Introduction

Antimicrobial resistance (AMR) is one of the top ten global health threats to humans [1]. In 2019, AMR was directly responsible for 1.2 million deaths and was associated with an estimated 4.6 million deaths worldwide [2]. AMR, due to extended-spectrum beta-lactamase (ESBL)- producing bacteria, has escalated over the past years, both in hospitals and in communities [3]. Worldwide, the rapid spread of resistant pathogens has been mainly attributed to the overuse of antibiotics in human medicine and animal husbandry in sub-Saharan African (SSA) regions [4]. Currently, many Gram-negative bacteria can produce ESBL enzymes, conferring resistance to penicillins, first-, second-, and third-generation cephalosporins, and aztreonam (but not carbapenems or cephamycins) [3]. ESBL genes, which were first reported among Klebsiella spp. and Escherichia coli, are rapidly spreading among other bacteria through plasmid-mediated horizontal gene transfer [5]. In addition, ESBL-encoding plasmids can also code for other non-beta-lactam resistance genes, leading to multi-drug resistance [6]. In SSA, infections caused by ESBL-producing bacteria, including E. coli (ESBL-EC) and Klebsiella pneumoniae (ESBL-KP), are of great concern. Most E. coli and K. pneumoniae are causative agents of infections such as bacteraemia, urinary tract infections, and diarrhoea, particularly among children, both in hospital and community settings [7, 8].

Admittedly, E. coli and K. pneumoniae isolated from the stools of patients are not typically associated with diarrhoea except for diarrheagenic E. coli such as STEC and EHEC [9]. However, other gastrointestinal bacteria, for example, Salmonella enterica, are associated with severe diarrhoea. In case such pathogens are ESBL producers, they can cause difficult-to-treat infections leading to life-threatening complications. Recent studies have highlighted faecal carriage as a significant reservoir of ESBL-producing bacteria in hospitals and communities [10, 11]. The colonization of the intestinal tract with ESBL-carrying bacteria has been shown to precede infections [12], and hence stool carriage of ESBL-EC and ESBL-KP is of medical importance. In SSA, studies have reported carriage of ESBL-EC and ESBL-KP in the intestinal tract of humans and animals, highlighting that also individuals without gastrointestinal symptoms can be carriers of ESBL [13,14,15,16].

This study is of descriptive nature, aiming to determine the prevalence, antibiotic resistance, and gene variation of ESBL-EC and ESBL-KP in children with and without diarrhoea attending a rural hospital and child clinic in Agogo, Ghana.

Materials and methods

Study site and study population

A cross-sectional study was conducted to determine the frequency, antibiotic resistance, and gene variation of ESBL-EC and ESBL-KP in children with and without diarrhoea. The study was conducted at the Agogo Presbyterian Hospital (APH) and a selected Child Welfare Clinic (CWC) where parents and guardians take their children under five years of age for routine check-ups in the Agogo community in the Asante Akyem municipal district, Ashanti region of Ghana. Between June and December 2019, children below five years of age living in Agogo and nearby communities were recruited and categorized into either of the following two groups: (1) children with diarrhoea or episodes of diarrhoea within the last 72 h; and (2) children without diarrhoea and symptom free, attending CWC for routine immunization and growth monitoring with no history of diarrhoea for at least one month before study enrolment.

Ethical approval

The study was approved by the Committee on Human Research, Publications, and Ethics at the School of Medical Science, Kwame Nkrumah University of Science and Technology, Kumasi, Ghana, and by the German Medical Association (CHRPE/AP/593/17 and (CHRPE/AP/119/22). All participants were informed about the purpose of the study. Written informed consent was obtained from the parent or guardian of each child before study enrolment.

Sample collection

Stool samples were collected in sterile containers. In cases where stool samples were not readily available, a team member followed up with the parent or guardian to obtain a stool sample within 24 h. All stool samples were transported in a cool box at 2–8 °C to the laboratory of the Kumasi Center for Collaborative Research in Tropical Medicine (KCCR) for analysis within 4 h of taking the sample. To monitor the temperature, one pack of ice cool aid was put in a Va-Q-bagi (Bereg-Kit, Switzerland) with a thermometer.

