Journal of Antimicrobial Chemotherapy Advance Access published July 3, 2015
J Antimicrob Chemother doi:10.1093/jac/dkv182
Declining macrolide resistance in Streptococcus pyogenes in Portugal (2007– 13) was accompanied by continuous clonal changes C. Silva-Costa, M. Ramirez* and J. Melo-Cristino on behalf of the Portuguese Group for the Study of Streptococcal Infections† Instituto de Microbiologia, Instituto de Medicina Molecular, Faculdade de Medicina, Universidade de Lisboa, Lisboa, Portugal *Corresponding author. Tel: +351-21-799-9460; Fax: +351-21-799-9459; E-mail: [email protected] †Members are listed in the Acknowledgements section.
Objectives: Macrolide resistance among Streptococcus pyogenes [group A streptococci (GAS)] in Portugal decreased between 1999 and 2006 and this decrease was accompanied by alterations in the prevalence of macrolide resistance phenotypes and clonal composition of the population. The aims of this study were to determine the macrolide resistance rate, resistance phenotypes and clones of GAS recovered from pharyngitis in 2007 –13 in Portugal. Methods: Antimicrobial susceptibility was tested by disc diffusion. Macrolide-resistant isolates were characterized by emm typing, T typing, PFGE profiling and MLST, and the presence of macrolide resistance determinants was determined by PCR. Results: We found continuing changes in macrolide resistance phenotypes and a persistent decline in overall erythromycin resistance, from 10% in 2007 to 1% in 2013. During this period there was a marked increase in emm11-ST403 cMLSB isolates, the disappearance of the emm3-ST315 M lineage and changes in the prevalence of previously identified GAS clones. Unexpectedly, the decline in erythromycin resistance and the decreasing prevalence of the MLSB phenotype were accompanied by a high consumption of long-acting and intermediate-acting macrolides, known to select for resistance and particularly for the erm(B) gene. Conclusions: The continuous decline in macrolide resistance detected since 2000, accompanied by a high clonal instability, emphasizes the importance of considering factors other than antibiotic consumption in explaining the prevalence of macrolide-resistant GAS.
Introduction Worldwide, there are large differences in macrolide resistance rates and in the prevalence of resistance phenotypes and clonal composition of macrolide-resistant group A streptococci (GAS) populations.1 Increases in macrolide resistance are often associated with high antimicrobial consumption and several studies have shown a correlation between decreases in macrolide resistance and decreases in antibiotic consumption. However, the circulating clones and the fluctuations in the clonal composition of the population can also be responsible for alterations in the prevalence of macrolide resistance and of macrolide resistance phenotypes.1 In Portugal, a steady decline in macrolide resistance in GAS was reported between 1999 and 2006, from 20% in 1999 to 12% in 2006, accompanied by fluctuations in macrolide resistance phenotypes and the clonal composition of the population.2 However, these fluctuations did not simply reflect fluctuations in the overall GAS population. In a study comparing macrolide-resistant and
-susceptible GAS populations in Portugal, we found that particular emm types and PFGE clusters were associated with macrolide resistance while others were associated with macrolide susceptibility, indicating that the macrolide-resistant population has its own dynamics.3 The aims of this study were to document potential changes in macrolide resistance and in the characteristics of resistant isolates during 2007 –13. We report ongoing fluctuations in macrolide resistance phenotypes and a continuing decline in overall erythromycin resistance, despite continued high macrolide use, and discuss their possible causes.
Methods Bacterial isolates and identification A total of 3634 Streptococcus pyogenes recovered from throat swabs and associated with a diagnosis of tonsillo-pharyngitis were collected from 32 microbiology laboratories in Portugal from January 2007 to December
# The Author 2015. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please e-mail: [email protected]
1 of 5
Downloaded from http://jac.oxfordjournals.org/ at University of Virginia on July 12, 2015
Received 24 February 2015; returned 26 April 2015; revised 7 May 2015; accepted 8 June 2015
Silva-Costa et al.
2013. The isolates were recovered mostly from paediatric patients (,18 years). The laboratories were asked to submit all non-duplicate S. pyogenes isolated from outpatients during the study period. The isolates were distributed in the study period as follows: 519 in 2007, 491 in 2008, 551 in 2009, 567 in 2010, 513 in 2011, 493 in 2012 and 500 in 2013. Isolates were identified to the species level as previously described.2 In this collection, 161 isolates (4.4%) were erythromycin resistant, and only these isolates were characterized further.
STs were determined using the goeBURST7 algorithm implemented in PHYLOViZ,8 with the complete S. pyogenes database available at spyogenes.mlst.net.
Antimicrobial susceptibility testing and macrolide resistance phenotype
Results
Antimicrobial susceptibility testing was performed and the macrolide resistance phenotype was determined as previously described.2,4
Statistical analysis Trends in macrolide resistance were evaluated using the Cochran – Armitage test for trend.9
Antimicrobial susceptibility testing and macrolide and tetracycline resistance phenotype
PFGE and MLST
Identification of resistance determinants
PCR determination of the macrolide and tetracycline resistance genotype Total bacterial DNA was isolated according to the methodology described by the CDC (http://www.cdc.gov/streplab/protocol-emm-type.html). PCRs to determine which of the macrolide and tetracycline resistance determinants were present were performed as described previously.2
T and emm typing
5
PFGE was performed as previously described. All isolates were analysed by MLST as previously described6 and allele and ST were attributed using the MLST database (spyogenes.mlst.net). The relationships between
All isolates tested carried a single macrolide resistance determinant (Table 1). In six of the tetracycline-resistant isolates we were unable to detect any of the resistance determinants tested.
