Prevalence of antimicrobial resistance and integrons in Escherichia coli from Punjab, Pakistan.

Antimicrobial resistance was studied in Escherichia coli strains isolated from urine samples of 457 patients suffering from urinary tract infection. High prevalence of class 1 integrons (43.56%), sulfamethoxazole resistance genes sul1 (45.54%) and sul2 (51.48%) along with occurrence of quinolone resistance genes was detected in multi drug resistance isolates.

Urinary tract infections (UTIs) are the most common bacterial infections and Escherichia coli is the most prevalent cause of UTIs (12, 15). Empirical treatment of UTIs in the developed countries is based on the susceptibility data originated from the local surveillance schemes (1, 7, 8). However, countries like Pakistan do not have any surveillance system to provide reference guidelines for empirical therapy. Thus errors are expected in most of the estimates for empirical therapy for UTIs in such conditions.

Integron system is a dynamic force in the evolution of multidrug resistance (MDR) and it helps bacteria to acquire novel combinations of resistance genes (6). Integrons are horizontally transferable genetic elements which play an important role in dissemination and accumulation of resistance genes in bacteria (15). A strong association between the presence of integrons and antimicrobial resistance has been established by recent studies on Uropathogenic E. coli (UPEC) in various regions of Asia, Europe and USA (1, 6, 13, 15, 17). However, the prevalence of integrons and related gene cassettes in E. coli strains isolated from UTI patients has not been previously reported from Pakistan.

The aim of this study was to determine the prevalence of antimicrobial resistance, integrons and related genetic elements among E. coli isolates from UTI patients from Pakistan. The study was conducted for duration of 1 year (June 2006- June 2007).

The prior approval was obtained from departmental ethical committee. Medical history of the participants was obtained with the help of a questionnaire after obtaining a signed informed consent from all the individuals who participated in the study. Clinically diagnosed UTI cases (n = 457), hospitalized in 11 different hospitals located mainly in 3 different cities of Punjab, were included in this study. Diagnoses was based on clinical history, physical examination, and urine tests including urine microscopy within first 1 hour of urine sample collection and subsequent two consecutive urine cultures. The median age of 280 (61.27%) female and 177 (38.73%) male patients was 40 and 45 years respectively. Exclusion criteria was as follows: (A) age < 15 years (B) hospital-acquired infection; (C) hospitalization for more than 48 hours before taking urine samples; (D) patients with permanent indwelling catheters; (E) patients with history of previous use of urinary catheter. However, cases with first time short term use of catheters after admission to the hospital were included and urine samples were taken from such cases within 6 hours of first catheterization.

All urine samples were collected before starting empirical antimicrobial therapy. Urine samples were collected from non-catheterized patients by midstream clean-catch method. Urine samples were collected directly from the catheter tubings into the sterile syringes in case of catheterized patients. All urine samples were processed on MacConkey agar medium (Oxoid, UK) with a standard wire loop and incubated at 37°C overnight. Significant growth was evaluated as ≥105 colony forming units (CFU)/ml for midstream urine, and ≥102–3CFU/ml for urine samples from catheterized patients (11). Urine samples not complying with the criteria of significant growth were excluded. Identification of E. coli was performed by standard biochemical methods. Bacterial isolates other than E. coli were excluded. Antimicrobial susceptibility of E. coli isolates was determined using standard disk diffusion techniques and strains were classified as susceptible (S), intermediate (I) and resistant (R) according to recommendations of CLSI (Clinical Laboratory Standards Institute) (5). Susceptibility to trimethoprim, sulfamethoxazole-trimethoprim, ampicillin, cefotaxime, ceftazidime, nitrofurantoin, nalidixic acid and ciprofloxacin was determined.

DNA was extracted from E. coli isolates by Phenol-Chloroform extraction method (9). All E. coli isolates were screened for the presence of class 1, 2 and 3 integrons with the help of PCR-RFLP method (17). Integrase PCR products were digested using Hinf1 according to the manufacturer’s instructions (Fermentas). Integrons were classified on the bases of the sizes of restricted fragments after treatment of hep amplified PCR products with Hinf1 (491bp fragment for class 1 integrons, 300 bp and 191 bp fragments for class 2 integrons, while 119 bp and 372 bp fragments for class 3 integrons) (17). Sulfamethoxazole resistance determinants sul1 and sul2 were identified by PCR amplification using gene specific primers as described previously (1). Ciprofloxacin and / or nalidixic acid resistant strains were screened for the plasmid associated quinolone resistance genes. Plasmids (extracted by using Miniprep (Qiagen) plasmid extraction kit) were used as templates to amplify internal fragments of the qnrA (580 bp), qnrB (264 bp) and qnrS (428 bp) (3).

