In vivo evolution to echinocandin resistance and increasing clonal heterogeneity in Candida auris during a difficult-to-control hospital outbreak, Italy, 2019 to 2022

A difficult-to-control outbreak of Candida auris is ongoing in a large tertiary care hospital in Liguria, Italy, where it first emerged in 2019. In a retrospective analysis, 503 cases of C. auris carriage or infection were observed between July 2019 and December 2022. Genomic surveillance identified putative cases that no longer occurred as part of one defined outbreak and the emergence of echinocandin (pandrug) resistance following independent selection of FKS1S639F and FKS1F635Y mutants upon prolonged exposure to caspofungin and/or anidulafungin.

A difficult-to-control outbreak of Candida auris is ongoing in a large tertiary care hospital in Liguria, Italy, where it first emerged in 2019. In a retrospective analysis, 503 cases of C. auris carriage or infection were observed between July 2019 and December 2022. Genomic surveillance identified putative cases that no longer occurred as part of one defined outbreak and the emergence of echinocandin (pandrug) resistance following independent selection of FKS1 S639F and FKS1 F635Y mutants upon prolonged exposure to caspofungin and/or anidulafungin.
Candida auris is an emerging multidrug-resistant (MDR) pathogenic yeast associated with outbreaks of invasive infections in healthcare settings worldwide, recently classified by the World Health Organization (WHO) as a pathogen of ‹critical priority' [1]. Over the past decade, C. auris emerged independently across the globe and has been reported in more than 45 countries on six continents. Consistent with this global trend, the European Centre for Disease Prevention and Control (ECDC) recently reported a marked increase of infection or carriage of C. auris in Europe during the period 2019 to 2021, with several countries experiencing sporadic detections or hospital outbreaks, such as Italy [2,3].
Here we report on the evolution of a difficult-tocontrol outbreak of C. auris ongoing since 2020 in a large tertiary care hospital in Italy, where genomic surveillance recognised the independent emergence of echinocandin-resistant genotypes and proved useful in identifying cases that no longer occurred as part of one defined outbreak.

Outbreak evolution
We performed a retrospective single-centre study investigating cases of C. auris at our hospital, hereafter referred to as HSM, a 1,200-bed teaching hospital representing the largest tertiary care facility in the region of Liguria, northern Italy, from 1 July 2019 to 31 December 2022. A case was defined as detection of C. auris from non-sterile and sterile body sites. Carriage was defined as the detection of C. auris from at least one non-sterile site (urine, skin and/or respiratory tract specimens) in the absence of clinical signs or symptoms of infection. Clinical isolates (infection) were those cultured to diagnose a disease state. Candidaemia was defined as illness in any patient who had C. auris isolated from at least one blood culture.
Following the first detection of a patient with candidemia and colonisation by C. auris in July 2019 at HSM [4], a subsequent increase in case numbers was recognised during 2020 in healthcare facilities in Liguria and in the neighbouring region of Emilia-Romagna [5]. Overall, at least 277 cases were reported in Liguria up to November 2021, mostly from our hospital, where multiple intensive care units (ICU) were affected. To contain the C. auris dissemination, a bundle of infection control interventions was implemented at HSM, including: (i) screening for skin carriage (combined axilla and groin skin swab) at admission to ICU for early identification of possible community-acquired cases; (ii) repeated weekly screening for carriage at skin, respiratory (whenever mechanically ventilated) and urine level during the ICU stay until first detection of C. auris; (iii) screening for skin carriage upon when  a C. auris-negative patient was discharged from the ICU and admitted to a different ward, with preventive contact precautions pending culture results; and (iv) implementation of strict contact precautions for colonised patients.
The retrospective analysis of all first-occurring episodes of C. auris carriage or infection observed at HSM from the first detection of C. auris in July 2019 [5] to December 2022 yielded 503 cases ( Figure 1A). Skin, respiratory or urine carriage was detected in 483 (96%) of the 503 cases, while first isolation from blood was registered in 20 (4%) cases ( Figure 1A). Among the 483 colonised patients, 85 (17.6%) subsequently developed candidaemia ( Figure 1B).
Previous characterisation of C. auris isolates (n = 10) collected during the early phases of the outbreak (i.e. following the sharp rise in colonisation and infection cases registered during the first wave of COVID-19 pandemic) revealed that all belonged to the clade I (South Asian) and showed a uniform high-level resistance to fluconazole and amphotericin B [6]. However, identification of caspofungin-resistant infection isolates with pandrug-resistant (PDR) phenotypes, and of cases without epidemiological links to those ICU mainly involved in the epidemic, revealed the dynamic nature and continued evolution of the outbreak. As such, additional phenotypic (antifungal susceptibility testing (AFST)) and molecular (whole genome sequencing (WGS)) investigations were carried out on 32 C. auris isolates, including: (i) two isolates resistant to echinocandins, (ii) seven isolates from patients with confirmed C. auris carriage/infection from samples collected within 0-24 h since hospital admission (i.e. suspected epidemiologically unrelated cases) and (iii) 23 isolates representative of the outbreak timespan ( Figure 1A). Details of the laboratory methods are appended in the Supplement.
When available, isolates cultured from a different specimen from the same patient were also characterised (n = 18) to allow for pairwise genomic comparison between echinocandin-susceptible and -resistant isolates and between surveillance and clinical isolates (mean separating days: 38 ± 40; median: 23; interquartile range (IQR): 12-64). Overall, 50 C. auris isolates were subjected to further investigation, together with 10 C. auris isolates previously characterised during the emergence of the outbreak [6].

