Relationship Between Rates of Antimicrobial Consumption and the Incidence of Antimicrobial Resistance in Staphylococcus aureus and Pseudomonas aeruginosa

Isolates From 47 French Hospitals Author(s): A.M. Rogues , MD, PhD, C. Dumartin , DPharm, B. Amado, A.G. Venier , MD, N. Marty , MD, PhD, P. Parneix , MD, J.P. Gachie , MD, MPH Source: Infection Control and Hospital Epidemiology, Vol. 28, No. 12 (December 2007), pp. 1389-1395 Published by: The University of Chicago Press on behalf of The Society for Healthcare Epidemiology of America Stable URL: http://www.jstor.org/stable/10.1086/523280 . Accessed: 25/10/2011 14:41
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

The University of Chicago Press and The Society for Healthcare Epidemiology of America are collaborating with JSTOR to digitize, preserve and extend access to Infection Control and Hospital Epidemiology.

http://www.jstor.org

infection control and hospital epidemiology

december 2007, vol. 28, no. 12

original article

Relationship Between Rates of Antimicrobial Consumption and the Incidence of Antimicrobial Resistance in Staphylococcus aureus and Pseudomonas aeruginosa Isolates From 47 French Hospitals
A. M. Rogues, MD, PhD; C. Dumartin, DPharm; B. Amadeo; A. G. Venier, MD; N. Marty, MD, PhD; P. Parneix, MD; J. P. Gachie, MD, MPH

objective. To investigate relationships between rates of antimicrobial consumption and the incidence of antimicrobial resistance in Staphylococcus aureus and Pseudomonas aeruginosa isolates from hospitals. methods. We conducted an observational study that used retrospective data from 2002 and linear regression to model relationships. Hospitals were asked to collect data on consecutive S. aureus and P. aeruginosa isolates, consumption rates for antibiotics (ie, anti-infectives for systemic use as dened by Anatomical Therapeutic Chemical class J01), and hospital characteristics, including infection control policies. Rates of methicillin resistance in S. aureus and rates of ceftazidime and ciprooxacin resistance in P. aeruginosa were expressed as the percentage of isolates that were nonsusceptible (ie, either resistant or intermediately susceptible) and as the incidence of nonsuceptible isolates (ie, the number of nonsuceptible isolates recovered per 1,000 patient-days). The rate of antimicrobial consumption was expressed as the number of dened daily doses per 1,000 patient-days. setting. tested. Data were obtained from 47 French hospitals, and a total of 12,188 S. aureus isolates and 6,370 P. aeruginosa isolates were

results. In the multivariate analysis, fewer antimicrobials showed a signicant association between the consumption rate and the percentage of isolates that were resistant than an association between the consumption rate and the incidence of resistance. The overall rate of antibiotic consumption, not including the antibiotics used to treat methicillin-resistant S. aureus infection, explained 13% of the variance between hospitals in the incidence of methicillin resistance among S. aureus isolates. The incidence of methicillin resistance in S. aureus isolates increased with the use of ciprooxacin and levooxacin and with the percentage of the hospitals beds located in intensive care units (adjusted multivariate coefcient of determination [aR2], 0.30). For P. aeruginosa, the incidence of ceftazidime resistance was greater in hospitals with higher consumption rates for ceftazidime, levooxacin, and gentamicin (aR2, 0.37). The incidence of ciprooxacin resistance increased with the use of uoroquinolones and with the percentage of a hospitals beds located in intensive care ( aR2, 0.28). conclusions. A statistically signicant relationship existed between the rate of uoroquinolone use and the rate of antimicrobial resistance among S. aureus and P. aeruginosa isolates. The incidence of resistant isolates showed a stronger association with the rate of antimicrobial use than did the percentage of isolates with resistance. Infect Control Hosp Epidemiol 2007; 28:1389-1395

The increasing prevalence of antimicrobial-resistant organisms is a major public health problem and is of particular concern for hospitals.1 Two problematic nosocomial pathogens are Staphylococcus aureus and Pseudomonas aeruginosa; both express multidrug resistance. Different pathways and associated factors are involved in the emergence and spread of antimicrobial resistance in these 2 bacteria. Rates of antimicrobial resistance and of isolation of resistant bacteria in hospitals are affected by various factors, including infection

control practices, usage patterns for antibiotics, local epidemiology, and the medical and surgical procedures performed at the hospital. Because colonization or infection with an antimicrobial-resistant microorganism in a hospital is therefore the result of several factors, infection control requires a multifaceted approach.2 The relative roles played by cross-transmission and antimicrobial selection pressure in the levels of bacterial resistance observed in hospital settings are still to be determined.

