An outbreak of Ralstonia pickettii bloodstream infection and clinical outcomes

Introduction: Ralstonia pickettii infections are rare and may be mistaken for other bacteria. This study aims to report a hospital outbreak of R. pickettii at a tertiary hospital, which was initially misidentified as Ralstonia insidiosa, along with its clinical consequences. Methodology: A bacteraemia outbreak occurred between August 14 and October 4, 2019, infecting 22 patients admitted to diverse intensive care units. All isolates were identified with the use of the automated VITEK 2 Compact system and were then subjected to a microbial identification system, matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS). Bacterial identification and genomic DNA typing was made using pulsed-field gel electrophoresis. Investigation covered all potential sources of the outbreak. Results: An index patient and five additional patients developed fever while receiving care. Blood cultures of these patients yielded R. insidiosa by the VITEK 2 Compact system. Culture isolates were then submitted to a reference centre for confirmation by the MALDI-TOF MS system, where the bacterium turned out to be R. pickettii . No pathogen was isolated in the commercial products except for three samples of unopened sterile distilled water. Despite its discontinuation, 16 new cases were identified, in which blood cultures grew R. pickettii by the MALDI-TOF MS system. Attempts to uncover the source of the outbreak failed. Clinical manifestation was confined to fever in all the patients. Conclusions : During this outbreak, R. pickettii infections ran a relatively mild course without clinical deterioration or mortality, possibly due to low virulence.


Introduction
Ralstonia pickettii is an aerobic, Gram-negative, non-fermentative bacillus of the Ralstonia genus having 16 species and three subspecies [1]. Previously named as Pseudomonas pickettii and Burkholderia pickettii, R. pickettii is the first identified and most common species of the genus. It is a water-borne bacterium that can grow at very low concentrations of nutrients [2]. In addition, it can survive at a wide range of temperatures (15-42 °C) and even pass through 0.2 μm filters, leading to outbreaks of infections via therapeutic fluids used in hospital settings [3]. Reportedly, outbreaks have often been associated with contamination of various therapeutic fluids and materials, such as sterile water, intravenous ranitidine, saline solution, narcotics, and magnesium vials by R. pickettii during the manufacturing of these products [2]. This bacterium has also been reported to cause pseudo-outbreaks by contamination of skin disinfectants and blood-culture bottles [1,2]. It has low virulence, but may result in severe infections such as meningitis, endocarditis, osteomyelitis in immunocompromised patients [1]. It may develop resistance to ciprofloxacin, trimethoprimsulphamethoxazole, piperacillin-tazobactam, imipenem/cilastatin, and ceftazidime and has the ability to form biofilms, posing challenges to treatment [1]. We aimed to report a hospital outbreak of R. pickettii with its clinical consequences at a tertiary hospital.

Hospital settings
This prospective study was performed at Kartal Koşuyolu Training and Research Hospital, a 465-bed tertiary referral center in İstanbul, Turkey, with three intensive care units (ICU) (adult cardiovascular surgical ICU, pediatric cardiovascular surgical ICU, and coronary ICU) as well as cardiovascular surgery clinics (one for pediatric cardiovascular surgery), gastroenterology surgery clinic, transplantation clinic, and cardiology clinics. The study was approved by the institutional review board and all study protocol conformed to the Declaration of Helsinki.

Patients
We identified a bacteraemia outbreak due to R. pickettii at Kartal Koşuyolu Training and Research Hospital between 14th August and 4th October 2019. The study included 22 patients whose blood cultures were positive for R. pickettii (Table 1).

Data collection
Data included patients' demographic features, symptoms, physical examination findings, laboratory results, comorbidities, treatments, daily visit records, and outcomes.

Microbiological characterization and antimicrobial susceptibility
All isolates were identified with the use of the automated VITEK 2 Compact system that utilize GNI cards (bioMérieux Vitek Hazelwood, Missouri, USA). Culture isolates were then submitted to Kartal Dr. Lütfi Kırdar Training and Research Hospital for confirmation by a microbial identification system, matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) (VITEK MS; bioMérieux, France). Susceptibility testing was conducted by the Etest (bioMe´rieux, France). As the breakpoints for the Ralstonia spp. have not been specified by the European Committee on Antimicrobial Susceptibility Testing (EUCAST), antimicrobial susceptibility testing was performed using breakpoints for similar species, including Pseudomonas spp., Burkholderia cepacia and Acinetobacter spp.(EUCAST, Clinical breakpoints -breakpoints and guidance, 25 th January 2020). The MICs were read after incubation overnight at 35 °C. The MICs for the bactericidal drugs (ertapenem, meropenem, ceftazidime, ciprofloxacin, levofloxacin, ceftriaxone, amikacin, and piperacillin-tazobactam) were defined as the point of intersection between the ellipse edge and the Etest strip, where there was complete inhibition of growth. The MIC for the bacteriostatic drug (trimethoprim-sulfametoxazole and tigecycline) was read at 80% inhibition.

