- Email: [email protected]
Antibiotic resistance in oral commensal streptococci from healthy Mexicans and Cubans: resistance prevalence does not mirror antibiotic usage
FEMS Microbiology Letters 217 (2002) 173^176
www.fems-microbiology.org
Antibiotic resistance in oral commensal streptococci from healthy Mexicans and Cubans: resistance prevalence does not mirror antibiotic usage J. Javier D|¤az-Mej|¤a, Alejandro Carbajal-Saucedo, Carlos F. Ama¤bile-Cuevas
Fundacio¤n LUSARA, Apartado Postal 102-006, 08930 Mexico D.F., Mexico Received 25 February 2002; received in revised form 13 June 2002; accepted 10 October 2002 First published online 6 November 2002
Abstract Antibiotic resistance genes might be maintained by selection pressures different from those which are responsible for initially selecting resistant bacteria. This possibility was suggested from a comparison of oral commensal streptococci isolated from healthy people not taking antibiotics. Resistance frequencies were similar for organisms from Mexico and Cuba despite significant differences in antibiotic usage in these two countries. Resistance to v 4 drugs was far more common in Mexico, the only detectable trend that can be related to the higher use of antibiotics in Mexico. If resistance is not uniquely maintained by antibiotics, then other environmental factors must also be at work. These need to be identified if a strategy to control antibiotic resistance is to be successful. 5 2002 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.
1. Introduction Antibiotic usage is the main pressure selecting for antibiotic resistance. It is not known if antibiotics are also necessary to maintain resistance in populations of bacteria [1,2]. Other environmental agents might substitute for antibiotics in selecting resistance phenotypes once resistant bacteria are established within a niche. In such cases, suspension of antibiotic use would not be enough to guarantee that the resistance phenotype would recede. Antibiotic selection is not limited to pathogenic organisms. Indigenous commensal bacteria are more frequently exposed to antibiotics. Most resistance surveys have focused on virulent organisms, especially human pathogens, but comparatively little is known about the prevalence of resistance in commensal bacteria. Indigenous £ora may act as a reservoir of resistance determinants, potentially transferable to infectious organisms [3].
* Corresponding author. Tel. :/Fax: +52 (55) 52195855. E-mail address : [email protected] (C.F. Ama¤bile-Cuevas).
Is antibiotic use correlated with the frequency of resistance among commensal bacteria ? To answer this question, we tested the resistance rates of oral streptococci isolated from healthy Mexicans and Cubans who had not taken antibiotics for at least 6 months prior to sampling. Mexico and Cuba have cultural and economical similarities, but their antibiotic usage patterns di¡er substantially. In Mexico, antibiotics are readily available and self-prescribed, circumstances that permit intense medical and agricultural abuse. In Cuba, antibiotics have been in short supply, prescription of most of them is tightly controlled, and they are not used in agriculture [4]. Although there are other countries with better-documented antibiotic controls, important societal di¡erences prevent a clear comparison between those countries and Mexico. Each country di¡ers in antibiotic use. In Mexico, the most commonly used antibiotics are L-lactams, ampicillin being the nation’s top-selling drug. In Cuba, the most commonly used antibiotic is tetracycline. Tetracycline ranks second or third in Mexico ([4] and www.americas. health-sector-reform.org). Despite the di¡erent antibiotic usage patterns between Mexico and Cuba, commensal isolates from both countries di¡er only slightly in resistance pro¢le. Usage control seems to have little e¡ect upon resistance, at least within oral streptococci. We presume that pressures other than antibiotics might be acting to maintain the prevalence of resistance determinants.
0378-1097 / 02 / $22.00 5 2002 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. PII : S 0 3 7 8 - 1 0 9 7 ( 0 2 ) 0 1 0 8 6 - 8
FEMSLE 10741 28-11-02
174
J.J. D|¤az-Mej|¤a et al. / FEMS Microbiology Letters 217 (2002) 173^176
2. Materials and methods
resistant
intermediate
80
CUBA
2.1. Isolation and characterization of strains
MEXICO
2.2. Polymerase chain reaction (PCR) detection of erm genes erm Genes were detected by PCR ampli¢cation using degenerate primers as previously described [6].
