Evaluation of a Stenotrophomonas maltophilia bacteremia cluster in hematopoietic stem cell transplantation recipients using whole genome sequencing

Background Stenotrophomonas maltophilia ubiquitously occurs in the hospital environment. This opportunistic pathogen can cause severe infections in immunocompromised hosts such as hematopoietic stem cell transplantation (HSCT) recipients. Between February and July 2016, a cluster of four patients on the HSCT unit suffered from S. maltophilia bloodstream infections (BSI). Methods For epidemiological investigation we retrospectively identified the colonization status of patients admitted to the ward during this time period and performed environmental monitoring of shower heads, shower outlets, washbasins and toilets in patient rooms. We tested antibiotic susceptibility of detected S. maltophilia isolates. Environmental and blood culture samples were subjected to whole genome sequence (WGS)-based typing. Results Of four patients with S. maltophlilia BSI, three were found to be colonized previously. In addition, retrospective investigations revealed two patients being colonized in anal swab samples but not infected. Environmental monitoring revealed one shower outlet contaminated with S. maltophilia. Antibiotic susceptibility testing of seven S. maltophlia strains resulted in two trimethoprim/sulfamethoxazole resistant and five susceptible isolates, however, not excluding an outbreak scenario. WGS-based typing did not result in any close genotypic relationship among the patients’ isolates. In contrast, one environmental isolate from a shower outlet was closely related to a single patient’s isolate. Conclusion WGS-based typing successfully refuted an outbreak of S. maltophilia on a HSCT ward but uncoverd that sanitary installations can be an actual source of S. maltophilia transmissions.


Background
Stenotrophomonas maltophilia, an intrinsically multidrug resistant gram negative pathogen, is widely distributed in aqueous habitats such as, sink drains, endoscopes and hemodialysis water within clinical settings [1,2]. Usually, this pathogen is not highly virulent in immunocompetent individuals but can cause severe infections including bacteremia, peritonitis and meningitis in immunocompromised hosts resulting in complications, e.g. septic shock, respiratory failure, tissue necrosis and septic thrombophlebitis [3]. Recipients of hematopoietic stem cell transplantations (HSCT) are especially at risk and suffer from pulmonary hemorrhage resulting in higher mortality rates compared to non-HSCT patients [4].
In recent years, an increasing incidence of S. maltophilia is reported on oncologic wards [5]. Moreover, several outbreaks and pseudo-outbreaks caused by this pathogen have been reported [6,7]. In addition to epidemiological investigations, different typing methods, e.g. pulse field gel electrophoresis (PFGE) and multi locus sequence typing (MLST) are used to identify the genetical relationship among different S. maltophilia isolates [8][9][10][11]. These techniques are useful in excluding outbreak scenarios, if MLST sequence types (ST) or PFGE patterns of isolated pathogens differ. In case of identical STs or highly similar PFGE patterns both methods reach their discriminatory limits. In these cases an outbreak is assumed, although more precise methods uncover and accidental cluster. Additionally, inter-laboratory comparability is limited using PFGE. Therefore, whole genome sequence-based typing (WGS) approaches are increasingly seen as gold standard method for highly discriminatory typing [12] and the superiority of WGS in comparison to other typing methods was already demonstrated for several bacterial pathogens [13][14][15][16][17]. Moreover, we could recently demonstrate the high reproducibility of WGS, which is another prerequisite for the applicability in clinical routine [18].
Here we evaluated a cluster of S. maltophilia bacteremia in recipients of allogeneic stem cell grafts using WGSbased typing.

Cluster detection, epidemiological investigations and infection control measures
The 1500-bed University Hospital Münster includes two HSCT-units each comprising 10 patient rooms, which are all HEPA-filtered and equipped with separated bathrooms for each patient (Fig. 1). In total, 193 patients were admitted to the HSCT-units during 2016. Patients usually receive an allogenic HSCT.
S. maltophilia bloodstream infections (BSI) were defined as one or more blood cultures positive for S. maltophilia obtained from patients with clinical signs of infection (fever >38°C, chills, hypotension) according to the critieria given by the European Centre for Disease Prevention and Control [19]. Four S. maltophilia BSI were detected in patients (P1, P3, P4, P6) in the HSCT ward between February and July 2016. As this number exceeds the average baseline of 0.5 S. maltophilia BSI per year detected during 2011-2015, an epidemiological investigation was initiated. Routine environmental screening of aqueous habitats (shower heads, shower outlets, washbasins and toilets) within patients' rooms in the HSCT unit, normally focusing on multidrug resistant

