Exploratory Evaluation of Chlorhexidine Decolonization and Skin Colonization Dynamics of Candida auris in ICU Patients: A Prospective Pilot Study

Abstract

Objective

Candida auris has emerged as a nosocomial pathogen in intensive care units (ICUs), and evidence for chlorhexidine-based decolonization remains limited. We report an exploratory pilot study describing skin colonization dynamics in four ICU patients with C. auris infection or colonization who received chlorhexidine decolonization alongside standard infection control measures.

Methods

Four consecutive C. auris-positive patients admitted to the ICU of a tertiary teaching hospital in Inner Mongolia, China, between January 31 and March 12, 2026, were enrolled. Two patients (intervention group) received twice-daily 2% chlorhexidine gluconate whole-body skin decolonization; two (non-intervention group) did not, based on family consent. All patients received identical baseline infection control measures. Skin swabs from the nares, axillae, groin, and external ear canals, together with environmental samples, were cultured serially. All isolates were identified by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS). For one patient, metagenomic next-generation sequencing (mNGS) of bronchoalveolar lavage (BAL) fluid was performed as part of routine clinical workup.

Results

Skin colonization burden declined progressively in both intervention patients: Case A fell from 34 colony-forming units (CFU)/swab at baseline to 4 CFU/swab by Day 12, and Case B from 10 CFU/swab (Day 4) to 5 CFU/swab by Day 6 (discharged on Day 10). Colonization burden did not decline in the non-intervention group (Case C: 24–30 CFU/swab in groin across Day 0–12). Clinical outcomes differed between groups, but the non-randomized design, baseline imbalance in infection status, and universal co-infection with multidrug-resistant organisms preclude any causal inference. C. auris was recovered from 1 of 93 environmental surveillance samples (a suction bottle) and was eliminated by targeted disinfection; no healthcare worker hand cultures were positive.

Conclusion

In this four-patient pilot study, twice-daily chlorhexidine decolonization was accompanied by a reduction in skin colonization burden, but the findings are hypothesis-generating only. The small sample size, non-randomized design, baseline differences, and lack of molecular typing limit interpretation. Adequately powered, preferably randomized, studies with whole-genome sequencing are needed.

Introduction

Candidozyma auris (formerly Candida auris) is a multidrug-resistant yeast first isolated in 2009 from the external ear canal of a patient in Japan.¹ Its emergence has been linked to multiple factors including climate adaptation and azole fungicide exposure in the environment.² Over the following decade it spread to more than 50 countries, driven by efficient nosocomial transmission, broad antifungal resistance,³ and an unusual ability to persist on inanimate surfaces.^4,5 The World Health Organization (WHO) now lists C. auris among its four critical-priority fungal pathogens.⁶

ICU patients are at increased risk. Prolonged stay, invasive devices, broad-spectrum antibiotics, and compromised immunity all favor acquisition,^7,8 and once the organism enters a unit it can spread rapidly via contaminated equipment and the environment, triggering outbreaks that are difficult to contain.^9,10 Skin colonization, particularly of the axillae, groin, and nares, acts as a reservoir for onward transmission and raises the risk of invasive disease in vulnerable hosts.^11,12

A central distinction in C. auris management is between colonization and infection. Colonization denotes presence of the organism on skin or mucosal surfaces without signs of invasive disease, whereas infection implies tissue invasion with clinical manifestations. The two states differ in pathophysiology, transmissibility, and clinical consequence: colonized patients may serve as reservoirs for transmission, while infected patients experience direct morbidity. In this pilot study we analyzed colonized and infected patients together because both require infection control measures to limit spread, because chlorhexidine acts on skin colonization irrespective of infection status, and because the sample size was too small to permit separate analysis. This combined analysis is acknowledged as a limitation, since baseline infection status confounds any comparison of clinical outcomes.

Chlorhexidine gluconate (CHG) is a broad-spectrum antiseptic widely used for decolonization of methicillin-resistant Staphylococcus aureus, but its efficacy against C. auris is less well established. An ex vivo porcine skin model found CHG largely ineffective,¹³ and a 2025 systematic review reported only a transient effect of CHG bathing on overall Candida colonization in ICU patients.¹⁴ Several outbreak reports and reviews, by contrast, suggest that CHG decolonization combined with enhanced environmental disinfection may contribute to outbreak control.^15,16 Real-world observational data from varied settings are therefore still needed to generate hypotheses for definitive trials.

