Abstract
Introduction
To evaluate antifungal susceptibility patterns and minimum inhibitory concentrations (MICs) of clinically important yeasts.
Materials and Methods
This laboratory-based observational study, involving prospective collection of yeast isolates and retrospective analysis of anonymized laboratory data, was conducted from March to December 2024 at a tertiary care cancer hospital. A total of 100 non-duplicate yeast isolates were included. Yeast isolates were initially identified using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry in the clinical microbiology laboratory. All isolates were reidentified using polymerase chain reaction–restriction fragment length polymorphism (PCR-RFLP) targeting the internal transcribed spacer region of fungal ribosomal DNA . The accuracy of PCR-RFLP was validated by molecular sequencing of selected isolates, including rare and uncommon species. Antifungal susceptibility testing was performed using the Sensititre YeastOne Antifungal Susceptibility Test system. MIC values were interpreted according to applicable Clinical and Laboratory Standards Institute species-specific clinical breakpoints. MIC50 and MIC90 values were calculated.
Results
Candida albicans (37%) and Candida tropicalis (32%) were the most frequently isolated species. Blood was the predominant specimen source (43%), followed by urine (37%). Echinocandins exhibited the lowest MIC values across most species, indicating consistent in vitro activity. Fluconazole showed the widest MIC range (0.12-256 µg/mL) and the highest MIC90 values, particularly among Candida glabrata and Candida krusei. Azole agents demonstrated strong positive correlations in MIC values, whereas amphotericin B showed weak correlations with other antifungal classes.
Conclusion
Non-albicans Candida species constituted a substantial proportion of yeast isolates and exhibited variable azole susceptibility. These findings highlight the importance of species-level identification and susceptibility-guided antifungal therapy. Ongoing local surveillance and routine susceptibility testing remain essential for optimizing antifungal therapy.
Introduction
Invasive fungal infections are a major cause of morbidity and mortality among hospitalized patients, particularly those with malignancies, immunosuppression, or prolonged hospital stays[1, 2]. Patients undergoing chemotherapy or intensive care are at increased risk of developing opportunistic fungal infections. Among these pathogens, Candida species are the leading cause of invasive fungal infections worldwide.
Over the past two decades, the epidemiology of candidiasis has changed substantially, with an increasing proportion of infections being caused by non-albicans Candida species. These species frequently exhibit reduced susceptibility to commonly used antifungal agents, particularly azoles, which may complicate the selection of appropriate antifungal therapy[3, 4].
Several surveillance studies have demonstrated geographic variation in the distribution of Candida species. However, local data from cancer centers in Pakistan remain limited, particularly with respect to minimum inhibitory concentration (MIC) distributions and cross-class MIC correlations. In several Asian countries, Candida tropicalis has emerged as one of the predominant causes of candidemia[5, 6].
Accurate identification of Candida species is essential because antifungal susceptibility profiles can vary substantially among species. Conventional phenotypic identification methods may not reliably distinguish closely related species, potentially affecting the selection of appropriate antifungal therapy. Molecular techniques, such as polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP), provide rapid and reliable identification of medically important yeast species[7, 8].
Antifungal susceptibility testing is critical for guiding appropriate therapy. The reference broth microdilution method recommended by the Clinical and Laboratory Standards Institute (CLSI) is widely used for determining antifungal susceptibility; however, commercial systems such as the Sensititre YeastOne Antifungal Susceptibility Test offer a more convenient alternative and have demonstrated good correlation with reference methods[9, 10]. Therefore, we aimed to characterize the species distribution, antifungal MIC patterns, and cross-drug MIC correlations among clinical yeast isolates obtained from a tertiary care cancer hospital.