From the Unite INSERM 657Pharmacoepidemiologie et evaluation de limpact des produits de sante sur les populations, IFR Sante Publique Universite Victor Segalen (A.-M.R., J.P.G.), the Centre de Coordination de la Lutte contre les Infections Nosocomiales du Sud-Ouest, CHU Groupe hospitalier Pellegrin (C.D., B.A., A.G.V., P.P.), the Service dHygiene Hospitaliere, CHU Groupe hospitalier Pellegrin, Bordeaux (A.M.R., J.P.G.), the Service de Microbiologie, ` ` Ho pital Purpan, Toulouse (N.M.), France. Received April 17, 2007; accepted August 3, 2007; electronically published November 1, 2007. 2007 by The Society for Healthcare Epidemiology of America. All rights reserved. 0899-823X/2007/2812-0013$15.00. DOI: 10.1086/523280

1390

infection control and hospital epidemiology

december 2007, vol. 28, no. 12

The use of antibiotic therapy to treat individuals has an ecological impact on all the patients hospitalized in the same facility.3,4 In most hospitals, only aggregated microbiologic and pharmacy data are available to researchers studying this relationship. If such aggregated data are available for many hospitals for a dened period (eg, a single year), the correlation of rates of consumption and incidences of resistance could prove helpful for comparative purposes. Only a few studies have taken such a collective approach.5 The purpose of this study was to analyze and correlate multicenter surveillance data on the incidence of antimicrobial resistance among S. aureus and P. aeruginosa isolates and the rate of antibiotic consumption in hospitals. A better knowledge of these relationships could help prioritize interventions for controlling antimicrobial resistance in the hospital setting.

the hospitals consumption rates for disposable gowns, as well as scrub and alcohol-based hand rub for hand disinfection. Measurement of Antibiotic Consumption Hospital pharmacies were requested to report the annual rate of consumption of antibiotics for systemic use (ie, antibiotics delivered to the various wards of the hospital for this purpose), as dened by group J01 of the Anatomical Therapeutic Chemical classication system. The use of antibiotics was expressed in dened daily doses (DDDs) per 1,000 patientdays for the whole hospital. The Anatomical Therapeutic ChemicalDDD classication from the World Health Organization, 2005 version, was used.6 The antibiotics that were considered to be used to treat infection with methicillinresistant S. aureus (MRSA) were glycopeptides, fosfomycin, rifampicin, fusidic acid, trimethoprim-sulfamethoxazole, and synergistin. Rates of Bacterial Resistance Each hospital was asked to collect data on consecutive S. aureus and P. aeruginosa isolates during the entire year of 2002. Microbiological data was pooled for each hospital. Data included information on the number of isolates that were resistant and the number of isolates from clinical specimens tested for susceptibility. The total number of resistant isolates was obtained by summing the numbers of intermediately susceptible and resistant isolates. We focused on MRSA and on P. aeruginosa with resistance to ceftazidime or ciprooxacin. The rate of antimicrobial resistance was expressed in 2 ways: the proportion (ie, percentage) of isolates that were nonsuceptible (ie, resistant or intermediately susceptible) and the number of nonsuceptible isolates per 1,000 patient-days (ie, incidence). Duplicate isolates were excluded; these were dened as isolates of the same organism with the same antimicrobial resistance pattern that were recovered from the same patient during the study period, regardless of the site from which they were isolated (eg, blood, sputum, urine, or wound). Statistical Analysis The outcome variables were the percentage of isolates with resistance and the incidence of resistance for S. aureus and P. aeruginosa. The associations between different hospital variables and resistance rates were tested in univariate analysis with the Spearman correlation test (r) and Kruskal-Wallis test. A multiple linear regression analysis was performed with a backward stepwise approach to identify factors associated with rates of resistance, after checking for hospital characteristics and infection control policies. Because the distribution was nonnormal, the incidence of antimicrobial resistance in P. aeruginosa isolates was logarithmically or squareroot transformed. Structure parameters, such as hospital type, number of beds in the hospital, and hospital areas, were included in the multivariate analysis as potential confounders.