Identification of E. coli and K. pneumoniae isolates and their antimicrobial susceptibility profiles

Stool samples were cultured on two MacConkey agar plates supplemented with 1 mg/L ceftazidime and 1 mg/L cefotaxime, respectively. Plates were incubated at 35–37°C for 18–24 hours in a normal atmosphere. Lactose-fermenting colonies (not more than three colonies), with typical morphology presumptive of E. coli and K. pneumoniae, were selected and sub-cultured on blood agar (Columbia Agar supplemented with 5% sheep blood) for isolation of pure colonies. The VITEK 2 Compact system, using Gram-negative bacteria identification (GN ID) cards and antibiotic susceptibility testing (AST) N214 cards, was used for identification and antimicrobial susceptibility profiling of the bacterial isolates (BioMerièux, Marcy L’ É toile, France). Tested antibiotics included penicillin (ampicillin, piperacillin/tazobactam, and ampicillin-sulbactam), carbapenems (meropenem, ertapenem, and imipenem), fluoroquinolones (ciprofloxacin), tetracyclines (tetracycline), aminoglycosides (gentamicin), and trimethoprim/sulfamethoxazole. Results were interpreted according to the guidelines of the European Committee on Antimicrobial Susceptibility Testing (EUCAST guidelines, version 10.0, 2020; http://www.eucast.org/clinical_breakpoints/).

ESBL-producing bacteria were further confirmed using the combined double-disk synergy test with cefotaxime and ceftazidime alone or in combination with clavulanic acid (Becton, Dickinson and Company, Sparks, MD, USA) as described by the EUCAST guidelines version 9.0 (2019). In this study, ESBL-producing isolates of E. coli and K. pneumoniae were considered MDR if they were resistant to at least three classes of antibiotics. Quality control of each batch of the MacConkey agar containing 1 mg/L ceftazidime and 1 mg/L cefotaxime was performed using E. coli ATCC 25922 and a blaCTX-M positive E. coli.

Polymerase chain reaction (PCR) and sequence detection of ESBL genotypes

For all confirmed ESBL-producing bacteria, a 10µL loopful of overnight pure colonies were transferred into saline. The solution was briefly vortexed, and the supernatant was discarded. The pellet was treated with 100 µL TE buffer (10:1) and heated at 95 °C for 5–10 min. The mixture was then centrifuged for 2 min. The supernatant containing the DNA was used for PCR analysis and sequencing. Molecular characterization of ESBL genes were performed by PCR for the presence of blaCTX-M (cefotaximase-Munich), blaTEM (Temoneira), and blaSHV (sulfhydryl variable enzyme) as described elsewhere [17, 18]. To differentiate blaCTX-M genes, previously designed specific target primers (Table 1) were used for PCR amplification [17]. The PCR product was then sent to Microsynth-Seqlab in Göttingen/Germany for sequencing. The resulting sequences were aligned and blasted using the Resfinder 4.1 software (https://cge.cbs.dtu.dk/services/ResFinder).

Table 1 Primers used for ESBL gene amplification by PCR
Full size table

Statistical analysis

Sociodemographic, clinical, and microbiological data collected from children enrolled in the study were entered and cleaned in Microsoft Excel. Statistical analyses were performed using the R language for statistical computing, version 4.0.2. All study variables were categorical and presented as frequencies with percentages. The Chi-squared test was used to compare count data against the null hypotheses. In cases where a cell had 5 values or less, Fisher’s exact test was applied.

Results

Characteristics of study participants

A total of 435 children were enrolled in the study. Of these, 47.1% (n/N = 205/435) presented with diarrhoea to the study hospital, and 52.8% (n/N = 230/435) were recruited among children without diarrhoea (Table 2). The number of female recruits (47.8%, n/N = 208/435) was slightly lower than that of male recruits (52.2%, n/N = 227/435). Diarrhoea samples were more common among males (58.5%, n/N = 120/205) than females (41.5%, n/N 85/205).

Table 2 Demographic characteristics of the study groups
Full size table

Table 2 shows the relationship between age, sex, and ESBL-producing isolates. ESBL-positive isolates were more common in children aged 49–60 months (55%, n/N = 22/40). The frequency was lowest among children between 0 and 12 months old. Age was not significantly associated with overall ESBL (p = 0.74), ESBL-EC (p = 0.52) or and ESBL-KP (p = 0.74) detection and ESBL-EC and ESBL-KP positivity were similarly distributed between male and female gender (p = 0.48).

