Nosocomial outbreak of Aeromonas hydrophila surgical site infections after spinal surgery: Identification and control

Background: Aeromonas hydrophila surgical site infections (SSIs) were diagnosed in April 2017 in four patients who had received spinal surgery. We launched an outbreak investigation to identify the source, and accordingly, preventive and control measures were implemented. Methods: Environmental samples and samples from the healthcare providers were collected for microbiological analysis. The clonal relatedness of A. hydrophila strains was determined by pulsed-field gel electrophoresis (PFGE) and multilocus sequence typing (MLST). Whole genome sequencing (WGS) of one clinical A. hydrophila isolate (AE456) was performed using an Illumina NovaSeq PE150. Results: We identified eight case patients with SSIs due to A. hydrophila in orthopaedic ward 2 (three males; median age, 58 years). Strict infection control measures were adopted, particularly contact precautions and unit disinfection. We also identified A. hydrophila from water in a fish tank. PFGE and MLST revealed identical patterns and STs among the 10 clinical A. hydrophila strains (clone A, ST 517), which were different from those of the strains from the fish tank (clone B, ST518). WGS of isolate AE456 revealed the presence of cepS , cphA and bla OXA-12 genes encoding resistance to β-lactams. All patients recovered after antimicrobial therapy and/or surgical debridement. After removal of the fish tank, no new case occurred, and the outbreak was stopped. Conclusions: Aeromonas hydrophila is rare, but severe, pathogen in surgical infections and caused long hospital stay and physical suffering. Strict measures, including environmental disinfection and contact precautions, are needed to prevent infection outbreak after surgery.


Introduction
Aeromonas hydrophila, which is mesophilic, gram negative, rod shaped, and fermentative, is a common fresh or brackish water-borne pathogen that can cause a variety of human diseases, including gastroenteritis, hepatobiliary infection, skin and soft-tissue infections, septicemia, meningitis and endocarditis [1][2][3][4]. A. hydrophila can produce rapidly progressing wound infections in healthy individuals after trauma and burns and is recognized as an agent of skin and soft-tissue infections associated with water exposure [5][6][7]. Surgical site infection (SSI) is a significant complication and is associated with increased duration of hospitalization, high healthcare costs, and poor patient outcomes [8].
SSIs due to A. hydrophila have rarely been reported. Daniel Tena and his colleagues reviewed 24 cases of SSIs due to Aeromonas species and found that SSIs occurred rapidly after abdominal or pelvic surgeries (91.3%), possibly due to endogenous sources [9].
Exogenous sources such as exposure of wounds to contaminated water or leech therapy have also been reported for SSIs due to A. hydrophila [ 10]. A. hydrophila has often been reported to cause infection outbreaks in fish stocks, [11] but outbreaks in humans are relatively rare. An outbreak of wound infections with A. hydrophila occurred after a "mud football" competition [12]. (MRI) showed a large amount of effusion deep in the wound. Wound drainage and/or peripheral blood samples were collected and sent for microbiological analysis. The following day, A. hydrophila was recovered from the pus or drainage samples. One patient also tested positive for A. hydrophila in blood samples.
An outbreak of nosocomial infection due to A. hydrophila was suspected. All patients in the same ward were checked carefully for possible SSIs. To obtain baseline prevalence data, hospital microbiological laboratory databases were reviewed for A. hydrophila isolates identified from 2008 to 2017, which was the period for which complete records were available. Two similar cases were identified in June and July 2016 in orthopaedic ward 2.
Physicians reviewed the 6 cases for clinical manifestations and risk factors for the acquisition of A. hydrophila SSIs in detail. The phenotypes and drug resistance patterns of the A. hydrophila isolates were compared by microbiologists. After discussion and combination of the data, infection control measures were taken, and the first environmental investigation was initiated but failed to identify the A. hydrophila isolate.
Unexpectedly, in May 2017, a seventh patient developed a postoperative SSI of A. hydrophila in the same ward, and in July 2017, an eighth patient was identified as having a postoperative A. hydrophila SSI and bloodstream infection in orthopaedic ward 6. He was previously hospitalized in orthopaedic ward 2 for two days after operation. A case was defined as any patient with symptoms and signs consistent with SSIs and   drainage or blood culture positive for A. hydrophila since 2008. SSI was defined as   erythema, induration, pain, and septic drainage from the surgical site. Postoperative A. hydrophila infection was defined as being nosocomial if the infection was acquired within 30 days after a surgical procedure [13]. A fish tank in the nurses station caught our attention. The second culture included samples from the decorative stones, filter gauze, inner walls and water in the fish tank.

