Genetic Study of Extended Spectrum Beta-Lactamase and Carbapenemase Producing Escherichia Coli Causing Sepsis among Egyptian Children

This is a cross-sectional study performed on 88 patients with sepsis. The isolated E. coli were identified by Gram stain and biochemically by the Microscan automated system. ESBL and carbapenemase producing E. coli were isolated on double disk diffusion and EDTA double disk, respectively. Polymerase chain reaction for ESBL and carbapenemase producing E. coli genes were performed. Bacterial susceptibility to antibiotics was tested. The initial results were measured through the 30-days of hospital admission. IRB approved the study.


INTRODUCTION
Development of antibiotics resistance is a major health problem all over the world.Among the common resistance pattern is the resistance of Escherichia coli (E.coli) to Extended-Spectrum Beta-Lactamase (ESBL).Recently, resistance to car-by porin loss.The second pathway is associated with the production of EBSL or of Amp C-type beta-lactamase; betalactamase production is capable of hydrolyzing carbapenems (carbapenemases) [1 -3].Carbapenem resistance pathways are linked to the presence of transferable genes carried on transferable elements such as plasmids and transposons that can be spread between the different species of Enterobacteriaceae [4].
Carbapenem resistance genes are classified to class A (KPC), B Metallo-beta-lactamases (IMP, VIM, NDM) and class D (OXA-48) serine carbapenemases.The resistance pattern of carbapenemase-producing E. coli (CPE) depends mainly on the type of gene responsible for carbapenemase enzyme production.Basically, carbapenem-resistant Enterobacteriaceae hydrolyze imipenem, ertapenem, meropenem, and doripenem [5].Amber class A carbapenemase is inhibited by boronic acid, clavulanic acid and Tazobactam.The class B β-lactamases has a broad spectrum of hydrolytic activity, including all penicillins, cephalosporins, and carbapenems with sensitivity only to aztreonam [5].The activity of this class of enzyme is not inhibited by clavulanic acid, Tazobactam, or sulbactam with inhibition of the enzyme by a chelator for zinc ions, a cofactor of its activity, such as Ethylenediaminetetraacetic acid (EDTA) [5,6].The amber class D has a peculiar pattern of resistance to ceftazidime and resistant to inhibition by clavulanic acid and Tazobactam [1].
The resistance pattern of CPE and the distribution of carbapenem resistance genes varies according to geographical regions [1 -4].KPC carbapenem resistance gene was found among USA, Greece, and Israel isolates [7], VIM in Greece [1,6], OXA-48 in North Africa and Turkey [5], and NDM in the Indian subcontinent and the Balkans [7].
The determinations of CPE by phenotypic methods are difficult and time consuming.Moreover, clinical isolates with carbapenemase activity rarely have an isolated pattern of resistance as it is usually associated with an Extended-Spectrum β-Lactamases (ESBLs), leading to a broader composite resistance profile [5].
The development of molecular methods for genetic determination of CPE has evolved to be an accurate, sensitive and rapid method for detection of such strains.The used techniques either single Polymerase Chain Reaction (PCR), multiplex PCR and sequencing provide precise identification of carbapene-mase resistance [8].
There are data concerning the association of E. coli with carbapenemase resistance among serious types of infections in adults [9].However, there is insufficient data about the association of CPE with sepsis in children.The aim of this study is to detect the prevalence of ESBL-and carbapenemasegenes among E. coli isolated from children with suspected septicemia.

Design, Setting and Population Studied
This cross-sectional study was carried out on children with suspected or diagnosed sepsis after two calendar days of admission to the pediatric intensive care unit in a tertiary-care university hospital, Egypt from January 2016 until January 2018.The study was approved by Mansoura Ethical Committee (code number: R/ 18.08.250).Informed consent was taken from patients' parents who accepted to participate in the study.Confidentiality of the data was considered.Patients, aged from 1 day to 16 years old, diagnosed to have sepsis according to the International Pediatric Sepsis Consensus Conference 2005 [10] were recruited in the study.

Data Collection
The clinical data of each child was obtained from the medical records consisting of demographic data, comorbid conditions, duration of hospital stay and the presence of devices such as mechanical ventilator, Central Venous Catheter (CVC) or urinary catheter at the time of bacterial blood culture collection.All patients started empirical broad-spectrum antibiotics until the results of blood cultures.Patients' survival was registered through 30-days of hospital admission.

