Emergence of Resistance to Quinolones and β-Lactam Antibiotics in Enteroaggregative and Enterotoxigenic Escherichia coli Causing Traveler’s Diarrhea

The objective of this study was to assess the antimicrobial resistance of enteroaggregative Escherichia coli (EAEC) and enterotoxigenic E. coli (ETEC) strains causing traveler’s diarrhea (TD) and to investigate the molecular characterization of antimicrobial resistance genes to third-generation cephalosporins, cephamycins, and quinolones. Overall, 39 EAEC and 43 ETEC clinical isolates were studied.

ABSTRACT

The objective of this study was to assess the antimicrobial resistance of enteroaggregative Escherichia coli (EAEC) and enterotoxigenic E. coli (ETEC) strains causing traveler’s diarrhea (TD) and to investigate the molecular characterization of antimicrobial resistance genes to third-generation cephalosporins, cephamycins, and quinolones. Overall, 39 EAEC and 43 ETEC clinical isolates were studied. The susceptibilities of EAEC and ETEC against ampicillin, amoxicillin-clavulanic acid, cefotaxime, imipenem, chloramphenicol, tetracycline, co-trimoxazole, nalidixic acid, ciprofloxacin, azithromycin, and rifaximin were determined. All genes encoding resistance determinants were detected by PCR or PCR plus DNA sequencing. The epidemiology of selected EAEC and ETEC strains was studied using multilocus sequence typing (MLST). The resistance to quinolones of EAEC and ETEC strains causing TD has significantly increased over the last decades, and high percentages have been found especially in patients traveling to India and sub-Saharan Africa. Sequence type 38 (ST38) and ST131, carrying the bla_CTX-M-15 and bla_CTX-M-27 genes, respectively, are highly prevalent among extended-spectrum β-lactamase (ESBL)-producing EAEC and ETEC strains. The cephamycinase ACT-20 is described in the present study for the first time in EAEC and ETEC strains causing TD in patients who had traveled to Central America. The percentages of resistance to azithromycin in EAEC and ETEC isolates from patients to Southeast Asia/India and Africa are above 25%. Meanwhile, rifaximin is still active against EAEC and ETEC, with the prevalence of resistant strains not being high. In conclusion, fluoroquinolones should no longer be considered the drugs of choice for the prevention or treatment in TD for travelers traveling to India and Africa. Azithromycin and rifaximin are still a good alternative to treat TD caused by EAEC or ETEC.

INTRODUCTION

Traveler’s diarrhea is the most frequent infection presented by travelers attending a travel medicine unit following a trip to low- or middle-income countries (1, 2). Enteroaggregative Escherichia coli (EAEC) and enterotoxigenic E. coli (ETEC) are two of the most frequent bacteria causing traveler’s diarrhea (TD), together with Shigella spp. and Campylobacter spp. (3–5). Other nonbacterial enteric pathogens identified as etiological agents of TD in minor proportions (between 28% and 35%) are Norovirus, Giardia, and Cryptosporidium spp. (5).

In the cases of TD where antimicrobial treatment is suggested, fluoroquinolones, azithromycin, and rifaximin are the recommended antibiotics (6, 7). In addition, fluoroquinolones have been considered an option in the prevention of TD in travelers at high risk, such as immunocompromised patients in whom chemoprophylaxis is considered essential. However, other views concerning the antibiotic use for TD (especially for mild and moderate diarrhea) have emerged recently, as antimicrobials have been shown to be an independent risk factor that predisposes travelers to contracting resistant strains, such as extended-spectrum β-lactamase (ESBL)-producing Enterobacteriaceae (8).

Information in the scientific literature regarding the antimicrobial susceptibilities of EAEC and ETEC strains is scarce; therefore, we do not know whether the guidelines recommending empirical treatment remain valid. In addition, the prevalence of ESBLs as a mechanism of resistance to third-generation cephalosporins has mainly been investigated in extraintestinal E. coli (ExPEC) strains, but there is little research into diarrheagenic E. coli. The main purpose of this study was to assess the antimicrobial resistance of EAEC and ETEC strains causing traveler’s diarrhea during the period of 2011 to 2017 and investigate the mechanisms of resistance to third-generation cephalosporins, cephamycins, and quinolones.

