Skip Navigation

Effect of Macrolide Consumption on Erythromycin Resistance in Streptococcus pyogenes in Finland in 1997–2001

  1. Miika Bergman1,
  2. Solja Huikko1,3,
  3. Marja Pihlajamäki1,
  4. Pekka Laippala3,4,a,
  5. Erkki Palva5,
  6. Pentti Huovinen1,
  7. Helena Seppälä1,2, and
  8. Finnish Study Group for Antimicrobial Resistance (FiRe Network)b
  1. 1National Public Health Institute, Antimicrobial Research Laboratory, Tampere
  2. 2Department of Ophthalmology, Turku City Hospital, Turku
  3. 3School of Public Health, University of Tampere, Tampere
  4. 4Research Unit, Tampere University Hospital, Tampere
  5. 5National Agency of Medicines, Helsinki, Finland
  1. Reprints or correspondence: Dr. Miika Bergman, National Public Health Institute, Antimicrobial Research Laboratory, Kiinamyllynkatu 13, 20520 Turku, Finland (mihebe{at}utu.fi).
 
Next Section

Abstract

The aim of this study was to investigate the association between regional macrolide resistance in Streptococcus pyogenes and macrolide use in Finland. During 1997–2001, a total of 50,875 S. pyogenes isolates were tested for erythromycin susceptibility in clinical microbiology laboratories throughout Finland. The local erythromycin resistance levels were compared with the regional consumption data of all macrolides pooled and, separately, with the use of azithromycin. The regional resistance rates of 1 year were compared with the regional consumption of the previous year and with the average rates of use for the 2 previous years. A linear mixed model for repeated measures was used in modeling the association. A statistically significant association existed between regional erythromycin resistance in S. pyogenes and consumption of macrolides; association with azithromycin use alone was not found.

The emergence of antimicrobial resistance poses a major challenge for the management of infections worldwide [1]. It is a generally accepted fact that antimicrobial consumption is a major promoter in the emergence and transmission of drug-resistant bacterial strains. There are, however, not many studies that have indisputably shown the connection. A great proportion of these are based on a relatively small number of tested strains or were performed within a relatively short time period. Moreover, most of the studies have been conducted in the hospital context, whereas larger, community-wide studies are relatively rare [2–9]. It is necessary to find out more about the resistance-consumption relationship in community settings [7, 10].

Macrolides are commonly used for treatment of respiratory infections, and they are recommended as second-line drugs for treating Streptococcus pyogenes infections. Macrolide resistance in S. pyogenes is, however, increasing in many countries [11–17], which has been postulated to be caused by the use—or overuse—of macrolides. Furthermore, it has been found that azithromycin, which is a long-acting macrolide administered once per day, would select resistance more effectively than macrolides administered 3 times per day (erythromycin) or twice per day (clarithromycin and roxithromycin) [4, 9, 18]. In Finland, azithromycin use has stabilized over the past few years, and it now constitutes approximately one-third of total macrolide consumption (figure 1) [19].

Figure 1
View larger version:
Figure 1

Macrolide and azithromycin consumption in Finland in 1995–2001. DDD, defined daily doses.

In our previous studies, we showed that regional outpatient erythromycin use in 1991 was linked to increased erythromycin resistance in S. pyogenes in Finland in 1992 [2] and that, when macrolide consumption in outpatient therapy was reduced through national prescribing recommendations, there was a steady decrease in the frequency of erythromycin resistance among S. pyogenes isolates [3]. In this study, we investigated the relationship between regional erythromycin resistance in S. pyogenes in 1997–2001 and local macrolide use in Finland, with special reference to azithromycin use.

Previous SectionNext Section

Materials and Methods

Macrolide resistance in S. pyogenes. The resistance data were collected annually for the years 1997–2001. The macrolide resistance data for S. pyogenes were obtained from the clinical microbiology laboratories belonging to the Finnish Study Group for Antimicrobial Resistance (FiRe), a national microbiology laboratory network. Data were obtained from 18 of the 22 central hospital districts in Finland (table 1). These districts represent ∼95% of the total Finnish population. Not all participating laboratories sent in their resistance data every year (table 1).