40 Erythromycin resistance
35
MLSB phenotype
Prevalence (%)
30
M phenotype
25 20 15 10 5
0 50
3
n=
3
49
51
n=
7 56
n=
1 55
n=
1
n=
n=
n=
51
9
1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 49
0
Figure 1. Erythromycin resistance and prevalence of macrolide resistance phenotypes in Portugal during the period 1998– 2013. Open circles and the continuous line represent the proportion of erythromycin-resistant GAS among those causing pharyngitis. The broken and dotted lines represent the proportions of each phenotype in the population. Filled triangles represent the proportion of isolates of the M phenotype. Open squares represent the proportion of isolates of the MLSB phenotype. The numbers below each year between 2007 and 2013 represent the total number of pharyngeal GAS isolates recovered. The shaded area represents previously published data.2,19
2 of 5
Downloaded from http://jac.oxfordjournals.org/ at University of Virginia on July 12, 2015
T typing (Group A, T-typing Antisera; Denka Seiken, Tokyo, Japan) was performed according to the manufacturer’s instructions. emm typing was performed as described by the CDC (http://www.cdc.gov/streplab/ protocol-emm-type.html).
During the 7 years of the study, the overall rate of erythromycin resistance was 4.4%, a lower value than the one reported in 2004 – 06 (13.2%). A decreasing trend in macrolide resistance was noted in this period (Figure 1; from 10% to 1%, Cochran – Armitage test for trend, P,0.001), which was due to the decline in isolates presenting both macrolide resistance phenotypes (P, 0.001 and P ¼ 0.004 for MLSB and M isolates, respectively; Figure 1). The majority (n ¼ 99) of the 161 macrolide-resistant isolates (61.5%) presented the cMLSB phenotype, 52 isolates (32.3%) presented the M phenotype and 10 isolates (6.2%) presented the iMLSB phenotype. Resistance to tetracycline was found in 57% (n¼ 92) of the macrolide-resistant isolates. Resistance to bacitracin was found in 12 isolates, all expressing the cMLSB phenotype.
Erythromycin-resistant GAS
Table 1. Characteristics of the 161 macrolide-resistant GAS in Portugal emm type (no. of isolates) 11 (62)
12 (25) 22 (16) 28 (12) 77 (9) 4 (8) 1 (7) 75 (6) 6 (5) Multiplee
PFGE clustera A56 (54) I7 (7) other (1) B17 (13) other (12) D11 (11) K5 (5) C12 (11) other (1) E9 (7) other (2) G8 (8) H7 (7) J6 (6) F8 (5) multiplef
T type (no. of isolates)b
Macrolide resistance phenotype [genotype] (no. of isolates)
11 (51), NT (2), 9 (1) 11 (7) 11 (1) 12 (10), NT (2), 9 (1) 12 (10), NT (2) 12 (11) 12 (5) 28 (11) 28 (1) 28 (4), 13 (3) 13 (2) 4 (8) 1 (7) 25 (6) 6 (4), 11 (1) multipleg
cMLSB [erm(B)] (54) cMLSB [erm(B)] (7) cMLSB [erm(B)] (1) M [mef(A)] (10) cMLSB [erm(B)] (3) M [mef(A)] (10) [mef(E)] (1); cMLSB [erm(B)] (1) (53) cMLSB [erm(B)] (11) cMLSB [erm(B)] (5) cMLSB [erm(B)] (11) cMLSB [erm(B)] (1) iMLSB [erm(TR)] (7) iMLSB [erm(TR)] (2) M [mef(A)] (8) M [mef(A)] (7) M [mef(A)] (6) M [mef(A)] (4), cMLSB [erm(B)] (1) cMLSB [erm(B)] (4), iMLSB [erm(TR)] (1); M [mef(A)] (6)
Tetracycline resistance phenotype [genotype] (no. of isolates)c R [tet(M)] (50), [ND] (3); S (1) R [tet(M)] (7) R [tet(M)] (1) S (13) R [tet(M)] (1); S (11) R [tet(M)] (10), [ND] (1) R [tet(M)] (4), [ND] (1) S (11) S (1) R [tet(O)] (7) R [tet(O)] (2) S (8) S (7) S (6) R [ND] (1); S (4) R [tet(M)] (4); S (7)
ST (no. of isolates)d [403 (49), 562 (5)] 403 (7) 403 (1) 36 (13) 36 (12) 46 (11) 46 (5) 52 (11) 52 (1) 63 (7) 63 (2) [39 (7), 658 (1)] [28 (2), 618 (1), 659 (4)] [657 (5), 150 (1)] [382 (3), 411 (1), 681 (1)] multipleh
a Each major lineage (n≥ 5) is designated by a capital letter, followed by a subscript number that indicates the number of isolates grouped in that lineage; whenever a PFGE cluster included ,5 isolates of that particular emm type, it was not discriminated and it was designated by ‘other’. b NT, non-typeable. c ND, not determined; did not yield any amplification product for the genes tested; R, resistant; S, susceptible. d STs enclosed by square brackets belong to the same clonal complex defined by the goeBURST algorithm (http://goeburst.phyloviz.net/), with the complete Spyogenes database available at spyogenes.mlst.net. In all cases, isolates of the same emm type presented the same ST or STs of the same clonal complex. e emm type (no. of isolates): 9 (4), 44 (1), 3 (1), 73 (1), 78 (1), 87 (1), 58 (1) and 2 (1). f Nine PFGE clusters. g T type (no. of isolates): 9 (5), 11 (1), 28 (1), 5/27/44 (1), 2 (1) and NT (2). h ST (no. of isolates): 75 (4), 406 (1), 25 (1), 409 (1), 331 (1), 62 (1), 55 (1) and 682 (1).
JAC
3 of 5 Downloaded from http://jac.oxfordjournals.org/ at University of Virginia on July 12, 2015
Silva-Costa et al.