RESULTS

The susceptibilities of 39 EAEC and 43 ETEC clinical isolates were determined by disk diffusion, Etest, or microdilution methods and are shown in Fig. 1 and in Table S1 in the supplemental material. Overall, EAEC showed greater resistance than ETEC, without significant differences. EAEC presented the following percentages of resistance: ampicillin (AMP), 56.4%; amoxicillin-clavulanic acid (AMC), 12.8%; cefotaxime (CTX), 12.8%; co-trimoxazole (SXT), 59%; chloramphenicol (CHL), 7.7%; tetracycline (TET), 51.3%; nalidixic acid (NAL), 43.6%; ciprofloxacin (CIP), 23%; and azithromycin (AZT), 23%. Meanwhile, the percentages of resistance of the ETEC clinical isolates were AMP, 48.9%; AMC, 7%; CTX, 14%; SXT, 44.2%; CHL, 11.6%; TET, 39.5%; NAL, 44.2%; CIP, 21%; and AZT, 14%. All EAEC and ETEC clinical isolates were susceptible to imipenem. Since no breakpoints are defined for rifaximin, we determined the MIC₅₀ and MIC₉₀ for EAEC and ETEC, being 8 and 16 μg/ml, respectively, for the two E. coli pathotypes.

FIG 1.

Percentages of resistance to different antibacterial agents in EAEC and ETEC strains. AMP, ampicillin; AMC, amoxicillin-clavulanic acid; CTX, cefotaxime; IMI, imipenem; SXT, co-trimoxazole; CHL, chloramphenicol; NAL, nalidixic acid; CIP, ciprofloxacin; TET, tetracycline; AZT, azithromycin.

The distribution of the percentages of resistance according to the E. coli pathotype and the geographical area visited is shown in Table 1. The resistance percentages were similar between EAEC and ETEC, with levels of resistance to cefotaxime greater than 38% in strains isolated from patients traveling to Southeast Asia/India than patients traveling to other areas. The prevalence of strains resistant to co-trimoxazole was higher in Africa than in Southeast Asia/India and Latin America. The prevalence of nalidixic acid-resistant strains was greater than 27% in all areas, being above 64% in Southeast Asia/India. The high level (>40%) of strains resistant to ciprofloxacin in Southeast Asia/India is worthy of mention, while in Africa, the percentage was intermediate (between 11 and 19%), and in Latin America, it was below 10%. Azithromycin-resistant strains were also more frequent in Southeast Asia/India than in Africa and Latin America, with resistance rates of 33.3%, 25%, and 9.1%, respectively, for EAEC and 28.6%, 11.1%, and 0%, respectively, for ETEC. It is important to highlight that 58% of the patients with TD from Southeast Asia/India visited India, and among these, the percentages of resistance to nalidixic acid were 75% and 71.4% for EAEC and ETEC, respectively; ciprofloxacin resistance rates were 62.5% and 43% for EAEC and ETEC, respectively; and rates of resistance to azithromycin were 37.5% and 28.6% for EAEC and ETEC, respectively (data not shown). However, statistical analysis was not performed due to the low population size obtained when stratifying the strains according to pathotype and geographical origin.

TABLE 1.

Percentages of resistance of EAEC and ETEC strains to different antimicrobial agents according to three geographical areas

Antimicrobial agent	% resistant isolates (no.) for:
	EAEC			ETEC
	Southeast Asia/India (n = 12)	Africa (n = 16)	Latin America (n = 11)	Southeast Asia/India (n = 14)	Africa (n = 18)	Latin America (n = 11)
Cefotaxime	33.3 (4)a	6.3 (1)a	0 (0)	42.9 (6)a	0 (0)	0 (0)
Co-trimoxazole	50(6)	81.3 (13)	36.4 (4)	35.7 (5)	61.1 (11)	27.3 (3)
Nalidixic acid	66.7 (8)	37.5 (6)	27.3 (3)	64.3 (9)	38.9 (7)	27.3 (3)
Ciprofloxacin	41.7 (5)	18.8 (3)	9.1 (1)	42.9 (6)	11.1 (2)	9.1 (1)
Azithromycin	33.3 (4)	25 (4)	9.1 (1)	28.6 (4)	11.1 (2)	0 (0)

^aESBL-producing strains.

