Virulence determinants, antifungal resistance, and genotype-phenotype associations in Candida bloodstream isolates: a three-year surveillance study from Northern India

Abstract

Candidemia is increasingly driven by non-albicans multidrug-resistant Candida species, yet the relationship between virulence determinants and antifungal susceptibility remains incompletely understood. In this three-year surveillance study (2022–2025), 357 non-duplicate Candida bloodstream isolates from a tertiary care centre in northern India were analysed to characterize species distribution, extracellular virulence phenotypes, virulence-associated genes, and antifungal susceptibility profiles. Species identification was performed using phenotypic methods and confirmed by PCR-RFLP and MALDI-TOF MS, while antifungal susceptibility testing followed CLSI M27-A3 guidelines. Non-albicans Candida species predominated (76.8%), with Candida tropicalis as the most frequent isolate, followed by Candida albicans and Candidozyma auris. Multiplex PCR demonstrated a high prevalence of virulence-associated genes (HWP1, ALS1, SAP2, and PLB1), with concordance between gene detection and phenotypic expression of adhesion, biofilm formation, and extracellular enzyme activity. Echinocandins retained potent in vitro activity across species, whereas C. auris exhibited a multidrug-resistant phenotype, including high fluconazole resistance and reduced susceptibility to amphotericin B and caspofungin. Species-stratified analyses revealed significant positive correlations between esterase and hemolysin activity and amphotericin B minimum inhibitory concentrations in C. albicans and C. tropicalis (p < 0.01). Multivariate analysis identified elevated amphotericin B MICs and specific virulence traits as independent predictors of strong biofilm formation. These findings highlight species-specific associations among virulence expression, genetic determinants, and the antifungal response, supporting integrated surveillance approaches to improve the management of invasive candidiasis.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-57636-6.

Introduction

Candidemia remains a major cause of morbidity and mortality among hospitalized patients worldwide, particularly in intensive care and immunocompromised populations¹. Despite advances in antifungal therapy and diagnostics, outcomes of invasive Candida bloodstream infections remain suboptimal, reflecting both host-related vulnerabilities and pathogen-specific determinants of virulence and antifungal response². Over recent decades, a marked shift in candidemia epidemiology has been observed, with non-albicans Candida species increasingly supplanting C. albicans as the predominant etiological agent in many regions³.

This epidemiological transition has important therapeutic implications, as non-albicans Candida species frequently exhibit variable or reduced susceptibility to commonly used antifungal agents⁴. In Asia and other low- and middle-income settings, species such as C. tropicalis and members of the C. parapsilosis complex are increasingly prevalent, in contrast to the dominance of C. glabrata in high-income countries^5–7. The global emergence of C. auris has further complicated candidemia management due to its multidrug-resistant phenotype, persistence in healthcare environments, and association with nosocomial outbreaks⁸. These trends underscore the need for continued regional surveillance integrating accurate species identification with antifungal susceptibility testing.

In addition to antifungal resistance, Candida pathogenicity is influenced by a repertoire of virulence factors, including extracellular hydrolytic enzymes such as phospholipases, proteinases, hemolysins, and esterases⁹. These factors contribute to host tissue invasion, nutrient acquisition, immune modulation, and survival within the bloodstream¹⁰. While virulence mechanisms have been extensively investigated in C. albicans, comparatively limited data are available on the expression of these phenotypes in non-albicans Candida species that cause invasive disease¹¹. Moreover, virulence traits are rarely examined alongside antifungal susceptibility in clinical surveillance studies.

Increasing evidence suggests that virulence expression and antifungal susceptibility may be interconnected rather than independent characteristics, potentially reflecting shared adaptive or stress-response pathways^12–14. However, systematic evaluations of the relationship between extracellular virulence phenotypes and antifungal susceptibility among bloodstream Candida isolates remain scarce, particularly in large clinical cohorts and in regions with a high burden of non-albicans species and C. auris¹⁵.

In this study, we conducted a three-year prospective surveillance of Candida bloodstream isolates at a tertiary care centre in northern India to characterize species distribution, extracellular virulence enzyme activity, virulence-associated genes, and antifungal susceptibility patterns. By integrating phenotypic and genotypic virulence profiling with standardized antifungal susceptibility testing and species-stratified analysis, we aimed to define the pathogenic and resistance landscape of invasive candidiasis and to explore associations between virulence determinants and antifungal response.

Material and method

Study design and clinical isolates

This prospective observational study was conducted from August 2022 to August 2025 at Sir Sunderlal Hospital, Institute of Medical Sciences, Banaras Hindu University, Varanasi, India. A total of 357 non-duplicate Candida isolates recovered from blood cultures of patients with suspected bloodstream infections were included. Only the first isolate per patient was analysed, and repeat isolates were excluded. Ethical approval was obtained from the Institutional Ethics Committee (Approval No. IMS/IEC/2024/7332). All methods were carried out in accordance with relevant guidelines and regulations. As the isolates were obtained as part of routine diagnostic procedures and no patient-identifiable information was used, the informed consent was waived.

Phenotypic identification of Candida species

Initial identification of yeast isolates was performed using standard mycological techniques. Colonies grown on Sabouraud’s dextrose agar (SDA) were examined for macroscopic morphology and subjected to Gram staining. Germ tube formation was assessed following incubation in human serum at 37 °C for 2 h. Species differentiation was further performed using CHROMagar Candida based on colony colour and morphology, and chlamydospore formation was evaluated on cornmeal agar supplemented with Tween 80¹⁶.

Molecular identification by PCR-RFLP

Genomic DNA was extracted from fresh yeast cultures using the cetyltrimethylammonium bromide (CTAB) method with mechanical disruption in a glass-bead mixer¹⁷. DNA concentration and purity were assessed spectrophotometrically. The internal transcribed spacer (ITS1-5.8S-ITS2) region was amplified using universal fungal primers ITS1 (5’-TCCGTAGGTGAACCTGCGG-3’) and ITS4 (5’-TCCTCCGCTTATTGATATGC-3’). PCR amplification was performed in a 25 µL reaction mixture containing PCR buffer, dNTPs, Taq DNA polymerase, primers, and template DNA. Cycling conditions included an initial denaturation at 95 °C for 5 min, followed by 30 cycles of denaturation at 95 °C for 30 s, annealing at 55 °C for 30 s, and extension at 72 °C for 1 min, with a final extension at 72 °C for 5 min¹⁸.

