Portal de Periódicos da CAPES

Antibiotic use and pneumococcal resistance to penicillin: the French experience

1999; Elsevier BV; Volume: 5; Linguagem: Inglês

10.1111/j.1469-0691.1999.tb00855.x

1469-0691

Didier Guillemot, Claude Carbón,

Primary Care and Health Outcomes

Antibiotic selection pressure is defined with reference to the environmental conditions promoting the emergence and spread of antimicrobial resistance, whatever the mechanism of resistance. Antibiotic exposure is considered as essential for the emergence of resistance, and there is universal agreement that increasing antimicrobial resistance is related to the selective pressure exerted by the use of antibiotics [1Tenover FC Hughes JM The challenges of emerging infectious diseases.JAMA. 1996; 275: 300-304Crossref PubMed Google Scholar]. Studies on antibiotic selective pressure are generally focused on only one aspect of the problem (genes, bacteria, individuals, population) (Figure 1). These aspects correspond to a specific scientific approach (genetic, cellular, clinical, epidemiologic). Now it is time to consider antimicrobial resistance as an environmental problem and to develop a pharmacoepidemiologic approach, taking into account the individual's antimicrobial exposure and cross-transmission as well as genetic, cellular, individual and population effects. This is required to improve knowledge on population dynamics of resistance, develop epidemiologic models which would predict the evolution of antimicrobial drug resistance and to find the best strategies for the optimal use of antibiotics with regard to the control of antimicrobial resistance. The lack of such studies has limited our knowledge and understanding of bacterial resistance dynamics in populations. As a result, policies regarding antibiotic treatment of outpatients have been hindered. In the community, the spread of drug-resistant Streptococcus pneumoniae is now particularly worrying, especially because of the increase of high-level third-generation cephalosporin resistance and increasing multiresistance. In France, the resistance rates for 1995 and 1996 were respectively: for penicillin, 65.4% and 70.3% (18.6% and 24.9% of intermediately resistant strains, 46.8% and 45.4% of fully resistant strains); for erythromycin, 57.5% and 68.5%; for tetracycline, 43.2% and 42.6%; and for trimethoprim–sulfamethoxazole, 47.5% and 50.9% [2Geslin P Fremaux A Sissia G Spicq C Georges S Development of resistance to beta‐lactams and other antibiotics of pneumococci isolated from acute otitis media in France: statement of the National Reference Center 1995–1996.Arch Pediatr. 1998; 5: 982-987Crossref PubMed Scopus (11) Google Scholar]. Many authors have suggested that the dominant factor in the spread of penicillin-resistant Streptococcus pneumoniae (PRSp) in the community is the increasing use of antibiotics [1Tenover FC Hughes JM The challenges of emerging infectious diseases.JAMA. 1996; 275: 300-304Crossref PubMed Google Scholar, 3Neu HC The crisis in antibiotic resistance.Science. 1992; 257: 1064-1073Crossref PubMed Scopus (2345) Google Scholar]. Ecological correlations have been found between invasive PRSp and antibiotic use in the community [4Baquero F Martinez‐Beltran J Loza E A review of antibiotic resistance patterns of Streptococcus pneumoniae in Europe.J Antimicrob Chemother. 1991; 28: 31-38Crossref PubMed Google Scholar], and time concordance can be observed. These analyses do not allow an estimate of the risk, and the chain of causality remains unclear. Futhermore, while carriage of pneumococci is the precondition for interindividual transmission, most epidemiologic studies estimating the risk of PRSp with β-lactam use have been focused on invasive strains. Few pharmacoepidemiologic studies have analyzed the risk factors associated with drug-resistant pneumococcal pharyngeal carriage. Among these risk factors, recent β-lactam use is the most consistent [5Kristinsson KG Hjalmarsdottir MA Steingrimsson O Increasing penicillin resistance in pneumococci in Iceland.Lancet. 1992; 339: 1606-1607Abstract PubMed Scopus (79) Google Scholar, 6Arason V Kristinsson K Sigurdsson J Stefansdottir G Molstad S Gudmundson S Do antimicrobials increase the carriage rate of penicillin resistant pneumococci in children? Cross sectional prevalence study.Br Med J. 1996; 313: 387-391Crossref PubMed Scopus (448) Google Scholar]. Nevertheless, these studies do not explain how antimicrobial drug use is a risk factor for human colonization by PRSp and for its diffusion throughout the population. From an epidemiologic point of view, the criteria for the causal nature of an association are the agreement with existing information (biological plausibility), the consistency of the association, the time sequence, the specificity of the association, and the strength of the association (quantitative strength, dose–response relationship and study design) [7Ström B Study designs for pharmacoepidemiology studies.in: Strom B Pharmacoepidemiology. 