Molecular and epidemiological analysis of a Burkholderia cepacia sepsis outbreak from a tertiary care hospital in Bangladesh

Background Burkholderia cepacia complex (Bcc) is a group of serious pathogens in cystic fibrosis patients and causes life threatening infections in immunocompromised patients. Species within the Bcc are widely distributed within the environment, can survive in the presence of disinfectants and antiseptics, and are inherently multidrug resistant (MDR). Methods Dhaka Medical College Hospital (DMCH) patients with a B. cepacia positive blood culture between 20 October 2016 to 23rd September 2017 were considered as outbreak cases. Blood stream infections (BSIs) were detected using BacT/ALERT 3D at DMCH. B. cepacia was isolated on chromogenic UTI media followed by MALDI-TOF. Minimum inhibitory concentration (MIC) of clinically relevant antibiotics was determined by agar dilution. Whole genome sequencing was performed on an Illumina MiSeq platform. Patients’ demographic and clinical data were collected. Patients’ clinical history and genomic data of the outbreak strains were merged to investigate possible outbreaks. Ninety-one B. cepacia genomes were downloaded from ‘Burkholderia Genome Database’ and the genomic background of the global strains were compared with our outbreak strains. Results Among 236 BSIs, 6.35% (15/236) were B. cepacia. Outbreak cases were confined to the burn critical care unit and, to a lesser extent, the paediatrics department. There was a continuum of overlapping cases at DMCH between 23 October 2016 to 30 August 2017. Core genome SNPs showed that the outbreak strains were confined to a single clade, corresponded to a common clone (ST1578). The strains were shown to be MDR and associated with a mortality of 31% excluding discharge against medical advice. MIC profiles of the strains suggested that antibiotics deployed as empirical therapy were invariably inappropriate. The genetic background of the outbreak strains was very similar; however, a few variations were found regarding the presence of virulence genes. Compared to global strains from the Burkholderia Genome Database, the Bangladeshi strains were genetically distinct. Conclusions Environmental surveillance is required to investigate the aetiology and mode of transmission of the B. cepacia outbreak. Systematic management of nosocomial outbreaks, particularly in resource limited regions, will mitigate transmission and will improve patients’ outcomes.

strains suggested that antibiotics deployed as empirical therapy were invariably inappropriate. The genetic background of the outbreak strains was very similar; however, a few variations were found regarding the presence of virulence genes. Compared to global strains from the Burkholderia Genome Database, the Bangladeshi strains were genetically distinct.

Introduction
The genus Burkholderia incorporates Gram negative, catalase-producing, lactose-nonfermenting), obligately aerobic bacilli and encompasses the species of the Burkholderia cepacia complex (Bcc), Burkholderia mallei, B. pseudomallei. and B. gladioli [1]. Bcc is composed of at least 20 different species, including B. cepacia, B. multivorans and B. cenocepacia [1,2]. Bcc was first described in mid-1980s, in cystic fibrosis (CF) patients as a cause of 'Cepacia Syndrome', characterized by lung function deterioration, bacteremia and death [3]. The members of Bcc are widely distributed in the environment, including water, soil, fruits and vegetables and can survive for prolong periods of time in moist environments, even in the presence of disinfectants and antiseptics. They are capable of colonizing fluids in the hospital such as irrigation solutions or intravenous fluids and, serve as potential source of nosocomial infections [4,5,6]. Bcc rarely cause diseases in healthy individuals and is mostly regarded as a serious pathogen in cystic fibrosis patients [7]. The pathogens are inherently multidrug resistant (MDR) and highly transmissible by direct contact [7][8][9].
Regardless of geographical location nosocomial outbreaks attributed by Bcc are frequent. Recent outbreaks of Bcc bacteraemia from India, Germany and USA, have been reported particularly in the immunocompromised patients [6,[10][11][12]. Bcc outbreaks are always linked to hospital environment and/or contaminated medical devices, intravenous fluids or antiseptic solutions [4][5][6][10][11][12][13]. We describe an outbreak of B. cepacia bacteraemia in burn's critical care units of Dhaka Medical College Hospital (DMCH) analysed by whole genome sequencing. The population structure and genetic relationship of B. cepacia isolated in this study were compared with 91 global B. cepacia deposited previously in the 'Burkholderia Genome Database' [14].

Ethical statement
Ethical consent was given by the Ethical Review Committee (ERC) of Dhaka Medical College Hospital (DMCH) prior to the start of the study and in accordance with the Helsinki Declaration [Memo no: MEU-DMC/ECC/2017/122] [15]. Required written consent was taken from all the participants for this study.

