Increasing Resistance in Commensal Escherichia coli, Bolivia and Peru

To the Editor: The global increase of antimicrobial-drug resistance, including resistance to the new and most potent antimicrobial agents, is a major public health concern. In low-resource countries, where bacterial infections are still among the major causes of death, especially for children, it is of particular concern (1). 
 
ANTRES (Towards Controlling Antimicrobial Use and Resistance in Low-income Countries—An Intervention Study in Latin America) is a research project on antimicrobial-drug use and resistance in low-resource countries of Latin America (see www.unifi.it/infdis/antres/default.htm). In 2002, the baseline ANTRES study showed a high rate of fecal carriage of Escherichia coli with acquired resistance to several antimicrobial drugs, especially older drugs (e.g., ampicillin, trimethoprim-sulfamethoxazole, tetracycline, streptomycin, and chloramphenicol), in preschool children from 4 urban settings in Bolivia and Peru (2). We report the results of a second cross-sectional study, conducted in 2005, that evaluated the evolution of antimicrobial-drug resistance in the studied areas. 
 
We studied healthy children 6–72 months of age from each of 4 urban areas: 2 in Bolivia (Camiri, Santa Cruz Department; Villa Montes, Tarija Department) and 2 in Peru (Yurimaguas, Loreto Department; Moyobamba, San Martin Department). The study design, sampling and inclusion criteria, methods, and ethical issues were the same as those of the baseline study (2). The study was carried out over 4 months (September–December 2005), the same seasonal period as in the previous study. No significant differences in sex ratios were found among children enrolled from the different areas, whereas minor differences were found for age. No statistical differences were found between the 2002 baseline study and the 2005 study results in terms of numbers of children (3,193 vs. 3,174) and sex ratios (0.94 vs. 0.95) (Table). Statistical analyses were performed by using Stata 9.0 (Stata Corp., College Station, TX, USA). Logistic regression models were used to compare the antimicrobial-drug resistance rates in 2002 and 2005, considering the combined influences of age, sex, city, and country. 
 
 
 
Table 
 
Antimicrobial drug–resistance rates of Escherichia coli as part of 
commensal flora in children, Bolivia and Peru, 2002 and 2005* 
*Expanded Table available online at www.cdc.gov/EID/content/14/2/338-T.htm. 
 
 
 
Data from the 2005 survey confirmed high resistance rates for ampicillin, trimethoprim-sulfamethoxazole, tetracycline, streptomycin, and chloramphenicol. The differences in resistance rates observed between 2002 and 2005 for these drugs, although sometimes statistically significant, are probably of limited epidemiologic relevance due to the high rates of antimicrobial-drug resistance found in the E. coli population in both surveys. The most relevant finding of the 2005 survey was the remarkable increase since 2002 in the resistance rates to fluoroquinolones and expanded-spectrum cephalosporins (Table). Molecular analysis showed that the dramatic increase in rates of resistance to expanded spectrum cephalosporins was mostly the result of dissemination of CTX-M-type extended-spectrum β-lactamase determinants (3). Concerning the association between sex and resistance rates, the higher resistance rates observed for some agents and in some settings for boys in the baseline study were not confirmed, except in 1 case (kanamycin in Camiri, p = 0.04) (2). Analysis by age (not performed for amikacin due to low numbers of resistant isolates) confirmed the occurrence of higher resistance rates for the youngest age group and an overall decreasing trend by age for all agents, except ciprofloxacin and gentamicin. For these 2 agents, resistance rates increased, although not significantly (p = 0.95 and p = 0.55, respectively) (2). Although we did not specifically address factors potentially involved in this phenomenon, we will address them in future investigations. 
 
Increasing resistance to fluoroquinolones and expanded-spectrum cephalosporins among E. coli clinical isolates has been observed in several parts of the world and complicates the management of infections (4,5). Recently, intestinal colonization with fluoroquinolone-resistant or extended-spectrum β-lactamase–producing E. coli of nonhospitalized persons has been described as an emerging phenomenon (6–9). Although the exact clinical implications of this phenomenon are not clearly established, colonization by these resistant strains is a public health threat at the community and hospital levels (8,9). The reasons for the increased prevalence of fecal carriage of these resistant E. coli strains by children from the studied areas are not clear. Data collected about household use of antimicrobial drugs excluded previous use of fluoroquinolones and expanded-spectrum cephalosporins (C. Kristiansson et al., unpub. data). The increased prevalence of resistant E. coli strains in preschool children most likely reflects increased exposure within a contaminated household setting, in the food chain, or both (6,8,10). 
 
