First report of Candida auris in Guangdong, China: clinical and microbiological characteristics of 7 episodes of candidemia

ABSTRACT

Candida auris is an emerging multidrug-resistant fungal pathogen worldwide. To date, it has not been reported in Guangdong, China. For the first time, we reported 7 cases of C. auris candidemia from two hospitals in Guangdong. The clinical and microbiological characteristics of these cases were investigated carefully. Two geographic clades, i.e. III and I, were found popular in different hospitals by whole genome sequencing analyses. All C. auris isolates from bloodstream were resistant to fluconazole, 5 of which belonged to Clade III harbouring VF125AL mutation in the ERG11 gene. The isolates with Clade I presented Y132F mutation in the ERG11 gene as well as resistance to amphotericin B. All isolates exhibited strong biofilm-forming capacity and non-aggregative phenotype. The mean time from admission to onset of C. auris candidemia was 39.4 days (range: 12 - 80 days). Despite performing appropriate therapeutic regimen, 42.9% (3/7) of patients experienced occurrences of C. auris candidemia and colonization after the first positive bloodstream. C. auris colonization was still observed after the first C. auris candidemia for 81 days in some patient. Microbiologic eradication from bloodstream was achieved in 85.7% (6/7) of patients at discharge. In conclusion, this study offers a crucial insight into unravelling the multiple origins of C. auris in Guangdong, highlighting great challenges in clinical prevention and control.

KEYWORDS:

• Candida auris
• candidemia
• whole-genome sequencing
• resistance
• outcome

Introduction

As a new pathogenic fungus, Candida auris has become a worldwide public health threat in recent years due to its multi-drug resistance and high outbreak potential, in which the bloodstream, the central nervous system, and internal organ gets affected [1, 2].

Studies have shown that candidemia is the most frequently reported infection caused by C. auris, with a crude mortality of 45% [3–5]. Along with the increasing incidence of C. auris infections, it has become a major bloodstream pathogen, even having a trend to surpass C. albicans in some healthcare facilities [6, 7].

Five geographic clades of C. auris were identified by their geographical origin to date, including Clade I (South Asia), Clade II (East Asia), Clade III (South Africa), Clade IV (South America), and Clade V (Iran) [8, 9]. These clades were found to differ from each other by thousands of single-nucleotide polymorphisms through whole genome sequencing (WGS) analyses [9, 10]. Studies showed that C. auris isolates of Clade I, III, and IV were mainly associated with invasive infections and outbreaks, while Clade II isolates presented a predilection for the ear [11].

Globally, more than 90% of C. auris were resistant to fluconazole, mostly due to mutations in the ERG11 gene [10, 12–14]. The susceptibilities of C. auris strains varied to other azoles, echinocandins, and amphotericin B, which were mainly attributed to geographical differences [11]. With the increasing resistance to azoles and amphotericin B, echinocandins are becoming the first-line therapy for C. auris infection [8]. However, treatment failures are common and usually associated with the development of resistance to echinocandins, mainly due to mutations in the FKS1 gene [11].

Given that geographical clonal strains differ widely and behave distinctly, exerting different pressures on clinical treatment and prevention, it becomes imperative to thoroughly investigate the local emergence of C. auris. Since the first C. auris isolate in China was reported in 2018, less than 100 cases of C. auris infection have been reported to date [12, 15]. Both clinical and basic studies of C. auris still need more attentions in China. In this study, we reported the first cases of C. auris candidemia in Guangdong, China. Furthermore, the clinical and microbiological characteristics of these cases were investigated carefully.

Materials and methods

Clinical data and isolates

Between January 2023 and July 2023, seven patients with C. auris candidemia were investigated from two hospitals in Guangdong. Patient information and 7 isolates from the first positive blood culture were collected. This study was approved by the institutional review board of the First Affiliated Hospital of Sun Yat-sen University.

C. auris identification

All isolates were identified as C. auris by matrix-assisted laser desorption ionization time of flight mass spectrometry (MALDI-TOF MS) (bioMérieux, France) with the IVD knowledge base V3.2 following the manufacturer’s instructions. They were further confirmed by DNA sequencing with ITS1/ITS4 primers as previously described [6]. The primers were ITS1 (5’-TCCGTAGGTGAACCTGCGG-3’) and ITS4 (5’- TCCTCCGCTTATTGATATGC-3’), respectively. The sequences were deposited in GenBank (OR708706 - OR708712).