We undertook a prospective pilot study to describe skin colonization dynamics in ICU patients with C. auris infection or colonization who received CHG decolonization together with environmental infection control measures, and to document environmental contamination and clinical outcomes over the study period. The study was designed to be exploratory and hypothesis-generating, not to establish efficacy.

Methods

Study Design and Setting

This prospective, observational pilot study was conducted in the ICU of the Affiliated Hospital of Inner Mongolia Medical University, a tertiary teaching hospital in Hohhot, Inner Mongolia, China, from January 31 to March 12, 2026. The study was approved by the Ethics Committee of the Affiliated Hospital of Inner Mongolia Medical University (approval No. KY2026081). All legally authorized representatives of the patients signed written informed consent. Given the small sample size and non-randomized design, the study was exploratory and hypothesis-generating in nature.

Patients and Group Allocation

Four consecutive patients with C. auris detected from clinical or surveillance cultures during the study period were enrolled. Patients were assigned to an intervention group (n=2, Cases A and B) or a non-intervention group (n=2, Cases C and D) based on family informed consent regarding chlorhexidine decolonization. Both groups received identical baseline infection control measures, including contact precautions, dedicated medical equipment, environmental disinfection, and hand hygiene compliance monitoring. The intervention group also received twice-daily 2% CHG whole-body skin decolonization, with particular attention to colonization-prone sites (nares, axillae, groin, and external ear canals).

Definitions of Colonization and Infection

Colonization and infection were determined by the attending physicians based on clinical assessment combined with microbiological findings, consistent with established criteria. C. auris colonization was defined as a positive culture from a non-sterile site (skin, nares, groin, axillae, external ear canals) without clinical signs of invasive fungal infection. C. auris infection was defined as a positive culture from a clinically relevant specimen (eg, BAL fluid, urine) accompanied by compatible clinical features (eg, fever, leukocytosis, radiographic infiltrates, or signs of organ dysfunction) and the absence of an alternative explanation, as judged by the treating clinician.^17,18 For urinary tract involvement, a quantitative culture threshold of ≥10⁷ CFU/mL with compatible urinary symptoms or pyuria was applied. Community-acquired C. auris was defined as detection within 48 hours of hospital admission, while nosocomial acquisition was defined as detection more than 48 hours after admission, in accordance with the Chinese national standard for diagnosis of nosocomial infections.¹⁹

Microbiological Methods

Skin and surveillance cultures were obtained using sterile rayon swabs moistened with sterile saline from four anatomical sites: nares, axillae, groin, and external ear canals. Clinical specimens (BAL fluid, urine) were collected as part of routine clinical care. All specimens were transported to the microbiology laboratory within 2 hours of collection and processed immediately.

Specimens were inoculated onto Sabouraud dextrose agar (SDA) supplemented with chloramphenicol (50 mg/L) to suppress bacterial contamination, and onto a chromogenic Candida agar (CHROMagar Candida) to facilitate preliminary differentiation of yeast species based on colony color and morphology. Cultures were incubated at 35°C under aerobic conditions for 48 hours; no extended incubation beyond 48 hours was performed. All yeast isolates recovered from skin, clinical, and environmental cultures were identified by MALDI-TOF MS (Bruker Daltonik, Bremen, Germany) using the direct smear method. MALDI-TOF MS identification was performed on all culture-positive isolates, including skin swabs, clinical specimens (BAL fluid, urine), and environmental samples, with identical processing workflows for screening and clinical isolates. All isolates achieved MALDI-TOF MS log scores ≥ 2.0, which is considered reliable species-level identification per Clinical and Laboratory Standards Institute (CLSI) guidelines; no supplementary molecular identification (eg, internal transcribed spacer [ITS] sequencing) was required. For one patient (Case A), mNGS was also performed on the BAL specimen as part of routine clinical workup for severe pneumonia. mNGS was performed on a domestic high-throughput sequencing platform using the PMseq workflow (BGI, Shenzhen, China), with library preparation from total extracted DNA, sequencing to a depth of approximately 112 million reads (Q30 ≥ 86%), and bioinformatic alignment against the PMseq Database (PMDB).