3. Results Our data indicate that Cuban isolates were more often resistant to sulfadiazine and trimethoprim than Mexican isolates (Fig. 1). The frequency of erythromycin resistance (intermediate+resistant) in Mexico is similar to that observed in Cuba, although Mexican isolates are two times more frequently fully resistant to erythromycin. The distribution of erm genes assessed by PCR detection, among resistant and intermediate strains within each country’s isolates, is very similar (Table 1). This suggests that the current breakpoints for erythromycin resistance and intermediate phenotypes do not accurately re£ect the existence of resistance determinants. Multi-resistance patterns are the most dramatic di¡erence between the countries. While most Cuban strains are resistant to only 0^3 drugs, Mexican strains were distributed across two groups: 60% susceptible and mono-resistant ; and 35% resistant to 3^6 drugs (Fig. 2).
60 isolates (%)
Urban Mexican (Mexico City) and Cuban (La Habana) volunteers were asked to complete a questionnaire on general health and antibiotic usage. An oral sample was taken, rubbing a transport swab (Culturette, BectonDickinson, Cockeysville, MD, USA) against the periodontal area. Swabs were streaked on Mitis-Salivarius agar (Difco, Detroit, MI, USA), and one of each morphologically di¡erent colony was isolated, subcultured and stored at 370‡C in glycerol-containing brain^heart infusion broth. 100 isolates from each country were included. These organisms were isolated from healthy volunteers who have not taken antibiotics, nor their housemates, for at least 6 months prior to the sampling process, and were not health-care workers. Volunteers were not visiting anyone at hospitals. Isolates from 29 Mexican volunteers and from 46 Cuban volunteers were included. Bacteria were identi¢ed using the API Strep 20 kit (bioMe¤rieux, Marcy l’Etoile, France). Susceptibility to ampicillin, cipro£oxacin, erythromycin, sulfadiazine, tetracycline and trimethoprim was assessed by the disk-di¡usion method, using sensi-disks (Becton-Dickinson) on plates of Mueller^ Hinton agar supplemented with 5% blood, following National Committee for Laboratory Standards guidelines [5].
40 20 0 SUL TMP TET
CIP AMP ERY
Fig. 1. Resistance pro¢les of Mexican and Cuban oral commensal streptococci. SUL, sulfonamide ; TMP, trimethoprim; TET, tetracycline ; CIP, cipro£oxacin ; AMP, ampicillin ; ERY, erythromycin. Inhibitory halo diameter breakpoints for AMP, ERY and TET, from the ‘Streptococcus spp. other than S. pneumoniae’ table, and for SUL, TMP and CIP from the ‘Staphylococcus spp.’ table [13]. The later resistance breakpoints upon these organisms might not be clinically relevant, and were used for comparison purposes.
4. Discussion Antibiotic resistance correlates with the use of antibiotics in the hospital setting. Enteric bacteria isolated from healthy Mexican children, but within hospitals, are frequently resistant [7]. Oral streptococci isolated from chronically medicated Mexican patients are much more frequently resistant than those reported here: 37% are resistant to erythromycin and 8% to cipro£oxacin, for instance (C.F. Ama¤bile-Cuevas, unpublished data). An even higher prevalence of resistance (e.g. 14.5% penicillin-resistant, 38.5% erythromycin-resistant) among oral streptococci isolated from healthy Greek children, not taking antibiotics but attending an orthopaedic clinic, was recently reported [8]. Interestingly, Greece had the highest numbers of antibiotic prescriptions among 13 European countries [9]. We wished to investigate if resistance to antibiotics can recede with our interventions if use is better controlled. Since Mexico and Cuba are comparable societies that differ in their use of antibiotics, they serve as natural laboratories to begin answering this question. Table 1 Presence of erm genes in erythromycin-resistant and intermediate oral streptococci
ERY resistant erm detected ERY intermediate erm detected
Mexico
Cuba
10 6 11 5
4 1 18 5
erm genes detected by PCR (see text). The prevalence of erm genes between resistant and intermediate organisms, within each country, is not signi¢cantly di¡erent (P s 0.05, M2 test).
FEMSLE 10741 28-11-02
J.J. D|¤az-Mej|¤a et al. / FEMS Microbiology Letters 217 (2002) 173^176
CUBA MEXICO
40
isolates (%)
30 20 10 0
0 1 2 3 4 5 6 number of antibiotics isolates are resistant
Fig. 2. Multi-resistance pro¢les of Mexican and Cuban oral commensal streptococci. Only fully resistant organisms were included ; breakpoints as in Fig. 1.
