Tropical Medicine and International Health doi:10.1111/j.1365-3156.2007.01969.x volume 13 no 1 pp 41 45 january 2008 Antimicrobial resistance surveillance among commensal Escherichia coli in rural and urban areas in Southern India Elizabeth Mathai 1, Sujith Chandy 2, Kurien Thomas 3, Belavendra Antoniswamy 4, Inbakumar Joseph 5, Matthews Mathai 6, Thomas L. Sorensen 7 and Kathleen Holloway 8 1 Department of Clinical Microbiology, Christian Medical College and Hospital, Vellore, India 2 Department of Clinical pharmacology, Christian Medical College and Hospital, Vellore, India 3 Department of Medicine, Christian Medical College and Hospital, Vellore, India 4 Department of Biostatistics, Christian Medical College and Hospital, Vellore, India 5 Department of Rural Unit for Health and Social Affairs, Christian Medical College and Hospital, Vellore, India 6 Department of Obstetrics and Gynaecology, Christian Medical College and Hospital, Vellore, India 7 Formerly Department of Communicable Disease Surveillance and Response, WHO, Geneva, Switzerland 8 Department of Medicines, Policy and Standards, WHO, Geneva, Switzerland Summary objective To assess antimicrobial resistance (AMR) in Tamil Nadu, India. methods Data on AMR of commensal and uropathogenic Escherichia coli were collected from one urban (Christian Medical College Hospital, Vellore) and one rural (CMCH Rural Unit for Health and Social Affairs) centre in Tamil Nadu at monthly intervals for 1 year. results Forty-two per cent of commensal E. coli was resistant to one or more of the tested antimicrobials. 8.4% were resistant to three drugs commonly used for the treatment of urinary tract infections, namely ampicillin, co-trimoxazole and nalidixic acid. 1.5% of isolates were resistant to nitrofurantoin. There was no significant difference between resistance rates in commensal E. coli collected in rural and urban areas. Resistance was more common in infecting than commensal strains. discussion Resistance to most antimicrobials is high both in urban and rural areas. Higher resistance to antimicrobials used widely for the treatment suggests that drug use contributes to it. Hence unnecessary use of antimicrobials must be avoided. Surveillance among commensal E. coli can be used to monitor changes in AMR over time. keywords Escherichia coli, antimicrobial resistance, Tamil Nadu Introduction Antimicrobial resistance (AMR) in bacteria causing common community acquired infections is increasing globally. Its impact on morbidity, mortality and cost of treatment is likely to be pronounced in developing countries where the burden of infectious diseases is higher and the population economically poorer. Irrational antimicrobial use is probably a major contributor to rising AMR (WHO 2001). However, there is inadequate community based data linking the two, and there is the possibility that AMR develops locally in response to selection pressure of antimicrobial agents to spread to distant areas (O Brien 2002). Community based surveillance could be used to scrutinize antimicrobial use and resistance, to evaluate temporal correlation between resistance and use, and to evaluate the effect of interventions to reduce resistance. Escherichia coli, a major part of the commensal flora of the gut, has the potential to cause a variety of infections. Commensal E. coli can act as reservoirs of resistance genes that easily transfer to other commensal E. coli, as well as other potential pathogenic bacteria (Leverstein-van Hall et al. 2002; Livermore 2003). Faecal E. coli therefore has been used as an indicator organism for AMR (Lester et al. 1990; Murray et al. 1990; Okeke et al. 2000; Bartoloni et al. 2006a). Escherichia coli is also the single most common bacterium causing urinary tract infections (UTI) (Gupta et al. 2001; Sahm et al. 2001). UTI is a frequent indication for antimicrobial therapy and hence a potential area for misuse contributing to development and spread of resistance in general, and in particular in E. coli, both pathogenic and commensal (Fluit & Schmitz 2001). Resistance to older generations of antimicrobials is high in most areas, and resistance to most newer antimicrobials has appeared in community acquired uropathogens (Sahm et al. 2001; Turnidge et al. 2002; Farrell et al. 2003; Karlowsky et al. ª 2008 Blackwell Publishing Ltd 41