A total of 50,875 S. pyogenes strains were tested during the 5-year study period. The number of annually tested strains per central hospital district varied from 46 to 6459, with an average of 678 isolates (table 1). The variation is mainly due to differences in the numbers of inhabitants in different central hospital districts.

Susceptibility testing was performed in participating laboratories by the use of the disk diffusion method paralleling the guidelines of the NCCLS [20]. Some of the laboratories used Iso-Sensitest agar (Unipath) as growth medium instead of Mueller-Hinton agar, which is recommended by the NCCLS as growth medium. All laboratories in the FiRe network participate in the quality-control programs of the National Public Health Services (Colindale, UK) or Labquality.

Macrolide consumption. Data on the regional consumption of macrolides were obtained from the National Agency of Medicines and from the Social Insurance Institution (tables 2 and 3) [19]. These sales statistics are based on sales from wholesalers to pharmacies on an annual basis. Antimicrobial consumption is expressed in defined daily doses per 1000 inhabitants per day.

Statistical analysis. The regional erythromycin resistance levels of S. pyogenes were compared with the pooled regional consumption data for all macrolides and, separately, with the use of azithromycin. The association between resistance and macrolide or azithromycin consumption was studied by a 2-pattern approach: the regional resistance rates of 1 year were compared with the regional consumption of the previous year, and to the average consumption of the 2 previous years. In the year 2000, the central hospital districts of Helsinki and Uusimaa were united. This administrative arrangement put a large proportion of Finland's population in the same central hospital district. As a result, the statistics concerning drug consumption and the resistance data of this new district could be obtained only in a combined form from the years 2000 and on. This reduced the number of the resistance-consumption study units by one in the years 2000 and 2001. In tables 1–3, the combined numbers are placed under the title Uusimaa.

The actual number of tested strains was not included in the statistical analysis; the fraction of resistant strains was used instead. By using this procedure, we wanted to avoid bias, which derives from the fact that the laboratories in the densely inhabited central hospital districts test more strains than do the laboratories in sparsely inhabited areas and, therefore, would get more emphasis. Moreover, data from Lapland for 1999 and 2000 were not included because the number of tested strains was considered to be too small (table 1).

A linear mixed model for repeated measures was used in modeling the association between resistance and the use of antimicrobials. The fraction of resistant strains was taken as the dependent variable; antimicrobial consumption and time were the explanatory variables. A random-effects model with intercept and time as random effects was fitted. The distribution of resistance was skewed, and a square root transformation was thus applied.

Mixed models were fitted by Proc Mixed in the SAS System for Windows, version 8.02 (SAS Institute). The level of statistical significance was set at .05.

Previous SectionNext Section

Results

During the 5-year study period, the rates of erythromycin resistance in S. pyogenes in the whole country of Finland were 9.2% in 1997, 17.1% in 1998, 11.3% in 1999, 8.1% in 2000, and 7.4% in 2001 (figure 2). During 1997–2001, the annual resistance rate in different central hospital districts varied from 1.2% to 39% (table 1). During 1995–2000, the total rates of annual macrolide use and azithromycin use varied from 1.15 to 2.85 and from 0.30 to 1.22 defined daily doses per 1000 inhabitants per day, respectively (tables 2 and 3).

Figure 2
View larger version:
Figure 2

Erythromycin resistance in Streptococcus pyogenes in Finland in 1990–2002. *Data not available.

Table 1
View larger version:
Table 1

Streptococcus pyogenes isolates from different central hospital districts in Finland and resistant isolates.

Table 2
View larger version:
Table 2

Macrolide use in different central hospital districts in Finland, 1995–2000.

Table 3
View larger version:
Table 3

Azithromycin use in different central hospital districts in Finland, 1995–2000.