Reference strains were used throughout the laboratory workflow as positive controls, including American Type Culture Collection (ATCC) strains for both antifungal susceptibility testing and MALDI-TOF MS identification, in accordance with established laboratory standards. All C. auris isolates from the four patients achieved MALDI-TOF MS log scores ≥ 2.0, providing reliable species-level identification; no ITS sequencing or other supplementary molecular identification was performed. For Case A, mNGS of the BAL specimen was performed as part of routine clinical diagnostic workup for severe pneumonia (not for species confirmation), and detected C. auris (14 sequence reads; 8.63% relative abundance among fungal sequences) together with other co-resident pathogens. The exclusion of non-auris Candida contamination was achieved through a combination of selective media (chloramphenicol-supplemented SDA inhibiting bacterial overgrowth and chromogenic agar providing species-specific color differentiation), comprehensive MALDI-TOF MS confirmation of all yeast isolates, and morphological screening of colony characteristics consistent with C. auris (smooth, convex, white to cream-colored colonies).^20,21 No mixed fungal growth was encountered during the study period.

Environmental Surveillance

Environmental surveillance was conducted as targeted monitoring for C. auris contamination of high-touch surfaces and medical equipment in patient vicinity. Environmental samples were collected using sterile swabs pre-moistened with sterile saline from predefined high-touch surfaces (eg, bed rails, infusion pumps, blood pressure cuffs, keyboards, suction bottles). Samples were inoculated onto selective media for fungal recovery, including chloramphenicol-supplemented SDA, chromogenic Candida agar, and standard agar media, with incubation at 35°C under aerobic conditions for 48 hours. Environmental surveillance was directed specifically at fungal pathogens, given the targeted nature of the monitoring for C. auris. Quantitative culture was performed, and all culture-positive environmental isolates underwent MALDI-TOF MS identification following morphological screening, using the same workflow as clinical isolates. A sample was considered to meet quality standards if no fungal growth was detected or if colony counts remained at or below 5.0 CFU/cm², in accordance with the Chinese national standard for hospital disinfection (GB 15982–2012), which classifies ICU as a Class II environment with a surface microbial limit of ≤5.0 CFU/cm²;²² samples failing to meet quality standards (eg, due to delayed transport exceeding 2 hours) were recollected.

Infection Control Measures

Upon laboratory confirmation of C. auris, contact precautions were initiated immediately following notification of the laboratory report, including placement of patients in single rooms or cohorting, use of dedicated medical equipment, and reinforcement of hand hygiene. Routine active screening for C. auris was not performed for all newly admitted ICU patients during the study period; surveillance cultures were obtained from the four enrolled patients and from environmental surfaces as described above. Daily environmental cleaning and disinfection were performed using 1000 mg/L chlorine-based disinfectant twice daily (once in the morning and once in the afternoon) for high-touch surfaces, including the patient bed unit, floor, bathroom, and medical equipment surfaces (eg, bed rails, bedside tables, call buttons, monitors, infusion pumps, door handles, blood pressure cuffs), in accordance with the Chinese expert consensus on healthcare-associated infection prevention and control of C. auris.²³ Terminal disinfection upon patient discharge, transfer, or death was performed as follows: (1) air disinfection by ultraviolet-C irradiation for at least 30 minutes followed by natural ventilation by opening windows; (2) bed unit disinfection using an ozone bed-unit disinfection device for 30 minutes; and (3) wipe disinfection of all environmental surfaces and medical equipment with 1000 mg/L chlorine-based disinfectant for a contact time of 30 minutes, followed by wiping with clean water. After terminal disinfection, targeted C. auris environmental surveillance cultures were obtained from the patient room and equipment; only rooms with negative C. auris cultures were approved for admission of new patients. Hand hygiene compliance was monitored and reinforced throughout the study period.

Data Collection and Outcomes

Data collected included patient demographics, admission source, infection status (colonization vs infection), acquisition type (community-acquired vs nosocomial), C. auris specimen source, co-infecting multidrug-resistant organisms, co-colonizing Candida species, antifungal therapy, ICU length of stay, and clinical outcome. Skin colonization dynamics were assessed by serial quantitative cultures from the four anatomical sites at predefined time points (Day 0 [baseline], Day 4, Day 6, Day 12, and Day 38). All four anatomical sites were sampled at each time point for enrolled patients; values recorded as “0” represent true negative cultures rather than missed sampling. Environmental monitoring results, including the number of samples collected, compliance with quality standards, and pathogens detected, were recorded.

