CRE: Carbapenem-resistant Enterobacteriaceae are resistant to carbapenems, regardless of mechanism (includes ESBL, AmpC, porins, carbapenemase production);

CPE: Carbapenemase-producing Enterobacteriaceae have a genetic element that codes for carbapenemase production e.g. NDM, KPC;

CRO: Carbapenem-resistant organisms include all gram negative organisms and all resistance mechanisms;

CPO: Carbapenemase-producing organisms includes all gram negative organisms that produce a genetically coded carbapenemase [8] .

A cross-section study was implemented in South Vietnam from September 2022 to May 2023 at The Care and Protection of Orphaned Children Center, whose area is about 2000 m², including a building consisting of three floors in which the first and general floors are sections used as bedrooms and patient-care section. On the general floor, there is a section used for preparing food in the kitchen section. There is a small chicken and vegetable farming section to supply food for feeding patients, almost children, residing in this center because they are orphans.

The majority of children in this center have a disability related to cerebral palsy; many of the children have experienced complications of immobility and infection, to be admitted to local hospitals and exposed to broad-spectrum antibiotics. This center has been operating since 2000 and caring for children of the range of ages from 4 to 26 years old. In this study, we used the word resident for the children living at this center.

All 20 healthcare workers (HCWs) and 67 residents (including children) participated in the study. We observed and determined 303 environmental samples in the Center of Care and Protection of Orphaned Children. But we randomly chose some samples in the environment, based on the Yamane formula [16] , was n = 175, $n=\frac{N}{1+N\cdot {\text{e}}^2}$, Where n = sample size required (n = 175), N = the total number of environmental samples was 307 (N) observed, and determined in the center of care and protection of orphan children, e = 0.05, allowable error (%).

This study received an approval from the Board of Directors on Ethics in Biomedical Research at Thien Phuoc Nhan Ai Center of Care and Protection of Disabled Children, Vietnam, under the approval number 026/2565. The written informed consent was obtained from HCWs and the guardians of children.

Most of the children residing at the center have a physical and (or) mental disability. The residents’ guardians participated in interviews to collect information. The HCWs at the Center are Catholic Sisters who provide nursing and personal care (wound care, bathing, feeding etc.) in 8 - 10 hour shifts and six days a week. All residents and staff of the center were included in this study (20 HCWs and 67 residents (most of them were children)).

For HCWs, all HCWs were female 100%, while the prevalence of female patients was 26.9 %. The range of HCW’s age was from 20 to 66 years old (Max = 66, Min = 20). Almost all HCWs were over or equal to 27 years old, and one was 20, while the children’s age was from 4 to 26 years old.

Most HCWs graduated high school (n = 9, 45%). Of 20 HCWs, two graduated (10%) from the Nursing College and were registered nurses, one (5%) was a pharmacist who graduated from Pharmacy College graduation, twelve children carers, one cleaner, two Food providers, one guard, one manager of the Center, one pharmacist, and two registered nurses.

For HCWs, all HCWs were female 100%, while the prevalence of female patients was 26.9 %. The range of HCW’s age was from 20 to 66 years old (Max = 66, Min = 20). Almost all HCWs were over or equal to 27 years old, and one was 20, while the children’s age was from 4 to 26 years old. (92.5% (n = 62) were under 20 years old and 7.5% (n = 5) were from 20 to 26 years old).

Most HCWs graduated high school (n = 9, 45%). Of 20 HCWs, two graduated (10%) from the Nursing College and were registered nurses, one (5%) was a pharmacist who graduated from Pharmacy College graduation, twelve children carers, one cleaner, two Food providers, one guard, one manager of the Center, one pharmacist, and two registered nurses. Most residents (children) haven’t approached the primary school graduation yet.

For residents (N = 67), 10% are children with normal health, while 90% have mental and physical disabilities ( most of them are children with cerebral palsy and other disorders (epilepsy, hydrocephalus, hyperactivity, mental retardation, osteogenesis imperfecta).

The proportion of HCWs with a duration of years for professional work of HCWs under five years was 50%, and 50% of HCWs have worked for more than five years, while more than 67% of residents have resided for over five years in the Center of Care and protection Of Orphan Children. This result implied that HCWs and orphan patients have lived, worked, and had many opportunities to contact for a long time.

There was 100% HCWs have worked seven days per week and 90% worked over 8 hours per day. This remarkable finding showed that HCWs had much time to contact patients in performing healthcare practice to patients in this healthcare setting. Thus, CRO transmission happened as an evitable event while HCWs cared for and treated patients at this center (as shown in part 3, 5 and 6 of Table 1 ).

CC = Children carer, Cln = Cleaner, FD = Food provider, Gd = Guard, Mgr of Ct = Manager of Center, Pharm = Pharmacist, RN = Registered nurse, No OCCU = No occupation; Drt of PW (HCW)/Rsd (Yrs) = Duration of Professional work (HCW)/residents (years), 7 ds/p W = 7 days per week, PW = Professipnal work, N of hrs for PW per D = Number of hours for professional work.

