Study Setting and Participants

The Burn Center is a 10-bed intensive care unit caring for pediatric and adult burn patients at a 547-bed academic tertiary care medical center in the Chicago metropolitan area, Illinois, USA. The unit accommodates ICU overflow from other services, including medical and surgical ICUs. The unit practices universal contact precautions (gowns, gloves, masks, and eye protection) for all patients, staff, and visitors to the unit. We abstracted patient data via retrospective review of the hospital electronic medical records.

The Institutional Review Board of Loyola University (Chicago, IL, USA) reviewed and approved the protocol for this study (LU218571). Informed consent was waived.

Case Identification and Investigation

The outbreak investigation, led by the infection prevention team, consisted of admission screening and weekly point prevalence surveys of all patients in the unit. We defined a hospital-acquired case of C. auris as any illness in patient who, after a negative C. auris admission screen, tested positive for C. auris on subsequent weekly point prevalence screens or in any clinical specimen. We defined colonized cases as patients who had C. auris identified from surveillance cultures but no detection of C. auris in any clinical specimens. Clinical cultures refer to blood, wound, respiratory, or urine cultures.

We conducted epidemiologic investigations to identify commonalities between cases, including healthcare workers, medical equipment, prior room occupancies, and exposure locations outside of the BICU, including the operating room, tub room, and procedural areas such as the interventional radiology and gastroenterology suites. We reviewed patients’ history of C. auris through query of the Illinois extensively drug-resistant organism registry (34).

Infection Control Measures

Universal contact precautions and masks are used for all BICU patients; further containment strategies implemented in response to this outbreak involved increased observation of isolation compliance, education of nursing and ancillary staff about C. auris transmission and control, environmental cleaning validation, enhanced environmental cleaning with ultraviolet (UV) light, observation and training regarding correct use of PPE, and proper hand hygiene. In addition, the local health department performed an infection control assessment and response to identify and address infection control gaps.

Microbiologic Identification of Cases

We isolated C. auris fungus from screening samples collected from the axilla and groin of patients using a BBL CultureSwab EZ Collection and Transport System (BD, https://www.bd.com). We then inoculated samples onto HardyCHROM Candida (Hardy Diagnostics, https://hardydiagnostics.com) and incubated aerobically, protected from light, at 35°C for 72 hours. We isolated C. auris from clinical samples submitted for routine diagnostic testing on standard microbiologic media including sheep blood agar, chocolate agar, inhibitory mold agar (BBL prepared plated media; BD), and blood culture media (BACTEC Plus Aerobic and Lytic Anaerobic media; BD). We performed species identification by using Biotyper matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry with the MBT Compass Library version 12 Revision K (Bruker Daltonics, https://bruker.com).

Whole-Genome Sequencing

We suspended available C. auris isolates in DNA/RNAshield (Zymo, https://zymoresearch.com) and transported them to the Regional Innovative Public Health Laboratory at Rush University Medical Center (Chicago, IL, USA). We extracted nucleic acids using the Cultured Cells DNA Kit and Maxwell extraction system (Promega, https://promega.com) and prepared sequencing libraries using 1 ng DNA extract and Nextera XT DNA Library Preparation Kit (Illumina, https://illumina.com). We barcoded genome libraries by using IDT for Illumina DNA/RNA UD Indexes (Illumina) and balanced using a small-scale sequencing run of an equivolume pool (Illumina iSeq). We subjected final libraries to 2 × 150 paired-end sequencing on NovaSeq6000 (Illumina). We submitted data to the National Center for Biotechnology Information Short Read Archive (Appendix Table 1).

Bioinformatic and Statistical Analysis

We downloaded all publicly available Illinois C. auris sequences for comparison to outbreak sequences (Appendix Table 1). We analyzed paired-end sequences with the MycoSNP-nf pipeline version 1.4 (https://github.com/CDCgov/mycosnp-nf) by using clade IV reference B11243 (Genbank accession no. GCA_003014415.1) and implemented on Terra as previously described (20,27,28), excluding isolates with estimated coverage depth <25. We determined C. auris clade using phylogenetics with clade I–IV reference sequences. We used SNP differences between all samples and the reference to build a neighbor-joining tree using MEGA 11 as previously described (29) with 1,000 bootstrap replicates (21,30). To compare SNP differences, we used SNP distance matrices with all available Illinois sequences. To identify potential antifungal-resistance mutations we used Snippy (31) to query for mutations in the FKS1, ERG11, TAC1b, MRR1, ERG3, and FUR1 genes. We compared mean SNP differences using Kruskal-Wallis nonparametric testing with adjusted significance between individual groups calculated using the Dunn multiple comparisons test.

Outbreak Investigation

Figure 1

Figure 1. Epidemiologic curve of Candida aurisoutbreak cases in burn intensive care unit, Illinois, USA, 2021–2023. Color indicates whether case was identified by screening or clinical isolates. PPS, point prevalence...

