Infection
Clinical and Epidemiological Study
Design of a Surveillance System of Antibiotic Use and Bacterial Resistance in German Intensive Care Units (SARI) E. Meyer, D. Jonas, F. Schwab, H. Rueden, P. Gastmeier, F.D. Daschner
Abstract Background: Data on antibiotic consumption and bacterial resistance are important for benchmarking, ensuring quality of antibiotic treatment and helping to understand the relationship between the use of antibiotics and the emergence of resistance. Methods: The SARI project is an ecological study that has established laboratory-based surveillance in German intensive care units (ICU). Resistance rates of 13 sentinel pathogens are reported and certain alert organisms are sent for genotyping and retesting of antimicrobial resistance. Results: The project, initiated in February 2000, now includes 35 ICUs generating a total of 266,013 patient days, 354,356 defined daily doses (DDD) and providing susceptibility data on 21,354 isolates. Pooled antibiotic usage density (AD = DDD/1,000 patient days) was highest for penicillins with lactamase inhibitor (AD 338.3) followed by quinolones (AD 155.5) and second-generation cephalosporins (AD 124.6). Total AD was calculated as 1,337 DDD/1,000 patient days. Resistance rates (RR) for laboratories testing according to the German Industrial Standard (DIN) were 19.3% for methicillin-resistant Staphylococcus aureus(MRSA), 9.5% for ciprofloxacinresistant Escherichia coli and 25.4% for imipenem-resistant Pseudomonas aeruginosa. 40% of the laboratories did not identify the extended spectrum lactamase production of a Klebsiella pneumoniae strain. Conclusion: Focusing on German ICUs, the SARI surveillance system provides a concept that produces a benchmark for the link between antibiotic resistance and consumption. Infection 2003; 31: 208–215 DOI 10.1007/s15010-003-3201-7
Introduction Antibiotics have contributed greatly to improvements in health. However, in addition to their benefits, their increased use and misuse in medicine has resulted in many microorganisms acquiring antimicrobial resistance. This leads to prolonged hospitalization, higher mortality, the need for more expensive and toxic drugs
208
and enhanced costs for health care that have to be borne by society [1]. A larger immunocompromised population, invasive procedures and selective pressures of treatment that often require the use of multiple antimicrobials are factors that make intensive care units (ICUs) a focus for the emergence and spread of many antimicrobial-resistant pathogens [2, 3]. Data from the surveillance system Intensive Care Antimicrobial Resistance Epidemiology (ICARE) and the Antimicrobial Use and Resistance (AUR) component of the US National Nosocomial Infections Surveillance (NNIS) System show that for most antimicrobial agents, the rate of use was highest in the intensive care area and that consumption ran parallel to the pattern seen for resistance [4]. Although there are striking differences in the epidemiology of antibiotic-resistant bacteria between the United States and Germany [5], the overall situation with regard to increasing antimicrobial resistance is similar [6]. Multicenter studies conducted in Germany by the Paul Ehrlich Gesellschaft for Chemotherapy (PEG) between 1990 and 2001 showed the percentage of methicillin-resistant Staphylococcus aureus (MRSA) to have increased from below 3% in 1990, to 12.9% in 1995, 15.2% in 1998 and 20.7% in 2001; the rate of ciprofloxacin-resistant Escherichia coli increased from below 1% in 1990, to 5.2% in 1995, 7.7% in 1998 and 14.5% in 2001 [7–9]. Only a small
E. Meyer (corresponding author), D. Jonas Institute of Environmental Medicine and Hospital Epidemiology, Freiburg University Hospital, Hugstetter Str. 55, D-79106 Freiburg, Germany; Phone: (+49/761) 270-5487, Fax: -5485, e-mail:
[email protected] F. Schwab, H. Rueden Institute of Hygiene, Free University Berlin, Germany; National Reference Center for Surveillance of Nosocomial Infections P. Gastmeier Division of Hospital Epidemiology and Infection Control, Hanover Medical School, Hanover, Germany; National Reference Center for Surveillance of Nosocomial Infections F.D. Daschner Institute of Enviromental Medicine and Hospital Epidemiology, Freiburg, Germany; National Reference Center for Surveillance of Nosocomial Infections Received: November 18, 2002 • Revision accepted: May 23, 2003
Infection 31 · 2003 · No. 4 © URBAN & VOGEL
E. Meyer et al. Antibiotic Use and Resistance Surveillance in Germany (SARI)
number of isolates from ICU patients was tested within the scope of the PEG-studies mentioned above and the resistance rates (RR) for S. aureus, E. coli, Klebsiella pneumoniae and Enterococcus faecalis were found to be higher than those of other inpatient isolates. In Germany, there are few data on resistance rates and antibiotic consumption in ICUs [10, 11]. However, such epidemiological surveillance data allow a comparison of antimicrobial use and antimicrobial resistance rates between hospitals and form the basis for the formulation of strategies to promote the prudent use of antimicrobials and infection control measures [12–14]. Experience in Scandinavian countries has demonstrated that guidelines and national intervention programs for the reduction of antibiotic consumption can reduce bacterial resistance rates without adverse effects on the level of health care [15]. The main objective of project SARI is to provide for the first time ICU-specific, national reference data both on antibiotic consumption and antimicrobial resistance rates. Furthermore, certain “alert” organisms are collected centrally for detailed investigation. Based on this work, SARI intends to develop guidelines for antibiotic use in German ICUs, to analyze mechanisms of resistance and to network with other European surveillance programs like Antibiotic Resistance Prevention and Control (ARPAC; www.abdn.ac.uk).This article describes the implementation of a unit-based method for the surveillance of antimicrobial use and resistance in 35 German ICUs.