Our findings support the need to continue monitoring the evolution of resistance in commensal E. coli, to evaluate the effects of these important reservoirs of resistance genes distributed in the community, to investigate the epidemiologic relationship with clinical isolates, and to define the role of the food supply. Investigation into whether carriage of resistant strains in adults correlates with data on antimicrobial-drug use in hospitals and in the community would also be of interest.

The fi rst confi rmed case of chikungunya fever on Rodrigues Island was reported in February, followed by 56 cases in March, 393 in April, and >80 in May (9).
The increase in CDRs reported for the Republic of Mauritius during the chikungunya epidemic is similar to the fi ndings reported for the neighboring island of Réunion (10). Excess deaths in Réunion and the Republic of Mauritius coincided with the epidemic curve of the chikungunya fever outbreak, which suggested an association between these 2 factors. No other events that may have negatively affected the health of persons living in the Republic of Mauritius in 2006 were reported. However, because information on cause of death is unavailable, studies are needed to confi rm that the chikungunya fever outbreak contributed to increased CDRs in 2006.

Increasing Resistance in Commensal Escherichia coli, Bolivia and Peru
To the Editor: The global increase of antimicrobial-drug resistance, including resistance to the new and most potent antimicrobial agents, is a major public health concern. In low-resource countries, where bacterial infections are still among the major causes of death, especially for children, it is of particular concern (1).
ANTRES (Towards Controlling Antimicrobial Use and Resistance in Low-income Countries-An Intervention Study in Latin America) is a research project on antimicrobial-drug use and resistance in low-resource countries of Latin America (see www.unifi . it/infdis/antres/default.htm). In 2002, the baseline ANTRES study showed a high rate of fecal carriage of Escherichia coli with acquired resistance to several antimicrobial drugs, especially older drugs (e.g., ampicillin, trimethoprimsulfamethoxazole, tetracycline, streptomycin, and chloramphenicol), in preschool children from 4 urban settings in Bolivia and Peru (2). We report the results of a second cross-sectional study, conducted in 2005, that evaluated the evolution of antimicrobialdrug resistance in the studied areas.
We studied healthy children 6-72 months of age from each of 4 urban areas: 2 in Bolivia (Camiri, Santa Cruz Department; Villa Montes, Tarija Department) and 2 in Peru (Yurimaguas, Loreto Department; Moyobamba, San Martin Department). The study design, sampling and inclusion criteria, methods, and ethical issues were the same as those of the baseline study (2). The study was carried out over 4 months (September-December 2005), the same seasonal period as in the previous study. No signifi cant differences in sex ratios were found among children enrolled from the different areas, whereas minor differences were found for age. No statistical differences were found between the 2002 baseline study and the 2005 study results in terms of numbers of children (3,193 vs. 3,174) and sex ratios (0.94 vs. 0.95) (Table). Statistical analyses were performed by using Stata 9.0 (Stata Corp., College Station, TX, USA). Logistic regression models were used to compare the antimicrobial-drug resistance rates in 2002 and 2005, considering the combined infl uences of age, sex, city, and country.
Data from the 2005 survey confi rmed high resistance rates for ampicillin, trimethoprim-sulfamethoxazole, tetracycline, streptomycin, and chloramphenicol. The differences in resistance rates observed between 2002 and 2005 for these drugs, although sometimes statistically signifi cant, are probably of limited epidemiologic relevance due to the high rates of antimicrobial-drug resistance found in the E. coli population in both surveys. The most relevant fi nding of the 2005 survey was the remarkable increase since 2002 in the resistance rates to fl uoroquinolones and expanded-spectrum cephalosporins (Table). Molecular analysis showed that the dramatic increase in rates of resistance to expanded spectrum cephalosporins was mostly the result of dissemination of CTX-M-type extended-spectrum βlactamase determinants (3). Concerning the association between sex and resistance rates, the higher resistance rates observed for some agents and in some settings for boys in the baseline study were not confi rmed, except in 1 case (kanamycin in Camiri, p = 0.04) (2). Analysis by age (not performed for amikacin due to low numbers of resistant isolates) confi rmed the occurrence of higher resistance rates for the youngest age group and an overall decreasing trend by age for all agents, except ciprofl oxacin and gentamicin. For these 2 agents, resistance rates increased, although not signifi cantly (p = 0.95 and p = 0.55, respectively) (2). Although we did not specifi cally address factors potentially involved in this phenomenon, we will address them in future investigations.
Increasing resistance to fl uoroquinolones and expanded-spectrum cephalosporins among E. coli clinical isolates has been observed in several parts of the world and complicates the management of infections (4,5). Recently, intestinal colonization with fl uoroquinolone-resistant or extended-spectrum β-lactamase-producing E. coli of nonhospitalized persons has been described as an emerging phenomenon (6)(7)(8)(9). Although the ex-act clinical implications of this phenomenon are not clearly established, colonization by these resistant strains is a public health threat at the community and hospital levels (8,9). The reasons for the increased prevalence of fecal carriage of these resistant E. coli strains by children from the studied areas are not clear. Data collected about household use of antimicrobial drugs excluded previous use of fl uoroquinolones and expanded-spectrum cephalosporins (C. Kristiansson et al., unpub. data). The increased prevalence of resistant E. coli strains in preschool children most likely refl ects increased exposure within a contaminated household setting, in the food chain, or both (6,8,10).
Our fi ndings support the need to continue monitoring the evolution of resistance in commensal E. coli, to evaluate the effects of these important reservoirs of resistance genes distributed in the community, to investigate the epidemiologic relationship with clinical isolates, and to defi ne the role of the food supply. Investigation into whether carriage of resistant strains in adults correlates with data on antimicrobial-drug use in hospitals and in the community would also be of interest. ‡Wald test applied to establish the statistical significance of parameters obtained from logistic regression analysis; NS, not significant; NA, not applicable (due to lack of variability of data).