PCR products were digested with the restriction enzyme MspI at 37 °C for 2 h. Restriction fragments were separated on 2% agarose gels and visualized under UV illumination. Species identification was based on characteristic banding patterns.

MALDI-TOF MS confirmation

Final species confirmation was performed using MALDI-TOF MS (Bruker Daltonics, Bremen, Germany). Spectra were analyzed using the Bruker Biotyper database. Identification scores ≥ 2.0 were considered reliable for species-level identification, scores between 1.7 and 1.99 for genus-level identification, and scores < 1.7 were considered unreliable. All MALDI-TOF analyses were performed at PGIMER, Chandigarh.

Preservation of isolates and reference strains

Clinical isolates were preserved in brain heart infusion (BHI) broth supplemented with 20% glycerol at − 80 °C until further testing. Reference strains used as quality controls included C. albicans ATCC 90,028, C. tropicalis ATCC 750, C. krusei ATCC 6258, C. parapsilosis ATCC 22,019, and C. auris CDC B11903.

Determination of phospholipase activity

Phospholipase activity was assessed using egg yolk agar as described by Cox et al., with minor modifications¹⁹. Briefly, yeast suspensions were adjusted to the McFarland 2 standard, and 5 µL aliquots were spot-inoculated onto egg yolk agar plates. Plates were incubated at 37 °C for up to 5 days. Phospholipase activity (Pz) was calculated as the ratio of colony diameter to the total diameter of the colony plus precipitation zone. Lower Pz values indicated higher enzymatic activity. Assays were performed in duplicate.

Determination of proteinase activity

Proteinase activity was evaluated using bovine serum albumin (BSA) agar medium as described earlier, with minor modifications²⁰. Standardized yeast suspensions were inoculated onto BSA agar plates and incubated at 37 °C for 5 days. Proteinase activity was determined by measuring the clear zone surrounding the colonies and expressed as Pz values. All experiments were conducted in triplicate.

Determination of esterase activity

Esterase production was assessed using the Tween 80 opacity test as outlined earlier, with minor modifications²¹. Yeast suspensions were inoculated onto Tween 80 agar plates and incubated at 30 °C for up to 10 days. Esterase activity was quantified as the ratio of colony diameter to the diameter of the opaque precipitation zone (Ez value). Lower Ez values indicated stronger activity.

Determination of hemolysin activity

Hemolysin production was evaluated on Sabouraud’s dextrose agar supplemented with 3% glucose and 7% sheep blood as described earlier²². Plates were inoculated with standardized yeast suspensions and incubated at 30 °C for 48–72 h. Haemolytic activity was quantified by calculating the haemolysis zone (Hz) ratio.

Antifungal susceptibility testing

Antifungal susceptibility testing was performed using the broth microdilution method according to CLSI M27-A3 guidelines, as outlined earlier^23,24. The antifungal agents tested included amphotericin B, fluconazole, voriconazole, itraconazole, posaconazole, caspofungin, micafungin, and anidulafungin. MIC endpoints were visually determined after 24–48 h of incubation at 35 °C, using 50% growth inhibition for azoles and echinocandins, and 100% inhibition for amphotericin B. CLSI species-specific clinical breakpoints were applied where available. For species lacking established CLSI breakpoints (e.g., C. pelliculosa and C. rugosa), interpretive criteria for C. albicans were applied, as previously described²³. This approach was adopted to enable comparative analysis across species in the absence of species-specific breakpoints, consistent with practices in epidemiological surveillance studies. All susceptibility assays were performed in triplicate, and discrepancies were resolved by repeating the tests.

Biofilm formation assay

Clinical Candida isolates were initially cultured on Sabouraud’s dextrose agar to confirm viability and purity, and biofilm formation was assessed by the method described earlier²⁵. Briefly, they were then inoculated into Brain Heart Infusion (BHI) broth at 37 °C and diluted 1:100. Standardised inocula were prepared by adjusting the suspension of the clinical isolates to approximately 1 × 10⁶ cells/mL. A 20 µL aliquot of each Candida suspension was added to sterile 96-well flat-bottom microtiter plates containing 180 µL of BHI broth. The plates were incubated at 37 °C without agitation for 24 h to allow biofilm development. After incubation, nonadherent cells were gently removed by washing the wells with sterile phosphate-buffered saline (PBS), and the attached biofilm biomass was quantified using crystal violet staining. Adherent Candida cells were heat-fixed at 60 °C, stained with 0.5% crystal violet, washed to remove excess stain, and air-dried. The dye was then dissolved using an ethanol-acetone mixture (80:20 v/v). Absorbance was recorded at 570 nm with an ELISA reader (MultiSkan SkyHigh, Thermo Scientific). This optical density directly reflects the amount of biofilm produced. Biofilm-forming ability was categorized based on the mean OD values relative to the control, as previously described in fungal biofilm studies. Isolates were classified as follows: (a) Non/weak biofilm producers: OD ≤ control; (b) Moderate biofilm producers: control < OD ≤ 2× control; (c) Strong biofilm producers: OD > 2× control.

Detection of virulence-associated genes by multiplex PCR

To complement phenotypic virulence profiling, selected virulence-associated genes were detected using multiplex PCR targeting hyphal wall protein 1 (HWP1), agglutinin-like sequence 1 (ALS1), secreted aspartyl proteinase 2 (SAP2), and phospholipase B1 (PLB1), based on previously validated protocols^26–28(Mukherjee et al., 2001; Dawoud et al., 2024). These genes were selected due to their established roles in adhesion, biofilm formation, tissue invasion, and extracellular enzymatic activity. PCR amplification was performed in a final reaction volume of 25 µL containing PCR buffer, MgCl₂, dNTPs, gene-specific primers, Taq DNA polymerase, and template DNA. Primer sequences and expected amplicon sizes were as follows: HWP1 (572 bp), ALS1 (239 bp), SAP2 (178 bp), and PLB1 (751 bp). Thermal cycling conditions consisted of an initial denaturation step: 1 cycle at 94 °C for 4 min; followed by 35 cycles of 94 °C for 30 s, 52 °C for 1 min, and 72 °C for 2 min. A final extension cycle was performed at 72 °C for 5 min of an initial denaturation at 94 °C for 4–5 min, followed by 30–35 cycles of denaturation at 94 °C for 30 s, annealing at 52–55 °C for 45–60 s, and extension at 72 °C for 1–2 min, with a final extension at 72 °C for 5–7 min (Table 1). PCR products were resolved on 2% agarose gel stained with ethidium bromide and visualized under UV transillumination. A 100 bp DNA ladder was used as a molecular size marker. Positive and negative controls were included in each run to ensure assay validity. Gene presence was recorded based on the detection of bands at the expected sizes.