Identification of isolates and antimicrobial susceptibility testing
The organisms in all the patient samples (n = 10) and tank samples (n = 2) were identified as A. hydrophila/caviae using a biochemical phenotypic identification system (Vitek 2 Compact, bioMérieux, Marcy l'Etoile, France). The organisms were identified as A.
hydrophila using matrix-assisted laser desorption ionization mass spectrometry-time of flight (MALDI-TOF MS). DNA sequencing analysis was performed using three pairs of primers, including 16S rRNA, gyrB and rpoB, as described previously [14,15] to further confirm the identity of the isolates. Antibiotic susceptibility tests were conducted for 12 isolates using the Vitek 2 Compact system and the disk diffusion method. The Clinical and Laboratory Standards Institute (CLSI) M45-A3 criteria were used to define susceptibility and resistance to the antibiotics tested [16].

Pulsed-field gel electrophoresis (PFGE)
Pulsed-field gel electrophoresis (PFGE) was conducted according to the standardized PulseNet PFGE protocol for pathogenic gram-negative enteric bacteria [17] with a slight modification. Briefly, whole-cell genomic DNA from lysed cell cultures of each isolate embedded in 1% agarose plugs (Bio-Rad, Richmond, CA, USA) were digested with the restriction enzyme XbaI (New England Biolabs, Ipswich, MA) and separated by electrophoresis through 1% pulsed-field-certified agarose (Bio-Rad) using a CHEF-Mapper instrument. Electrophoretic switch times of 4-40 s were used with a 6 V/cm current and a switch angle of 120° at a constant temperature of 14 °C. PFGE patterns were interpreted using the criteria proposed by Tenover et al [18]. An isolate is designated genetically indistinguishable if its restriction pattern has the same numbers of bands and the corresponding bands are the same apparent size; An isolate is considered unrelated to the outbreak strain if there are seven or more band differences between the outbreak pattern and that of the test isolate.

Detection of virulence genes and multilocus sequence typing (MLST)
Polymerase chain reactions (PCRs) using previously described primers and conditions were conducted to detect the virulence genes encoding heat-stable enterotoxin (ast), haemolysin (ahh1), hemolysin (asa1), cytotoxic enterotoxin (act), enolase (eno) and components of the type III secretion system (ascV and aexT) [19]. Six housekeeping genes (gyrB, groL, gltA, metG, ppsA and recA) were chosen for the MLST analysis according to the method previously reported by Martino. Amplified PCR products corresponding to the expected sizes were sequenced as previously described [19]. Each unique allelic profile, as defined by the allele numbers of the 6 loci, was assigned a sequence type (ST) number.

Whole genome sequencing (WGS) and genome assembly
Resistance to aztreonam and ceftazidime of one A. hydrophila isolate collected from patient 4 developed after one week of empirical therapy with ceftriaxone. WGS of the latter resistant A. hydrophila isolate 5 (AE456) was performed using an Illumina NovaSeq PE150. Illumina PCR adapter reads and low quality reads from the paired-end were filtered by the step of quality control using Readfq (vision 10). All good quality paired reads were assembled using the SOAP denovo (http://soap.genomics.org.cn/soapdenovo.html) and ABySS into a number of scaffolds. We used the ARDB (Antibiotic Resistance Genes Database) to perform the antibiotic resistant genes analyses.