me th ods
Data Collection An observational study design was adopted, which used a questionnaire mailed to 96 public hospitals and 123 private hospitals belonging to the South-West Regional Coordinating Center for Nosocomial Infection Control. The hospitals participated in this study voluntarily and were required to have participated in the regional laboratory network for bacterial resistance. Thus, the laboratories of the hospitals that participated performed microbiological testing and interpreted results in accordance with the same guidelines, those of the French Society of Microbiology. Hospitals were asked to report aggregated annual data for 2002. Data were collected retrospectively from administrative, pharmacy, and laboratory computerized databases. If a hospital tested fewer than 10 S. aureus or P. aeruginosa isolates for susceptibility during the year, then the hospital was excluded from the study. Hospital Characteristics For each hospital, administrative data were recorded: the type and number of beds, number of patient-days (ie, periods of more than 24 hours), and the number of admissions. This data was collected for the whole hospital and for each hospital ward (ie, areas where a patient stayed at least 1 night in the hospital). The different hospital inpatient wards were classed in groups according to the French administrative denition: medical, surgery, intensive care, obstetric, rehabilitation, long-term care, and psychiatry. The mean length of stay was obtained by dividing the number of patient-days by the number of admissions. The day of admission and the day of discharge were counted together as 1 patient-day. In addition, the questionnaire collected information about infection control measures implemented in the hospital. These included the hospitals organization, use of infection control teams, guidelines for barrier precautions, and policies for controlling the spread of bacterial resistance, as well as

rates of antibiotic use and incidence of resistance in french hospitals

1391

table 1.

Consumption Rates for the Main Classes of Antibacterials for Systemic Use in French Hospitals Dened daily doses per 1,000 patient days, median (range)

Class Penicillins Any Combination of penicillins, including b-lactamase inhibitors Cephalosporins Any First- and second-generation cephalosporins Ceftazidime Fluoroquinolones Any Ciprooxacin Macrolides Aminoglycosides Imidazoles Combination of sulfonamides and trimethoprim Glycopeptides
a

All hospitals (N p 47) 277.4 (12.3-513.5) 198.9 (8.1-424.6) 37.4 (7.8-157.6) 5.5 (0-104) 1.8 (0-19) 59.8 16.4 26.0 11.9 10.7 6.1 4.1 (8.5-455) (0-163) (0-87.5) (1.6-77.2) (0-95.9) (0-22.6) (0-27.2)

Public hospitals (n p 31) 246.5 (12.2-509.6) 177.1 (8.1-420.5) 29.2 (7.9-88.5) 4.9 (0-17.6) 2.1 (0-16) 59.3 15.4 27.4 10.9 9.4 6.4 4.1 (8.5-129.9) (0-54) (1.8-76.8) (1.6-67.9) (0-26.7) (0-22.6) (0-27.2)

Private hospitals (n p 16) 311.6 (152-513) 240.8 (117.5-424.6) 47.7 (15.9-157.6) 27.8 (0.3-104) 1.1 (0-19) 76.8 16.6 23.8 16.0 16.2 5.4 4.3 (35.3-455) (0-163) (0-87.5) (3.3-77.1) (3-95.7) (0-20.3) (0.4-18.4)

Pa .05 .005 .001 .001 .37 .10 .63 .71 .11 .01 .55 .50

Wilcoxon test.

We retained those variables that seemed to be statistically associated with rates of resistance on the basis of a threshold P value of .20. The backward stepwise selection was used with a P value of .05 as the signicance level for removing variables from the model. The nal model included all covariates that were statistically signicant. Data analysis was performed using Stata software, version 7.0 (Stata).

median was 468 DDDs per 1,000 patient-days. The consumption rates for the main antibiotic classes are presented in Table 1. Bacterial Resistance A total of 12,188 S. aureus isolates and 6,370 P. aeruginosa isolates were tested in the participating hospitals. The median percentage of S. aureus isolates that were MRSA at each hospital was 41%, and the incidence of resistance was 0.87 isolates per 1,000 patient-days. The median percentage of P. aeruginosa isolates that were resistant to ceftazidime at each hospital was 17%, and the incidence of ceftazidime resistance was 0.2 isolates per 1,000 patient-days; the mean percentage of P. aeruginosa isolates resistant to ciprooxacin at each hospital was 34%, and the incidence of ciprooxacin resistance was 0.43 isolates per 1,000 patient-days. The incidence of resistance was higher in public hospitals than in private hospitals (Table 2). The percentage of S. aureus isolates that were MRSA was signicantly positively correlated with the percentage of ciprooxacin resistance in P. aeruginosa isolates (r, 0.34; P p .01). The incidence of methicillin resistance in S. aureus was signicantly positively correlated with the incidences of ciprooxacin resistance (r, 0.70; P p .001) and ceftazidime resistance (r, 0.51; P p .001) in P. aeruginosa isolates. There were statistically signicant relationships between rates of antimicrobial resistance in S. aureus or P. aeruginosa isolates and certain hospital characteristics. The percentage of S. aureus isolates that were MRSA increased with the length of hospital stay (r, 0.38; P p .008) and with the percentage of the hospitals beds located in the rehabilitation service (r, 0.48; P p .001). The incidence of MRSA increased with the