Our data suggest that the maintenance of resistance in bacterial populations freed from antibiotic selection does not seem to correlate with overall changes in antibiotic usage patterns, at least between Mexico and Cuba. Few studies on commensal bacteria outside hospital settings are available for comparison. The volunteers we selected are among the least exposed to antibiotics in an urban population (see inclusion criteria above). At least in Mexico, such individuals are di⁄cult to ¢nd. In an unpublished survey, 4.2% of Mexicans have had antibiotics in a period of 2 weeks before the survey (C. Izazola, personal communication). It would be very interesting to assess the minimal exposure necessary to introduce resistance into the indigenous £ora. Could living with someone receiving antibiotic treatment, or having a single visit to a hospital, exert a change in such £ora? Once a resistance determinant is established, its prevalence does not seem a¡ected by overall antibiotic usage. This has been observed even in animal husbandry: after years of having eliminated the use of avoparcin, glycopeptide-resistant enterococci are still being detected in farms [10,11]. In our study, high sulfonamide and trimethoprim resistance frequencies, and carrying erm genes in the oral streptococcal isolates, are very similar despite di¡erences in antibiotic availability in Mexico and Cuba. Possibly the ¢tness cost of carrying resistance genes is negligible. Alternatively, other selective pressures might be more important. For instance, resistance genes can take other roles in bacteria or be linked to genes that would incur a greater cost to ¢tness if lost. Trying to ascertain the nature of such possible pressures, we asked the volunteers about the number of dental ¢llings they carried, as mercury released from dental amalgams has been proposed to co-select for antibiotic resistance [12]. Cubans bear slightly more ¢llings per individual than Mexicans do, but no relationship between the number of reported ¢llings and resistance prevalence
175
was found, nor any change in mercury resistance in the isolates (data not shown). Although the frequency of resistance to individual drugs does not seem to be a¡ected by the overall use of antibiotics in the community, the accumulation of resistance determinants might be increased by greater antibiotic use (Fig. 2). Mexican strains seem to be distributed in two groups, one susceptible to all or most drugs tested, and another resistant to three or more antibiotics. Antibiotic abuse might be selecting, in addition to resistant strains, isolates with the ability to develop and/or acquire resistance more readily. Multi-resistance is, indeed, one of the most feared aspects of antibiotic resistance, as it dramatically reduces the therapeutic options when treating infections. If the availability and abuse of antibiotics is actually selecting and/or maintaining multi-resistance, even in non-treated individuals, enforced policies of rational antibiotic use are de¢nitely needed. Also, further research is necessary on non-antibiotic agents and conditions that might be maintaining resistance determinants among bacterial populations.
Acknowledgements This work received partial support from a grant from NIH/NIAID to A. Salyers and S. Levy. We thank X. Garc|¤a, who made possible the sampling of Cuban volunteers, and A. Fuentes and G. Jime¤nez for helping with the sampling process. A. Summers, J. Davies, M. Roberts and M. Chicurel gave helpful comments, as well as the editor and reviewers.
References [1] Salyers, A.A. and Ama¤bile-Cuevas, C.F. (1997) Why are antibiotics resistance genes so resistant to elimination ? Antimicrob. Agents Chemother. 41, 2321^2325. [2] Heinemann, J.A., Ankenbauer, R.G. and Ama¤bile-Cuevas, C.F. (2000) Do antibiotics maintain antibiotic resistance? Drug Discov. Today 5, 195^204. [3] Levy, S.B. (1986) Ecology of antibiotic resistance determinants. In: Antibiotic Resistance Genes: Ecology, Transfer, and Expression (Levy, S.B. and Novick, R.P., Eds.), pp. 17^30. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. [4] Calva, J.J., Niebla-Pe¤rez, A., Rodr|¤guez-Lemoine, V., Santos, J.I. and Ama¤bile-Cuevas, C.F. (1996) Antibiotic usage and antibiotic resistance in Latin America. In: Antibiotic Resistance : From Molecular Basics to Therapeutic Options (Ama¤bile-Cuevas, C.F., Ed.), pp. 73^97. RG Landes/Chapman and Hall, Austin, TX. [5] National Committee for Clinical Laboratory Standards (1996) Performance Standards for Antimicrobial Disk Susceptibility Tests, 6th edn. NCCLS, Wayne, PA. [6] Arthur, M., Molinas, C. and Mabilat, C. (1993) PCR detection of erm erythromycin resistance genes by using degenerate oligonucleotide primers. In: Diagnostic Molecular Microbiology (Persing, D.H., Smith, T.F., Tenover, F.C. and White, T.J., Eds.), pp. 534^538. American Society for Microbiology, Washington, DC. [7] Calva, J.J., Sifuentes-Osornio, J. and Cero¤n, C. (1996) Antimicrobial