Materials and Methods
Study Design and Sample Size
This laboratory-based observational study involved the prospective collection of yeast isolates and retrospective analysis of anonymized laboratory data. It was conducted from March to December 2024 at a tertiary care cancer hospital. A total of 100 non-duplicate clinical yeast isolates were included during the study period. The sample size was determined based on the feasibility of collection within the defined study period and is acknowledged as a limitation of the study. Ethical approval was obtained from Shaukat Khanum Memorial Trust Review Board (protocol number: IRB-23-46, date: 11.12.2023). Patient consent was waived by the Institutional Review Board because the patient data were analyzed retrospectively and all identifiers were removed before analysis.
Identification of Yeasts
Yeast isolates obtained from clinical specimens, including blood, urine, pus, tissue, and body fluids, were included in the study. All isolates were cultured on Sabouraud dextrose agar and incubated at 35-37 °C for 24-48 hours before identification and antifungal susceptibility testing.
Initial species identification was performed using matrix-assisted laser desorption ionization–time-of-flight mass spectrometry in the clinical microbiology laboratory. All isolates were subsequently reidentified using PCR-RFLP targeting the internal transcribed spacer (ITS) region of fungal ribosomal DNA, a widely accepted universal DNA barcode marker for fungi[7, 8]. PCR amplification was performed using the universal primers ITS1 and ITS4. The amplified products were digested with the restriction enzyme MspI to generate species-specific fragment patterns, which were analyzed by agarose gel electrophoresis and compared with reference profiles for species identification.
Rare and uncommon yeast species, such as Kodamaea ohmeri and Dirkmeia churashimaensis, were confirmed by Sanger sequencing.
Reference strains of Candida albicans (ATCC 10231), Candida parapsilosis (ATCC 90018), Candida glabrata (ATCC 15126), Candida tropicalis (ATCC 1369), and Candida auris (CDC B11903) were used as quality-control strains.
Antifungal Susceptibility Testing
Antifungal susceptibility testing of clinically significant yeasts was performed using the Sensititre™ YeastOne™ system (Thermo Fisher Scientific), which provides standardized, reproducible, and quantitative MIC results for multiple antifungal agents[9]. The antifungal panel included fluconazole, voriconazole, itraconazole, posaconazole, 5-flucytosine, amphotericin B, and the echinocandins caspofungin, micafungin, and anidulafungin.
A broad range of Candida species, including Candida albicans, Candida famata, Candida glabrata, Candida guilliermondii, Candida krusei, Candida orthopsilosis, Candida parapsilosis, Candida pelliculosa, and Candida tropicalis, as well as non-Candida yeasts such as Papiliotrema laurentii, Cryptococcus neoformans, and Dirkmeia churashimaensis, were tested.
A yeast suspension equivalent to 0.5 McFarland was prepared from a pure 24-hour culture. Subsequently, 20 µL of the suspension was transferred into 11 mL of YeastOne broth. Using the Sensititre™ AIM system, 100 µL of the inoculated broth was dispensed into each well of the Sensititre plate. The plates were sealed and incubated at 35 °C in ambient air and were read after 24 hours using the Sensititre™ Vizion system.
Individual MIC values were recorded for each antifungal agent. MIC results obtained using the Sensititre™ YeastOne™ system were interpreted using species-specific CLSI clinical breakpoints outlined in CLSI M60, Performance Standards for Antifungal Susceptibility Testing of Yeasts, with categorical interpretations performed only when established breakpoints were available [10, 11].
CLSI clinical breakpoints have been established primarily for selected Candida species and specific antifungal agents. For non-Candida yeasts and organism–drug combinations for which validated breakpoints were unavailable, MIC values were reported descriptively without categorical interpretation. Where applicable, epidemiological cutoff values were considered to distinguish wild-type from non-wild-type isolates.
Statistical Analysis
MIC values were log2-transformed before analysis. MIC50, MIC90, and geometric mean MIC values were calculated for each antifungal agent.
Isolates with missing MIC values for a given antifungal agent were excluded from the calculations for that specific drug. All analyses were performed at the overall isolate level and were not stratified by clinical setting (e.g., intensive care unit, inpatient, or outpatient).
Pearson correlation coefficients were calculated to assess relationships among log2-transformed MIC values for azole and echinocandin agents. Amphotericin B was excluded from the correlation analysis because cross-class MIC correlations with azoles have limited clinical interpretability and no validated CLSI clinical breakpoint is available for amphotericin B in Candida species.
Results
Species Distribution and Specimen Sources
A total of 100 yeast isolates were included in the study. Blood (43%) and urine (37%) were the predominant specimen sources, followed by pus, tissue, tips, and other specimens.
The species distribution was as follows: Candida albicans (n = 37, 37%), Candida tropicalis (n = 32, 32%), Candida krusei (n = 6, 6%), Candida famata (n = 5, 5%), Candida parapsilosis (n=5, 5%), Candida glabrata (n = 4, 4%), and Candida pelliculosa (n = 3, 3%). Less frequently isolated species included Cryptococcus neoformans (n = 2), Candida orthopsilosis (n = 1), Candida guilliermondii (n=1), Papiliotrema laurentii (formerly Cryptococcus laurentii) (n = 1), Kodamaea ohmeri (n = 1), and Dirkmeia churashimaensis (n = 1). One mixed yeast infection was also identified. No clinical Candida auris isolates were identified during the study period.
Overall Antifungal Susceptibility Profile
The overall in vitro antifungal activity against all isolates is summarized in Table 1 and Figure 1.
The distribution of MIC values (Figure 1) demonstrated substantial variability for fluconazole, whereas the echinocandins exhibited tightly clustered low MIC values, indicating consistent in vitro activity.
Amphotericin B showed relatively narrow variability, suggesting consistent antifungal activity.
Fluconazole MICs ranged from 0.12 to 256 µg/mL, with an MIC50 of 1 µg/mL, an MIC90 of 4 µg/mL, and a geometric mean MIC of 1.56 µg/mL. Voriconazole, itraconazole, and posaconazole demonstrated low MIC50 values (≤0.06 µg/mL) with narrow MIC distributions. All three echinocandins—caspofungin, micafungin, and anidulafungin—showed excellent in vitro activity, with MIC50 values of 0.03 µg/mL and low geometric mean MICs. Amphotericin B retained consistent activity, with an MIC50 of 0.25 µg/mL.
MIC analysis included 89–90 isolates per antifungal agent, depending on the availability of valid MIC results. Isolates with missing MIC values were excluded only from the analysis of the respective drug.
CLSI categorical interpretations were applied only to Candida species for which established clinical breakpoints were available. For non-Candida yeasts, including Cryptococcus and other rare species, MIC values are presented descriptively because validated clinical breakpoints are unavailable.
Based on applicable CLSI species-specific clinical breakpoints, most Candida albicans isolates were categorized as susceptible to echinocandins and azoles for which applicable CLSI clinical breakpoints were available. Reduced fluconazole susceptibility was observed primarily among non-albicans Candida species. For Candida glabrata, fluconazole results were categorized as susceptible‑dose dependent or resistant according to CLSI criteria.
Candida albicans: Isolates generally demonstrated low MICs for azoles and echinocandins. Fluconazole MICs ranged from 0.12 to 8.0 µg/mL, with an MIC50 of 1 µg/mL and an MIC90 of 4 µg/mL. All echinocandins demonstrated potent activity, with MIC50 values of ≤0.03 µg/mL (Table 2). Amphotericin B also demonstrated consistent activity.
Candida tropicalis: Isolates demonstrated preserved susceptibility to most antifungal agents. Fluconazole MICs ranged from 0.5 to 8.0 µg/mL, with MIC50 and MIC90 values of 1 µg/mL and 4 µg/mL, respectively.
Candida krusei: Isolates demonstrated intrinsically elevated fluconazole MICs (1–128 µg/mL), with an MIC50 of 64 µg/mL and an MIC90 of 128 µg/mL. Voriconazole and the echinocandins demonstrated lower MIC values.
Candida glabrata: Isolates exhibited intrinsically reduced susceptibility to fluconazole, with MICs ranging from 4 to 256 µg/mL and a high geometric mean MIC of 11.31 µg/mL.
Cryptococcus neoformans: Isolates showed a narrow fluconazole MIC range (0.5–1.0 µg/mL).
Other non-Candida species and rare yeasts demonstrated variable MIC distributions across the antifungal agents tested. For these yeasts, MIC values are presented descriptively only and should not be interpreted as indicating clinical susceptibility.
Most Candida albicans and Candida tropicalis isolates remained susceptible to fluconazole and the echinocandins, whereas Candida glabrata predominantly exhibited susceptible‑dose dependent or resistant fluconazole phenotypes. Candida krusei was intrinsically resistant to fluconazole.
Percentages were calculated only for isolates with valid MIC results for which categorical interpretation was possible. Therefore, the totals may not correspond to the total number of isolates within a species because isolates with missing MIC values or without applicable CLSI breakpoints were excluded from the categorical analysis.
Comparative Analysis
Figure 2 presents a bubble plot showing the geometric mean MICs of the antifungal agents across yeast species.
Species-specific comparisons showed higher fluconazole MIC values for Candida glabrata and Candida krusei, whereas the echinocandins maintained low MIC values across most species. This pattern highlights the variability in azole susceptibility and the consistently low MICs observed with echinocandins.
Antifungal MIC Correlation Analysis
Figure 3 illustrates the pairwise Pearson correlation coefficients among antifungal MIC values. Strong correlations were observed only between certain azole pairs (fluconazole–itraconazole r = 0.78, fluconazole–posaconazole r = 0.78, itraconazole–posaconazole r = 0.86). Other azole pairs showed weaker correlations (e.g., voriconazole–itraconazole r = 0.20, voriconazole–posaconazole r = 0.20). Echinocandin correlations ranged from weak to moderate (r = −0.02 to 0.65). Amphotericin B was excluded from the correlation analysis because cross-class MIC correlations with azoles were considered to have limited clinical interpretability in the absence of validated clinical breakpoints and treatment-outcome data.
Discussion
The antifungal susceptibility patterns observed in this study support the continued role of echinocandins in the management of invasive candidiasis. Consistently low MIC values across most Candida species indicate preserved in vitro activity of this antifungal class, consistent with existing evidence demonstrating their potent activity and the globally low prevalence of echinocandin resistance, which is largely attributable to the low frequency of FKS gene mutations[12, 13]. The relatively higher MICs observed in the Candida parapsilosis complex are consistent with its intrinsically reduced susceptibility, which is associated with naturally occurring polymorphisms in the β-(1,3)-D-glucan synthase enzyme rather than acquired resistance mechanisms[12, 13].
The categorical susceptibility findings (Table 3) further support the preserved activity of echinocandins against the major Candida species. However, the greater variability in azole susceptibility, particularly among Candida glabrata and Candida krusei, underscores the importance of species-level identification and susceptibility-guided therapy.
Echinocandin resistance and antifungal susceptibility trends among Candida species have been widely studied in recent years, highlighting the emergence of resistance mechanisms and the need for continued surveillance[14]. Amphotericin B maintained stable in vitro activity across most isolates in this study, consistent with previous reports indicating preserved susceptibility[15].
Overall, these findings are consistent with the role of echinocandins in empirical therapy for high-risk cancer patients, with subsequent therapy guided by species-level identification and local susceptibility patterns[16].
The antifungal MIC correlation analysis further supported these findings, with strong associations only between certain azole pairs (fluconazole–itraconazole, fluconazole–posaconazole, itraconazole–posaconazole), while other azole pairs were weakly correlated. Echinocandin correlations ranged from weak to moderate (r = −0.02 to 0.65). Amphotericin B was excluded from the correlation analysis because cross-class correlations with azoles have limited clinical interpretability in the absence of validated clinical breakpoints and treatment-outcome data. These findings highlight the value of considering species-specific susceptibility patterns when interpreting antifungal MIC data[17, 18].
