Introduction
The emergence and spread of multidrug-resistant (MDR) bacteria are serious global public health threat. In 2019, an estimated number of 4.95 million deaths were associated with antimicrobial resistance (AMR), with 1.27 million directly attributable to MDR bacteria [
1]. The highest burden is in Western sub-Saharan Africa, with 27.3/100 000 AMR-attributable and 114.8/100 000 AMR-associated deaths [
1], the same region where AMR surveillance is minimal and resistance data limited. The health care costs of AMR reach nearly
$1.2 Trillion in the “High-AMR” Case [
2]. According to the World Bank estimation/projection, by 2030, there will be an annual increase of
$0.22 trillion of the extra health expenditure incurring in the low-AMR settings [
2].
The emergence of carbapenem resistance substantially limits the therapeutic options at hospitals worldwide since carbapenems belong to our last resort antibiotics [
3,
4]. Regrettably, carbapenem resistance among Enterobacterales has experienced a dramatic increase and global spread over the past decade, with reports both from hospital and community settings [
4]. Infections due to carbapenem resistant Enterobacterales (CRE) are associated with increased mortality rates [
5,
6,
7]. Management of patients with CRE infections necessitates the use of combination therapies, which typically involve various types of antibiotics, such as tigecycline and colistin [
8,
9]. In low- and middle-income countries (LMICs) resistance data are scare due to the challenges in detecting carbapenemases in microbiology laboratories [
10,
11].
Two primary mechanism account for carbapenem resistance among bacteria: the first mechanism involves a reduction of membrane permeability by porin loss often associated with production of ESBL or AmpC-β-lactamase. The second mechanism entails the production of carbapenemase enzymes capable of hydrolyzing carbapenems [
12]. However, although carbapenemase and AmpC-β-lactamase production are not routinely identified in clinical microbiology laboratories in LMICs, several recent studies across African hospitals have reported carbapenemase production among Enterobacterales [
13,
14,
15,
16,
17,
18,
19]. Carbapenem resistance at rates of approximately 45% have been reported among
E. coli and
Klebsiella pneumoniae isolates from clinical specimens in two tertiary hospitals in Nigeria [
14]. A phenotypic resistance rate of 23.3% and genotypic resistance rate of 43.1% to carbapenem were reported in Uganda [
16]. Data on AmpC-β-lactamase are accumulating gradually also from Africa. In a recent study, a high rate of AmpC-β-lactamase was reported in five referral hospitals in Khartoum, with highest rates (49.3%) among
Acinetobacter baumannii [
20].
In Burkina Faso, data on prevalence of CRE or AmpC-β-lactamase production among clinical isolates are limited. One study conducted in a referral teaching hospital, identified 17 carbapenemase-producing strains, four in clinical samples and 13 fecal carriage isolates among 443 Gram negative bacteria [
18]. Another reported NDM genes in clinical
E. coli isolates [
21]. Recent studies in the country have discovered carbapenemase-producing bacteria or the related genes to be prevalent in the hospital waste waters implying their likely presence in clinical samples as well [
22,
23]. In the current study, we assessed the prevalence of carbapenemase and AmpC-β-lactamase production among ESBL-producing
E. coli (ESBL-Ec) and ESBL-producing
Klebsiella spp. (ESBL-K) isolated from clinical samples and their co-resistance to non-betalactam anibiotics and distribution in hospitals of primary, secondary and tertiary health care in Burkina Faso.
Discussion
Carbapenemase and AmpC-β-lactamase detection is not routine practice in microbiology laboratories in Africa, despite several studies reporting the presence of carbapenemase and AmpC-β-lactamase-producing Enterobacterales [
14,
20,
24,
25,
26,
27,
28,
29]. In our study, the prevalence of carbapenemase production was 5.3%. Twenty-five (25) isolates, primarily
E. coli (76%), were carbapenemase producers, most of them originating in urine samples. Our results fall within the prevalence range of 2.6–6.7% reported in North Africa, but remain somewhat lower than the range of 9.0–60.0% reported in sub-Saharan Africa in a review from 2015 [
30]. In Burkina Faso, a study conducted at Sanou Sourou teaching hospital in Bobo-Dioulasso reported a prevalence as low as 0.9% for carbapenemase producers among Gram-negative bacteria in clinical specimen, most of them urine samples [
18], and another reported NDM, VIM, OXA-48 and KPC genes among E. coli isolated from clinical patient samples [
21]. In contrast, a study conducted in two tertiary hospitals in northwestern of Nigeria reported a prevalence of 39% among 248 ESBL and non-ESBL
E. coli and
Klebsiella pneumoniae clinical isolates. These isolates were primarily obtained from urine samples [
14]. These variations in prevalence could be attributed to several factors, including differences in sample types, methods employed for carbapenemase detection, and the geographical regions in which the study was conducted. Nonetheless, in all the studies, carbapenemase-producing isolates were predominantly urinary pathogens. This can be attributed to two evident reasons: first, urine samples are among the most common samples investigated in microbiological laboratories, and second, urinary tract infections constitute the most prevalent symptomatic manifestation of intestinal colonization by MDR bacteria.
Both Class B (NDM and VIM) and class D (OXA-48-like) carbapenemases were detected in our study. NDM carbapenemase-producing Enterobacterales were the most frequent findings (76%), consistent with results of several earlier studies from sub-Saharan Africa [
16,
21,
30,
31]. In our study, we identified carbapenemase-producing isolates in three hospitals, each representing different levels of the healthcare system. CHU-YO, a referral hospital at the highest level of specialized care facilities, showed substantial prevalence for both. This may be attributed to more abundant use of antibiotics, use of broader spectrum antibiotics, and prolonged duration of drug treatments as patients are referred from medical centers, regional hospitals, or medical clinics where they have already received antimicrobial treatment. Indeed, the misuse or overuse of antibiotics during hospitalization can contribute to the selection of MDR bacteria [
32,
33].
Production of AmpC-β-lactamases has been described in Africa at various rates [
20,
29,
34] for example at a rate of 49.3% in Sudan [
20], 15.2% in Nigeria [
29], 2.5% in Ethiopia [
24] and 36.5% in Uganda [
35]. In our study, exploring 473 clinical ESBL-PE isolates, co-production of AmpC-β-lactamases was observed in 5.3% of isolates. Co-production of AmpC-β-lactamase at a rate of 5.2% was seen in a previous study from India [
26] and 22% in Iran [
36]. In a study from Ethiopia 3.6% of AmpC-β-lactamase positive isolates co-produced ESBL (5/139) [
24]. Many other studies also report either separately ESBL and AmpC-β-lactamase production in various LMIC countries [
24,
26,
34]: variations may be attributed to differences in sample sizes, types of study areas, and phenotypic methods used for AmpC-β-lactamase detection, which can yield differing results [
29]. All these data unmistakably confirm the presence of AmpC-β-lactamase in clinical isolates. Consequently, AmpC-β-lactamase detection should be implemented in routines of hospitals also in Africa. This is of utmost importance since AmpC-β-lactamase production in bacteria can lead to challenges and failures of treatment, with increased morbidity and mortality [
29].
A total of 473 ESBL-isolates (356 E. coli and 117 Klebsiella spp.) were tested against various antibiotics. The resistance rates to regimen with potential efficacy against ESBL-producers, i.e. piperacillin + tazobactam, and amoxicillin + clavulanic acid, remained high.
ESBL-producing Enterobacterales have been investigated in numerous studies in Africa, revealing its generally high prevalences. For example, in a study in Kenya with rates of 42% versus 45% in urban versus rural communities and 70% and 63% in urban versus rural hospitals [
38]. In 2017–2020 in Ghana, 50% of
E. coli and 59% of
Klebsiella sp in urinary samples were ESBL-producing [
39], and in Tanzania approximately 20% [
40]. A recent study from Burkina Faso showed ESBL-PE prevalences of
3.2% among E. coli isolates from urine in pregnant women versus 35.4% among clinical isolates [
41]
. Other reports from LMICs include for example those from Kano, northwest Nigeria [
42], Khartum, Sudan [
20], Ethiopia [
24], and Algeria [
17], and India [
28] The high ESBL-PE rates in LMICs and their potential co-resistance should draw the attention of clinicians who often prescribe β-lactams in prophylaxis or infection management in health care facilities [
43].
Although carbapenem-producing Enterobacterales often carry both ESBL and carbapenemase genes, not all CPE strains can be covered by a study on ESBL-PE, such as the present one. However, as most of the CPEs do, it was of interest to see in Burkina Faso, how large proportion of ESBL-PE actually are CPE strains. Most (76%) of the identified CPE were
E. coli, with NDM as the most common carbapenemase type, reaching a prevalence of approximately 4% among our ESBL-PEs. The respective resistance rates with disc diffusion method proved somewhat higher, reaching almost 20 % for ertapenem among ESBL-Ec. Similar findings were reported in previous studies across Africa, resistance rates of 0 to 14.7% to imipenem and meropenem have been reported in recent studies in Togo, Nigeria, Ethiopia and Sudan among ESBL or AmpC-β-lactamase-producing isolates [
20,
24,
42,
44]. Our results were lower compared to 58.6% resistance against imipenem reported in India [
28]. Earlier studies in Burkina Faso and Ghana did not find resistance to carbapenems [
18,
45], the differences potentially attributed to antibiotic consumption in various study areas [
46]. For instance, in India carbapenem consumption is higher [
47] than in Burkina Faso, where their use is more controlled and the drugs are very expensive. The presence of resistance could be attributed to the co-presence of genetic determinants of resistance to carbapenems with those of other antibiotics commonly used in our hospitals [
22].
Among our ESBL-PE isolates, co-resistance was recorded particularly against aminoglycosides, fluoroquinolones, and sulfonamides. Among aminoglycosides, high resistance rates were recorded against all regimen tested, except amikacin. This finding is in line with those reported in similar studies in Burkina Faso [
18], in Togo [
44], in Sudan [
20], in Ethiopia [
24], in Algeria [
17] and India [
28,
48]. Interestingly, our resistance rates appear high with respect to similar studies, which show wide variations in resistance level [
42,
45]. Our high rates could be associated with misuse or overuse of antibiotics in hospitals, communities or in farms since these antibiotics are routinely used and have easy access [
49].
Amikacin, fosfomycin and chloramphenicol emerged as the most effective antibiotics in vitro. Our results align with previous research in Burkina Faso in 2021 [
18], Algeria in 2019 [
17], and Nigeria in 2014 [
42], yet contrasting the higher resistance to amikacin observed in Ghana in 2013 [
45]. Apart from having fosfomycin as one of the alternative drugs in cystitis (IDSA guideline), these three antibiotics are not the ones initially preferred for many infections but rather represent reserve antibiotics in current practice. Indeed, when treating infections caused by ESBL- or carbapenemase-producing Enterobacterales it is often necessary to resort to less effective reserve antibiotics.
Material and Methods
Study Design and Period
The prospective study was carried out in five hospitals in Burkina Faso from January 2020 to June 2022. ESBL-producing E. coli and Klebsiella spp. strains isolated from urine, blood and pus samples were collected from each hospital over a 12-month period. All isolates were characterized in the CRUN microbiology laboratory.
Sampling and Sites Description
The health system in Burkina Faso comprises three levels. The first level encompasses peripheral health care facilities and primary hospitals. The second level comprises regional hospitals and certain medical clinics, which serve as reference facility for primary hospitals. The third level includes national and teaching hospitals, representing the highest level of referral care and offering specialized services [
50].
Sampling was carried out in five hospitals, each selected to represent different levels of the health system: 1) Yalgado Ouédraogo teaching hospital (CHU-YO), a tertiary hospital located in the capital city, Ouagadougou; 2) The regional hospital of Koudougou (CHR-KDG); 3) the El-Fateh Suka medical clinic in Ouagadougou, both categorized as secondary hospitals, and two medical centers, 4) CMA Saint Camille de Nanoro a rural medical center and 5) CMA évangélique Source de vie in Ouagadougou, the last two representing primary health care. This diverse selection of healthcare units allowed a comprehensive sampling approach. A total of 473 clinical isolates were collected, comprising 356 ESBL-Ec and 117 ESBL-K strains. These isolates were collected from various clinical specimen, including urine (n=303), pus (n=140), and blood culture (n=30). The isolates were collected in tryptic soy agar tubes and kept at room temperature until transferred to the Clinical Research Unit of Nanoro (CRUN) microbiology laboratory for analysis.
Bacterial Isolation and Identification
At CRUN microbiology lab, isolates were plated on ESBL-selective chromogenic culture media (CHROMagarTM ESBL). Isolates Identification of the isolates was verified using API 20E (Biomérieux France) according to the manufacturer’s instructions.
ESBL Production Test
All isolates identified were tested for ESBL production using the double disk synergy test between 3rd generation cephalosporins (ceftriaxone and ceftazidime) and 4th generation cephalosporin (cefepime) disks and amoxicillin-clavulanic acid disk according to the CLSI 2022 guidelines. The presence of ESBL production was indicated by the presence of synergistic inhibition zone between ceftazidime, ceftriaxone and/or cefepime and the amoxicillin-clavulanic acid disk.
Carbapenemase Production Test
A total of 30 isolates (22 E. coli and 8 Klebsiella spp.) that had meropenem inhibition zone diameter less than 22 mm in the AST, were investigated for production of the five main carbapenemases (OXA48-like, NDM, KPC, VIM and IMP) using the immunochromatographic test O.K.N.V.I. RESIST-5 (CORIS BioConcept, Belgium), according to the manufacturer's instructions.
AmpC-β-Lactamase Production
Bacterial isolates with cefoxitin inhibition zone diameter less than 18 mm were considered presumptive AmpC-β-lactamase producers. A total of 92 presumptive AmpC-β-lactamase producers bacterial isolates (73 ESBL-E. coli and 19 ESBL-Klebsiella spp.) were tested for AmpC-β-lactamase production using MH agar supplemented with cloxacillin at 4µg/l. A bacterial suspension prepared with fresh colonies (McFarland 0.5) was inoculated on to the entire surface of the MH agar supplemented with cloxacillin at 4µg/l and a disk of cefoxitin was placed on the plate. The test was positive if the inhibition zone diameter around cefoxitin disk was ≥18 mm.
Antimicrobial Susceptibility Test
Antimicrobial susceptibility test (AST) was performed using the disk diffusion method on Mueller Hinton (MH) agar as described by Bauer et
al., 1966. In total, 473 ESBL-producing isolates (356
E. coli and 117
Klebsiella spp.) were tested. The results were interpreted according to the American Clinical and Laboratory Standards Institute (CLSI) 2022 guidelines. 15 antibiotics listed in
Table 4 were tested.