Clonal characterization
Discussion The decreasing trend in macrolide resistance already detected in the period 1999–20062 continued in this study period (2007 –13), reaching the lowest value of macrolide resistance in GAS recorded in Portugal (1% in 2013). Ecological studies have associated changes in macrolide consumption with the prevalence of resistant isolates.10 – 12 Macrolide consumption in Portugal increased by 35% from 3.20 DDDs/1000 inhabitants/day in 1998 to a peak of 4.34 in 2005.13 From then on, consumption remained high and approximately constant until 2009.13 Surprisingly, this was accompanied by a decrease in overall macrolide resistance in GAS, as documented here. It could be argued that there was insufficient selective pressure to maintain a high macrolide resistance rate in GAS, but consumption in Portugal has remained well above the value of 2 DDDs/1000 inhabitants/day believed to trigger an increase in macrolideresistant GAS.12 Although we have no data concerning macrolide consumption since 2009, it seems unlikely that a rapid decline in macrolide use, especially of long- and intermediate-acting macrolides, would explain the variations seen in 2009–13. The reasons for this are 2-fold: (i) any changes in macrolide use are expected to have a delayed effect on resistance rate;12 and (ii) mathematical models predict that, once resistance has been selected, only drastic reductions in antibiotic use can be hoped to lead to reduced resistance rates,14 and this was certainly not the case. In contrast, we would contend that overall macrolide consumption must have remained high and at levels of use above the threshold for selection of resistance, as was documented until 2009. Recently, decreases in macrolide resistance rates in European countries have been reported, although different explanations for this decline have been presented. While in some countries, such as France and Belgium, a decline in the consumption of these antibiotics accompanied the decrease in macrolide resistance, in other countries, such as Spain, Greece and Portugal, a decline in macrolide resistance has been reported without considerable changes in macrolide consumption.1 Moreover, even decreasing levels of macrolide consumption can be accompanied by increases in macrolide resistance, as was the case in Slovenia.1 Changes in clonal composition, already described in the period 1998 –2006,2,5 continued in the period 2007 – 13. The most notable changes were the increase in emm11-ST403 isolates and
4 of 5
Acknowledgements Members of the Portuguese Group for the Study of Streptococcal Infections Teresa Vaz, Marı´lia Gia˜o, Rui Ferreira, Iryna Klyeshtorna (Centro Hospitalar do Barlavento Algarvio), Ana Buschy Fonseca (Hospital de Cascais), Henrique Oliveira (Centro Hospitalar de Coimbra), Ana Cristina Silva, Hermı´nia Costa, Maria Fa´tima Silva, Maria Ame´lia Afonso (Centro Hospitalar de Entre Douro e Vouga), Margarida Pinto, Odete Chantre, Joa˜o Marques, Isabel Peres, Isabel Daniel, Teresa Ferreira, Isabel Lourenc¸o, Cristina Marcelo (Centro Hospitalar de Lisboa Central), Lurdes Monteiro, Luı´s Marques Lito (Centro Hospitalar Lisboa Norte), Teresa Marques, Cristina Toscano, Judite Batista, Teresa Morais, Maria Ana Pessanha, Elsa Gonc¸alves (Centro Hospitalar Lisboa Ocidental), Paulo Lopes, Luı´sa Felı´cio, Angelina Lameira˜o (Centro Hospitalar de Vila Nova de Gaia/Espinho), Ana Paula Mota, Margarida Tomaz, Hermı´nia Marques (Centro Hospitalar do Alto Ave), Maria Helena Ramos, Ana Paula Castro (Centro Hospitalar do Porto), Fernando Fonseca, Carla Leite, Sı´lvia Reis (Centro Hospitalar da Po´voa do Varzim/Vila do Conde), Grac¸a Ribeiro, Luı´sa Boaventura, Catarina Chaves, Teresa Reis (Hospitais da Universidade de Coimbra), Nuno Canhoto, Teresa Afonso (Hospital Central do Funchal), Teresa Pina, Helena Peres (Hospital Curry Cabral, Lisboa), Ilse Fontes, Paulo Martinho (Hospital de Santa Luzia, Elvas), Ana Domingos, Gina Marra˜o, Jose´ Grossinho (Hospital de Santo Andre´, Leiria), Manuela Ribeiro, Helena Gonc¸alves (Hospital de Sa˜o Joa˜o, Porto), Maria Alberta Faustino, Maria Carmen Iglesias, Adelaide Alves (Hospital de Braga), Maria Paula
Downloaded from http://jac.oxfordjournals.org/ at University of Virginia on July 12, 2015
The characteristics of the 161 GAS isolates included in this study are summarized in Table 1. PFGE analysis identified 11 major lineages, including 146 isolates (90.7%). MLST analysis identified three novel alleles among the genes used in MLST, one in gki (gki116) and two in gtr (gtr91 and gtr92). The gtr allele 92 presented a 1 nt deletion that was predicted to result in a truncated form of the protein. New allele combinations producing novel STs were also noted and were assigned numbers ST657, ST658, ST659, ST681 and ST682. The new alleles and the novel STs were submitted to the S. pyogenes MLST database (spyogenes.mlst.net). The prevalence of the major emm types was not stable during the study period (Figure S1, available as Supplementary data at JAC Online); this instability was also noted in the previously studied period (1998– 2006).2,5
the disappearance of the emm3-ST315 lineage. Overall, the genetic lineages identified among macrolide-resistant GAS in Portugal appear to be disseminated in other European countries.15 – 17 The situation documented here is similar to that in neighbouring Spain, where a steady decline in macrolide resistance from 2005 onwards was detected, without considerable changes in consumption of this antibiotic. However, in contrast to Portugal, the decline in macrolide resistance in Spain was attributed to the dynamic behaviour of four major clones, all also identified among macrolide-resistant GAS in Portugal: emm12-ST36, emm11-ST403, emm28-ST52 and emm4-ST39.17 Analysis of the prevalence of the major emm types circulating in Portugal (Figure S1 and data not shown) indicated that between 1998 and 2013 each emm type was dominant for periods of 3 – 4 years, without a clear increase or decrease in the proportion of a single clone accompanying the decrease in macrolide resistance. Macrolide consumption still stands as the main driver of macrolide resistance in streptococci.10 – 12 However, the association between clonal fluctuations in the population and macrolide resistance phenotypes must also play a key role in the dynamics of the macrolide-resistant GAS population, as already recognized.1 Here, we documented a decline in macrolide resistance among GAS associated with pharyngitis to a level comparable to the lowest in Europe (including countries with low macrolide consumption), in spite of high macrolide use.10,13,18 These changes were accompanied by variations in the clonal composition of the population, which are part of the natural dynamics of macrolideresistant GAS and independent of the behaviour of the general GAS population, reinforcing the importance of considering both the clonal structure of the bacterial population and antibiotic consumption when explaining the prevalence of resistant isolates.