Table 1.

Primers used for PCR for the detection of the investigated genes.

Genes	Primers (5’-3’)	Amplicon Size	Reference
HWP1	F: CCA TGT GAT GAT TAC CCA CA	572	26
	R: GCT GGA ACA GAA GAT TCA GG
ALS1	F: CCA TCA CTG AAG ATA TCA CCACA	239	26
	R: TGG AGC TTC TGT AGG ACT GGTT
SAP 2	F: AAC AAC AAC CCA CTA GAC ATC ACC	178	27
	R: TGA CCA TTA GTA ACT GGG AAT GCT TTA GGA
PLB1	F: ATGATTTTGCATCATTTG	751	28
	R: AGTATCTGGAGCTCTAC

Statistical analysis

Quantitative variables were summarized using mean and standard deviation (SD) or median and interquartile range (IQR), as appropriate. Categorical variables were expressed as counts and percentages. Data distribution was assessed using the Shapiro-Wilk test. As isolate-level averaged measurements approximated a normal distribution, parametric statistical methods were applied. Biofilm biomass was analysed on a species-wise basis using isolate-level mean values derived from triplicate measurements. Comparisons between biofilm biomass and control values were performed using one-sample t-tests. Associations between categorical variables, including Candida species, biofilm strength categories, virulence enzyme activity, and antifungal susceptibility categories, were assessed using the chi-square (χ²) test or Fisher’s exact test, as appropriate. Species with insufficient sample size were excluded from inferential testing. To identify independent predictors of strong biofilm formation, multivariate logistic regression analysis was performed. Variables entered into the model included amphotericin B MIC category (≥ 1 µg/mL vs. < 1 µg/mL), Candida species, and extracellular virulence enzyme activity (esterase and hemolysin). Adjusted odds ratios (ORs) with 95% confidence intervals (CIs) were calculated. Model fit was evaluated using the Hosmer-Lemeshow goodness-of-fit test, and model explanatory power was assessed using Nagelkerke R². All statistical tests were two-tailed, and a p-value < 0.05 was considered statistically significant. Statistical analyses were performed using SPSS version 25 (IBM Corp., Armonk, NY, USA) and GraphPad Prism version 10.0.

Results

To provide an integrated assessment of invasive candidiasis, the distribution of species, extracellular virulence enzyme activity, antifungal susceptibility patterns, and their interrelationships were analyzed across all bloodstream isolates.

Species distribution

A total of 357 Candida isolates were recovered from bloodstream infections during the study period. Species identification was performed using phenotypic methods and confirmed by PCR-RFLP and MALDI-TOF MS. Non-albicans Candida species predominated, accounting for 76.8% of isolates. C. tropicalis was the most frequently identified species (42.6%), followed by C. albicans (23.2%) and C. auris (7.6%). Less common species included C. pelliculosa, C. krusei, C. parapsilosis, C. glabrata, and others, reflecting substantial species diversity among bloodstream isolates. Phenotypic identification using microscopy, germ tube testing, chromogenic media, and micromorphological features enabled the initial classification of Candida bloodstream isolates (Fig. 1). Based on this preliminary identification, the overall distribution of Candida species was subsequently analyzed to delineate epidemiological patterns among bloodstream infections (Fig. 2).

Fig. 1.

Phenotypic identification of Candida species. Representative phenotypic identification of Candida species. (A) Gram-stained yeast cells (100× magnification). (B) Germ tube formation by C. albicans following incubation in human serum (40× magnification). (C) Colony morphology of Candida species on CHROMagar Candida, demonstrating species-specific colour differentiation. (C1) C. albicans: light green coloured, (C2) C. tropicalis: blue to metallic blue coloured, (C3) C. auris: pinkish to cream coloured, (C4) C. krusei: purplish coloured (C5) C. rugosa: Dark green, (C6) C. glabrata: cream to white colonies, (C7) C. orthopsilosis: creamy pink, (C8) C. parapsilosis: cream to pale pink colonies (C9) C. pelliculosa, and (C10) C. guilliermondii: small pink to purple colonies.

Fig. 2.

Species distribution of Candida bloodstream isolates. Distribution of Candida species isolated from bloodstream infections over the study period (August 2022-August 2025). (A) Proportion of C. albicans and non-albicans Candida species among all bloodstream isolates. (B) Absolute number of isolates for each Candida species identified.

Non-albicans Candida species predominated, accounting for 76.8% of isolates. C. tropicalis was the most frequently isolated species (42.6%; n = 152), followed by C. albicans (23.24%; n = 83), C. pelliculosa (11.5%; n = 41), C. auris (7.6%; n = 27), C. krusei (6.2%; n = 22), C. parapsilosis (3.36%, n = 11), C. glabrata (2.52%, n = 9), C. rugosa (1.96%, n = 7), C. orthopsilosis (1.12%, n = 4), and C. guilliermondii (0.28%, n = 1).

The species distribution patterns observed among Candida bloodstream isolates (Fig. 2) underscored the predominance of non-albicans Candida species and the diversity of organisms implicated in candidemia. To further characterize these isolates at the species level, detailed phenotypic features were examined using micromorphological methods (Fig. 3).

Fig. 3.

Micromorphological features of selected Candida species. Micromorphological features of selected Candida species on cornmeal agar with Tween 80. morphology on cornmeal agar A-C.albicans, B-C.tropicalis, C-C.auris, D-C.parapsilosis, E-C.pelliculosa, F- C.glabrata, G-C.krusei, H-C.rugosa.

Given the potential for phenotypic overlap among closely related species, molecular confirmation was subsequently performed using PCR-RFLP analysis. PCR-RFLP analysis of the ITS region provided definitive species-level confirmation for Candida bloodstream isolates (Fig. 4). This approach enabled accurate discrimination among closely related, phenotypically overlapping species, ensuring robust classification across the study cohort. Molecular confirmation formed the foundation for subsequent analyses of extracellular virulence determinants and antifungal susceptibility, allowing species-specific evaluation of pathogenic potential and resistance patterns.