Methods SARI is based on the methodology of the Intensive Care Antimicrobial Resistance Epidemiology Project (ICARE), a surveillance system designed to give a national estimate of the prevalence of antimicrobial-resistant organisms in hospitals and the type and quantity of antimicrobials used in these hospitals [4, 16, 17]. SARI focuses on ICUs, whereas ICARE also collects data on non-ICU and outpatient areas. In addition to the system used by ICARE, the participants are asked to send selected pathogens to the study laboratory for retesting and genotyping. This is done in order to obtain additional information on the accuracy of the testing carried out by the laboratory and indications on transmission frequencies of resistant pathogens.
Participating Hospitals Hospitals already participating in the ICU surveillance component of the German Hospital Infection Surveillance System (KISSICU) were invited to take part in project SARI. Participation is voluntary and is not standardized for case mix (type and size of ICU).
Pharmacy Data Once a month, the participating ICUs report use of all oral and parenteral antibiotics in grams. Monthly orders are recorded by the pharmacies of each hospital and transmitted electronically to the study center. For reasons of monthly accounting, we assume these data to be complete and accurate. Thus, no further validation is undertaken. The amount of antimicrobial drugs is standardized by conversion to defined daily doses (DDD) according to the Anatomical Therapeutic Chemical Classification system (ATC) as defined by the World Health Organization (WHO; www.whocc.no). A DDD is defined as the assumed average maintenance dose per day for a drug used for its main indication in adults, but does not necessarily reflect the recommended or prescribed daily dose. Antimicrobial use density (AD) is calculated as the DDD per 1,000 patient days, based on the formula AD =
antibiotic use (g) • defined daily dose (g)
1,000 patient day
and is not only calculated for all SARI ICUs, but is also stratified by type of ICU (medical, surgical-neurosurgical, interdisciplinary).
Resistance Data Each month, 20 microbiology laboratories serving the 35 ICUs report data on the number of tested and resistant isolates for 13 sentinel bacterial species (Table 1). The methods used for reporting comprised questionnaires filled in by hand or web-based masks. Resistance data are checked for plausibility and completeness by the study center in Freiburg. No further validation is undertaken. Laboratories use interpretive National Committee for Clinical Laboratory Standards [18] (NCCLS, n = 7) or the German Industrial Standard [19] (Deutsche Industrienorm DIN 58940, n = 13). In the case of differences between both standards, the resistance defining breakpoint concentration is usually one dilution step higher according to NCCLS than to DIN. All non-duplicate clinical isolates are counted, regardless of whether they are associated with hospital or community-acquired infection or colonization. A “copy” strain is defined as an isolate
Table 1 Antibiotic-pathogen combinations for which susceptibility data are collected.
Pathogen
Antimicrobial agents
S. aureus S. pneumoniae Coagulase-negative staphylococci E. faecalis, E. faecium E. coli, K. pneumoniae
Oxacillin, vancomycin, teicoplanin, gentamicin, ciprofloxacin Penicillin (= oxacillin 1µg), cefotaxime, vancomycin, erythromycin, ciprofloxacin or ofloxacin Quinupristin/dalfopristin, vancomycin, teicoplanin Vancomycin, teicoplanin, for E. faecium: quinupristin/dalfopristin Cefotaxime or ceftazidime or cefriaxone, imipenem, amikacin, piperacillin/tazobactam, ampicillin/ sulbactam, amoxicillin/clavulanic acid, ciprofloxacin Imipenem, amikacin, ciprofloxacin Ceftazidime, piperacillin/tazobactam, imipenem, amikacin, ciprofloxacin Ceftazidime, piperacillin/tazobactam, amikacin, ciprofloxacin, trimethoprim/sulfamethoxazole
E. cloacae, Citrobacter, S. marcescens P. aeruginosa, A. baumannii S. maltophilia
Infection 31 · 2003 · No. 4 © URBAN & VOGEL
209
E. Meyer et al. Antibiotic Use and Resistance Surveillance in Germany (SARI)
of the same species of bacteria showing the same antimicrobial susceptibility pattern in the same patient throughout a 1-month period, whatever the site of isolation.