Out of 435 children, 178 (40.9%) carried ESBL-producing isolates (ESBL-EC and ESBL-KP) (Table 3), and no difference in ESBL carriage between children with and without diarrhoea was observed (p = 0.75). In total, 187 ESBL-producing isolates (168 ESBL-EC and 19 ESBL-KP) were identified from 178 ESBL-positive children. Nine of the children had both ESBL-producing E. coli and K. pneumoniae (2.05%, n/N = 9/435). It was observed that almost all children without diarrhoea who were positive for ESBL bacteria (98.9%, n/N = 91/92) had ESBL-EC, whereas ESBL-EC was present in 89.5% (n/N = 77/86) of children with diarrhoea who were positive for ESBL bacteria. Almost all the ESBL-KP isolates (n = 13) were observed in children that had diarrhoea (86.6%, n/N = 13/15).

Table 3 ESBL-EC and ESBL-KP positive children in diarrhoea and diarrhoea group
Full size table

Genotypic identification of ESBL genes among isolates

Figure 1 shows genotypic characterization identified three different beta-lactamase genes (blaCTX-M, blaSHV, and blaTEM) among 187 ESBL-EC and ESBL-KP isolates from 178 children. The majority were blaCTX-M positive (95.2%, n/N = 178/187), while a few carried blaCTX-M/blaTEM (1.6%, n/N = 3/187), blaSHV (1.1%, n/N = 2/187), or blaTEM (1.1%, n/N = 2/187). In two of the phenotypically confirmed ESBL isolates, none of the three genes (blaCTX-M, blaSHV, and blaTEM) were identified.

Fig. 1
figure 1

Distribution and frequency of different β-lactamase genes among the ESBL isolates

Full size image

Among isolates carrying the prevalent ESBL gene blaCTX-M, the majority of them were of group blaCTX-M-1 (96.1%, n/N = 171/178) (Fig. 1) whilst only seven were of group blaCTX-M-9 (3.9%, n/N = 7/178). Sequencing of the PCR product revealed blaCTX-M-15 (92.4%, n/N = 158/171) as the most common type of the blaCTX-M-1 group, and the remaining types were blaCTX-M-3 (4.1%, n/N = 7/171) and blaCTX-M-28 (3.5%, n/N = 6/171). Among the genes in group blaCTX-M-9, more than half were blaCTX-M-27 (57.1%, n/N = 4/7) while the remaining were blaCTX-M-14 (28.6%, n/N = 2/7) and blaCTX-M-14b (14.3%, n/N = 1/7).

In terms of distribution of ESBL genes identified, blaCTX-M genes (n = 178) were equally distributed between the diarrhoea (50%, n/N = 89/178) and non-diarrhoea (50%, n/N = 89/178) stool types. blaTEM (n = 2) and blaSHV (n = 2) were rare and isolated from non-diarrhoea and diarrhoea stools, respectively. Among the blaCTX-M genes, blaCTX-M-1 was also distributed uniformly between the diarrhoea (52.0%, n/N = 89/171) and non-diarrhoea (48.0%, n/N = 82/171) samples whilst type blaCTX-M-9 was all isolated from non-diarrhoea stool samples (100.0%, n/N = 7/7).

In terms of the distribution of beta-lactamase genes in ESBL-EC and ESBL-KP, blaCTX-M-15 were found to be the most common type in both bacterial species, with 85% (n/N = 143/168) and 79% (n/N = 15/19) abundance in ESBL-EC and ESBL-KP, respectively. The genes blaCTX-M-3, blaCTX-M-28, and blaSHV-12 was found in both organisms, whereas blaCTX-M-27, blaCTX-M-14 and 14b, and blaTEM were only found in ESBL-EC.