Increased clonal heterogeneity among outbreak isolates
We performed a global core-genome single nucleotide polymorphism (SNP) phylogenetic analysis (see the Supplement for methodological details) including sequence data of C. auris isolates collected in the early (n = 10; 2019-2020) and late (n = 50; 2020-2022) stages of the outbreak recognised at HSM [6]. Among these 60 isolates from HSM, two were from patients initially diagnosed with C. auris infection or colonisation at two hospitals located in the same region (Liguria), who were later transferred to our facility. Sequence data of international isolates (n = 110) from validated outbreak benchmark datasets was also included in that analysis [7,8]. All outbreak isolates from HSM belonged to the clade I (South Asian), representing a monophyletic group (Ob4) within subclade Ic; outbreak cases from the United States (US) (Ob1-3) were reliably resolved within the same subclade ( Figure 2). In Supplementary Figure S1, we provide results of a phylogenetic analysis carried out with a global collection of C. auris genomes for the clade assignation.
Analysis of loci commonly involved in antifungal resistance revealed the presence of multiple alterations associated with the population structure at subclade level (e.g. CDR1 V704L , ERG11 K143R , TAC1B A 640V ). Furthermore, we identified a genetic signature uniquely represented in outbreak isolates from HSM (i.e. HMG1 P238H ). Screening for the HMG1 P238H allele using a comprehensive multinational dataset including 2,283 publicly available C. auris genomes revealed that this variant was detectable in a single record (i.e. excluding sequenced isolates from HSM), representing the first C. auris strain (RCPF-1821) isolated in Russia in 2017 [9]. Consistently, phylogenetic analysis resolved RCPF-1821 close to outbreak isolates from HSM ( Figure  2). The Supplement lists the inclusion criteria adopted to establish the genome dataset and related accession numbers.

Antifungal susceptibility profiles
We carried out AFST by broth microdilution following the Clinical and Laboratory Standards Institute guidelines and interpreted as per the US Centers for Disease Control's tentative breakpoints (see the Supplement for the microdilution method and interpretation criteria) [10]. All tested isolates were resistant to amphotericin B and fluconazole and showed low MIC values for all other tested azoles but voriconazole ( Table 1). The resistance profile for these antifungal agents was overall stable, consistent with susceptibility data obtained during the initial phase of the outbreak [6]. On the other hand, although most isolates retained susceptibility against echinocandins, evolution towards caspofungin resistance was observed in some cases, with a variable cross-resistance to micafungin and anidulafungin (Table 1).

In vivo evolution to echinocandin resistance
In late 2022, two cases of caspofungin-resistant C. auris candidaemia were observed. Both patients had previously tested positive for skin carriage and experienced prolonged exposure to echinocandins due to intraabdominal candidiasis after abdominal surgery ( Altogether, these findings suggested that echinocandin resistance resulted from antifungal pressure and emerged independently, following the selection of different FKS1 genotypes (Table 2).