2nd edn. John Wiley & Sons Ltd, Chichester1994: 15-29Google Scholar]. In the case of studies of PRSp and β-lactam use, most of these criteria are satisfied. To test the hypothesis that PRSp carriage is related to daily dose and duration of β-lactam use, we conducted a study of pharyngeal carriage of S. pneumoniae and antimicrobial use in children. This 1995 study was published at the beginning of 1998 [8Guillemot D Carbon C Balkau B et al.Low dosage and long treatment duration of β‐lactam: risk‐factors for the carriage of penicillin‐resistant Streptococcus pneumoniae.JAMA. 1998; 279: 365-370Crossref PubMed Scopus (531) Google Scholar]. It explored PRSp pharyngeal carriage and oral β-lactam use during the preceding month, the dose and the duration of treatment in a random sample of 941 children, 3–6 years old. The daily dose in mg/kg was calculated for the last used antimicrobial drug generic and coded as a high or low daily dose, with reference to French clinical recommendations As recent studies suggested efficacy in community respiratory infectious diseases with an antibiotic given for 5 or fewer days, we considered more than 5 days to be a long duration of treatment. We calculated the percentile of the daily dose for each antibiotic generic and defined two classes with reference to the median (greater or lower than the median). Considering children with pneumococcal carriage who used a β-lactam during the previous month, the median daily dose in mg/kg was 46.2 for amoxycillin, 31.8 for cefaclor, 43.5 for cefadroxil, 41.6 for cefatrizin, 23.4 for cefuroxime axetil, 8.4 for cefpodoxime proxetil and 8.9 for cefixime. Seventy-seven per cent of the S. pneumoniae isolates from children who had not taken a β-lactam had a penicillin G MIC lower than 0.1 mg/L. All of the S. pneumoniae isolates from children who had taken a daily dose of β-lactam lower than the median had a penicillin G MIC higher than 0.1 mg/L. In contrast, all S. pneumoniae isolates from children who had a daily dose of β-lactam higher than the median had a penicillin G MIC lower than 0.1 mg/L (p=0.003) (Figure 2). These two groups did not have different mean durations of treatment, or different mean numbers of daily doses. Grouped according to the combined value for daily dose and duration of aminopenicillin and cephalosporin of the last β-lactam, the percentage of low dose with a long duration was 46.6%, and that of high dose or short duration 39.1%. As compared with no use, low dose with a long duration of treatment was associated with an increased risk of PRSp carriage, for both aminopenicillin and cephalosporin (Figure 3).Figure 3Odds ratio for PRSp carriage according to daily dose and duration of the last β-lactam used during the previous 30 days.View Large Image Figure ViewerDownload (PPT) Even with all epidemiologic criteria, the chain of causality remains unclear. The two determinants for the spread of PRSp in the community are antibiotic selective pressure in individuals and interindividual transmission [9Soares S Kristinsson KG Musser JM Tomasz A Evidence for the introduction of a multiresistant clone of serotype 6B Streptococcus pneumoniae from Spain to Iceland in the late 1980s.J Infect Dis. 1993; 168: 158-163Crossref PubMed Scopus (295) Google Scholar]. Several factors could help to explain the antibiotic selective pressure: bacterial antagonism in regulating the bacterial flora of the human pharynx [10Johanson WG Blackstock R Pierce AK Sanford JP The role of bacterial antagonism in pneumococcal colonization of the human pharynx.J Lab Clin Med. 1970; 75: 946-952PubMed Google Scholar, 11Sanders W Bacterial interference. I. Occurrence among the respiratory tract flora and characterization of inhibition of group A streptococci by viridans streptococci.J Infect Dis. 1969; 120: 698-707Crossref PubMed Scopus (119) Google Scholar, 12Crowe CC Sanders W Longley S Bacterial interference. II. Role of the normal throat flora in prevention of colonization by group A streptococcus.J Infect Dis. 1973; 128: 527-532Crossref PubMed Scopus (136) Google Scholar], horizontal gene transfer [13Coffey TJ Dowson CG Daniels M et al.Horizontal transfer of multiple penicillin‐binding protein genes, and capsular bio‐synthetic genes, in natural populations ot Streptococcus pneumoniae.Mol Microbiol. 1991; 5: 2255-2260Crossref PubMed Scopus (273) Google Scholar], or the selection of mutants of S. pneumoniae, due to a low tissue concentration of antibiotic [14Negri MC Morosini MI Loza E Baquero F In vitro selective antibiotic concentrations of beta‐lactams for penicillin‐resistant Streptococcus pneumoniae populations.Antimicrob Agents Chemother. 1994; 38: 122-125Crossref PubMed Scopus (77) Google Scholar]. Another possibility could be the following: when a population of children is exposed to an oral β-lactam, the increased risk of PRSp carriage might be the result of a difference in the average duration of carriage between PRSp and PSSp. This difference induces an increased transmission of PRSp throughout the population, as suggested in recent studies, where PRSp carriage persisted longer than that of fully susceptible clones [15Yagupsky P Porat N Fraser D et al.Acquisition, carriage, and transmission of pneumococci with decreased antibiotic susceptibility in young children attending a day care facility in southern Israel.J Infect Dis. 1998; 177: 1003-1012Crossref PubMed Scopus (149) Google Scholar]. Can optimal use of antibiotics decrease PRSp evolution? In order to answer this question, we have first to look at antibiotic prescription in the community. An audit of anti-infective prescribing in office-based medical practice in the Loiret, a French administrative 'department', showed that respiratory tract infections with a presumed viral etiology accounted for 36% of prescriptions. Furthermore, a high percentage of antibiotic prescriptions in children were underdosed as compared to clinical recommendations and, whatever the clinical hypothesis, the duration of treatment was close to 8 days [16Guillemot D Carbon C Balkau B et al.Inappropriateness and variability of antibiotic prescription, among French office based physicians.J Clin Epidemiol. 