Phenotype-genotype correlations among carbapenem-resistant Enterobacterales recovered from four Egyptian hospitals with the report of SPM carbapenemase

Carbapenem-resistant Enterobacterales (CRE), currently listed by the World Health Organization (WHO) as top priority critical pathogens, are a major global menace to human health. In low- and middle-income countries (LMICs) the threat is mounting fueled by selective pressures caused by antibiotic abuse and inadequate diagnostic resources. This study phenotypically and genotypically characterized carbapenem resistance among 115 Enterobacterales isolates including 76 Klebsiella (K.) pneumoniae, 19 Escherichia (E.) coli, 14 Shigella (S.) sonnei, 5 Enterobacter (E.) cloacae, and 1 Proteus (P.) mirabilis. Ninety-three isolates (80.9%) were carbapenem-resistant with an alarming 57.5% carbapenem non-susceptibility in isolates collected from the outpatient department. Molecular characterization of the carbapenemases (CPases) encoding genes showed that blaNDM (80.5%) was the most prevalent; it was detected in 62 isolates (54 K. pneumoniae, 6 E. coli and 2 S. sonnei), followed by blaVIM (36.4%) which was observed in 28 isolates (24 K. pneumoniae, 3 E. coli and 1 E. cloacae). Other CPases included blaKPC (28.6%; in 20 K. pneumoniae, 1 E. coli and 1 S. sonnei), blaOXA-48 (26%; in 17 K. pneumoniae, 1 E. coli,1 E. cloacae and 1 P. mirabilis), blaIMP (6.5%; in 5 K. pneumoniae) and blaSPM (1.3%; in K. pneumoniae). Notably more than half of the Enterobacterales isolates (54.5%) co-harboured more than one CPase-encoding gene. Co-existence of blaNDM and blaVIM genes was the most dominant (31.2%), followed by association of blaNDM and blaKPC (24.7%), then blaVIM and blaKPC (13%). Moreover, the effects of different genotypes on meropenem MIC values were assessed, and a statistically significant difference between the genotype (Ambler classes A and B) and the genotype (Ambler classes B and D) was recorded. The current findings may serve for a better understanding of the context of CRE in Egypt, associated drivers and CPases.

were the main consumers, with a surge in the use of "last-resort" antibiotics like carbapenems and polymyxins [3]. As a consequence, carbapenem-resistant Enterobacterales (CRE) endemic and epidemic emergence has been widely reported in hospitals [4,5] and infrequently in community settings [6,7]. Production of CPases comprises the main mechanism of carbapenem resistance in Enterobacterales. This further emphasizes the criticality of CRE, since CPases-encoding genes are typically found on mobile genetic elements (plasmids, transposons, etc.) which facilitates their dissemination [8]. Moreover, studies have shown the association of these genes with other AMR encoding genes such as extended-spectrum beta-lactamases (ESBLs), genes mediating resistance to aminoglycosides and fluoroquinolones which results in the upsurge of multidrug-resistant phenotypes [9]. Since CPasesencoding genes in Enterobacterales show wide diversity according to the geographic location, choosing the appropriate antibiotic is highly dependent upon rapid and accurate diagnostic tests. For instance, therapeutic decisions involving beta-lactamase inhibitors may fail against CPases, such as oxacillinases and metallo-betalactamases (MBLs), which necessitates identifying CPases on both the phenotypic and genomic levels [10]. Accordingly, LMICs encounter huge challenges due to scarce financial resources required for capacity building of efficient clinical laboratories to perform molecular-based tests. Relying solely on phenotypic methods to routinely report carbapenem resistance poses several challenges. First, the delay encountered to obtain the results of conventional culture-based approaches, which may force clinicians to rely on local hospital susceptibility data, as well as regional and national surveillance studies to guide their empirical treatment decisions. Besides, the obtained susceptibility patterns do not distinguish between carbapenem resistance due to carbapenemases production and that caused by other combined mechanisms such as extended spectrum beta-lactamases, AmpC and reduced permeability. As a result, carbapenems may be indicated for treatment even though a CPase is not present, hence unfavourable outcomes of potential antimicrobial resistance may arise. In Egypt, several studies were conducted to explore the prevalence of CRE strains among inpatients, nonetheless data concerning community acquired carbapenem resistance is lacking [11]. Considering all these challenges, we aimed to catch a glimpse of phenotype-genotype correlations among carbapenem-resistant Enterobacterales isolates from four Egyptian hospitals. We investigated the dissemination and the co-occurrence of CPases-encoding genes among the collected CRE isolates. Additionally, we tested the correlation between the identified genotypes and features such as the isolated organism, specimen source and department of isolation.

Carbapenem susceptibility testing and MIC determination
The susceptibilities of the Enterobacterales isolates to imipenem (IPM) and meropenem (MER) were determined by the disc diffusion method according to the Clinical and Laboratory Standards Institute (CLSI) guidelines [14]. The susceptibility patterns of the isolates were interpreted according to the CLSI M07-A11 protocol [15]. Isolates were considered as CRE if they were found resistant or intermediate to one or both carbapenems (IPM and MER). Micro-broth dilution method was used to determine the minimum inhibitory concentration (MIC) of meropenem (AstraZeneca, UK) according to the CLSI guidelines [14]. Negative control wells contained only culture media to ensure sterility, while positive controls were inoculated with the organism, and E. coli ATCC 25922 was used as a reference strain. Following a 24 h incubation at 37 °C, the MIC readings were recorded and interpreted according to the CLSI M07-A11 protocol [15].

PCR amplification of carbapenemases genes
Bacterial DNA was extracted using GeneJet ™ DNA extraction kit (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer's instructions and stored at − 20 °C for future use. PCR was performed using previously described primers (Thermo Fisher Scientific, Waltham, MA, USA) to amplify carbapenemasesencoding genes from Ambler class A: bla KPC [16], class B metallo-beta-lactamases (MBLs): bla VIM , bla IMP , bla SPM [17], bla NDM [16], and class D oxacillinases: bla OXA-48like [16]. To easily interpret the expected bands and to avoid obtaining amplification products of the same size in each PCR tube, 2 multiplex reactions were performed for the detection of the 6 genes under investigation.

Statistical analysis
All analyses were carried out using R statistical platform (https:// www.r-proje ct. org) in R-studio version 1.4.1106. Several R-packages were used in data analysis and visualization including readxl, ggplot2, polycor and lares.
Fisher's Exact (FE) test of independence was employed to analyze the associations between nominal variables. Moreover, the correlation between these variables was performed using the Spearman's rank correlation and strength of the association was expressed as Spearman's correlation coefficient (r s ) between − 1 and + 1. Kruskal-Wallis (KW) test was used to compare the medians for non-normally distributed quantitative data and Mann-Whitney test was applied as a post-hoc test using Bonferroni correction method for multiple comparisons. For all statistical analyses, p-values ≤ 0.05 were considered statistically significant.

Molecular analysis of CPase genes
The results of multiplex PCR assays are shown in Additional file 1:

Features associated with the carriage of CPases-encoding genes
We explored the correlation between the detected bla NDM , bla VIM , bla KPC , bla OXA-48 , and bla IMP genes and the documented features of the isolates (Fig. 3). Interestingly, bla NDM , bla VIM , and bla KPC (Fig. 3A-C, respectively) demonstrated statistically significant (p-value ≤ 0.05) positive correlations with the same features, namely, the count of genes detected in combination, K. pneumoniae isolates from sputum cultures, and recovery from inpatients. Likewise, the count of genes detected in combination was also positively correlated with both bla OXA-48 and bla IMP (r s = 0.316 and 0.313, respectively) (Fig. 3D, E, respectively). On the other hand, statistically significant negative correlations were obtained between certain features and the genes bla NDM , bla VIM , and bla KPC . For instance, isolates from urine samples obtained from outpatients were less likely to harbour bla NDM , bla VIM , and bla KPC genes.

Discussion
To tackle the AMR crisis, the WHO assembly adopted a global action plan with five main objectives [18]. One goal is to strengthen the knowledge and evidence base  through surveillance and research. Surveillance data can monitor the drivers of AMR and track trends across multi-sectors [19,20]. Consequently, interpretation of such data endorses the implementation of infection control programs [21,22], besides guiding the choice of resistance testing methods and treatment decisions [10]. Egypt's healthcare-associated infections (HAI) surveillance system, established in 2011, is a leading prospective, standardized system in WHO's Eastern Mediterranean Region (EMR) [23,24]. However, limited studies have reported the increasing burden of CRE in Egypt using the system's surveillance data [5]. On the other hand, the academic network strived to provide research data to fill the knowledge gap regarding the severity of CRE problem in Egypt [11]. Routine testing for CRE resistance mechanisms is still not mandated in many clinical laboratories [25]. Consequently, they are unable to differentiate carbapenem non-susceptibility caused by carbapenemases production from resistance because of other mechanisms such as overexpression of AmpC enzymes and membrane impermeability. The present study aimed to detect CPases, the most important mechanism of carbapenem resistance, among 115 Enterobacterales isolates recovered from inpatients and outpatients in 4 major Egyptian hospitals. K. pneumoniae was the most frequently recovered isolate (66.1%), followed by E. coli (16.5%), S. sonnei (12.2%), E. cloacae (4.3%) and P. mirabilis (0.9%). In outpatients, E. coli and S. sonnei isolates dominated (43% and 30%, respectively). Strikingly, carbapenem resistance was detected in 80.9% of the Enterobacterales isolates. This level of resistance is higher than recent studies in Egypt describing the isolation of CRE with the percentages 46% [26], 59% [27], 68.8% [28] from cancer patients and 48% from intensive care units [5]. In the study performed by Hassuna et al. [29], very high frequency of carbapenem resistance (95%) was detected among the K. pneumoniae isolates. Comparable rates were observed in this study, in which K. pneumoniae cultured from sputum and pus specimens obtained from inpatients were highly enriched with carbapenem resistance 96-100%. In agreement with previous studies [30] we found significant differences in carbapenem resistance between Enterobacterales isolated from inpatients compared to outpatients, which may be due to the selective pressure of excessive consumption of antibiotics in hospitals. It has been estimated that 75% of patients treated mostly for respiratory infections receive empirical carbapenem therapy [31]. Furthermore, a study involving 97,910 strains of K. pneumoniae found a correlation factor of 0.427 (p < 0.01) between carbapenem consumption intensity and rate of carbapenem resistance [32]. In agreement with these observations, the high carbapenem resistance rates in this study underline a pressing need for a national antimicrobial stewardship plan in Egypt to optimize the utilization of carbapenems as treatment options, especially in respiratory tract infections. The WHO provides a practical implementation guidance for the recommendations targeting health care facilities regarding the prevention and control of CRE, comprising how to utilize surveillance results, and put in place antibiotic stewardship programmes [33]. Next, we molecularly characterized the CPases-encoding genes in our isolates. Amid the positive genotypes, bla NDM (80.5%) was predominant, followed by bla VIM (36.4%), bla KPC (28.6%), bla OXA-48 (26%) and bla IMP (6.5%). Similar percentages were recently reported in Egypt [29,34,35]. Even though, the bla SPM was formerly detected in Pseudomonas aeruginosa isolates [36][37][38], here we report the first-time detection of the rare bla SPM in combination with both bla VIM and bla OXA-48 from a K. pneumoniae isolate in Egypt. Our results uncovered substantial co-existence of CPases-encoding genes (54.5%), this may be attributed to the fact that these genes are carried on mobile genetic elements and thus are easily transferred within different hospital settings [39]. Twenty-four isolates (31.2%) co-harboured bla NDM and bla VIM genes. This percentage is higher than in previous studies from Egypt by Kamel et al. [26] and Khalil et al. [40] in which such association was detected in only 1 (3%) and 4 (8.7%) isolates, respectively. Whereas more current investigations reported higher rates ranging between 69% [41] and 100% [29]. On the other hand, association of bla NDM and bla KPC was detected in 19 (24.7%) isolates which is comparable to results by Ramadan et al., (33.3%) [35], but higher than Ragheb et al., (5%) [41], Khalil et al., (8.7%) [40] and El-Kholy et al., (10%) [38]. Additionally, bla VIM and bla KPC were simultaneously present in 10 (13%) isolates, in agreement with previous reports with roughly the same incidences [40,41]. Researchers have formerly assessed the effects of different genotypes on carbapenem MIC values [42]. Similarly, our results record a statistically significant difference between the genotype (Ambler classes A and B) and the genotype (Ambler classes B and D). This finding may be due to the low level hydrolytic activity of bla OXA-48 [43]. It is noteworthy that 16 (13.9%) isolates with phenotypic carbapenem-resistant profiles were negative for the tested CPases genes. In accordance with other studies [7] this may be attributed to types of carbapenemases different than those tested in this study. Besides, the presence of other combined mechanisms such as extended spectrum beta-lactamases, AmpC and reduced permeability to carbapenems caused by porin mutations [44,45].
What is more alarming, is the excessive rates (57.5%) of carbapenem resistance in isolates collected from the outpatient department. Even though, 10 out of the 23 non-susceptible isolates were negative for the genes used in this study, yet the resistance level conferred by the positive CPases genes (32.5%) is worrisome. Regrettably, CRE strains could circulate in community settings and the resulting infections may easily propagate especially with the lack of proper hygiene measures [7]. Additionally, in many LMICs, including Egypt, patients can purchase antibiotics without a physician's prescription [46]. In a recent study conducted among 563 Egyptian medical students, about 77.7% of the participants were self-medicated [47].
A major goal of this study was to detect features associated with carriage of CPases-encoding genes. Successful strategies for preventing the transmission of these problematic strains require thorough and meaningful characterization of these infections in respect to causative organism, department of isolation and specimen source. Such approach may prompt the improvement of the applied infection control measures, provide guidelines for effective empiric therapy, and regulate the use of novel combinations involving beta-lactamase inhibitors. In agreement with previous studies [10], bla NDM and bla KPC were strongly correlated to K. pneumoniae isolated from sputum cultures recovered from inpatients. Correspondingly, improved infection control regimens are needed, primarily with patients suffering from respiratory tract infections caused by K. pneumoniae. A recent study highlighted the association between contamination of hospital's environment and transmission of carbapenemresistant K. pneumoniae in ICUs [48]. Furthermore, long term carriage of bla KPC , for up to 3 years, was defined in patients discharged from a hospital with an outbreak of KPC producing K. pneumoniae [49]. Of note, all the investigated CPases-encoding genes, except bla SPM , were significantly associated with the overall count of genes detected in combination. This implies the high likelihood of the co-existence of the investigated CPases-encoding genes which raises deep concerns about the simultaneous transmission of such genes.
Even though sample-based surveillance enabled us to report the most frequent types of infections caused by CRE, however this study had some limitations. First, the number of clinical isolates was small. Hence, it was insufficient to draw firm conclusions about the true resistance rates except for K. pneumoniae [50]. Second, isolates were obtained anonymously from the microbiology laboratories, so no data were available on the patient's microbiological history and prior antibiotic use. Finally, with respect to outpatients, a common practice in Egypt is to seek antimicrobials without being tested for susceptibility, thus the examined isolates are more likely to be from patients who have failed first line and even second line treatments, which might have caused an overestimation of the resistance rates in this study.

Conclusion
This study underlines the escalating dilemma of CRE prevalence in Egyptian hospitals and sheds the light on the rising incidences in community settings. Here we elucidated the prevalent CPases-encoding genes present either in combination or as single genes. To address this challenge, ongoing identification, and riskbased surveillance of CRE are indispensable. Equally crucial is the urgent implementation of a national antimicrobial stewardship plan, to diminish the current burden of CRE in Egyptian hospitals and circumvent its emergence in community settings.