Fig. 4.

Molecular identification by PCR-RFLP. PCR-RFLP analysis of ITS1-ITS4 amplicons digested with MspI restriction enzyme for species-level identification of Candida isolates. Representative banding patterns corresponding to different Candida species are shown. M: 100 bp DNA ladder; lanes 1, 2, 3, 20, 21, 22,24, 25,30, 34: C. albicans; lanes 8,14, 16, 19, 27, 32: C. tropicalis; lanes 17, 18, 23, 29, 33: C. auris; lanes 4, 5: C. rugosa; lanes 12: C. glabrata; lanes 15: C. krusei; lane 9: C. guilliermondii; lanes 10, 11: C. pelliculosa; lane 6, 24, 26, 28, 31: C. parapsilosis, lane 7: unidentified.

Extracellular virulence enzyme activity

All isolates of C. rugosa and C. guilliermondii exhibited negative results for phospholipase activity. Conversely, the highest levels were observed in C. albicans, with 85.5% of isolates showing very strong activity and 1.2% showing positive activity. Among non-albicans Candida species, phospholipase activity was expressed to varying degrees. Strong or positive phospholipase activity was observed in subsets of C. tropicalis (20.3%), C. auris (14.8%), C. krusei (36.4%), C. orthopsilosis (22.2%), C. parapsilosis (25%), and C. pelliculosa (54.5%) isolates. However, the majority of isolates belonging to these species demonstrated either weak or absent phospholipase activity, indicating substantial interspecies heterogeneity in enzyme expression.

All isolates of C. orthopsilosis, C. guilliermondii, and C. parapsilosis had robust haemolysin activity. Strong haemolysin activity was observed in 96.38%, 93.42%, and 95.12% of isolates of C. albicans, C. tropicalis, and C. pelliculosa, respectively. No affirmative activity was observed in C. utilis, C. glabrata, or C. orthopsilosis. Certain isolates of C. albicans, C. tropicalis, C. auris, and C. krusei exhibited positive haemolysin activity, while several isolates from different species demonstrated negative results.

Significant proteinase activity was detected in C. orthopsilosis, C. guilliermondii, and C. parapsilosis, as well as in 78.31% of C. albicans and 75.65% of C. tropicalis.

Esterase activity was maximal in C. albicans and C. tropicalis, while all isolates of C. orthopsilosis exhibited robust positive activity. In C. auris, 77.77% had robust esterase activity, whereas 63.63% of C. krusei isolates tested negative. C. rugosa and C. parapsilosis exhibited pronounced positive esterase activity at substantial rates. Reference strains demonstrated expected enzyme activity profiles, validating assay performance. Enzyme activity categories were assigned according to predefined Pz/Ez/Hz thresholds.

Figure 5 summarises the overall activity of virulence enzymes among Candida bloodstream isolates. However, given the marked diversity in species composition, a species-level analysis was necessary to better understand interspecies variability in virulence phenotypes. Accordingly, Fig. 6 depicts the distribution of phospholipase, proteinase, esterase, and hemolysin activity across major Candida species, highlighting differences in the proportion of isolates exhibiting strong, moderate, or absent enzymatic activity.

Fig. 5.

Extracellular virulence enzyme activity. Representative extracellular virulence enzyme activity among Candida bloodstream isolates. (A) Phospholipase activity on egg yolk agar. (B) Proteinase activity on bovine serum albumin agar. (C) Esterase activity demonstrated by the Tween 80 opacity test (D). Hemolysin activity on glucose-enriched sheep blood agar. Precipitation or haemolysis zones surrounding colonies indicate enzymatic activity.

Fig. 6.

Distribution of extracellular virulence enzyme activity across Candida species. Charts depict the proportion of bloodstream isolates of major Candida species exhibiting strongly positive, positive, or negative activity for (A) phospholipase, (B) proteinase, (C) esterase, and (D) hemolysin, based on established Pz/Ez/Hz thresholds. Data are expressed as percentages of isolates within each species. Only species with adequate sample sizes are shown.

Statistical analysis revealed significant interspecies differences in extracellular virulence enzyme activity among Candida isolates. Chi-square analysis demonstrated highly significant variation in phospholipase (χ² = 118.55, p < 0.0001) and esterase activity (χ² = 105.79, p < 0.0001), while proteinase (χ² = 26.85, p < 0.001) and hemolysin activity (χ² = 23.85, p < 0.001) also differed significantly across species. These findings underscore marked heterogeneity in virulence-associated enzyme expression among Candida species. Further, pairwise Fisher’s exact tests revealed that Candida albicans exhibited significantly higher phospholipase activity than both C. auris (OR = 34.02, p < 0.0001) and C. glabrata (OR = 20.71, p = 0.00016). Hemolysin activity was also significantly higher in C. albicans than in C. auris (OR = 7.62, p = 0.0065). No statistically significant differences in proteinase or esterase activity were observed in these pairwise comparisons.

Distribution of virulence-associated genes

To further characterize the molecular basis of virulence among Candida bloodstream isolates, multiplex PCR was performed to detect key virulence-associated genes (HWP1, ALS1, SAP2, and PLB1). Overall, a high prevalence of adhesion- and invasion-related genes was observed across both C. albicans and non-albicans Candida species. Among C. albicans isolates, HWP1 and ALS1 were the most frequently detected genes, consistent with their established role in hyphal formation and epithelial adherence. SAP2 was detected in a substantial proportion of isolates, consistent with the observed proteinase activity, whereas PLB1 detection corresponded to phospholipase-producing isolates.

Among non-albicans Candida species, particularly C. tropicalis and C. auris, ALS1- and HWP1-homolog-associated amplification patterns were frequently observed, indicating conserved adhesion mechanisms across species. SAP2 detection rates were moderate, while PLB1 was variably present, consistent with the heterogeneous phospholipase activity observed phenotypically (Fig. 7).

Fig. 7.