2006). Inadequate data on local resistance patterns can lead to wrong choices of antimicrobial therapy and increase development of resistance. There is very little information on the rate of development of resistance in E. coli in different areas and the groups of drugs that can be used for interventions. Putative differences in resistance patterns between commensal and uropathogenic E. coli remain to be elucidated. We therefore attempted to establish a surveillance system for monitoring AMR in E. coli in order to compare it with antimicrobial use in southern India. Methods Surveillance data on AMR of E. coli were collected from one urban and one rural area in Tamil Nadu, south India. Christian Medical College Hospital (CMC), Vellore was the urban centre and its Rural Unit for Health and Social Affairs situated about 30 km away from the main hospital was the rural centre. We obtained 45 isolates from each area every month as monthly data points to understand seasonal and other variations. To detect a change of 20% over 24 months, with 80% power and 5% level of significance, the minimum required sample was 960. Therefore approximately 40 samples were collected from each centre every month, a number feasible in terms of cost and sufficient to give reasonably precise point estimates each month to measure trends in resistance of E. coli. Commensal E. coli were used as the indicator organisms in the surveillance, based on the assumption that resistance trends in these E. coli would reflect differences in the use of antimicrobials in the two settings and could also act as markers for resistance trends in the pathogenic E. coli. Using commensal E. coli ensured that the required numbers were obtained every month. These E. coli were collected from pregnant women because they form a healthy population regularly attending the clinics; thus the sample was representative of the community. Pregnant women are less likely to take antimicrobials and hence the commensal flora obtained from these women is not likely to be under selection pressure. Information on prior antimicrobial use was not collected from the subjects as such information will be based on memory recall and may not be accurate. Samples were collected on 2 days every week from both the centres during the period August 2003 July 2004. Women with symptoms of UTI and those with a positive nitrate or leucocyte esterase test were instructed to collect midstream clean catch urine (MSU). E. coli isolated in pure growth and in counts above 1000 CFU ml were included as infecting strains. For obtaining commensal E. coli, women who did not satisfy the above mentioned criteria were instructed to collect urine samples without precautions for MSU and also wipe the perianal area with a piece of sterile tissue paper and deposit the same in the urine sample. This method was followed as it was found to be more acceptable to women than producing faecal specimens. Sterile bottles were used for collecting both types of samples. All samples were transported to the laboratory in CMC within 2 4 h. Samples were plated on Maconkery agar. E. coli was identified by standard biochemical tests such as oxidase negativity, acid and in most cases gas without H2S in TSI medium, ability to ferment mannite, indole production in peptone broth and inability to utilise citrate or split urea. Susceptibility to antimicrobials was tested following NCCLS guidelines (NCCLS 2002) on Mueller Hinton agar. Disc strengths were as recommended by NCCLS. Reference strains were tested along with each batch, for quality control. Only one E. coli isolate per person was included in the study. Approximately 100 isolates were tested per month. Resistance data were analysed on a monthly basis for the commensal E. coli. Informed consent was obtained from the participants. The study was approved by the institutional review board and WHO ethics committee. Results A surveillance system was established successfully and data were collected at 12 time points. A total of 1095 commensal E. coli and 29 infecting E. coli were tested. 585 of the commensal E. coli were obtained from urban women and the remainder from rural women. 463 (42%) of the commensal isolates showed resistance to one or more drugs (Table 1); 92 (8.4%) were resistant to all three drugs commonly used for the treatment of UTI, namely ampicillin, co-trimoxazole and nalidixic acid. This was the commonest multidrug resistant phenotype. Nalidixic acid resistance not associated with ampicillin or co-trimoxazole resistance occurred in 104 (37%) of the 284 resistant isolates. Except for one isolate with gentamicin resistance, all other aminoglycoside and cephalosporin resistances occurred in strains resistant to one or more of the three drugs mentioned above. Only 16 (1.5%) of the isolates were resistant to nitrofurantoin. Although resistance was generally higher in the urban area, the difference was not statistically significant. Overall, 18% of isolates were resistant to ampicillin; >20% were resistant to each tetracycline, nalidixic acid and co-trimoxazole. 61.8% of the 199 ampicillin resistant isolates were susceptible to an amoxicillin-clavulanic acid combination. Minor variations over time occurred in prevalence of resistance to most antimicrobials (Figure 1). Resistance 42 ª 2008 Blackwell Publishing Ltd
