Antimicrobial Resistance in Fecal Escherichia coli Isolates from Healthy Urban Children of Two Age Groups in Relation to Their Antibiotic Therapy

Abstract

The study was performed in the Czech Republic during 2007 to 2009. Of Escherichia coli isolates from 275 children aged 6 weeks, 36% (n = 177) were resistant to 1 to 7 antibiotics. Of isolates from 253 children aged 6 to 17 years, 24% (n = 205) were resistant to 1 to 5 antibiotics. There was no significant difference in the prevalences of antibiotic-resistant E. coli isolates between these groups of children, even though the consumptions of antibiotics were quite different.

TEXT

Since the introduction of antibiotics, the resistance of human intestinal flora has become a common finding. In 1967, nonpathogenic strains of Escherichia coli with multiple transferable antibiotic resistances were isolated from the feces of a small number of infants in Dublin, Ireland, who had not been in hospital and had not been treated with antibiotics (7). From that time, the presence of antibiotic-resistant E. coli isolates among children has been studied in various countries, and it was found that E. coli strains colonizing healthy children in communities may be resistant far more often in some regions than in others. The aim of our work was to characterize and analyze antibiotic resistance in fecal E. coli strains isolated from two groups of urban children differing in age while also comparing the antibiotic practices applied for these children.

The two groups of healthy children examined were from the town of Karvina in the Czech Republic. Karvina has about 60,000 inhabitants living in an urban habitat. One group consisted of 275 children aged 6 weeks, and the other group consisted of 253 children aged 6 to 17 years. The children were examined during 2007 to 2009. Samples were taken by individual rectal swabs, which were transported to the laboratory and placed overnight in buffered peptone water at 37°C. The samples were then cultivated for E. coli on a chromogenic medium selective for E. coli and coliform bacteria (CM0956; Oxoid, United Kingdom). One colony of each plate was tested for susceptibility to 12 antimicrobial agents by the disk diffusion method as described elsewhere previously (5). For E. coli isolates found to be resistant to one or more of the antibiotics, PCR was used to detect specific antibiotic resistance genes, the integrase genes int1 and int2, variable regions of class 1 and class 2 integrons, and gene cassettes within class 1 and 2 integrons, as described elsewhere previously (5). The samples from buffered peptone water were enriched in MacConkey broth and subcultivated onto MacConkey agar (MCA) containing cefotaxime (2 mg liter⁻¹) to detect E. coli strains with ex-tended-spectrum beta-lactamase (ESBL) and subsequently onto MCA supplemented with ciprofloxacin (0.06 mg liter⁻¹) to isolate fluoroquinolone-resistant E. coli isolates. These E. coli isolates were further characterized; specifically, the transferabilities of their ESBL and plasmid-encoded quinolone resistance genes was tested by conjugation as described elsewhere previously (1, 2). Plasmids of ESBL and quinolone-resistant strains were characterized by replicon typing and EcoRV digestion (1, 2). All E. coli strains were identified by using the API 10S test (bioMérieux, France).

The antibiotics given to the children throughout their lives until the time of examination were searched for by using the archives of the proper pediatricians. The pediatricians in the Czech Republic are obligated to record and archive all therapies used for children. The term “antibiotic application” was defined as one administration of a usual dosage of some antibiotic prescribed for a therapeutic or prophylactic purpose to one child. The impact of antibiotics was expressed as the total number of antibiotic applications for those children examined.

A total of 177 E. coli isolates were obtained from 275 children aged 6 weeks. Sixty-three (36%) isolates were resistant to 1 to 7 antibiotics (Table 1). From 253 children aged 6 to 17 years, a total of 205 E. coli isolates were obtained. Forty-nine (24%) isolates were resistant to 1 to 5 antibiotics. There was no significant difference in the total prevalences of antibiotic-resistant E. coli isolates between children aged 6 weeks and children aged 6 to 17 years. The prevalences of antibiotic-resistant E. coli for individual antibiotics were significantly higher for children aged 6 weeks than for children aged 6 to 17 years (applicable for ampicillin, trimethoprim-sulfamethoxazole, cephalothin, and nalidixic acid), or the differences were not significant.

Table 1.

Prevalences of antibiotic-resistant Escherichia coli isolates from urban children of two age groups

Antibiotic	No. of resistant isolates (prevalence [%]) in children aged:		P valuea
	6 wk (n = 177)	6–17 yr (n = 205)
Ampicillin	50 (28)	34 (17)	0.0061
Tetracycline	32 (18)	25 (12)	NS
Sulfonamides	29 (16)	21 (10)	NS
Streptomycin	24 (14)	20 (10)	NS
Sulfamethoxazole-trimethoprim	13 (7)	6 (3)	0.0476
Cephalothin	10 (6)	2 (1)	0.0090
Nalidixic acid	9 (5)	3 (1)	0.0430
Chloramphenicol	7 (4)	4 (2)	NS
Amoxicillin-clavulanic acid	4 (2)	2 (1)	NS
Ciprofloxacin	1 (1)	0 (0)	NS
Gentamicin	1 (1)	0 (0)	NS
Ceftazidime	1 (1)	0 (0)	NS

^aDifferences in prevalences were compared by using the chi-square test and Fisher's exact test. Statistical significance is indicated at a P value of ≤0.05. NS, not significant.

The antibiotic-resistant E. coli isolates from both groups of children were highly variable in their resistance patterns. The bla_TEM, bla_SHV, and bla_CTX-M-1 genes were detected in isolates with ampicillin resistance. The only proven tetracycline resistance determinants among E. coli isolates resistant to tetracycline were the tetA and tetB genes. The sul1, sul2, and sul3 genes were found among the isolates resistant to sulfonamides. Resistance to streptomycin was linked to the strA gene and the aadA1, aadA2, and aadA5 gene cassettes. Resistance to chloramphenicol was linked to the cat and cmlA genes. Various types of class 1 integrons (with the gene cassettes aadA1, aadA2, dfr17, dfr1-aadA1, dfr17-aadA5, and dfr12-orf-aadA2) and class 2 integrons (with sat-aadA1, dhfr1-sat-aadA1, and estX-sat-aadA1) were identified among the antibiotic-resistant E. coli isolates.

Two ESBL-producing E. coli isolates were obtained from one 6-week-old child and one 11-year-old child. Both isolates contained the bla_CTX-M-1 gene. The bla_CTX-M-1 genes were localized on a 30-kb conjugative plasmid of the IncN group along with the tetracycline resistance gene tetA and on a 95-kb conjugative plasmid of the IncI1 group along with the class 1 integron containing the dfr1-aadA1 gene cassettes and the cib gene for colicin production, respectively. Another E. coli isolate, from a 10-year-old child, selected using MCA with ciprofloxacin had the plasmid-mediated quinolone resistance gene qnrS1. The qnrS1 gene was transferred by conjugation and was located on a 45-kb plasmid of the IncX1 group.

Only five antibiotic applications were recorded for children at the age of 6 weeks (Table 2). A total of 1,965 antibiotic applications were recorded for children aged 6 to 17 years. Beta-lactam antibiotics were used most frequently, followed by macrolides and sulfonamides with trimethoprim. The use of other antibiotics was exceptional.

Table 2.

Antibiotics given to the children examined during their lives up to the time of examination

Antibiotic	No. of antibiotic applications for children aged:
	6 wka	6–17 yr
Beta-lactams
Penicillins (penicillin, penamecillin, phenoxymethylpenicillin, oxacillin)		491
Aminopenicillins (ampicillin, amoxicillin)		311
Aminopenicillins with beta-lactamase inhibitors (ampicillin + sulbactam, amoxicillin + clavulanic acid)	3	376
Narrow-spectrum cephalosporins (cephalexin, cefadroxil)		59
Expanded-spectrum cephalosporins (cefaclor, cefprozil, cefuroxime, including axetilcefuroxime)		200
Broad-spectrum cephalosporins (cefotaxime, ceftazidime, ceftriaxone)		5
Macrolides (erythromycin, spiramycin, roxithromycin, clarithromycin, azithromycin)	1	387
Sulfonamides, trimethoprim (sulfamethoxazole + trimethoprim, trimethoprim)		122
Aminoglycosides (neomycin, gentamicin, netilmycin)	1	5
Others (clindamycin, nitrofurantoin, vancomycin, bacitracin)		9
Total	5	1,965

^aOne antibiotic application means an application of a usual therapeutic dose of some antibiotic prescribed during a child's single therapeutic situation.

Commensal E. coli strains in children represent an important reservoir of antibiotic resistance determinants. Relatively low prevalences of antibiotic-resistant E. coli were observed for the Czech children compared to the findings from a large number of other studies on this topic, and they were in good agreement with results of one recent study from Germany (4).

At 36%, we consider the high prevalence of antibiotic-resistant E. coli strains in Czech children aged 6 weeks to be disturbing. These children had scarcely been exposed directly to antibiotics, and yet they were frequently colonized by antibiotic-resistant E. coli. The range of sources for antibiotic-resistant E. coli is very limited for this age group of children. The fetal intestine is sterile and bathed in swallowed amniotic fluid. Following delivery, multiple bacteria challenge the intestine of the newborn (3). Microbes from the vaginal canal and perineum enter the mouth and stomach of vaginally delivered infants, and within few minutes after birth, the gastric content of the newborn is influenced by and reflects the cervical flora of the mother. All children in our study were born in maternity hospitals with high hygienic standards. Such deliveries proceed with a physician in attendance, and during the time important for bacterial colonization of the newborns' gastrointestinal tracts, the children are in nearly exclusive contact with their mothers. Mothers and their children usually leave the maternity hospitals in the Czech Republic when the child is 4 days old, and the child must be colonized primarily at that time. We strongly believe that the sources for the colonization of 6-week-old children with antibiotic-resistant E. coli isolates were their mothers.

Among the children aged 6 to17 years surveyed, there was a 24% prevalence of antibiotic-resistant E. coli, even though the children had been treated with antibiotics during their lifetimes a total of 1,965 times. The prevalences of E. coli isolates resistant to tetracycline, streptomycin, and chloramphenicol were 12%, 10%, and 2%, respectively, even if no such antibiotics had been administered to these children. In the United Kingdom as well, healthy children 8 to 9 years old were shown to carry bacteria resistant to antibiotics to which children were not usually exposed (6). The authors of that study considered that bacteria resistant to ceftazidime, chloramphenicol, and tetracycline may be coselected by exposure to other antibiotics used for children or may be acquired from family members, pets, other children, or food.

The E. coli isolates harboring the bla_CTX-M-1 gene, responsible for the production of ESBL, and the qnrS1 gene have been documented for the first time in children anywhere in the world. Various associations among the plasmids and beta-lactamase genes have been observed. The associations between bla_CTX-M-1 and IncN plasmids, bla_CTX-M-1 and IncI1 plasmids, and qnrS1 and IncX1 plasmids have been observed recently in the Czech Republic among human and animal E. coli isolates (1, 2).