Antimicrobial resistance of ESBL-EC and ESBL-KP isolates

ESBL-EC and ESBL-KP of this study showed resistance to antibiotics commonly used in Ghana (Fig. 2a and b). All isolates showed resistance to ampicillin. However, both ESBL-EC and ESBL-KP isolates were susceptible to meropenem and imipenem. Both ESBL-EC and ESBL-KP isolates showed the most resistance against tetracycline, with 85.1% (n/N = 143/168) and 73.6% (n/N = 14/19), followed by sulfamethoxazole-trimethoprim, against which the isolates showed 73.8% (n/N = 124/168) and 84.2% (n/N = 16/19) resistance, respectively. ESBL-EC isolates showed higher resistance against ciprofloxacin (31.5%, n/N = 53/168) compared to ESBL-KP isolates, which showed only 5.2% (n/N = 1/19) resistance. Whereas, ESBL-KP showed higher resistance towards ampicillin-sulbactam (84.2%, n/N = 16/19), gentamicin (63%, n/N = 12/19), and piperacillin-tazobactam (36.8%, n/N = 7/19) compared to ESBL-KP isolates, which showed only 34% (n/N = 57/168), 14.2% (n/N = 24/168), and 9.5% (n/N = 16/168) resistance, respectively. Notably, MDR was observed in 74% (n/N = 138/187) of the total isolates, with 41.7% (n/N = 78/187) being resistant to three classes of antibiotics, 22.4% (n/N = 42/187) being resistant to four classes of antibiotics, and 9.6% (n/N = 18/187) being resistant to five classes of antibiotics. MDR isolates were slightly higher in ESBL-KP (84.2%, n/N = 16/19) compared to ESBL-EC (78.5%, n/N = 132/168).

Fig. 2
figure 2

Antimicrobial resistance among ESBL-EC and ESBL-KP isolates from diarrhoea and non-diarrhoea stool samples of children. Abbreviations: Ampicillin (AMP), Ciprofloxacin (CIP), Ertapenem (ETP), Gentamicin (GM), Imipenem (IPM), Meropenem (MEM), Sulfamethoxazole-Trimethoprim (SXT), Tetracycline (TE), Piperacillin-Tazobactam (TZP), Ampicillin-Sulbactam (SAM). aMultidrug resistance I defined as resistance to three antibiotics. bMultidrug resistance II defined as resistance to four antibiotics. cMultidrug resistance III defined as resistance to five antibiotics

Full size image

Discussion

Carriage of ESBL-producing Klebsiella pneumoniae and Escherichia coli is one of the drivers of nosocomial and community infections, globally [7, 8]. Notwithstanding, it has received little attention in Ghana, particularly infections in children. This study explored ESBL-EC and ESBL-KP in children up to five years of age in Agogo, Ghana.

Our data shows that both children with and without diarrhoea are ESBL carriers, serving as a potential transmission reservoir. The overall prevalence of ESBL producers among the study population was 40.9% (Table 2). Out of which, 37.5% were ESBL-EC and only 3.4% were ESBL-KP. This finding is almost similar to the frequencies of ESBL producers among the adults from Ghana in the previous two studies, which reported frequencies of 41.5% and 49.3%, respectively [19, 20]. Interestingly, a previous study [14] carried out on young children in the same hospital as this study reported a much higher prevalence of ESBL-EC (61%) compared to the findings in this study. However, Falgenhauer et al. investigated hospitalized children in Agogo, who likely were more exposed to risk factors for carriage of ESBL, such as nosocomial carriage, antibiotics use, and longer hospitalization [21]. No age association with carriage of ESBL in children was found in our study (Table 2); however, a study in Tanzania reported high ESBL carriage in infants (0–3 months) compared to the older groups [15]. It could be due to mother-to-child transmission of the ESBL-producing isolates [22] but it may also reflect on antibiotic usage among children, especially for the management of diarrhoea [23].

Genomic characterization of the isolates from this study revealed blaCTX-M-15 as the most prevalent gene type in both diarrhoea and non-diarrhoea stool samples (Table 3), which is consistent with other studies [14, 15, 24] and subsequently confirms the current global state of the blaCTX-M-15 in community and hospital settings [25, 26]. This could be a consequence of the dissemination of group blaCTX-M-1, specifically blaCTX-M-15, which happens as a result of the mobilization of genetic platforms such as plasmids and transposons. It has been identified as the most dominant ESBL enzyme and is found in clinical isolates, community isolates, environmental isolates, and farm animals [14]. This is worrisome as the gut is the main reservoir for many Enterobacterales bacteria.