Table 1 Resistance of 1095 commensal Escherichia coli Urban (n = 585) Rural (n = 510) Total (n = 1095) Ampicillin 118 20 81 16 199 18 Co-trimoxazole 151 26 112 22 263 24 Cefuroxime 15 3 6 1 21 2 Gentamicin 12 2 14 3 26 2 Chloramphenicol 31 5 30 6 61 6 Tetracycline 153 26 103 20 256 23 Nalidixic acid 165 28 119 23 284 26 Ciprofloxacin 27 5 13 3 40 4 100 90 80 70 60 50 40 30 20 10 Figure 1 Pattern of resistance among commensal Escherichia coli in rural and urban areas over 12 months in Vellore. 0 Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Cot ampi tetra nal genta cfr chloro cipro Table 2 Resistance patterns of 29 Escherichia coli causing urinary tract infection compared with that of commensals Pathogenic E. coli (n = 29) Commensal E. coli (n = 1095) Ampicillin 15 48 199 18 Co-trimoxazole 11 34 263 24 Cefuroxime 5 17 21 2 Gentamicin 4 10 26 2 Chloramphenicol 6 21 61 6 Tetracycline 10 34 256 23 Nalidixic acid 18 59 284 26 Ciprofloxacin 10 31 40 4 was higher in infecting than commensal flora (Table 2). By disc approximation test (NCCLS 2002) 18 (1.6%) commensal E. coli and 3 (10.3%) infecting strains (P < 0.001) were ESBL producers. Only four (19%) of these isolates were from rural areas (P < 0.05). Discussion We documented the pattern of AMR among commensal and infecting E. coli during a 1-year period. Prevalence of resistance among commensal E. coli in this study is similar to that currently found in infecting strains in high income countries (Farrell et al. 2003; Zhanel et al. 2006). Infecting strains in this study were more resistant than commensals, and resistance rates were higher against antimicrobials which have been in use longer (Mathai et al. 2004), such as co-trimoxazole, tetracyclines and ampicillin. This suggests that it probably takes many years before significant effects of drug use manifest at a community level. For most antimicrobials tested, there was no significant difference in resistance rates between urban and rural centres. However, the pattern of antimicrobial use may be different in these areas. Recent data suggest that prevalence of resistance can be high even in areas where antimicrobial use is minimal (Bartoloni et al. 2004) and is probably due to spread of resistant bacteria. Appearance of ESBL producing strains in the community (Valverde et al. 2004) is of concern. Prevalence of ESBL producing strains was ª 2008 Blackwell Publishing Ltd 43
higher in the urban area, probably reflecting higher use of cephalosporin. Thus the use of antimicrobials, especially the newer ones, must be controlled, and the spread of resistant clones addressed. There were month wise variations in resistance rates of most antimicrobials. The trends for different antimicrobials appeared to be similar. This is probably because resistant isolates were resistant to more than one antimicrobial and the numbers of resistant bacteria per month was small. There was no significant increase in overall resistance during the period of study. A longer period of observation is therefore required to document changes in resistance patterns over time. A system for surveillance of AMR in a given area can function as an early warning system for increasing resistance (Sorberg et al. 2002) and help to plan and monitor interventions. Although many developed countries have good systems to monitor trends in AMR, there are no such established methods in low income countries. E. coli is a good indicator organism to monitor resistance (Lester et al. 1990; Murray et al. 1990; Okeke et al. 2000; Bartoloni et al. 2006a). In this study we modified this basic concept to improve data collection. Perianal swabbing was more acceptable and practical than collecting stool samples for obtaining E. coli. Pregnant women, healthy representatives of the community, are less likely to be on antimicrobials and more easily accessed for collection of samples through antenatal clinics. Almost all pregnant women in the area we studied attend antenatal check ups, which renders them representative of the community. Gender differences in carriage of E. coli are unlikely. This is an innovation which we feel is practical for surveillance. Since the same method will be used for long term surveillance, changes that are noticed will be valid. However, the system has the potential for selecting pathogenic E. coli as urine can be inhibitory to commensal E. coli (Hull & Hull 1997). This is not a limitation to its use for surveillance to monitor changes over a period of time. Only one isolate per person was used. Although this can lead to under estimation of the actual prevalence of a resistance pool (Bartoloni et al. 2006b), it may not be a limitation for surveillance. Acknowledgement The study was funded by WHO, Geneva. References Bartoloni A, Bartalesi F, Mantella A et al. (2004) High prevalence of acquired antimicrobial resistance unrelated to heavy antimicrobial consumption. Journal of