res ults
Participating Hospitals Data were obtained from 47 hospitals that volunteered to participate. These hospitals had a combined total of 15,953 beds and a total of 4,719,281 patient-days; 18 of the hospitals had more than 300 beds. These hospitals had a median capacity of 150 beds (range, 43-1,418 beds). The total number of beds included 4,575 medical service beds in 44 hospitals, 3,708 surgical beds in 41 hospitals, 472 intensive care beds in 35 hospitals, 970 obstetrics beds in 29 hospitals, 862 psychiatric beds in 14 hospitals, 1,765 rehabilitation beds in 29 hospitals, and 2,979 long-term care beds in 26 hospitals. The hospitals were classied according to the French administrative denition; 31 hospitals (including 3 teaching hospitals) were classied as public hospitals, and 16 were classied as private hospitals. All of these hospitals had an infection control committee, but 6 of them had no infection control team, and only 18 (38%) of them had an infection control practitioner. Policies for controlling the spread of bacterial resistance in the hospital by use of contact precautions had been implemented for more than 2 years by a majority of the hospitals (39 [84%]). The overall antibiotic consumption rates for the hospitals varied between 113 and 1,004 DDDs per 1,000 patient-days; the

1392

infection control and hospital epidemiology

december 2007, vol. 28, no. 12

table 2. Hospitals

Antimicrobial Resistance in Staphylococcus aureus and Pseudomonas aeruginosa Isolates in French Median value (range)

Bacteria, variable S. aureus No. of isolates tested for susceptibility Percentage of isolates resistant to methicillin No. of methicillin-resistant isolates recovered per 1,000 patient-days P. aeruginosa No. of isolates tested for susceptibility Percentage of isolates resistant to ceftazidime No. of ceftazidime-resistant isolates recovered per 1,000 patient-days Percentage of isolates resistant to ciprooxacin No. of ciprooxacin-resistant isolates recovered per 1,000 patient-days

All hospitals (N p 47) 151 (14-2043) 41 (5-86) 0.87 (0.11-2.04) 80 (11-1249) 17 (5-57) 0.18 (0.02-1.71) 34 (5-62) 0.43 (0.02-1.72)

Public hospitals (n p 31) 226 (63-2043) 44 (5-66) 1.02 (0.16-2.04) 106 (13-249) 15 (5-50) 0.18 (0.02-0.87) 37 (14-62) 0.44 (0.12-1.72)

Private hospitals (n p 16) 62 (14-342) 35 (13-86) 0.58 (0.11-1.62) 40 (11-187) 19 (5-57) 0.21 (0.03-1.71) 31 (5-57) 0.41 (0.03-1.71)

P .08 .01 .30 .77 .11 .38

percentage of a hospitals beds located in the rehabilitation service (r, 0.28; P p .05) and with the percentage of beds located in intensive care units (r, 0.38; P p .004). A greater number of beds in intensive care units was associated with higher incidences of P. aeruginosa isolates resistant to ceftazidime (r, 0.48; P p .006) and to ciprooxacin (r, 0.40; P p .004). The hospitals with an infection control practitioner had a lower percentage of S. aureus isolates that were MRSA, compared with hospitals without an infection control practitioner (35.5% vs 44.6% of isolates; P p .001). The incidence of MRSA was higher in hospitals that had implemented policies to control the spread of bacterial resistance within the hospital (median [range], 1.02 [0.1-2] vs 0.72 [0.2-2] resistant isolates per 1,000 patient days; P p .04). Similarly, the incidence of ciprooxacin resistance in P. aeruginosa was higher in hospitals that had implemented these policies (median [range], 0.5 [0.1-1.7] vs 0.32 [0.1-0.9] resistant isolates per 1,000 patient days; P p .02). Rate of Antibiotic Consumption and MRSA The overall rate of antibiotic consumption, not including the antibiotics used to treat MRSA infection, explained 13% of the variation in the incidence of methicillin resistance in S. aureus isolates in the hospitals (coefcient, 0.001 [95% condence interval, 0.0001-0.0023]; adjusted multivariate coefcient of determination [aR2], 0.13). The percentage of S. aureus isolates that were MRSA was positively correlated with ciprooxacin consumption, but not with the consumption of other antibiotics (r, 0.31; P p .03), whereas the incidence of MRSA correlated with the consumption of ciprooxacin ( r, 0.37; P p .008), levooxacin (r, 0.33; P p .02), ceftazidime (r, 0.36; P p .01), and carbapenem ( r, 0.32; P p .02). In the multivariate analysis, after adjustment for hospital char-