Hospital setting and outbreak case definition
We carried out this study between 20 October 2016 and 23 rd September 2017 in DMCH which is the largest public hospital in Bangladesh. A total of 527 blood culture samples received by the clinical microbiology laboratory were included in this study. Detailed microbiological analysis of isolates, including Whole Genome Sequencing (WGS), performed at Cardiff University, were undertaken retrospectively. Patients with a B. cepacia positive blood culture were included as outbreak cases. Patients' demographic (name, age, sex, locality and socio-economic condition) and clinical data (clinical symptoms or reason for hospitalization, ward, type of infection, admission date, sample collection date, outcome and antimicrobial therapy) were collected. Clinical data was collected following positive culture results from the local laboratory. Additional clinical follow up information was not available in the current study, except for outcome (discharge, death, discharge against medical advice {DAMA}). The study outline is described in

Multi locus sequence typing (MLST)
B. cepacia (n = 15) isolated in this study was submitted to pubMLST to assign as sequence types (STs) based on 7 loci [20]. Global B. cepacia (n = 91) genomes were downloaded from 'Burkholderia Genome Database' of which 82 strains could not be matched with previously described STs [14]. We also submitted the 82 genomes from database to pubMLST [20]. A minimum spanning phylogenetic tree was constructed with 106 B. cepacia genome using Ridom SeqSphere+ (Ridom GmbH, Münster, Germany) based on 7 B. cepacia MLST genes.

Core genome phylogenetic analysis
We mapped single nucleotide polymorphisms (SNPs) of global B. cepacia from database (n = 91) and our outbreak strains (n = 15) with the first isolated strain in our study, dm93 (used as reference genome) using Snippy (3.2) [21]. SNPs alignemnt was performed to build a high-resolution phylogeny. We constructed a maximum likelihood (ML) tree using FastTree (2.1.3). We employed Interactive Tree of Life (iTOL) to visualize the tree [22].

Investigating possible outbreaks
Data extracted from patients' clinical history and WGS data of the outbreak strains were evaluated to investigate possible outbreaks. Likely epidemiological links on transmission (patient to patient connectivity) was predicted if there is overlapping of hospital stay among the outbreak cases (Fig 2). Detailed epidemiological information and subsequent infection control measures were not available as the microbiological information was only available retrospectively. Infection Prevention and Control (IPC) measures were not implemented at the time due to the limited and untimely microbiological information at the DMCH.

Accession numbers
WGS data of all B. cepacia in this study were deposited in the National Centre for Biotechnology Information under accession numbers stated below.
The first case was admitted to paediatrics department of DMCH on the 23 rd of October 2016. According to hospital records, the patient was a suspected case of sepsis, and blood was sent for culture and sensitivity on the 2 nd of November 2016 which was positive for B. cepacia. The rest of the cases were confined to the burn HDU or the burn paediatric HDU or the burn ICU (Fig 2; Table 1). Excluding DAMA (5/135), in the burn unit, the mortality rate for patients

Clonal relationship with global B. cepacia
B. cepacia isolated in this study were assigned as a single novel ST, ST1578 (Fig 3). All strains (n = 106) were grouped into 9 clusters (Fig 3). The Bangladeshi strains shared ST clusters with strains from Australia, USA, South Korea, Thailand, UK and Malaysia. Interestingly, STs differed according to geographical area (Fig 3). Core genome alignment also suggested that the Bangladeshi B. cepacia are genetically closer to environmental strains of Australian origin rather than human strains isolated from cystic fibrosis patients in the UK (LMG16656.fsa nt), Thailand (LO6.fasta), and other Asian strains (Fig 4).

Outbreak investigation
We documented this outbreak between October 2016 and August 2017 at DMCH (Fig 2). All outbreak cases had clinical signs of sepsis at the time of sample collection and patients in burn critical units were undergoing artificial ventilation and had central venous catheter lines and urinary catheters at the time of diagnosis. There was a continuous overlapping of patients at DMCH during the outbreak period (Table 1; Fig 2). Core genome SNPs showed that the outbreak strains were confined to a single clade (Fig 4; S1 Data), corresponded to a common clone (ST1578) (Fig 3).