Whole genome sequencing (WGS) based typing
To determine the clonal relationship of S. maltophilia strains isolated from blood cultures, the isolates were subjected to WGS using the Illumina MiSeq platform (Illumina Inc., San Diego, USA) as described previously [13]. Retrospectively identified colonization isolates were not subjected to WGS as they were no longer available in contrast to BSI isolates, that are stored for a longer time period. Using SeqSphere + software version 2.0 beta (Ridom GmbH, Münster, Germany), all coding regions were extracted and compared in a gene-by-gene approach (core genome multilocus sequence typing, cgMLST) using S. maltophilia K279a strain (GenBank accession number AM743169.1) as a reference sequence. Instead of a published cgMLST scheme, which is not yet available, this ad hoc scheme was used to differentiate the cluster. SeqSphere + software was used to display the clonal relationship in a minimumspanning tree (MST). For backwards compatibility with classical molecular typing, i. e. MLST, the MLST sequence types (ST) were extracted from the WGS data in silico.

Epidemiological investigation and susceptibility testing
Chronological order and spacial distribution of isolated S. maltophilia is displayed in Fig. 1. Isolates were obtained from blood culture samples of patients on both wards. Two of the four isolates (isolated from P4 and P6) were detected within short intervals in two patients both admitted to room 1 on ward 2. Of the four patients suffering from S. maltophilia bacteremia, three (P1, P4 and P6) were previously tested positive for S. maltophilia in stool samples or anal swabs. Additionally, two colonized patients (anal swabs; P2, P5) admitted to ward 2 were uncovered by the retrospective analysis (Table 1). One environmental isolate (E1) could be detected in a shower outlet of a patient room that was a patient with S. maltophilia infection. Susceptibility testing revealed all patient isolates to be susceptible to TMP-SMX, except isolate P3. The environmental isolate was resistant to TMP-SMX (Table 1).

Clinical characteristics and outcome of patients with S. maltophilia bacteremia
All patients with S. maltophilia BSI suffered from acute myeloid leukemia as the underlying disease. P2 and P5, who were only colonized by S. maltophilia, suffered from chronic lymphatic leukemia and myeloproliferative neoplasm. Except P1, all patients had already received allogenic SCT at the time of S. maltophilia detection. P1 and P6 developed an acute respiratory distress syndrome. P1, P3, and P6 died due to S. maltophilia bacteremia.

WGS-based typing
We analyzed four bloodstream isolates and the environmental isolate by WGS. Comparison via cgMLST, based on 1876 genes present in all isolates, revealed no genetical relationship among isolates originating from patients (Fig. 2). In contrast, S. maltophilia strains isolated from P3 and E1 showed a similar genotype with only four alleles difference, suggesting a relationship between these two isolates. P3 and E1 harbored the MLST ST 94. All other isolates harbored new MLST STs.

Discussion
S. maltophilia has emerged as an important nosocomial pathogen associated with increased mortality rates in patients suffering from hematological malignancies and receiving hematopoietic stem cell transplantations [20,21]. In this study, WGS was used to evaluate a cluster of four S. maltophilia BSI in patients partially suffering from pulmonary dysfunction associated with the infection due to this pathogen [4]. Susceptibility testing, normally used as a first indication of nosocomial transmission of bacterial strains, could not exclude a spread of one single clone. WGS and subsequent cgMLST analysis, methods that have shown to provide detailed information in evaluating the epidemiology in nosocomial clusters of other pathogens [22], excluded a genetical relation and therefore an outbreak scenario. A number of different sources are possibly responsible for the increasing number of S. maltophilia BSI. Due to its multidrug resistant nature, selection of this pathogen can easily occur in patients after receiving broad-spectrum antibiotics [23]. Moreover, S. maltophilia infections occur frequently in combination with the distinct immunocompromised status of HSCT recipients [24]. On the other hand, several environmental sites such as sink drains could be detected as sources of S. maltophilia within hospital settings [25].
To what extent these habitats are origins for hospitalacquired colonizations or infections of S. maltophilia could not yet be shown in detail. In this study we documented a genetical relation between an environmental and a patient blood stream infection isolate by analyzing WGS data. Although there was no local proximity, S. maltophilia could be transmitted from the environment to a patient or vice versa, at least giving the possibility of further spread via aqueous habitats in the HSCT unit.

Conclusion
WGS-analysis can be used to precisely refute an outbreak of S. maltophilia. Additionally, WGS-based typing documented sanitary installations in HSCT units to be an actual source or result of transmission between environmental and human habitats. Hence, in addition to classical infection control strategies, monitoring of aqueous habitats has to be established within the rooms of HSCT recipients in order to prevent transmission of these multidrug resistant organisms, as recently also shown for P. aeruginosa [26].