Multiplex PCR detection of virulence-associated genes in Candida bloodstream isolates. Representative agarose gel electrophoresis image showing amplification of virulence-associated genes using multiplex PCR targeting PLB1 (751 bp), HWP1 (572 bp), ALS1 (239 bp), and SAP2 (178 bp). PCR products were resolved on a 2% agarose gel and visualized under UV illumination. Distinct bands corresponding to the expected amplicon sizes indicate the presence of respective virulence genes among Candida isolates. The rightmost lane represents a 100 bp DNA ladder (molecular size marker), with fragment sizes indicated (100–1000 bp). Multiple lanes demonstrate co-amplification of genes within individual isolates, reflecting the coexistence of adhesion (HWP1, ALS1) and enzyme-associated (SAP2, PLB1) virulence determinants. Variation in band intensity across lanes reflects differences in amplification efficiency and/or gene copy number among isolates.

Gene distribution patterns showed concordance with extracellular enzyme activity assays. Isolates demonstrating strong proteinase and phospholipase activity were more likely to harbour SAP2 and PLB1 genes, respectively, whereas strong biofilm producers were frequently associated with HWP1 and ALS1 positivity. These findings indicate that molecular virulence determinants are widely distributed among bloodstream isolates and support the phenotypic observations of extracellular enzyme production and biofilm formation.

Antifungal susceptibility profiles

While Figs. 5 and 6 describe the distribution of extracellular virulence phenotypes among Candida bloodstream isolates, antifungal susceptibility testing was performed to evaluate patterns of therapeutic response across species. The antifungal susceptibility profiles of the isolates, including MIC₅₀ and MIC₉₀ values and categorical susceptibility interpretations, are summarized in Fig. 8. Species-wise MIC₅₀ and MIC₉₀ values for amphotericin B, azoles, and echinocandins demonstrated marked interspecies variability (Fig. 8A). Echinocandins exhibited low MIC₅₀ and MIC₉₀ values across all major Candida species, whereas elevated MICs were observed for azoles and amphotericin B, particularly among C. auris isolates.

Fig. 8.

Antifungal susceptibility patterns of Candida bloodstream isolates. (A-C) Species-specific minimum inhibitory concentration (MIC) distributions showing MIC₅₀ and MIC₉₀ values (µg/mL) for eight antifungal agents against major Candida species isolated from bloodstream infections: (A) Candida tropicalis (n = 152), (B) Candida albicans (n = 83), and (C) Candida auris (n = 27). Antifungal agents tested included amphotericin B, fluconazole, voriconazole, itraconazole, posaconazole, caspofungin, micafungin, and anidulafungin. MIC₅₀ and MIC₉₀ represent the concentrations inhibiting 50% and 90% of isolates, respectively. (D) Overall antifungal susceptibility profile of all bloodstream Candida isolates categorized as susceptible (S), intermediate/susceptible dose-dependent (I/SDD), or resistant (R) according to Clinical and Laboratory Standards Institute (CLSI) interpretive breakpoints. Bars represent the percentage distribution of isolates for each antifungal agent across susceptibility categories.

In contrast, azole susceptibility varied substantially by species. While C. albicans largely retained azole susceptibility, C. tropicalis exhibited broader MIC distributions and higher MIC₉₀ values, suggesting greater population-level heterogeneity. C. auris exhibited a distinctly multidrug-resistant phenotype, with high MIC₉₀ values for fluconazole, amphotericin B, and caspofungin, consistent with reduced susceptibility across multiple antifungal classes.

These MIC distribution patterns highlight the limitations of species-agnostic therapy and reinforce the need for routine antifungal susceptibility testing to guide effective treatment. In contrast, C. auris displayed a multidrug-resistant phenotype. A high proportion of isolates were resistant to fluconazole (88.9%), with notable resistance also observed to amphotericin B (29.6%) and caspofungin (48.1%). Although micafungin and anidulafungin retained relatively higher activity, reduced susceptibility to echinocandins was evident in a subset of isolates. A categorical interpretation of antifungal susceptibility according to CLSI breakpoints further highlighted species-specific resistance patterns (Fig. 8B). Overall susceptibility to echinocandins was preserved, while high rates of fluconazole resistance and reduced amphotericin B susceptibility were observed among C. auris isolates.

Intrinsic or acquired resistance patterns were observed among other non-albicans Candida species. C. krusei demonstrated expected resistance to fluconazole, while C. pelliculosa showed decreased susceptibility across multiple azole agents but retained susceptibility to echinocandins. C. glabrata and C. parapsilosis isolates largely remained susceptible to azoles and echinocandins, though species-specific MIC shifts were noted. (Detailed antifungal susceptibility data are provided in Supplementary Table S1.) The quality control MIC ranges for reference strains were within the CLSI-recommended limits.

Collectively, these findings highlight the continued efficacy of echinocandins as first-line agents for candidemia and underscore the need for routine species-level identification and antifungal susceptibility testing, particularly in settings with a high prevalence of multidrug-resistant species such as C. auris.

Correlation between virulence factors and antifungal susceptibility

Given the observed heterogeneity in both virulence enzyme expression and antifungal susceptibility among Candida species, we next examined potential associations between extracellular virulence phenotypes and minimum inhibitory concentrations against antifungals. Species-stratified correlation analysis revealed moderate but statistically significant associations between specific virulence enzyme activities and antifungal MICs (Fig. 9). Notably, positive correlations were observed between esterase and hemolysin activity and amphotericin B MICs in C. albicans and C. tropicalis, whereas correlations involving azoles and echinocandins were weaker and inconsistent.

Fig. 9.

Heatmap representation of Spearman’s rank correlation coefficients. Heatmap representation of Spearman’s rank correlation coefficients between extracellular virulence enzyme activity (phospholipase, proteinase, esterase, and hemolysin) and antifungal minimum inhibitory concentrations (MICs) among major Candida species. (A) Candida albicans. (B) Candida tropicalis. Positive correlations indicate higher MICs associated with increased enzyme activity, whereas negative correlations indicate lower MICs associated with increased enzyme activity. Only statistically significant correlations (p ≤ 0.05) were interpreted.

In C. albicans, esterase and hemolysin activities showed moderate but statistically significant positive correlations with amphotericin B MICs (r = 0.482 and r = 0.314, respectively; p < 0.01), indicating reduced polyene susceptibility among isolates with higher enzyme expression. Esterase activity also showed significant negative correlations with the MICs of itraconazole, posaconazole, caspofungin, and micafungin (p < 0.05), suggesting greater susceptibility to azoles and echinocandins in isolates with elevated esterase activity.