Erythromycin-resistant GAS
Pinheiro, R. Semedo (Hospital Dr Jose´ Maria Grande, Portalegre), Luı´sa Gonc¸alves, Olga Neto (Hospital dos SAMS, Lisboa), Luı´sa Sancho (Hospital Dr Fernando da Fonseca, Amadora/Sintra), Jose´ Diogo, Ana Rodrigues (Hospital Garcia de Orta, Almada), Elmano Ramalheira, Raquel Diaz, So´nia Ferreira (Centro Hospitalar Baixo Vouga), Jose´ Miguel Ribeiro, Isabel Vale, Ana Carvalho (Centro Hospitalar de Tondela—Viseu), Maria Anto´nia Read, Margarida Monteiro, Valquı´ria Alves (Hospital Pedro Hispano, Matosinhos).
Funding This work was partially supported by Fundac¸a˜o para a Cieˆncia e Tecnologia, Portugal (PTDC/SAU-ESA/72321/2006).
Transparency declarations
Supplementary data Figure S1 is available as Supplementary data at JAC Online (http://jac. oxfordjournals.org/).
References 1 Silva-Costa C, Fria˜es A, Ramirez M et al. Macrolide-resistant Streptococcus pyogenes: prevalence and treatment strategies. Expert Rev Anti Infect Ther 2015; 13: 615–28.
6 Enright MC, Spratt BG, Kalia A et al. Multilocus sequence typing of Streptococcus pyogenes and the relationships between emm type and clone. Infect Immun 2001; 69: 2416–27. 7 Francisco A, Bugalho M, Ramirez M et al. Global optimal eBURST analysis of multilocus typing data using a graphic matroid approach. BMC Bioinformatics 2009; 10: 152. 8 Francisco AP, Vaz C, Monteiro PT et al. PHYLOViZ: phylogenetic inference and data visualization for sequence based typing methods. BMC Bioinformatics 2012; 13: 87. 9 Altman DG. Practical Statistics for Medical Research. Boca Raton, FL: Chapman & Hall/CRC, 1999. 10 Van Heirstraeten L, Coenen S, Lammens C et al. Antimicrobial drug use and macrolide-resistant Streptococcus pyogenes, Belgium. Emerg Infect Dis 2012; 18: 1515– 8. 11 Seppa¨la¨ H, Klaukka T, Vuopio-Varkila J et al. The effect of changes in the consumption of macrolide antibiotics on erythromycin resistance in group A streptococci in Finland. Finnish Study Group for Antimicrobial Resistance. N Engl J Med 1997; 337: 441–6. 12 Granizo JJ, Aguilar L, Casal J et al. Streptococcus pyogenes resistance to erythromycin in relation to macrolide consumption in Spain (1986–1997). J Antimicrob Chemother 2000; 46: 959–64. 13 Adriaenssens N, Coenen S, Versporten A et al. European Surveillance of Antimicrobial Consumption (ESAC): outpatient macrolide, lincosamide and streptogramin (MLS) use in Europe (1997– 2009). J Antimicrob Chemother 2011; 66 Suppl 6: vi37–45. 14 Austin DJ, Kristinsson KG, Anderson RM. The relationship between the volume of antimicrobial consumption in human communities and the frequency of resistance. Proc Natl Acad Sci USA 1999; 96: 1152 –6. 15 Malli E, Tatsidou E, Damani A et al. Macrolide-resistant Streptococcus pyogenes in Central Greece: prevalence; mechanism and molecular identification. Int J Antimicrob Agents 2010; 35: 614– 5.
2 Silva-Costa C, Pinto FR, Ramirez M et al. Decrease in macrolide resistance and clonal instability among Streptococcus pyogenes in Portugal. Clin Microbiol Infect 2008; 14: 1152 –9.
16 Ardanuy C, Domenech A, Rolo D et al. Molecular characterization of macrolide- and multidrug-resistant Streptococcus pyogenes isolated from adult patients in Barcelona, Spain (1993 – 2008). J Antimicrob Chemother 2010; 65: 634–43.
3 Silva-Costa C, Fria˜es A, Ramirez M et al. Differences between macrolide resistant and susceptible Streptococcus pyogenes: the importance of clonal properties in addition to antibiotic consumption. Antimicrob Agents Chemother 2012; 56: 5661 –6.
17 Montes M, Tamayo E, Mojica C et al. What causes decreased erythromycin resistance in Streptococcus pyogenes? Dynamics of four clones in a Southern European region from 2005 to 2012. J Antimicrob Chemother 2014; 69: 1474– 82.
4 Melo-Cristino J, Fernandes ML. Streptococcus pyogenes isolated in Portugal: macrolide resistance phenotypes and correlation with T types. Portuguese Surveillance Group for the Study of Respiratory Pathogens. Microb Drug Resist 1999; 5: 219–25.
18 Richter SS, Heilmann KP, Dohrn CL et al. Increasing telithromycin resistance among Streptococcus pyogenes in Europe. J Antimicrob Chemother 2008; 61: 603–11.
5 Silva-Costa C, Ramirez M, Melo-Cristino J. Identification of macrolideresistant clones of Streptococcus pyogenes in Portugal. Clin Microbiol Infect 2006; 12: 513– 8.
19 Silva-Costa C, Ramirez M, Melo-Cristino J et al. Rapid inversion of the prevalences of macrolide resistance phenotypes paralleled by a diversification of T and emm types among Streptococcus pyogenes in Portugal. Antimicrob Agents Chemother 2005; 49: 2109– 11.
5 of 5
Downloaded from http://jac.oxfordjournals.org/ at University of Virginia on July 12, 2015
M. R. has received honoraria for serving on the speaker’s bureau of Pfizer and for consulting from GlaxoSmithKline. J. M.-C. has received research grants administered through his university and received honoraria for serving on the speaker’s bureaus of Pfizer, Novartis and Gilead. C. S.-C.: none to declare. No company or financing body was involved in the decision to publish.