The mechanisms of resistance to third-generation cephalosporins are usually associated with the production of ESBLs. This analysis was determined in all 11 isolates presenting cefotaxime resistance by the disk diffusion test (diameters obtained ranged from 8 to 18 mm, considering the criteria resistance [R], ≤ 22 mm; intermediate [I], 23 to 25 mm; and susceptibility [S], ≥ 26 mm) and a positive result by double-disk synergy test. The MIC of cefotaxime was determined, showing a range from 6 to >256 μg/ml in EAEC isolates and from 6 to 96 μg/ml in ETEC isolates. All the isolates were positive for CTX-M ESBL; eight of them belonged to CTX-M-15 and three to the CTX-M-27. The multilocus sequence type (MLST) analysis generated a high heterogeneity of types. Indeed, it was of note that three strains from India belonging to sequence type 38 (ST38) (two EAEC and one ETEC) carried the bla_CTX-M-15 gene; however, the plasmid typing was K for one EAEC strain, FIB (allele 33) and FII (allele 1) for the other EAEC strain, and Y for the ETEC strain. In addition, two ETEC strains carrying the bla_CTX-M-27 gene belonging to the high-risk clone ST131 had the same plasmid type profile (FIB, allele 20; FIA, allele 2). One of the patients with TD caused by the ST131 clone had visited India, while another had traveled to Vietnam and Cambodia (Table 2). Moreover, the bla_CTX-M-27 gene was also found in a singleton (ST1193) strain of EAEC carrying the same plasmid replicon types as the ST131-belonging strains but with different alleles. Five EAEC and three ETEC isolates were resistant to amoxicillin-clavulanic acid, with MICs from 12 to 64 μg/ml (data not shown). The presence of genes encoding OXA, TEM, SHV, and plasmid mediated AmpC-type β-lactamases was determined. The bla_ACT-20 gene was detected in two strains, one EAEC and one ETEC, showing the highest MICs for amoxicillin-clavulanic acid of 24 and 64 μg/ml, respectively. The MIC for cefoxitin for both strains was >256 μg/ml. These patients had visited Guatemala and the Dominican Republic (Table S1). The remaining six strains (four EAEC and two ETEC), with an MIC of 12 μg/ml to amoxicillin-clavulanic acid and MICs between 6 and 16 μg/ml to cefoxitin, presented the bla_OXA-1-like gene, which explains the moderate level of resistance to amoxicillin-clavulanic acid. Two EAEC strains isolated from patients who had traveled to India also harbored the bla_SHV-like gene, and none of the overall strains were found to carry the bla_TEM gene.

TABLE 2.

Main features of the EAEC and ETEC clinical isolates carrying ESBLs

Isolate no. by type	Geographical origin	MLST		Plasmid typing (allele no.)	CTX-M type	CTX MIC (μg/ml)
		ST	CC
EAEC
5	India	ST1193	Singleton	FIB (10), FIA (6)	CTX-M-27	8
11	India	ST38	ST38	FIB (33), FII (1)	CTX-M-15	>256
20	India	ST7615	Singleton	I1, K, B/O	CTX-M-15	6
70	India	ST38	ST38	K	CTX-M-15	128
84	Togo	ST44	ST10	I1, FIA (6)	CTX-M-15	>256
ETEC
36	India/Nepal	ST23	ST23	FIB (not described)	CTX-M-15	64
38	India	ST1284	Singleton	FIA (20)	CTX-M-15	96
39	India	ST131	ST131	FIA (2), FIB (20)	CTX-M-27	6
43	Vietnam/Cambodia	ST131	ST131	FIA (2), FIB (20)	CTX-M-27	12
102	China	ST5584	Singleton	Y	CTX-M-15	12
107	India	ST38	ST38	Y	CTX-M-15	64

