Clinical outcomes, molecular epidemiology and resistance mechanisms of multidrug-resistant Pseudomonas aeruginosa isolated from bloodstream infections from Qatar

Abstract Background Bloodstream infections (BSIs) caused by multidrug-resistant (MDR)-Pseudomonas aeruginosa are associated with poor clinical outcomes, at least partly due to delayed appropriate antimicrobial therapy. The characteristics of MDR-P. aeruginosa bloodstream isolates have not been evaluated in Qatar. Our study aimed to examine in vitro susceptibility, clinical and molecular characteristics, and mechanisms of resistance of MDR-P. aeruginosa bloodstream isolates from Qatar. Materials and methods We included all MDR-P. aeruginosa isolated from blood cultures taken between October 2014 and September 2017. Blood cultures were processed using BD BACTEC™ FX automated system. BD Phoenix™ was used for identification, Liofilchem® MIC Test Strips for MIC determination. Whole-genome sequencing was performed using the Illumina-HiSeq-2000. Results Out of 362 P. aeruginosa bloodstream isolates, 16 (4.4%) were MDR. The median patient age was 55 years (range 43–81) and all patients presented with septic shock. Most patients received meropenem (12/16) and/or colistin (10/16). Clinical response was achieved in eight patients, and five patients died within 30-days. MDR-P. aeruginosa isolates belonged to 13 different sequence types. All isolates were non-susceptible to cefepime and ciprofloxacin. The most active agents were colistin (16/16) and aztreonam (10/16). Seven isolates produced blaVIM, and four possessed genes encoding extended-spectrum β-lactamases. Aminoglycoside modifying enzymes were present in 15/16, transferable qnr-mediated quinolone resistance gene was detected in 3/16, and the novel ciprofloxacin modifying enzyme CrpP-encoding gene in one isolate. Conclusion MDR-P. aeruginosa BSIs are relatively uncommon in Qatar but are highly resistant, harbour multiple resistance genes, and are commonly associated with unfavourable clinical outcomes. Colistin was the only agent with consistent activity against the study isolates. Key messages MDR-P. aeruginosa constituted <5% of P. aeruginosa blood isolates over three years. Typical risk factors for MDR infections were highly prevalent in the study population and overall clinical outcomes are consistent with those previously reported. Colistin was the only agent with consistent antibacterial activity against the study isolates.


Introduction
Pseudomonas aeruginosa possess a remarkable array of intrinsic and acquired antimicrobial resistance mechanisms, often expressed simultaneously and resulting in multidrug-resistant (MDR) phenotypes [1]. Risk factors for MDR-P. aeruginosa infection includes prior antimicrobial therapy, the presence of indwelling medical devices, neutropenia, mechanical ventilation, and previous gut colonization with MDR-P. aeruginosa [2]. Bloodstream infections (BSIs) caused by MDR-P. aeruginosa are associated with poor clinical outcomes, including prolonged hospitalization, increased healthcare costs, and high mortality [3]. Such outcomes are at least in part due to delayed appropriate antimicrobial therapy and the limited availability of effective treatment options [4]. However, the clinical, microbiological, and molecular characteristics of bloodstream MDR-P. aeruginosa isolates have not been evaluated in Qatar. Our aim was to investigate the clinical characteristics and outcomes of MDR-P. aeruginosa BSIs from Qatar, assess their in-vitro susceptibility, and to investigate their molecular epidemiology and resistance mechanisms.

Materials and methods
Patients were identified prospectively from routine clinical specimens received by the Division of Microbiology at Hamad Medical Corporation (HMC) in Doha, Qatar. The facility provides routine and tertiary diagnostic services for all primary health centres and public hospitals across the whole country. MDR-P. aeruginosa were defined as isolates with in-vitro resistance to !1 agent from !3 antipseudomonal classes of antimicrobials [5]. Consecutive patients with blood cultures yielding growth of MDR-P. aeruginosa during the period from October 2014 to September 2017 were included. Clinical data were retrieved from the electronic healthcare system. Blood cultures were processed using BD BACTEC TM FX automated system (Becton Dickinson, USA). Bacterial identification and initial antimicrobial susceptibility testing were performed on BD Phoenix TM (Becton, Dickinson and Company, Franklin Lakes, New Jersey, United States). Liofilchem V R MIC gradient strips (Liofilchem, Roseto degli Abruzzi, Italy) were used for minimum inhibitory concentration (MIC) determination. Broth microdilution was used for colistin susceptibility testing (ComASP Colistin, Liofilchem, Roseto degli Abruzzi, Italy). Escherichia coli ATCC 25922, E. coli ATCC 35218 and P. aeruginosa ATCC 27853 were used as controls. Clinical Laboratory Standards Institute (CLSI) breakpoints were used to interpret susceptibility results [6]. Intermediate and resistant categories were grouped as non-susceptible for all reported antimicrobial agents.

Results
Over the study period, MDR-P. aeruginosa constituted 16 (4.4%) out of 362 episodes of P. aeruginosa bacteraemia. The included isolates were from blood cultures from Hamad General Hospital (12/16), Rumailah Hospital (3/16), and the National Centre for Cancer Care and Research (1/16). No MDR-P. aeruginosa were isolated from blood cultures from other HMC facilities during the study period.
The median patient age was 55 years (range 43-81) and the majority were males 93.8% (15/16). Nine patients were in an intensive care unit (ICU) at the time of MDR-P. aeruginosa bacteraemia ( Table 1). All patients presented with septic shock and the majority had multiple risk factors for MDR infections, including hospitalization or outpatient hospital attendance  Table 2).