In C. tropicalis, phospholipase and proteinase activity were positively correlated with amphotericin B MICs (r = 0.256 and r = 0.209; p ≤ 0.01), while proteinase activity showed significant negative correlations with multiple azoles and echinocandins (p ≤ 0.01). These associations indicate a species-specific linkage between virulence expression and antifungal response.

In contrast, correlations in C. auris and less prevalent species were weak and inconsistent, likely reflecting limited sample sizes and the predominance of intrinsic resistance mechanisms. Accordingly, these findings were interpreted as exploratory.

Notably, isolates harbouring the HWP1 and ALS1 genes demonstrated a greater propensity for strong biofilm formation, consistent with their known roles in adhesion and biofilm maturation. Similarly, the presence of the SAP2 and PLB1 genes was associated with increased proteinase and phospholipase activity, respectively, suggesting functional expression of these virulence determinants. Although formal correlation analysis between gene presence and antifungal MICs was limited, isolates that were simultaneously positive for virulence genes and exhibited strong biofilm formation tended to display elevated amphotericin B MICs, particularly in C. albicans and C. tropicalis. These observations further support a potential interplay between virulence mechanisms and antifungal tolerance. These findings suggest that virulence expression and antifungal susceptibility may represent interconnected phenotypic traits in certain Candida species, although the observed associations do not imply causality.

Biofilm-forming capacity across Candida species

Based on predefined biofilm formation criteria, marked inter-species variability in biofilm-forming capacity was observed among Candida bloodstream isolates (Table 2). Among C. albicans isolates (n = 83), 53 (63.9%) were classified as strong biofilm producers, 22 (26.5%) as moderate producers, and 8 (9.6%) as weak producers. Similarly, C. tropicalis (n = 152) showed a predominance of strong biofilm producers: 81 isolates (53.3%) classified as strong, 52 (34.2%) as moderate, and 19 (12.5%) as weak biofilm formers.

Table 2.

Distribution of biofilm-forming capacity among Candida bloodstream isolates.

Species	Total isolates (n)	Weak biofilm producers n (%)	Moderate biofilm producers n (%)	Strong biofilm producers n (%)
Candida albicans	83	8 (9.6%)	22 (26.5%)	53 (63.9%)
Candida tropicalis	152	19 (12.5%)	52 (34.2%)	81 (53.3%)
Candidozyma auris	27	4 (14.8%)	12 (44.4%)	11 (40.8%)
Candida pelliculosa	41	8 (19.5%)	19 (46.3%)	14 (34.2%)
Candida krusei	22	5 (22.7%)	10 (45.5%)	7 (31.8%)
Candida glabrata	9	4 (44.4%)	3 (33.3%)	2 (22.3%)
Candida parapsilosis	11	3 (27.3%)	5 (45.4%)	3 (27.3%)
Candida rugosa	7	2 (28.6%)	3 (42.8%)	2 (28.6%)

Biofilm strength was classified based on mean optical density values relative to control wells. Weak biofilm producers were defined as OD ≤ control, moderate biofilm producers as control < OD ≤ 2× control, and strong biofilm producers as OD > 2× control.

In contrast, C. glabrata exhibited comparatively reduced biofilm-forming ability, with 4 of 9 isolates (44.4%) classified as weak biofilm producers and only 2 isolates (22.3%) demonstrating strong biofilm formation. Other non-albicans species, including C. pelliculosa, C. krusei, and C. parapsilosis, predominantly exhibited moderate biofilm-forming capacity, with strong biofilm producers accounting for approximately one-third of isolates in each species. C. auris showed a mixed biofilm phenotype, with 11 of 27 isolates (40.8%) classified as strong biofilm producers and 12 (44.4%) exhibiting moderate biofilm formation.

Comparison of biofilm strength categories across Candida species demonstrated a highly significant association between species identity and biofilm-forming capacity (χ² = 46.3, df = 14, p < 0.0001). Strong biofilm producers were disproportionately represented among C. albicans and C. tropicalis, whereas C. glabrata and C. parapsilosis showed a higher proportion of weak and moderate biofilm producers. When biofilm categories were dichotomized into strong versus non-strong producers, the association remained significant (χ² = 31.8, df = 7, p < 0.0001), indicating a species-dependent trend toward robust biofilm formation.

Overall, these findings indicate that robust biofilm formation is common among invasive Candida isolates but varies substantially in magnitude and distribution across species.

Biofilm formation and amphotericin B susceptibility

When biofilm-producing categories were aligned with amphotericin B susceptibility data, a species-dependent pattern was observed. Among C. albicans and C. tropicalis, isolates classified as strong biofilm producers, representing 53/83 and 81/152 isolates, respectively, were more frequently associated with elevated amphotericin B MICs compared with weak or moderate biofilm-producing isolates. This observation is consistent with species-stratified correlation analyses that demonstrate positive associations between biofilm-associated virulence traits, including esterase and hemolysin activities, and amphotericin B MICs.

In contrast, isolates classified as weak biofilm producers-such as 4/9 C. glabrata isolates and 8/41 C. pelliculosa isolates- generally exhibited lower amphotericin B MICs, indicating greater polyene susceptibility in the absence of extensive biofilm formation. C. auris was an exception to this trend, with elevated amphotericin B MICs observed across both moderate (12/27) and strong (11/27) biofilm producers, suggesting that intrinsic resistance mechanisms, rather than biofilm-associated tolerance, predominantly influence the antifungal response in this species.

Collectively, these findings suggest that biofilm-associated tolerance to amphotericin B is most pronounced in C. albicans and C. tropicalis, whereas antifungal resistance in C. auris appears to be largely independent of biofilm-forming capacity.

Multivariate analysis of biofilm formation and amphotericin B susceptibility

To further evaluate the relationship between biofilm formation and antifungal susceptibility, multivariate logistic regression analysis was performed using strong biofilm formation as the dependent variable. After adjustment for species and extracellular virulence enzyme activity, amphotericin B MIC ≥ 1 µg/mL was independently associated with strong biofilm formation (adjusted OR 2.84, 95% CI 1.62–4.98, p = 0.0003). C. albicans (adjusted OR 3.91, p < 0.0001) and C. tropicalis (adjusted OR 2.67, p = 0.002) were independently associated with strong biofilm production, whereas C. auris was not.