Statistics Pharmacy and microbiology data are analyzed quarterly using SAS version 8.01 software (SAS). A pooled mean RR is calculated for each combination of antimicrobial-resistant organism by pooling the data from all SARI ICUs using interpretive standards (NCCLS or DIN). For each ICU, individual resistance rates and key percentile distribution are calculated and pooled over time. ICUs can compare their own resistance data with those of other ICUs by looking at the median, the 25th and 75th percentiles and mean RR. The number of resistant isolates (number of isolates tested against a specified antimicrobial per month) are divided by the total number of isolates for which susceptibility testing had been done. The pooled number of grams for each antibiotic applied by ICUs is divided by the number of grams per DDD for a specified antibiotic, divided by the number of patient days of this ICU and multiplied by 1,000 to derive the number of DDD per 1,000 patient days or the antimicrobial use density (AD). For comparison of antibiotic use, all ICUs feed back their own AD, as well as mean antimicrobial usage rates and key percentiles. Values above the 75th percentile are interpreted as high compared to the rest of SARI ICUs, but do not necessarily imply a problem. Differences between antibiotic use density in project ICARE and SARI were tested by one-sample Wilcoxon rank test.
Validation Testing In March 2001, a challenge panel was sent to all microbiology laboratories. The panel of ten characterized strains included four American Type Culture Collection (ATCC)-type strains and six strains which had been carefully characterized previously [20]. Without further specification, the participants were asked to identify the species and to test the antimicrobial susceptibility as in the case of isolates of clinical importance. Testing errors were determined by comparing the clinically categorized MIC of the isolates (i.e. susceptible, intermediate or resistant) with the results from the participating laboratories. A summary of the results was reported to all the laboratories testing according to the same standard (DIN or NCCLS).
Centralized Susceptibility Testing All participants were asked to send selected clinical isolates for retesting and genotyping by the study laboratory, i.e. MRSA, vancomycin-resistant enterococci (VRE) and staphylococci, quinupristin/dalfopristin-resistant enterococci and staphylococci, cephalosporin-resistant K. pneumoniae, Pseudomonas aeruginosa and Enterobacteriaceae resistant to amikacin, quinolones or carbapenems. There were no regulations for the shipping of samples or types of media. Methicillin resistance was confirmed by growth on MuellerHinton agar supplemented with 4% NaCl and 6 µg/ml oxacillin (Heipha, Heidelberg, Germany). Agar plates were incubated at 35 °C for 24 h. S. aureus was identified by Staphytect Plus-Test DR 850 M (Oxoid, Wesel, Germany). In case of ambiguities, isolates were identified by mecA, femB duplex PCR [20, 21] . Cephalosporin-resistant K. pneumoniae were retested in the Vitek GNS-500 card (bioMérieux, Nürtingen, Germany). The MICs of the remaining gram-negative bacteria were tested by
210
Mueller-Hinton agar dilution, according to NCCLS. For clinical categorization of the determined MIC, different breakpoint concentrations used in laboratories testing according to DIN or NCCLS were taken into account. Furthermore, an interassay MIC variation of ± 1 dilution step was allowed. Therefore, only strains with a MIC of more than two dilution steps below the resistance breakpoint were counted as erroneous.
Centralized Genotyping In order to determine the unit-based diversity of resistant pathogens, isolates were typed using published protocols. MRSA isolates were genotyped with pulsed-field gel electrophoresis (PFGE) as described [22, 23]. The gram-negative bacteria were genotyped by means of amplified fragment length polymorphism (AFLP) [24]. In case of P. aeruginosa, K. pneumoniae and E. coli, published protocols were employed [25–27].
Results Participating Hospitals SARI was presented to all KISS hospitals in 1999 and they were invited to take part in it. The project started in February 2000 and initially included 12 ICUs; by 2001, the number had increased to 35 ICUs. Twenty hospitals, representing 35 ICUs, submitted data to SARI and the ICU component of the KISS system. These ICUs are separated geographically and represent different categories of bed-size and teaching affiliation (Table 2). Until June 2002, ICUs supplied data on a total of 744 months (12 ICUs on 29 months, two on 27, seven on 18, two on 17, six on 16, three on 15, one on 14 and two ICUs on 13 months).
Validation Testing All the participating laboratories returned their identification and susceptibility testing results. Seventeen of the 20 laboratories correctly identified the nine different species of the ten strains sent in. One laboratory wrongly identified
Table 2 Characteristics of SARI ICUs.