Statistical Analysis

Given the exploratory nature of this pilot study and the extremely small sample size (n=4), no formal statistical hypothesis testing was performed. Continuous variables are presented descriptively. Skin colonization burden is reported as CFU per swab at each time point. Given the absence of inferential statistics, the findings are descriptive and hypothesis-generating only.

Results

Patient Characteristics

Four patients with C. auris detected during the study period were enrolled (Table 1 and Figure 1). The intervention group comprised Case A (56-year-old male with C. auris infection, BAL fluid positive, community-acquired) and Case B (72-year-old male with C. auris colonization, groin skin positive, nosocomial). The non-intervention group comprised Case C (70-year-old male with C. auris infection, urine positive, community-acquired) and Case D (64-year-old male with C. auris infection, BAL fluid positive, nosocomial). All four patients were co-infected with other multidrug-resistant organisms, including carbapenem-resistant Acinetobacter baumannii (CRAB), carbapenem-resistant Enterobacterales (CRE), and carbapenem-resistant Pseudomonas aeruginosa (CRPA). Three patients (Cases A, B, and C) also had co-colonization or co-infection with other Candida species, including Nakaseomyces glabratus (formerly Candida glabrata), Candida tropicalis, and Candida albicans. None of the four patients received antifungal therapy during the study period. The reasons for the absence of antifungal therapy varied: Case D died rapidly within 5 days of ICU admission before antifungal therapy could be initiated; for the remaining patients, the attending physicians determined that antifungal therapy was not indicated based on clinical assessment, given that C. auris was considered to represent colonization or low-burden infection in the context of the patients’ overwhelming bacterial co-infections and multiple life-limiting comorbidities. ICU length of stay ranged from 5 to 41 days. Case D, with an ICU stay of only 5 days, was not available for longitudinal assessment of skin colonization dynamics.

Table 1.

Clinical Characteristics of the Four Patients with Candida auris Colonization or Infection

Characteristic	Case A	Case B	Case C	Case D
Age (years)/Sex	56/M	72/M	70/M	64/M
Admission source	External hospital	Rehabilitation dept (in-hospital transfer)	External hospital	Respiratory dept (in-hospital transfer)
Infection status	Infection	Colonization	Infection	Infection
Acquisition type	Community	Nosocomial (contact)	Community	Nosocomial
C. auris specimen	BAL fluid (mNGS also performed)	Groin skin (active screening)	Urine	BAL fluid
Other MDR organisms	CRPA, CRE (BAL); CRE (wound)	CRAB (BAL); CRE (urine)	CRAB, CRE, CRPA (sputum/urine/pleural fluid)	CRAB (sputum)
Other Candida species	N. glabratus (sputum); C. tropicalis (urine)	C. albicans (BAL); C. tropicalis (urine)	None	None
Decolonization	Yes (2% CHG, twice daily)	Yes (2% CHG, twice daily)	No	No
Antifungal therapy	None	None	None	None
ICU LOS (days)	37	27	41	5
Clinical outcome	Transferred to county hospital	Discharged	Discharged against medical advice	Death

Abbreviations: BAL, bronchoalveolar lavage; CRAB, carbapenem-resistant Acinetobacter baumannii; CRE, carbapenem-resistant Enterobacterales; CRPA, carbapenem-resistant Pseudomonas aeruginosa; DAMA, discharge against medical advice; ICU LOS, intensive care unit length of stay; MDR, multidrug-resistant; N. glabratus, Nakaseomyces glabratus.

Figure 1.

Study timeline and patient enrollment. Schematic representation of the study period (January 31 to March 12, 2026), showing the enrollment of the four patients with Candida auris colonization or infection, group allocation (intervention vs non-intervention), and key clinical events (ICU admission, discharge, transfer, death).

C. auris Identification

All yeast isolates recovered from skin swabs, clinical specimens (BAL fluid, urine), and environmental samples were identified by MALDI-TOF MS with high-confidence species-level identifications (log score ≥ 2.0) for all four patients; no supplementary molecular identification was required. For Case A, mNGS was also performed on the BAL specimen as part of routine clinical workup for severe pneumonia. mNGS detected C. auris (14 sequence reads; 8.63% relative abundance among fungal sequences) and Nakaseomyces glabratus (10 reads; 5.90% relative abundance), together with several co-resident bacterial pathogens including Stenotrophomonas maltophilia, Achromobacter xylosoxidans, Pseudomonas aeruginosa, Corynebacterium striatum, and Klebsiella pneumoniae. The mNGS finding was consistent with the culture-based MALDI-TOF MS identification of C. auris; however, mNGS was performed for clinical diagnostic purposes only and does not constitute an independent confirmation of isolate identity. Species identification of all isolates rested on MALDI-TOF MS (log score ≥ 2.0). All screening and clinical isolates were processed using identical workflows, and no discrepant identifications were observed.

