Outbreak of carbapenem-resistant Acinetobacter baumannii in post–haematopoietic stem cell transplantation patients potentially caused by drinking water contamination

Summary

Background

Carbapenem-resistant Acinetobacter baumannii (CRAB) infections pose a significant threat to haematopoietic stem cell transplantation (HSCT) recipients. This study characterized CRAB isolates from post-HSCT patients and the ward environment to investigate potential reservoirs of clonal transmission.

Presentation of the case

Six HSCT recipients with haematological malignancies were colonized or infected with CRAB. Five patients were identified through throat swab screening, whereas one was detected by urine culture. Of the 42 Acinetobacter baumannii isolates recovered from the HSCT ward environment, 27 (64.3%) were traced to the centralized drinking water supply, including buckets used for cooling and storing boiled water (N = 2), patients’ drinking water (N = 13), and water bottles (N = 12).

Whole-genome-based core genome multi-locus sequence typing identified a predominant clonal cluster (cluster A; N = 18) involving A. baumannii isolates from the HSCT ward. This cluster comprised five patient isolates (from five of the six CRAB-positive HSCT recipients) and 13 environmental isolates. Molecular epidemiological analysis revealed that 8 of the 13 environmental isolates (61.5%) originated from the centralized drinking water supply, including one isolate from buckets used for cooling and storing boiled water and seven isolates from patients’ drinking water or water bottles.

Conclusion

These findings confirm the circulation of a clonal CRAB strain within the HSCT ward and demonstrate cross-transmission between patients and the environment. Contaminated containers used for storing and transporting drinking water may serve as potential reservoirs for CRAB, facilitating its dissemination within the ward. Strict adherence to standardized potable water management protocols is imperative to prevent CRAB transmission in high-risk units.

Introduction

Acinetobacter baumannii (AB) is a notorious opportunistic pathogen and a leading cause of severe healthcare-associated infections (HAIs) [1]. This bacterium demonstrates remarkable environmental resilience, enabling it to persist on dry surfaces and facilitating transmission through contaminated environments or the hands of healthcare personnel. Common environmental reservoirs in HAIs include sinks, shower facilities, infusion pumps, resuscitation equipment, pillows, mattresses, and water supply systems [[2], [3], [4], [5], [6], [7], [8]]. AB is associated with a wide spectrum of clinical manifestations, including bloodstream infections, pneumonia, urinary tract infections, and skin and soft-tissue infections [9]. Over the past two decades, AB has disseminated extensively, emerging as one of the most prevalent and challenging pathogenic micro-organisms worldwide. [10]. Patients at the highest risk for AB infection include those who are immunocompromised, undergo invasive medical procedures, or receive immunosuppressive therapies, such as haematopoietic stem cell transplantation (HSCT) [11,12]. Therefore, effective infection prevention and control (IPC) measures are essential to reduce the burden of AB infections in these vulnerable populations.

Methods

Bacterial strain collection and identification

From December 2020 to March 2021, biological samples were collected from six patients in a HSCT ward and from 30 patients across 17 other hospital wards. In addition, 369 environmental samples obtained were collected from the HSCT ward. AB was identified using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Antimicrobial susceptibility testing was performed using the broth microdilution method, with Escherichia coli ATCC 25922 as the quality control strain, in accordance with guidelines of the Clinical and Laboratory Standards Institute. Whole-genome sequencing (WGS) was performed on carbapenem-resistant Acinetobacter baumannii (CRAB) isolates from the 30 patients and 16 AB isolates recovered from environmental samples.

Patient characteristics

Epidemiological and clinical data were collected for 36 patients from hospital medical records. Information included the date of infection, underlying medical conditions, surgical history, and antibiotic usage.