Characteristics Type of hospital • University • Teaching affiliation • Others
13 (37%) 6 (46%) 6 (17%)
Type of ICU • Medical • Interdisciplinary • Surgical-neurosurgical
10 14 11
No. of ICU beds • 6–11 • 12–20 • 21–26
15 15 5
ICU: intensive care unit
Infection 31 · 2003 · No. 4 © URBAN & VOGEL
E. Meyer et al. Antibiotic Use and Resistance Surveillance in Germany (SARI)
the E. coli strain as Citrobacter Table 3 freundii. Two laboratories each Pooled means of antimicrobial usage rates (DDD/1,000 patient-days) and percentiles of the distriincorrectly identified two sepabution, all SARI ICUs (n=35), February 2000 through June 2002. rate strains; Staphylococcus intermedius instead of S. aureus, Percentile Enterococcus hirae instead of Group of antimicrobial agenta No. DDD Pooled mean 75th 50th 25th Enterococcus faecium, Corynebacterium spp. instead of Penicillins with extended spectrum 18,707 70.3 114.9 65.5 25.8 Penicillins with lactamase inhibitor 89,991 338.3 491.4 248.4 165.1 Acinetobacter baumanii and Second-generation cephalosporins 33,101 124.6 166.6 102.4 59.4 Brevibacterium instead of Third-generation cephalosporins 29,078 109.5 145.9 106.1 82.9 Stenotrophomonas maltophila. Carbapenems 22,258 83.7 115.6 63.0 37.2 The laboratories reported Glycopeptides 11,265 42.3 55.9 28.9 18.4 530 of 560 susceptibility results Quinolones 41,396 155.5 215.2 130.7 91.0 of interest. Of these, 508 correa antimicrobials are grouped according to the WHO ATC classification (www.whocc.no); DDD: desponded to the expected refined daily doses sults. Discrepancies (4%) mainly concerned testing resistance to piperacillin/tazobactam, ceftazidime and the cording to the WHO classification) and a total antibiotic usaminoglycosides. age density of 1,337 DDD/1,000 patient days. Pooled an40% of the laboratories did not identify the extendedtibiotic usage rates (AD) were highest for penicillins with spectrum -lactamase production of the K. pneumoniae lactamase inhibitor, followed by quinolones and secondstrain, with a MIC of ceftazidime that decreased by more generation cephalosporins. Table 3 shows antibiotic usage than two dilution steps in the presence of clavulanic acid. density for exemplary antibiotics grouped according to the WHO ATC classification. Stratifying by type of ICU shows differences in the use Retesting of Resistant Pathogens of some antibiotic groups (Table 4). For instance, gly540 gram-negative isolates, which had been classified as flucopeptides are used almost twice as often in surgical as in oroquinolone-resistant by the laboratories, were retested interdisciplinary ICUs. by the study laboratory. However, 7.7% of these isolates revealed a MIC unambiguously below the breakpoint concentration that categorizes the isolate as resistant. Resistance rates (RR) After retesting 441 isolates, primarily indicated as carSusceptibility data were reported on 24,494 isolates of 13 senbapenem resistant, 7.0% turned out not to be resistant. tinel pathogens from February 2000 to June 2003. RR and Analysis of the data revealed that two laboratories had erpercentile distribution are provided for ICUs according to roneously filled in the forms accompanying each isolate the method of resistance testing (DIN or NCCLS) used.The posted to them. In the end, only three susceptible isolates six most frequent pathogens isolated in the participating remained that had been falsely designated as resistant. ICUs were S. aureus, E. coli, coagulase-negative staphyloIn the case of 235 amikacin-resistant strains, 15.7% cocci (CNS), E. faecalis, P. aeruginosa and K. pneumoniae. were mistakenly identified as resistant. Identification and Table 5 gives an example of inter-ICU comparison of semethicillin resistance was confirmed in 608 of 624 MRSA lected pathogens from SARI-ICUs testing according DIN. isolates by the study laboratory. In order to estimate coverage of the Table 4 received isolates, the numbers of P. Mean antibiotic use density (DDD per 1,000 patient days) by type of ICU, February 2000 aeruginosa isolates obtained were comthrough June 2002. pared with the figures reported for calMean antibiotic use density by type of ICU culation of resistant rates. 2,198 imipenem-resistant P. aeruginosa strains Group of antimicrobial agenta Medical Interdisciplinary Surgical were reported for calculation of resisPenicillins with extended spectrum 95.7 78.5 43.5 tance rates, but only 243 were sent to the Penicillins with lactamase inhibitor 308.0 326.5 372.1 study laboratory. Similarly, only 239 of Second-generation cephalosporins 114.5 123.0 133.6 2,259 reported ciprofloxacin-resistant P. Third-generation cephalosporins 116.1 93.8 118.4 aeruginosa strains were sent in to the Carbapenems 74.3 63.8 108.7 study laboratory. Glycopeptides 37.4 30.4 56.9
Antibiotic Usage Data from 266,013 patient days were collected on a total of 354,356 DDD (ac-
Infection 31 · 2003 · No. 4 © URBAN & VOGEL
Quinolones a
141.4
172.3
151.2
antimicrobials are grouped according to the WHO ATC classification (www.whocc.no)
211
E. Meyer et al. Antibiotic Use and Resistance Surveillance in Germany (SARI)
Table 5 Pooled mean antimicrobial resistance rates and key percentiles of selected pathogens from SARI ICUs reporting according DIN (n = 20), February 2000 through June 2002.