The first clinical C. auris isolate in a BICU patient was identified from blood culture in 2021. Three additional case-patients with hospital-acquired C. auris were identified in the subsequent 2 months (Figure 1). Admission screening in the BICU was initiated 3 months after the first case-patient was detected (Figure 1). A fifth case-patient, who originally screened negative on admission, was identified from a wound culture 83 days after admission; that case was notable because it was the first confirmed hospital acquisition of C. auris. A point prevalence survey 5 months after first case detection identified a sixth case. Weekly point prevalence screening was initiated 6 months after first case detection; 22 additional cases were identified 6–21 months after first case identification (Figure 1). Weekly point prevalence surveys were discontinued 28 days after discharge of the last patient with C. auris.

We reviewed case records to identify documented epidemiologic links; specifically, common locations (rooms), procedures, and staff exposures. Intensive observation of infection control practices throughout the BICU identified breaches that may have contributed to C. auris transmission, including poor hand hygiene compliance, improper PPE donning and doffing, cluttered patient care areas preventing thorough environmental cleaning, poor auditing of environmental cleaning, and inconsistent cleaning and disinfection practices for shared equipment. Shared equipment within the unit included bladder scanners, forced-air patient warming devices, vascular Dopplers, EKG machines, point-of-care ultrasounds, and recliners. Particular attention was given to staff whose patient care activities were extended to areas in the medical center outside of the BICU, including physical, occupational, speech and respiratory therapy, and radiology staff. Infection control breaches were observed by the hospital infection control team and by the local health department during the infection control assessment performed 6 months after the first case of C. auris was detected.

Outbreak Mitigation Measures

Early in the outbreak, a multidisciplinary C. auris response team, including staff from infection prevention, environmental services, nursing, facilities management, BICU physicians and hospital leadership, convened to create and implement a structured plan. All patients in the BICU with positive C. auris culture results were placed on contact precautions; signs were placed on the patients’ room doors, and their electronic medical records were flagged for C. auris and an isolation order. The team implemented outbreak mitigation measures universally in the BICU and centered on communication, education, and process improvement, focusing on environmental cleaning and hand hygiene. Education on C. auris transmission and necessary precautions were extended to the BICU nursing staff, with special attention on ancillary groups, particularly those also providing care to units outside of the BICU: environmental service, respiratory therapy, physical and occupational therapy, food and nutrition services, radiology, and pastoral care.

The team reviewed cleaning responsibilities between nursing and environmental service, including method of cleaning and frequency. Standard cleaning practices include floor and surface cleaning with a disinfectant effective against C. auris, bleach-wipe cleaning of equipment, and a log to track cleaning of shared equipment. Storage cabinets were installed in patient rooms. Black-light audits on discharge cleans were required on every terminal discharge to monitor cleaning practices, and environmental service staff received coaching when cleaning failures were identified. Germicidal ultraviolet disinfection was performed in patient rooms and above the unit’s nursing station beginning 9 months after first case detection. Thirteen months after first case detection, terminal cleaning of patient rooms incorporated high-intensity UV disinfection.

Unobtrusive-observer audits revealed that overall hand hygiene compliance was 78%–93% during the outbreak period. Most observations were of nursing staff; the greatest opportunities for improvement in compliance were among patient transporters (32% compliance), food and nutrition services (35% compliance), and physicians (67% compliance). The team increased hand hygiene promotion signage and efforts to normalize just-in-time coaching for hand hygiene and PPE breaches among staff and visitors.

Patient Characteristics

Figure 2

Figure 2. Case timeline for outbreak of Candida aurisin burn intensive care unit (BICU), Illinois, USA, 2021–2023. Horizontal bars indicate BICU admission duration for each patient; bar colors indicate genomic...

During the 21-month investigation, 28 patients were colonized or infected with C. auris (Table); 4 patients had invasive C. auris before admission screening. The average patient age was 49 years (range 16–81 years). Most patients were admitted with burns (64%), 9 patients (32%) were admitted with soft-tissue infections, and 1 patient was on medical ICU service. None of the patients had a history of C. auris infection, determined by chart review and query of the Illinois extensively drug-resistant organism registry. Seven patients were admitted from outside hospitals or had a hospitalization <30 days before the BICU admission; none were admitted from skilled nursing facilities. The mean length of stay in the BICU before identification of C. auris was 26 days (range 7–83 days). C. auris was identified in clinical cultures from 13 patients, some of which had C. auris in multiple cultures; 8 had C. auris identified in blood culture, 6 in respiratory culture, 8 in wound culture, and 3 in urine culture. The mean total length of stay in the BICU was 67 days (Figure 2).