Epidemiological findings
Among the 34 patients with positive A. hydrophila cultures isolated in the microbiological laboratory since 2008, eight were current outbreak cases with postoperative SSIs, five had skin and soft-tissue infections, two had diarrhea, nine had a hepatobiliary infection, five had a urinary infection, and five were colonized. Eight case patients were all hospitalized in orthopaedic ward 2.
The A. hydrophila infection appeared to be intermittent, as shown in Fig. 1. From June 2016 to July 2017, i.e., from the first case to the last case recorded eight patients tested positive for A. hydrophila in fluids from the surgical sites or blood cultures, and the clinical characteristics of these patients are shown in Table 1. Three patients were male (37.5%), and five were female; the median age of the patients was 60 years (range, 42 to 68 years). The case patients had been admitted for various underlying medical conditions, including thoracic spinal stenosis (n = 2), scoliosis (n = 1), lumbar spinal stenosis (n = 4), and a giant cell tumor of the thoracic vertebra (n = 1). The types of surgery included laminectomy or tumor excision, decompression, bone graft fusion and internal fixation.
Drainage tubes were placed in the surgical site after the operations, and intravenous cefuroxime was used to prevent infection. SSIs occurred 3-6 days after operation, 0.5-13 hours after removal of drainage tubes. The clinical manifestations included fever (n = 6), purulent drainage fluid (n = 3), pain in the wounds (n = 2), and hypotension and oliguria (n = 1). Six patients underwent wound debridement and irrigation. Two or three drainage tubes were placed in the wounds for continuous irrigation-suction with saline and hydrogen peroxide. All the patients received imipenem or ertapenem after obtaining the antimicrobial susceptibility test results of A. hydrophila. All patients recovered and were subsequently discharged from the hospital after 11-59 days (median, 27 days). isolates from the fish tank were most closely related to A. hydrophila (> 99% homology for three fragments, namely, the 16S rRNA, gyrB and rpoB gene sequences).
Ten strains of A. hydrophila collected from the patients were resistant to cefuroxime, cefoxitin, ceftriaxone, and trimethoprim-sulfamethoxazole and susceptible to cefepime, imipenem, meropenem, and amikacin, as shown in Table 2. Two out of the 10 strains were resistant to ceftazidime, ciprofloxacin and levofloxacin. The antibiotic susceptibility patterns of the 2 isolates collected from the fish tank were similar to those from patients, except the susceptibility to cefoxitin, cefuroxime and ceftriaxone. It is very difficult to determine the source and transmission route of an outbreak, even if the isolates were identified from patients or the environment [24,25]. There were 6 nursing workers who were the only people who had direct contact with the patients and the fish tank. These workers participated in the postoperative care of the patients and usually helped empty the drainage bags. These nursing workers also cleaned the fish tank and changed water. A large plastic box was used to hold the aseptic wound dressing packages in the ward. To throw the dead fishes away and change water, the nursing workers sometimes removed the dressing change packages and poured the water and dead fishes into the box temporarily. The packages were wrapped in cloth with a bending plate, gauze, tweezer and scissor. On the second day after the operation, the doctors disinfected the patients' wounds and replaced the dressing gauze with new gauze from the packages. Therefore, if the box had not been dried and sterilized after holding water and dead fishes, the dressing packages toward the bottom of the box would absorb the residual water and be contaminated.
Alternatively, direct transmission may also be caused by nursing workers lacking hand hygiene; when these workers poured the drainage fluid every day, the bacteria, via retrograde motion, may have entered the wound through the drainage tube. The reason for this speculation is that the skin of the patients' surgical incisions healed well, and a large amount of fluid was found under the muscular layer or subcutaneous tissue layer during debridement.
However, there remain many questions regarding the transmission hypothesis. First, a large number of Aeromonas species, including A. hydrophila, were isolated from the fish tank, but only A. hydrophila ST517 could be identified in patients. A plausible reason for this finding may be the different antimicrobial susceptibility patterns between the strains.
The clinical A. hydrophila strains were resistant to cefuroxime used for infection prophylaxis after surgery, but Aeromonas species collected from the fish tank were sensitive to cefuroxime and may be uncultured strains. Second, an exact match of PFGE patterns and STs between A. hydrophila strains from the fish tank and A. hydrophila strains isolated from infected patients could not be established. This finding may be due to the wide diversity and large number of strains detected in the fish tank. It is difficult to identify pathogenic bacteria by culture and based on colony morphology.
The pathogenicity of Aeromonas species appears to be due to the ability of the bacteria to produce several virulence factors that are highly heterogeneously present among clinical isolates [26]. In this study, all the A. hydrophila isolates from patients harbored multiple virulence genes. A. hydrophila isolates from the fish tank carried less virulence genes tested than the clinical isolates. Antimicrobial resistance was more commonly observed among clinical isolates than among environmental isolates, as previously described [27].
Saavedra et al has demonstrated that extensive use of antimicrobial drugs for prophylaxis and treatment unquestionably plays a role in the increased number of resistant A.
hydrophila strains [28]. Resistance to cefotaxime can develop during therapy [29,30]. There are some limitations associated with this study. First, we could not determine the specific link between these outbreak cases and the fish tank. As a possible source of the outbreak, the fish tank remained a concern because a large amount of A. hydrophila was recovered from the tank. Another limitation of the investigation was that we did not use a selective culture medium with cefuroxime or ceftriaxone to identify the pathogen. In Declarations contributors in writing the manuscript. All authors read and approved the final manuscript.

Funding
This work was financially supported by Peking University Third Hospital. The funder had no role in study design, data collection, data analysis, data interpretation, and writing of the manuscript.

Availability of data and materials
The datasets are available from the corresponding author under reasonable request.

Ethics approval and consent to participate
Permission to use the information in the medical records of the patients and the A.
hydrophila isolates for research purposes was given by the Ethics Committee of Peking University Third Hospital. Based on the approval of Ethics Committee, all informations were collected only by patient's code and their identity was not disclosed. Patient's information was in private.

Consent for publication
Not applicable.