Discussion
The ECDC reported in November 2022, that C. auris emerged progressively in Europe during 2020 and 2021, a consequence of sporadic (France, Germany) or multiple (Greece, Italy) outbreaks; regional endemicity was also reported (Spain) [2]. In most scenarios, however, available information uniquely relied on epidemiological data and lacked implementation of genomic surveillance to track the spread of C. auris and monitor evolution of the outbreaks. Here we have provided a genomic-informed snapshot of a large C. auris outbreak in Europe, gathering new insights about its clonal heterogeneity and in vivo evolution towards PDR phenotypes.
In Italy, the first C. auris outbreak was recognised in our facility during 2020 [4,6]. Despite the bundle of infection control interventions implemented at HSM to contain the C. auris dissemination, the epidemic proved difficult to control and additional cases were increasingly recognised throughout 2021 and 2022.
Overall, 503 C. auris cases were identified from July 2019 to December 2022, mostly contributed by carriage, with a total of 105 cases of candidaemia. The ability to preferentially colonise cutaneous surfaces is a distinguishing feature of C. auris, leading to an increased risk of intrahospital transmission and development of candidaemia, which is considerably enhanced by the simultaneous colonisation of skin and other sites (multiple site colonisation) [11].
Genomic surveillance traced the HSM outbreak mainly to the clonal expansion of a single lineage, without evidence of phylogeographical mixing of multiple clades, and proved fundamental in dissecting the clonal heterogeneity of C. auris. Indeed, we identified distinctive genetic signatures (e.g. CDR1 V704L , ERG11 K143R , TAC1B A64 0V , HMG1 P238H ) that altogether could help to molecularly trace the dissemination of this lineage at national or international level.
Determination of SNPs separating paired carriage and clinical isolates from the same patient proved useful to infer the baseline genetic diversity in this outbreak, strengthening data obtained from the few similar evaluations carried out on sporadic cases in Australia and on outbreak cases in the US [12,13]. Using the estimate of the baseline genetic diversity as a point of reference to assess transmissions, we inferred that genetically heterogeneous subpopulations were present and that at least some cases no longer occurred as part of a defined outbreak in our setting. These findings suggest that C. auris could be more widely disseminated than expected, as the transfer of patients colonised by C. auris between facilities may have been unreported. The most recent evidence shows that C. auris has been detected in four Italian regions: Liguria, Emilia Romagna, Piedmont and Veneto [14]; nevertheless, regions other than Liguria notified only a limited number of cases (n = 64) in the period from November 2021 to December 2022. Concerns about possible establishment of C. auris as a healthcare-associated pathogen in Europe, as in the epidemiological scenario in the US [15], have recently been raised by the ECDC [2].
It is of concern that we identified echinocandin-resistant isolates which evolved in vivo upon prolonged exposure to anidulafungin or anidulafungin and caspofungin, which led to the independent emergence of different FKS1 mutants. Unlike FKS1 S639Y , commonly observed in echinocandin-resistant C. auris that have so far been reported [16,17], FKS1 F635Y represents a novel genotype recently identified in two isolates in India [18]. These alterations accounted for a distinct echinocandin-resistance profile and were shown to contribute a differential response to echinocandin treatment in a murine model of disseminated infection, suggesting that the FKS1 genotype could be a more accurate predictor of treatment response than echinocandin MICs [18].
Our investigation has some limitations, since the absence of characterisation of isolates from healthcare facilities located in other regional areas prevented us from ruling out the introduction of multiple C. auris lineages, and from establishing representative genetic signatures of the clone(s) circulating in Italy.

Conclusion
The evolution towards PDR phenotypes calls for close monitoring of antifungal resistance in patients with prolonged exposure to echinocandins. Prompt implementation of genomic surveillance and antifungal stewardship programmes is critical to contain the selection and spread of PDR C. auris.

Ethical statement
The study was conducted according to the principles stated in the Declaration of Helsinki, and it was approved by the Ethics Committee of the Liguria region (N. CER Liguria 275/2022, 308/2022). Specific informed consent for this study was waived due to the retrospective nature of the analyses.

Funding statement
This study was supported by a research grant from Gilead Italia (Fellowship Program 2021 -grant no. 13372).

Data availability
Raw Illumina reads and draft genomes of C. auris isolates sequenced in this study have been deposited in NCBI databases under BioProject no. PRJNA655187.

Conflict of interest
Outside the submitted work, MB reports research grants and/or personal fees for advisor/consultant and/or speaker/ chairman from BioMérieux, Cidara, Gilead, Menarini, MSD, Pfizer, and Shionogi. Outside the submitted work, DRG reports investigator-initiated grants from Pfizer, Shionogi and Gilead Italia, and speaker/advisor fees from Pfizer and Tillotts Pharma. Outside the submitted work, AM reports investigator-initiated grant from Gilead Italia and research grant from DiaSorin. Outside the submitted work, VDP reports research grant from Seegene Inc. The other authors have no conflicts of interests to disclose.