Genomic Analysis of Outbreak Isolates

To investigate the genetic relationship of C. auris among cases, we conducted WGS on available isolates from the BICU outbreak, including isolates from 22 (79%) of 28 case-patients and 8 longitudinal isolates from 3 of those patients (Figure 2). Isolates from the remaining 6 case-patients were not available for analysis. We also sequenced available contemporaneous isolates from the same facility (n = 19) or another healthcare facility within the medical system (n = 31) to estimate diversity of hospital C. auris isolates and identify potential links outside of BICU. Of 80 isolates, 78 (97.5%) had sufficient genome quality for analysis; all were C. auris clade IV. Comparison with publicly available C. auris sequences from Illinois (n = 364) revealed that genomes from the BICU and related facilities were interspersed throughout other Illinois C. auris sequences (Appendix Figure 1), indicating multiple transmissions to the BICU facility from the broader diversity in the region.

Figure 3

Figure 3. Genomic analysis of outbreak and contemporaneous contextual Candida auris isolates in outbreak of C. aurisin BICU, Illinois, USA, 2021–2023. A) Neighbor-joining SNP-based phylogenetic tree of sequences...

Figure 4

Figure 4. Pairwise SNP distances within different Candida aurispopulations in study of outbreak in burn intensive care unit, Illinois, USA, 2021–2023. Black points (which appear as lines for the large...

BICU isolates formed 4 clusters within Illinois sequences (Figure 3, panel A). Clusters contained 3–6 unique patients and closely related genomes (0–8 SNP differences) (Figure 3, panels B-E). Three BICU isolates did not cluster closely with any other BICU or Illinois isolate, differing by 7–39 SNPs from the closest non-BICU and 21–49 SNPs from the closest BICU isolate (Figure 3, panel A; Appendix Figure 1). We identified a mean 1.9 (range 0–8) SNP differences within BICU clusters, which was not significantly different from SNP differences from isolates collected from the same patient (p>0.9999) (Figure 4). However, SNP differences among all BICU isolates were significantly higher (mean 34.6; p<0.0001), but not significantly different from, mean SNP differences among isolates collected within the same timeframe elsewhere within the medical center (mean 35.8) or all Illinois (mean 37.9). Those findings indicate that BICU clusters were very closely related among isolates within the cluster but not more closely related among clusters than for other regional isolates, consistent with multiple independent introductions from regional C. auris followed by within-unit spread.

Integrated Genomic and Epidemiologic Investigation of Outbreak Clusters

Clusters 2 and 4 contained exclusively BICU isolates (Figure 3, panels C, E), whereas clusters 1 and 3 contained sequences collected within the medical center but outside of the BICU (Figure 3, panels B, D). In one instance, in cluster 1, Figure 3, panel B), 1 sequence from the same facility collected 4 months before the first BICU case was identical to the first BICU case’s genome (patient 1). No epidemiologic links to BICU patients in this cluster were identified, and facility stays were separated by 123 days. In cluster 3 (Figure 3, panel D), 1 sequence from another unit within the facility and 1 sequence from elsewhere in the medical system collected 1 month before and 1 month after the first BICU case clustered with BICU isolates (patients 14, 15, 16, 18). No epidemiologic links were identified between the cases from the same facility. The case-patient from elsewhere in the medical system had been hospitalized at the BICU facility the month before, overlapping with other patients in this cluster. In addition, this patient and another BICU patient with C. auris from this cluster were both exposed to the same healthcare worker (speech therapist) during overlapping timeframes. Contextual isolates from institutions outside of the medical system did not fall in BICU clusters (>8 SNPs) (Figure 3; Appendix Figure 1).

BICU sequences clustered by collection date; clusters generally contained isolates collected within 3–6 months of each other (Figure 3, panels B–E). Clustering patients had overlapping BICU admission dates in all but 1 instance (Figure 2). In the exception, a patient (patient 9) from cluster 1 was admitted to the BICU 91 days after other patients in cluster 1 were discharged, but the C. auris isolate was not sequenced. This patient was then discharged to the same long-term acute care hospital as another patient in cluster 1 (Figure 3, panel B). The isolate belonging to cluster 1 was identified in an admission screening culture 8 months later, when the patient was admitted to another unit within the BICU facility. Thus, C. auris transmission may have occurred either on the BICU or at the long-term acute care hospital.

Analysis of Longitudinal Outbreak Isolates

Longitudinal isolates were collected from 3 patients over 51 (n = 2), 63 (n = 6), and 135 (n = 3) days and included both clinical and screening isolates. All longitudinal sequences clustered closely with other sequences from the same patient. Six of 11 longitudinal sequences were identical to another sequence from the same person; clinical and screening isolates often identical to one another (Figure 3, panels C–E). Specimens collected from the same person had a mean 1.2 (range 1–4) SNP differences (Figure 4).

We investigated mutations in antifungal resistance–associated genes to look for longitudinal acquisition of antifungal resistance mutations. We identified a mutation associated with azole resistance in the TAC1b gene (I187T) in the last isolate collected from 1 patient; the 2 isolates obtained from this person earlier did not contain this mutation (Figure 3, panel D) (32,33). The patient received voriconazole therapy after collection of the isolates lacking the mutation but before the emergence of the TAC1b mutation. The isolate was not subjected to phenotypic antifungal susceptibility testing.