Exclusion criteria were defined as non-consent, known CRO/CPO colonization or infection in 6 months prior to the study or active infection at the time of the study. However, none of the residents or HCW fulfilled any exclusion criteria hence all were included.

Residents and HCW were allocated a participant code, and isolates were allocated a laboratory code for tracking.

The samples from residents and HCW were collected on a sterile rectal cotton sampling swab inserted approximately 1 cm into the anal canal and rotated for 10 seconds. The swab was then put into individual vials of Amies transport medium and transported on dry ice to a reference hospital within 6 hours of collection.

The environmental samples included surfaces and furnishings in the center using a sterile cotton swab. An area of approximately 10 cm × 10 cm was sampled on environmental surfaces. The swabs were placed in Amies transport medium and processed in the same way as for the rectal swabs.

We showed the relationship between variables, the characteristics in our study in the Conceptual Framework as shown in Figure 1 .

The sample swabs were plated onto MELAB Chromogenic CARBA agar plates and incubated at 37o Celsius. Purity cultures of suspect colonies were performed on blood agar and followed species identification by Card NID, antimicrobial susceptibility testing, and phenotypic carbapenemase detection using the NMIC-500 CPO Detect panel of the BD Phoenix TM Automated Microbiology System.

The reference strains in the NMIC 500 test to confirm susceptibilities are E. coli ATCC® 25922, K. pneumoniae ATCC BAA-1705^TM, and P. aeruginosa ATCC^TM 27853.

We discovered thirty-six CRO isolates in HCWs, residents, and the environment. However, we chose twelve CROs detected from HCWs and residents who had close contact with HCWs: HCWs feed food or bath residents, etc. CROs belonged to the same species with a high ratio of identical antibiotic-resistant phenotypes between CRO isolates and the number of CROs chosen to perform WGS available to our study financial resources. So, we selected twelve CRO isolates from HCWs and residents (as shown in Table 2 ), to determine whether the potential CRO transmission/contamination occurred between HCWs and residents in these chosen CROs. Extraction and concentration of DNA was performed in Vietnam (as shown in Table 2 ) and the samples sent to Charles River Laboratories in Australia for sequencing.

“I” means intermediate and “R” means resistance, “X” means the MICs of antibiotics are in the range of sensitive or intermediate, “N” is an antibiotic not recommended for the treatment of infections.

The genomic DNA was quantified with Qubit High-sensitivity assays using a Qubit 4.0 Fluorometer (Thermo Fisher Scientific, USA). The samples were purified using AMPure XP (Beckman Coulter, USA) with a 1X volume ratio to elute samples with low EDTA Tris-HCl buffer. Samples were normalized to 50-200ng and processed using the Illumina DNA Library Prep kit (Illumina, USA) according to the manufacturer’s protocols using 5 PCR cycles for indexing. The performance of Quantification and size estimation of the libraries were on both the Qubit 4.0 Fluorometer (Thermo Fisher Scientific, USA) and the 4200 Tape Station System (Agilent, USA). The samples were normalized to 4nM and then pooled into a new microfuge tube. The pooled library was diluted to 1nM and sequenced on the MiniSeq Sequencer (2 × 150 bp paired-end reads) (Illumina, USA) using MiniSeq High Output 300 Cycle flowcell.

To determine CRO transmission, WGS was performed on 12 isolates from residents and HCWs who had close contact (feeding/bathing/care), comparing isolates of the same species with similar antibiotic-resistant phenotypes (≥ 80%) profiles. In detail, we used 25 antibiotics to determine the susceptibility or resistance of CROs in our study ( Table 2 ). So, the ratio of identical antibiotic-resistant phenotypes

Moreover, we added two random outgroups of Escherichia coli (FORC_044 and strain_2313) and used K12 as the reference genome for SNP calling. Most strains (6276064, 6276065, 6276066, 6276074, FORC_044, strain_2313) have from 40,000 to 80,000 different SNPs pairwise between any two genomes. Therefore, they are unrelated isolates. Two samples (6276065 (H006) and 6276066 (C4003)) had only 16 different SNPs and are closely strongly related, and the number of genomes of E. coli is about 4,608,319 genomes. Hence, the SNP difference between 6276065 and 6276066 was 0.0003% (16/4,608,319). So, the genomic identity between 6276065 and 6276066 was 99.9997% (100% - 0.00034%). This result is evidence of a cross-contamination event between H006 and C4003.

Based on the Phylogram of E. coli isolates of Figure 2 , the 6276065 (H006) and 6276066 (C4003) isolates were from an identical clone, as described on Figure 2 . This is clear evidence for E. coli cross-contamination/transmission between a child and a healthcare worker. There was a cross-contamination event between H006 and C4003.

To determine if E. coli transmission between the 6276065 (H006) and 6276066 (C4003) were linked epidemiologically we investigated the characteristics of Demography/ antibiotic use/ procedures related to and performed on the H006 (an HCW) and a patient (C4003) (as shown in part (a) (H006 (an HCW) and a resident (C4003)) in Table 5 ).