FEMSLE 10741 28-11-02
176
J.J. D|¤az-Mej|¤a et al. / FEMS Microbiology Letters 217 (2002) 173^176
resistance in fecal £ora: longitudinal community-based surveillance of children from urban Mexico. Antimicrob. Agents Chemother. 40, 1699^1702. [8] Ioannidou, S., Tassios, P.T., Kotsovili-Tseleni, A., Foustoukou, M., Legakis, N.J. and Vatopoulos, A. (2001) Antibiotic resistance rates and macrolide resistance phenotypes of viridans group streptococci from the oropharynx of healthy Greek children. Int. J. Antimicrob. Agents 17, 195^201. [9] Mo«lstad, S., Lundberg, C.S., Karlsson, A.-K. and Cars, O. (2002) Antibiotic prescription rates very markedly between 13 European countries. Scand. J. Infect. Dis. 34, 366^371. [10] Borgen, K., SYrum, M., Kruse, H. and Wasteson, Y. (2000) Persistence of vancomycin-resistant enterococci (VRE) on Norwegian broiler farms. FEMS Microbiol. Lett. 191, 255^258.
[11] Borgen, K., Simonsen, G.S., Sundsfjord, A., Wasteson, Y., Olsvik, Z. and Kruse, H. (2000) Continuing high prevalence of VanA-type vancomycin-resistant enterococci on Norwegian poultry farms three years after avoparcin was banned. J. Appl. Microbiol. 89, 478^485. [12] Summers, A.O., Wireman, J., Vimy, M.J., Lorscheider, F.L., Marshall, B., Levy, S.B., Bennett, S. and Billard, L. (1993) Mercury released from dental ‘silver’ ¢llings provokes an increase in mercuryand antibiotic-resistant bacteria in oral and intestinal £oras of primates. Antimicrob. Agents Chemother. 37, 825^834. [13] National Committee for Clinical Laboratory Standards (1998) Performance Standards for Antimicrobial Susceptibility Testing, 8th Informational Supplement. NCCLS, Wayne, PA.
FEMSLE 10741 28-11-02
www.fems-microbiology.org
Antibiotic resistance in oral commensal streptococci from healthy Mexicans and Cubans: resistance prevalence does not mirror antibiotic usage J. Javier D|¤az-Mej|¤a, Alejandro Carbajal-Saucedo, Carlos F. Ama¤bile-Cuevas
Fundacio¤n LUSARA, Apartado Postal 102-006, 08930 Mexico D.F., Mexico Received 25 February 2002; received in revised form 13 June 2002; accepted 10 October 2002 First published online 6 November 2002
Abstract Antibiotic resistance genes might be maintained by selection pressures different from those which are responsible for initially selecting resistant bacteria. This possibility was suggested from a comparison of oral commensal streptococci isolated from healthy people not taking antibiotics. Resistance frequencies were similar for organisms from Mexico and Cuba despite significant differences in antibiotic usage in these two countries. Resistance to v 4 drugs was far more common in Mexico, the only detectable trend that can be related to the higher use of antibiotics in Mexico. If resistance is not uniquely maintained by antibiotics, then other environmental factors must also be at work. These need to be identified if a strategy to control antibiotic resistance is to be successful. 5 2002 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.
1. Introduction Antibiotic usage is the main pressure selecting for antibiotic resistance. It is not known if antibiotics are also necessary to maintain resistance in populations of bacteria [1,2]. Other environmental agents might substitute for antibiotics in selecting resistance phenotypes once resistant bacteria are established within a niche. In such cases, suspension of antibiotic use would not be enough to guarantee that the resistance phenotype would recede. Antibiotic selection is not limited to pathogenic organisms. Indigenous commensal bacteria are more frequently exposed to antibiotics. Most resistance surveys have focused on virulent organisms, especially human pathogens, but comparatively little is known about the prevalence of resistance in commensal bacteria. Indigenous £ora may act as a reservoir of resistance determinants, potentially transferable to infectious organisms [3].
* Corresponding author. Tel. :/Fax: +52 (55) 52195855. E-mail address : [email protected] (C.F. Ama¤bile-Cuevas).