JAC
J Antimicrob Chemother doi:10.1093/jac/dkv182
Declining macrolide resistance in Streptococcus pyogenes in Portugal (2007– 13) was accompanied by continuous clonal changes C. Silva-Costa, M. Ramirez* and J. Melo-Cristino on behalf of the Portuguese Group for the Study of Streptococcal Infections† Instituto de Microbiologia, Instituto de Medicina Molecular, Faculdade de Medicina, Universidade de Lisboa, Lisboa, Portugal *Corresponding author. Tel: +351-21-799-9460; Fax: +351-21-799-9459; E-mail: [email protected] †Members are listed in the Acknowledgements section.
Objectives: Macrolide resistance among Streptococcus pyogenes [group A streptococci (GAS)] in Portugal decreased between 1999 and 2006 and this decrease was accompanied by alterations in the prevalence of macrolide resistance phenotypes and clonal composition of the population. The aims of this study were to determine the macrolide resistance rate, resistance phenotypes and clones of GAS recovered from pharyngitis in 2007 –13 in Portugal. Methods: Antimicrobial susceptibility was tested by disc diffusion. Macrolide-resistant isolates were characterized by emm typing, T typing, PFGE profiling and MLST, and the presence of macrolide resistance determinants was determined by PCR. Results: We found continuing changes in macrolide resistance phenotypes and a persistent decline in overall erythromycin resistance, from 10% in 2007 to 1% in 2013. During this period there was a marked increase in emm11-ST403 cMLSB isolates, the disappearance of the emm3-ST315 M lineage and changes in the prevalence of previously identified GAS clones. Unexpectedly, the decline in erythromycin resistance and the decreasing prevalence of the MLSB phenotype were accompanied by a high consumption of long-acting and intermediate-acting macrolides, known to select for resistance and particularly for the erm(B) gene. Conclusions: The continuous decline in macrolide resistance detected since 2000, accompanied by a high clonal instability, emphasizes the importance of considering factors other than antibiotic consumption in explaining the prevalence of macrolide-resistant GAS.
Introduction Worldwide, there are large differences in macrolide resistance rates and in the prevalence of resistance phenotypes and clonal composition of macrolide-resistant group A streptococci (GAS) populations.1 Increases in macrolide resistance are often associated with high antimicrobial consumption and several studies have shown a correlation between decreases in macrolide resistance and decreases in antibiotic consumption. However, the circulating clones and the fluctuations in the clonal composition of the population can also be responsible for alterations in the prevalence of macrolide resistance and of macrolide resistance phenotypes.1 In Portugal, a steady decline in macrolide resistance in GAS was reported between 1999 and 2006, from 20% in 1999 to 12% in 2006, accompanied by fluctuations in macrolide resistance phenotypes and the clonal composition of the population.2 However, these fluctuations did not simply reflect fluctuations in the overall GAS population. In a study comparing macrolide-resistant and
-susceptible GAS populations in Portugal, we found that particular emm types and PFGE clusters were associated with macrolide resistance while others were associated with macrolide susceptibility, indicating that the macrolide-resistant population has its own dynamics.3 The aims of this study were to document potential changes in macrolide resistance and in the characteristics of resistant isolates during 2007 –13. We report ongoing fluctuations in macrolide resistance phenotypes and a continuing decline in overall erythromycin resistance, despite continued high macrolide use, and discuss their possible causes.
Methods Bacterial isolates and identification A total of 3634 Streptococcus pyogenes recovered from throat swabs and associated with a diagnosis of tonsillo-pharyngitis were collected from 32 microbiology laboratories in Portugal from January 2007 to December
# The Author 2015. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please e-mail: [email protected]
1 of 5
Downloaded from http://jac.oxfordjournals.org/ at University of Virginia on July 12, 2015
Received 24 February 2015; returned 26 April 2015; revised 7 May 2015; accepted 8 June 2015
Silva-Costa et al.
2013. The isolates were recovered mostly from paediatric patients (,18 years). The laboratories were asked to submit all non-duplicate S. pyogenes isolated from outpatients during the study period. The isolates were distributed in the study period as follows: 519 in 2007, 491 in 2008, 551 in 2009, 567 in 2010, 513 in 2011, 493 in 2012 and 500 in 2013. Isolates were identified to the species level as previously described.2 In this collection, 161 isolates (4.4%) were erythromycin resistant, and only these isolates were characterized further.
STs were determined using the goeBURST7 algorithm implemented in PHYLOViZ,8 with the complete S. pyogenes database available at spyogenes.mlst.net.
Antimicrobial susceptibility testing and macrolide resistance phenotype
Results
Antimicrobial susceptibility testing was performed and the macrolide resistance phenotype was determined as previously described.2,4
Statistical analysis Trends in macrolide resistance were evaluated using the Cochran – Armitage test for trend.9
Antimicrobial susceptibility testing and macrolide and tetracycline resistance phenotype
PFGE and MLST
Identification of resistance determinants
PCR determination of the macrolide and tetracycline resistance genotype Total bacterial DNA was isolated according to the methodology described by the CDC (http://www.cdc.gov/streplab/protocol-emm-type.html). PCRs to determine which of the macrolide and tetracycline resistance determinants were present were performed as described previously.2
T and emm typing
5
PFGE was performed as previously described. All isolates were analysed by MLST as previously described6 and allele and ST were attributed using the MLST database (spyogenes.mlst.net). The relationships between
All isolates tested carried a single macrolide resistance determinant (Table 1). In six of the tetracycline-resistant isolates we were unable to detect any of the resistance determinants tested.