Clinical manifestations and treatment outcomes of candidemia vary according to the infecting Candida species, emphasizing the importance of species-level identification in guiding appropriate clinical management[19].
Consistently low MIC values across most Candida species indicate preserved in vitro activity of echinocandins, with no substantial evidence of reduced susceptibility in the isolates tested[20].
Azole antifungals, particularly fluconazole, demonstrated considerable variability in MIC values. Recent multicenter studies have reported increasing azole resistance among Candida species, particularly Candida parapsilosis[21]. Species-specific differences in efflux pump activity and alterations in ergosterol biosynthesis may contribute to reduced azole susceptibility, further emphasizing the importance of early species identification and susceptibility-guided antifungal therapy[22]. The high frequency of Candida tropicalis in our cohort is consistent with reports from Pakistan and India, where Candida tropicalis has been identified as a leading Candida species, including among patients with hematologic malignancies[23, 24].
It is also important to note that Candida glabrata exhibits intrinsically reduced susceptibility to fluconazole, which should be considered when interpreting its elevated MICs. Its occurrence in oncology populations may be associated with factors such as neutropenia, prolonged hospitalization, broad-spectrum antibiotic exposure, and intensive healthcare exposure, although these factors were not specifically evaluated in our study.
From a clinical perspective, the susceptibility findings support the use of an echinocandin as appropriate empirical therapy for high-risk oncology patients with suspected invasive candidiasis, particularly given the consistently low echinocandin MICs observed among the major Candida species. Subsequent therapy should be individualized based on species identification, susceptibility results, and clinical response, with de-escalation considered when clinically appropriate[16].
Rare yeast species identified in this study, including Dirkmeia churashimaensis, Kodamaea ohmeri, and Papiliotrema laurentii, are uncommon but recognized opportunistic human pathogens, particularly in immunocompromised and oncology patients. Reported infections include fungemia and catheter-related bloodstream infections, while Dirkmeia churashimaensis has also been associated with an outbreak of bloodstream infection[25-27]. The Dirkmeia churashimaensis and Kodamaea ohmeri isolates identified in our study were confirmed by sequencing. This molecular confirmation reduces the likelihood of species misidentification and supports the reliability of these uncommon findings. However, because these species were represented by single isolates and clinical outcome data were unavailable, their clinical significance in our cohort cannot be determined.
Study Limitations
This study has several limitations. First, it was conducted at a single center and included a relatively small sample size, which may limit the generalizability of the findings. Second, clinical outcome data were unavailable, precluding assessment of the relationship between antifungal susceptibility results and treatment response or patient outcomes. Third, the absence of CLSI clinical breakpoints for several species-drug combinations restricted categorical interpretation for some isolates. The lack of stratification by clinical setting may also limit the interpretation of setting-specific antifungal resistance patterns. Despite these limitations, the study provides valuable local epidemiological and antifungal susceptibility data that may inform clinical practice and future surveillance efforts. The recovery of single isolates of rare yeast species, including Dirkmeia churashimaensis, also limited species-specific assessment of their clinical significance and susceptibility patterns.
The taxonomy of Candida famata has undergone substantial revision, with isolates historically identified as Candida famata now largely assigned to Debaryomyces hansenii and related taxa. Therefore, species-level interpretation of the Candida famata isolates should be approached with caution[28].
Conclusion
This study highlights the predominance of non-albicans Candida species and the variability in antifungal susceptibility patterns in a tertiary care oncology setting. Echinocandins demonstrated consistent in vitro activity, whereas azole resistance remains a concern. These findings underscore the importance of species-level identification and routine antifungal susceptibility testing to guide appropriate antifungal therapy.