Is antibiotic use correlated with the frequency of resistance among commensal bacteria ? To answer this question, we tested the resistance rates of oral streptococci isolated from healthy Mexicans and Cubans who had not taken antibiotics for at least 6 months prior to sampling. Mexico and Cuba have cultural and economical similarities, but their antibiotic usage patterns di¡er substantially. In Mexico, antibiotics are readily available and self-prescribed, circumstances that permit intense medical and agricultural abuse. In Cuba, antibiotics have been in short supply, prescription of most of them is tightly controlled, and they are not used in agriculture [4]. Although there are other countries with better-documented antibiotic controls, important societal di¡erences prevent a clear comparison between those countries and Mexico. Each country di¡ers in antibiotic use. In Mexico, the most commonly used antibiotics are L-lactams, ampicillin being the nation’s top-selling drug. In Cuba, the most commonly used antibiotic is tetracycline. Tetracycline ranks second or third in Mexico ([4] and www.americas. health-sector-reform.org). Despite the di¡erent antibiotic usage patterns between Mexico and Cuba, commensal isolates from both countries di¡er only slightly in resistance pro¢le. Usage control seems to have little e¡ect upon resistance, at least within oral streptococci. We presume that pressures other than antibiotics might be acting to maintain the prevalence of resistance determinants.
0378-1097 / 02 / $22.00 5 2002 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. PII : S 0 3 7 8 - 1 0 9 7 ( 0 2 ) 0 1 0 8 6 - 8
FEMSLE 10741 28-11-02
174
J.J. D|¤az-Mej|¤a et al. / FEMS Microbiology Letters 217 (2002) 173^176
2. Materials and methods
resistant
intermediate
80
CUBA
2.1. Isolation and characterization of strains
MEXICO
2.2. Polymerase chain reaction (PCR) detection of erm genes erm Genes were detected by PCR ampli¢cation using degenerate primers as previously described [6].
3. Results Our data indicate that Cuban isolates were more often resistant to sulfadiazine and trimethoprim than Mexican isolates (Fig. 1). The frequency of erythromycin resistance (intermediate+resistant) in Mexico is similar to that observed in Cuba, although Mexican isolates are two times more frequently fully resistant to erythromycin. The distribution of erm genes assessed by PCR detection, among resistant and intermediate strains within each country’s isolates, is very similar (Table 1). This suggests that the current breakpoints for erythromycin resistance and intermediate phenotypes do not accurately re£ect the existence of resistance determinants. Multi-resistance patterns are the most dramatic di¡erence between the countries. While most Cuban strains are resistant to only 0^3 drugs, Mexican strains were distributed across two groups: 60% susceptible and mono-resistant ; and 35% resistant to 3^6 drugs (Fig. 2).
60 isolates (%)
Urban Mexican (Mexico City) and Cuban (La Habana) volunteers were asked to complete a questionnaire on general health and antibiotic usage. An oral sample was taken, rubbing a transport swab (Culturette, BectonDickinson, Cockeysville, MD, USA) against the periodontal area. Swabs were streaked on Mitis-Salivarius agar (Difco, Detroit, MI, USA), and one of each morphologically di¡erent colony was isolated, subcultured and stored at 370‡C in glycerol-containing brain^heart infusion broth. 100 isolates from each country were included. These organisms were isolated from healthy volunteers who have not taken antibiotics, nor their housemates, for at least 6 months prior to the sampling process, and were not health-care workers. Volunteers were not visiting anyone at hospitals. Isolates from 29 Mexican volunteers and from 46 Cuban volunteers were included. Bacteria were identi¢ed using the API Strep 20 kit (bioMe¤rieux, Marcy l’Etoile, France). Susceptibility to ampicillin, cipro£oxacin, erythromycin, sulfadiazine, tetracycline and trimethoprim was assessed by the disk-di¡usion method, using sensi-disks (Becton-Dickinson) on plates of Mueller^ Hinton agar supplemented with 5% blood, following National Committee for Laboratory Standards guidelines [5].
40 20 0 SUL TMP TET
CIP AMP ERY
Fig. 1. Resistance pro¢les of Mexican and Cuban oral commensal streptococci. SUL, sulfonamide ; TMP, trimethoprim; TET, tetracycline ; CIP, cipro£oxacin ; AMP, ampicillin ; ERY, erythromycin. Inhibitory halo diameter breakpoints for AMP, ERY and TET, from the ‘Streptococcus spp. other than S. pneumoniae’ table, and for SUL, TMP and CIP from the ‘Staphylococcus spp.’ table [13]. The later resistance breakpoints upon these organisms might not be clinically relevant, and were used for comparison purposes.
4. Discussion Antibiotic resistance correlates with the use of antibiotics in the hospital setting. Enteric bacteria isolated from healthy Mexican children, but within hospitals, are frequently resistant [7]. Oral streptococci isolated from chronically medicated Mexican patients are much more frequently resistant than those reported here: 37% are resistant to erythromycin and 8% to cipro£oxacin, for instance (C.F. Ama¤bile-Cuevas, unpublished data). An even higher prevalence of resistance (e.g. 14.5% penicillin-resistant, 38.5% erythromycin-resistant) among oral streptococci isolated from healthy Greek children, not taking antibiotics but attending an orthopaedic clinic, was recently reported [8]. Interestingly, Greece had the highest numbers of antibiotic prescriptions among 13 European countries [9]. We wished to investigate if resistance to antibiotics can recede with our interventions if use is better controlled. Since Mexico and Cuba are comparable societies that differ in their use of antibiotics, they serve as natural laboratories to begin answering this question. Table 1 Presence of erm genes in erythromycin-resistant and intermediate oral streptococci
ERY resistant erm detected ERY intermediate erm detected
Mexico
Cuba
10 6 11 5
4 1 18 5
erm genes detected by PCR (see text). The prevalence of erm genes between resistant and intermediate organisms, within each country, is not signi¢cantly di¡erent (P s 0.05, M2 test).
FEMSLE 10741 28-11-02
J.J. D|¤az-Mej|¤a et al. / FEMS Microbiology Letters 217 (2002) 173^176
CUBA MEXICO
40
isolates (%)
30 20 10 0
0 1 2 3 4 5 6 number of antibiotics isolates are resistant
Fig. 2. Multi-resistance pro¢les of Mexican and Cuban oral commensal streptococci. Only fully resistant organisms were included ; breakpoints as in Fig. 1.
Our data suggest that the maintenance of resistance in bacterial populations freed from antibiotic selection does not seem to correlate with overall changes in antibiotic usage patterns, at least between Mexico and Cuba. Few studies on commensal bacteria outside hospital settings are available for comparison. The volunteers we selected are among the least exposed to antibiotics in an urban population (see inclusion criteria above). At least in Mexico, such individuals are di⁄cult to ¢nd. In an unpublished survey, 4.2% of Mexicans have had antibiotics in a period of 2 weeks before the survey (C. Izazola, personal communication). It would be very interesting to assess the minimal exposure necessary to introduce resistance into the indigenous £ora. Could living with someone receiving antibiotic treatment, or having a single visit to a hospital, exert a change in such £ora? Once a resistance determinant is established, its prevalence does not seem a¡ected by overall antibiotic usage. This has been observed even in animal husbandry: after years of having eliminated the use of avoparcin, glycopeptide-resistant enterococci are still being detected in farms [10,11]. In our study, high sulfonamide and trimethoprim resistance frequencies, and carrying erm genes in the oral streptococcal isolates, are very similar despite di¡erences in antibiotic availability in Mexico and Cuba. Possibly the ¢tness cost of carrying resistance genes is negligible. Alternatively, other selective pressures might be more important. For instance, resistance genes can take other roles in bacteria or be linked to genes that would incur a greater cost to ¢tness if lost. Trying to ascertain the nature of such possible pressures, we asked the volunteers about the number of dental ¢llings they carried, as mercury released from dental amalgams has been proposed to co-select for antibiotic resistance [12]. Cubans bear slightly more ¢llings per individual than Mexicans do, but no relationship between the number of reported ¢llings and resistance prevalence
175
was found, nor any change in mercury resistance in the isolates (data not shown). Although the frequency of resistance to individual drugs does not seem to be a¡ected by the overall use of antibiotics in the community, the accumulation of resistance determinants might be increased by greater antibiotic use (Fig. 2). Mexican strains seem to be distributed in two groups, one susceptible to all or most drugs tested, and another resistant to three or more antibiotics. Antibiotic abuse might be selecting, in addition to resistant strains, isolates with the ability to develop and/or acquire resistance more readily. Multi-resistance is, indeed, one of the most feared aspects of antibiotic resistance, as it dramatically reduces the therapeutic options when treating infections. If the availability and abuse of antibiotics is actually selecting and/or maintaining multi-resistance, even in non-treated individuals, enforced policies of rational antibiotic use are de¢nitely needed. Also, further research is necessary on non-antibiotic agents and conditions that might be maintaining resistance determinants among bacterial populations.
Acknowledgements This work received partial support from a grant from NIH/NIAID to A. Salyers and S. Levy. We thank X. Garc|¤a, who made possible the sampling of Cuban volunteers, and A. Fuentes and G. Jime¤nez for helping with the sampling process. A. Summers, J. Davies, M. Roberts and M. Chicurel gave helpful comments, as well as the editor and reviewers.
References [1] Salyers, A.A. and Ama¤bile-Cuevas, C.F. (1997) Why are antibiotics resistance genes so resistant to elimination ? Antimicrob. Agents Chemother. 41, 2321^2325. [2] Heinemann, J.A., Ankenbauer, R.G. and Ama¤bile-Cuevas, C.F. (2000) Do antibiotics maintain antibiotic resistance? Drug Discov. Today 5, 195^204. [3] Levy, S.B. (1986) Ecology of antibiotic resistance determinants. In: Antibiotic Resistance Genes: Ecology, Transfer, and Expression (Levy, S.B. and Novick, R.P., Eds.), pp. 17^30. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. [4] Calva, J.J., Niebla-Pe¤rez, A., Rodr|¤guez-Lemoine, V., Santos, J.I. and Ama¤bile-Cuevas, C.F. (1996) Antibiotic usage and antibiotic resistance in Latin America. In: Antibiotic Resistance : From Molecular Basics to Therapeutic Options (Ama¤bile-Cuevas, C.F., Ed.), pp. 73^97. RG Landes/Chapman and Hall, Austin, TX. [5] National Committee for Clinical Laboratory Standards (1996) Performance Standards for Antimicrobial Disk Susceptibility Tests, 6th edn. NCCLS, Wayne, PA. [6] Arthur, M., Molinas, C. and Mabilat, C. (1993) PCR detection of erm erythromycin resistance genes by using degenerate oligonucleotide primers. In: Diagnostic Molecular Microbiology (Persing, D.H., Smith, T.F., Tenover, F.C. and White, T.J., Eds.), pp. 534^538. American Society for Microbiology, Washington, DC. [7] Calva, J.J., Sifuentes-Osornio, J. and Cero¤n, C. (1996) Antimicrobial
FEMSLE 10741 28-11-02
176
J.J. D|¤az-Mej|¤a et al. / FEMS Microbiology Letters 217 (2002) 173^176
resistance in fecal £ora: longitudinal community-based surveillance of children from urban Mexico. Antimicrob. Agents Chemother. 40, 1699^1702. [8] Ioannidou, S., Tassios, P.T., Kotsovili-Tseleni, A., Foustoukou, M., Legakis, N.J. and Vatopoulos, A. (2001) Antibiotic resistance rates and macrolide resistance phenotypes of viridans group streptococci from the oropharynx of healthy Greek children. Int. J. Antimicrob. Agents 17, 195^201. [9] Mo«lstad, S., Lundberg, C.S., Karlsson, A.-K. and Cars, O. (2002) Antibiotic prescription rates very markedly between 13 European countries. Scand. J. Infect. Dis. 34, 366^371. [10] Borgen, K., SYrum, M., Kruse, H. and Wasteson, Y. (2000) Persistence of vancomycin-resistant enterococci (VRE) on Norwegian broiler farms. FEMS Microbiol. Lett. 191, 255^258.
[11] Borgen, K., Simonsen, G.S., Sundsfjord, A., Wasteson, Y., Olsvik, Z. and Kruse, H. (2000) Continuing high prevalence of VanA-type vancomycin-resistant enterococci on Norwegian poultry farms three years after avoparcin was banned. J. Appl. Microbiol. 89, 478^485. [12] Summers, A.O., Wireman, J., Vimy, M.J., Lorscheider, F.L., Marshall, B., Levy, S.B., Bennett, S. and Billard, L. (1993) Mercury released from dental ‘silver’ ¢llings provokes an increase in mercuryand antibiotic-resistant bacteria in oral and intestinal £oras of primates. Antimicrob. Agents Chemother. 37, 825^834. [13] National Committee for Clinical Laboratory Standards (1998) Performance Standards for Antimicrobial Susceptibility Testing, 8th Informational Supplement. NCCLS, Wayne, PA.
FEMSLE 10741 28-11-02