40 Erythromycin resistance
35
MLSB phenotype
Prevalence (%)
30
M phenotype
25 20 15 10 5
0 50
3
n=
3
49
51
n=
7 56
n=
1 55
n=
1
n=
n=
n=
51
9
1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 49
0
Figure 1. Erythromycin resistance and prevalence of macrolide resistance phenotypes in Portugal during the period 1998– 2013. Open circles and the continuous line represent the proportion of erythromycin-resistant GAS among those causing pharyngitis. The broken and dotted lines represent the proportions of each phenotype in the population. Filled triangles represent the proportion of isolates of the M phenotype. Open squares represent the proportion of isolates of the MLSB phenotype. The numbers below each year between 2007 and 2013 represent the total number of pharyngeal GAS isolates recovered. The shaded area represents previously published data.2,19
2 of 5
Downloaded from http://jac.oxfordjournals.org/ at University of Virginia on July 12, 2015
T typing (Group A, T-typing Antisera; Denka Seiken, Tokyo, Japan) was performed according to the manufacturer’s instructions. emm typing was performed as described by the CDC (http://www.cdc.gov/streplab/ protocol-emm-type.html).
During the 7 years of the study, the overall rate of erythromycin resistance was 4.4%, a lower value than the one reported in 2004 – 06 (13.2%). A decreasing trend in macrolide resistance was noted in this period (Figure 1; from 10% to 1%, Cochran – Armitage test for trend, P,0.001), which was due to the decline in isolates presenting both macrolide resistance phenotypes (P, 0.001 and P ¼ 0.004 for MLSB and M isolates, respectively; Figure 1). The majority (n ¼ 99) of the 161 macrolide-resistant isolates (61.5%) presented the cMLSB phenotype, 52 isolates (32.3%) presented the M phenotype and 10 isolates (6.2%) presented the iMLSB phenotype. Resistance to tetracycline was found in 57% (n¼ 92) of the macrolide-resistant isolates. Resistance to bacitracin was found in 12 isolates, all expressing the cMLSB phenotype.
Erythromycin-resistant GAS
Table 1. Characteristics of the 161 macrolide-resistant GAS in Portugal emm type (no. of isolates) 11 (62)
12 (25) 22 (16) 28 (12) 77 (9) 4 (8) 1 (7) 75 (6) 6 (5) Multiplee
PFGE clustera A56 (54) I7 (7) other (1) B17 (13) other (12) D11 (11) K5 (5) C12 (11) other (1) E9 (7) other (2) G8 (8) H7 (7) J6 (6) F8 (5) multiplef
T type (no. of isolates)b
Macrolide resistance phenotype [genotype] (no. of isolates)
11 (51), NT (2), 9 (1) 11 (7) 11 (1) 12 (10), NT (2), 9 (1) 12 (10), NT (2) 12 (11) 12 (5) 28 (11) 28 (1) 28 (4), 13 (3) 13 (2) 4 (8) 1 (7) 25 (6) 6 (4), 11 (1) multipleg
cMLSB [erm(B)] (54) cMLSB [erm(B)] (7) cMLSB [erm(B)] (1) M [mef(A)] (10) cMLSB [erm(B)] (3) M [mef(A)] (10) [mef(E)] (1); cMLSB [erm(B)] (1) (53) cMLSB [erm(B)] (11) cMLSB [erm(B)] (5) cMLSB [erm(B)] (11) cMLSB [erm(B)] (1) iMLSB [erm(TR)] (7) iMLSB [erm(TR)] (2) M [mef(A)] (8) M [mef(A)] (7) M [mef(A)] (6) M [mef(A)] (4), cMLSB [erm(B)] (1) cMLSB [erm(B)] (4), iMLSB [erm(TR)] (1); M [mef(A)] (6)
Tetracycline resistance phenotype [genotype] (no. of isolates)c R [tet(M)] (50), [ND] (3); S (1) R [tet(M)] (7) R [tet(M)] (1) S (13) R [tet(M)] (1); S (11) R [tet(M)] (10), [ND] (1) R [tet(M)] (4), [ND] (1) S (11) S (1) R [tet(O)] (7) R [tet(O)] (2) S (8) S (7) S (6) R [ND] (1); S (4) R [tet(M)] (4); S (7)
ST (no. of isolates)d [403 (49), 562 (5)] 403 (7) 403 (1) 36 (13) 36 (12) 46 (11) 46 (5) 52 (11) 52 (1) 63 (7) 63 (2) [39 (7), 658 (1)] [28 (2), 618 (1), 659 (4)] [657 (5), 150 (1)] [382 (3), 411 (1), 681 (1)] multipleh
a Each major lineage (n≥ 5) is designated by a capital letter, followed by a subscript number that indicates the number of isolates grouped in that lineage; whenever a PFGE cluster included ,5 isolates of that particular emm type, it was not discriminated and it was designated by ‘other’. b NT, non-typeable. c ND, not determined; did not yield any amplification product for the genes tested; R, resistant; S, susceptible. d STs enclosed by square brackets belong to the same clonal complex defined by the goeBURST algorithm (http://goeburst.phyloviz.net/), with the complete Spyogenes database available at spyogenes.mlst.net. In all cases, isolates of the same emm type presented the same ST or STs of the same clonal complex. e emm type (no. of isolates): 9 (4), 44 (1), 3 (1), 73 (1), 78 (1), 87 (1), 58 (1) and 2 (1). f Nine PFGE clusters. g T type (no. of isolates): 9 (5), 11 (1), 28 (1), 5/27/44 (1), 2 (1) and NT (2). h ST (no. of isolates): 75 (4), 406 (1), 25 (1), 409 (1), 331 (1), 62 (1), 55 (1) and 682 (1).
JAC
3 of 5 Downloaded from http://jac.oxfordjournals.org/ at University of Virginia on July 12, 2015
Silva-Costa et al.