WGS and phylogenetic analysis

WGS and phylogenetic analysis of each isolate was performed as previously described with some modifications [16].

Briefly, genomic DNA was extracted from samples using the Yeast Genomic DNA Rapid Extraction Kit (Sangon Biotech, China) according to the manufacturer’s instructions. Libraries were constructed for the DNA samples using a Nextera XT DNA Library Prep Kit (Illumina Inc., USA). Library quality was assessed by Qubit dsDNA HS Assay kit and High Sensitivity DNA kit (Agilent, USA) on an Agilent 2100 Bioanalyzer. Library pools were then sequenced on an Illumina NextSeq 550 system (Illumina Inc., USA) using a 75 bp, single-end sequencing kit. The quality of reads was assessed by Fastqc (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/). The sequence read was trimmed by Trim Galore v0.6.0 (http://www.bioinformatics.babraham.ac.uk/projects/trim_galore/), assembled de novo using SPAdes v.3.9.0, and assessed by QUAST v5.0.2. Contigs of smaller size (<1000 bp) were excluded.

Single nucleotide polymorphism (SNP) analysis was performed using Snippy v4.6.0 pipeline (https://github.com/tseemann/snippy). C. auris strain B11220 (GenBank accession no. GCA_003013715.2) was selected as the reference genome, and 66 sequenced C. auris isolates from each of the five established clades were added to the analysis. A maximum likelihood phylogeny was constructed from the SNP data using IQ-TREE 2.2.2.7 with a TVMe + ASC + R3 model. The ultrafast bootstrap value was set to 1000.

The whole-genome sequencing project was deposited in the NCBI Sequence Read Archives (SRA) and GenBank under accession no. PRJNA1025033.

In vitro antifungal susceptibility testing

Nine commonly antifungal agents (Shanghai Aladdin Bio-Chem Technology Co., Ltd., China) were included. The broth microdilution was performed according to document M27-A3 of the Clinical and Laboratory Standards Institute (CLSI) [17].

To date, there are no species-specific antifungal susceptibility breakpoints established for C. auris. The antifungal resistance for C. auris was analyzed according to the tentative minimum inhibitory concentrations (MIC) breakpoints published by the CDC (https://www.cdc.gov/fungal/candida-auris/c-auris-antifungal.html). The strains of C. parapsilosis (ATCC 22019) and C. krusei (ATCC 6258) were used as quality controls.

Sequencing of resistance-associated mutations

The ERG11 gene and FKS1 gene of C. auris were further investigated using a modified protocol previously described by AlJindan et al [18]. The sequence of ERG11 gene was amplified using the primers ERG11aF (5’-ATGGCCTTGAAGGACTGCATCGT-3’) and ERG11aR (5’-TTAGTAAACACAAGTCTCTCTTTTCTCCCA-3’). FKS1 gene was amplified using forward (5’-ATGTCTTACGATAACAATCACAACTAC-3’), and reverse primers (5’-AGTAAGATTCGGCCAACTTAGCAG-3’).

PCR amplification was conducted in a 50-µL reaction mixture containing 5 µL 10 × PCR buffer, 5 µL templates, 1 µL forward primer, 1 µL reverse primer, 0.5 µL Taq enzyme, 4 µL dNTP mixture, and 33.5 µL double-distilled water. The PCR amplification condition was as follows: 1 cycle of 94°C 5 min; 35 cycles of 94°C 30 s, 62°C (ERG11 gene) / 57°C (FKS1 gene) 30 s, 72°C 1 min; 1 cycle of 72°C 7 min. The PCR products were subjected to Sanger sequencing in two directions (Sangon Biotech, China). Sequencing data were analyzed using the Mega 7.0 software.

The sequences were deposited in GenBank (OR493453 - OR493466).

Biofilm formation

XTT (2,3-bis(2-methoxy-4-nitro-5-sulfo-phenyl)-2H-tetrazolium-5-caboxanilide) reduction assay was used to assess the biofilm-forming capacity of each isolate as previously described with some modifications [19, 20].

