Prolonged carriage of ESBL-producing enterobacterales and potential cross-transmission among residents in geriatric long-term care facilities

Previous studies indicated residents in geriatric long-term care facilities (LTCFs) had much higher prevalence of extended-spectrum β-lactamase-producing Enterobacteriaceae (ESBL-E) carriage than the general population. Most ESBL-E carriers are asymptomatic. The study tested the hypothesis that residents with ESBL-E carriage may accumulate inside geriatric LTCFs through potential cross-transmission after exposure to residents with prolonged ESBL-E carriage. 260 residents from four Japanese LTCFs underwent ESBL-E testing of fecal specimens and were divided into two cohorts: Cohort 1,75 patients with ≥ 2 months residence at study onset; Cohort 2, 185 patients with < 2 months residence at study onset or new admission during the study period. Three analyses were performed: (1) ESBL-E carriage statuses in Cohort 1 and Cohort 2; (2) changes in ESBL-E carriage statuses 3–12 months after the first testing and ≥ 12 months after the second testing; and (3) lengths of positive ESBL-E carriage statuses. Compared with the residents in Cohort 1, a significantly larger proportion of residents in Cohort 2 were positive for ESBL-E carriage (28.0% in Cohort 1 vs 40.0% in Cohort 2). In the subsequent testing results, 18.3% of residents who were negative in the first testing showed positive conversion to ESBL-E carriage in the second testing, while no patients who were negative in the second testing showed positive conversion in the third testing. The maximum length of ESBL-E carriage was 17 months. The findings indicated that some residents acquired ESBL-E through potential cross-transmission inside the LTCFs after short-term residence. However, no residents showed positive conversion after long-term residence, which indicates that residents with ESBL-E carriage may not accumulate inside LTCFs. Practical infection control and prevention measures could improve the ESBL-E prevalence in geriatric LTCFs.

The present study focused on asymptomatic LTCF residents with carriage of ESBL-E identified in their feces. In previous studies, LTCF residents had much higher prevalence of ESBL-E carriage than the general population 12,13 . We hypothesized that in geriatric LTCFs, asymptomatic residents with ESBL-E carriage may accumulate through potential cross-transmission after exposure to residents with prolonged ESBL-E carriage. To test this hypothesis, we performed multiple ESBL-E testing with long-term follow-up for continuous residents in LTCFs to assess (1) the proportion of residents who acquired ESBL-E through potential cross-transmission inside the facilities and (2) the length of ESBL-E carriage among ESBL-E-positive residents.

Methods
Study settings and participants. We conducted a cohort study among residents receiving long-term care in 4 geriatric LTCFs in Japan. The facilities were selected because they had secure linkage with their own specific back-up hospitals where residents were typically transferred for any medical needs. All 4 LTCFs were anonymized in accordance with the ethics protocol of the study. Brief profiles of the facilities are shown in Table 1.
The study participants were 260 older adults residing in the 4 LTCFs who underwent ESBL-E carriage testing during the study period (August 2018 through March 2020). Residents who did not undergo ESBL-E testing were excluded from the study. Background data for the 260 residents, such as age, sex, and general condition, were not obtained because of the ethics protocol employed in the study.
In Japan, geriatric LTCFs are classified into two main types: (1) geriatric health services facilities and (2) geriatric special nursing homes. The former are intermediate facilities between hospitals and nursing homes, with a primary focus on rehabilitation. These facilities typically have a goal of returning patients to home-based care, although some residents may require long-term care for years. The latter provide daily life support, including endof-life care [23][24][25] . The study included three geriatric health services facilities and one geriatric special nursing home. Study design. The first ESBL-E testing was performed on all 260 residents < 1 month after study onset or initiation of residence. The 260 residents who underwent the first ESBL-E testing were classified into two cohorts ( Fig. 1A): Cohort 1 contained residents who had been residing in an LTCF for ≥ 2 months at study onset, and Cohort 2 contained residents who had been residing in an LTCF for < 2 months at study onset or initiated residence in an LTCF during the study period. First, we compared the ESBL-E carriage statuses (positive/negative) between residents in Cohort 1 and Cohort 2 (Fig. 1A, Analysis Part 1).