Clonal characterization
Discussion The decreasing trend in macrolide resistance already detected in the period 1999–20062 continued in this study period (2007 –13), reaching the lowest value of macrolide resistance in GAS recorded in Portugal (1% in 2013). Ecological studies have associated changes in macrolide consumption with the prevalence of resistant isolates.10 – 12 Macrolide consumption in Portugal increased by 35% from 3.20 DDDs/1000 inhabitants/day in 1998 to a peak of 4.34 in 2005.13 From then on, consumption remained high and approximately constant until 2009.13 Surprisingly, this was accompanied by a decrease in overall macrolide resistance in GAS, as documented here. It could be argued that there was insufficient selective pressure to maintain a high macrolide resistance rate in GAS, but consumption in Portugal has remained well above the value of 2 DDDs/1000 inhabitants/day believed to trigger an increase in macrolideresistant GAS.12 Although we have no data concerning macrolide consumption since 2009, it seems unlikely that a rapid decline in macrolide use, especially of long- and intermediate-acting macrolides, would explain the variations seen in 2009–13. The reasons for this are 2-fold: (i) any changes in macrolide use are expected to have a delayed effect on resistance rate;12 and (ii) mathematical models predict that, once resistance has been selected, only drastic reductions in antibiotic use can be hoped to lead to reduced resistance rates,14 and this was certainly not the case. In contrast, we would contend that overall macrolide consumption must have remained high and at levels of use above the threshold for selection of resistance, as was documented until 2009. Recently, decreases in macrolide resistance rates in European countries have been reported, although different explanations for this decline have been presented. While in some countries, such as France and Belgium, a decline in the consumption of these antibiotics accompanied the decrease in macrolide resistance, in other countries, such as Spain, Greece and Portugal, a decline in macrolide resistance has been reported without considerable changes in macrolide consumption.1 Moreover, even decreasing levels of macrolide consumption can be accompanied by increases in macrolide resistance, as was the case in Slovenia.1 Changes in clonal composition, already described in the period 1998 –2006,2,5 continued in the period 2007 – 13. The most notable changes were the increase in emm11-ST403 isolates and
4 of 5
Acknowledgements Members of the Portuguese Group for the Study of Streptococcal Infections Teresa Vaz, Marı´lia Gia˜o, Rui Ferreira, Iryna Klyeshtorna (Centro Hospitalar do Barlavento Algarvio), Ana Buschy Fonseca (Hospital de Cascais), Henrique Oliveira (Centro Hospitalar de Coimbra), Ana Cristina Silva, Hermı´nia Costa, Maria Fa´tima Silva, Maria Ame´lia Afonso (Centro Hospitalar de Entre Douro e Vouga), Margarida Pinto, Odete Chantre, Joa˜o Marques, Isabel Peres, Isabel Daniel, Teresa Ferreira, Isabel Lourenc¸o, Cristina Marcelo (Centro Hospitalar de Lisboa Central), Lurdes Monteiro, Luı´s Marques Lito (Centro Hospitalar Lisboa Norte), Teresa Marques, Cristina Toscano, Judite Batista, Teresa Morais, Maria Ana Pessanha, Elsa Gonc¸alves (Centro Hospitalar Lisboa Ocidental), Paulo Lopes, Luı´sa Felı´cio, Angelina Lameira˜o (Centro Hospitalar de Vila Nova de Gaia/Espinho), Ana Paula Mota, Margarida Tomaz, Hermı´nia Marques (Centro Hospitalar do Alto Ave), Maria Helena Ramos, Ana Paula Castro (Centro Hospitalar do Porto), Fernando Fonseca, Carla Leite, Sı´lvia Reis (Centro Hospitalar da Po´voa do Varzim/Vila do Conde), Grac¸a Ribeiro, Luı´sa Boaventura, Catarina Chaves, Teresa Reis (Hospitais da Universidade de Coimbra), Nuno Canhoto, Teresa Afonso (Hospital Central do Funchal), Teresa Pina, Helena Peres (Hospital Curry Cabral, Lisboa), Ilse Fontes, Paulo Martinho (Hospital de Santa Luzia, Elvas), Ana Domingos, Gina Marra˜o, Jose´ Grossinho (Hospital de Santo Andre´, Leiria), Manuela Ribeiro, Helena Gonc¸alves (Hospital de Sa˜o Joa˜o, Porto), Maria Alberta Faustino, Maria Carmen Iglesias, Adelaide Alves (Hospital de Braga), Maria Paula
Downloaded from http://jac.oxfordjournals.org/ at University of Virginia on July 12, 2015
The characteristics of the 161 GAS isolates included in this study are summarized in Table 1. PFGE analysis identified 11 major lineages, including 146 isolates (90.7%). MLST analysis identified three novel alleles among the genes used in MLST, one in gki (gki116) and two in gtr (gtr91 and gtr92). The gtr allele 92 presented a 1 nt deletion that was predicted to result in a truncated form of the protein. New allele combinations producing novel STs were also noted and were assigned numbers ST657, ST658, ST659, ST681 and ST682. The new alleles and the novel STs were submitted to the S. pyogenes MLST database (spyogenes.mlst.net). The prevalence of the major emm types was not stable during the study period (Figure S1, available as Supplementary data at JAC Online); this instability was also noted in the previously studied period (1998– 2006).2,5
the disappearance of the emm3-ST315 lineage. Overall, the genetic lineages identified among macrolide-resistant GAS in Portugal appear to be disseminated in other European countries.15 – 17 The situation documented here is similar to that in neighbouring Spain, where a steady decline in macrolide resistance from 2005 onwards was detected, without considerable changes in consumption of this antibiotic. However, in contrast to Portugal, the decline in macrolide resistance in Spain was attributed to the dynamic behaviour of four major clones, all also identified among macrolide-resistant GAS in Portugal: emm12-ST36, emm11-ST403, emm28-ST52 and emm4-ST39.17 Analysis of the prevalence of the major emm types circulating in Portugal (Figure S1 and data not shown) indicated that between 1998 and 2013 each emm type was dominant for periods of 3 – 4 years, without a clear increase or decrease in the proportion of a single clone accompanying the decrease in macrolide resistance. Macrolide consumption still stands as the main driver of macrolide resistance in streptococci.10 – 12 However, the association between clonal fluctuations in the population and macrolide resistance phenotypes must also play a key role in the dynamics of the macrolide-resistant GAS population, as already recognized.1 Here, we documented a decline in macrolide resistance among GAS associated with pharyngitis to a level comparable to the lowest in Europe (including countries with low macrolide consumption), in spite of high macrolide use.10,13,18 These changes were accompanied by variations in the clonal composition of the population, which are part of the natural dynamics of macrolideresistant GAS and independent of the behaviour of the general GAS population, reinforcing the importance of considering both the clonal structure of the bacterial population and antibiotic consumption when explaining the prevalence of resistant isolates.
