Successful treatment of infective endocarditis due to pandrug-resistant Klebsiella pneumoniae with ceftazidime-avibactam and aztreonam

Pandrug-resistant (PDR) K. pneumoniae refractory to conventional treatment has been reported worldwide, causing a huge burden on the healthcare system, patient safety and the economy. K. pneumoniae is a prominent opportunistic pathogen causing hospital-acquired and community-acquired infections, but is rarely associated with infective endocarditis. Currently, there are sparse data guiding the optimal regimen when commonly used antibiotics fail, notably for the treatment of endocarditis infections. Here we report our experience in treating a 40-year-old female with PDR K. pneumoniae infection of cardiovascular implantable electronic device (CIED) and right-sided infective endocarditis. Initial susceptibility testing of the incriminated pathogen showed an apparent susceptibility to colistin but the prolonged course of colistin, gentamicin and meropenem did not resolve the infection. However, the synergistic combinations of aztreonam with ceftazidime-avibactam was able to overcome resistance and clear the infection rapidly. Genome sequencing showed that the PDR K. pneumoniae isolate belongs to the international high-risk clone ST14. The isolate harbored genes encoding NDM-1, OXA-48, CTX-M-14b, SHV-28 and OXA-1, explaining resistance to all β-lactams, including carbapenems. It carried the armA gene conferring resistance to all clinically important aminoglycosides and had alterations in GyrA, ParC and MgrB, explaining resistance to ciprofloxacin and colistin.

www.nature.com/scientificreports/ The emergence of co-resistance to β-lactams, aminoglycosides, quinolones, colistin and tigecycline in K. pneumoniae isolates poses a serious therapeutic challenge due to limited treatment options. In recent years, the incidences of pan-drug resistant (PDR) K. pneumoniae infections refractory to conventional treatment have been reported globally, causing a significant increase in long-term hospitalizations, morbidity and mortality 11 . Resistance to last-resort carbapenems in this species is mainly mediated by the production of β-lactamases, notably those belonging to the KPC, NDM, VIM and OXA-48-like type enzymes. In Saudi Arabia, K. pneumoniae is the most contributing organism to carbapenem resistance among all Enterobacterales, increasing from 0 to 33.3% in 10 years (2007-2016) 12 . Resistance to carbapenems in K. pneumoniae isolates from Saudi Arabia are mainly associated with the acquisition of OXA-48-like and NDM carbapenemases, although few recent studies have reported the detection of KPC carbapenemases in Klebsiella spp. isolates 13,14 .
Combination antibiotic therapy has been used as an option to treat patients with life-threatening PDR K. pneumoniae infections 15,16 . However, limited studies showed evidence-based combination antibiotic therapy to treat patients infected by PDR K. pneumoniae. This is particularly true for endocarditis, for which treatment options are already limited by the localization and characteristics of the infection. Most of the available information is driven from in vitro studies and just a few numbers of clinical in vivo studies 17 . The recent approach of therapeutic options such as the β-lactam-β-lactamase inhibitors ceftazidime-avibactam (CAZ/AVI) or meropenem-vaborbactam combinations showed potent inhibitors activity against class A and D carbapenemase producers (e.g. KPC, OXA-48-like, GES) but were ineffective against class B carbapenemase producers (e.g. NDM, VIM, IMP) 18 . Clinical trials and in vitro studies have demonstrated the activity of CAZ/AVI against ESBL-, AmpC-, KPC-and OXA-48-producing pathogens [18][19][20][21] . The combination of CAZ/AVI has been approved to be used as a therapeutic option to treat adults with complicated urinary tract infections, hospital-acquired pneumonia and other infections caused by MDR gram-negative pathogens 22 . Other studies have shown that CAZ-AVI plus aztreonam (ATM) can be an effective therapeutic combination against metallo-β-lactamases (MBLs) 41 . Here, we report the successful treatment of CIED and right-sided IE due to carbapenemase OXA-48-and NDM-producing K. pneumoniae strain ST14 using CAZ-AVI plus ATM only.

Results
Case record, diagnostic and antibiotic treatment. The patient is a 40-year-old female with known rheumatic heart disease since childhood who had a bioprosthetic mitral valve replacement in 2015. In July 2017, the patient had the mechanical valve replaced due to valve thrombosis and an insertion of a dual-chamber pacemaker to treat arrhythmia. In October 2017, she developed prosthetic mitral valve endocarditis due to Enterococcus faecalis that was cleared by six weeks of treatment with a combination of ampicillin and ceftriaxone. In January 2019, the patient had another infective endocarditis affecting the tricuspid valve, and the pacemaker leads due to a carbapenem-resistant K. pneumoniae, for which she was managed during ten days in two different hospitals before being transferred to our specialized cardiac center following several cardiac arrests from septic pulmonary emboli (Fig. 1). The antibiotic susceptibility performed in the transferring hospital suggested apparent susceptibility to gentamicin and colistin, and so, the patient was put on colistin (2.5 million IU twice daily) and gentamicin (7 mg/kg once daily) intravenously (IV), to which 500 mg meropenem twice daily due to acute kidney injury was added at day one. However, repeated blood cultures showed persistent bacteremia with a K. pneumoniae strain resistant to all tested antimicrobials, including those used for treatment (Table 1). A synergy of antibiotic combinations showed that only CAZ/AVI with ATM was an effective option for treatment (Table 1). Accordingly, the therapy was changed on day five to CAZ/AVI (2.5 gm every 8 h) and ATM (2 gm every eight hours) IV. Her fever resolved, and bacteremia cleared right after starting this combination, while inflammatory markers and creatinine levels improved over a couple of weeks. At 42 days of therapy, the pacer device and the lead tip were replaced. The culture from the tip of the lead did not grow any bacteria, and consequently, the antibiotic treatment was stopped at day 50 ( Fig. 1). Although the source of the bacteremia was not found, the follow up at six months after discharge showed that the patient was completely healthy and asymptomatic.
Genome sequence analysis. Genome sequencing showed that the carbapenem-resistant K. pneumoniae SA-KpST14 isolate harbored five different plasmids ( Table 2). In silico analyses identified the strain as sequence type (ST)14 and detected genes explaining resistance to β-lactams, aminoglycosides, quinolones, phenicols and fosfomycin, as shown in Table 1. Otherwise, the genetic disruption of the mgrB regulator by insertion sequence IS5 explained resistance to colistin.

Discussion
Right-sided IE accounts for 5-10% of all IE cases, and more than 50% of these cases are due to intravenous drug use 6 . The presence of an implantable endovascular device imposes a higher risk of developing gram-negative endocarditis along with hospital stay, history of invasive procedures and other risk factors for developing gram-negative bacteremia in general 8 . The weaker adhesion ability of gram-negative bacteria (GNB) has been attributed to the low prevalence of CIED and IE caused by non-HACEK GNB 35 . Presentation can be associated with septic emboli to the lungs as in the presented case 6 . Liver abscess and urinary tract infection are the most common source when bacteremia is present 36 . Mortality rates vary depending on the pathogen, virulence factors, complications and the valve involved, reaching up to 49% in one article 6,37 . Urgent surgical interventions as a first-choice treatment is recommended for uncontrolled infection to prevent complications, including heart failure and embolic events 35,38 .
In Enterobacterales, PDR is observed among carbapenemase-producing bacteria, especially among K. pneumoniae, as it can easily acquire mobile genetic elements through horizontal gene transfer 39 . The risk of acquiring carbapenemase-producing bacteria increases in a patient with prior surgery, extended hospital stays and the presence of wounds 40 .
Latest Infectious Diseases Society of America (IDSA) guidelines on the treatment of carbapenem-resistant enterobacterales (CREs) in general recommended the use of CAZ/AVI with ATM or cefiderocol monotherapy for MBL-producers (e.g., NDM, VIM or IMP) and CAZ/AVI monotherapy or cefiderocol monotherapies for OXA-48-like producers 41 . While second-line options included tigecycline, eravacycline, colistin and fosfomycin in limited indications and they recommended against combinations of antimicrobials when a β-lactam is susceptible 41 .
However, the guideline did not focus on cases of CIED or IE but rather gave a recommendation for infections outside the urinary tract in general, and no recommendations were given for conditions where two or more resistance genes are detected within the same species 41 .
Therapeutic options used in the literature with successful results for carbapenemase co-producing (NDM-1 and OXA-48 like) K. pneumoniae includes the use of ATM in combination with CAZ/AVI as this has shown promising results in NDM producing Enterobacterales in-vitro and in-vivo. It is thought that this is due to the efficacy of ATM against MBLs in general with the addition of the effect of the avibactam component in CAZ/AVI on extended-spectrum β-lactamases, and ambler class A, and D carbapenmases, which are often co-produced by some strains 22,42 . The presence of serine β-lactamases along with NDM-1 gene was detected in up to 30% in one study for which the combination of CAZ/AVI and ATM has shown synergy in vitro and in vivo 43 . Several studies have proposed this effect 42,[44][45][46][47] , and other β-lactamase inhibitors (Clavulanate and Tazobactam) has also been tested with ATM showing variable degrees of successful results 48 . It is worth mentioning that an in-vitro study was comparing the synergy of CAZ/AVI plus ATM with Meropenem-Vaborbactam plus ATM in NDM-1 non-OXA-48 like co-producer E. coli and K. pneumoniae strains which showed similar synergy against these CREs 44 . A single product formulation of aztreonam-avibactam is currently in phase III clinical trial, which will address many issues with using a two-drug combination like susceptibility testing and epidemiologic surveillance data 43 . Clinicians should be aware that resistance to CAZ/AVI in K. pneumoniae may emerge while on treatment, and meropenem susceptibility may be restored as previously reported by mutations in the omega loop of bla KPC in the carbapenemase-producing strain, which may require testing MICs every time a phenotypic or genotypic alteration occur 49 .
In the presented case, we started the CAZ/AVI combination with ATM on the 4 th day of hospital admission, which resulted in rapid clearance of bacteremia (in one day). The decision to start this combination was guided by the available literature at that time (February 2019) and synergy testing using gradient diffusion strips showing positive synergy results. The fact that the patient avoided the indicated open heart surgery for valve replacement just by using this combination proves that it is an effective antimicrobial combination in similar cases.

Materials and methods
Bacterial isolate and antimicrobial susceptibility testing. The K. pneumoniae strain, namely SA-KpST14, was recovered from a 40-year-old female with rheumatic heart disease (RHD) at Prince Sultan Military Medical City, Riyadh, Saudi Arabia. The initial antimicrobial susceptibility testing was performed using the VITEK-2 system (BioMerieux, Brussels, Belgium). Etest (bioMérieux, Durham, NC) was used to determine the MICs for aztreonam, imipenem, ceftazidime/avibactam and synergetic activities of antibiotic combinations listed in Table 1. Antimicrobial MIC interpretations were in accordance to CLSI guidelines. A zone of hope for antimicrobial activity was defined based on the definition of synergy (1 plus 1 equals more than 2); hence, a zone of hope is defined as (0 plus 0 equals more than 1) 50 . MICs of colistin were confirmed by broth dilution methods done according to CLSI guidelines. Carba-R test using GeneXpert system (Cepheid, USA) was initially used to detect the presence of carbapenemase-resistance genes.
Complete genome sequencing. Genomic DNA (gDNA) of the K. pneumoniae SA-KpST14 strain was extracted from an overnight culture on LB agar using the QIAamp DNA Mini Kit (QIAgen, Germany) according to the manufacturer instructions. The quality and purity of the extracted DNA were checked using the Nanodrop 2000 spectrophotometer (Thermofisher, USA) and Qubit 3.0 Fluorometer with the dsDNA HS (High sensitivity) kit (Thermofisher, USA). Short reads sequences were generated on the Illumina MiSeq platform using the Nextera-XT library preparation kit (Illumina, San Diego, CA). Long reads sequencing with the Oxford Nanopore Technology (ONT) were generated using the ligation sequencing kit SQK-LSK109 (Oxford Nanopore Technologies, Ltd., UK) on the MinION sequencer (Oxford Nanopore Technologies, Ltd., UK). Identification of antibiotic resistance, virulence genes and plasmid replicon typing. Identification of antibiotic resistance genes and virulence factors were determined with ABRicate (https:// github. com/ tseem ann/ abric ate) (version 0.9.8) using the ResFinder (version 2.1) 54 , Comprehensive Antimicrobial Resistance (CARD) 55 , virulence factors (VFDB) 56 and Kaptive (version 0.7.3) 57 databases. Basic plasmid characteristics were determined using the PlasmidFinder (version 1.3) 58 software. Plasmid maps were drawn using the Blast Ring Image Generator (BRIG) 59 software and Easyfig 60 .

Ethical approval. This study was approved by the Cardiac Research Department of Prince Sultan Cardiac
Center in Riyadh, Saudi Arabia (Reference number R21004). We confirm that all research in this study was performed in accordance with the relevant guidelines and regulations after obtaining informed consent for conducting and publishing this study. All rules and regulations of ICH-GCP and the Declaration of Helsinki were followed.
Informed consent. Informed written consent was taken from the patient to conduct and publish this study and no personal data will be disclosed or breached beyond the principal investigator.