Occurrence and distribution of extended-spectrum β-lactamase in clinical Escherichia coli isolates at Ho Teaching Hospital in Ghana

Objective This study determined the occurrence and distribution of Extended Spectrum β-Lactamase (ESBL) genotypes of E. coli isolates in Ho Teaching Hospital, Ghana. Design A cross-sectional study. Setting A single centre study was conducted at Ho Teaching Hospital of Ghana. Participants Patients who visited Ho Teaching Hospital Laboratory with the request for culture and susceptibility testing. Main outcome measure Escherichia coli were isolated, and Extended-Spectrum β-Lactamase genes were detected. Results Of the 135 isolates, 56(41.5%,95% CI: 33.1% – 50.3%) were ESBL producers. More males, 14(58.3%), produced ESBL than females, 42(37.8%). The ESBL prevalence was highest among the elderly who were 80 years and above 3(100.0%), with the least prevalence among patients within 50–59 years and 0–9 years age bracket, representing 4(25.0%) and 3(27.3%), respectively. The total prevalence of ESBL was marginally higher among out-patients (41.8% 95% CI: 31.9% – 52.2%) compared to in-patients [40.5% 95% CI: 24.8% – 57.9]. BlaTEM-1 was the predominant ESBL genotype obtained from 83.9% (47/56) of the confirmed ESBL producing isolates, with the least being TOHO-1 4(7.1%). The co-existence of 2 different ESBL genes occurred in 19(33.9%) of the isolates. The single and quadruple carriage were 16(28.6%) and 3(5.4%), respectively. The highest co-existence of the ESBL genotypes was recorded for blaTEM-1 and blaCTXM-1 15(26.8%), followed by blaTEM-1, blaCTXM-1 and blaSHV-73 [12(21.4%)]. Conclusion The high prevalence of ESBL-producing E. coli isolates with multiple resistant gene carriage is a threat to healthcare in the study area. Funding This research received no external funding.


INTRODUCTION
Extended-Spectrum β-Lactamases (ESBLs) are enzymes that confer resistance to β-lactam antibiotics, such as cephalosporins, aztreonam, penicillin and monobactams. The enzymes were described and first detected in Klebsiella and later in Escherichia coli and other Gramnegative rods. 1,2 There are different ESBL enzyme types, but the commonest one includes Temoneira (TEM β-lactamase), Cefotaximase (CTX-M β-lactamase) and Sulfhydryl variable (SHV β-lactamase). 3 These ESBLs may have emerged due to the previous usage of 3rd and 4th generation cephalosporins, recent hospitalisation, re-cent use of antibiotics, patients over 60 years of age, diabetes, among others. 4,5 The enzymes have substrate specificity and an affinity to the cephalosporin types introduced into hospital use. These enzymes are plasmidborne and are spread from one organism to another by conjugation. The enzymes also spread from person to person and in hospital wards from where they spread to communities. This ESBL producing E. coli is a threat to public health as it has been associated with increasing morbidity and mortality. 6 In a study in Indonesia, Sianipar et al. 7 reported a mortality rate of 30.6% in patients infected with ESBL-producing bacteria.
In the United Kingdom, the mortality rate from bacteraemia infection caused by ESBL-producing E. coli was 60.9% as against 27.3% from those that did not produce ESBL. 8 In their systematic review and meta-analysis in Ethiopia, Tufa et al. 9 reported a death rate of 33.3% and 100.0% due to infections caused by ESBL-producing gram-negative isolates in two separate studies. Infection with ESBL-producing organisms is severe in developing countries like Ghana. There is a scarcity of information on the subject, coupled with misuse of drugs and high antibiotic resistance rates. 10 Some studies have reported the prevalence of ESBL producing E. coli, especially among patients with urinary tract and bloodstream infections. A study in China undertaken by Quan et al. 11 reported an ESBL prevalence of 55.5% in E. coli isolated from patients with communityonset bloodstream infections. In their study at Bhubaneswar in India, Singh and other co-researchers 12 also reported 61.1% ESBL prevalence in E. coli isolates among patients admitted to the intensive care unit of a tertiary hospital. Varying levels of ESBL production were recorded in the Potohar region of Pakistan (40.0%) 13 Wireko et al. 23 The emergence and dissemination of ESBLs in E. coli isolates in health care settings are of great concern, as they lead to treatment failures with cephalosporins. 6,10 Generally, there are limited comprehensive data on ESBL-producing Enterobacteriaceae in African countries since most health facilities cannot test for the presence of these enzymes. This study aimed at determining E. coli ESBL occurrence and genotypes circulating in Ho Teaching Hospital (HTH) of Ghana.