Two phenotypes could be defined among quinolone-resistant EAEC (17 strains) and ETEC (20 strains) strains. One was nalidixic acid resistant (NAL^r) but ciprofloxacin intermediate or susceptible (CIP^i-s) (considering the criteria resistant [R], <24 mm; intermediate [I], 24 to 25 mm; and susceptible [S], ≥ 26 mm). The second phenotype corresponded to the strains which were resistant to both nalidixic acid and ciprofloxacin (NAL^r CIP^r). Both chromosome- and plasmid-mediated quinolone resistances were found (Table 3). All eight EAEC strains with the NAL^r CIP^i-s phenotype showed a mutation in the gyrA gene, whereas only one EAEC strain with the NAL^r CIP^r phenotype presented a mutation in amino acid codon Ser-83 of the gyrA gene. The remaining eight strains with the NAL^r CIP^r phenotype showed the following mechanisms of resistance: four strains had a mutation in the same position in the gyrA gene and in the amino acid codon Ser-80 of the parC gene, two strains had the same double mutation plus the qnrS gene, and two strains also had this double mutation and the aac(6′)-Ib-cr gene. In the 11 ETEC strains with the Nal^r CIP^i-s phenotype, the mechanisms of resistance to quinolones found were eight strains with a mutation in the amino acid codon Ser-83 of the gyrA gene, one with a mutation in the amino acid codon Asp-87, and only the qnrS gene was detected in the last strain (the only one NALⁱ CIPⁱ). The mechanisms of resistance to quinolones in the nine ETEC strains with the NAL^r CIP^r phenotype were one strain with a mutation in the amino acid codon Ser-83 of the gyrA gene; three strains with a double mutation in the gyrA and parC genes, as mentioned above; two strains with a gyrA gene mutation and the presence of the qnrS gene; and finally, three strains with the a double mutation and the aac(6′)-Ib-cr gene.

TABLE 3.

Different mechanisms of resistance to quinolones in EAEC and ETEC isolates

Quinolone resistance mechanism(s)	No. of quinolone-resistant ETEC isolates by phenotype
	Nalr Cipi-s (n = 8)	Nalr Cipr (n = 9)	Nalr Cipi-s (n = 11)	Nalr Cipr (n = 9)
gyrA mutation	8	1	10	1
gyrA + parC mutations	0	4	0	3
gyrA mutations + qnrS	0	0	0	2
gyrA + parC mutations + qnrS	0	2	0	0
gyrA + parC mutations + aac(6′)-Ib-cr	0	2	0	3
qnrS	0	0	1 (Nali Cipi)	0

DISCUSSION

ETEC and EAEC cause not only TD but also high morbidity in children in developing countries, mainly in those under 5 years of age (9). More than 50% of the patients attending the tropical medicine unit of our hospital presented TD, and antimicrobial therapy is needed due to the severity or persistence of the symptoms in around 35% of those in whom diarrhea is caused by ETEC or EAEC (10, 11). Nowadays, with the incorporation of rapid diagnostic tests based mainly on multiplex PCR, the etiology of the TD can be determined on the same working day, and therefore, a more adequate treatment can be implemented (12). Knowledge of the antimicrobial susceptibility of the most frequent etiological agents causing TD, such as EAEC or ETEC, will help in administering the most adequate treatment before the antimicrobial susceptibility of the bacteria isolated is generated.

