Vascular access infection by Staphylococcus aureus from removed dialysis accesses

Abstract Hemodialysis patients are particularly vulnerable to Staphylococcus aureus infection, with the vascular access serving as the site of entry for this formidable pathogen. Patients with arteriovenous grafts (AVGs) and tunneled‐cuffed catheters (TCCs) are at elevated risk of S. aureus infection. In this study, we investigated the correlation between the clinical characteristics of S. aureus vascular access infection (VAI), molecular profiles, and the biofilm formation abilities of clinical isolates of S. aureus. We collected samples of methicillin‐resistant S. aureus (MRSA), methicillin‐sensitive S. aureus (MSSA), and methicillin‐sensitive S. argenteus (MSSAg) from patients with S. aureus VAI and patients with other infections. The molecular profiles of the clinical isolates were determined using disk diffusion testing and molecular typing. The biofilm formation ability was determined by microtiter plate assay. In total, 63 S. aureus and 10 S. argenteus isolates were identified: 40 MRSA, 23 MSSA, and ten MSSAg. MRSA was highly prevalent (77.8%) in TCC isolates and was multidrug resistant. Of the 40 MRSA isolates, ST239‐SCCmec III was the predominant clone. SCCmec type IV was the predominant type (35%) in isolates from AVGs, while SCCmec type III was prevalent in TCC infection and showed significantly higher biofilm formation ability than types IV and V. In dialysis VAI by S. aureus, patients with TCC were more often infected with MRSA than patients with AVG, and MRSA in TCC–VAI was predominantly SCCmec type III, which had the strongest drug resistance and biofilm formation ability.

soft tissue, and device-related infection (Tong, Davis, Eichenberger, Holland, & Fowler, 2015). S. aureus can survive under extreme conditions and form long-lasting colonization in human tissue or biofilms on the surfaces of foreign devices. The biofilm is produced by a combination of host factors (e.g., fibrinogen, fibrin, fibronectin, and extracellular polysaccharides) and microbial products (e.g., glycocalyx or "slime") and has a critical role in bacterial antimicrobial resistance and recalcitrant infection.  (Kang et al., 2015).
In this study, we investigated the association between the clinical characteristics of S. aureus VAI, microbial features (including epidemiologic and SCCmec classification), and the biofilm formation ability of S. aureus clinical isolates from surgical specimens.

| Patient characteristics
This study was performed at a territory referral hospital in Taiwan between September 2013 and December 2016. We prospectively collected information on 33 consecutive patients with S. aureus VAI who required removal of AVGs and TCCs. In addition, 14 patients hospitalized between September 2013 and December 2016 with S. aureus infection from diseases other than VAI including empyema, port-A catheter infection, chronic ischemic leg, infected endocarditis, and other surgical wound infections were also analyzed for comparison. Patients with poor compliance and those who declined to be part of this study were excluded. All patients were provided explanations about the procedures individually, and informed consent was obtained prior to performing the procedures.

According to the Centers for Disease Control and Prevention
(CDC), the epidemiologic definition of healthcare-associated infection is a set of infections in which the patients had hospitalization, surgery, dialysis, residence in a long-term care facility, or usage of indwelling catheters in the preceding 12 months. Typically, patients undergoing renal replacement therapies return to community dialysis clinics after their acute illnesses have stabilized. To discuss the possible source of VAI in patients with hemodialysis, the isolates were classified as healthcare-associated hospital-onset (HAHO) if the culture was obtained >3 calendar days after a hospital admission (admission day is considered day 1) or healthcare-associated community-onset (HACO) if the culture was obtained as outpatient or ≤3 calendar days after a hospital admission with major healthcare risk factors (dialysis, hospitalization, surgery, or long-term care residence within 1 year prior to collection of the positive MRSA culture or presence of central venous catheter within 2 days prior) (Gualandi et al., 2018;Nguyen et al., 2013).
In all cases of hemodialysis VAI, antibiotic therapy was initiated with broad-spectrum antibiotics. Based on the results of the culture and sensitivities, conversion to an appropriate antibiotic was indicated. The infected hemodialysis vascular accesses were removed in the event of uncontrolled access-related bacteremia and severe complications (progressing to pseudoaneurysms or necrotizing fasciitis).

| Microbiological investigations
73 bacterial isolates from 47 patients' blood and contaminated device samples were cultured under laboratory standards. The samples were routinely cultured on blood agar at 37°C overnight. Strain identification was performed using standard biochemical (phenotypic) procedures using bioMérieux's API Staph system.