Second, we assessed the changes in ESBL-E carriage statuses in residents from the first testing who subsequently underwent a second testing and/or a third testing (Fig. 1B, Analysis Part 2). The second testing was performed 3-12 months after the first testing and the third testing was conducted ≥ 12 months after the second testing for residents who had been continuously residing in the LTCFs. The second testing and third testing were not performed in residents who had been discharged from the LTCFs at the time of each testing. In this analysis, we identified the proportions of residents who had changed from negative to positive for ESBL-E carriage status in the results for the second testing and third testing.
Third, we assessed the lengths of ESBL-E carriage from one testing to the next testing among residents who underwent the second testing and third testing (Fig. 1B, Analysis Part 3). In this analysis, we measured the durations of positive test results in the same residents. If all three tests revealed positive results, the lengths of ESBL-E carriage were measured from the first testing through third testing. The range (minimum to maximum) for the lengths of ESBL-E carriage were determined.  26 . This medium was able to grow ESBL-E simultaneously with carbapenem-resistant Enterobacteriaceae (CRE) and/or AmpC β-lactamase-producing bacteria. Following aerobic incubation at 37 °C for 18-24 h, colonies growing on either side of the adapted screening agar were regarded as suspect for any of the above bacteria, and underwent further analysis. ESBL-E and AmpC β-lactamase-producing bacteria were identified using an ESBL + AmpC Detection Set™ (MAST Group, Bootle, Merseyside, UK). CRE was identified using a Vitek™ 2 automated instrument for infectious disease testing (bioMérieux SA, Marcy l'Etoile, France) for its resistance against imipenem and meropenem. Bacterial species were identified by the Vitek™ 2 automated instrument (bioMérieux); these species were subsequently confirmed using Matrix-Assisted Laser Desorption/ Ionization Time of Flight Mass-Spectrometry (MALDI-TOF MS; Bruker Daltonics Inc., Bremen, Germany) 27 . Additional analyses were performed to determine the clonal relationships between bacterial isolates identified in the first and second ESBL-E carriage testing. First, the Vitek™ 2 automated instrument was preliminary employed to test antimicrobial susceptibilities to 10 agents with potential activity against ESBL-E: tazobactam/ piperacillin, cefmetazole, ceftazidime, cefepime, aztreonam, gentamicin, amikacin, ciprofloxacin, levofloxacin, and trimethoprim/sulfamethoxazole. When the minimal inhibitory concentrations against an antimicrobial agent differed more than fourfold among bacterial isolates, these strains were determined as having different susceptibilities. Bacterial isolates with three or more different antimicrobial susceptibilities were determined as different strains. In this manner, we identified strains with a similar pattern of antimicrobial susceptibility between the first and second ESBL-E carriage testing. Pulsed-field gel electrophoresis (PFGE) were then employed to characterize the clonal relationships among these strains 28 . Genomic DNA was digested with XbaI enzyme (Takara Bio, Otsu, Japan), and the resulting DNA fragments were separated by PFGE in a CHEF-DR III system (Bio-Rad, Hercules, CA, USA) with pulses ranging from 5.3 to 49.9 s for 19.7 h at 6 V/cm and 14˚C. The banding patterns were analyzed with UPGMA clustering method with CLIQS™ (Totallab Ltd, Newcastle, UK) 29 .
Analysis part 1. Cohort 1 and Cohort 2 contained 75 and 185 residents, respectively ( Of the 36 residents who underwent both the second testing and third testing ( Fig. 2; Table 5), 6 (16.7%) were positive for ESBL-E carriage in the third testing. Among the 11 residents who were positive for ESBL-E carriage in the second testing, 6 (54.4%) remained positive in the third testing. Furthermore, among the 25 residents negative for ESBL-E carriage in the second testing, no residents showed positive conversion in the third testing.     www.nature.com/scientificreports/ Among 33 residents who were positive for ESBL-E carriage in both first and second testing, 22 (66.7%) were carriers with specific bacterial strains whose antimicrobial susceptibilities were similar between the first and second testing: 44 strains in 22 pairs (a pair comprised strains isolated in the first and second testing). All strains were E. coli. PFGE was subsequently employed to compare the clonality of these strains, resulting in the detection of 15 pairs with each similar PFGE pattern. Specific E. coli strains with a similar specific PEGE pattern were found in 3 different residents in Facility C (Fig. 3).

Discussion
The present study had three key findings. First, the prevalence of ESBL-E carriage was significantly higher in Cohort 2 than in Cohort 1 (40% vs 28%), which indicates that ESBL-E was most likely to be carried into LTCFs through admission of new residents, who provided a large contribution to the increased total prevalence of ESBL-E carriage among LTCF residents. Second, 18% (15/82) of residents showed positive conversion to ESBL-E carriage in the second testing, which indicates that some residents may have acquired ESBL-E through potential cross-transmission inside the LTCFs after short-term residence (3-12 months after admission). Additional PFGE analysis indicated that 3 different residents in the same LTCF were ESBL-E carriers of a specific strain whose clonality was similar. Third, some residents maintained ESBL-E carriage for > 1 year (maximum, 17 months), which indicates that these residents may be high-risk triggers for outbreaks inside LTCFs through resident-toresident transmission as a potential ESBL-E reservoir. However, we subsequently found that no residents exhibited positive conversion to ESBL-E carriage after long-term residence (> 12 months after admission). This finding indicated that, from the long-term perspective, residents with ESBL-E carriage were less likely to accumulate inside LTCFs. Thus, the present findings highlight the importance for appropriate implementation of practical ESBL-E infection control and prevention measures by care providers in geriatric LTCFs, with the expectation of disappearing ESBL-E from the facilities.
Using the exact same LTCF settings employed in the present study, we recently reported that the prevalence of methicillin-resistant Staphylococcus aureus (MRSA) in the nasal cavity of LTCF residents was approximately 10% 23 . However, the prevalence of ESBL-E carriage among residents was much higher in the present study (36.5%; 95/260 residents). We further reported that MRSA may be imported into LTCFs via transfer of residents rather than spread by potential cross-transmission inside LTCFs 23 . Similar to MRSA, the present study indicated that ESBL-E were most likely to be carried into LTCFs through admission of new residents. However, major differences between MRSA and ESBL-E may be the frequency of potential cross-transmission and the length of carriage. Compared with MRSA 23 , ESBL-E was more likely to become widespread inside geriatric LTCFs through potential cross-transmission among residents and less likely to disappear spontaneously after shortterm residence. Our additional PFGE analysis results support potential cross-transmission among residents. These differences may contribute to the higher prevalence of ESBL-E compared with MRSA. Given that ESBL-E can be transmitted during the excrement disposal process or fecal contamination, care providers in geriatric LTCFs should pay thorough attention to adherence for infection prevention, especially for residents requiring diaper disposal.
The present results indicated that residents with ESBL-E carriage may diminish during long-term residence. This finding was supported by a previous study. Overdevest et al. 11 conducted a surveillance study and suggested that ESBL-E could be predicted to disappear from LTCFs over time. They also reported that the lengths of ESBL-E carriage differed in accordance with the strain types, with ESBL-Escherichia coli of sequence type O25:ST131 having the longest carriage period before its disappearance from LTCFs 11 . In the present study, ESBL-Escherichia coli was only identified in residents with prolonged ESBL-E carriage for > 12 months (data not shown), although we did not identify sequence types. We further found that 17% of residents had acquired ESBL-E through potential cross-transmission inside the LTCFs within short-term residence after admission. Although there are no previous studies to support this finding, some possible hypotheses can be proposed. First, new residents admitted to LTCFs may have had multiple risk factors for ESBL-E acquisition, such as episodes of recent antibiotic use and/or previous hospitalization [12][13][14][15]18,19,30,31 . Our selected LTCFs had their own linkages with specific back-up hospitals. Second, ESBL-E may have been most infectious immediately after being carried into the facilities, and then gradually become less infectious. This may be associated with the duration from previous antibiotics use. Importantly, our findings highlight that the risk of ESBL-E acquisition inside geriatric LTCFs may further increase with admission of large numbers of new residents.
Care providers in geriatric LTCFs should consider that residents are most likely to have ESBL-E in their feces on their initial admission to the facilities, especially those admitted from hospital settings. Furthermore, residents with short-term residence should be considered as high-risk residents for ESBL-E acquisition. A patient traceability with alarm system for ESBL-E carriage between specific back-up hospitals and their receiving LTCFs may be effective in preventing ESBL-E transmission inside LTCFs. However, the standard precautions, such as thorough hand hygiene and appropriate use of gloves, should be of the greatest importance for all care providers in geriatric LTCFs. A previous study showed that enhanced infection control measures mainly based on thorough adherence to standard precautions led to subsidence of an outbreak of ESBL-producing bacteria inside an LTCF 32 . Several intervention programs may improve adherence to standard precautions among care providers in geriatric LTCFs 2,3,33 . In particular, care providers should be careful about daily care in the excrement disposal process and fecal contamination for residents requiring diapers. Such residents comprised the majority in our study settings. To our knowledge, no previous studies have focused on older adults with diapers and infections. This situation may warrant further assessment of infection control and prevention strategies against ESBL-E.
The primary limitation of the present study was incomplete information on the background characteristics of the residents, such as sex, age, general condition, medical history. Among these factors, general condition (activities of daily living and nutritional status) and medical history (comorbidities, use of antimicrobials, and  www.nature.com/scientificreports/ duration of prior hospitalization) may affect the prevalence of ESBL-E carriage 3,6,13,30 , which would result in potential confounders of our findings. Unfortunately, the institutional ethics review board did not grant approval to collect this information without obtaining written informed consent from each resident even though the information was anonymized. To confirm the present findings, further studies should be conducted with complete information for residents, including potential factors associated with ESBL-E carriage status. Second, a large number of residents did not undergo the second testing and third testing. These residents could not undergo the second testing and third testing because they had been already discharged from the LTCFs at the time of the testing. Third, we performed ESBL-E testing only three times with different intervals. Multiple tests with regular intervals and longer follow-up may be required to accurately assess the status of ESBL-E carriage. Fourth, we were unable to obtain information on the prevalence of ESBL-E carriage in region-specific general populations surrounding each LTCF and among patients hospitalized in each specific back-up hospital, which may have affected our results. Fifth, the study participants comprised a small number of residents from a limited number of LTCFs. Studies with larger numbers of residents from a greater number of LTCFs located in various regions (both rural and urban areas) are warranted to further confirm our findings. Sixth, we could not identify sequence types of specific bacterial strains. Finally, a standard infection prevention and control protocol was not formally established in all four LTCFs during the study period, which may have affected the prevalence of ESBL-E carriage.

Conclusions
ESBL-E was most likely to be introduced into LTCFs through admission of new residents. Furthermore, some residents acquired ESBL-E through potential cross-transmission inside the LTCFs within a short period (3-12 months) after admission. Although ESBL-E carriage could be maintained for > 1 year, no residents exhibited positive conversion to ESBL-E carriage upon long-term residence of > 12 months after admission, which indicates that residents with ESBL-E carriage may not accumulate inside LTCFs. Practical infection control and prevention by care providers could decrease the prevalence of ESBL-E in geriatric LTCFs.