Background
The worldwide dissemination of plasmid-borne extended-spectrum beta-lactamases (ESBLs) in
Enterobacteriaceae has become a major global public health problem [
1]. Inappropriate use of antibiotic is the predominant selection pressure for antibiotic resistance and aids horizontal transfer of bacterial resistance [
2]. The spread of multi-resistant strains of bacteria in the community is compounded with a paucity of new classes of antibiotics effective against Gram-negative bacteria like
Escherichia coli[
3].
E. coli is the most common bacterial cause of urinary tract infections (UTI’s) in women and also a major cause of catheter induced UTI, surgical site infections, intra-abdominal infections and sepsis in all age groups and meningitis in neonates [
4]. The recent spread of New Delhi metallo-beta-lactamase-1 (NDM-1) producing bacteria emphasises the fact that resistant bacteria respect no geographical boundaries [
5].
E. coli forms part of the bacteria commensal flora of the human gut. It has been identified as the predominant reservoir of antibiotic resistance genes [
2]. Once acquired, these resistance genes are stable and are easily transferable to pathogenic bacteria [
6]. The published studies on antibiotic resistance in commensal
E. coli in resource-poor settings are limited as resistance monitoring is resource intensive, requires dedicated staff and a well-equipped laboratory [
7‐
9].
During pregnancy many anatomical and physiological changes occur, which also affect the urinary tract. Asymptomatic bacteriuria is reported to occur in 2–10% of pregnancies and
E. coli causes 70–80% of them [
10]. Most infections are endogenous in origin [
10]. Asymptomatic bacteriuria can have serious consequences for the foetus and/or mother [
10]. However, knowledge about the risk factors for carriage of resistant
E. coli in the women of reproductive age group is lacking in general and especially from India. Screening the commensal
E. coli in antenatal women for antibiotic resistance pattern will provide guidelines for empiric therapy especially for sick new-borns. Molecular analysis of these isolates will provide prevalence of antibiotic resistant genes present in
E. coli in the community. The aim of the study is to ascertain the factors associated with the carriage of resistant commensal
E. coli in pregnant women and further to characterize the genotypic determinants of antibiotic resistance in them.
Discussion
In the present study it was found that out of the 710 E. coli isolated from women attending antenatal clinics in Ujjain, India 94% were resistant to at least one of the antibiotic groups tested. The statistically significant factors associated with carriage of MDR E. coli were better education, overcrowding (family size more than 10 members) history of antibiotic use and/or hospitalization in the last four weeks.
Quinolone resistance is usually chromosomally mediated [
20]. All 35 multi-resistant isolates showed presence of
gyrA and
parC genes, which are chromosomally mediated. The
aac(6
′
)-Ib-cr gene accounted for 95% of the PMQR genes. Similar findings in pathogenic
E. coli have been reported elsewhere in India [
21] and in
E. coli isolated from hospital wastewater in our setting [
22]. Plasmid mediated resistance to cephalosporins was largely due to
bla
CTX-M -15 which is in keeping with other studies done in India [
21,
23]. The
bla
TEM-1 is not been subtyped therefore no comment can be made for its co-relation with ESBL production. It is interesting to note that
bla
SHV was not detected. The presence of genes coding for extended spectrum of beta lactamases and plasmid mediated quinolone resistance in commensal
E. coli is disconcerting.
Studies from India [
7,
24] and other low-middle income countries [
25] have reported lower rates of resistance (range of 32 to 63%) to a single antibiotic compared to the present study (94%). The difference in the resistance rates could be due to variation in the geographical area of the study [
7,
8,
26] and also is dependent on the pattern of antibiotic prescribing in the local community [
27,
28]. The ESBL rate (15%) in the present study is lower than that reported among pathogenic
E. coli (69%) isolated from patients with clinical infections in the same geographical area [
29].
The majorities of women enrolled in the study were either illiterate or had only a primary education (Table
1). Women with higher education had a higher likelihood of carriage of MDR
E. coli. This could be a chance association and needs further validation. But, women with higher education may also be socio-economically better off and thus more likely to seek health-care and consequently have a greater exposure to antibiotics [
30]. Most self-employed women enrolled in the study were daily wage workers, and consequently, probably, had limited opportunity to seek medical care, limiting their antibiotic exposure.
Exposure to hospital environment especially hospitalization in intensive care units is a major risk factor for carriage of multi-drug resistant bacteria especially in resource poor settings where hospitals can have high infection rates and spread of multi-drug resistant pathogens [
25,
31]. A multitude of factors including poor infrastructure of hospitals, low compliance with hand-hygiene, heavy workload with understaffing, overcrowding, lack of or poorly functioning infection control programme contribute to the problem [
31].