Positive esterase and hemolysin activity also remained significant predictors of strong biofilm formation. The overall model demonstrated good fit (Hosmer-Lemeshow p = 0.61) and moderate explanatory power (Nagelkerke R² = 0.34), indicating that biofilm formation is influenced by both antifungal susceptibility and virulence-associated phenotypes in a species-dependent manner.

Forest plot visualization of the multivariate logistic regression model (Fig. 10) demonstrated that amphotericin B MIC ≥ 1 µg/mL, C. albicans, C. tropicalis, and positive esterase and hemolysin activity were independently associated with strong biofilm formation, as indicated by adjusted odds ratios with confidence intervals not crossing unity. In contrast, C. auris did not show a statistically significant association with strong biofilm production after adjustment for antifungal susceptibility and virulence factors.

Fig. 10.

Multivariate logistic regression analysis of factors associated with strong biofilm formation among Candida bloodstream isolates. Forest plot showing adjusted odds ratios (aORs) with 95% confidence intervals for independent predictors of strong biofilm formation. Covariates included Candida species, amphotericin B MIC category (≥ 1 µg/mL vs. < 1 µg/mL), and extracellular virulence enzyme activity (esterase and hemolysin). The vertical reference line indicates an odds ratio of 1. Odds ratios are displayed on a logarithmic scale.

Discussion

This study provides an integrated analysis of species distribution, extracellular virulence phenotypes, and antifungal susceptibility patterns among Candida bloodstream isolates collected over a three-year period at a tertiary care centre in northern India. The predominance of non-albicans Candida species, particularly C. tropicalis, and the emergence of multidrug-resistant C. auris are consistent with regional and global epidemiological trends²⁹. By combining phenotypic virulence profiling with standardized antifungal susceptibility testing, this study extends routine candidemia surveillance beyond species identification and resistance patterns to include pathogenic traits that may influence host-pathogen interactions.

The predominance of non-albicans Candida species observed in this study aligns with multiple reports from the Indian subcontinent and other low- and middle-income settings, where C. tropicalis frequently replaces C. albicans as the leading cause of candidemia. This epidemiological shift has important clinical implications, as non-albicans species often exhibit heterogeneous antifungal susceptibility profiles and variable virulence attributes³⁰. The detection of C. auris in a substantial proportion of bloodstream isolates further underscores its expanding clinical footprint and highlights the need for continued surveillance and stringent infection prevention measures.

Extracellular virulence enzyme activity was widely expressed across Candida species, indicating that pathogenic potential is not restricted to C. albicans. While C. albicans exhibited high levels of phospholipase, proteinase, esterase, and hemolysin activity, comparable virulence phenotypes were also observed in C. tropicalis, C. auris, and members of the C. parapsilosis complex. These findings support growing evidence that non-albicans Candida species possess substantial phenotypic capacities for tissue invasion, immune modulation, and survival in the bloodstream, which may contribute to their increasing clinical relevance⁴.

Among the virulence factors assessed, hemolysin and esterase activity were particularly prevalent across multiple species. Haemolysin production may facilitate iron acquisition and persistence in the bloodstream, while esterase activity has been associated with lipid utilization and host cell interaction. The high frequency of these phenotypes across both albicans and non-albicans species suggests that they may represent conserved adaptive traits in invasive Candida infections rather than species-specific virulence markers³¹.

Antifungal susceptibility testing demonstrated preserved in vitro activity of echinocandins against most Candida species, supporting their continued role as first-line agents for the treatment of candidemia³². In contrast, azole susceptibility varied considerably across species, with C. auris isolates showing markedly reduced susceptibility. Amphotericin B showed variable activity, with elevated MICs in subsets of C. albicans (10.74% resistant), C. tropicalis (7.89% resistant), and C. auris (29.62% resistant). These findings reflect the complex, species-dependent landscape of antifungal resistance encountered in contemporary clinical practice³³.

The antifungal resistance profile of C. auris observed in this study, characterized by high fluconazole resistance and reduced susceptibility to amphotericin B and caspofungin, is consistent with reports from multiple geographic regions^34,35. The combination of multidrug resistance, persistence in healthcare environments, and the capacity for nosocomial transmission makes C. auris a particularly challenging pathogen to manage in invasive candidiasis³⁶. These findings reinforce the importance of accurate species identification, routine susceptibility testing, and robust infection control practices in healthcare settings.

The present study provides a comprehensive evaluation of candidemia by integrating species distribution, extracellular virulence phenotypes, antifungal susceptibility profiles, and molecular virulence determinants. The predominance of non-albicans Candida species, particularly C. tropicalis, is consistent with regional epidemiological trends reported across Asia and other resource-limited settings, where environmental, clinical, and antifungal usage patterns favour non-albicans species emergence^37,38. A key strength of this study lies in the combined assessment of phenotypic and genotypic virulence determinants. The high prevalence of HWP1 and ALS1 genes, particularly among strong biofilm-producing isolates, underscores the central role of adhesion-mediated pathogenicity in bloodstream infections. These findings are consistent with previous reports demonstrating that HWP1 facilitates hyphal attachment to host tissues, while ALS family proteins mediate adherence and biofilm integrity³⁹. Similarly, the detection of SAP2 and PLB1 genes in a substantial proportion of isolates aligns with the observed extracellular proteinase and phospholipase activity. These enzymes contribute to host tissue invasion, immune evasion, and nutrient acquisition, thereby enhancing pathogenic potential⁴⁰. The concordance between gene detection and phenotypic expression in the present study supports the functional relevance of these virulence factors in invasive candidiasis.

The observed association between virulence traits and antifungal susceptibility further highlights the complex interplay between pathogenicity and drug response. In particular, the positive correlation between esterase/hemolysin activity and amphotericin B MICs in C. albicans and C. tropicalis suggests that increased virulence expression may be linked to reduced polyene susceptibility⁴¹. This may reflect adaptive stress-response mechanisms, alterations in membrane composition, or biofilm-associated antifungal tolerance⁴².

The multidrug-resistant phenotype of C. auris observed in this study is consistent with global reports and represents a significant clinical concern. Unlike other species, antifungal resistance in C. auris appears largely independent of classical virulence phenotypes, suggesting that intrinsic resistance mechanisms, including efflux pumps and target-site modifications, may predominate⁴³.