Overall, in this study, the antimicrobial resistance rate of EAEC was slightly higher than that of ETEC, without significant differences. However, for both, the resistance to the classical and less expensive antibiotics used in developing countries, such as ampicillin, co-trimoxazole, and tetracycline, was greater than 39%. On stratifying the ETEC and EAEC strains according to the geographical area visited by the patient with TD, it was of note that the strains from Latin America were less resistant than those from Southeast Asia/India or Africa, but significance could not be calculated since the population was not large enough. This reflects the situation of antimicrobial resistance in different countries in Latin America versus Southeast Asia/India. Southeast Asia and India present high rates of resistance to the most available and inexpensive antibiotics, including quinolones, whereas in Latin America, ETEC and EAEC strains remain susceptible to these antimicrobial agents (13–19). The high prevalence of quinolone-resistant EAEC and ETEC isolates from Southeast Asia and India is worthy of mention, with rates of resistance for ciprofloxacin at higher than 40% for both EAEC and ETEC. In a previous study performed by our group during the period from 2001 to 2007, the overall percentages of resistance were 15 and 22% for nalidixic acid and were 4 and 8% for ciprofloxacin for EAEC and ETEC, respectively (20). Therefore, a significant increment (P < 0.0001 for both nalidixic acid and ciprofloxacin among EAEC strains and P = 0.0013 for nalidixic acid and P = 0.0062 for ciprofloxacin among ETEC strains) has been observed, being more dramatic in strains isolated from patients who had traveled to Southeast Asia/India, especially India, where 53.3% of the total strains were resistant to ciprofloxacin.

Four (44%) out of nine EAEC strains and two (33%) out of six ETEC strains with an MIC of azithromycin greater than 16 μg/ml showed an MIC of ≥256 μg/ml. Azithromycin reaches rectal concentrations of a mean of 133 μg/g with a single 1-g dose (21); therefore, it is above the MICs of most EAEC (89%) and ETEC (95%) strains with an MIC of azithromycin less than 256 μg/ml. The activity of rifaximin against EAEC and ETEC remained unchanged compared to previous studies (22, 23). Rifaximin is a nonabsorbable antibiotic, reaching a fecal concentration of 7,961 μg/g with a dose of 800 mg daily for 3 days (24), which is far above the MIC₉₀ that we found for EAEC and ETEC.

The main mechanisms of resistance to cefotaxime are ESBLs. Different ESBLs have been described to date, with the main types being TEM-type, SHV-type, and CTX-M-type, with the CTX-M-type being the most currently extended ESBL at a global level (25–28). Travelers have been shown to be potential carriers of the ESBL-producing Enterobacteriaceae in the intestinal tract, facilitating the dissemination of these microorganisms between countries (8, 28–33). The two most prevalent STs detected were ST38 carrying CTX-M-15 and ST131 carrying CTX-M-27. CTX-M-15-producing EAEC strains belonging to ST38 from India causing TD have been described previously, demonstrating that ST38 is a successful EAEC group (34, 35); however, in our study, this ST was also found in ETEC strains. The ST5584 ETEC strain carrying the bla_CTX-M-15 gene isolated from a traveler to China was not found among a collection of E. coli strains causing diarrhea in China, reported in a recent study (36). Different replicon type profiles (FIB/FII, K, and Y) were found in these three ST38 EAEC strains. In addition, CTX-M-15 and other types of CTX-M-producing E. coli ST38 clones, mainly ExPEC, have been detected in Saudi Arabia. In fact, some uropathogenic E. coli (UPEC) ST38 strains were also described as carrying the aggR gene, a main feature of EAEC (37). In China, the UK, Bangladesh, and Nigeria, ST38 was found among the most frequent STs in a collection of EAEC strains (38, 39).

CTX-M-27-producing E. coli ST131 strains have been described in several countries (40), and the bla_CTX-M-27 gene has been detected in EPEC and EIEC strains isolated in China (36). This bla_CTX-M-27 gene has also been detected in one EAEC strain isolated from surface water (41). However, as far as we know, its presence in ETEC strains has not been reported. In this study, both ETEC strains came from Southeast Asia/India, and the bla_CTX-M-27 gene was located in an IncF plasmid, as expected (42–44).