Genotyping
To prove the simplified assumption of ICARE [17] that high RR on ICUs are exclusively the Percentile result of either crossPathogen Resistant to No. tested Pooled mean 75th 50th 25th transmission or overuse S. aureus Oxacillin 2,516 19.3 30.7 14.3 6.0 of antimicrobials, certain Vancomycin 2,386 0.0 0.0 0.0 0.0 isolates from ICUs with Teicoplanin 1,450 0.0 0.0 0.0 0.0 both high RR and high Ciprofloxacin 2,363 20.7 30.4 18.5 8.1 AD were investigated E. faecalis Vancomycin 1,741 0.1 0.0 0.0 0.0 and assigned to genoTeicoplanin 1,099 0.1 0.0 0.0 0.0 types in a unit based E. faecium Vancomycin 598 2.7 2.8 0.0 0.0 manner (Table 6). HowQuinupristin/dalfopristin 387 4.7 10.5 0.0 0.0 ever, it should be K. pneumoniae Cefotax./ceftr./cefta. 1,099 6.3 14.0 2.3 0.0 pointed out that collaboImipenem 752 0.1 0.0 0.0 0.0 rating laboratories failed Ciprofloxacin 830 7.5 10.6 6.3 1.7 to send a considerable Amikacin 716 0.6 0.0 0.0 0.0 proportion of the rePiperacillin/tazobactam 778 9.1 12.3 3.8 0.0 quested isolates. Ampicillin/sulbactam 697 21.8 31.3 18.9 2.1 In the case of E. coli Cefota./ceftr./cefta. 2,393 1.0 1.7 0.3 0.0 cephalosporin usage and Imipenem 1,584 0.0 0.0 0.0 0.0 MRSA, in two out of five Ciprofloxacin 1,897 9.5 13.9 10.1 4.8 ICUs the predominant Amikacin 1,447 1.0 1.7 0.0 0.0 genotype was found in Piperacillin/tazobactam 1,828 5.9 11.8 5.1 1.2 Ampicillin/sulbactam 1,611 23.7 53.3 29.5 10.5 less than half of all strains investigated. However, P. aeruginosa Ceftazidime 1,494 15.3 21.7 12.3 4.9 on three ICUs the frePiperacillin/tazobacam 1,463 23.1 37.2 21.6 7.2 Imipenem 1,334 25.4 28.8 19.1 6.0 quencies were as high as Amikacin 1,027 7.7 10.7 3.4 0.0 79% to 93%, which is inCiprofloxacin 1,496 18.0 33.5 15.5 7.6 dicative of increased transmission rates of MRSA and simultaneTable 6 ously high antimicrobial usage. Similarly, varying percentages Predominance of genotypes in a setting of simultaneously high of predominance were found on ICUs with high imipenem resistance rates and antibiotic use density (DDD/1,000 patient usage and high rates of imipenem-resistant P. aeruginosa. days) above the median. ICUa
No. of strains tested
Frequencies of predominant genotypes (%)
Cephalosporin usage and MRSA
1 2 3 5 6
86 17 14 29 33
79 41 43 93 79
Imipenem usage and resistant P. aeruginosa
1 3 4 7 8 9 a
11 9 9 11 7 8
18 55 22 18 28 25
ICUs coded by study numbers; MRSA: methicillin-resistant Staphylococcus aureus; ICU: intensive care unit
212
Discussion The German Government, the EU Commission and the WHO recognize the importance of studying the emergence and determination of resistance and the need for strategies to control it [28]. Data of the European Antimicrobial Surveillance System (EARRS) show a marked difference in antimicrobial RR between different European countries, resulting most likely from differences in antibiotic consumption and differences in hospital infection control practice [6].As a consequence, general policies to control resistance need to be specifically tailored to countries, hospitals and high-risk areas like ICUs. We present the first German surveillance system for the evaluation of antimicrobial use and bacterial resistance rates in ICUs. The importance of comparing national data for the purpose of quality management was understood and appreciated by all the ICUs that volunteered to take part in project SARI.