There was a statistically significant association between regional erythromycin resistance in S. pyogenes and regional use of macrolides (table 4). This was shown both in the comparison of the resistance to the consumption of the previous year (P = .016) and the mean rate of use during the 2 previous years (P = .003). The estimates were positive, indicating that the higher the previous drug consumption had been, the higher the level of resistance. A statistically significant association was not found between erythromycin resistance and the use of azithromycin (table 5) with regard to both use during the previous year (P = .186) and the mean rate of use during the 2 previous years (P = .119).

Table 4
View larger version:
Table 4

Connection between erythromycin resistance and total macrolide use.

Table 5
View larger version:
Table 5

Connection between erythromycin resistance and azithromycin use.

The linear change of resistance over the time period of this study was not significant—that is, when controlling for the drug consumption, time as such did not explain the level of resistance (tables 4 and 5).

Previous SectionNext Section

Discussion

This study is among the first in which the relationship between outpatient antimicrobial use and resistance has been investigated over a period of several years. Although on a national level, macrolide use remained rather steady during the study period, we found a statistically significant association between regional erythromycin resistance in S. pyogenes and the regional use of macrolides. This parallels our previous findings in Finland [2] and the findings of Granizo et al. [4], who found a strong relationship between erythromycin resistance in S. pyogenes and the use of macrolides in Spain. However, in contrast to the results obtained by Granizo et al. [4], regional azithromycin use alone did not have a significant association with the regional resistance levels in our study. In previous studies of community-acquired Streptococcus pneumoniae, there has also been found an association between erythromycin resistance and total macrolide use in Finland [6], Spain [9, 21], and Germany [18], as well as between erythromycin resistance and consumption of long-acting macrolides in Spain [9, 21] and Germany [18].

Cižman [10] suggests that the association between long-acting macrolides and the selection of macrolide resistance may be explained by the concept of selective windows: when an antimicrobial agent has a low maximum concentration and a long half-life, it has a longer selective window and therefore is more likely to select resistance. Azithromycin belongs to these long-acting macrolides and would therefore be a main contributor in the selection of macrolide resistance. Although our study does not support this suggestion, it is possible that azithromycin use levels need to exceed a certain threshold to trigger the emergence of resistance [22, 23]. Another reason for the discrepancy may be that the use of macrolides for S. pyogenes infections has decreased in Finland in the past few years [24], which means that macrolide-resistant isolates are selected only when carriers of S. pyogenes are treated with macrolides for infections caused by other respiratory pathogens. Because of the convenience in dosing, it is possible that azithromycin is more commonly used for very young children than for school-age children. Because school-age children are known to have a higher pharyngeal carriage rate of S. pyogenes than are younger children, it may be that the selection pressure caused by azithromycin is not as strong as that of the other macrolides in the case of S. pyogenes in Finland.

Time itself did not explain the level of resistance in our study; when controlling for the drug consumption, there was no significant linear change in the resistance level. Nevertheless, although the P value for time was not significant, there might be a trend toward decreased resistance over this time period.

In many previous studies, the antimicrobial use-resistance relationship has been studied by a relatively straightforward approach. A common way to analyze the association seems to be the calculation of nonparametric correlations between the 2 variants [4, 6, 18]. In this study, we have used a linear mixed model for repeated measures in the statistical analysis. In this kind of modeling, we are able to consider the nature of repeated measures and allow subject-specific parameters.

However, there are some factors in this study setting that may have biased the results. First, the figures used for the antimicrobial use are figures of sales from different central hospital districts. They do not represent the actual amount of antimicrobials really used in the area. Second, all the antimicrobial drugs sold in a central hospital district are not necessarily used in that area, because patients can freely choose the pharmacy where they buy their medication. Third, people may freely travel, and in favorable circumstances, they may transfer resistant strains even without the effect of selection pressure by antimicrobial drugs.

The time gap between the introduction of a new antimicrobial agent and subsequently emerging resistance is not well known. By use of 2 study approaches—that is, by comparing resistance data to the consumption of the previous year and to the mean consumption of 2 previous years—we wanted to find out which of these 2 methods would better explain the level of resistance. Our findings show that the models of comparing resistance data to the previous year's and the 2 previous years' consumption lead to similar outcome. It would have been interesting to study the relationship also within shorter time periods—for example, by the use of monthly data—because the emergence of resistance does not, at least in some cases, necessarily need a long time span. However, the data available to us did not provide this opportunity, because the consumption rates are currently published only annually. In addition, the consumption figures are based on the sales from wholesalers to pharmacies; therefore, the reliability of the consumption data concerning shorter time periods than a year would be weaker. The question about the most ideal study approach remains to be solved. To better understand the temporal relationships between drug consumption and resistance trends, epidemiological research should be closely connected to biological as well as mathematical methods.

The relationship between antimicrobial resistance and use is very complex. However, by using this research frame and these statistical methods, we can claim with some confidence that the connection really exists. In addition, in this study, we have taken advantage of the availability of Finnish national data on antimicrobial resistance and use; similar data are not available to this extent in all countries of the world. We hope that, with this study, we have taken a step toward a better understanding of the complicated question of how resistance emerges.

Previous SectionNext Section

Finnish Study Group For Antimicrobial Resistance (Fire Network)

Anja Kostiala-Thompson and Merja Rautio (Jorvi Hospital, Espoo); Risto Renkonen and Anna Muotiala (MedixDiacor Laboratory Service, Helsinki); Martti Vaara and Petteri Carlson (Helsinki University Hospital, Helsinki); Hannele Jousimies-Somer (Mehiläinen Hospital, Helsinki); Katariina Pekkanen (Yhtyneet Laboratoriot Oy, Helsinki); Jukka Korpela and Ritva Heikkilä (Central Hospital of Kanta-Häme, Hämeenlinna); Suvi-Sirkku Kaukoranta and Heikki Kaukoranta (Central Hospital of North Karelia, Joensuu); Pirkko Hirvonen and Antti Nissinen (Central Hospital of Central Finland, Jyväskylä); Pekka Ruuska (Central Hospital of Kainuu, Kajaani); Henrik Jägerroos (Central Hospital of Lapland, Rovaniemi); Martti Larikka (Central Hospital of Länsi-Pohja, Kemi); Simo Räisänen (Central Hospital of Central Ostrobothnia, Kokkola); Ulla Larinkari and Benita Forsblom (Central Hospital of Kymenlaakso, Kotka); Marja-Leena Katila and Ulla Kärkkäinen (Kuopio University Hospital, Kuopio); Hannu Sarkkinen and Pauliina Kärpänoja (Central Hospital of Päijät-Häme, Lahti); Maritta Kauppinen and Seppo Paltemaa (Central Hospital of South Karelia, Lappeenranta); Päivi Kärkkäinen (Mikkeli Central Hospital, Mikkeli; Savonlinna Central Hospital, Savonlinna); Sylvi Silvennoinen-Kassinen (Deaconess Institution in Oulu, Oulu); Markku Koskela (Oulu University Central Hospital, Oulu); Marja-Liisa Klossner and Sini Pajarre (Central Hospital of Satakunta, Pori); Sinikka Oinonen and Virpi Ratia (Central Hospital of Seinäjoki, Seinäjoki); Paul Grönroos (Koskiklinikka, Tampere); Risto Vuento and Oili Liimatainen (Tampere University Hospital, Tampere); Maj-Rita Siro (Health Center Pulssi, Turku); Erkki Eerola and Raija Manninen (University of Turku, Turku); Olli Meurman (Turku University Central Hospital, Turku); Marko Luhtala (Central Hospital of Vaasa, Vaasa); and Katrina Lager (Antimicrobial Research Laboratory, National Public Health Institute, Turku).

Previous SectionNext Section

Footnotes

  • Financial support: The Academy of Finland (to M.B., P.H., and M.P.) and a special government grant (EVO grant) from the Turku City Hospital (to H.S.).

  • ↵a Deceased.

  • ↵b Members of the study group are listed at the end of the text.

  • Received July 25, 2003.
  • Accepted December 21, 2003.
Previous Section
 

references

  1. ↵
    1. World Health Organization
    . WHO Global Strategy for Containment of Antimicrobial Resistance. Geneva, Switzerland: WHO; 2001.
  2. ↵
    1. Seppälä H,
    2. Klaukka T,
    3. Lehtonen R,
    4. Nenonen E,
    5. Huovinen P
    . Outpatient use of erythromycin: link to increased erythromycin resistance in group A streptococci. Clin Infect Dis 1995;21:1378-85.
    Abstract/FREE Full Text
  3. ↵
    1. Seppälä H,
    2. Klaukka T,
    3. Vuopio-Varkila J,
    4. et al
    . The effect of changes in the consumption of macrolide antibiotics on erythromycin resistance in group A streptococci in Finland. N Engl J Med 1997;337:441-6.
    CrossRefMedlineWeb of Science
  4. ↵
    1. Granizo JJ,
    2. Aguilar L,
    3. Casal J,
    4. Dal-Re R,
    5. Baquero F
    . Streptococcus pyogenes resistance to erythromycin in relation to macrolide consumption in Spain (1986–1997). J Antimicrob Chemother 2000;46:959-64.
    Abstract/FREE Full Text
    1. Granizo JJ,
    2. Aguilar L,
    3. Casal J,
    4. Garcia-Rey C,
    5. Dal-Re R,
    6. Baquero F
    . Streptococcus pneumoniae resistance to erythromycin and penicillin in relation to macrolide and beta-lactam consumption in Spain (1979–1997). J Antimicrob Chemother 2000;46:767-73.
    Abstract/FREE Full Text
  6. ↵
    1. Pihlajamaki M,
    2. Kotilainen P,
    3. Kaurila T,
    4. Klaukka T,
    5. Palva E,
    6. Huovinen P
    . Macrolide-resistant Streptococcus pneumoniae and use of antimicrobial agents. Clin Infect Dis 2001;33:483-8.
    Abstract/FREE Full Text
  7. ↵
    1. Garcia-Rey C,
    2. Aguilar L,
    3. Baquero F,
    4. Casal J,
    5. Martin JE
    . Pharmacoepidemiological analysis of provincial differences between consumption of macrolides and rates of erythromycin resistance among Streptococcus pyogenes isolates in Spain. J Clin Microbiol 2002;40:2959-63.
    Abstract/FREE Full Text
    1. Bronzwaer SL,
    2. Cars O,
    3. Buchholz U,
    4. et al
    . A European study on the relationship between antimicrobial use and antimicrobial resistance. Emerg Infect Dis 2002;8:278-82.
    MedlineWeb of Science
  9. ↵
    1. Garcia-Rey C,
    2. Aguilar L,
    3. Baquero F,
    4. Casal J,
    5. Dal-Re R
    . Importance of local variations in antibiotic consumption and geographical differences of erythromycin and penicillin resistance in Streptococcus pneumoniae. J Clin Microbiol 2002;40:159-64.
    Abstract/FREE Full Text
  10. ↵
    1. Cižman M
    . The use and resistance to antibiotics in the community. Int J Antimicrob Agents 2003;21:297-307.
    CrossRefMedlineWeb of Science
  11. ↵
    1. Martin JM,
    2. Green M,
    3. Barbadora KA,
    4. Wald ER
    . Erythromycin-resistant group A streptococci in schoolchildren in Pittsburgh. N Engl J Med 2002;346:1200-6.
    CrossRefMedlineWeb of Science
    1. Alos JI,
    2. Aracil B,
    3. Oteo J,
    4. Gomez-Garces JL
    . Significant increase in the prevalence of erythromycin-resistant, clindamycin- and miocamycin-susceptible (M phenotype) Streptococcus pyogenes in Spain. J Antimicrob Chemother 2003;51:333-7.
    Abstract/FREE Full Text
    1. Dicuonzo G,
    2. Fiscarelli E,
    3. Gherardi G,
    4. et al
    . Erythromycin-resistant pharyngeal isolates of Streptococcus pyogenes recovered in Italy. Antimicrob Agents Chemother 2002;46:3987-90.
    Abstract/FREE Full Text
    1. Hoban D,
    2. Waites K,
    3. Felmingham D
    . Antimicrobial susceptibility of community-acquired respiratory tract pathogens in North America in 1999–2000: findings of the PROTEKT surveillance study. Diagn Microbiol Infect Dis 2003;45:251-9.
    CrossRefMedlineWeb of Science
    1. Petinaki E,
    2. Kontos F,
    3. Pratti A,
    4. Skulakis C,
    5. Maniatis AN
    . Clinical isolates of macrolide-resistant Streptococcus pyogenes in central Greece. Int J Antimicrob Agents 2003;21:67-70.
    CrossRefMedlineWeb of Science
    1. Reinert RR,
    2. Lutticken R,
    3. Bryskier A,
    4. Al-Lahham A
    . Macrolide-resistant Streptococcus pneumoniae and Streptococcus pyogenes in the pediatric population in Germany during 2000–2001. Antimicrob Agents Chemother 2003;47:489-93.
    Abstract/FREE Full Text
  17. ↵
    1. Sauermann R,
    2. Gattringer R,
    3. Graninger W,
    4. Buxbaum A,
    5. Georgopoulos A
    . Phenotypes of macrolide resistance of group A streptococci isolated from outpatients in Bavaria and susceptibility to 16 antibiotics. J Antimicrob Chemother 2003;51:53-7.
    Abstract/FREE Full Text
  18. ↵
    1. Reinert RR,
    2. Al-Lahham A,
    3. Lemperle M,
    4. et al
    . Emergence of macrolide and penicillin resistance among invasive pneumococcal isolates in Germany. J Antimicrob Chemother 2002;49:61-8.
    Abstract/FREE Full Text
  19. ↵
    1. National Agency for Medicines and the Social Insurance Institution
    . Finnish statistics on medicines 1995–2001. Helsinki: National Agency for Medicines and the Social Insurance Institution; 2002.
  20. ↵
    1. NCCLS
    . Performance standards for antimicrobial disc susceptibility tests. 6th ed. Wayne, PA: NCCLS; 1997.
  21. ↵
    1. Baquero F
    . Evolving resistance patterns of Streptococcus pneumoniae a link with long-acting macrolide consumption? J Chemother 1999;11((Suppl 1)):35-43.
    MedlineWeb of Science
  22. ↵
    1. Baquero F
    . Trends in antibiotic resistance of respiratory pathogens: an analysis and commentary on a collaborative surveillance study. J Antimicrob Chemother 1996;38((Suppl A)):117-32.
    Abstract/FREE Full Text
  23. ↵
    1. Austin DJ,
    2. Kristinsson KG,
    3. 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.
    Abstract/FREE Full Text
  24. ↵
    1. Rautakorpi UM,
    2. Klaukka T,
    3. Honkanen P,
    4. et al
    . Antibiotic use by indication: a basis for active antibiotic policy in the community. Scand J Infect Dis 2001;33:920-6.
    CrossRefMedlineWeb of Science
| Table of Contents

This Article

  1. Clin Infect Dis. (2004) 38 (9): 1251-1256. doi: 10.1086/383309
  1. AbstractFree
  2. » Full Text (HTML)Free
  3. Full Text (PDF)Free

Classifications

Share

  1. Email this article

Navigate This Article

Current Issue

  1. March 1, 2012 54 (5)
  1. Clinical Infectious Diseases
  1. Alert me to new issues

Most