acteristics and infection control variables, the percentage of S. aureus isolates that were MRSA increased with the use of ciprooxacin and with the percentage of a hospitals beds located in the rehabilitation service (aR2, 0.39). The incidence of MRSA in the hospital increased with the use of ciprooxacin and levooxacin and with the percentage of a hospitals beds that were in intensive care units (aR2, 0.30). Rate of Antibiotic Consumption and AntimicrobialResistant P. aeruginosa In the multivariate analysis, the percentage of P. aeruginosa isolates that were antimicrobial resistant was more likely to be signicantly associated with the incidence of resistance, compared with the rate of antibiotic consumption. between the rate consumption and the percentage of isolates that were resistant than an association between the consumption rate and the incidence of resistance. The incidence of ceftazidimeresistant P. aeruginosa was associated with the consumption of ceftazidime (r, 0.31; P p .02), ciprooxacin (r, 0.35; P p .01), levooxacin (r, 0.43; P p .01), macrolides (r, 0.40; P p .004), gentamicin (r, 0.32; P p .02), and/or rst- or second-generation cephalosporins (r, 0.28; P p .05). The incidence of ciprooxacin-resistant P. aeruginosa was correlated with the consumption of uoroquinolones (r, 0.44; P p .001) and ceftazidime (r, 0.44; P p .001). Multiple linear regression analysis yielded a nal model that included the use of ceftazidime, levooxacin, and gentamicin; rates of consumption of these antimicrobials were all independently and positively associated with the incidence of ceftazidime resistance in P. aeruginosa. The nal model explained 37% of the variance of the dependent variable (aR2, 0.37). The incidence of ciprooxacin resistance in P. aeruginosa increased with the use of uoroquinolones and the percentage of a hospitals beds located in intensive care (aR2, 0.28).

rates of antibiotic use and incidence of resistance in french hospitals

1393

dis c us s i on
This study was based on an ecological design, and it attempted to correlate multicenter surveillance data on antibiotic use with rates of antimicrobial resistance in S. aureus and P. aeruginosa isolates, with the aim of interhospital comparison. There were differences in the incidence of antimicrobial resistance in the 47 hospitals. Rates of resistance were inuenced by the distribution of the various hospital wards within the hospital, and the rate of methicillin-resistance among S. aureus isolates was strongly associated with hospital type (ie, public or private). We showed that hospitals with a high incidence of methicillin resistance in S. aureus isolates also had a high prevalence or incidence of antimicrobial-resistant P. aeruginosa. This suggests that factors affecting the rates of methicillin resistance among the S. aureus isolates in a given hospital might also inuence antimicrobial resistance rates for P. aeruginosa. A signicant relationship existed between the rate of uoroquinolone use and the rates of antimicrobial resistance in S. aureus and P. aeruginosa isolates. That relationship persisted when hospital characteristics, infection control measures, and exposure to other antimicrobial agents were taken into account. The incidence of resistant isolates showed a stronger association with the rate of antimicrobial consumption than did the percentage of isolates with resistance; incidence seemed to be a more relevant indicator than percentage for the study of these ecological relationships. There are some potential limitations to our method that should be considered. First, a study based on aggregated data that uses group-level analytic methods alone does not establish a causal relationship between antibiotic exposure and antimicrobial resistance. The analysis of aggregated data may be limited by ecological bias, which is the failure of groupleveleffect estimates to reect the biological effect at the level of the individual or patient.7 Second, in this cross-sectional study, data were collected for only 1 year, and therefore do not reect longer-term trends in antimicrobial use or resistance. The temporal relationship between antibiotic exposure and resistance may vary according to the time periods examined. However, some studies showed that selection for resistant strains takes a few months in a hospital setting less than 1 year.8-11 Finally, in a multicenter study it is difcult to collect information on all potential confounders, such as the existence of an outbreak, patient case-mix, or collection of clinical microbiology specimens, which were not standardized at our hospitals. Notwithstanding their pitfalls, ecological studies play a potentially useful role in studies of infectious agents because they allow for measurement of the total effect of an exposure. This is important because the total effect of antibiotic exposure encompasses not just the direct effects on the individual who receives the antibiotic therapy but also the indirect effects mediated by effects on transmissibility or on the likelihood of the transmission of susceptible organisms.7

Our ndings were consistent with previous observations that have found a relationship between antimicrobial use and the prevalence of MRSA.3,11-13 The rst multivariate analysis conducted in 50 Belgian hospitals showed that the use of ceftazidime and cefsulodin, broad-spectrum penicillins with b-lactamase inhibitors, and/or quinolones was associated independently with a high incidence of MRSA in the hospital.14 In 15 international hospitals, Westh et al.15 found correlations between the prevalence of MRSA and the therapeutic consumption of b-lactam combinations, carbapenems, and, of course, glycopeptides. They showed that hospitals with a high prevalence of MRSA were those with low consumption of blactamasesensitive antibiotics. Muller et al.16 demonstrated that the selective pressure caused by the use of antimicrobials is an independent risk factor for MRSA acquisition. Furthermore, when relationships between antimicrobial use and MRSA prevalence were analyzed by time-series analysis, temporal relationships were found between the monthly percentage of S. aureus isolates that were MRSA and previous rates of use of macrolides, third-generation cephalosporins, and/or uoroquinolones.11 Several investigations offer evidence suggesting that consumption of uoroquinolones may predispose patients to infection or colonization with MRSA.16-18 The excretion of high concentrations of quinolones in the nares and skin might eliminate normal skin ora and promote colonization by MRSA strains that are quinolone-resistant.19 In 2002, in France, susceptibility to ciprooxacin was found in less than 10% of MRSA isolates.20 Furthermore, exposure to subinhibitory concentrations of quinolones might lead to an increase in the ability of MRSA to adhere to skin.21,22 Recent reports demonstrated that the restriction of ciprooxacin use resulted in a reduced incidence of MRSA in the intensive care unit, as well as in the hospital as a whole.23,24 Two other comments could be made here. First, although the spread of MRSA in the hospital setting is associated with clonal dissemination, we did not nd a relationship between rates of antimicrobial resistance and infection control practices developed in the hospital. This could not be explained by importation of community-associated strains of MRSA into our hospitals, because in France, although MRSA is present in the community and in nursing homes, most MRSA infections are still acquired in the hospital. Our ndings provide evidence for the difculties of measuring infection control practices at the hospital level. Even if hand hygiene is the best way of preventing the transmission of microorganisms in hospital settings, it is difcult to demonstrate the association between consumption rates of alcohol-based hand rub and appropriate adherence to infection control measures by the staff.25 Although many factors are associated with MRSA acquisition, programs aiming to control or improve antimicrobial prescribing practices could be added to infection control measures if the latter used alone were not proven fully effective in controlling MRSA. Society for Healthcare

1394

infection control and hospital epidemiology

december 2007, vol. 28, no. 12

Epidemiology of America guidelines discuss uoroquinolone consumption as a risk factor for MRSA acquisition and suggest that uoroquinolone use should be curtailed in areas with a high incidence of MRSA.26 Second, limitations in analyzing the relationship between consumption and resistance in some projects could be due to the unit choice for analyzing resistance (ie, prevalence vs incidence).27,28 Schwaber et al.29 showed that, from the public health perspective, rates are more appropriate than proportions for measuring the burden of resistance. Case-control studies have identied uoroquinolone use as an individual risk factor for the acquisition of multidrugresistant P. aeruginosa. Hospital rates of uoroquinolone use were predictive of the incidence of uoroquinolone-resistant P. aeruginosa.17,30 Higher hospital-level consumption rates for ciprooxacin, levooxacin, and moxioxacin are each associated with an increased proportion of hospital-acquired P. aeruginosa isolates being nonsuceptible to ciprooxacin.31,32 In another study, a decrease in ciprooxacin susceptibility among P. aeruginosa isolates correlated signicantly with increased use of levooxacin and antipseudomonal cephalosporins. However, some results should be interpreted with caution, because the use of certain antibiotics could be both a cause and a consequence of the emergence of resistance; for example, higher rates of consumption of ceftazidime and aminoglycosides in hospitals with a higher incidence of uoroquinolone-resistant P. aeruginosa is likely to be a result of the high prevalence of uoroquinolone-resistant organisms and the use of alternative drugs to treat infection with these organisms. Use of these non-uoroquinolone drugs may not be a cause of the increasing prevalence of uoroquinoloneresistant P. aeruginosa.33 In Mohr et al.33, as in our study, there was no statistical relationship between the overall rate of antibiotic use and the emergence of antimicrobial resistance in P. aeruginosa, suggesting that the development of resistance is associated with the use of individual agents, rather than with the overall rate of antibiotic consumption.33 Despite the limitations of this kind of ecological study, our results were consistent with recent observations. These results support efforts to reduce prescriptions of selected antimicrobial drug classes, such as uoroquinolones, in the hospital setting. We believe that the approach used to examine antimicrobial use and resistance in our study could be helpful in determining how to prioritize efforts and resources between policies to regulate antibiotic use and other interventions, such as infection control practices, but these factors are difcult to measure. Further study is required to determine the key parameters of infection control practices.

Address reprint requests to Anne-Marie Rogues, MD, PhD, Unite INSERM 657, Universite Bordeaux 2, France (anne-marie.rogues@chu-bordeaux.fr).

re f e re n c e s
1. Turnidge J, Christiansen K. Antibiotic use and resistanceproving the obvious. Lancet 2005; 365: 548-549. 2. Schlaes DM, Gerding DN, John JF, et al. Society for Healthcare Epidemiology of America and Infectious Diseases Society of America Joint Committee on the Prevention of Antimicrobial Resistance: guidelines for the prevention of antimicrobial resistance in hospitals. Infect Control Hosp Epidemiol 1997; 18:275-291. 3. Muller AA, Mauny F, Bertin M, et al. Relationship between spread of methicillin-resistant Staphylococcus aureus and antimicrobial use in French university hospital. Clin Infect Dis 2003; 36: 971-978. 4. Loeb MB, Craven S, McGeer AJ, et al. Risk factors for resistance to antimicrobial agents among nursing home residents. Am J Epidemiol 2003; 157:40-47. 5. Monnet DL. Toward multinational resistance surveillance systems in Europe. Int J Antimicrob Agents 2000; 15:91-101. 6. World Health Organization (WHO). Guidelines for ATC Classication and DDD Assignment. WHO Collaborating Centre for Drug Statistics Methodology, Norwegian Institute of Public Health: Oslo; 2005. Available at: http://www.whocc.no/atcddd/. Accessed October 22, 2007. 7. Harbarth S, Harris AD, Carmeli Y, Samore MH. Parallel analysis of individual and aggregated data on antibiotic exposure and resistance in gram-negative bacilli. Clin Infect Dis 2001; 33:1462-1468. 8. Monnet DL, Lopez-Lozano JM. Campillos P, Burgos A, Yague A, Gonzalo N. Making sense of antimicrobial use and resistance surveillance data: application of ARIMA and transfer function models. Clin Microbiol Infect 2001; 7(suppl 5):29-36. 9. Lopez-Lozano JM, Monnet DL, Yague A, et al. Modelling and forecasting antimicrobial resistance and its dynamic relationship to antimicrobial use: a time series analysis. Int J Antimicrob Agents 2000; 14:21-31. 10. Lepper PM, Grusa E, Reichl H, Hogel J, Trautmann M. Consumption of imipenem correlates with b-lactam resistance in Pseudomomas aeruginosa. Antimicrob Agents Chemother 2002; 46:2920-2925. 11. Monnet DL, MacKenzie FM, Lopez-Lozano JM, et al. Antimicrobial drug use and methicillin-resistant Staphylococcus aureus, Aberdeen, 1996-2000. Emerg Infect Dis 2004; 10:1432-1441. 12. Hill DA, Herford T, Parratt D. Antibiotic usage and methicillin-resistant Staphylococcus aureus: an analysis of causality. J Antimicrob Chemother 1998; 42: 676-677. 13. Graffunder EM, Venezia RA. Risk factors associated with nosocomial methicillin-resistant Staphylococcus aureus (MRSA) infection including previous use of antimicrobials. J Antimicrob Chemother 2002; 49: 999-1005. 14. Crowcroft NS, Ronveaux O, Monnet DL, Mertens R. Methicillin-resistant Staphylococcus aureus and antimicrobial use in Belgian Hospitals. Infect Control Hosp Epidemiol 1999; 20:31-36. 15. Westh H, Scheel Zinn C, Thamdrup Rosdahl V; the SARISA Study Group. An international multicenter study of antimicrobial consumption and resistance in Staphylococcus aureus isolates from 15 hospitals in 14 countries. Microbial Drug Resistance 2004; 10:169-176. 16. Muller PT, Mauny F, Talon D, Donnan A, Harbarth S, Bertrand X. Effect of individual- and group-level antibiotic exposure on MRSA isolation: a multilevel analysis. J Antimicrob Chemother 2006; 58:878-881. 17. MacDougall C, Harpe SE, Powell JP, Johnson CK, Edmond MB, Polk RE. Pseudomonas aeruginosa, Staphylococcus aureus and uoroquinolone use. Emerg Infect Dis 2005; 11:1197-1204. 18. Weber SG, Gold HS, Hooper DC, Karchmer AW, Carmeli Y. Fluoroquinolones and the risk for methicillin-resistant Staphylococcus aureus in hospitalized patients. Emerg Infect Dis 2003; 9:1415-1422. 19. Acar JF, Goldstein FW. Trends in bacterial resistance to uoroquinolones. Clin Infect Dis 1997; 24(suppl 1):S67-73. 20. Conseil scientique de lONERBA. Bacterial resistance to antibiotics: data

acknowledgments
We are grateful to the hospitals that participated in the survey. Potential conicts of interest. All authors report no conicts of interest relevant to this article.

rates of antibiotic use and incidence of resistance in french hospitals

1395

21.

22.

23.

24.

25.

26.

27.

from the National Observatory of Bacterial Resistance Epidemiology (ONERBA). Med Mal Infect 2005; 35:155-169. Harbarth S, Liassine N, Dharan S, Herrault P, Auckenthaler R, Pittet D. Risk factors for persistent carriage of methicillin-resistant Staphylococcus aureus. Clin Infect Dis 2000; 31:1380-1385. Bisognano C, Vaudaux PE, Lew DP, et al. Increased expression of bronectin-binding proteins by uoroquinolone-resistant Staphylococcus aureus exposed to subinhibitory levels of ciprooxacin. Antimicrob Agents Chemother 1997; 41:906-913. Cook PP, Catrou P, Gooch M, Holbert D. Effect of reduction in ciprooxacin use on prevalence of methicillin-resistant Staphylococcus aureus rates within individual units of a tertiary care hospital. J Hosp Infect 2006; 64:348-351. Charbonneau P, Parienti JJ, Thibon P, et al. Fluoroquinolone use and methicillin-resistant Staphylococcus aureus isolation rates in hospitalized patients: a quasi experimental study. Clin Infect Dis 2006; 42:778-784. Talon D, Muller A, Thouverez M, Bertrand X. MRSA incidence as an indicator of infection control practices: do the results reect the reality? J Hosp Infect 2004; 57:265-266. Muto CA, Jernigan JA, Ostrowsky BE, et al. Society for Healthcare Epidemiology of America guideline for preventing nosocomial transmission of multidrug-resistant strains of Staphylococcus aureus and Enterococcus. Infect Control Hosp Epidemiol 2003; 24:362-386. MacKenzie F, Bruce J, Struelens MJ, et al. Antimicrobial drug use and

28.

29.

30.

31.

32.

33.

infection control practices associated with the prevalence of methicillinresistant Staphylococcus aureus in European hospitals. Clin Microbiol Infect 2007; 13:269-276. White RL. How do measurements of antibiotic consumption relate to antibiotic resistance? In: Gould IM, van der Meer JWN, eds. Antibiotics Policies: Theory and Practice. New York: Springer, 2005:75-103. Schwaber MJ, De-Medina T, Carmeli Y. Epidemiological interpretation of antibiotic resistance studieswhat are we missing? Nat Rev Microbiol 2004; 2:979-983. Paramythiotou E, Lucet JC, Timsit JF, et al. Acquisition of multidrugresistant Pseudomonas aeruginosa in patients in intensive care units: role of antibiotics with pseudomonal activity. Clin Infect Dis 2004; 38:670-677. Polk RE, Johnson CK, McClish D, Wenzel RP, Edmond MB. Predicting hospitals rates of uoroquinolone-resistant Pseudomonas aeruginosa use in US hospitals and their surrounding communities. Clin Infect Dis 2004; 39:497-503. Ray GT, Baxter R, DeLorenze GN. Hospital-level rates of uoroquinolone use and the risk of hospital-acquired infection with ciprooxacin-nonsusceptible Pseudomonas aeruginosa. Clin Infect Dis 2005; 41:441-449. Mohr JF, Jones A, Ostrosky-Zeichner L, Wanger A, Tillotson G. Associations between antibiotic use and changes in susceptibility patterns of Pseudomonas aeruginosa in a private, university-afliated teaching hospital: an 8-year-experience: 1995-2002. Int J Antimicrob Agents 2004; 24: 346-351.