Skin Colonization Dynamics

Skin colonization dynamics are summarized in Table 2 and Figure 2. At baseline (Day 0), all patients with available data had localized colonization at a single anatomical site: Case A had C. auris detected in nares (34 CFU/swab), Case C in groin (24 CFU/swab), while Case B had no detectable colonization at baseline. Case D was not yet admitted at the Day 0 time point.

Table 2.

Serial Skin Colonization Burden (CFU/Swab) by Anatomical Site in the Four Patients

Time	Case A (Intervention)	Case B (Intervention)	Case C (Non-Intervention)	Case D (Non-Intervention)
Day 0 (Baseline)	Nares: 34	0	Groin: 24	Not admitted
Day 4	Groin: 20	Groin: 10	Groin: 30	Not admitted
Day 6	Nares: 10	Groin: 5	Groin: 25, Axilla: 20	Not admitted
Day 12	Axilla: 4, Ear: 4	Discharged (Day 10)	Groin: 19	Not admitted
Day 38	Transferred (Day 27)	Discharged	DAMA (Day 38)	Death (Day 35)

Notes: All four anatomical sites (nares, axillae, groin, external ear canals) were sampled at each time point for enrolled patients; values recorded as “0” represent true negative cultures. Case D was not available for serial skin sampling due to death after 5 days in the ICU; C. auris was detected in bronchoalveolar lavage (BAL) fluid on Day 33 (first detection), and the patient died on Day 35.

Abbreviations: ND, no data available. CFU, colony-forming unit.

Figure 2.

Serial skin colonization burden in the four patients. Line graph showing the Candida auris colonization burden (CFU/swab) at each sampling time point (Day 0, Day 4, Day 6, Day 12, Day 38) for each patient. Values represent the highest CFU/swab among all anatomical sites (nares, axillae, groin, external ear canals) at each time point. Case D was not available for serial skin sampling due to death after 5 days in the ICU.

In the intervention group, Case A showed a progressive decline in colonization burden from 34 CFU/swab (nares, Day 0) to 20 CFU/swab (groin, Day 4), 10 CFU/swab (nares, Day 6), and 4 CFU/swab (axilla and ear, Day 12), with the last measured value at 4 CFU/swab on Day 12; no skin cultures were obtained at the time of ICU transfer on Day 27. Case B had no detectable colonization at baseline, became positive on Day 4 (groin, 10 CFU/swab), and subsequently showed declining burden to 5 CFU/swab (groin, Day 6); the patient was discharged on Day 10. The anatomical sites of detection varied across time points within each patient.

In the non-intervention group, Case C showed persistent colonization with burden fluctuating between 24 and 30 CFU/swab in the groin, with new detection in the axilla (20 CFU/swab) on Day 6, and groin colonization of 19 CFU/swab on Day 12, before discharge against medical advice on Day 38. Case D had C. auris detected in BAL fluid (25 CFU) on Day 33 (first detection); the patient died on Day 35 after 5 days in the ICU. No skin surveillance cultures were obtained given the brief ICU stay.

Interpretable longitudinal analysis of skin colonization dynamics was limited to three patients (Cases A, B, and C), as Case D was not available for serial sampling. The observed trends in skin colonization burden in the intervention group should be interpreted with caution: the study cannot distinguish between changes attributable to chlorhexidine and natural fluctuation over time. The intermittent detection of C. auris from isolated anatomical sites at different time points raises the possibility that repeated negative cultures may reflect sampling variability rather than true microbiological clearance.

Environmental Monitoring

A total of 100 environmental samples were collected during the study period (Table 3), of which 93 (93.0%) met quality standards (defined as no fungal growth or colony counts below predefined thresholds). C. auris was detected in 1 of 93 quality-compliant samples (1.1%), recovered from the surface of a suction bottle on February 12. Other pathogens detected in non-compliant samples included Staphylococcus epidermidis (blood pressure cuff, January 31) and Acinetobacter baumannii and other organisms (infusion pump, keyboard, February 4). Following targeted disinfection of the contaminated suction bottle and reinforcement of environmental cleaning protocols, no further C. auris was detected in environmental samples from February 14 through March 12 (30 samples, 100% compliance, no C. auris detected). Hand hygiene compliance monitoring throughout the study period showed no hand contamination among healthcare workers despite the presence of environmental contamination.