Plasmid-mediated Quinolone Resistance in Salmonella enterica, United Kingdom
To the Editor: Fluoroquinolones are broad-spectrum antimicrobial drugs used to treat many clinical infections. Salmonellosis is treated with fl uoroquinolones only in elderly or immunocompromised patients, but these drugs are also used for treating patients with enteric fever, invasive disease, or long-term salmonellae carriage. High-level fl uoroquinolone resistance is uncommon, but reduced susceptibility is increasing.
Since 1998, plasmid-mediated quinolone resistance encoded by qnr genes A, B, and S that confer lowlevel resistance to nalidixic acid and reduced susceptibility to ciprofl oxacin has been identifi ed in several enterobacterial species, including Salmonella. Their clinical importance is in facilitating resistance to potentially lethal levels of quinolone. Additionally, qnr genes are often associated with strains that produce extended-spectrum β-lactamases.
We recently reported identifi cation of qnr genes in Salmonella in the United Kingdom (1). Most isolates were associated with the Far East. Two isolates of S. Virchow were part of an outbreak associated with imported cooked chicken from Thailand. During October 2006-April 2007, we monitored qnr genes in nontyphoidal salmonellae isolated in the United Kingdom that expressed reduced susceptibility to ciprofl oxacin (MIC 0.125-1.0 μg/mL) with concomitant susceptibility to nalidixic acid (MIC <16 μg/mL). This resistance phenotype is a useful marker for the qnr gene as the sole quinolone resistance determinant (1).
Recent studies showed that isolates of Salmonella spp. and Escherichia coli with decreased susceptibility to ciprofl oxacin (MICs >0.06 μg/mL and 0.5 μg/mL, respectively), but with susceptibility or intermediate resistance to nalidixic acid (MIC 8-16 μg/mL and 4-8 μg/mL, respectively), all had qnrA or qnrS genes but lacked mutations in the topoisomerase genes (2,3). Strains with ciprofl oxacin MICs >1 μg/mL were also included to monitor involvement of qnr genes in development of high-level ciprofl oxacin resistance. Breakpoint concentrations used are based on long-term studies within the Health Protection Agency Laboratory of Enteric Pathogens. Ciprofl oxacin Etest (AB Biodisk, Solna, Sweden) results were interpreted according to manufacturer's procedures. A total of 45 Salmonella spp. strains were tested. Screening for qnr genes by multiplex PCR identifi ed 37 isolates with qnrS and 2 carrying qnrB variants (Table) (4). However, the qnrB primer pair in this multiplex did not fully match all qnrB gene variants. PCR and sequencing using primers FQ1 and FQ2 (5) and qnrS-F and qnrS-R (1), were used to identify specifi c qnrB and qnrS gene variants.
The qnrS1-positive salmonellae belong to serotypes Typhimurium