Study area and design
A cross-sectional study design was used for the study. It was done at the HTH and the School of Basic and Biomedical Sciences research laboratory, the University of Health and Allied Sciences, Ho, over 24 months (20 th April 2018-19 th April 2020). The HTH is a 241-bed capacity tertiary medical facility located in the capital of the Volta Region, Ho.
The microbiology department of the hospital receives a request for various microbiological analyses from various units and departments of the hospital, including out patient's department, male ward, female ward, paediatric ward, accident and emergency unit, dental unit, dialysis unit, and ear, nose and throat units, among others.
The bacterial isolates E. coli isolates were cultured from various clinical specimens, including urine, high vaginal swabs and blood on MacConkey agar and blood agar. Growths suspected of E. coli were confirmed using Gram stain reaction, triple sugar fermentation test, citrate test, urease test, indole test, Voges Proskauer and methyl red test. Organisms isolated and confirmed as E. coli were inoculated into 80% glycerol-Mueller Hinton broth, stored in a -80 o C freezer, and later used for other tests. Control organisms were Escherichia coli (ATCC 25922) and Klebsiella pneumoniae (NCTC 13442).

Phenotypic detection of ESBL Screening of E. coli isolates for ESBL production
Screening E. coli for ESBL production was done by subjecting the isolates to cefpodoxime (10µg), ceftazidime (30µg) and ceftriaxone (30µg), according to the procedures described by CLSI 2018. 24 After removing the stored isolates from the freezer, the surface was aseptically scraped and subcultured onto Nutrient agar. After overnight incubation, the E. coli isolates from Nutrient agar were emulsified in buffered phosphate saline. The turbidity was measured with a DensiCHEK plus densitometer (Biomerieux, U.S.) to obtain 0.5 McFarland standards. The isolates exhibited a zone of inhibition of ≤17mm, ≤22mm or ≤25mm to at least one of the three antimicrobial agents; cefpodoxime, ceftazidime or ceftriaxone, respectively, as screen positive for ESBL production. The positive screen isolates were subjected to ESBL confirmation by the combination disc method.

Confirmation of ESBL production by combination disc method
Screen positive E. coli isolates were confirmed for ESBL production using the combination disc (Oxoid, UK) method. After removing the stored isolates1 from the freezer, the surface was aseptically scraped and subcultured onto Nutrient agar. After overnight incubation, E. coli isolates from Nutrient agar were emulsified in buffered phosphate saline, and the turbidity was adjusted to 0.5 McFarland standards. The culture was used to seed the Muller-Hinton agar surface evenly. The antibiotic discs were then applied. After overnight incubation, the zones of inhibition around the cefpodoxime disc alone and that of cefpodoxime plus clavulanic acid were measured and recorded. ESBL production was detected when the difference between the zone of inhibition around the cefpodoxime alone and around cefpodoxime/clavulanic acid was 5mm or greater and interpreted according to the procedure of Clinical and Laboratory Standards Institute (CLSI). 24

Molecular detection of ESBL DNA Extraction
An overnight culture of the isolates in 30ml Luria Bertani broth (Oxoid, UK) in a shaking incubator was used for the genomic DNA extraction using high molecular weight phenol-chloroform extraction method 25 , except that Tris EDTA (TE) was used as the elution buffer. The harvested DNA was incubated at 4 o C for two days to resuspend the pellet into a translucent viscous gel. The concentration of the extracted DNA was measured using a Nanodrop spectrophotometer (Thermo Scientific), and the viscous DNA was stored under -24 o C.