Infectious Diseases 189, 1291 1294. Bartoloni A, Benedetti M, Pallecchi L et al. (2006a) Evaluation of a rapid screening method for detection of antimicrobial resistance in the commensal microbiota of the gut. Transactions of the Royal Society of Tropical Medicine and Hygiene 100, 119 125. Bartoloni A, Pallecchi L, Benedetti M et al. (2006b) Multidrugresistant commensal Escherichia coli in children, Peru and Bolivia. Emerging Infectious Diseases 12, 907 913. Farrell DJ, Morrissey I, De Rubeis D, Robbins M & Felmingham D (2003) A UK multicentre study of the antimicrobial susceptibility of bacterial pathogens causing urinary tract infection. Journal of Infection 46, 94 100. Fluit AC & Schmitz FJ (2001) Bacterial resistance in urinary tract infections: how to stem the tide. Expert Opinion Pharmacother 2, 813 818. Gupta K, Sahm DF, Mayfield D & Stamm WE (2001) Antimicrobial resistance among uropathogens that cause community-acquired urinary tract infections in women: a nationwide analysis. Clinical Infectious Diseases 33, 89 94. Hull RA & Hull SI (1997) Nutritional requirements for growth of uropathogenic Escherichia coli in human urine. Infection and Immunity 65, 1960 1961. Karlowsky JA, Hoban DJ, Decorby MR, Laing NM & Zhanel GG (2006) Fluoroquinolone-resistant urinary isolates of Escherichia coli from outpatients are frequently multidrug resistant: results from the North American Urinary Tract Infection Collaborative Alliance-Quinolone Resistance study. Antimicrobial Agents and Chemotherapy 50, 2251 2254. Lester SC, del Pilar Pla M, Wang F, Perez Schael I, Jiang H & O Brien TF (1990) The carriage of Escherichia coli resistant to antimicrobial agents by healthy children in Boston, in Caracas, Venezuela, and in Qin Pu, China. New England Journal of Medicine 323, 285 289. Leverstein-van Hall MA, Box AT, Blok HE, Paauw A, Fluit AC & Verhoef J (2002) Evidence of extensive interspecies transfer of integron-mediated antimicrobial resistance genes among multidrug-resistant Enterobacteriaceae in a clinical setting. Journal of Infectious Diseases 186, 49 56. Livermore DM (2003) Bacterial resistance: origins, epidemiology, and impact. Clinical Infectious Diseases 36, S11 S23. Mathai E, Thomas RJ, Chandy S, Mathai M & Bergstrom S (2004) Antimicrobials for the treatment of urinary tract infection in pregnancy: practices in southern India. Pharmacoepidemiology and Drug Safety 13, 645 652. Murray BE, Mathewson JJ, DuPont HL, Ericsson CD & Reves RR (1990) Emergence of resistant faecal Escherichia coli in travelers not taking prophylactic antimicrobial agents. Antimicrobial Agents and Chemotherapy 34, 515 518. NCCLS (2002) Performance Standards for Antimicrobial Disk Susceptibility Testing, Wayne, Pennsylvania, USA. O Brien TF (2002) Emergence, spread, and environmental effect of antimicrobial resistance: how use of an antimicrobial anywhere can increase resistance to any antimicrobial anywhere else. Clinical Infectious Diseases 34(Suppl. 3), S78 S84. 44 ª 2008 Blackwell Publishing Ltd
Okeke IN, Fayinka ST & Lamikanra A (2000) Antibiotic resistance in Escherichia coli from Nigerian students, 1986 1998. Emerging Infectious Diseases 6, 393 396. Sahm DF, Thornsberry C, Mayfield DC, Jones ME & Karlowsky JA (2001) Multidrug-resistant urinary tract isolates of Escherichia coli: prevalence and patient demographics in the United States in 2000. Antimicrobial Agents and Chemotherapy 45, 1402 1406. Sorberg M, Farra A, Ransjo U et al. (2002) Long-term antibiotic resistance surveillance of gram-negative pathogens suggests that temporal trends can be used as a resistance warning system. Scandinavian Journal of Infectious Diseases 34, 372 378. Turnidge J, Bell J, Biedenbach DJ & Jones RN (2002) Pathogen occurrence and antimicrobial resistance trends among urinary tract infection isolates in the Asia Western Pacific Region: report from the SENTRY Antimicrobial Surveillance Program, 1998 1999. International Journal of Antimicrobial Agents 20, 10 17. Valverde A, Coque TM, Sanchez-Moreno MP, Rollan A, Baquero F & Canton R (2004) Dramatic increase in prevalence of faecal carriage of extended-spectrum beta-lactamase-producing Enterobacteriaceae during nonoutbreak situations in Spain. Journal of Clinical Microbiology 42, 4769 4775. WHO (2001) Global Strategy for Containment of Antimicrobial Resistance, World Health Organisation, Geneva. Zhanel GG, Hisanaga TL, Laing NM et al. (2006) Antibiotic resistance in Escherichia coli outpatient urinary isolates: final results from the North American Urinary Tract Infection Collaborative Alliance (NAUTICA). International Journal of Antimicrobial Agents 27, 468 475. Corresponding Author Elizabeth Mathai, Department of Clinical Microbiology, Christian Medical College, Ida Scudder Road, Vellore 632004, India. Tel.: +41 22 7777239; E-mail: mathaim@yahoo.com ª 2008 Blackwell Publishing Ltd 45