Bacterial identification and genomic DNA typing by pulsed-field gel electrophoresis
Pulsed-field gel electrophoresis (PFGE) was performed as previously described by Durmaz et al. [4]. Briefly, bacterial cells embedded in 1 % low-meltingpoint agarose (Bio-Rad Lab, Hercules, CA, USA) plugs were lysed with lysozyme and proteinase K, and then chromosomal DNA was digested with 40 U SmaI (Thermo Scientific-Fermantas Corporation, Vilnius, Lithuania). Fragmented DNA samples were The gel was stained with ethidium bromide (5 μg/mL), visualized under UV light, and photographed using ChemiDoc MP Imaging System (BIO RAD Company, United Kingdom). Analysis of PFGE patterns was made using BioNumerics software version 7.5 (Applied Maths, Saint-Matins-Latem, Belgium) and compared using a Dice coefficient with a tolerance of 1.5% and an optimization of 1%. Isolates with identical patterns were considered to be genotypically 'indistinguishable', while those that differed by 1-3 bands were defined as 'closely related' and 4-6 bands as 'possibly related'. 'Unrelated' or 'different' strains indicated those that differed by ≥ 7 bands [4].
Both antimicrobial susceptibility testing and PFGE identification were performed in 13 isolates recovered from the patients and in three isolates recovered from unused sterile distilled water.
Investigation covered all potential sources of the outbreak. The first six cases of R. picketti were detected in the mid-August, when the weather was quite hot and dry. In accordance with institutional infection control measures, there was no evidence for the infection source, such as plants, organic matter, leakages of any kind, moisture, ventilation/heating problems. Eighty commercial samples from different wards and from commercial products commonly used in infected patients were taken for culture and tested for R. pickettii, including distilled water (10 samples), chlorhexidine and alcohol-solutions (10 samples), povidone-iodine (10 samples), magnesium (10 samples), heparin (10 samples), and saline solution (30 samples, of which, after the outbreak 10 isolates were documented). Although the study patients had fever, 10 unused blood-culture bottles were examined to exclude a pseudo-outbreak.

Investigation into the source of the outbreak
A total of 22 patients were included in this study, during which time 9,419 patients were admitted to our hospital. On August 14, a female (index) patient who had been admitted 64 days earlier to the cardiovascular surgery ICU for repair of the mitral and tricuspid valves developed an increased body temperature of 38.3 °C (Table 1). Blood samples were obtained from a central venous catheter and a peripheral vein and were collected in four bottles (2 aerobic and 2 anaerobic), two of which yielded R. insidiosa by the VITEK 2 Compact system at our centre.
Initially, it was thought to be a contamination related to the laboratory procedures, which led to a comprehensive review thereof. However, during the same week, blood cultures of five additional patients grew R. insidiosa. Of these, three patients had been in the same ICU as the index patient, while the other two had been receiving care at different wards. An investigation into the outbreak was initiated.
No pathogen was isolated from the commercial products except for three samples of unopened sterile distilled water on September 4, upon which the use of sterile distilled water was terminated. Routine weekly surveillance samples of urine and respiratory secretions were collected from all patients admitted to the ICU. None of the patients with bloodstream infections (BSIs) had pulmonary involvement, nor any growth in respiratory secretions or urine.
Culture isolates were then submitted to Kartal Dr. Lütfi Kırdar Training and Research Hospital for a more specific identification and confirmation by the MALDI-TOF MS system (VITEK MS; bioMérieux, France). In the following days, despite the discontinuation of the sterile distilled water, 16 new cases were identified. Investigation was extended to different sources, but the source of the outbreak could not be identified.
We then reviewed the procedures that the affected patients had received, which pointed to the common use of 500 mL saline solutions as part of an intravenous procedure (Table 2). Subsequently, 10 samples of saline solutions of diverse lot numbers were found to be free of contamination.

Isolation of R. pickettii
The isolates of the index patient and the following five patients, diagnosed as R. insidiosa by the VITEK 2 Compact system at our centre, turned out to be R. pickettii as the culprit bacterium. Blood samples of the subsequent 16 patients also grew R. pickettii by the MALDI-TOF MS system at the reference centre. Antimicrobial susceptibility testing was performed in 13 isolates recovered from the patients and in three isolates recovered from unused sterile distilled water. As the breakpoints have not been specified for Ralstonia spp., the breakpoints given for similar species were used to interpret pathogen susceptibility. Thus, one strain isolated from the distilled water No. 2 was interpreted to be resistant to ertapenem, and three others isolated from three patients (Patients no. 3, 4, 7 - Table  2) to amikacin.
All tested isolates were found to have higher MIC values than those for ciprofloxacin and levofloxacin. They were also found susceptible to the other antimicrobials that were tested (meropenem, trimethoprim-sulfametoxazole, tigecycline, ceftazidime, ceftriaxone and piperacillin-tazobactam, Table 3).
In genotypic analysis, all the 13 isolates were interpreted as "indistinguishably" similar, and three isolates from distilled water were interpreted as "different" (Figure 1).