A total of 101 non-repetitive E. coli isolates (representing 7–10 isolates from each hospital) were isolated from UTI patients. The results indicate a high incidence of antimicrobial resistance along with related genetic elements among E. coli isolates from UTI patients. In order of descending prevalence, 89 (88.12%) of 101 E. coli isolates were resistant to sulfamethoxazole-trimethoprim, 87 (86.14%) were resistant to trimethoprim, 85 (84.16%) were resistant to ceftazidime, 85 (84.16%) were resistant to nalidixic acid, 79 (78.22%) were resistant to ampicillin, 79 (78.22%) were resistant to nitrofurantoin, 77 (76.24%) were resistant to cefotaxime, and 35 (34.65%) were resistant to ciprofloxacin. Majority of isolates were resistant to ≥3 antimicrobials tested. Our observed antimicrobial resistances were either similar or higher than the resistance levels previously reported in various other regions. Sulfamethoxazole-trimethoprim resistance (88.12%) in this study was similar to previously reported levels (up to 90 %) among UPEC isolates (12). The prevalence of trimethoprim resistance in our study (86.14%) is clearly higher than resistance levels reported in the previous studies (25% to 68%) in Asia, Africa and South America (10). Our observed trends in sulfamethoxazole resistant allele distributions show that sul2 is the most common resistance gene (sul2 > sul1) in sulfamethoxazole-trimethoprim resistant isolates. The prevalence of sul1 & sul2 was 45.54% (n = 46) and 51.48% (n = 52) respectively among 101 E. coli isolates.

A high level of antimicrobial resistance (84.16%) against nalidixic acid was observed during current study compared to a resistance level of 26.5% reported by Siddiqui (14). However, resistance was low against ciprofloxacin (34.65%) in our study as compared to 95% resistance to fluoroquinolones in enterobacteriaceae strains reported by Fluit et al. in Europe (7). However, Goldstein (8) reported even lower level of fluoroquinolones resistance in E. coli (<10% resistance) in France as compared to our study. Our observed prevalence of plasmid associated quinolone resistance genes, qnrB, qnrA and qnrS, was 4 (4.70%), 3 (3.53%) and 3 (3.53%) respectively out of 85 nalidixic acid resistant strains. Similarly, 2 (5.71%), 1 (2.86%) and 2 (5.71%) of 35 ciprofloxacin resistant strains were positive for qnrB, qnrA and qnrS respectively. This study constitutes the first report of quinolone resistance determinants among Uropathogenic E. coli isolates from Pakistan and the results are in consistent with recent studies from Europe and Asia (4, 16).

Nitrofurantoin is one of the most appropriate antibacterial agents for empirical therapy of UTIs because it is highly concentrated in the urine and it is administered orally. However, the resistance level in our isolates (78.22%) was significantly high against this antimicrobial as well. Our observed resistance level against cefotaxime was 76.24% whereas 84.16% of isolates were resistant to ceftazidime which is very high level of resistance as compared to up to 5% resistance levels reported in some developed countries for cefotaxime and ceftazidime (8). Susceptibility to penicillin is also decreasing globally (8). We observed a high level of resistance (78.22%) against ampicillin. Saffer et al. (13) reported a similarly high level of resistance (82–100%) against ampicillin & similar trend (100% ampicillin resistance) was described by Nicolle (12). E. coli isolates sensitive to penicillins and / or cephalosporins were not screened in the current study for detection of ESBL (Extended Spectrum Beta Lactamases) production. This is a limitation of the current study as E. coli isolates producing ESBL are classified as resistant to these antimicrobials even if the in vitro tests indicate susceptibility (2, 3). We will further investigate the prevalence of ESBL in these isolates in future study.

The presence of integrons in E. coli isolates is a serious risk factor for spreading antimicrobial resistance. The prevalence of integrons ranging from 22% to 59% has been reported in clinical E. coli isolates in several previous studies on UPEC in Europe and Asia (6, 15). We also observed a similar trend of prevalence of class 1 integrons (44 (43.56%) of 101 isolates) in our study. Arrangement of various genotypes’ combinations revealed that the independent occurrence of int1 among the studied samples was relatively rare as evident by a small number, i.e., 4 (3.96%) of 101 isolates (Table 1). However relatively higher number possessed sul2 without int1 and sul1 (23 (22.77%) of 101 isolates) as compared to occurrence of sul1 without int1 and sul2 (4 (3.96%) of 101 isolates). Moreover, the occurrence of class 1 integrons was noted in almost all those isolates which were positive for sul1. These results are in accordance with the previous studies (15). The occurrence of a significant number of sul2 alleles was independent of the presence of integrons. Perhaps these sul2 alleles are carried on transposons and / or plasmids which may also transfer them horizontally (1). On the other hand no indication of presence of class 3 integrons was noted among the studied isolates, while class 2 integrons were detected in 3 (2.97%) isolates (Table 1) as compared to 5 to 15 % prevalence of class 2 integrons reported from other regions (15).

Table 1

Resistance gene combinations (of sul1, sul2, int1, int2, int3) found in 101 E. coli isolates cultured from UTI patients’ urine samples.

Gene combinations	Frequency (%)
int1 alone	4 (3.96%)
int1 + sul1	13 (12.87%)
int1 + sul2	0 (0%)
int1 + sul1+ sul2	27 (26.73%)
int2 alone	3 (2.97%)
int2 + sul1	0 (0%)
int2 + sul2	0 (0%)
int2 + sul1+ sul2	0 (0%)
int3 alone	0 (0%)
sul1 alone	4 (3.96%)
sul2 alone	23 (22.77%)
sul1 + sul2	2 (1.98%)

The results indicate that antimicrobial resistance level in E. coli isolates from UTI patients in Pakistan is relatively higher for most of the antimicrobials tested compared to the developed countries. The data also suggested that urine cultures and susceptibility tests can not be neglected in order to avoid worrisome trend of development of resistance to most commonly used antimicrobials for treatment of UTIs. The prevailing trend of co-occurrence of class 1 integrons and antimicrobial resistance genes in our study is an additional threat for spread of the antimicrobial resistance traits which may further complicate future strategies for empirical therapy of UTIs.