Detection of ESBL genes by Polymerase Chain Reaction
E. coli ESBL genotypes were determined using Polymerase Chain Reaction (PCR). ESBL genotypes tested for include BlaTEM-1, BblaCTXM-1, BlaCTXM-825, BlaCTXM-914, BlaSHV-73 and TOHO-1. The primer sequence and their annealing temperatures are listed in Table 1. The PCR conditions were: initial denaturation at 94 o C for 30 seconds, actual denaturation at 94 o C for 30 seconds, 45 seconds of annealing at 62 o C for BlaCTXM-914, and 55 o C for BlaCTXM-1, BlaCTXM-825, BlaTEM-1, BlaSHV-73 and TOHO-1. Initial and final extensions were done at 68 o C for 60 seconds per kilobase and 5 minutes, respectively. PCR reaction volume was 12.5µl for each primer pair and the components of the reaction were 6.25µl of one taq quick-load 2X master mix with standard buffer, 0.25µl each of forward and reverse primers and 4.75µl of nuclease-free water. Thermocycling was done for 30 cycles.

Loading of amplicons
Using a micropipette, the resulting PCR product was loaded into agarose wells. The first and second lanes were loaded with 6µl of 100bp ladder and 10µl of a known DNA positive for CTXM, SHV, TEM or TOHO. The last lane was loaded with a known negative sample. The remaining lanes were loaded with samples under investigation.

Visualisation of amplicons
After the electrophoresis, the resulting gel was visualised using an image analysis system (UVITEC Cambridge, United Kingdom). The images were photographed with a digital camera (Kodak, Japan), as shown in Figure 1, to detect the BlaCTXM-1 gene.   Less than half (<50%) of the isolates obtained from blood, ear swab, high vaginal swab and urine were confirmed to be carrying ESBL genes. Details of these results are presented in Table 4.
Generally, there was an increasing undulating trend of ESBL across the increasing age categories for males from the least burden of 0.0% among the age group 0-9 years through to 100.0% among those aged 80 years and above. A similar pattern of rising and fall ESBL-positivity was observed across the female age groups with a trough at age 50 -59 years (8.3%) and crest at age group 80 years and above (Figure 2).   Figure 3). . Figure 3 Occurrence of ESBL genotypes in E. coli isolates in Ho Teaching Hospital.
The highest burden of ESBL genotypes was found among those between 30 and 39 years, and 40 and 49 years except BlaCTXM-914, which was not detected in the isolates obtained from these age brackets. The lowest frequency of the ESBL genes was obtained from patients between 10 and 19 years. Except for the ESBL genotype of the BlaCTXM-1 gene, where proportions were higher among the females than the males, the male gender had higher percentages of all the other five ESBL genes in this study. Similarly, inpatients had a greater percentage of the ESBL genotypes except for BlaCTXM-825, even though the frequency of the carriage was higher in the out-patients. The total ESBL prevalence was, however, higher in out-patients 41(41.8%) than among the in-patients 15(40.5%) ( Table 5).

0.00=not detected
The ESBL genes BlaSHV-73, BlaCTXM-825, BlaCTXM-914, and TOHO-1 were predominantly seen in non-urine samples than in urine samples. All the 6 ESBL genotypes under investigation were found in the isolates obtained from wound and urine. BlaCTX-M-1 was the only genotype carried by the isolates obtained from pleural aspirate ( Table 6).  It was found that 16 of the ESBL producing isolates had one ESBL encoding gene. However, most isolates (19) carried two different ESBL genes, with 3 isolates having four different ESBL genes. Fifteen of the 56 E. coli isolates carried BlaTEM-1 and BlaCTXM-1 genes. Details of these results are presented in Table 7.