Discussion
Although B. cepacia is considered almost exclusively a pathogen for CF patients [3,7], our study reported B. cepacia bacteraemia predominantly from burns patients. B. cepacia are considered opportunistic human pathogens and can be transmitted to patients via environmental contamination or person-to-person contact [4][5][6]8,9]. Burns patients are generally more susceptible to infection due to impaired immune function. Risk factors for sepsis in burns include >20% of total body surface area (TBSA), inhalation injury, delayed burn wound excision, increased length of hospital stay, use of artificial medical devices and ICU admission [24][25][26][27][28]. Patients in burn ICUs are more vulnerable to septicaemia than general ICU patients [29]. Our findings demonstrate an outbreak of bacteraemia in DMCH caused by a single clone of B. cepacia ST1578, mostly confined to patients admitted in burn critical care units (Fig 2). The outbreak cases had a mean hospital stay of 43.9 days and straddled each other implying transmission via direct patient contact and/or patient to hospital environment and vice versa. Clinically, this outbreak was associated with a mortality rate of 31% (Fig 2; Table 1).
Previous studies show that burn sepsis accounts for 50-60% of deaths in burn patients [30,31]. In this study, mortality due to burn sepsis with bacteria other than B. cepacia (60.17%) was significantly higher than B. cepacia sepsis (33.33%) (p<0.1). A limitation of the current study was the lacked sufficient clinical information to access and to investigate why the mortality rate was significantly lower for B. cepacia cases. At the time of enrolment many of these patients were on "inadequate" antibiotics based on our susceptibility data. However, there are limitations to the interpretation of AST results for B. cepacia and we lack data on whether antibiotics were changed later during hospital stay. It is possible there were other factors favouring Table 2. MIC value of B. cepacia in this study against relevant antibiotics (n = 15). survival in the group infected with B. cepacia, including possibly relatively lower pathogenicity of B. cepacia as a cause of bacteraemia. B. cepacia has been described as a cause of pseudo-outbreaks [32]. Although this possibility should be considered here the timing of infections suggest this was not the case, and a persistent environmental source of cross contamination on the burn unit was a more likely cause. Unsurprisingly, radical infection control programs can mitigate the spread of infections and improve patient outcomes in burn units [33]. No IPC intervention was undertaken here because the outbreak was identified retrospectively. MDR Gram-negative bacteria infections have become a serious challenge in health care settings as a result of both intrinsic and acquired resistance mechanisms, limiting therapeutic options [8]. B. cepacia is of concern due to their intrinsic resistance to clinically relevant antibiotics such as aminoglycosides and polymyxins [34]. In this study, all B. cepacia showed very high MIC breakpoints to amikacin and gentamicin mediated by resistance genes ant, aph and aadA2 and/or overexpression of efflux pump such as ArmA, CoeA (Fig 4; S1 Data) [34][35][36]. Burkholderia spp. are typically resistance to colistin due to a unique intrinsic amino arabinose biosynthesis operon [33]. Genome sequencing also identified resistance genes such as adeF, qacH, tetC and sul1 (Fig 4; S1 Data) which contribute to phenotypic resistance of B. cepacia ( Table 2). Mutations in ampD are associated with the upregulation of ß-lactams degrading enzymes, PenB and AmpC. This mechanism has been found to be one of the causes of ß-lactam resistance in B. cepacia [18,34]. Mutations in ampD were identical in all B. cepacia analysed in our study and all the isolates had high MIC value for amoxicillin-clavulanate; however, MICs for piperacillin-tazobactam and cephalosporins were variable (Table 2). Perhaps enzymatic degradation was mainly responsible for putative ß-lactam resistance in this study although we did not evaluate the expression of ß-lactamase such as PenB, AmpC or PenA in relation to AmpD mutations [18,34]. Although the empirical antibiotics were shown  to be ineffective (Table 1), the relationship between patients' morbidity/mortality and antibiotics prescription cannot be fully explored as datasets are incomplete.

Strain ID AUG a Pip-Taz a CRO a CAZ a CTX a CEF a IMP a MEM a CIP a LEVO a AK a GEN a Fos a SXT-TRM a Tige a Cl
Virulence of B. cepacia is typically related to adhesins, invasins, intracellular pathogenicity, antiphagocytic factors, secretory and signalling systems [19,37]. A set of virulence genes in relation to all steps in pathogenesis were identified in Bangladeshi outbreak strains B. cepacia (Fig 5; S1 Data). However, whether other virulence determinants are associated with the ability to cause bacteraemia, and the role of patient factors, could not be determined by this current study.
B. cepacia identified in this study belonged to a novel clonal type ST1578 (Fig 3). The genetic background of the outbreak strains was very similar; however, a few variations were found regarding the presence of virulence genes (Fig 5; S1 Data). Compared to global strains from the database, resistance genes for aminoglycosides (ant(2'')-Ia, aph(6)-Id, aph(3'')-Ib, aadA2), fluoroquinolones (qacH), tetracycline (tet(C)) and sulphonamide (sul1) were found only among the strains isolated in this study (Fig 4; S1 Data). Virulence gene, aai was absent in our outbreak strain which was common in both human strains, LMG16656.fsa nt and LO6.fasta, (from Burkholderia Genome Database), but the overall virulence pattern of the outbreak strains was most similar to LMG16656.fsa nt (human strain isolated from the UK) than LO6.fasta (human strain isolated from Thailand). Compared to B. cepacia LO6.fasta, flagellar protein (lfg, lfh, lfi), secretory system (clpV1, vsc, iagB, spaR), regulatory protein (prrA) were absent in the Bangladeshi strains (Fig 5; S1 Data). Thus, it can be inferred that the Bangladeshi outbreak strains identified in this study are genetically distinct from global strains retrieved from the database.
Although we did not fully evaluate the epidemiological link between patients and the environment, epidemiological and molecular data suggest that the outbreak clone was circulating within DMCH over a protracted period (Fig 2). It is disconcerting that antibiotic pressure might have enhanced the elevation in antibiotic resistance during the outbreak [35,36]. Identification of environmental sources of the outbreak followed by patient management can reduce the risk of infections in vulnerable populations [4][5][6][10][11][12][13]29].