The main enzymatic mechanisms of E. coli associated with the acquisition of resistance to AMC include (i) hyperproduction of plasmid-mediated class A β–lactamases, such as TEM-1 and SHV-1; (ii) plasmid-mediated AmpC-type β–lactamase (p-AmpC); (iii) chromosomal AmpC β–lactamase (c-AmpC); (iv) production of inhibitor-resistant TEM (IRT) β–lactamases; and (v) plasmid-mediated β–lactamase OXA-1 (45). Among the EAEC and ETEC strains resistant to AMC, a plasmid-mediated AmpC (ACT-20) was detected only in the two strains with the highest MIC. This type of p-AmpC has previously been found in a strain of Enterobacter hormaechei isolated from dog feces (46), but so far, it has not been described in bacteria causing infections in humans. The EAEC and ETEC strains with moderate resistance to AMC presented an OXA-1 enzyme that is currently the most frequently found mechanism of resistance to AMC (47).

The acquisition of resistance to quinolones in E. coli can be either chromosome or plasmid mediated. Chromosomal mutations generating resistance to quinolones are those associated mainly with the gyrA and parC genes encoding the A subunits of DNA gyrase and topoisomerase IV, respectively, which are the protein targets of these antibacterial agents. In addition, mutations that produce an overproduction of an efflux pump or a decreased expression of a gene encoding a porin can also reduce the accumulation of the quinolone, and hence, increase resistance. The plasmid-mediated mechanisms of resistance to quinolones are related to the presence of the following three genes/gene types: (i) the qnr genes, which protect the protein target of the binding of the quinolones; (ii) the aac(6′)-Ib-cr gene, which produces the acetylation of a radical group of some quinolones generating a decrease in activity; and (iii) the qepA or opxAB genes, which are quinolone efflux pumps (48).

In this study, the NAL^r CIP^i-s phenotype shown by both ETEC and EAEC strains was mainly associated with a mutation in the gyrA gene, with the exception of one ETEC strain showing a NALⁱ CIPⁱ phenotype that did not have any mutation in the gyrA gene but presented the qnrS gene. The qnr gene was not detected in a previous study performed with ETEC and EAEC strains resistant to quinolones (20). In addition, in the present study, the NAL^r CIP^r phenotype was related to a double mutation in the gyrA and parC genes alone or together with the qnrS or aac(6′)Ib-cr gene. In a study performed in India, the main mechanisms of resistance to quinolone in ETEC were also amino acid changes in GyrA and ParC. They did not find any Qnr determinant, but 65% of the strains presented the aac(6′)Ib-cr gene (49).

In summary, our results strengthen the message that resistance to quinolones and third-generation cephalosporins has increased in EAEC and ETEC strains causing of TD, mainly in patients traveling to India and Africa, and especially sub-Saharan Africa. In addition, the ST38 and ST131 carrying the bla_CTX-M-15 and bla_CTX-M-27 genes, respectively, are highly prevalent in ESBL-producing EAEC and ETEC strains. The cephamycinase ACT-20 has also been described for the first time in EAEC and ETEC strains causing TD in patients who had traveled to Central America. The percentages of resistance to azithromycin in EAEC and ETEC isolates from patients to Southeast Asia/India and Africa are above 25%; however, the high concentration of azithromycin reached in the intestinal tract can surpass the MIC of most of azithromycin-resistant strains. Meanwhile, rifaximin is still active against EAEC and ETEC strains, and strains with an MIC of >32 μg/ml were not found. However, it is not recommended as empirical treatment for inflammatory febrile diarrhea due to its nonabsorbable nature.

The preliminary data obtained regarding the prevalence of resistance to quinolones challenge the recommendation of use of this antibiotic in the treatment of TD in patients visiting or coming from the geographical areas studied, especially India and Africa, although further studies must be done in order to elucidate the prevalence of resistance to fluoroquinolones in larger collections of EAEC and ETEC strains causing TD, as well as in other etiological agents of this infectious disease. In addition, it must be taken into account that in vitro susceptibility testing does not always correlate with lack of success in clinical practice (50).

MATERIALS AND METHODS

Bacterial isolates.