Infection 31 · 2003 · No. 4 © URBAN & VOGEL
E. Meyer et al. Antibiotic Use and Resistance Surveillance in Germany (SARI)
In the United States, the joint project ICARE of the Centers for Disease Control and Prevention (CDC) Hospital Infections Program and the Emory University Rollins School of Public Health began in 1996 at a subset of hospitals participating in the CDC’s National Nosocomial Infection Surveillance System (NNIS). Data collection ended in 1999. It has now been succeeded by the Antimicrobial Use and Resistance (AUR) component of NNIS. Hospitals report data from all three areas: these include at least one ICU, all inpatient and all outpatient areas. The key parameters of antimicrobial use can be correlated with antimicrobial resistance levels and tracked through the hospital’s quality improvement indicator process. For instance, as in a previously published study, an association was shown between the use of vancomycin or third-generation cephalosporin and an increased prevalence of VRE in 126 US ICUs [29]. We adopted the idea of a national surveillance system from ICARE, but modified its design. Firstly, since ICARE and other data show that higher usage rates for many antimicrobials run in parallel with higher resistance rates, especially in ICUs, and because of the workload involved, SARI focuses only on these high-risk areas. Secondly, for purposes of national and international comparison, in project SARI each antibiotic has been assigned a DDD following the recommendation of the WHO. ICARE converted their pharmacy data into their own DDDs, which must be taken into account when comparing antibiotic usage data. Thirdly, a simple causative explanation of antimicrobial resistance through use of antibiotics would suggest a statistical correlation between both factors. However, besides antimicrobial use, increased RR might be due to transmissions of resistant pathogens between patients. Thus, high RR in the presence of low AD may be caused by crosstransmission of resistant pathogens.This should be reflected by the predominance of certain genotypes and lower species diversity. In contrast, high RR in the presence of a high AD would primarily suggest an overuse of antimicrobials and not the above-average transmission of pathogens. Hence, a higher genodiversity should be expected. For this reason, selected resistant pathogens are genotyped in project SARI. Dramatic differences in antimicrobial resistance may exist between different ICUs within individual hospitals and may depend on both antimicrobial use and infection control practices. Even in the absence of epidemiological patient data, unit-based genotyping can be a first indirect indicator of transmissions, which is helpful for the allocation of scarce resources. Furthermore, retesting and confirmation of a reported antimicrobial susceptibility can provide an indication of the reliability of the resistance rates communicated by the participants. Various objections might be raised to this study design. Firstly, as was also the case in the ICARE project, the data are unit and laboratory-based.This gives an overview of the
Infection 31 · 2003 · No. 4 © URBAN & VOGEL
ecological situation in a larger number of ICUs, but does not permit a patient-based differentiation between colonization and infection, origin of acquisition of the resistant pathogen or the influence of variables like frequency of microbiological cultures.The ecological study design can lead to the formation of a hypothesis but does not ultimately prove a causative relationship that could be obtained in patient-based study designs. However, a patient-based design would require personnel and financial resources beyond the reach of this work and most health care institutions. Secondly, in contrast to the United States, laboratories in Germany do not apply one antimicrobial testing standard. They either test according to NCCLS or DIN 58940. The different breakpoints specified by NCCLS or DIN necessitate separate analysis for both of the interpretation standards. Thirdly, bias should be considered with regard to the following points: ICUs differ according to the number of months they participated. However, these differences will diminish over time. A selection bias is likely because a notable number of resistant strains requested for genotyping was not sent. In addition, misclassification of sensibility testing may bias the resistance data. The main objective of this work on SARI is to describe the project design and test the validity of sampling methodology. Initial results on antibiotic consumption and RR, of which only a few examples are described here, will be discussed elsewhere. Only two aspects of the data are discussed in more detail: the validity of resistance data and the distinction between cross-transmission and antimicrobial usage as the origin of high resistance rates. Similarly to ICARE, the proficiency of the microbiology laboratories of the ICUs in testing antimicrobial susceptibility was investigated with a challenge panel of wellcharacterized strains [30]. Since this panel included strains that are resistant to the antimicrobials of interest, this provided additional proof that resistant strains were actually recognized by the participating microbiology laboratories. Validation testing showed accurate results for bacterial identification and susceptibility testing in 97.5% and 95.8%, respectively. Yet, an alarming finding was that only 40% of the laboratories reported the ESBL(extended spectrum lactamase)-positive strain. The problem of underreporting this new emerging resistance has already been described elsewhere [31]. The evaluation and correlation of antibiotic consumption and antimicrobial resistance is hampered by numerous confounding risk factors for resistance in ICUs other than the use of antibiotics. Ecological constraints, such as antimicrobial usage, can give rise to resistant pathogens. Moreover, the lack of infection control can increase the number of resistant isolates resulting from cross-transmission, which leads to the predominance of certain genotypes and a lowered diversity. Due to the principle of unit-based, rather than patientbased data collection, the occurrence of a distinct trans-
213
E. Meyer et al. Antibiotic Use and Resistance Surveillance in Germany (SARI)
mission event cannot be proven. However, in the absence of epidemiological patient data, genodiversity can serve as an indirect marker of the transmission frequencies in different ICUs. MRSA was chosen for a first comparison, because meaningful numbers of typed isolates and ICUs could only be achieved with this most frequent resistant species. It might be argued that MRSA is unsuitable for use in such a comparison of different ICUs, because the diversity in hospitals and their catchment population can be limited to few genotypes [32, 33]. However, a marked variation in the range of diversity was found between different ICUs, even in the case of MRSA. In addition, similar findings were demonstrated in ICUs with high imipenem consumption and high resistance rates of P. aeruginosa resistance rates, a bacterial species of high genodiversity. Surveillance is an integral part of controlling resistance. Local, national and international surveys are needed to identify, monitor and study the epidemiology of the emergence and spread of resistant isolates. This is the first report on a novel surveillance system, with a modified study design from ICARE, for both antibiotic consumption and antimicrobial resistance in intensive care settings in Germany.The epidemiological data obtained by implementation of this surveillance system serve the participating ICUs and other ICUs as benchmark data, and form the basis for infection control, prudent antibiotic use and quality management.
7.
8.
9.
10. 11. 12.
13.
14.
15.
Acknowledgments This study was supported by a grant from the German Ministry of Science and Education (01 KI 9907).We would like to thank all the study members of SARI and B. Schroeren-Boersch, B. Spitzmueller, T. Chojnacki and D. Lawrie-Blum, Institute of Environmental Medicine and Hospital Epidemiology, Freiburg University Hospital.
16.
17.
References 1.
2.
3. 4.
5.
6.
Barenfanger J, Short MA, Groesch AA: Improved antimicrobial interventions have benefits. J Clin Microbiol 2001; 39: 2823–2828. Archibald L, Phillips L, Monnet D, McGowan JE Jr, Tenover F, Gaynes R: Antimicrobial resistance in isolates from inpatients and outpatients in the United States: increasing importance of the intensive care unit. Clin Infect Dis 1997; 24: 211–215. Fridkin SK: Increasing prevalence of antimicrobial resistance in intensive care units. Crit Care Med 2001; 29(suppl 4): N64–8. Fridkin SK, Steward CD, Edwards JR, Pryor ER, McGowan JE Jr, Archibald LK, Gaynes RP, Tenover FC: Surveillance of antimicrobial use and antimicrobial resistance in United States hospitals: project ICARE phase 2. Project Intensive Care Antimicrobial Resistance Epidemiology (ICARE) hospitals. Clin Infect Dis 1999; 29: 245–252. Harbarth S, Albrich W, Goldmann DA, Huebner J: Control of multiply resistant cocci: do international comparisons help? Lancet Infect Dis 2001; 1: 251–261. Bronzwaer SL, Cars O, Buchholz U, Molstad S, Goettsch W, Veldhuijzen IK, Kool JL, Sprenger MJ, Degener JE: A European study
214
18.
19.
20.
21.
22.
on the relationship between antimicrobial use and antimicrobial resistance. Emerg Infect Dis 2002; 8: 278–282. Kresken M, Hafner D: Drug resistance among clinical isolates of frequently encountered bacterial species in central Europe during 1975-1995. Study Group Bacterial Resistance of the PaulEhrlich-Society for Chemotherapy. Infection 1999; 27 (Suppl 2): 2–8. Kresken M. Hafner D, Schmitz F-J, Wichelhaus T. für die Studiengruppe: Resistenzsituation bei klinisch wichtigen Infektionserregern gegenüber Antibiotika in Deutschland und im mitteleuropäischen Raum. Bericht über die Ergebnisse einer multizentrischen Studie der Arbeitsgemeinschaft Empfindlichkeitsprüfungen & Resistenz der Paul-Ehrlich-Gesellschaft für Chemotherapie e.V. aus dem Jahr 2001. Antiinfectives Intelligence. Bonn, 2003. Kresken M, Hafner D: Resistenzsituation bei klinisch wichtigen Infektionserregern gegenüber Chemotherapeutika in Mitteleuropa. Chemotherapie Joural 2000; 9: 51–86. Albrich WC, Angstwurm M, Bader L, Gartner R: Drug resistance in intensive care units. Infection 1999; 27 (Suppl 2): 19–23. Daschner F, Langmaack H, Wiedemann B: Antibiotic resistance in intensive care unit areas. Infect Control 1983; 4: 382–387. Elliott TS, Lambert PA: Antibacterial resistance in the intensive care unit: mechanisms and management. Br Med Bull 1999; 55: 259–276. Monnet DL: Toward multinational antimicrobial resistance surveillance systems in Europe. Int J Antimicrob Agents 2000; 15: 91–101. Weinstein RA: Controlling antimicrobial resistance in hospitals: infection control and use of antibiotics. Emerg Infect Dis 2001; 7: 188–192. Monnet DL, Sorensen TL: Interpreting the effectiveness of a national antibiotic policy and comparing antimicrobial use between countries. J Hosp Infect 1999; 43: 239–242. National Nosocomial Infections Surveillance (NNIS) System Report, Data Summary from January 1992-June 2001, issued August 2001. Am J Infect Control 2001; 29: 404–421. Monnet DL, Archibald LK, Phillips L, Tenover FC, McGowan JE Jr, Gaynes RP: Antimicrobial use and resistance in eight US hospitals: complexities of analysis and modeling. Intensive Care Antimicrobial Resistance Epidemiology Project and National Nosocomial Infections Surveillance System Hospitals. Infect Control Hosp Epidemiol 1998; 19: 388–394. National