Determination of ESBL Production by Double-Disc Synergy Test (DDST)
Escherichia coli strains resistant to ceftazidime (30 μg), cefotaxime (30 μg) were subjected to determination of ESBLs production by DDST.Briefly, overnight culture of E. coli adjusted to 0.5 McFarland standard was subculture on Muller-Hinton agar plate with application of disks containing ceftazidime, ceftazidime (30 μg) + clavulanic acid (10 μg), cefotaxime, and cefotaxime (30 μg + clavulanic (10 μg) pairs were placed with 20 mm space between them.Plates were incubated overnight at 37ºC.The increase of ≥5 mm inhibition zone of growth in ceftazidime/clavulanic acid and cefotaxime/clavulanic compared to ceftazidime and cefotaxime was regarded as an ESBLs producing isolate [12].The following strains were used as positive control strains for ESBL, E. coli ATCC 25922 for blaCTX-M-and E. coli AF427133.1 for TEM and SHV beta lactamase (Grenole Cedex2, France).E. coli ATCC 25922 was used as a negative control strain.

Determination of Carbapenemases Production by Combined Disk Test (CDT) and Boronic Acid Discs
Resistant strains of E. coli strains to imipenem (10 μg) and/or to meropenem (10 μg) were subjected to further determination of carbapenemase by CDT and boronic acid discs.The CDT with positive results determines the production of Metallo-β-Lactamase (MBL) while the positive boronic acid disc detects presence of class A carbapenemase as previously described [13].
CDT was performed by the use of two plates of Muller -Hinton agar plated with standard suspension of E. coli as previously described in DDST, then four antibiotics discs imipenem (10 μg), meropenem (10 μg) were added to each plate.Then over two antibiotics discs on one plate, 10μL of EDTA was added and on the other plate 10 μL of aminophenyl boronic acid (APBA) was added.The plates were incubated overnight at 37ºC for 24 hours.The increase of ≥5 mm in zone diameter around discs with the β-lactamase inhibitor (APBA or EDTA), as compared with the carbapenem discs alone, was considered to be a positive result.The followings strains were used as positive control strains; E. cloacae JMI10526 for blaIMP, Acinetobacter baumannii AB5 for blaVIM, K. pneumoniae ATCC strain BAA-1705 for bla KPC, K. pneumoniae ATCC strain BAA-2146 for blaNDM-1, and E. coli ATCC BAA-2523 for blaOXA-48), E. coli ATCC 25922 for blaCTX-M-and E. coli AF427133.1 for TEM and SHV beta lactamase (Grenole Cedex2, France).

Polymerase Chain Reaction (PCR) for ESBLs and Carbapenemase Genes
DNA extraction of E. coli was performed by boiling method in sterile distilled water.Briefly, colonies were used from nutrient agar, washed with 1 ml of 1X Tris-EDTA buffer (10 mM Tris-HCl, 1 mM EDTA [pH 8]) and centrifuged.The deposit was suspended in 0.5 ml sterile distilled water then was boiled in a water bath at 100ºC for 15 minutes.After, another 10 minutes of incubation at room temperature, centrifugation was performed at 11,500 X g for 5 minutes.The supernatant was decanted and the pellet was stored at -20ºC until use for further DNA analysis.The used primers for determination of ESBLs and carbapenemase genes are listed in Table 1 [8, 14 -16].Positive control strains for each genotype were used by previous strains described in the phenotypic methods.
Ready to use kit for amplification was supplied from Qiagen.The total reaction volume was 50 μL with 0.2 μM concentration of each specific primer and 2 μL extracted DNA for each target gene, PCR amplification is carried out in a 50 μL reaction volume.Amplification was performed by PCR (Biosystem).Initial denaturation step was performed at 95°C for 10 minutes followed by, 35 cycles composed of 45 seconds of denaturation at 94°C, 45 seconds of primer annealing at the specific temperature for each primer pair, then 50 seconds for extension at 72°C and final extension was performed at 72°C for 7 minutes.The amplification products were held at 4°C.Electrophoresis was performed for 30 minutes by the use of 1.5% agarose gel stained with ethidium bromide.The products were visualized by ultraviolet light and compared to ladder marker a 100 bp DNA size marker (Invitrogen, UK).

