and AST-YS07 cards following the instruction of the manufacturer at the Ethiopian Public Health Institute situated in the same campus with the Department of Medical Laboratory Science. Briefly, inoculum suspensions for the VITEK 2 were prepared in sterile saline at turbidity equal to 2.0 McFarland standards, as measured using a DensiChek instrument (bioMérieux). The YST-21343 and AST-YS07 cards were automatically filled with the prepared culture suspension, sealed, and incubated by the VITEK 2 instrument. The cards were incubated at 35.5 °C for 18 h, and data were collected at 15-min intervals during the entire incubation period and final identification and break point minimum inhibitory concentration (MIC) values used to categories Candida species, susceptible, intermediate or resistant to each drug were investigated using version 7.0 software, an advanced expert system (AES) designed to evaluate the results produced by the VITEK 2 system. Identification of each isolate down to a species level assigned by the VITEK 2 compact system as excellent, very good, good, acceptable, or low- discrimination was taken as correct identification.

Quality assurance

Clinical samples were inoculated after the performance of culture media and their sterility was checked following standard procedure. A standard strain of C. albicans (ATCC 10231) was used to evaluate the performance of the VITEK machine.

Data analysis and interpretation

All data from the investigation were coded and analyzed using SPSS version 20. Descriptive statistics such as frequency and percentage of Candida species were calculated.

Ethical clearance

All ethical deliberations and responsibilities were appropriately addressed, and the study was conducted after the approval of the Internal Review Board (IRB) of the Department of Medical Laboratory Sciences (DRERC/323/18/MLS) and after permission letter was obtained from Ethiopian Public Health Institute and Arsho Advanced Medical Laboratory private limited company. Informed written consent was obtained from the study participant ahead of data collection. Each participant was given the right to refuse or to participate in the study and even to withdraw him/her from the study at any time during the course of the study. All information recorded from the study subjects were kept confidentially. An assent form was complted and signed by a family member and/or adult guardian for patients under the age of 16 years.

Results

A total of 209 yeasts were isolated from 776 different clinical samples (Table 1). Among the isolates, 84(40.2%) were recovered from vaginal swab while 52(24.9) were isolated from the oropharyngeal swab. The remaining 42(20.1%), 19(9.1%) and 12(5.7%) were isolated from sputum, corneal, and nail scrapings, respectively. The distribution of Candida with respect to clinical sample was variable where C. albicans, C. famata, and C. kefyr recovered in all clinical specimens. Of 209 yeasts recovered, 104(49.8%) were C. albicans, 90 (43.1%) were NAC species, and 15 (7.2%) were yeasts other than Candida species. Among NAC species, C. krusei 15.6%, C. famata 14.4%, C. rugosa 11.1%, and C. lusitaniae 10.0% were the commonest isolates. Out of yeasts other than Candida species Cryptococcus laurentii represented 66.7% of this group of yeasts.

Table 1 Distribution of yeast isolates per clinical samples

The proportion of C. albicans to NAC species and the first three predominant Candida species in this study in comparison with earlier studies between and within countries are presented in Table 2. The three predominant Candida species in our study in their descending order were C. albicans, C. krusei, and C. famata.

Table 2 The proportion of C. albicans to NAC species and the first 3 dominant Candida species isolated in the current study compared to earlier studies between and with countries