1) Characteristics of Demography/antibiotic use/procedures related to and performed on H006 (HCW) and C4003 (resident, child):

The H006 did not have a history of antibiotic use of around one year, hospital admission out of the center, surgical operation, or peripheral or urethral catheter use before participating in our study. In contrast, C4003, 5-year-old female child, was admitted to a pediatric hospital to treat pneumonia, and the type of antibiotics used to treat pneumonia was Augmentin 1000 mg two times around 60 days before participating in our study (as shown in Table 5 ).

2) There was antibiotic resistance loss of gentamycin and norfloxacin in the child sample (as shown in Table 2 )

BEF = Before.

This data comes from annotation of the genomes using Patric ( https://www.bv-brc.org/ ).

3) The child sample has gained four virulence genes that lost antibiotic resistance (as shown in Table 7 ).

4) The mutation effect of ydbK changed several amino acids and might cause a change in antibiotic resistance.

From these points of view determined above, the gain of antibiotic resistance made a lot more sense:

We can list some significant notes, including

a) There were very few mutations, four mutations, between the C3004 isolate and HCW (H006) isolate. It suggested that a short evolution time existed between two isolates.

b) Strong selective pressure (for example, antibiotic therapy) for this functional mutation happened shortly. It suggested the HCW (H006) isolate gained antibiotic resistance from an ancestor, which was antibiotic-sensitive.

c) Child sample has maintained the ancestral sensitivity. HCW (H006) and (C4003) might have been contaminated from a common source (Child-infected HCW) or conversely. It meant that if the child had the ancestral strain (which was also most similar to the public reference genome), It suggested the infection from the Child (C4003) to HCW (H006) OR a common infection source.

Continuously, we compared the genomics and applied the MALDI-based typing of the 12 CRO samples. Specifically, we focused on the CRO samples, as shown in Table 2 . We suspected that some infections may be transmission/contamination events within an orphanage due to antibiotic profiling of these isolates demonstrated high resistance ( Table 2 ).

To track the CRO transmission/contamination in six A. baumanii isolates, with WGS code as shown in Table 7 underwent with WGS. We applied the Snippy 4.6.0 software to count SNPs (single nucleotide polymorphism) different between the suspected isolates and the Orthoani software tool for ANI (average nucleotide identity) comparison to determine whether two or more isolates are the same or differentiated species. We also used the Acinetobacter nosocomialis M2 reference genome (GCF_005281455.1_ASM528145v1) to compare these isolates. The number of genomes of Acinetobacter nosocomialis is 3,940,614 genomes.

So, the rate of genome different (as shown in Table 8 ).

It means that the rate of genome identity

Lee’s Study showed that when the ratio of ANI (average nucleotide identity) of two or more strains with the same species is over > 95%, two or more species belong to a unique clone [17] . Based on research results of Lee about the rate of ANI, our study results determined that the 6276067_ isolate was A. seifertii, the 6276069_isolate was A. baumannii, the 6276071, 6276070, and 6276073_isolate belonged to an identical clone.

To improve the taxonomic accuracy of the Acinetobacter sp. samples typing of three isolates, including the 6276071, 6276070, and 6276073_isolate, we calculated the ratio of average nucleotide identity (ANI) of these three isolates by using Ortho ANI.

The ratios of ANI between 6276070, 6276071, 6276073_isolate and A. nosocomial were 97.6676, 97.4891, and 97.8075% ( Table 9 ). All these ratios were over 95% [17] , while the ratios of ANI between 6276070, 6276071, 6276073_isolate, and A. baumanii (or A. seifertii) were below 95%. So, the 6276070, 6276071, and 6276073_isolate were A. nosocomialis and belonged to an identical clone. This data clearly shows that 6276070 (C3022), 6276071 (C3011) and 6276073 (H009) are very closely related and link health care worker H009 to residents (children) C3022 and C3011 by a contamination event.

The results of phylogram of A. nosocomialis isolates, including 6276070 (C3 022), 6276071 (C3 011) and 6276073 (H009) showed all 3 isolates belonge to an identical clone (as described in Figure 3 ).

Based on part (a): 6276070 and part (b): 6276071, and 6276073 in Table 5 , three isolates, including 6276070 (C3 022), 6276071 (C3 011), and 6276073 (H009), did not have risk factors: antibiotic use, hospital stay in and out of Center of Care and Protection of Orphan Children, the past surgical operation, peripheral and urethral catheter use. These three isolates were closely associated with a contamination event described above. It suggests that the origin of transmission may come from the environment.

We detected some carbapenem-resistant genes in 12 CRO samples, including the OXA-48 family, TEM family, CTX-M family, NDM family, OXA-1 family, OXA-51 family, ADC family (Acinetobacter spp), and other antimicrobial-resistant genes, and show in detail the antimicrobial-resistant genes in another paper.