Briefly, all isolates were cultured on sabouraud dextrose agar (SDA) plates for 48 hours (h) at 28°C to obtain single colonies. Then, these colonies were standardized to 1 × 10⁶ cells/mL in RPMI-1640 culture medium, 200 μL of which were added to each well of a 96-well flat-bottomed microtiter plate (Corning, USA) and incubated for 24 h at 37°C. After washing twice carefully with PBS, the biofilm-coated wells were air-dried and 150 μL of XTT (Sigma, USA) - coenzyme Q₀ (MedChemExpress, USA) solution (final concentration: 0.25 mg/mL XTT and 40 μg/mL coenzyme Q₀) were added and incubated for 3 h at 37°C. Then, 100 μL of the solution was transferred to a new 96-well plate, and the optical density (OD) was measured at 450 nm using a spectrophotometer (Siemens, German).

Both C. albicans (ATCC 14053) and C. parapsilosis (ATCC 22019) were used as quality controls. RPMI-1640 medium without cells served as blank control. Each isolate was tested in five independent experiments.

Determination of aggregation behaviour

Aggregate-forming capacity was assessed as previously described [21]. Briefly, C. auris was cultured on SDA plate for 24 h at 37°C. A single colony was emulsified in 20 μL of sterile water on a microscope slide, and then the entire specimen was examined under the microscope at a magnification of × 400. The isolates that primarily exhibited large, non-dispersible cell clusters were categorized as aggregative isolates [21, 22].

Statistical analyses

Analysis of variance test was used for statistical analysis of biofilm-forming capacity.

Results

Clinical characteristics

Seven patients with C. auris candidemia were studied (Table 1). There were four males and three females with a median age of 59.6 (range: 40 - 77) years, 71.4% (5/7) of which were from ICU. Case 7 was the only patient without a comatose condition during hospitalization. All patients had hypertension, urinary catheterization, and recent intake of broad-spectrum antibiotics. The dosage and number of days for antibiotics before C. auris candidemia were listed in Table S1.

Table 1. Clinical characteristics of 7 cases with C. auris candidemia.

Case	1	2	3	4	5	6	7
Hospital/Department^a	A/ICU	A/ICU	A/ICU	A/NR	A/ICU	B/ICU	B/HT
Sex (yr)^b/Age	F/76	M/59	M/77	F/52	M/65	M/48	F/40
First isolation of C. auris from bloodstream
Interval time after admission (days)	24	12	80	58	42	29	31
Colonization before candidemia	Yes	No	Yes	Yes	Yes	No	No
Infection before candidemia	No	No	No	No	No	No	No
Clinical status
Immunosuppressive state	Yes	No	Yes	Yes	No	No	Yes
Renal impairment	No	No	No	Yes	No	No	No
Diabetes mellitus	No	No	No	Yes	No	No	Yes
Hypertension	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Cancer	Yes	No	No	No	No	No	Yes
Coma	Yes	Yes	Yes	Yes	Yes	Yes	No
Invasive procedures
Peripherally inserted central catheters	Yes	No	Yes	Yes	Yes	Yes	No
Central venous catheters	No	No	No	No	No	No	No
Urinary catheterization	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Parenteral nutrition	No	No	No	No	No	No	No
Arterial catheters	No	No	No	No	No	No	No
Surgery within 30 days	Yes	No	No	No	Yes	No	No
Recent intake of broad-spectrum antibiotics	Yes	Yes	Yes	Yes	Yes	Yes	Yes

^a ICU: Intensive care unit, NR: Neurology department, HT: Hematology department; ^bF: female, M: male.

The mean time from admission to onset of C. auris candidemia was 39.4 days, ranging from 12 to 80 days. It was found that 57.1% (4/7) of patients had C. auris colonization before candidemia. All patients had at least one invasive procedure, and the majority (71.4%, 5/7) had peripherally inserted central catheter (PICC) catheterization.

Phylogenetic analysis

The draft assemblies of these 7 isolates varied in length from 12.3–14.5 Mb, with a mean ± standard deviation scaffold count of 3787 ± 1261, GC content of 45.4% ± 0.2%, scaffold N₅₀ value of 12.4 ± 0.7 kb, and coverage depth of 102-fold ± 23-fold for quality-trimmed reads (Table 2). The accession numbers for the assembly and raw reads for individual isolates are presented in Table 2.