WGS and bioinformatic analysis

Genomic DNA was extracted using a plant genomic DNA kit (Tiangen, DP305, Beijing, China), and DNA quantity and integrity were verified by spectrophotometry and agarose gel electrophoresis, respectively. Sequencing libraries were prepared with the TruePrep DNA Library Prep Kit V2 for Illumina (Vazyme, Nanjing, China) and sequenced on the Illumina NovaSeq platform (Illumina Inc., San Diego, CA, USA) to generate 150-bp paired-end reads. Raw reads were quality assessed with FastQC and trimmed with Trim Galore, followed by de-novo assembly using Unicycler. Genome annotation was performed with Prokka, and ribosomal RNA (rRNA) and transfer RNA (tRNA) genes were identified using RNAmmer and tRNAscan-SE, respectively. Functional annotation was conducted against the NR, Swiss-Prot, KEGG, COG, VFDB, PHI, TCDB, and CAZy databases; signal peptides, transmembrane domains, and subcellular localization were predicted with SignalP, TMHMM, and PSORTb, respectively. Type III secretion system effectors were predicted using EffectiveT3, CRISPR/Cas systems with CRISPR-Casfinder, and secondary metabolite biosynthetic gene clusters with antiSMASH. Core genome multi-locus sequence typing (cgMLST) was performed using Ridom SeqSphere+ (Ridom GmbH, Münster, Germany), with the validated AB scheme comprising 2390 targets, using strain ACICU (GenBank accession no. NC_010611) as the reference genome; isolates differing by ≤10 alleles were considered clonally related [[13], [14], [15], [16], [17]]. WGS, multi-locus sequence typing (MLST), and cgMLST analyses were performed by Zhejiang Tianke High Technology Development Co., Ltd. (Hangzhou, China).

Presentation of the case

Case description

On 4^th January 2021, an initial cluster of CRAB infections was identified among post-HSCT patients on the HSCT ward. The IPC team conducted an on-site investigation and collected environmental samples. A total of six post-HSCT patients were confirmed to be have a CRAB infection.

The six patients were admitted between late November and December 2020 and ranged in age from 14 to 42 years. Five patients underwent HSCT during this admission, while one had previously undergone HSCT at the same hospital. All patients underwent routine admission screening, including throat and anal swab sampling, and received broad-spectrum antibiotics during their hospital stay. Prior to the outbreak, routine screening identified only one patient with Enterobacter cloacae from a throat swab, while no pathogens were detected in the remaining patients (Table I).

Table I.

Basic information of 6 post–haematopoietic stem cell transplantation (HSCT) patients with CRAB-positive isolates

Case	Admission date	Diagnosis on admission	Sex	Age	Time of HSCT	Antibiotic usea	Specimen of CRAB positivea	Specimen collection time of CRAB positive	Specimen of other pathogen positive∗	MLST	cgMLST
1B	2/11/20	Acute myelocytic leukaemia	F	42	15/12/20	Yes	Sputum	26/12/20	No	ST958	Cluster A
2B	3/11/20	Thalassaemia	M	13	17/12/20	Yes	Sputum	30/12/20	No	STnovel	No in cluster
3B	27/11/20	Acute lymphoblastic leukaemia	F	20	17/12/20	Yes	Sputum	31/12/20	No	STnovel	Cluster A
4B	4/12/20	Thalassaemia	F	21	25/12/20	Yes	Sputum	31/12/20	No	ST1806	Cluster A
5B	14/12/20	Thalassaemia	M	14	Before admission	Yes	Urine	4/1/21	No	ST1806	Cluster A
6B	20/12/20	Acute lymphoblastic leukaemia	M	41	24/12/20	Yes	Sputum	5/1/21	12/29, Sputum, Enterobacter cloacae	ST1806	Cluster A

CRAB, carbapenem-resistant Acinetobacter baumannii; MLST, multi-locus sequence typing; cgMLST, core genome multi-locus sequence typing.

^aHaematopoietic stem cell transplantation and symptoms appeared before CARB was positive.

HSCT ward design and management

The HSCT ward consisted of 18 single-bed laminar airflow isolation rooms and six multi-bed rooms, providing a total of 31 beds. The ward maintained consistently high occupancy rates throughout the study period (Figure 1). Facilities included medical offices, nursing stations, a meal preparation room, storerooms, and an item transfer room.

Figure 1.

Bed distribution map of the HSCT ward. From no. 1 to no. 20, a total of 18 single-bed laminar air flow (LAF) isolation units (highlighted in yellow) and 6 multi-bed units, (highlighted in green. Numbers in red font are carbapenem-resistant Acinetobacter baumannii (CRAB)–positive patients’ beds. HSCT, haematopoietic stem cell transplantation.

The meal preparation room contained one sink and one refrigerator, while two electric water heaters provided centrally supplied boiled water for drinking and daily hygiene. A stem cell collection and processing room was located adjacent to the ward entrance. Patient and caregiver meals were prepared within the hospital, and all personal belongings, daily necessities, and catering items were disinfected before being transported to the ward.