EAEC and ETEC clinical isolates causing TD were investigated in this study. The bacterial isolates were collected from 2011 to 2017. These strains were isolated from patients who were travelers and had diarrhea at the time they visited at the tropical medicine unit in our hospital. None of the patients required hospital admission. The stool samples were collected during the acute phase of diarrhea and were processed within 2 h of collection. The stool specimens were cultured for E. coli and other bacterial enteropathogens using conventional methods. Single-colony subcultures of all different colonial morphotypes growing on MacConkey agar were identified by conventional criteria (51). These colonies were tested by PCR to detect EAEC and ETEC, as described elsewhere (52).

Antimicrobial susceptibility testing.

The susceptibilities of EAEC and ETEC against ampicillin, amoxicillin-clavulanic acid, cefotaxime, imipenem, chloramphenicol, tetracycline, co-trimoxazole, nalidixic acid, and ciprofloxacin were determined by disk diffusion following the European Committee on Antimicrobial Susceptibility Testing (EUCAST) recommendations. Meanwhile, the MICs of amoxicillin-clavulanic acid, cefotaxime, cefoxitin, and azithromycin were determined by the Etest method, and the MIC of rifaximin was obtained using the microdilution method according to EUCAST guidelines (53). E. coli ATCC 25922 and E. coli ATCC 35218 were used as controls. The Clinical and Laboratory Standards Institute (CLSI) and EUCAST breakpoints were used to define resistance to nalidixic acid and ciprofloxacin, respectively. The breakpoints of azithromycin considered were those described by EUCAST for Salmonella enterica serovar Typhi (MIC, ≤16 μg/ml for wild-type isolates).

Detection of β–lactam and quinolone resistance mechanisms.

A double-disk synergy test was carried out in the cefotaxime-resistant isolates in order to confirm the extended-spectrum β-lactamase (ESBL) carriage (54). The detection of ESBL genes (bla_CTX-M, bla_SHV, bla_OXA, and bla_TEM genes) was carried out by PCR and DNA sequencing under previously described conditions (55). In addition, strains resistant to amoxicillin-clavulanic acid were tested for cefoxitin susceptibility by Etest in order to confirm the presence of cephamycinases, which were also detected by PCR and DNA sequencing using specific primers, as described previously (56).

To determine the quinolone resistance mechanisms, mutations in the quinolone resistance-determining region of the gyrA and parC genes were detected by PCR, and sequencing was performed as described elsewhere (57, 58). The purified PCR products visualized in gels were processed for DNA sequencing and analyzed in an automatic DNA sequencer (ABI 377; PerkinElmer, Emeryville, CA, USA) using the BigDye Terminator cycle sequencing kit (version 3.1; PerkinElmer). Detection of the qnr genes screening for the qnrA, qnrB, qnrC, qnrD, and qnrS genes was performed by multiplex PCR using a combination of specific primers (59). Bacterial strains positive for each qnr gene were used as positive controls and were run in each batch of samples tested. Detection of the aac(6′)-Ib-cr gene was performed using specific primers described previously (60).

Plasmid typing and MLST.

Replicon typing was then performed in the strains carrying the bla_CTXM genes to determine the potential plasmids carrying this resistance gene; the primers used were designed by Carattoli et al. in 2006 (42) but adapted amplification protocols for commensal and pathogenic E. coli isolates described by Johnson et al. were employed (61), or the PCR-based replicon typing kit was used (Diatheva, Cartoceto, Italy). In the same set of strains, the multilocus sequence typing (MLST) was determined analyzing by amplification seven housekeeping genes (adk, fumC, icd, purA, gyrB, recA, and mdh) (62). The database available at https://enterobase.warwick.ac.uk/ was used for assigning sequence types (STs) and clonal complexes (CCs). Strains carrying plasmids from incompatibility group IncF were further characterized following the plasmid MLST website (https://pubmlst.org/plasmid/) protocol, developed by Keith Jolley and sited at the University of Oxford (63).

Statistical analysis.

Data of the present study are presented as frequencies. The prevalence of resistance was compared to previous data using the binomial test (64). Proportions were compared using a chi-square test or Fisher’s exact test if the application conditions of the chi-square test were not met. Significance was set at P < 0.05. The analysis was carried out using Stata (65).