Erythromycin-resistant GAS
Pinheiro, R. Semedo (Hospital Dr Jose´ Maria Grande, Portalegre), Luı´sa Gonc¸alves, Olga Neto (Hospital dos SAMS, Lisboa), Luı´sa Sancho (Hospital Dr Fernando da Fonseca, Amadora/Sintra), Jose´ Diogo, Ana Rodrigues (Hospital Garcia de Orta, Almada), Elmano Ramalheira, Raquel Diaz, So´nia Ferreira (Centro Hospitalar Baixo Vouga), Jose´ Miguel Ribeiro, Isabel Vale, Ana Carvalho (Centro Hospitalar de Tondela—Viseu), Maria Anto´nia Read, Margarida Monteiro, Valquı´ria Alves (Hospital Pedro Hispano, Matosinhos).
Funding This work was partially supported by Fundac¸a˜o para a Cieˆncia e Tecnologia, Portugal (PTDC/SAU-ESA/72321/2006).
Transparency declarations
Supplementary data Figure S1 is available as Supplementary data at JAC Online (http://jac. oxfordjournals.org/).
References 1 Silva-Costa C, Fria˜es A, Ramirez M et al. Macrolide-resistant Streptococcus pyogenes: prevalence and treatment strategies. Expert Rev Anti Infect Ther 2015; 13: 615–28.
6 Enright MC, Spratt BG, Kalia A et al. Multilocus sequence typing of Streptococcus pyogenes and the relationships between emm type and clone. Infect Immun 2001; 69: 2416–27. 7 Francisco A, Bugalho M, Ramirez M et al. Global optimal eBURST analysis of multilocus typing data using a graphic matroid approach. BMC Bioinformatics 2009; 10: 152. 8 Francisco AP, Vaz C, Monteiro PT et al. PHYLOViZ: phylogenetic inference and data visualization for sequence based typing methods. BMC Bioinformatics 2012; 13: 87. 9 Altman DG. Practical Statistics for Medical Research. Boca Raton, FL: Chapman & Hall/CRC, 1999. 10 Van Heirstraeten L, Coenen S, Lammens C et al. Antimicrobial drug use and macrolide-resistant Streptococcus pyogenes, Belgium. Emerg Infect Dis 2012; 18: 1515– 8. 11 Seppa¨la¨ H, Klaukka T, Vuopio-Varkila J et al. The effect of changes in the consumption of macrolide antibiotics on erythromycin resistance in group A streptococci in Finland. Finnish Study Group for Antimicrobial Resistance. N Engl J Med 1997; 337: 441–6. 12 Granizo JJ, Aguilar L, Casal J et al. Streptococcus pyogenes resistance to erythromycin in relation to macrolide consumption in Spain (1986–1997). J Antimicrob Chemother 2000; 46: 959–64. 13 Adriaenssens N, Coenen S, Versporten A et al. European Surveillance of Antimicrobial Consumption (ESAC): outpatient macrolide, lincosamide and streptogramin (MLS) use in Europe (1997– 2009). J Antimicrob Chemother 2011; 66 Suppl 6: vi37–45. 14 Austin DJ, Kristinsson KG, Anderson RM. The relationship between the volume of antimicrobial consumption in human communities and the frequency of resistance. Proc Natl Acad Sci USA 1999; 96: 1152 –6. 15 Malli E, Tatsidou E, Damani A et al. Macrolide-resistant Streptococcus pyogenes in Central Greece: prevalence; mechanism and molecular identification. Int J Antimicrob Agents 2010; 35: 614– 5.
2 Silva-Costa C, Pinto FR, Ramirez M et al. Decrease in macrolide resistance and clonal instability among Streptococcus pyogenes in Portugal. Clin Microbiol Infect 2008; 14: 1152 –9.
16 Ardanuy C, Domenech A, Rolo D et al. Molecular characterization of macrolide- and multidrug-resistant Streptococcus pyogenes isolated from adult patients in Barcelona, Spain (1993 – 2008). J Antimicrob Chemother 2010; 65: 634–43.
3 Silva-Costa C, Fria˜es A, Ramirez M et al. Differences between macrolide resistant and susceptible Streptococcus pyogenes: the importance of clonal properties in addition to antibiotic consumption. Antimicrob Agents Chemother 2012; 56: 5661 –6.
17 Montes M, Tamayo E, Mojica C et al. What causes decreased erythromycin resistance in Streptococcus pyogenes? Dynamics of four clones in a Southern European region from 2005 to 2012. J Antimicrob Chemother 2014; 69: 1474– 82.
4 Melo-Cristino J, Fernandes ML. Streptococcus pyogenes isolated in Portugal: macrolide resistance phenotypes and correlation with T types. Portuguese Surveillance Group for the Study of Respiratory Pathogens. Microb Drug Resist 1999; 5: 219–25.
18 Richter SS, Heilmann KP, Dohrn CL et al. Increasing telithromycin resistance among Streptococcus pyogenes in Europe. J Antimicrob Chemother 2008; 61: 603–11.
5 Silva-Costa C, Ramirez M, Melo-Cristino J. Identification of macrolideresistant clones of Streptococcus pyogenes in Portugal. Clin Microbiol Infect 2006; 12: 513– 8.
19 Silva-Costa C, Ramirez M, Melo-Cristino J et al. Rapid inversion of the prevalences of macrolide resistance phenotypes paralleled by a diversification of T and emm types among Streptococcus pyogenes in Portugal. Antimicrob Agents Chemother 2005; 49: 2109– 11.
5 of 5
Downloaded from http://jac.oxfordjournals.org/ at University of Virginia on July 12, 2015
M. R. has received honoraria for serving on the speaker’s bureau of Pfizer and for consulting from GlaxoSmithKline. J. M.-C. has received research grants administered through his university and received honoraria for serving on the speaker’s bureaus of Pfizer, Novartis and Gilead. C. S.-C.: none to declare. No company or financing body was involved in the decision to publish.
JAC