Each patient was permitted one designated caregiver. Before entering the ward, caregivers underwent viral screening for SARS-CoV 2 and influenza and were required to sign informed confirming adherence to IPC measures, including support with patient hygiene (oral, skin, and perineal care) and restrictions on leaving the ward. Healthcare workers entered the ward only when required for patient care and used appropriate personal protective equipment.

Water supply management in HSCT ward

The hospital’s domestic water supply was provided by the municipal network. Floors 1–6 received direct municipal water, while floors 7 and above, including the HSCT ward on the 21st floor, were supplied through a secondary pressurization system with a storage tank. Each patient unit had a dedicated bathroom with a handwashing sink and shower; however, no water heating devices or water storage faucets were present.

Patients were required to use centrally supplied boiled water for drinking and hygiene activities, including oral, skin, and perineal care; tap water was not permitted for these purposes. Boiled water was prepared in the meal preparation room by heating secondary pressurized tap water to 100°C. The boiled water was stored in uncovered food-grade polypropylene buckets for cooling or transferred directly into kettles for immediate use. At scheduled times each day, the staff transported the cooled boiled water and freshly boiled water to patient rooms and transferred it into patients’ personal kettles with minimal contact.

Environmental surfaces in the meal preparation room were cleaned and disinfected twice daily. Water storage buckets, ladles, and kettles were cleaned before use and disinfected after use with 500-mg/L chlorine-containing disinfectant.

Environmental investigation in HSCT ward

Baseline assessment (4^th January 2021): The initial environmental survey included 141 samples. Microbiological analysis identified 8 Acinetobacter spp. and 17 AB complex isolates (Table S1).

Expanded surveillance (6^th January 2021): Following the initial findings, environmental surveillance was expanded. Among 49 samples collected, 5 Acinetobacter spp. and 8 AB complex isolates were recovered (Table S2).

Postintervention evaluation (12^th January 2021): To assess the effectiveness of IPC interventions, environmental sampling was performed in the stem cell collection and processing room. A total of 44 samples were collected, and neither Acinetobacter spp. nor AB complex was detected (Table S3).

Postintervention evaluation (18^th January 2021): By this time, seven patients had been transferred to the haematology ward, followed by terminal disinfection of the vacated rooms. A second comprehensive environmental survey was conducted across the HSCT ward, including both terminally disinfected and occupied rooms. Of 136 environmental samples collected, two Acinetobacter spp. and 15 AB complex isolates were identified, indicating persistent environmental contamination despite IPC interventions (Table S4).

Overall survey period (4^th–18^th January 2021): Across four rounds of environmental sampling, 369 samples were collected from the HSCT ward over four sampling rounds. A total of 73 Gram-negative strains were recovered, including 42 AB complex isolates. These isolates were recovered from buckets used for cooling and storing boiled water, iron-alloy kettles containing freshly boiled water in the meal preparation room, patients’ drinking water bottles and drinking water, personal digital assistant (PDA) (a mobile handheld device used by nurses for barcode scanning of medical orders), infusion pumps and infusion stands, bed unit surfaces, patients’ sanitary ware, and caregivers’ hands, suggesting widespread environmental contamination and potential cross-transmission of the AB complex (Table II).

Table II.

The distribution of 42 strains of Acinetobacter baumannii complex in the environment of haematopoietic stem cell transplantation ward

Specimen	Number	Proportion (%)
Patient’s drinking water	13	30.95
Patient’s water bottles	12	28.57
Objects of bed units	6	14.29
Patient’s/family members’ hands	4	9.52
Patients’ sanitary ware	2	4.76
Infusion pump and pole	2	4.76
buckets for cooling and storing boiled water,	2	4.76
Shared medical supplies	1	2.38
Total	42	100

Terminal disinfection verification (2^nd February 2021): Following terminal disinfection of all patient rooms, 95 environmental samples were collected. Neither Acinetobacter spp. nor AB complex was detected, demonstrating successful environmental decontamination (Table S5).

Immediate infection control interventions

Following the initial environmental findings on 6^th January 2021, a second round of environmental sampling was undertaken alongside targeted IPC measures. As the AB complex was detected in water-related samples, interventions focussed on the centralized drinking water supply system, specifically the buckets used for cooling and storing boiled water and the iron-alloy kettles used to dispense freshly boiled water. The practise of cooling and storing boiled water was discontinued, and only freshly boiled water was provided to patients for drinking and daily hygiene. Strict hand hygiene was reinforced for healthcare workers and family caregivers, and enhanced cleaning and disinfection of kettles and environmental surfaces were implemented.

On 6^th January, blood, urine, and throat swabs were collected from all patients, including the six colonized patients. All samples were negative for Acinetobacter spp. and AB complex. Repeat surveillance over the following remained negative, confirming successful decolonization.

Over the subsequent three weeks, patients were transferred in batches to two haematology wards, new admissions to the HSCT ward were suspended, and terminal disinfection was performed sequentially in all patient rooms, with standardized decontamination of the remaining ward areas. No new cases were identified between 6^th and 31^st January 2021. The final environmental sampling on 31 January detected neither Acinetobacter spp. nor AB complex in any sample. The HSCT ward subsequently reopened to new admissions, and no additional cases were reported during the following month.

WGSGS of AB

To investigate CRAB transmission within the HSCT ward, WGS was performed on 52 AB isolates, comprising six isolates from the HSCT ward, 16 representative environmental isolates (selected from 42 recovered isolates), and 30 CRAB isolates from patients in 17 other hospital wards collected between December 2020 and March 2021.

cgMLST identified six clonal clusters, with isolates differing by ≤10 alleles considered clonally related (Figure 2). Of the 22 HSCT ward isolates, 18 belonged to cluster A, including five patient and 13 environmental isolates, while four were singletons. Eight of the 13 environmental cluster A isolates were water associated (six from patients’ drinking water, one from a buckets used for cooling and storing boiled water, and one from a patient’s water bottle). The remaining isolates were recovered from shared nursing PDAs, infusion pumps and stands, bed unit surfaces, patients’ sanitary ware, and caregivers’ hands (Table III). Although the isolate from patient 2B was a singleton, it differed by only 15 allele from the nearest cluster A isolate.

Figure 2.

Fifty-two CRAB phylogenetic trees. Samples 1B–6B isolates were from 6 post-HSCT recipients, and E1–E16 AB isolates were environmental samples from the HSCT ward; 7B–36B isolates were from 30 patients in 17 other wards. The phylogenetic trees were classified based on allele differences ≤10, resulting in a total of A–F, clonal clusters. Each cluster was colour coded and enclosed with dotted lines, ensuring that each cluster contained at least 3 isolates. CRAB, carbapenem-resistant Acinetobacter baumannii; HSCT, haematopoietic stem cell transplantation.

Table III.

MLST and cgMLST typing of 22 Acinetobacter baumannii in haematopoietic stem cell transplantation ward

Genotyping	Specimen	Number
cgMLST
Cluster A
	Patient’s drinking water	6
	Patient’s throat swab	4
	Patient’s urine	1
	cooling and storing boiled water, bucket 1	1
	Patient’s water bottle	1
	PDA (a mobile handheld device used by nurses for barcode scanning of medical orders)	1
	Infusion pump and pole	1
	Objects of bed units	1
	Patients’ sanitary ware	1
	Family members’ hands	1
Non-cluster
	Patient’s drinking water	2
	cooling and storing boiled water, bucket 3	1
	Patient’s throat swab	1
MLST
ST958
	Patient’s drinking water	8
	cooling and storing boiled water, buckets 1 and 3	2
	Patient’s throat swab	1
	Patient’s water bottle	1
	Infusion pump and pole	1
	Objects of bed units	1
	Patients’ sanitary ware	1
	Family members’ hand	1
ST1806
	Patient’s urine	1
	Patient’s throat swab	2
STnovel
	Patient’s throat swab	2
	PDA (a mobile handheld device used by nurses for barcode scanning of medical orders)	1

MLST, multi-locus sequence typing; cgMLST, core genome multi-locus sequence typing; PDA, personal digital assistant.

MLST showed that cluster A comprised 13 ST958 isolates, three ST1806, and two novel sequence types (STs); while three of the four singleton isolates were also ST958 (Table III). The temporal distribution and specimen sources of the HSCT isolates are shown in Figure 3. These findings demonstrate dominant clonal CRAB strains in the HSCT ward, with evidence of cross-transmission between patients and the hospital environment. Contamination of the buckets used for cooling and storing boiled water during the centralized water supply process likely served as a key environmental reservoir.

Figure 3.

Distribution of 22 CRAB samples in the haematopoietic stem cell transplantation ward for cgMLST analysis. Yellow, samples of cluster A; blue, samples not part of cluster A. CRAB, carbapenem-resistant Acinetobacter baumannii; cgMLST, core genome multi-locus sequence typing.

ST958 was the predominant ST in the HSCT ward, whereas ST1806 was the most common among the 30 CRAB isolates recovered from 17 other wards (17/30; Table III). cgMLST assigned these isolates to five distinct clonal clusters (clusters B–F), all genetically distinct from cluster A (Table S6), indicating limited spread of the HSCT outbreak clone beyond the ward.

Seventeen of the 30 isolates (56.7%) belonged to clusters B–F, with each cluster including patients from two to three wards (Table S7). Nine of these 17 patients (52.9%) had been previously hospitalizations at other healthcare facilities, suggesting both intrahospital and interhospital transmission of CRAB.

Discussion

This study identified contaminated buckets used for cooling and storing boiled water during the centralized drinking water supply process as the likely environmental reservoir facilitating dissemination of a clonal CRAB strain within the HSCT ward. Importantly, neither the municipal water supply nor the hospital’s secondary pressurization system was implicated; contamination was confined to the ward’s centralized drinking water distribution process.

Environmental contamination is a well-recognized reservoir for CRAB outbreaks. Previous studies have identified multi-drug-resistant AB on computer keyboards and ventilator condensate in a paediatric intensive care unit [18]. Similarly, during a CRAB outbreak in a UK neurosurgical ward, the outbreak strain was recovered from 53% of environmental surfaces [6]. In another neurosurgical intensive care unit, contaminated faucet aerators generated splashes during handwashing by healthcare personnel, leading to environmental contamination and subsequent patient colonization [19]. However, reports implicating drinking water containers as the primary source of CRAB outbreaks are scarce [14].

Previous studies have reported AB in sinks, faucets, and even hospital water supplies, resulting in HAI outbreaks [[20], [21], [22], [23], [24], [25], [26]]. To our knowledge, this is the first study to demonstrate genetically related AB isolates in patients’ drinking water, drinking water containers, and patients within an HSCT ward, providing evidence of transmission between these reservoirs and the healthcare environment.

The persistence of CRAB in water-related reservoirs and on shared medical equipment highlights the importance of robust water management programs, strict environmental cleaning, and standardized disinfection of reusable equipment in units caring for immunocompromised patients. Elimination of environmental contamination following terminal disinfection further supports the effectiveness of these interventions in controlling CRAB transmission.

WGS confirmed a dominant clonal CRAB strain circulating within the HSCT ward and demonstrated cross-transmission between patients and the environment. Although ST208 has previously been reported as the predominant clone in China, ST958 predominated in our outbreak, highlighting regional variation in circulating CRAB lineages [16,27]. These findings also demonstrate the value of cgMLST, which provides greater discriminatory power than Pulsed-field gel electrophoresis (PFGE) and conventional MLST for investigating CRAB outbreaks and transmission pathways [28,29].

This study has several limitations. Resistance genes were not analysed for the 52 sequenced isolates, AB samples were not analysed, limiting characterization of the molecular mechanisms underlying antimicrobial resistance. In addition, environmental sampling did not include all potential water reservoirs, such as the hospital water storage tank, faucet aerators, sink and shower drains, or the incoming municipal water supply. Future investigations should incorporate comprehensive sampling of both hospital water systems and point-of-use water sources to better define contamination pathways and guide targeted prevention strategies.

In conclusion, our findings demonstrate that contaminated cooled boiled water storage buckets served as reservoirs for CRAB transmission, thereby facilitating cross-transmission between patients and the ward environment in the HSCT ward.

CRediT authorship contribution statement

Xiaoyu Luo: Writing – review & editing, Writing – original draft, Data curation. Yinghua Zhang: Writing – review & editing, Writing – original draft, Methodology. Shihui Guo: Investigation, Data curation. Qiaoling Wei: Investigation, Data curation. Qunxiu Guo: Investigation, Data curation. Lanlan Li: Investigation, Data curation. Jingjing Ya: Investigation, Data curation. Juan Huang: Investigation, Data curation. Jinai Qin: Supervision, Project administration, Funding acquisition, Conceptualization.

Data availability

The whole-genome sequencing data in this study have been deposited in GenBank under BioProject ID PRJNA1182049.

Ethical approval

Ethical permission was approved by the Ethics Committee of the First Affiliated Hospital of Guangxi Medical University, 530,021, Nanning, Guangxi, China, with the reference no. 2024-E0873.

Conflict of interest statement

All authors declare that they have no actual or potential competing or financial interests or personal relationships that might lead to a conflict of interest.

Appendix A. Supplementary data

The following is the Supplementary data to this article: