Characterization of multidrug-resistant Gram-negative bacilli isolated from hospitals effluents: first report of a bla_OXA-48-like in Klebsiella oxytoca, Algeria

Abstract

The antibiotic susceptibility profile and antimicrobial resistance determinants were characterized on Gram-negative bacilli (GNB) isolated from Algerian hospital effluents. Among the 94 isolates, Enterobacteriaceae was the predominant family, with Escherichia coli and Klebsiella pneumoniae being the most isolated species. In non-Enterobacteriaceae, Acinetobacter and Aeromonas were the predominant species followed by Pseudomonas, Comamonas, Pasteurella, and Shewanella spp. The majority of the isolates were multidrug-resistant (MDR) and carried different antimicrobial resistance genes including bla_CTX-M, bla_TEM, bla_SHV, bla_OXA-48-like, bla_OXA-23, bla_OXA-51, qnrB, qnrS, tet(A), tet(B), tet(C), dfrA1, aac(3)-IIc (aacC2), aac(6′)-1b, sul1, and sul2. The qacEΔ1-sul1 and intI2 signatures of class 1 and class 2 integrons, respectively, were also detected. Microarray hybridization on MDR E. coli revealed additional resistance genes (aadA1 and aph3strA, tet30, mphA, dfrA12, bla_cmy2, bla_ROB1, and cmlA1) and classified the tested strains as commensals, thus highlighting the potential role of humans in antibiotic resistance dissemination. This study is the first report of bla_OXA-48-like in Klebsiella oxytoca in Algeria and bla_OXA-23 in A. baumannii in Algerian hospital effluents. The presence of these bacteria and resistance genes in hospital effluents represents a serious public health concern since they can be disseminated in the environment and can colonize other hosts.

Introduction

Antimicrobial resistance resulting from intensive and inappropriate use of antibiotics represents an increasing challenge for health care services worldwide. In Gram-negative bacilli (GNB), this resistance is a cause for concern, owing to the emergence and dissemination of different resistance genes identified either in clinical setting or in the environment. Within the environment, aquatic ecosystems have been documented as an important vehicle and reservoir of antibiotic resistance [1]. Different studies have shown that hospital effluents contain high levels of antibiotic-resistant bacteria and contribute to antibiotic resistance dissemination in the environment [2, 3]. Moreover, the non-metabolized fraction of consumed antibiotics excreted directly in wastewater can exert a selective pressure promoting the selection of these resistant bacteria [2]. Generally, hospital effluents are discharged into municipal sewage and collected via wastewater treatment plants (WWTP) or eliminated directly into the environment without any treatments. The latter practice has been criticized and a pre-treatment has been recommended [4].

The resistant bacteria present in hospital wastewater could be a source of genes encoding various resistance mechanisms including extended spectrum β-lactamases (ESBLs), carbapenemases, and plasmid mediated quinolones resistance (PMQRs) [3, 5, 6]. These genes may be transmitted to other bacteria present in sewage and the environment via the aquatic system and the food chain. The spread of these bacteria in the natural environment can have significant consequences on ecological balance and animal and public health [2]. The present study aimed to identify and characterize resistance mechanisms of multidrug-resistant GNB found in hospital effluents and to improve our understanding on the role of hospital effluents as a source of a large diversity of antibiotic resistance genes.

Material and methods

Bacterial sampling

Twenty wastewater samples were collected from untreated effluents from four hospitals in Algeria: two in Bejaia and two in Tizi Ouzou provinces. Sampling was carried out during the period of maximum hospital activity, in 2011 and 2012 in Bejaia and Tizi Ouzou, respectively. In Bejaia, the samples were collected from non-functional WWTP and sewer pipes from Khelil Amrane (H1) and Amizour (H2) hospitals, respectively. In 2011, the H1 hospital had an annual capacity of 160 beds and admitted 5255 patients and the H2 hospital had a capacity of 209 beds and admitted 9360 patients. In Tizi Ouzou, samples were collected from sewer pipes from Nadir Mohammed hospital (H3) and Tigzirt hospital (H4). In 2012, the H3 hospital had 1042 beds and admitted 40,170 patients and the H4 hospital was the only establishment located in a rural region. It had a capacity of 35 beds and admitted 638 patients annually. All the hospitals included in this study were located close to the Mediterranean Sea with approximate distances of 0.4, 8, 33, and 44 km for H4, H1, H2, and H3, respectively. In addition, the H3 hospital was located at about 15 km from Oued Sebaou and the H2 at around 15 km from Oued Soummam, the main rivers of the provinces of Tizi Ouzou and Bejaia, respectively, which flow into the Mediterranean Sea.

Microbiological analysis

A volume of 200 ml of each sample was collected in sterile glass flasks and transported to the laboratory in a cooler, and samples were processed within 2 h upon arrival. Serial dilutions of 10⁻¹ to 10⁻⁴ of each sample were performed. One hundred microliters of each sample and their corresponding dilutions were plated on MacConkey agar supplemented with 2 μg/ml of cefotaxime followed by incubation at 37 °C for 18–24 h. A single colony from colonies sharing the same appearance by visual inspection (morphology, color, shape, size) was selected and purified. After purification, the isolates were stored in a nutrient broth containing 10% glycerol (v/v) at − 80 °C. Rapid identification was performed using the API 20E test strip (Biomerieux, France) for Enterobacteriaceae according to the manufacturer’s instructions in conjunction with conventional biochemical testing [7]. Non-Enterobacteriaceae bacteria were identified by sequencing the 16S rRNA and rpoB according to the Laboratoire de Santé Publique du Québec (LSPQ) reference methods or gyrB for Shewanella spp. [6].

The bacterial diversity of different genera of each hospital was evaluated using Shannon diversity index. The BPMSG diversity online calculator website https://bpmsg.com/academic/div-calc.php was used to calculate this index.

Antimicrobial susceptibility testing and extended-spectrum β-lactamase (ESBL) detection

Antimicrobial susceptibility was performed using the disk diffusion method and the results were interpreted according to the Clinical and Laboratory Standards Institute (CLSI, 2014) guidelines [8]. The antibiotics (Bio-Rad, France) tested were as follows: ticarcillin (TIC, 75 μg), ticarcillin-clavulanic acid (TIM, 75/10 μg), amoxicillin-clavulanic acid (AMC, 20/10 μg), ertapenem (ETP, 10 μg), imipenem (IPM, 10 μg), ceftazidime (CAZ, 30 μg), cefepime (FEP, 30 μg), aztreonam (ATM, 30 μg), gentamicin (GEN, 10 μg), amikacin (AMK, 30 μg), ciprofloxacin (CIP, 5 μg), tetracycline (TET, 30 μg), and trimethoprim-sulfamethoxazole (SXT, 1.25/23.75 μg). The ESBL production was detected by the CLSI confirmatory test using both CTX (30 μg) and CAZ (30 μg) disks alone and in combination with clavulanic acid (CA, 10 μg) [8].

Detection of antibiotic resistance genes

DNA extraction was performed by the MagAttract® DNA Mini M48 kit using the biorobot M48 (Qiagen, USA) according to the manufacturer’s instructions. For β-lactam resistance genes, PCR amplification targeting bla_CTX-M, bla_SHV, and bla_TEM genes was carried out in a 50 μl final volume using Phusion DNA Polymerase (New England Biolabs, USA) [9, 10]. For the carbapenemases encoding genes which included bla_OXA-23, bla_OXA-24, bla_OXA-51, and bla_OXA-58, multiplex PCR was performed in a 50 μl final volume using HotStar Taq enzyme (Qiagen, USA) [11, 12]. The bla_oxa-48-like genes were amplified using simplex PCR with the same amplification conditions.

Genes encoding resistance to non-β-lactam antibiotics were amplified in a 25 μl volume using AmpliTaq DNA polymerase (ThermoFisher scientific, USA) according to the manufacturer’s guidelines. The non-β-lactam resistance genes included qnrA, qnrB, and qnrS encoding resistance to quinolones [13, 14], acc(6′)-lb and aac(3)-IIc (aacC2) encoding resistance to aminoglycosides, dfrA1encoding resistance to trimethoprim, tet(A), tet(B), and tet(C) encoding resistance to tetracycline, and sul1, sul2, and sul3 encoding resistance to sulfonamides [15, 16]. In addition, qacEΔ1-sulI and intI2 genetic markers for class 1 and class 2 integrons, respectively, were also amplified in the same conditions [17, 18]. The primers used in this study and PCR conditions are listed in Table 1.

Table 1.

Primers and PCR conditions used in this study

Primer	Target	Control strain	Primer sequence (5′–3′)	PCR conditions	Amplicon size (pb)	Reference
SHVF2 SHVR	bla SHV	ATCC BAA200	GGTTAT GCG TTA TAT TCG CC TTA GCG TTG CCA GTG CTC	40 cycles of 30 s at 98 °C, 10 s at 98 °C, 30 s at 65 °C, 30 s at 72 °C; 10 min at 72 °C; 4 °C	867	[9]
TEMF TEMR	bla TEM	ATCC BAA197	ATG AGT ATT CAA CAT TTC CG CTG ACA GTT ACC AAT GCT TA	30 cycles of 30 s at 98 °C, 10 s at 98 °C, 30 s at 58 °C, 30 s at 72 °C;10 min at 72 °C; 4 °C	867	[9]
CTX-M-U1 CTX-M-U2	bla CTx-M	ID098246	ATG TGC AGY ACC AGT AAR GTK ATG GC TGG GTR AAR TAR GTS ACC AGA AYC AGC GG	30 cycles of 30 s at 98 °C, 10 s at 98 °C, 30 s at 65 °C, 30 s at 72 °C;10 min at 72 °C; 4 °C	593	[10]
OXA-P1 OXA-P2	bla OxA-48	LSPQ4078	ATG CGT GTA TTA GCC TTA TC CTA GGG AAT AAT TTT NTC CT	10 min at 20 °C; 30 cycles of 15 min at 95 °C, 60 s at 94 °C, 60 s at 57 °C; 10 min at 72 °C; 4 °C	759	This study
OXA-23-like-FWD OXA-23-like-REV	bla OXA-23	MA092246	GAT CGG ATT GGA GAA CCAGA ATT TCT GAC CGC ATT TCC A		501	[11]
OXA-24-like-FWD OXA-24-like-REV	bla OXA-24	MA09250	GGT TAG TTG GCC CCC TTA AA AGT TGA GCG AAA AGG GGA TT		246	[11]
OXA-51-like-FWD OXA-51-like-REV	bla OXA-51	LSPQ4139	TAA TGC TTT GAT CGG CCT TG TGG ATT GCA CTT CAT CTT GG		353	[12]
OXA-58-like-FWD OXA-58-like-REV	bla OXA-58	LSPQ4139	AAG TAT TGG GGC TTG TGC TG CCC CTC TGC GCT CTA CAT AC		599	[11]
QnrAm-F QnrAm-R	qnrA	J53pMG252	AGA GGA TTT CTC ACG CCA GG TGC CAG GCA CAG ATC TTG AC	10 min at 95 °C and 35 cycles of amplification consisting of 1 min at 95 °C, 1 min at 54 °C and 1 min at 72 °C and 10 min at 72 °C	580	[13]
qnrB-forward qnrB-reverse	qnrB	J53pMG298	GAT CGT GAA AGC CAG AAA GG ACG ATG CCT GGT AGT TGT CC	94 °C for 4 min, 35 cycles of 94 °C of 30 s, 58 °C for 30 s, and 68 °C for 45 s, and a final step of 68 °C for 10 min	469	[14]
qnrS-forward qnrS-reverse	qnrS	J53pMG306	ACG ACA TTC GTC AAC TGC AA TAA ATT GGC ACC CTG TAG GC		417	[14]
aac-forward aac-reverse	acc (6′)-lb	Salmonella SA20042859	TTG CGA TGC TCT ATG AGT GGC TA CTC GAA TGC CTG GCG TGT TT		482	[15]
fDFRa rDFRa	dfrA1	S17-1 lamda pir	AAG AAT GGA GTT ATC GGG AAT G GGG TAA AAA CTG GCC TAA AAT TG		391	[16]
fAACa rAACa	aacC2(aac(3)-IIa)	R176	CGG AAG GCA ATA ACG GAG TCG AAC AGG TAG CAC TGA G		740	[16]
fTETa rTETa	tet(A)	SAS1393	GTG AAA CCC AAC ATA CCC C GAA GGC AAG CAG GAT GTA G		888	[16]
fTETb rTETb	tet(B)	CT4afooB	CCT TAT CAT GCC AGT CTT GC ACT GCC GTT TTT TCG CC		774	[16]
fTETc rTETc	tet(C)	pBR322	ACT TGG AGC CAC TAT CGA C CTA CAA TCC ATG CCA ACC C		881	[16]
fSULa rSULa	sulI	SAS1393	TTC GGC ATT CTG AAT CTC AC ATG ATC TAA CCC TCG GTC TC		822	[16]
fSULb rSULb	sulII	RSF1010	CGG CAT CGT CAA CAT AAC C GTG TGC GGA TGA AGT CAG		722	[16]
fsul3 rsul3	sulIII	sul3	GAG CAA GAT TTT TGG AAT CG CTA ACC TAG GGC TTT GGA		790	This study
fQACe rSUL1	qacEΔ1-sul1	SAS1393	ATC GCA ATA GTT GGC GAA GT GCA AGG CGG AAA CCC GCG CC		797	[17]
intI2L intI2R	int2	PT508-124.61	CAC GGA TAT GCG ACA AAA AGG T GTA GCA AAC GAG TGA CGA AAT G		788	[18]

DNA microarray analysis of Escherichia coli isolates

Nine representative multidrug-resistant (MDR) E. coli isolates were analyzed by microarray hybridization. The E. coli Maxivirulence version 3.1 microarray was used as previously described [19]. This microarray can detect 348 virulence genes and 98 antibiotic resistance genes and can determine putative E. coli pathotypes based on virulence gene profiles [20] and phylogenetic groups (A1, B1, B2, and D) based on the pattern of presence or absence of chuA, TspE4.C2, and yjaA [21].

Results

Bacterial diversity

A total of 94 Gram-negative bacilli were cultured from selective plates. The distribution of these isolates was as follows: 25 isolates were from H1, 20 from H2, 22 from H3, and 27 from H4. Among the 26 different species identified, Enterobacteriaceae was the most predominant family with a prevalence of 78% (n = 73). The identified species from this family included E. coli (n = 19), Klebsiella pneumoniae (n = 14), Citrobacter freundii (n = 9), Klebsiella oxytoca (n = 7), Enterobacter cloacae (n = 5), Citrobacter braakii (n = 4), Morganella morganii (n = 3), Citrobacter youngae (n = 2), Proteus vulgaris (n = 2), Providencia rettegeri (n = 2), Cedecea lapagei (n = 1), Enterobacter aerogenes (n = 1), Enterobacter agglomerans (n = 1), Enterobacter asburiae (n = 1), Enterobacter sakazakii (n = 1), and Kluyvera sp. (n = 1). The prevalence of non-Enterobacteriaceae GNB was 22% (n = 21). This group included Acinetobacter baumannii (n = 5), Aeromonas hydrophila (n = 5), Aeromonas sp. (n = 3), Aeromonas veronii (n = 2), Pseudomonas aeruginosa (n = 2), Aeromonas sobria (n = 1), Comamonas aquatica (n = 1), Pasteurella spp. (n = 1), and Shewanella xiamenensis (n = 1). The predominant species from these hospitals were K. pneumoniae (13/25), E. coli (13/20), E. coli (5/22), and K. oxytoca (6/27) for H1, H2, H3, and H4, respectively. In addition, 10/21 non-Enterobacteriaceae isolates (47.6%) were collected from Nadir Mohamed hospital (H3). Shannon diversity index was different between the effluents of different hospitals and revealed greater genera diversity for effluents of H3 (2.07), followed by H4 (1.7), H1 (1.4), and H2 (1.2).

Antibiotic susceptibility of hospital effluents isolates

For β-lactam antibiotics, the highest resistance rate was observed with TIC representing 98.6% and 66.6% for Enterobacteriaceae and non-Enterobacteriaceae, respectively, followed by CAZ with 86.3% and 47.6% for Enterobacteriaceae and non-Enterobacteriaceae, respectively. For non-β-lactam antibiotics, the highest resistance rate observed with ciprofloxacin was 74% and 62% for Enterobacteriaceae and non-Enterobacteriaceae, respectively. All the isolates were resistant to at least two antibiotic classes and the MDR phenotype was observed in 91/94 (96.8%) isolates exhibiting resistance to 3 to 8 antibiotic classes. The MDR phenotype was observed in 72/73 isolates of the Enterobacteriaceae family (98.6%) and in 19/21 isolates in the non-Enterobacteriaceae families (90.5%) which were resistant to three or more antimicrobial classes. In addition, the ESBL phenotype was detected in 45/73 of Enterobacteriaceae (61.5%) and in 9/21 of non-Enterobacteriaceae GNB (42.8%).

Antibiotics resistance mechanisms

Among the genes encoding class A β-lactamases tested in this study, bla_CTX-M, bla_TEM, and bla_SHV were detected in 27/94 (28.7%), 17/94 (18.1%), and 14/94 (14.9%) isolates, respectively. The bla_CTX-M gene was detected only in the Enterobacteriaceae family including E. coli (n = 7) and K. pneumoniae (n = 7), K. oxytoca (n = 3), C. youngae (n = 2), C. freundii (n = 2), C. braakii (n = 1), E. aerogenes (n = 1), E. agglomerans (n = 1), E. asburiae (n = 1), M. morganii (n = 1), and P. vulgaris (n = 1). The bla_TEM gene was also detected in this family and was detected in E. coli (n = 5) and K. pneumoniae (n = 3) followed by K. oxytoca (n = 1), C. Youngae (n = 1), E. agglomerans (n = 1), and P. rettegeri (n = 1). The bla_SHV gene was detected in K. pneumoniae (n = 9), E. cloacae (n = 3), E. agglomerans (n = 1), and E. coli (n = 1). Of the 21 non-Enterobacteriaceae isolates, only bla_TEM was detected in five isolates including A. hydrophila (n = 2), Aeromonas spp. (n = 1), A. veronii (n = 1), and A. baumannii (n = 1).

Regarding carbapenemases encoding genes, bla_OXA-48-like was detected in K. oxytoca (n = 2). The bla_OXA-416 gene was also detected in S. xiamenensis isolate and the genome was fully sequenced and documented in a separate study [6]. The bla_OXA-23 gene was detected in two A. baumannii isolates whereas the bla_OXA-51 gene was detected in all the five isolates. The bla_OXA-24 and bla_OXA-58 genes tested negative in all isolates.

For genes conferring resistance to quinolones, the qnrB gene was detected in E. coli (n = 6), C. freundii (n = 5), K. pneumoniae (n = 4), K. oxytoca (n = 3), C. youngae (n = 2), E. cloacae (n = 2), E. asburiae (n = 1), M. morganii (n = 1), P. vulgaris (n = 1), and E. aerogenes (n = 1). However, the qnrS gene was detected only in K. oxytoca (n = 1) and A. baumannii (n = 2) isolates, whereas the qnrA gene tested negative in all isolates. The qnrVC6 variant was also detected in S. xiamenensis isolate as previously documented [6]. Genes conferring resistance to tetracycline tet(A), tet(B), and tet(C) were detected in 12/73 Enterobacteriaceae isolates (16.4%) and 8/21 non-Enterobacteriaceae isolates (38%). The dfrA1gene was found in Enterobacteriaceae isolates and distributed as follows: C. freundii (n = 3), E. coli (n = 4), K. oxytoca (n = 1), P. vulgaris (n = 1), and P. rettgeri (n = 2). In the non-Enterobacteriaceae isolates, this gene was detected in A. hydrophila (n = 2), A. sobria (n = 1), and A. baumannii (n = 2). The sul1gene was found in 14/73 Enterobacteriaceae isolates (19.2%) and in 10/21 non-Enterobacteriaceae isolates (47.6%), and the sul2 gene was detected in 33 /73 Enterobacteriaceae isolates (45.2%) and in 3/21 non-Enterobacteriaceae isolates (14.3%).

With regard to genes conferring resistance to aminoglycosides, the aac(3)-IIc (aacC2) gene was detected in 36/73 Enterobacteriaceae isolates (49.3%) and 3/21 non-Enterobacteriaceae isolates (14.3%), and the aac(6′)-1b gene was found in 28/73 Enterobacteriaceae isolates (38.3%) and in 8/21 non-Enterobacteriaceae isolates (38%). The qacEΔI-sul1 conserved segment of class 1 integron was detected in 9/73 Enterobacteriaceae isolates (12.3%) and in 1/21 non-Enterobacteriaceae isolates (4.8%), and the intI2 gene was found only in the Enterobacteriaceae group. The different gene combinations of MDR isolates are described in Table 2 and Table 3.

Table 2.

Antimicrobial resistance profiles and resistance genes detected in MDR Enterobacteriaceae isolates

Resistant

Intermediate

Susceptible

TIC ticarcillin, AMC amoxicillin-clavulanic acid, ETP ertapenem, IMP imipenem, CAZ ceftazidime, FEP cefepime, ATM aztreonam, GEN gentamicin, AMK amikacin, CIP ciprofloxacin, TET tetracycline, SXT trimethoprim-sulfamethoxazole

Table 3.

Antimicrobial resistance profile, resistance genes detected in MDR non-Enterobacteriaceae isolates

Resistant

Intermediate

Susceptible

TIC ticarcillin, AMC amoxicillin-clavulanic acid, ETP ertapenem, IPM imipenem, CAZ ceftazidime, FEP cefepime, ATM aztreonam, GEN gentamicin, AMK amikacin, CIP ciprofloxacin, TET tetracycline, SXT trimethoprim-sulfamethoxazole

Phylogenetic groups and virulence profile of MDR E. coli

Escherichia coli isolates tested by microarray were classified as commensal and were assigned to the phylogenetic groups A (n = 4), D (n = 3), and B1 (n = 2). However, some virulence genes commonly found in E. coli pathotypes were detected [22]. They included chuA, iha, cvaC, fimH, fyuA, iroN, iss, iutA, kpsM II, traT, ompA, ompT, TSPE4.C2, and yjaA. This analysis also detected other resistance genes not tested by PCR. They included aadA1 and aph3strA conferring resistance to aminoglycosides, tet30 conferring resistance to tetracycline, mphA conferring resistance to macrolides, dfrA12 conferring resistance to trimethoprim, bla_cmy2 and bla_ROB1 conferring resistance to β-lactams, and cmlA1 conferring resistance to chloramphenicol. The detail microarray results are presented in Table 4.

Table 4.

Phylogenitic groups, virulence genes, and antimicrobial resistance genes, detected by microarray, of tested Escherichia coli isolates

Isolate	PG	Virulence genes	Resistance genes
T3	B1	fimH/kpsM II/traT/ompA/ompT/TSPE4.C2	blaCTX-M/aacC2/sul1/tet(A)/tetR/tet30/mphA/dfrA12
T9	A	iutA/traT/ompA	blaCTX-M/aacC2/tet30/tetR/mphA
T10	D	chuA/fimH/fyuA/iutA/traT/ompA/TSPE4.C2	blaCTX-M/blacmy2/blaROB1/aacC2/sul1/mphA
T18	A	ompA	blaCTX-M/qnrB/sul2/tet(A)/tet30/tetR/int1
W1	D	chuA/cvaC/fimH/iroN/iss/iutA/traT/ompA	blaTEM/blacmy2/blaROB1/aadA1/dfrA1/sul2/tet(A)/tet30/mphA/tetR/int2
W4	D	chuA/iha/cvaC/fimH/iroN/iss/iutA/traT/ompA/ompT/ompA/TSPE4.C2	blaTEM/blacmy2/qnrB/blaROB1/aacC2/aadA1/dfrA1/int1sul1/sul2/tet(A)/tetR/aph3strA/mphA/cmlA1/int2
W30	A	cvaC/fimH/traT/ompA/yjaA	bla CTX-M
T15	A	fimH	blaTEM/qnrB/aacC2/aadA1/blaCTX-M/int1/sul2/tet(A)/tet30/tetR
W34	B1	cvaC/fimH/iroN/iss/traT/TSPE4.C2	blacmy2/aadA1/int1/sul1/blaROB1/mphA/cmlA1/dfrA12/int2

PG phylogenitic group

Discussion

Hospital effluents represent an important source of antibiotics and MDR bacteria carrying several antimicrobial resistance genes and mobile genetic elements (MGEs). This poses a significant threat to public health and the environment, especially when these effluents are discharged in the environment (rivers, lakes, and seas, etc.) without any prior treatments. In this context, our study investigated cephalosporin resistant Enterobacteriaceae and non-Enterobacteriaceae GNB isolated from hospital effluents in Algeria.

In this work, the predominant species isolated were E. coli followed by K. pneumoniae. These bacteria are both pathogenic and commensal organisms and are frequent causes of nosocomial infections [23, 24]. In addition, A. baumannii and A. hydrophila were the predominant species among bacteria other than Enterobacteriaceae isolated in our study. A. baumannii has recently emerged as a leading cause of hospital associated outbreaks infections resulting from contamination and transmission in hospital environments [25]. A. hydrophila was also reported among the most frequent cause of human diseases among Aeromonas spp. [26]. The high levels of resistance to the β-lactam antibiotics (TIC and CAZ), and to the non-β-lactam-antibiotics (CIP and SXT), observed in this study, may be associated to the selective pressure exerted by these antibiotics commonly used in antibiotherapy. The fact that most of the isolates were MDR is worrisome. This is because MDR encoding genes are commonly carried on MGE such as plasmids, integrons, and transposons [27].

In our study, bla_CTX-M was the most predominant gene among the β-lactamase encoding genes detected. In fact, in the last decade, bla_CTX-M has explosively disseminated and becomes the most common resistance gene around the world [28]. Previous studies in Bejaia (site of our study) have also reported the dissemination of bla_CTX-M in patients [29] and in wild fish from the Mediterranean Sea [30]. This result is similar to that of some studies carried out on hospital effluents, particularly in Romania [31] and Poland [3]. However, the prevalence of these genes was lower compared to that reported in Enterobacteriaceae isolated from Zemirli hospital effluents, in Algeria [5].

Interestingly, genes encoding carbapenemases including bla_OXA-23 and bla_OXA-48-like were detected for the first time in Algerian hospital effluents in A. baumannii, and K. oxytoca and S. xiamenensis, respectively. The emergence of such genes is considered as a threat in the treatment of GNB infections and can lead to MDR and extensive drug resistance; for example, MDR A. baumannii is recognized as the hardest nosocomial pathogen to control and treat [32]. The fact that bla_OXA-23 was previously reported in A. baumannii isolated from hospitals located on the Mediterranean coast [33] and from seabream fish from the Mediterranean Sea [34] underlines the probable role of hospital effluents in the transmission of such genes from hospitals into the environment. These genes may also be derived from other sources such as municipal sewage and animal husbandry [2]. Interestingly, S. xiamenensis OXA-416 producing which is a known reservoir of antimicrobial resistance genes carried several other resistance genes clustered on a transposon harbored by a pSx1 plasmid were detected and documented in a separate work [6]. The bla_OXA-48 gene was also recently reported in K. pneumoniae outbreaks in Algeria [35] and this also indicates a potential dissemination between patients and the environment.

The genes qnrB, qnrS, tet(A), tet(B), tet(C), dfrA1, aacC2, aac(6′)-1b, sul1, and sul2 conferring resistance to non-β-lactam-antibiotics detected in this study are common in GNB and were also reported in different regions of Algeria [36]. The integrons were previously detected in high proportions of Enterobacteriaceae strains isolated from hospital effluents [37]. In our study, the identification of class 1 and class 2 integrons suggests that the resistance genes could be associated with MGEs and emphasizes their role in the dissemination of antibiotic resistance.

Microarray analysis performed on representative MDR E. coli isolates revealed the presence of additional resistance genes not tested by PCR. This highlights the importance of other methods such as microarray and genomic technology in the detection of resistance genes and MGEs. Previous studies on resistant Enterobacteriaceae isolates conducted on effluents from Algiers’s hospital have reported the same resistance genes as in our study [5, 38] with the exception of bla_ROB1, bla_OXA-48, bla_OXA-51, bla_OXA-23, mphA, tet30, and cmlA1.

The E. coli phylogenetic groups A, B1, and D identified in this study had a low level of virulence and a high level of antimicrobial resistance, which is in agreement with previous studies [38]. This could be associated to the acquisition of antibiotic resistance [39]. The presence of these commensal MDR E. coli in hospital effluents underlines the role of humans as a source of antibiotic resistance in the environment.

In Algeria as well as many other developing countries, the use of antibiotics for human health and animal husbandry is not properly monitored and hospital effluents are continually discharged into the environment without pre-treatment. Selective pressure exerted by these antibiotics has several repercussions on public health and the environment, especially on the spread of antibiotic resistance. Our study revealed that Algerian hospital effluents carry MDR bacteria with several resistance genes, which could be transferred to different parts of the hydrological system such as surface water (used in agriculture) and sea water where they could reach aquatic animals as well as other neighboring countries. One cannot also exclude a possible transmission of MDR determinants to waterborne pathogens and their dissemination in the environment, which could lead to increased morbidity and mortality.

Owing to the global nature of the antimicrobial resistance problem, it is imperative for all countries to work together within a framework to better manage anthropogenic effluents, including hospital effluents, in order to limit the spread of MDR bacteria and antibiotic resistance genes.