The antifungal profile of Candida isolates

The in vitro antifungal susceptibility profile of Candida species to the five antifungal drugs expressed as MICs in μg/ml is shown in Table 3. Among 194 C. albicans and NAC isolates, the MIC values of C. famata to all drugs tested and C. ciferrii to echinocandins were not determined by the VITEK. Regardless of Candida species identified, 85.6% were susceptible while 10% were resistant to fluconazole. The remaining 3.9% were intermediate to the agent. Hundred percent of C. krusie, 20% C. rugosa, 25% of C. pelliculosa, 17% C. lipolytica and 17% C. ciferrii demonstrated resistant to the agent. Furthermore, 2% C. albicans, 28.6% C. dubliniensis, 10% of C. rugosa, 17% C.lipolytica, and 17% C. ciferrii were intermediate to fluconazole. Voriconazole showed the greatest antifungal activity against Candida isolates in which all Candida isolates (99.4%) exhibiting susceptibility while one (0.6%) isolate of C. ciferrii was found out to be resistant. The susceptibility and the resistance rate of Candida isolates to both caspofungin and micafungin were the same being 96%). Four isolates of C. albicans, two isolates of C. rugosa and one isolate of C. lipolytica demonstrated resistance to the drugs. Similarly, 86.2% of yeast isolates were susceptible to flucytosine, 6.6% were resistant while 7.2% were intermediate to it. One isolate of C. albicans and 11 isolates of C. krusei were resistant to the drug.

Table 3 Over all percentage antifungal drug profile of Candida species

Break points of different antifungal drugs against Candida Isolates are shown in Table 4. The MIC values of fluconazole, voriconazole, caspofungin, micafungin and, flucytosine against C. albicans were variable. Relatively, higher fluconazole MICs were found in NAC species than C. albicans. All Isolates of C. krusei were resistant to fluconazole at a concentration of ≥16 μg/ml. The potency of micafungin was better than that of caspofungin. Most isolates were susceptible to micafungin at the lowest dilution drug tested (0.06 μg/ml).

Table 4 Break points of different antifungal drugs against Candida Isolates

Discussion

Rapid and accurate identification of Candida species down to the species level is of great importance for the selection of appropriate antifungal agents and for patient management. In the present study, 18 different species of yeasts compromising of 15 Candida species and three yeasts other than Candida species were identified. While Candida albicans was the most frequently isolated yeast in the present study, its prevalence was lower than a study conducted in Northwest Ethiopia [16] and higher than a study conducted in central Ethiopia [17]. Candida albicans as the most frequently isolated species was also reported by Sasso et al [30], Mnge et al [31], and Zeng et al [32]. C. parapsilosis as the most frequently isolated yeast was depicted in a study conducted by Sahal and Bilkay [33].

A preponderance of C. albicans compared with NAC species varies between and with countries. A preponderance of C. albicans was observed in our study. Our finding was consistent with previous studies conducted elsewhere [16,17,18, 21,22,23, 25,26,27]. However, a preponderance of NAC species were demonstrated by many earlier studies [19, 20, 24, 28, 29]. While C. albicans is still the predominant species in Ethiopian studies, the occurrence of NAC species in the current study was higher than those studies [16, 17] underlining the shifting trend of Candida infections towards NAC species. A shift of Candida albicans towards non-albicans Candida species observed in the present study could be due to an improved detection rate of non-albicans Candida species or a true prevalence change. Our result is in line with a study conducted by Ghazi et al [34] that demonstrated a shift of Candida albicans towards non-albicans Candida species in the Middle East and North Africa. In the present study, C. krusei was the 2nd dominant species. Our result was in line with that of Bitew and Abebaw [17] and Mohndas and Ballal [19], but in contradiction with many other studies where C. glabrata or C. tropicalis was reported as a 2nd predominant NAC species [16, 18,19,20, 22, 23, 25,26,27,28,29]. A choice of fluconazole as a first line drug for the management of candidiasis in Ethiopia [35] may be a risk factor for a higher isolation rate of C. krusei in this study. Treatment of fungal infections by azole antifungals such as fluconazole was suggested to enhance the selection of resistant Candida species by shifting infection and/or colonization to more intrinsically resistant species chiefly of C. krusei or C. glabrata [36]. The lower isolation rate of C. glabrata compared to C. krusei which is also intrinsically resistant to fluconazole in the current study is unclear. The geographical difference can be a contributing factor for differences in the distribution of NAC species is suggested by Falags et al [37]. This can be demonstrated by the fact that the prevalence of C. glabrata is high in America, North and Central Europe while C. tropicalis is commonly isolated in South America and Asia [37]. C. famata was the third predominant Candida species in the present study. Similar results were obtained by many previous studies [38,39,40]. These studies have demonstrated that strains of C. guilliermondii, C. lusitaniae, C. fermentati, C. intermedia, and C. palmioleophila have been reported as C. famata by phenotypic methods such as the VITEK 2 compact system, and such drawback of the VITEK 2 could be a possible explanation for the higher isolation rate of C. famata in the current study. Therefore, Candida species identified as C. famata by phenotypic methods should be confirmed by molecular methods.