1998; 51: 61-68Abstract Full Text Full Text PDF PubMed Scopus (74) Google Scholar]. Furthermore, based on national representative samples, we showed that, in France, respiratory tract infections with a presumed viral etiology were the most frequent infections associated with antibiotic use in 1980–81. The increase in these antibiotic-treated infections between 1980–81 and 1991–92 was 86.2% for adults and 115.3% for children [17Guillemot D Maison P Carbon C et al.Trends in antimicrobial drug use in the community between 1981 and 1992, in France.J Infect Dis. 1998; 177: 492-497Crossref PubMed Scopus (101) Google Scholar]. In order to control drug-resistant S. pneumoniae, community-wide education programs for clinicians and the general public are important. Such programs should focus on the appropriate use of antibiotics. Our investigations identified three main areas for improving antimicrobial drug prescribing in order to control PRSp: (1) to reduce the number of useless prescriptions in respiratory tract infections with a presumed viral etiology, especially rhinopharyngitis, which is the main community infection associated with antibiotic prescription; (2) to increase the prescribed daily dose of antimicrobials to the recommended levels; and (3) to reduce the duration of treatment. A retrospective analysis based on the French IMS—Dorema data showed that, in France, antibiotic prescription in rhinopharyngitis is not a recent practice and did not increase very much between 1984 and 1995 (Figure 4) [18Obervatoire National des Prescriptions et Consommations des Médicaments.in: Etude de la prescription et de la consommation des antibiotiques en ambulatoire (en France). Obervatoire National des Prescriptions et Consommations des Médicaments, Saint Denis1998Google Scholar], suggesting that antibiotic prescription in this indication in general practice has not changed very much. It may be more difficult to change old habits than to change new fashions. The obvious need is for long-term educational efforts at different levels, to make both general practitioners and patients accept a reduction in the number of prescriptions. Another, parallel, option is to modify some more practical aspects of prescription. For this, we need first to assess the effect of changing antibiotic use in terms of dosage and duration of treatment. In order to distinguish a possible natural decrease in resistance from the impact of such an intervention, the methodology must be an epidemiologic trial. This is the main objective of the AUBEPPIN project. We plan to select at random two geographic areas in the Loiret. The first will be the site of an intervention aiming to increase the dosage above that of the French clinical recommendations and to reduce the duration of treatment to 5 days, by education programs for clinicians and pharmacists. The second will be the control group without any intervention. The intervention will be conducted between October 1999 and April 2000. Before and after the intervention, nasopharyngeal carriage of S. pneumoniae will be surveyed in the two groups, each containing 1500 healthy 3–6-year-old schoolchildren, selected by sampling schools at random in each group. The parents of included children will complete, each month, a short questionnaire on physician consultations and diagnosis as stated by the practitioner, and drug consumption—including the type of drug, the duration of treatment and the number of daily doses—over 6 months. The main criterion will be the PRSp to PSSp ratio, with an expected reduction of 30%. Futhermore, in order to compare the duration of carriage, nasopharyngeal cultures will be obtained every 2 weeks for 6 months from those children who were the pneumococcal carriers at the beginning of the study. In order to take into account the inter-individual transmission inside the school, there will be a genetic analysis of the strains. At least two types of important information can be expected from this project: •the reversibility of the development of penicillin G resistance among S. pneumoniae isolates when the conditions of use of antibiotics are modified.•the effect of a simple message—'to use optimal dosage and short duration of treatment'—that could be added to other actions aimed at developing optimal use of antibiotics, such as reducing the number of undue prescriptions. In any case, these types of intervention would lead to decreased costs and, hopefully, have a positive impact on bacterial resistance.