Committee for Clinical Laboratory Standards: Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically (4th edn). Approved Standard M7-A4. Wayne, PA, NCCLS, 1997. DIN 58940-4 Medical microbiology - Methods for the determination of susceptibility of pathogens (except mycobacteria) to antimicrobial agents - part 4: Evaluation classes of the minimum inhibitory concentration. Beuth Verlag, Berlin 1998, pp 1–11. Frank U, Jonas D, Lupke T, Ribeiro-Ayeh B, Schmidt-Eisenlohr E, Ruden H, Daschner FD: Antimicrobial susceptibility among nosocomial pathogens isolated in intensive care units in Germany. Eur J Clin Microbiol Infect Dis 2000; 19: 888–891. Towner KJ, Talbot DC, Curran R, Webster CA, Humphreys H: Development and evaluation of a PCR-based immunoassay for the rapid detection of methicillin-resistant Staphylococcus aureus. J Med Microbiol 1998; 47: 607–613. Tenover FC, Arbeit RD, Goering RV, Mickelsen PA, Murray BE, Persing DH, Swaminathan B: Interpreting chromosomal DNA restriction patterns produced by pulsed-field gel electrophoresis:
Infection 31 · 2003 · No. 4 © URBAN & VOGEL
E. Meyer et al. Antibiotic Use and Resistance Surveillance in Germany (SARI)
23.
24.
25.
26.
27.
28.
criteria for bacterial strain typing. J Clin Microbiol 1995; 33: 2233–2239. Winters MA, Goering RV, Boon SE, Morin R, Sorensen M, Snyder L: Epidemiological analysis of methicillin-resistant Staphylococcus aureus comparing plasmid typing with chromosomal analysis by field inversion gel electrophoresis. Med Microbiol Lett 1993; 2: 33–41. Jonas D, Spitzmuller B, Weist K, Rueden H, Daschner F: Comparison of PCR-based methods for typing Escherichia coli. Clin Microbiol Infect 2003 (in press). Speijer H, Savelkoul PH, Bonten MJ, Stobberingh EE, Tjhie JH: Application of different genotyping methods for Pseudomonas aeruginosa in a setting of endemicity in an intensive care unit. J Clin Microbiol 1999; 37: 3654–3661. van der Zwet WC, Parlevliet GA, Savelkoul PH, Stoof J, Kaiser AM, Koeleman JG, Vandenbroucke-Grauls CM: Nosocomial outbreak of gentamicin-resistant Klebsiella pneumoniae in a neonatal intensive care unit controlled by a change in antibiotic policy. J Hosp Infect 1999; 42: 295–302. Vos P, Hogers R, Bleeker M, Reijans M, van de Lee T, Hornes M, Frijters A, Pot J, Peleman J, Kuiper M: AFLP: a new technique for DNA fingerprinting. Nucleic Acids Res 1995; 23: 4407–4414. Proposal for a Council recommendation on the prudent use of antimicrobial agents in human medicine. Official Journal of the European Communities L 34/13, 13502-01. 2001.
Infection 31 · 2003 · No. 4 © URBAN & VOGEL
29. Fridkin SK, Edwards JR, Courval JM, Hill H, Tenover FC, Lawton R, Gaynes RP, McGowan JE Jr: The effect of vancomycin and thirdgeneration cephalosporins on prevalence of vancomycin-resistant enterococci in 126 U.S. adult intensive care units. Ann Intern Med 2001; 135: 175–183. 30. Steward CD, Wallace D, Hubert SK, Lawton R, Fridkin SK, Gaynes RP, McGowan JE Jr, Tenover FC: Ability of laboratories to detect emerging antimicrobial resistance in nosocomial pathogens: a survey of project ICARE laboratories. Diagn Microbiol Infect Dis 2000; 38: 59–67. 31. Tenover FC, Mohammed MJ, Gorton TS, Dembek ZF: Detection and reporting of organisms producing extended-spectrum beta-lactamases: survey of laboratories in Connecticut. J Clin Microbiol 1999; 37: 4065–4070. 32. O'Neill GL, Murchan S, Gil-Setas A, Aucken HM: Identification and characterization of phage variants of a strain of epidemic methicillin-resistant Staphylococcus aureus (EMRSA-15). J Clin Microbiol 2001; 39: 1540–1548. 33. Jonas D, Towner KJ, Loerwald M, Shunburne L, Schmidt-Wieland T, Daschner FD: Diversity of Staphylococcus aureus strains isolated from two European regions with different prevalences of methicillin resistance. Eur J Clin Microbiol Infect Dis 2002: 21: 880–883.
215