Discussion
The present study describes MDR-P. aeruginosa bacteraemia over 3 years in Qatar. The proportion of MDR from all bloodstream isolates was 4.4%. This is small compared with reports from other regions such as Spain and Italy [2,14]. Not surprisingly, co-morbidities were highly prevalent in our patients and all had severe clinical presentations including septic shock [15]. However, our overall mortality was (5/16) at 30days. Similarly, high mortality rates have been consistently reported with P. aeruginosa bacteraemia [16,17].
Antimicrobial resistance mutations can result in loss of fitness and reduced P. aeruginosa virulence [1]. However, from previous studies of invasive MDR-P. aeruginosa disease the production of potent virulence factors, such as exoU type III secretion system, have been linked to poor clinical outcomes in association with MDR-P. aeruginosa infections, but in the present study all cases of death express common virulence factor genes such as exoT and exoY with variable expressions of exoS and exoU which cannot be reliably linked to mortality (Table 3) [18]. Other important risk factors for mortality in patients with P. aeruginosa bacteraemia include multiple co-morbidity states, critical illness, shock, and older age, all of which were common in our patients [19,20].
Similar to the previous studies, colistin was the only agent with consistent in-vitro activity against MDR-P. aeruginosa bloodstream isolates included in the present study [21,22]. However, the clinical use of colistin is fairly problematic given its toxicity and the continuing uncertainty about its appropriate dosing [23]. The high prevalence of resistance to antipseudomonal b-lactam agents in this study is associated with the presence of class A ESBL (i.e.; bla VEB-9 , bla CTX-M-15 ), MBL, and PDC enzymes. Vietnamese extended-spectrum b-lactamase (VEB) and Verona integron-encoded metallo-b-lactamase (VIM) enzymes are established in P. aeruginosa from our region [24]. bla NDM and bla IMP , both of which are occasionally identified in MDR-P. aeruginosa from our region, were not detected in this study [24,25]. The majority of isolates were susceptible to aztreonam, which is not susceptible to the hydrolytic activity of MBLs or the narrow-spectrum OXA b-lactamases found in this study [26]. Ceftazidime/avibactam and ceftolozane/tazobactam were active against 8/16 MDR-P. aeruginosa isolates reported here. The study pre-dated their availability for clinical use in Qatar and none of the patients in this cohort was treated with these agents. Clinical data on the use of ceftazidime/avibactam and ceftolozane/tazobactam for the treatment of patients with infections caused by MDR-P. aeruginosa are mainly derived from retrospective studies [27,28]. Though the data are encouraging, their utility depends on the local epidemiology and prevailing P. aeruginosa resistance mechanisms [29]. Predictably, neither agent was active against VIM-producing MDR-P. aeruginosa isolates reported here (Tables 1 and 4).
All isolates included in this study were resistant to ciprofloxacin. Rates of fluoroquinolone resistance in MDR-P. aeruginosa are usually very high [30]. Fluoroquinolone resistance is predominantly mediated by well-characterized mutations in the gyrA (T83I), parE (A473V), and topoisomerase IV encoding genes, in addition to upregulation of efflux mechanisms [1,31]. Two types of transferable quinolone resistance genes, Qnr and CrpP, were also detected in this study. Qnr-encoding gene confers fluoroquinolone through target protection, while CrpP-mediates antibiotic modification through phosphorylation [32]. A variety of Qnr-encoding genes in P. aeruginosa have been described from multiple countries in the Arabian Peninsula and North Africa [33]. However, to our knowledge, the presence of CrpP-encoding gene in P. aeruginosa isolates from this region has not been previously reported.
In-vitro susceptibility of MDR-P. aeruginosa included in this study to aminoglycosides ranged from 31.3% for gentamicin to 43.8% for amikacin. Aminoglycoside resistance is driven mainly through the production of a variety of well-established AME [1]. Considering the antimicrobial agents available for clinical use at the time of the study, five (25%) isolates were only susceptible to colistin and an aminoglycoside. However, the efficacy of monotherapy with either of these for BSIs is questionable; their combined use is associated with an increased risk of toxicity [23].
A wide variety of sequence types were represented in this study's MDR-P. aeruginosa bloodstream isolates, including the high-risk ST233 and ST357 [34]. Other previously reported high-risk P. aeruginosa clones such as ST235, ST111, and ST175 were notably absent [1].  T83I   T83I   T83I   T83I   S80I, T83I   T83I   T83I   S80I   T83I   T83I   T83I   T83I   T83I   T83I  Conclusions MDR-P. aeruginosa BSIs are relatively uncommon in Qatar, representing less than 5% of P. aeruginosa blood isolates over three years. Typical risk factors for MDR infections were highly prevalent in the study population and overall clinical outcomes are consistent with those previously reported. Its multiple clinical limitations notwithstanding, colistin was the only agent with consistent antibacterial activity against the study isolates. Alternatives, such as newer b-lactamb-lactamase inhibitor combinations and aminoglycosides, are active against half of the isolates or less.

Ethical approval
This study was approved by the Research Ethics Committee (Protocol number IRGC-01-51-033) at Hamad Medical Corporation (HMC), Doha, Qatar, with a waiver for informed consent.