| Antibiotic susceptibility testing
The antimicrobial susceptibility of S. aureus was determined by the disk diffusion method on Mueller-Hinton agar using the following antibiotics: clindamycin, erythromycin, fusidic acid, oxacillin, penicillin, trimethoprim-sulfamethoxazole, and tigecycline in accordance with the standard of the Clinical and Laboratory Standards Institute (CLSI) (CLSI, 2013).

| Genomic DNA extraction
A single colony from a S. aureus isolate and from a S. argenteus isolate were inoculated in brain heart infusion (BHI) broth for 16 hr and 1 ml of overnight culture was harvested using centrifugation at 16,500 x g for 5 min. Bacterial cells were suspended in 1 ml of ultrapure water and heated at 100°C for 15 min. The supernatant containing the DNA was stored at 4°C until further use.

| Erythromycin resistance gene detection
The presence of several antibiotic resistance genes was identified using multiplex PCR with the 16S rDNA gene as an internal control.

| Biofilm formation assay
The biofilm production of S. aureus isolates at different timepoints was determined quantitatively by using a modified microtiter plate assay as described previously (Stepanović, Vuković, Dakić, Savić, & Švabić-Vlahović, 2000). The bacterial isolates were grown in tryptic soy broth (TSB) supplemented with 0.25% glucose at 37°C overnight. aureus reference strain, ATCC 29213, was used as the positive control. Wells were gently washed with sterile physiological saline, fixed using 99% methanol, and dried at room temperature. The adherent cells were stained with 0.1% crystal violet, and the bound dye was dissolved using 33% glacial acetic acid per well. Finally, the optical density (OD) of each well was measured at 570 nm using an ELISA reader. Based on the obtained OD value, the strength of biofilm formation was calculated according to the standard Christensen et al. (1985) cut-off value classification formula as follows: optical density cut-off value (ODc) = average OD of negative control +3 × standard deviation (SD) of negative control. Strains were classified as follows: OD ≤ ODc nonadherent; ODc <OD ≤ 2 × ODc weakly adherent; 2 × ODc < OD ≤ 4 × ODc moderately adherent; and 4 × ODc < OD strongly adherent.   Table 1. The patients in the AVG and TCC groups were more diabetic than other groups. The patients with other sources of infection had similar comorbidities as AVG and TCC groups except diabetic.

| Antimicrobial susceptibility testing
We collected a total of 39 S. aureus isolates from 28 hemodialysis patients (AVGs and TCCs) as well as 24 isolates from 14 patients with other diseases or surgical infections. In addition, ten S. argenteus isolates were collected from five hemodialysis patients. The isolates were confirmed to be MRSA with oxacillin resistance by susceptibility testing and were confirmed to be mecA positive by

| SCCmec typing
The SCCmec type classification of the 40 MRSA isolates is shown in The resistance profiles of the MRSA isolates differed among the various SCCmec types (Tables 2). SCCmec type III isolates displayed a high resistance rate (>80%) to six of the tested antibiotics except for tigecycline (8.3%, 1/12). Moreover, the resistance rate of SCCmec type III isolates to SXT was significantly higher than that of other SCCmec type isolates. All of the SCCmec type II isolates were resistant to clindamycin, erythromycin, oxacillin, and penicillin. 87.5% of the SCCmec type V isolates were resistant to clindamycin and erythromycin as well as 100% were resistant to oxacillin and penicillin.
SCCmec type IV, which is prevalent in AVG isolates was predominantly resistant to β-lactams including penicillin (100%, 14/14) and oxacillin (92.9%, 13/14). Moreover, the macrolide resistance genes that demonstrated resistance to erythromycin, including erm(A), erm(B), erm(C), and msr(A), were prevalent in 58.73% (37/63) of the S. aureus isolates and mainly occurred among MRSA isolates (Table 2). Within the erythromycin-resistant S. aureus collection, the erm(A) gene was predominant (45.95%, 17/37) followed by the erm(C) gene (37.84%, 14/37). In addition, the combination of erm(A) and erm(C) was found in 13.51% (5/37) of the erythromycin-resistant S. aureus isolates and occurred only in SCCmec type II MRSA isolates. Whereas the erm(A) gene was mainly found in SCCmec type III MRSA isolates, erm(B) and erm (C) genes were found in SCCmec type IV and V isolates. Macrolide resistance by efflux due to the msr(A) gene was found in MSSA isolates and one mecA-positive MRSA isolate that was sensitive to oxacillin. Age (
For instance, ST59 isolates carried SCCmec type IV and V elements, whereas ST45 isolates carried SCCmec type II, IV, and V elements.