Table 3.

Environmental Surveillance Results During the Study Period

Date	Samples (n)	Compliant (n)	Compliance Rate (%)	Pathogen Detected	Non-Compliant Sites
Jan 31	16	15	93.8	S. epidermidis	Blood pressure cuff
Feb 4	33	28	84.8	A. baumannii etc.	Infusion pump, keyboard
Feb 6	6	6	100.0	None	–
Feb 12	15	14	93.2	C. auris	Suction bottle
Feb 14–Mar 12	30	30	100.0	None	–
Total	100	93	93.0	–	–

Notes: A sample was considered to meet quality standards if no fungal growth was detected or if colony counts remained at or below 5.0 CFU/cm², in accordance with the Chinese national standard for hospital disinfection (GB 15982–2012), which classifies ICU as a Class II environment with a surface microbial limit of ≤5.0 CFU/cm². Candida auris was detected in 1 of 93 quality-compliant samples (1.1%).

Abbreviations: A. baumannii, Acinetobacter baumannii; C. auris, Candida auris; S. epidermidis, Staphylococcus epidermidis.

Clinical Outcomes

Clinical outcomes are summarized in Table 1. In the intervention group, Case A was transferred from the ICU after 37 days, and Case B was discharged after 27 days. In the non-intervention group, Case C was discharged against medical advice after 41 days with persistent skin colonization, and Case D died after 5 days in the ICU. Interpretation of these outcome differences is limited by baseline differences between groups: the intervention group included one infected and one colonized patient, while the non-intervention group included two infected patients. All four patients were also co-infected with other multidrug-resistant organisms and had multiple severe comorbidities, further complicating any outcome comparison. No causal inference regarding the effect of chlorhexidine decolonization on clinical outcomes can be drawn from these data.

Discussion

In this four-patient pilot, twice-daily chlorhexidine decolonization was accompanied by a fall in skin colonization burden in the two intervention patients. These findings are exploratory: the sample size is small, allocation was non-randomized, the groups differed at baseline, and no molecular typing was performed. Clinical outcomes also differed between groups, but baseline infection status, universal co-infection with multidrug-resistant organisms, and the five-day ICU stay of Case D preclude any causal inference.

The decline in colonization burden we observed is consistent with chlorhexidine’s known action on fungal cell membranes, though the study cannot separate a drug effect from natural fluctuation, and intermittent detection from different sites means some negative cultures may reflect sampling variability rather than true clearance. Our results differ from those of Johnson et al, who found chlorhexidine largely ineffective in an ex vivo porcine skin model.¹³ Several factors may account for this difference. Our protocol used twice-daily application with thorough coverage of high-yield sites, likely achieving higher local concentrations than a single experimental application. Both intervention patients had localized colonization, which may respond better than the extensive colonization seen in outbreaks. The 2025 clinical study that reported only transient effects of chlorhexidine bathing focused on Candida colonization broadly rather than on C. auris specifically,¹⁴ limiting direct comparison. Schelenz et al, describing the first European hospital outbreak, found decolonization ineffective in patients with extensive colonization, suggesting that colonization burden may influence treatment response.⁹

Environmental surveillance identified C. auris on a single surface (the suction bottle; 1 of 93 quality-compliant samples, 1.1%); no C. auris was detected in subsequent surveillance after targeted disinfection. The environmental yield was low, and surveillance targeted fungi rather than providing comprehensive bacterial and fungal assessment. Hand cultures from healthcare workers remained negative throughout the study period. C. auris has been reported to persist on inanimate healthcare surfaces in other settings,^24,25 but the present surveillance data are descriptive only and do not allow any inference regarding environmental persistence, transmission pathways, or the relative contribution of specific infection control measures in our setting; such inferences would require molecular typing and whole-genome sequencing, which were not performed.

Clinical outcomes are reported descriptively and should not be interpreted as evidence of comparative efficacy. Both intervention patients left the ICU alive (one transferred, one discharged), while both non-intervention patients had poor outcomes (one discharged against medical advice, one died). These differences, however, are confounded at baseline: the intervention group contained one infected and one colonized patient, whereas the non-intervention group contained two infected patients, one of whom died within 5 days of ICU admission. All four patients carried other multidrug-resistant organisms and had multiple severe comorbidities. None received antifungal therapy: Case D died before treatment could start, and in the remaining three the attending team judged antifungal therapy not indicated given the polymicrobial bacterial burden and the absence of invasive candidiasis. No causal inference regarding the effect of chlorhexidine on clinical outcomes can therefore be drawn.

This study has several limitations. The sample size of four precludes statistical inference, and all findings are descriptive. Allocation was non-randomized and based on family consent, introducing selection bias that is further compounded by baseline imbalance in infection status, colonization burden, and comorbidity. Case D died after five days, leaving longitudinal data for only three patients. For Case A, mNGS of the BAL specimen was performed as part of routine clinical diagnostic workup and incidentally detected C. auris nucleic acid; species identification of the cultured isolate rested on MALDI-TOF MS alone. mNGS is a clinical diagnostic tool for pathogen detection and cannot substitute for molecular typing or strain characterization; no whole-genome sequencing was performed, so strain relatedness, transmission chains, and clonality could not be assessed. Pooling colonized and infected patients was unavoidable at this sample size but further confounds outcome comparison. Intermittent detection from different anatomical sites means that some negative cultures may reflect sampling variability rather than true clearance, and the absence of a sham-decolonization control makes it impossible to attribute observed changes solely to chlorhexidine. The single-center design also limits generalizability.

Despite these limitations, the study provides descriptive real-world observations that may inform future investigations. The bundle of patient isolation, dedicated equipment, contact precautions, environmental disinfection, and hand hygiene aligns with current expert recommendations. Kenters et al, reporting an International Society of Antimicrobial Chemotherapy (ISAC) expert meeting, emphasized the need for multi-faceted strategies for C. auris control in healthcare institutions,¹² and Centers for Disease Control and Prevention (CDC) guidance likewise recommends contact precautions, hand hygiene, and environmental disinfection, all components of our protocol.^26,27

Definitive evaluation will require larger, ideally randomized, studies with whole-genome sequencing to trace transmission, standardized sampling protocols applied consistently across anatomical sites and time points, and sample sizes adequate to detect clinically meaningful differences. Multicenter collaboration would help overcome the case-volume constraint at single institutions. Because C. auris so often coexists with other multidrug-resistant organisms, future studies might also examine multi-target decolonization strategies.

Conclusions

In this four-patient pilot, twice-daily chlorhexidine decolonization was accompanied by a fall in skin colonization burden in the two intervention patients. Clinical outcomes differed between groups, but the non-randomized design, small sample size, baseline imbalance, universal co-infection with multidrug-resistant organisms, and absence of molecular typing rule out any causal claim. Environmental contamination was detected once (1 of 93 samples) and cleared by targeted disinfection; hand cultures remained negative throughout. The findings are hypothesis-generating only, and adequately powered, preferably randomized, studies with whole-genome sequencing are needed to determine whether chlorhexidine decolonization has a role in C. auris control.

Funding Statement

Affiliated Hospital of Inner Mongolia Medical University Research Fund (YSXH2025KYF64).

Abbreviations

C. auris, Candida auris; ICU, intensive care unit; CFU, colony-forming unit; WHO, World Health Organization; CDC, Centers for Disease Control and Prevention; CHG, chlorhexidine gluconate; MALDI-TOF MS, matrix-assisted laser desorption/ionization time-of-flight mass spectrometry; mNGS, metagenomic next-generation sequencing; BAL, bronchoalveolar lavage; SDA, Sabouraud dextrose agar; CRAB, carbapenem-resistant Acinetobacter baumannii; CRE, carbapenem-resistant Enterobacterales; CRPA, carbapenem-resistant Pseudomonas aeruginosa; CLSI, Clinical and Laboratory Standards Institute; PMDB, PMseq Database; ATCC, American Type Culture Collection; ITS, internal transcribed spacer; ISAC, International Society of Antimicrobial Chemotherapy.

Ethics Approval

This study was approved by the Ethics Committee of the Affiliated Hospital of Inner Mongolia Medical University (approval No. KY2026081). All legally authorized representatives of the patients signed written informed consent.

Author Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

All authors declare no conflicts of interest in this work.