Remarkably, we found blaCTX-M-28 in both diarrhoea and non-diarrhoea stool samples, and to the best of our knowledge, this is the first time blaCTX-M-28 has been reported in Ghana. Unlike blaCTX-M-15, which has been frequently documented, blaCTX-M-28 has been sparingly described in SSA [27]. However, this gene has been reported in other parts of the world, such as Tunisia and Bosnia [28, 29]. blaCTX-M-28 may be underreported due to its close sequence similarity to blaCTX-M-15, which differs in only one nucleotide position [27]. The identification of blaCTX-M-28 in this study suggests either transmission of this gene from another country through movement of people [10] or a homoplastic mutation of that gene in an ancestor cell, leading to an identical sequence of blaCTX-M-28. We also identified blaCTX-M-27 in four of the ESBL-producing E. coli found in non-diarrhoea stool samples, indicating the diverse circulation of the blaCTX-M group in Agogo. The presence of other blaCTX-M groups besides blaCTX-M-15 emphasizes the dynamics and ongoing spread of ESBL genes in Agogo.

Due to the dissemination of these resistance genes, most developing countries are using broad-spectrum antibiotics as the clinical algorithm for managing patients with infection [30]. According to the latest guidelines from the Infectious Diseases Society of America, carbapenems, fluoroquinolones, and cotrimoxazole are recommended for managing most ESBL infections. Similarly, in Ghana, these are some of the common antibiotics used for treating ESBL infections; however, high resistance against fluoroquinolones and cotrimoxazole has already been reported [31, 32]. The antibiotic resistance profiles from our study also show high resistance to these antibiotics (Fig. 2). Among them, the most ineffective antibiotics were cotrimoxazole and tetracycline. This is of no surprise, as these two antibiotics have been used in Ghana for many years, and resistance against them has been reported in a wide range of bacteria [14, 33, 34]. The higher rates of resistance of ESBL-producers to cotrimoxazole, tetracycline, and ciprofloxacin are in accordance with other studies [12, 35], which could be because genes for resistance to these antibiotics are present on the same plasmid [12].

In contrast, all ESBL-EC and ESBL-KP were susceptible to the two carbapenems, meropenem and imipenem, which agrees with earlier studies in Nigeria and Tanzania [15, 36]. This might be due to the low or no usage of these antibiotics in Ghana [37].

We also observed a high prevalence of MDR (> 75%) for both ESBL-EC and ESBL-KP isolates in our study (Fig. 2), which is not surprising as ESBL-producing bacteria are frequently associated with co-resistance to other antimicrobial agents [14, 38]. This high rate of MDR could be a result of the extensive use of antibiotics in human medicine [39], in veterinary medicine [40] as well as other factors such as HGT and mobile genetic elements. Therefore, the ready availability of antibiotics in the local markets must be monitored. Most importantly, monitoring the resistance pattern exhibited by the MDR isolates and its impact on patient management is essential, especially in children.

In this study, sample size was not calculated prior to the study, and hence most of the calculations are exploratory and of a descriptive nature. Due to the low number of samples, e.g., for ESBL-KP, the results have to be interpreted with caution. Risk factors for the acquisition of the carriage of ESBL-producing E. coli and K. pneumoniae were not assessed. Data on antibiotic usage in children was not assessed in this study. Also, a test to ascertain whether the genes found were plasmid- or chromosomal-associated was not done. This is important in order to understand the impact of the resistance genes on the spread of AMR. Plasmid genes, for example, are mobile and easy to transfer to other bacteria and bacterial species.

Conclusions

This study highlights the high frequency of stool carriage of ESBL-EC and ESBL-KP among children with or without diarrhoea in the Agogo community. Our study highlights the importance of this population as a possible reservoir and suggests that it may pose a risk for the transmission of drug resistance throughout the wider community. The study also found blaCTX-M-15 as the most prevalent gene in both diarrhoea and non-diarrhoea stools of children. It also reports for the first time the presence of blaCTX-M-28 in Ghana. The high frequency of ESBL genes found in this study is alarming, considering the limited diagnostic and treatment options that are available in resource-poor countries such as Ghana. The steady and continued increase of ESBL-producing bacteria and the associated antibiotic resistance has the potential to further select for even more resistant pathogens, such as carbapenem-resistant bacteria. We recommend the need for routine screening of ESBL-producing pathogens to optimize the use of antibiotics, and we encourage additional studies to evaluate these emerging genes and their risk factors in Ghana.