Data Analysis
Statistical analysis was done using SPSS software version 21 (IBM SPSS Statistics for Windows, IBM Corp., Armonk, NY, USA).Qualitative variables were presented as number and percent and chi-square was used for the comparison between different groups.Continuous variables were presented as mean ± SD (standard Deviation) for parametric data and Student ttest was used to compare two means.P value < 0.05 was considered statistically significant.et al., 2011 [8] bp: Base Pair

RESULTS
Data from 112 patients were collected; 88 patients were eligible to our study and 24 were excluded due to different causes as incomplete clinical data, improperly preserved samples or samples improperly taken.Samples were collected from admitted patients to neonatal ICU (39 patients with mean age is 12 days), PICU (21 patients with mean age 24 months) and surgical ICU (28 patients with mean age 3 months).Fortysix males (52.3%) and 42 females (47.7%) were included in the study.Escherichia coli strains resistant to ceftazidime (30 μg), cefotaxime (30 μg) were subjected to determination of ESBLs production E. coli and strains that were resistant to imipenem (10 μg) and/or to meropenem (10 μg), were subjected to determination of carbapenemase production and there was some cross-production as some of the isolates came from the same patient (Table 2).Phenotypic testing shows that out of the 88 E. coli isolates, 49 (55.68%) were positive for ESBL production and 30 (34.09%) were positive for carbapenemase production and 20 isolates were positive for both ESBL and carbapenemas.
Clinical characteristics associated with infection with ESBL and carbapenemase, are summarized in Table 3.There was significantly lower survival among patients infected with ESBL producing E. coli than those with ESBL non-producing E. coli and lower survival among patients infected with carbapenemase-producing E. coli than those with carbapenemase non-producing E. coli.Neither type of bloodstream infection, usage of CVC, urinary catheter nor endotracheal tube significantly differed between ESBL producing and nonproducing E. coli infected groups, also they didn't significantly differ between carbapenemase producing and non-producing E. coli infected groups.
The clinical symptoms of sepsis are shown in Table 4.There were non-significant differences between ESBL producing and non-producing E. coli and between carbapenemase producing and non-producing E. coli regarding clinical and laboratory signs of infection.Table 5 shows statistically nonsignificant correlations between infection with ESBL producing E. coli and infection with carbapenemase producing E. coli and the clinical data in the studied patients;    Antibiotic sensitivity pattern was shown in Table 7.All ESBL producing E. coli were resistant to ceftazidime.High resistance was detected to cefotaxime, cefoxitin, imipenem and meropenem.Sensitivity to amikacin, ciprofloxacin, gentamicin, Piperacillin/Tazobactam and Sulfamethoxazole/trimethoprim was 67.3%, 67.3%, 59.2%, 71.4% and 67.3%; respectively.Meanwhile, all carbapenemase producing E. coli were resistant to imipenem, meropenem and ceftazidime in addition, high resistance to cefotaxime, cefoxitin was observed.Sensitivity to amikacin, ciprofloxacin, gentamicin, Piperacillin/Tazobactam and Sulfamethoxazole/trimethoprim was 83.3%, 86.7%, 70%, 90% and 83.3% respectively.Antibiotic sensitivity pattern among individualized ESBL producing E. coli and CPE genes revealed that isolates with SHV gene were more sensitive to Piperacillin/Tazobactam and Cefepime and isolates with IMP gene was more sensitive to Piperacillin/Tazobactam and Ciprofloxacin.
A logistic regression analysis was employed to create a model that can predict 30-days mortality for patients with ESBL producing E. coli and CPE infections.We assessed the presence of risk factors for infection as insertion of central venous catheter, urinary catheter or endotracheal tube, patient age, primary blood stream infection, positive ESBL producing E. coli colonization and positive CPE colonization.The model can correctly predict the fatal outcome for 61.8% of the patients and recovery for 88.9% with a total success rate of 78.4%.Table 8 shows the logistic regression coefficient, Wald chi-square test which tests the unique contribution of each variable, the odds ratio and 95% confidence interval for odds ratio for each of the risk factors.Patient age, insertion of CVC,

Zaki et al.
positive colonization with ESBL producing E. coli and CPE had statistically significant partial effects.Positive colonization with ESBL producing E. coli was found to be the most statistically significant risk factor for fatal outcome in the studied group.

DISCUSSION
ESBL producing E. coli and CPE strains are worldwide leading pathogens in community and hospital acquired infections [17,18], that need continuous surveillance of antimicrobial resistance and sensitivity patterns.β-lactam anti- This study was conducted to detect the production of ESBL and carbapenemase among E. coli isolated from children with hospital acquired bloodstream infection, admitted to pediatric intensive care unit from January 2016 until January 2018.
Phenotypic testing of ESBL and carbapenemase production by the isolated E. coli showed that out of the 88 isolates; 49 patients (55.68%) were ESBL producers and 30 (34.09%) were carbapenemase producers.This finding is consistent with the high proportion of ESBL production reported in a study conducted in Egypt also, showing more than half (54.5%) of isolated E. coli positive for ESBL [20] and in New Zealand as about half of E. coli isolates produced ESBL [21].Similar results in the USA [22] and India [23] have been reported as prevalence rates of ESBL were from 35.0 to 42% for E. coli.Higher rates were even reported in other studies from India (>80%), China (>60%) [24].While in other researches, all the tested E. coli were ESBL-positive; in Saudi Arabia [25], Ethiopia [26] and Senegal [27].In this study, carbapenemase production among isolated E. coli was higher than those previously studied from clinical isolates in New Zealand (15%) [21].Meanwhile, in another study, carbapenemase production among E. coli was higher (45.2%) [27].These high rates of ESBL producing E. coli and CPE in Egypt are probably related to the overuse of antibiotics, improper administration of antimicrobial agents in trial therapies for the management of any febrile illness or due to lack of proper infection control practices (that can lead to an emergence of resistant organisms).Moreover, the trained attitudes to use thirdgeneration cephalosporins have been described as the most important precipitating factor in the emergence of ESBL and carbapenemase producing E. coli the presenting study, it was found that risk factors for sepsis as CVC, urinary catheter or endotracheal tube insertion were not significantly associated with ESBL or carbapenemase-producing E. coli indicating that usage of these invasive devices is not a reliable significant risk factor for infection with ESBL producing E. coli or CPE.It may be due to hospital rules in practicing strict infection control measures or due to the small number of patients included in this study.Although some authors reported that, bladder catheterization or other invasive medical devices are important risk factors for infection with ESBL-producing E. coli [29 -31], others reported urinary catheter insertion is not a risk factor for infection with ESBL producing E. coli [32].Furthermore, endotracheal tube insertion and mechanical ventilation are independent risk factors for carbapenemase-producing E. coli [33].
In the existing study, none of the clinical manifestations of sepsis statistically differed between ESBL producing and nonproducing E. coli infected groups and also didn't differ between carbapenemase producing and non-producing E. coli groups.Therefore, the physician, depending on the clinical examination alone, can't predict whether the patient is infected with ESBL or carbapenemase producing or nonproducing E. coli.This is concordant with another study that revealed non-significant differences in patients' clinical characteristics [34].However, in another study, there was a great difference concerning to disease severity and clinical signs between the ESBL-positive and ESBL-negative groups [35].
It was observed that a significant increase in mortality among ESBL producing E. coli infected patients versus noninfected.This is similar to other authors who explored a significantly high mortality rate in patients infected with ESBL producing E. coli [36,37].Also, in a survey that included 400 bloodstream isolated pathogens, more than 60% of neonates infected with ESBL producing E. coli died opposed to 35.7% who were infected with other isolates [38].
Widespread use of carbapenems has led to the appearance of CPE isolates [39].Gram-negative bacteria including E. coli develop carbapenems resistance by increasing the production of Carbapenemases and/or decrease in permeability of the outer membrane.KPC is Class A carbapenemase, while Metallobeta-lactamases (MBL) include VIM, IMP and NDM.Class D Carbapenemases are OXA like enzymes [40,41].
In this study, among the ESBL E. coli producers, the most frequently detected gene was SHV; it was in 30 isolates (61.22%); it was detected alone in 20 isolates (40.81%) and in combination with CTX-M in 10 isolates (20.41%).The next most frequent detected gene was TEM (38.78%) and then CTX-M comes (20.41%).In another study by Abdallah et al, TEM and SHV presented in about half, one fifth of isolates; respectively.In the same study, CTX-M type was the most common β-lactamase-encoding gene represent about 90% of the ESBL-producing E. coli [21] and this was higher than that in our study (20.41%).In contrast to this study, In a study conducted by Abdallah et al, the genetic analysis showed that CTX-M was present in 96% of ESBL E. coli in adult infections, also in a recent study the most common ESBLs genes detected were again CTX-M in multi-drug resistant gram negative bacilli in infected febrile neutropenic cancer patients [42].These discrepancies can be attributed to the differences in age categories in the first study and the different patients' clinical criteria in the second one.It was observed that TEM and SHV are important factors in decreasing the susceptibility of ESBL E. coli producers to third-generation cephalosporin [43].
Among 30 carbapemase producing E. coli, IMP was the most frequent gene; it was detected in 19 isolates (63.33%); 8 isolates (26.67%) had IMP only and 11 (36.67%) in combination with other CPE genes.OXA-48 and NDM were detected (20%) and 3 (10%) isolates; respectively.This was similar to a study by Kamel et al. [42], as one E. coli isolate harbored more than one type of metallo beta lactamase (VIM, NDM).The high prevalence of IMP gene in this study is probably due to excessive use of Imipenem in clinical therapy.A study conducted in Saudi Arabia, a near regional area to Egypt, showed different results as neither IMP, VIM or KPC was present in the carbapenemase-positive isolates.Meanwhile, the prevalence rates of OXA-48-like is 58.1% and of NDM-type is 41.9% and this was higher than in the present study [44].
In the present study, all ESBL and carbapenemase-producing E. coli were resistant to ceftazidime and all carbapenemase producers were resistant to Imipenem and Meropenem.High resistance was detected to cefotaxime, cefoxitin also (79.6% and 81.6% among ESBL producers respectively and 36.7% and 33.3% among carbapenemase producers; respectively).A study in Senegal showed that all strains were resistant to ceftriaxone or cefotaxime (99.9%) [27].In the present study low rates of resistance to amikacin was detected, which is consistent with findings from Korea [45], Taiwan [31], Japan [46] and Senegal [27].
In the present study, high resistance rates were observed among ESBL producers towards Imipenem and Meropenem; 81.6% and 83.7% respectively.This was comparable to resistance rates observed for Imipenem (60%), and Meropenem (30%) in Saudi Arabia [45].Meanwhile, all Carbapenemase producers in our study were resistant to Imipenem and Meropenem.Whereas, relatively high sensitivity of ESBL E. coli producers to sulfamethoxazole/trimethoprim complex (78.4%), and ciprofloxacin (75%) were observed in contrast to other authors that showed high resistance rates with sulfamethoxazole/trimethoprim complex (85.7%) and ciprofloxacin (72%) [27].Colistin, amikacin, fosfomycin, and temocillin are the only few remaining antibiotics used to treat infections caused by carbapenemase-producing Gram-negative bacilli.Proper combination therapy with two or more drugs is better than using monotherapy associated with a better survival rate [47,48].
In the current study, a logistic regression analysis was employed to create a model that can predict 30-days mortality for patients with ESBL producing E. coli and CPE infections.Risk factors for infection were assessed as the presence of invasive devices as CVC, urinary catheter or endotracheal tube, patient age, primary blood stream infection, positive ESBL producing E. coli colonization and positive CPE colonization.The model can correctly predict the fatal outcome for 61.8% of the patients and recovery for 88.9% with a total success rate of 78.4%.Positive colonization with ESBL producing E. coli was 6 times risk of mortality when compared to carbapenemase producing so should be cautiously and vigorously managed.According to our knowledge, this is the first study to do this analysis.A recent study by Komatsu and coauthors performed only univariate analysis, revealing multiple significant risk factors of death but they couldn't complete the multivariate analysis because there were only ten deaths [49].In agreement to these results, other studies observed that patients with infection due to ESBL producing E. coli tended to have poorer outcomes [49,50].