The present study demonstrates that biofilm formation is a prominent and species-dependent virulence trait among Candida bloodstream isolates. A majority of C. albicans and C. tropicalis isolates were classified as strong biofilm producers, whereas other non-albicans species exhibited more heterogeneous biofilm phenotypes. These findings are consistent with the established role of biofilms in promoting persistence within the host, facilitating immune evasion, and enhancing tolerance to environmental stressors⁴⁴. Importantly, the statistically significant interspecific differences observed in biofilm strength underscore that biofilm formation is not a universal characteristic of invasive Candida isolates but rather a regulated, species-specific phenotype⁴⁵.

Strong biofilm-forming species in this cohort also exhibited higher frequencies of extracellular virulence enzyme activity, particularly esterase and hemolysin production. These enzymes are known to contribute to surface adherence, nutrient acquisition, and biofilm maturation, suggesting coordinated regulation of virulence determinants. The coexistence of robust biofilm formation with elevated enzymatic activity in C. albicans and C. tropicalis supports the concept of a multifactorial virulence program rather than isolated pathogenic traits⁴⁶. In contrast, species with weaker biofilm-forming capacity, such as C. glabrata, showed lower enzyme activity, further reinforcing this association⁴⁷.

Species-stratified correlation analysis revealed moderate associations between specific extracellular virulence phenotypes and antifungal susceptibility, most notably between esterase and haemolysin activity and amphotericin B MICs in C. albicans and C. tropicalis⁴¹. These associations suggest that virulence expression and antifungal response may represent interconnected phenotypic traits in certain species. However, the observed correlations do not imply causality and should be interpreted cautiously, particularly given the exploratory nature of the analysis.

Alignment of biofilm categories with antifungal susceptibility revealed that strong biofilm producers, particularly among C. albicans and C. tropicalis, were more frequently associated with elevated amphotericin B MICs. This observation suggests that biofilm-associated structural or metabolic adaptations may contribute to reduced susceptibility to polyenes in these species. Biofilm-mediated sequestration of antifungal agents altered membrane sterol composition, and the activation of stress responses may collectively reduce the efficacy of amphotericin B⁴⁸. However, this relationship was not universal across species. Notably, Candidozyma auris demonstrated moderate to strong biofilm formation alongside consistently elevated MICs to amphotericin B and other antifungal classes, irrespective of biofilm category²³. This dissociation suggests that intrinsic resistance mechanisms, rather than biofilm-associated tolerance alone, play a dominant role in shaping antifungal response in this species. The ability of C. auris to combine biofilm formation with multidrug resistance likely contributes to its persistence in healthcare settings and underscores the challenges of its clinical management. The biological basis for the observed associations between virulence enzyme activity and amphotericin B susceptibility remains speculative. Potential explanations may include shared stress-response pathways, alterations in cell membrane composition, or adaptive phenotypic responses that influence both virulence expression and antifungal tolerance.

This study has some limitations as well. It was conducted at a single centre and did not incorporate patient-level clinical data or outcome measures, precluding assessment of the clinical impact of virulence phenotypes or antifungal resistance. Additionally, the correlation analyses were exploratory and not adjusted for multiple testing. Despite these limitations, the study provides valuable surveillance data from a large cohort of bloodstream isolates. Taken together, these findings indicate that biofilm formation, extracellular enzyme expression, and antifungal susceptibility are interconnected but species-dependent determinants of Candida pathogenicity. Strong biofilm formation, in combination with elevated virulence enzyme activity, may identify isolates with reduced susceptibility to amphotericin B, particularly among C. albicans and C. tropicalis. These results reinforce the importance of integrating biofilm assessment into virulence surveillance studies and highlight the continued relevance of echinocandins as first-line agents for invasive candidiasis in settings with a high burden of biofilm-associated and resistant Candida species.

Besides, the clinical implications of these findings are substantial. First, the high prevalence of NAC species and the observed variable susceptibility patterns strongly argue against empirical fluconazole monotherapy in settings with similar epidemiology. Echinocandins should remain the preferred empirical agents for suspected candidemia, particularly in critically ill patients. Second, the significant resistance observed in C. auris underscores the urgent need for routine antifungal susceptibility testing. Reliance on classical susceptibility assumptions is inadequate in the context of rapidly evolving resistance profiles. Laboratories should prioritize rapid identification and susceptibility testing for bloodstream Candida isolates to inform timely therapeutic adjustments. Third, the presence of C. auris has important implications for infection control. Its ability to persist on environmental surfaces, form biofilms, and cause healthcare-associated outbreaks necessitates strict adherence to infection prevention measures, including contact precautions, environmental decontamination, and active surveillance in high-risk units.

Conclusion

This large-scale surveillance study provides comprehensive insight into the evolving epidemiology, virulence characteristics, and antifungal susceptibility patterns of Candida bloodstream isolates from northern India. The predominance of non-albicans Candida, particularly C. tropicalis, alongside the emergence of multidrug-resistant C. auris reflects a significant shift with important therapeutic and infection-control implications. Our findings demonstrate that extracellular virulence enzyme production is not restricted to C. albicans but is widely expressed among several non-albicans Candida species. Importantly, significant correlations between elevated enzyme activity, particularly esterase and hemolysin, and increased amphotericin B MICs in C. albicans and C. tropicalis suggest that virulence traits may contribute to reduced antifungal susceptibility. These results support the concept that pathogenic potential and antifungal resistance are interconnected and species-specific. The high susceptibility of most isolates to echinocandins reinforces their role as first-line agents for treating candidemia; however, the substantial resistance observed in C. auris underscores the need for routine species-level identification and antifungal susceptibility testing. Continuous regional surveillance integrating virulence profiling and resistance monitoring is essential to inform antifungal stewardship strategies and optimize patient outcomes. Future studies incorporating molecular characterization of resistance mechanisms and virulence determinants will further enhance our understanding of invasive Candida infections.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Shabnam Kumari: Investigation, Formal analysis, Methodology, Data curation, Visualization, Writing- original draft. Zinnu Rain: Investigation, Data curation, Methodology, Writing-original draft. Pradyot Prakash: Resources, Writing-review & editing, Supervision. Ashish Kumar Singh: Conceptualization, Design, Supervision, Data curation, Visualization, Methodology, Formal analysis, Investigation, Software, Writing-review & editing. Ragini Tilak: Supervision, Conceptualization, Writing-review & editing, Validation, Resources.

Data availability

All data generated or analysed during this study are included in this article and its supplementary information files. However, minor data generated and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.