In the present study a significantly higher MDR carriage among women living in larger families (more than 10 members), OR 4.97 (95% CI 1.51 to 16.33; P = 0.008) was found. Similar results have been reported in postmenopausal women in our setting [
32]. Over-crowding combined with inadequate hygiene and sanitary facilities and sewage disposal, would lead to greater sharing of commensal flora in families and communities [
8,
32]. In a study among primary school children in rural Tamil Nadu over crowding in classrooms was associated with increased carriage of antibiotic resistant commensal
E. coli[
24]. Thus, the dynamics of bacterial transmission due to over crowding is worthy of further investigation.
The rate and amount of antibiotic use in community an important determinant of increasing antibiotic resistance in bacteria [
27,
28]. A meta analysis and review on the effect of antibiotic prescribing in a primary-care setting on antimicrobial resistance in individual patients included five studies of urinary tract infections (14348 participants), the pooled odds ratio (OR) for resistance was 2.5 (95% CI 2.1 to 2.9) within two months of antibiotic treatment [
33]. The OR at one month (only one study) was 6.1 (95% CI 2.8 to 13.4) [
33]. Recent antibiotic use was associated with increased risk of carriage of resistant
E. coli, OR 1.8 (95% CI 1.5 to 2.3) in a study from Indonesia [
25].
The ESBL production is the commonest mechanism of resistance to cephalosporins in
Enterobacteriaceae[
2]. Phenotypic detection of ESBL production was in 15% isolates in the present study, which is much lower than the 51% reported in a study from China [
34].
The
bla
CTX-M-15 and
bla
TEM −1 coded genes are widespread in
E. coli[
2]. The detection of
bla
CTX-M-15 and
bla
TEM-1 genes as shown in Figure
1 is similar to that in Latin America [
9]. In the present study it has been noted that a majority of MDR
E coli isolates (48%) were carrying
aac(6
′
)-Ib-cr gene. These isolates also showed high MIC to ciprofloxacin and ofloxacin probably due to chromosomally mediated resistance (
par A and
par C) (Table
3). Co-residence of
bla
CTX-M-15,
qnr and
aac(6
′
)-Ib-cr genes on the same plasmid has been reported elsewhere [
35].
PMQR genes code for low levels of quinolone resistance [
36]. The presence of PMQR gene in MDR isolates was 4%, which is similar to that in a study from Republic of Korea [
16]. It is interesting to notice that chromosomal mediated quinolone resistance coded by mutation in QRDR regions of
gyr A and
par C genes was present in all MDR
E coli isolates in the present study. A study in Ghana on quinolone-resistant
E. coli in the faecal flora also reported the presence of
qnr genes and amnio acid substitutions at
gyrA and
par C genes [
37].
An analysis of resistance pattern of commensal
E. coli can be helpful in deciding empiric treatment of choice for suspected neonatal infections. Neonatal sepsis with multi-drug resistant gram negative organisms resistant to ampicillin, third generation cephalosporins and gentamicin caused high mortality (26%) in neonates in a neonatal intensive care unit (NICU) in India [
38]. The current choices of therapy in our NICU are in line with the most sensitive patterns described ie third generation cephalosporins along with an aminoglycoside (Table
2). However, the last resort antibiotics (eg. vancomycin and imipenem) are now increasingly being used in view of increasing resistance [
29] in our NICU to salvage septic neonates.
The most important strength of this study is the combination of epidemiological and molecular methods for the identification of factors associated with carriage of resistant bacteria. Antenatal women are a population, which is at a low risk for receiving antibiotics. Despite this a history of antibiotic therapy was identified as an important risk factor associated with carriage of resistant bacteria. The fact underscores the importance of prudent antibiotic use. A structured questionnaire was used which could have resulted in us missing other factors related to carriage of resistant E. coli like environmental factors and food chain related factors.
Competing interests
The authors declared that they have no competing interests.
Authors’ contributions
AP, SC, KM, RM and CSL participated in the conception and design of the study. KM and AP collected data in the field. SC and RM carried out the molecular genetic studies, participated in the sequence alignment. AP and SC performed the statistical analysis and drafted the manuscript. CSL coordinated the study. AP, SC, KM, RM and CSL revised the paper critically for substantial intellectual content. All authors read and approved the final manuscript.