Table 2. Characteristics and accession numbers of whole genomes of C. auris in the present study.

Isolate	Isolation source	Genome size (Mb)	No. of reads	No. of contigs	No. of scaffolds	Scaffold N50 (bp)	Scaffold GC content (%)	Genome accession no.	BioSample accession no.	SRA accession no.
Case 1	Blood	13.5	18055561	4248	4248	12527	45.59	JAWIVU000000000	SAMN37711175	SRR26315287
Case 2	Blood	12.7	11431255	2923	2923	12902	45.51	JAWIVV000000000	SAMN37711176	SRR26315286
Case 3	Blood	12.8	16403973	2812	2812	13051	45.54	JAWIVW000000000	SAMN37711177	SRR26315285
Case 4	Blood	14.5	23411900	5569	5569	11327	45.34	JAWIVX000000000	SAMN37711178	SRR26315284
Case 5	Blood	13.3	16438242	3982	3982	12527	45.67	JAWIVY000000000	SAMN37711179	SRR26315283
Case 6	Blood	14.2	26517252	4941	4941	11588	45.09	JAWIVZ000000000	SAMN37711180	SRR26315282
Case 7	Blood	12.3	15623713	2031	2031	12950	45.23	JAWIWA000000000	SAMN37711181	SRR26315281

Two geographic clades were found through phylogenetic analysis of WGS (Figure 1A). Five isolates from Hospital A were identified as strains of Clade III (South Africa). Both isolates from Hospital B belonged to Clade I (South Asia).

Figure 1. Phylogenetic analysis of 7 Candida auris isolates originating from bloodstream. (A) Phylogenetic tree showing the genetic relationships among isolates representing five distinct clades. The isolates from 7 patients were highlighted in blue. (B) Analysis of the pairwise SNP differences between different isolates in this study.

Table S2 listed a series of variants identified from WGS among these isolates. For Clade III, 31 SNPs among the 5 isolates (Cases 1-5) were detected. The number of SNP difference between Case 1 and Case 3 was 1 (Figure 1B). The maximum number of pairwise SNP differences between isolates with Clade III was 25, suggesting a high degree of genetic relatedness (Figure 1B). For Clade I (Cases 6-7), the number of pairwise SNP differences between isolates was 7 (Figure 1B).

Further, we analyzed the genetic relationships between the isolates in this study and some isolates reported in China. As shown in Figure 2, the 5 strains with Clade III were obviously related to the first strain of GZ0008 isolated from the same hospital. And both strains from Hospital B were closely related to the strain BJCA003 from Beijing.

Figure 2. Phylogenetic tree showing the genetic relationships between Candida auris isolates in this study and some isolates reported in China. The isolates from 7 patients were highlighted in bold. The GZ0008|Guangzhou (Genome accession no. was JAWWNA000000000) was the first strain isolated from the Hospital A in Guangdong.

Antifungal susceptibility

The antifungal susceptibility results of C. auris isolates are shown in Table 3. According to the tentative MIC breakpoints, all isolates were resistant to fluconazole (MIC range: ≥ 64 μg/mL) and sensitive to echinocandins (MIC range: 0.06 - 0.25 μg/mL), respectively. Notably, both isolates with Clade I were resistant to amphotericin B with a MIC of 2 μg/mL.

Table 3. In vitro antifungal susceptibility profile and main resistant mutants of C. auris isolates.

	Case
	1	2	3	4	5	6	7
Agents MIC (μg/mL)
FLC	≥64	≥64	≥64	≥64	≥64	≥64	≥64
VRC	0.25	0.5	0.5	0.5	0.5	0.25	0.25
ITC	0.25	0.25	0.25	0.25	0.25	0.25	0.25
POS	0.12	0.12	0.12	0.12	0.12	0.12	0.06
CAS	0.12	0.12	0.12	0.12	0.06	0.12	0.12
ANF	0.12	0.12	0.25	0.25	0.12	0.12	0.12
MCF	0.12	0.12	0.12	0.12	0.06	0.06	0.06
FCT	0.03	0.06	0.06	0.03	0.06	0.12	0.12
AMB	1	1	1	1	1	2	2
Resistance gene/GenBank No.
ERG11	VF125AL/OR493453	VF125AL/OR493454	VF125AL/OR493455	VF125AL/OR493456	VF125AL/OR493457	Y132F/OR493458	Y132F/OR493459
FKS1	ND/OR493460	ND/OR493461	ND/OR493462	ND/OR493463	ND/OR493464	ND/OR493465	ND/OR493466

MIC, minimal inhibitory concentration; FLC, fluconazole; VRC, voriconazole; ITC, itraconazole; POS, posaconazole; CAS, caspofungin; ANF, anidulafungin; MCF, micafungin; FCT, 5-flucytosine; AMB, amphotericin B; ND: non-detected.

For other triazoles, posaconazole (MIC range: 0.06 - 0.12 μg/mL) presented lower MICs than voriconazole (MIC range: 0.25 - 0.5 μg/mL). In addition, the MICs of C. auris to 5-flucytosine ranged from 0.03–0.12 μg/mL.

Analysis of resistance-associated mutations

All isolates were found to have fluconazole-resistant associated mutation in the ERG11 region (Table 3). Five isolates with Clade III contained the VF125AL mutation, while the other 2 isolates with Clade I contained the Y132F mutation.

No mutation within the FKS1 region was detected in any of these echinocandins-susceptible isolates.

Biofilm-forming capacity and aggregation behaviour

As shown in Figure 3A, each isolate exhibited strong biofilm-forming capacity through XTT reduction assay. The biofilm formation of 57.1% (4/7) of C. auris isolates was comparable to that of C. albicans (P > 0.05). In comparison to C. parapsilosis, 57.1% (4/7) of C. auris isolates showed higher biofilm-forming capacities (P < 0.05). No significant difference in biofilm-forming capacity was observed between isolates of Clade I and III (P = 0.8384).

Figure 3. The biofilm-forming capacity and aggregation behaviour of Candida auris isolates originating from bloodstream. (A) The biofilm-forming capacity of each isolate detected by XTT reduction assay. OD₄₅₀: the optical density at 450 nm; QC1: Candida albicans (ATCC 14053); QC2: Candida parapsilosis (ATCC 22019); NC: the negative control. (B) The aggregation capacity assessed by aqueous suspension on sterile microscopy slides.

In addition, all 7 isolates exhibited growth as individual, budding yeast cells, displaying a non-aggregative phenotype (Figure 3B).

Therapy and outcome

Prior to the onset of C. auris candidemia, 42.9% (3/7) of patients had been treated with antifungal drugs, including caspofungin (50 mg QD IV) and triazoles (200 mg QD PO) (Figure 4). For Case 3, C. auris candidemia occurred after discontinuing caspofungin for 7 days.

Figure 4. Treatment regimens and outcomes for 7 patients with Candida auris candidemia. Candidemia was considered in patients with positive blood culture of Candida spp. C. auris colonization was considered in patients with C. auris strain isolated from specimens without leading to clinical infection. NA: not available; Gray bars: period of hospital stay, with numbers denoting hospital days.

Echinocandins were administered as the main antifungal drugs for C. auris candidemia among all patients except for Case 7 (Figure 4). The patient in Case 7 was given a 27-day therapy of voriconazole (200 mg QD PO) with a MIC of 0.5 μg/mL. Two patients died due to severe underlying diseases.

Despite performing appropriate fungal therapy based on susceptibility testing results, 42.9% (3/7) of patients experienced a recurrence of C. auris candidemia and colonization after the first C. auris candidemia. For Case 4 and Case 5, C. auris colonization was still observed for 29 and 81 days after the first C. auris candidemia, respectively. Except for Case 6, 85.7% (6/7) of patients attained microbiologic eradication from bloodstream at discharge.

Discussion

Studies have shown that C. auris spreads rapidly among susceptible patients once it is introduced into a healthcare facility [11]. Since its first description in Japan in 2009, C. auris infections have been reported in over 40 countries across six continents, with several nosocomial outbreaks occurring in different countries [12]. However, this fungus has not been reported to date in Guangdong Province, China. According to the Guangdong Fungal Disease Surveillance Network, C. auris has been isolated from clinical specimens successively beginning from 2022 in Guangdong. Remarkably, C. auris candidemia has been appeared continuously in two hospitals since 2023. Take Hospital A for example, the first C. auris strain was isolated from a patient with a history of hospitalization in South Africa where C. auris with Clade III was found to have a dramatic increase in recent years [23]. The five strains from Hospital A all belonged to Clade III, the same as the first isolation (Genome accession no. JAWWNA000000000). The interval was 149 days between the first isolation and the first candidemia. Prior to the hospital visit, the 5 patients had neither any history of overseas travel nor visited any cities where C. auris had been reported in China [12]. The maximum number of pairwise SNP differences between isolates with Clade III was 25 (Figure 1B). As shown in Figure 2, they were obviously related to the first strain. We speculated the five strains of C. auris belonged to the hospital-acquired transmission, originating from the first strain.

In this study, both isolates (28.6%, 2/7) from Hospital B belonged to Clade I. To date, only two isolates with Clade I have been reported in mainland China, both of which were from Beijing [15, 24]. Similar to the strain of BJCA003 recently reported by Xu et al. [24], both isolates with Clade I presented Y132F mutation in the ERG11 gene and resistance to amphotericin B (Table 3). Phylogenetic analysis also showed that both isolates were closely related to the BJCA003 (Figure 2). Chow et al.’s study showed that a genetic distance of ≤ 12 SNPs between patients was indicative of recent transmission [25]. The number of pairwise SNP differences between Clade I isolates was 7 (Figure 1B), suggesting they originated from the same cloning strain. However, clinical data showed that before admission to Hospital B, both patients (Case 6 and Case 7) had C. auris colonization in other hospitals in Guangdong. Our study suggested a complex original source of C. auris might exist in Guangdong. In further studies, we will investigate the epidemiological relations of C. auris strains collected from different hospitals in Guangdong.

Several risk factors were thought to be associated with invasive C. auris infection, including hospital stay longer than 10 - 15 days, renal impairment, use of mechanical ventilation, central venous catheterization, total parenteral nutrition, sepsis, and previous use of antifungal medicines [1]. In this study, the mean time from admission to onset of C. auris candidemia (39 days) was similar to the findings in some studies (31 - 38 days) [26–29], while it was longer than those reported by other series (18 - 27 days) [7, 30–33].

Remarkably, there was no complete agreement regarding the risk factors of C. auris candidemia among different studies. For ICU patients with C. auris colonization, Garcia-Bustos et al. [34] reported several independent predictors of C. auris candidemia, including total parenteral nutrition, sepsis, central venous catheters, arterial catheters, renal impairment, previous surgery, previous exposure to antifungal agents, and multifocal colonization, but only multifocal colonization was found to be an independent risk factor in Briano’s study [35]. Compared to other non-auris candidemia cases, Rudramurthy et al. [36] found that longer stay in ICU, underlying respiratory illness, surgery, medical intervention, and antifungal exposure were more common in those of C. auris candidemia; however, factors such as parenteral nutrition and recent surgery were significantly less common in the C. auris group in Simon’s study [29]. Such a difference among these studies might be attributed to the limited amount of case reports and information, as well as different distributions of baseline characteristics in different populations. Table S3 listed major clinical characteristics of cases of C. auris candidemia reported in the last decade [7, 24, 26–34, 36–45]. As shown in Table S3, the most commonly clinical characteristics included ICU stay, underlying medical comorbidities, invasive procedure, and recent intake of broad-spectrum antibiotics, which were also observed in most of the patients here. Interestingly, we noted all the patients suffered from hypertension rather than diabetes mellitus. Vaseghi et al. found hypertension was the most prevalent comorbidity in coronavirus-associated C. auris infection [46]. Further studies are needed to determine whether hypertension plays a role in C. auris candidemia.

The data on C. auris colonization might be significantly underestimated due to the limitation of available conventional methods. On SDA plates, the colonies of C. auris were white to cream-coloured and smooth, which might be misidentified as other Candida spp. such as C. parapsilosis [47]. And C. auris isolates were misidentified mostly as C. haemulonii or Rhodotorula glutinis, by automated yeast identification systems such as Vitek2 [11]. It might be the reason why C. auris had not been detected before candidemia among patients (Case 6 and Case 7) with an earlier history of C. auris colonization. On the other hand, its difficult identification in the clinical laboratory, leading to erroneous diagnosis and inappropriate medication, was thought to favour the development of resistance acquisition to multiple drugs [48, 49]. A series of studies showed that C. auris infections were more commonly detected in patients receiving antifungal therapy, which added to the evidence that the emergence of C. auris was linked to alterations in treatment pressure [50, 51].

In line with high resistance rates of C. auris to fluconazole [12], all isolates here were resistant to fluconazole with MICs ≥ 64 μg/mL. Two common mutations in the ERG11 gene were found to be associated with geographic clades: VF125AL with Clade III, Y132F with Clade I, IV and V [10, 14]. Our data presented similar results. The resistance rates of C. auris to amphotericin B varied significantly due to geographical differences [8]. Chow et al. reported 33.3% (33/99) of C. auris strains resistant to amphotericin B in the United States [25]. To date, the majority of C. auris strains in China exhibited low MICs for amphotericin B [12]. Tian et al. reported that the resistance rate was 1.1% (1/93) for amphotericin B in Shenyang, China [52]. In this study, both isolates with Clade I were resistant to amphotericin B, showing higher MICs compared to those with Clade III.

C. auris is notorious for its persistence in the environment due to some factors such as biofilm formation, thermoresistant capacity and partial resistance to commonly used disinfectants [1, 5]. Detecting biofilm production is important since biofilms play an important role in catheter-associated candidemia [53]. Invasive candidiasis, especially catheter-associated candidemia, is thought to originate after dispersion and subsequent hematogenous dissemination from a biofilm formed inside a catheter [54]. We found all strains had strong biofilm-forming capacity. On the other hand, biofilms can form a physical barrier against antifungal agents and disinfectants, contributing to the persistence of the infection [5]. Simon et al. reported that patients with C. auris candidemia had a high microbiologic recurrence [29]. Here, 42.9% (3/7) of patients recurred C. auris candidemia and colonization despite performing appropriate fungal therapy. It was hard to achieve C. auris decolonization in some patients, suggesting a high risk of recurrent infection.

A strong association was noted between the aggregation behaviour and the biofilm-forming capacity of C. auris. In some studies, non-aggregating isolates were considered to exhibit a higher biofilm-forming capacity and virulence in comparison with aggregate-forming isolates [20, 22]. To note, most of clinical isolates were found to exhibit a non-aggregative phenotype while the preponderance of aggregative phenotype was observed in the colonizing isolates [55]. In this study, all the isolates presented not only strong biofilm-forming capacity but also non-aggregative phenotype.

Several limitations nevertheless deserve mention. Firstly, limited by the small sample size, our data cannot provide strong evidence for predicting the risk factors of C. auris candidemia. Secondly, the transmission patterns among these patients remain uncertain due to the lack of surveillance cultures for C. auris, including different body sites of the patients, hospital environment and healthcare staff. Thirdly, WGS analyses remain relatively preliminary since these sequences were mainly employed for investigating geographic clades and the possible source of C. auris in this study. Lastly, the identification of Candida spp. is still conducted using the Vitek 2 YST system in some laboratories of primary health care institutions, which might underestimate the prevalence of C. auris candidemia in Guangdong. And we did not include patients with other invasive Candida infections, culture-negative sepsis, or colonization, limiting our awareness of the severity of C. auris infection.

In summary, our study was the first to reveal the existence of C. auris in Guangdong, China. Clinical characteristics and strains of C. auris candidemia collected from two hospitals were thoroughly investigated. At least two geographic clades of fluconazole-resistant C. auris, namely III and I, were prevalent in this region. Moreover, it was hard to completely achieve C. auris decolonization despite performing appropriate therapeutic regimen. Albeit just the tip of the iceberg, our study puts an insight into the epidemiology of C. auris in this region. In the future, we will pay more attention to the prevention of subsequent, widespread transmission and the possible emergence of resistance upon treatment.

Acknowledgments

We acknowledge Vision Medicals Co., Ltd., Guangzhou, China for providing technical guidance in this study.

Disclosure statement

No potential conflict of interest was reported by the author(s).