In the present study, the drug sensitivity patterns of all yeast isolates were tested against five antifungal drugs. The MIC for C. famata to all the drugs tested and C. ciferrii for echinocandins were not provided by the VITEK. This could probably be due to the fact that the MIC values of the drugs against the isolates are not present in the system database and/or the VITEK 2 system may take a longer incubation period to determine MIC endpoints of the drugs against the isolates.

In our study, 85.6, 10.5, and 3.9% of Candida species were susceptibly, resistant and, intermediated to fluconazole, respectively. In view of the overall high level of susceptibility of Candida species to the antifungal demonstrated that the drug is still considered to be an active drug against Candida species, regardless of the indiscriminate use of the agent for therapy and prevention in Ethiopia [35]. However, C. krusei recognized to be a naturally non-susceptible to fluconazole [41] was found out to be 100% resistant to it. Our result was compatible with our previous study [17] and another study [42] but in disagreement with the result of Elfeky et al [18] who reported that as high as 60% of the species are susceptible to the agent. Based on our in vitro drug susceptibility testing result, we suggest a search for alternative agents is essential when treating fungal infections caused by C. krusei. Voriconazole was the most active drug against yeast isolates among all drugs tested. All yeast isolates were 100% susceptible to the antifungal except one species of C. ciferrii which was resistant to the agent giving the overall percentage susceptibility of 99.4%. While inhibition of ergosterol synthesis is a common mode of action to all azole antifungal drugs [42) susceptibility difference between fluconazole and voriconazole against C. krusei in our study is unclear. On the other hand, contrary to the notion that C. glabrata is intrinsically non-susceptible to fluconazole, all isolates in our study were sensitive to all azole drugs tested. More or less similar results were reported by many studies [17, 27, 43].

In our previous study [17], the in vitro susceptibility of all yeast isolates were 100% to both micafungin and caspofungin. However, in this study a resistance rate of 4% to each drug was noted. C. rugosa (20%), C. lipolytica (16.7%), and C. albicans (3.8%) were the isolates where resistance strains to the two drugs were observed. While the number of isolates categorized as susceptible, intermediate, and resistance to both drugs were the same, micafungin was more potent than caspofunginn supporting the work of Ostrosky-Zeichne et al [44]. Further, while the mode of action of the two antifungals is the same that is preventing fungal cell wall synthesis by inhibiting the enzyme that synthesizes β-glucan [36], a discrepancy in potency observed in our study is not clear.

Of 181 Candida isolates in which MIC values provided by the machine 86.2, 6.6% and, 7.1% were susceptible, resistant, and intermediate to flucytosine, respectively. Surprisingly, about 78.6% of C. krusei was resistant to the drug. Percentage drug resistance noted in this study was more than two fold than from that of our previous study [17]. Regarding the drug resistance profile of yeast other than Candida species no MIC values were obtained for both C. laurentii and T. mucoid. C. neoformans was 100% susceptible to fluconazole and flycytosine, but MIC values were not obtained against the remaining drugs tested.

Although attempts were made to collect demographic data and history of each patient from request forms filled out by physicians, unfortunately, the recording of these parameters in the test request form was not consistent. Consequently, demographic data and the clinical history of each patient were not documented for this study. The mechanisms responsible for the acquisition of antifungal resistance in our clinical isolates were not determined due to resource shortage and this can be considerd as another limitation of this study.

Conclusions

The present study demonstrated the distribution of Candida species in different clinical specimens where the isolation rate of non-albicans Candida species was comparable to Candida albicans. The high resistance rate of C. krusei to fluconazole and flucytosine may demonstrate that the treatment of candidiasis empirically is questionable.