| Biofilm formation ability
To assess the biofilm formation ability of S. aureus isolates, we conducted a biofilm formation assay in microtiter plates and incubated for 4, 8, and 24 hr. All of the isolates in this study were biofilm producers with strong adherence after 24 hr. In addition, the biofilm formation ability of MRSA isolates differed among the various SCCmec types (Figure 1). The HAHO-MRSA isolate types II and III showed similar biofilm formation; however, the HAHO-MRSA type III isolates produced significantly higher biofilms than type IV and V

| D ISCUSS I ON
The higher incidence of infection in the hemodialysis population is because of impaired host immunity due to uremia-related neutrophil dysfunction and the hemodialysis procedure itself (inadequate water, unfamiliarity with dialyzer use, and repetitive break of skin integrity) (Jaber, 2005 Huang & Chen, 2011). In addition, livestock-associated MRSA (LA-MRSA) was limited mainly to SCCmec type IV and V, which originated from human-associated MSSA receiving the SCCmec element (Price et al., 2012). Nevertheless, in our study, HAHO-MRSA was discovered to carry type III as well as type IV elements, whereas HACO-MRSA isolates had both type IV and type V elements with type IV elements predominating. Our observation is in line with the finding of Wang et al. (2015), in which SCCmec type IV isolates accounted for HACO-and HAHO-MRSA infections. The increasing incidence of SCCmec IV HA-MRSA in Taiwan has been reported previously (Chen, Huang, Chiu, Su, & Lin, 2005;Huang et al., 2007).
This increase indicates that CA-MRSA is spreading into the hospital setting and may be replacing the conventional HA-MRSA strains, which have significant clinical and public health implications (Otter & French, 2011  the erm(A) gene was predominant in MRSA isolates and was correlated with SCCmec type III. In addition, they also found a relationship between erm(C) and SCCmec type IV. However, in our study, the erm(C) gene was found in SCCmec types II, IV and V indicating that this gene was not specifically related to any SCCmec type.
ST239-III-MRSA, which is a wellknown pathogen spread worldwide that has been associated with healthcare-associated infections, was the predominant multidrug-resistant clone found in our TCCs and surgical patients during the study period. Nevertheless, the dominant clone in our study, ST239-III-t037, which was reported by Chen, Liu, Jiang, Chen, and Wang (2010)

| Study limitations
The major limitation of this investigation is that it was a nonrandomized study with a small number of patients. The sample size of 47 patients was small, and multiple isolates from single patients were analyzed, which may cause bias and overexaggeration of the results.
Despite this limitation, we attempted to identify the relationship between the different types of vascular accesses and microbiology

(a) (b) (c)
profiles along with biofilm formation abilities. This is the first prospective study that focused on S. aureus VAI and compared it with other hospitalized S. aureus surgical infections. The microbiology profiles can provide useful information regarding the management of patients undergoing hemodialysis.

| CON CLUS IONS
In S. aureus dialysis vascular access infections, the isolates from TCCs had a higher percentage of methicillin resistance than the isolates from AVGs. MRSA in AVG-VAI was mostly SCCmec type IV. MRSA in TCC-VAI was similar to other hospitalized surgical infections and was predominantly SCCmec III, which had the strongest drug resistance and biofilm formation.

ACK N OWLED G EM ENT
We thank Wiley Editing Services for editing this manuscript. This study was supported by grants from the Chang Gung Memorial Hospital, Chiayi, Taiwan (grant numbers: CMRPG6E0423, CMRPG6H0121, CMRPG6H0291, CMRPG6H0292, and CMRPG6H0293).

CO N FLI C T O F I NTE R E S T S
The authors have no conflict of interest to declare.

AUTH O R S CO NTR I B UTI O N
Conceiving and designing the study: Chu C, Huang YK, and Lin CL; interpreting the data and writing manuscripts: Wong MY, Huang YK, Chu C, and Kao CC; obtaining funding: Huang YK, Tseng YH, and Tung CW. All authors read and approved the final manuscript.

E TH I C S S TATEM ENT
This study was approved by the Institutional Review Board (IRB) of the Chang Gung Memorial Hospital (IRB Nos: IRB101-41888 and IRB-8482B).

DATA ACCE SS I B I LIT Y
The authors declare that the experimental data published in this paper are made accessible upon request for interested readers.