DISCUSSION
Extended-spectrum β-lactamase mediated antimicrobial resistance by E. coli isolates has been a global problem. The proportion of E. coli developing resistance differs from country to country and even differs within localities in the same country.
The ESBL situation has not been studied at the Ho municipality, but high resistant proportions of E. coli were reported from other clinical laboratories. There is, therefore, the need to monitor this occurrence and understand the spread and distribution of ESBL mediated resistance genes in E. coli.
In the current study, we recorded an ESBL prevalence of 41.5%. The prevalence reported in this study was lower than those reported in other jurisdictions. A study by Mahamat et al. 28 32 Other studies recorded low prevalence compared to the finding in this study. 33,34 The burden and distribution of ESBL are dependent on heavy antibiotic use, geographical locations and hospital to hospital variations 35,36 , and transmission from community, livestock, animals, and environment 37 and weak infection control measures instituted by various hospitals. 38 Other plausible reasons accounting for this variation in prevalence, according to Abrar et al. 39 , are the availability of the antibiotics and socio-economic status of the society. This variation in the prevalence may also be due to differences in the type of sample used for analysis and the time of sample collection. 36 The finding of this study, however, is consistent with the 40.8% prevalence reported by Seyedjavadi et al. 36 in their study to determine the relationship among blaTEM, blaSHV and blaCTX-M genotypes in patients suffering from urinary tract infections. The high ESBL prevalence may lead to treatment failure and narrowing the antibiotic spectrum as the enzyme destroys cephalosporins, resulting in therapeutic difficulties. The difficulty in detecting ESBLs and the inconsistencies in reporting ESBLs results have led to a surge in resistance patterns. To curb the spread of antibiotic resistance, there is the need to carry out regular antibiotic usage and resistance surveillance to track the changes in antibiotic susceptibility and formulate appropriate control strategies.
Our study investigated a variety of β-lactamases among the 56 E. coli that were phenotypically positive using a combination disc method. These genotypes were  36,42 . This observation, however, is a threat to public health and will negatively affect treatment outcomes in our healthcare settings. The more the resistance genes on one isolate, the more resistant the organism will be against different antimicrobials that will be used to treat the infection caused by the organism. The carriage of multiple antibiotic resistant genes by an organism has been reported due to the misuse of third-generation cephalosporins. 36 In this study, patients aged 80 years or more were mostly infected with the ESBL-producing E. coli, followed by those within 40-49 years age bracket (52.6%), with the least being 25% for those within 50 and 59 years. No significant relationship existed between the rate of E. coli infection and the age of the patients. The outcome of our study is at variance with a report by Seyedjavadi et al. 36 , where ageing is a risk factor associated with beta-lactamase mediated resistance in patients infected with E. coli. Recovery of ESBL-producing E. coli isolates in the aged could be due to the compromised immunity associated with ageing.
Even though not statistically significant, the male gender (58.3%) was mostly infected with the ESBL producing E. coli isolates compared to the female counterparts (37.8%). Contrary to our findings, Al-Garni et al. 43 reported female preponderance in their study to assess the risk factors and molecular features of extended-spectrum β-lactamase producing bacteria at the southwest of Saudi Arabia, even though the female majority was also not statistically significant. The proportion of women and men with ESBL infections indicates that gender is a non-significant factor for the acquisition of infection. 43 Despite the findings in this study that high frequency of ESBL producing isolates were obtained from urine samples (38/98) compared to other samples, the prevalence in pleural aspirate was 100.0%. Other studies recorded the most prevalence in blood samples 38 and sputum. 39 The high prevalence of ESBL producing isolates from blood and sputum was contrary to our finding in this study. Most of the ESBL producing E. coli isolates were obtained from pleural aspirate and wound swab. Most E. coli isolates are commensals in the gut. They can become pathogenic in immunocompromised hosts, causing intestinal and extraintestinal diseases in humans and animals. 44

Limitation of the study
The authors acknowledged some limitations in the study. We did not have access to patients' presenting complaints or their initial differential diagnosis. Patients long term history of antibiotic usage were not also known. The study did not investigate the presence of other ESBL genotypes that might be present.

CONCLUSION
This occurrence of ESBL-producing E. coli isolates in the various samples was high, with blaTEM-1 being the predominant ESBL genotype among the six genotypes identified. Multiple carriage of ESBL genotypes was observed and is of great cause of concern as it may lead to limited therapeutic options. Since antibiotic resistance pattern varies from one region to another, there is also the need to monitor the resistance pattern in other parts of the country.