The characteristics of the patients
The BSI outbreak affected 22 patients (11 males, 11 females), with 37 R. pickettii isolates recovered from 36 aerobe and one anaerobe blood cultures. Twenty-five blood samples were obtained from the peripheral veins, and 12 blood samples were obtained from the central catheters. The isolates of the index patient and the following five patients were initially misidentified as R. insidiosa by the VITEK 2 Compact system, but later were identified as R. pickettii by the MALDI-TOF MS system. Eleven patients were admitted to the cardiovascular intensive care unit, 10 of whom underwent cardiac surgery. The remaining 11 patients were admitted to one of the four different departments ( Table 1).
The mean age of patients was 47.8 years (range 3-87 years). The mean hospital stay was 40 days (range 4-177 days). All patients exhibited fever as the main indicator of the outbreak, with no signs of sepsis.
Fifteen patients had central venous catheters, which were kept in place during the BSI outbreak due to the absence of clinical deterioration. Five patients died from causes other than infection (Table 1).
A total of 15 patients received intravenous empirical antibiotic therapy duo to the suspicion of BSI until culture results ( Table 1). One of the seven patients (Patient no. 22- Table 1) who did not receive intravenous antibiotic therapy was diagnosed with brain-death, while the other patients survived and were discharged in good condition before obtaining the results of blood cultures. Oral amoxicillin-clavulanic acid was initiated in one of the six patients at discharge. The mean hospital stay of these six patients was 6.2 days.
During hospitalization, repeat blood cultures were obtained from 14 patients and all were tested negative. Of these, two patients with histories of both caesarean section and pulmonary hypertension had recurrent fever on days 11 and 12 of hospitalization, respectively, and had positive cultures. One of these patients underwent  lung transplantation 63 days after the initial positive culture and was well after transplantation (Patient no. 11- Table 1). All patients were followed up until January 2021 without any signs of recurrent infections.

Discussion
R. pickettii infections are rare and are often mistaken for other bacteria, including Ralstonia species, Burkholderia cepacia complex, and Pseudomonas fluorescens. MALDI-TOF mass spectrometry provides true identification of Ralstonia spp. at the genus level [1,5]. Our isolates were initially misidentified as R. insidiosa by the VITEK 2 Compact system, but later MALDI-TOF MS enabled true identification of R. pickettii.
To date, there have been reports that R. pickettii infections are associated with contamination of fluids during the production process. Therefore, once an outbreak of R. pickettii is identified, contamination of solutions used for therapeutic purposes should be suspected and investigated [2,6]. Our investigation into the therapeutic solutions used at our hospital pointed to the sterile distilled water as the possible culprit. Although weekly surveillance cultures were collected from the patients in the ICU, respiratory secretions of the patients were free from R. pickettii. On PFGE testing, 13 outbreak-related strains were shown to be indistinguishable, while the three strains isolated from the sterile distilled water were of different genotypes. We were not able to identify the real source of the outbreak. Given that the saline solution was the sole common agent used in all patients, we considered saline solutions to be the probable source of the outbreak. Also, since the infected patients were admitted in different departments, the transmission by either health care workers or environmental contamination was less likely. So far, patient-to-patient transmission has not been reported [1].
Infusion-related BSIs are rarely encountered and their incidence is unknown, because routine bacteriological sampling is not performed [7]. Our major shortcoming was that we did not perform bacteriological sampling from infusates at the time the patients developed fever.
Six patients who did not receive antibiotic treatment during hospitalization were discharged. At their followup visits, they had no complaints of fever, nor any signs of infection. One patient later underwent coronary artery bypass grafting and had an uneventful recovery.
A similar and relatively favourable outcome of an outbreak with R. pickettii was also reported in four patients at a haemodialysis centre [11]. Of the four patients, two had catheters and both underwent catheter removal and received intravenous antibiotic therapy. The remaining two patients received oral antibiotic therapy alone.
We had two female patients who had been admitted with pulmonary hypertension in the postpartum period (Table 1). Their fever initially responded to antibiotic treatment, but later the fever recurred. Both patients had positive blood cultures for R. pickettii at the time of the initial and recurrent fevers.
Ralstonia spp. have been reported to cause serious respiratory tract infections such as cavitary pneumonia and lung abscesses, especially in patients with cystic fibrosis and chronic obstructive pulmonary disease [16,17,23,24]. At our centre, routine weekly surveillance samples of respiratory secretions are collected from all ICU patients. None of our patients with BSIs had pulmonary involvement, nor any growth in respiratory secretions.

Was the condition a real outbreak or transient bacteraemia?
During this outbreak, despite the presence of positive blood cultures in all affected patients, the infection did not result in clinical deterioration or mortality. In particular, irrespective of the presence or removal of catheters or the presence or absence of antimicrobial treatment, the outcome did not differ with regard to mortality among affected patients. This may raise the question of whether the condition was a real outbreak or transient bacteraemia. In addition, it is likely that, had it really been an outbreak, it might have ended up with transient bacteraemia due to low virulence or relatively competent immune system of the patients.

Limitations
The main limitation to the present study is that infusion samples were not sent for microbiologic examination, which would have increased the chance to find out the source of the outbreak.