CONCLUSION
Neither risk factors for infection nor clinical manifestations can differentiate between ESBL and carbapenemase producing E. coli.SHV is the most frequently detected gene of ESBL producing E. coli and IMP the most frequently detected for carbapenemase production.Piperacillin/Tazobactam is the candidate drug to start within children with sepsis and suspected ESBL or carbapenemase-producing E. coli infection.Infection with ESBL producing E. coli was 6 times associated with mortality when compared to carbapenemase producing E. coli.Future studies on multicenter hospitals on a larger number of patients are recommended to validate the results.

ETHICS APPROVAL AND CONSENT TO PARTICIPATE
The protocol and methods employed were reviewed and approved by the Institutional Review Board, Faculty of medicine, Mansoura University, Mansoura, Egypt (code number: R/ 18.08.250).

HUMAN AND ANIMAL RIGHTS
No animals/humans were used for studies that are the basis of this research.

Table 2 . ESBL producing E. coli and Carbapenemase producing E. coli isolates in the studied patients. ESBL Producing E. coli Total P value
ESBL: Expanded Spectrum B-Lactamase Chi Square Test

Table 4 . Clinical signs and laboratory tests of sepsis in the studied groups.
CRT: Capillary Refill Time, ESBL: Extended Spectrum B-Lactamase, CRP: C-Reactive Protein *Chi-Square test

Table 8 . Logistic regression analysis for associated factors of 30-days mortality in the studied group.
Blood Stream Infection, CVC: Central Venous Catheter, ESBL: Expanded Spectrum B-Lactamase, ETT: Endo-tracheal Tube.B= logistic regression coefficient, P= P value calculated using Wald chi-Square test (Wald χ 2 ), OR= Odds Ratio, 95% CI = 95% Confidence Interval for odds ratio.biotics are the most common class of antibacterial agents used to treat bacterial infections due to their broad antibacterial spectrum and excellent safety profile [17].Carbapenems are the first-choice treatment for ESBL producing E. coli especially with the increase of ESBL clinical isolates reports expressing multidrug resistance [19].The development of CPE, is causing a threat to health, leaving few treatment options. BSI: