h g o Klare, Dietlitide Badstiibner, Carola Koristabel atid WoEfgang Witte

ORIGINAL ARTICLE Identification of enterococci and determination of their glycopeptide resistance in German and Austrian clinical microbiology laboratories Clin Microbial Infect 1999; 5: 535-539 h g o Klare, Dietlitide Badstiibner, Carola Koristabel atid WoEfgang Witte Robert Koch Institute, Wernigerode, Germany Objective: To determine exactly how German and Austrian routine laboratories perform tests for the identification of enterococci and determination of their glycopoptide resistance. Methods: Six enterococcal test strains with different types of glycopeptide resistance (Enterococcus faecium, VanA; E. faecalis, VanA; E. faeciurn, VanB; E. faecalis, VanB; E. gallinarum, VanC1; E. casseliflavus, VanC2) were sent as anonymous isolates to 73 clinical microbiology laboratories (65 in Germany; eight in Austria). The participating laboratories had to identify the strains up to the species level and to determine their antibiotic susceptibilities to the glycopeptides vancomycin and teicoplanin by the test method(s) that are used daily in the corresponding laboratories. Results: The analysis of the results received from 62 laboratories (56 from Germany, six from Austria) demonstrated that the most used routine method in susceptibility testing was the agar diffusion test, followed by the Etest, and the microbroth dilution procedure. The majority of participants had no difficulties in susceptibility testing with the VanAtype strains. However, the agar diffusion test was often not able to recognize clearly the VanB and VanC strains; some problems also arose with VanB isolates in the Etest. With the microbroth dilution method, the corresponding type of glycopeptide resistance was correctly determined in the majority of enterococcal test strains. Difficulties also arose in identification, especially with the VanC strains (E. gallinarum and E. casseliflavus), which were often falsely identified as E. faecium. The reasons for these errors are obviously based on the lack of important tests (such as motility and presence of a yellow pigment) in some commercially available identification test kits. Key words: Enterococci, identification, glycopeptide resistance, multicenter study INTRODUCTION The emergence of glycopeptide resistance in enterococci threatens the chemotherapy of enterococcal infections, especially in Enterococcrts fuecitrm, where resistance to ampicillin is frequent. The incidence of glycopeptide-resistant enterococci (GRE) has increased remarkably during the last 10 years in US hospitals (up to 16% in intensive care units and up to 12% in other hospital units in 1997 [l]). Although less frequent, serious but sporadic hospital infections and nosocomial outbreaks with G R E have also been reported in Europe, including Germany [2]. Correct detection of glycopeptide resistance in the clinical laboratory is a prerequisite for chemotherapy and prevention of nosocomial spread. Corresponding author and reprint requests: lngo Klare, Robert Koch Institute, Wernigerode Branch, BurgstraRe 37, D-38855-Wernigerode, Germany Tel: +49 3943 679247 Fax: +49 3943 679207 E-mail: klarei@rki.de Revised version accepted 10 March 1999 The NCCLS has recently published its latest minimal inhibitory concentration (MIC) interpretative criteria for glycopeptides [3], which also correspond approximately to the DIN standard [4]. In enterococci, six different genotypes of glycopeptide resistance are known: tma mediates high-level vancomycin resistance (with MIC 2128 mg/l in most isolates); also, for tland, MICs for vancomycin above 32 mg/l have been described [5,6]. For t m B, the vancomycin MICs are in the range of the breakpoint; however, they can also be up to 1024 mg/l [5]. The vuna gene cluster also mediates resistance to teicoplanin whereas vunb and vund do not. The three intrinsic vanc genotypes (vunc1, imcz, ttznc3) o f particular enterococcal species (E. gullinarum, E. cassel$uvus, E. Juvescens) confer low-level resistance to vancomycin (and, simultaneously, susceptibility to teicoplanin [5]) with no clinical significance. However, there are also reports concerning strains of these latter species that were high-level glycopeptide resistant as a result of picking up the vuna gene cluster [7]. Two previous studies, one in the UK and one in the USA [8,9], revealed that detection of resistance 535

536 Clinical Microbiology and Infection, Volume 5 Number 9, September 1999 genotypes other than vana in the clinical laboratory can be problematic; only 5040% of GRE with the vanb genotype were classified as resistant to vancomycin. The present study reports on the capacity of German and Austrian clinical microbiology laboratories to correctly detect GRE. MATERIALS AND METHODS Study design Six GRE strains possessing different genotypes of glycopeptide resistance were sent as anonymous isolates to 73 laboratories (65 in Germany, eight in Austria), where they were checked. The characteristics of these strains are shown in Table 1; the genotypes of glycopeptide resistance had been demonstrated by polymerase chain reaction [lo]. The laboratories performed bacterial identification and antibiotic susceptibility testing according to their routine procedures. They returned a questionnaire on the species detected, the results from the susceptibihty test method used (difision zones in agar difision test, MICs in dilution methods and in Etest), the criteria applied for interpretation, and the derived susceptibility category. The laboratory data received from the participants were analyzed by comparison with the known enterococcal species and their glycopeptide susceptibilities, and are summarized in Tables 2 and 3. RESULTS Of the 73 clinical microbiology laboratories that received the enterococcal strains, 62 (56 from Germany and six &om Austria) returned the questionnaires with their test results for analysis (85%). Most of the 62 institutions applied one definite susceptiblity method, but some performed up to four. The majority of the participating laboratories had used the agar diffusion test (Table 2). When VanA strains were tested for vancomycin, they were always classified correctly. In contrast, when VanA isolates were tested for teicoplanin, E. faecium was classified correctly as resistant by all laboratories, but E. faecalis was classified correctly in only 74% of cases. Additionally, the agar difision test often failed to class* VanB strains correctly (especially when testing vancomycin). The Etest classified the E. jiecium VanA type strain 70/90 correctly; about one-third of the laboratories determined the VanA E. jiecalis isolate 1528 as intermediate or susceptible to teicoplanin by this method. For the detection of VanB strains, the Etest performed slightly better than the agar difhsion test (especially in the case of vancomycin). With vancomycin and teicoplanin MICs originating h m microbroth dilution test v a d - and vanbencoded resistance in enterococci were detected most often correctly in comparison to the other tests checked (Tables 1 and 2). Within the group of VanC strains, a broad range of in-vitro resistance and susceptibility against vancomycin was determined with all antibiotic susceptibility methods by the participants (Table 2). The teicoplanin susceptibility of the VanC strains was determined correctly by nearly all participants (Tables 1 and 2). Some difficulties also arose in bacterial identification of the six test strains (with the exception of E. faeculis 1528; see Table 3). The majority of false identifications occurred with the VanC isolates: E. gallinarum UW 701/95 was correctly identified in only 36%, and E. casselijlavus UW 703/95 in 59%. Additionally, in both VanC strains, a broad spectrum of false species was determined (Table 3). However, a small number of laboratories also falsely identified E. faecalis UW 700/95 and the E,faecium strains (Table 3). DISCUSSION Glycopeptides such as vancomycin and teicoplanin are important 'reserve' antibiotics for treatment of infections with ampicillin-resistant enterococci or in cases of Table 1 Characterization of test strains with different genotypes of glycopeptide resistance used in the ring study Resistance phenotype: MIC (mg/l) and antibiotic susceptibility category (S, I, R)" for the glycopeptides Genotype of glycopepdde Species Strain no. Vancomycin Teicoplanin resistance E. faen'um 70/90 512-1024 (R) 128-256 (R) vana E. faecalis 1528 256-512 (R) 64-128 (R) V a d E. faecium UW 699/95 16-128 (R) 0.25-1 (S) vanb E. faecalis UW 700/95 32-256 (R) 0.25-1 (S) vanb E. gallinarum UW 701/95 8-16 (I-R) 0.25-1 (S) VanCj E. casselijlavus UW 703/95 4-8 (S-I) 0.25-1 (S) vancz " S, sensitive; I, intermediate; R, resistant.

Klare et al: Determination of glycopeptide resistance in enterococci 537 Table 2 Reports h m 62 clinical microbiology laboratories on classification of glycopeptide susceptibility for six enterococcal control strains tested Vancomycin Teicoplanin Agar Microbroth *gar Microbroth di fusion dilution Etest dihsion dilution Etest Strain Genome Data' S I R S I R S I R S I R S I R S I R E. faen'um uana 1 35 20 24 32 18 17 70190 2 35 2 0 24 32 18 17 3 (100) (100) (100) (100) (100) (100) E. faecufis uana 1 34 20 23 31 18 16 1528 2 34 2 0 23 3 5 2 3-3 1 5 2 3 1 1 3 (100) (100) (100) (10) (16) (74) - (17) (83) (13) (18) (69) E. faecium vanb 1 35 20 24 32 17 17 UW 699/95 2 1 6 5 14 1 2 1 7 2 6 1 6 2 8 3 1 1 7 1 7 3 (46) (14) (40) (5) (10) (85) (8) (25) (67) (88) (9) (3) (100) (100) E. faecufis vanb 1 35 20 24 32 17 17 UW 700/95 2 1 2 7 16 1-19 7 2 15 3 1 1-1 7 1 6-1 3 (34) (20) (46) (5) - (95) (29) (8) (63) (97) (3) - (100) (94) - (6) E. gallinarum vancf 1 35 20 24 32 17 17 UW 701/95 2 2 2 8 5 5 9 6 6 1 2 6 3 1 1-1 7 1 7 3 (63) (23) (14) (25) (45) (30) (25) (50) (25) (97) (3) - (100) (100) E. carsefijlavus vancr 1 35 20 24 32 17 17 UW 703/95 2 3 3 2-1 4 2 4 1 4 8 2 3 1 1-1 6 1-1 7 3 (94) (6) - (70) (10) (20) (59) (33) (8) (97) (3) - (94) (6) - (100) "Data: 1, total number of laboratories performing the test: 2, classification as sensitive, intermehate, resistant (&om left to right); 3, percentage (in parentheses). The correct classification groups are in bold type. Table 3 Identification of six enterococcal test strains by 58 clinical-microbiological laboratories from Germany and Austria E. E. E. E. E. E. E. E. Enterococcus Lactoroms Aerococcus Strain Genotype faerium feculis galfinarum cassefijlavus flauesceus durans auium porcinlrs spp. lactis mutans E. faecium vana 55 1 1 1 - - 70/90 (94) - (2) (2) - E. faecualis vana - 58 1528 - (100) - E. faecium vanb 53 1 2 1-1 - UW 699/95 (91) (2) (3) (2) - (2) - E. Jaecalis vufa 1 56 UW 700/95 (2) (96) - - E. gallinarum vancl 14 2 21 19 1-1 uw 701/95 (24) (3) (36) (33) (2) - (2) 1 - (2) (2) - The correct enterococcal species are indicated by bold type penicillin allergy. Enterococci of the VanA type are cross-resistant between vancomycin and teicoplanin. This type is the most common and clinically most important. In contrast, infections by VanB- and VanCtype strains can be treated with teicoplanin. Thus, susceptibility testing to vancomycin and teicoplanin should be performed. In this connection, the correct species identification can give valuable indications on the corresponding type of glycopeptide resistance: the VanC types are characterized by low-level vancomycin

538 Clinical Microbiology and Infection, Volume 5 Number 9, September 1999 resistance as a species property of E. gallinarum (VanCi), E. casseliflavus (Van&), and E. flavescens (vanc3). However, there also exist hgh-level vancornycin/ teicoplanin-resistant VanC strains [7]. Among the different antibiotic susceptibility methods used by the participants of this study, the agar difhsion test was the most common. However, in the German description for this test [ll], the agar diffusion test is not recommended for the susceptibility testing of glycopeptides because of its bad correlation between inlbitory zones and the corresponding MICs. Here, dilution tests (e.g. the microbroth dilution method [3,12]) are recommended. Nevertheless, it must be noted that the resistance to vancomycin in VanB strains can be expressed at very different levels; for example, the MICs can range between 4 and 1024 mg/l [5], but for the majority of strains they are 16-32 mg/l. Thus, VanB isolates and also (low-level resistant) VanC strains are not easy to detect. Difficulties with the agar diffusion test in determining vancomycin resistance in enterococci (especially of the VanB or VanC type) have also been described by other authors [8,9,13-161. The majority of these authors, however, reported no problems in cases of the high-level glycopeptide resistance type VanA. Concerning the Etest, there are different experiences reported in the literature: Endtz et al [13] could detect all enterococci correctly, including the VanB- and VanC-type GRE; in contrast, Schulz and Sahm [16] reported that the Etest must be interpreted with caution in these enterococcal species. It remains open whether the laboratories participating in our study correctly followed the manufacturer s recommendations for the Etest with regard to the bacterial inoculum and the time of incubation (at least 24 h). Furthermore, weak growth within the inhibition zones has to be considered for detection of VanB-type GRE (A. Bolmstrom, personal communication). To be totally sure, in critical cases (besides an exact enterococcal species determination), MICs should be determined by microbroth dilution, e.g. by using fiozen or lyophilized MIC rows for vancomycin and teicoplanin (e.g. by MIC rows from Feinchemie Sebnitz GmbH, Germany [17]), and additionally the PCR for the corresponding van gene(s) should be performed (e.g. by multiplex PCR [IS]). With the Vitek automated system and other methods (agar dilution, rapid microbroth dilution) that were rarely used in the present study (one or two laboratories used one of these three methods per test strain), we could not derive a classification. However, other authors observed a substantial influence of the growth medium used on the expression of the vancomycin resistance in the Vitek test (191; for example, these authors found that the brain-heart infusion broth used in this test can support good enterococcal growth but does not sufficiently support the expression of glycopeptide resistance among certain strains. From our point of view, the microbroth ddution test seems to be the most correct method in daily routine use for determining the exact type of enterococcal glycopeptide resistance at the present time. In this connection, the detection of enterococcal glycopeptide resistance genes by PCR can be a valuable confirmation. Enterococci of the E. gallinarum group (E. gallinarum, E. casseliflavus, E. flavescens) are all motile andwith the exception of E. gallinarum-also yellow pigmented [ZO]. However, tests for these characteristics are often not included in commercially available identification systems (e.g. API STREP). This fact also explains the differences in species determination of the VanC strains (e.g. as E. casseliflavus misidentified as E.&ecium strains). However, by testing these two reactions one can adequately discriminate ths group of enterococci from other enterococcal species. Thus, tests for motility and presence of the yellow pigment should be included in commercial identification systems. Acknowledgment We wish to thank all the participants of this multicenter study for their active cooperation and support. References 1. Martone WJ. Spread of vancomycin-resistant enterococci: why did it happen in the United States? Infect Control Hosp Epidemiol 1998; 19: 539-45. 2. Hare I, Witte W, Reinhardt A, Just H-M, EBinger U, Hoffler D. Ausbriiche von Infektionen mit uuna-positiven, high-level glycopeptidresistenten Enterocoafaecium in Deutschland [abstract 1541. In: Abstracts der 48. Jahrestagung der Deutschen Gesellschafi fur Hygiene und Mikrobiologie ev, Bonn, Deutsche Gesellschaft fir Hygiene und Mikrobiologie, 1996: 177. 3. National Committee for Clinical Laboratory Standards. Methods for ddution antimicrobial susceptibility tests for bacteria that grow aerobically. NCCLS document M7-A4, 4th edn. Approved standard. Wayne, PA NCCLS, 1997. 4. Deutsches Institut fir Normung ev. DIN 58940 Methoden zur Empfindlichkeits-bestimmung von mikrobiellen Krankheitserregern gegen Chemotherapeutika. Teil4: Bewertungsstufen der minimalen Hemmkonzentration. Berlin: Bed-Verlag, 1998. 5. Arthur M, Courvalin P. Genetics and mechanism of glycopeptide resistance in enterococci. Antimicrob Agents Chemother 1993; 37: 1563-71. 6. Perichon B, Reynolds P, Courvalin? VanD type glycopeptideresistant Enterococw faecium BM 4339. Antimicrob Agents Chemother 1997; 41: 201618. 7. Dub-Malen S, Blaimont B, Wauters G, Courvalin P. Emergence of high-level resistance to glycopeptides in Enterococcus gallinarum and Enteroroms casselijlauus. Antimicrob Agents Chemother 1994; 38: 1675-7. 8. Snell JJS, Brown DFJ, Perry SF, George R. Antimicrobial susceptibility testing of enterococci: results of a survey conducted by the United Kingdom National External Quality Assessment

Klare et al: Determination of glycopeptide resistance in enterococci 539 Scheme for Microbiology. J Antimicrob Chemother 1993; 32: 401-11. 9. Tenover FC, Tokars J, Swenson J, Paul S, Spitalny K, Jarvis W. Ability of clinical laboratories to detect antimicrobial agentresistant enterococci. J Clin Microbiol 1993; 31: 1695-9. 10. Hare I, Konstabel C, Bastrop R. Simple and rapid extraction of enterococcal DNA suitable for PCR of vancomycin resistance genes by use of ion exchanger Chelex 100 Resin. In Gobel U, ed. Methodische Entwicklungen in der mikrobiologischen Nukleinsaure-Diagnostik, 3./4. Poster-Workshop, December 1995/December 1996, Berlin, Hyg Mikrobiol (Abstract-Band). Reinbeck: Einhorn-Verlag, 1997: 33-4. 11. Deutsches Institut fur Normung ev. DIN 58940: Methoden zur Empfindlichkeitsprtifung von bakteriellen Krankheitserregern (auber Mykobakterien) gegen Chemotherapeutika. Teil 3: Agardiffusionstest. In: Deutsches Institut fur Normung ev, ed. DIN-Taschenbuch Medizinische Mikrobiologie und Immunologie-Normen und weitere Unterlagen. Berlin: Beuth-Verlag, 1992: 354-60. 12. Deutsches Institut fur Normung ev. DIN 58940: Methoden zur Empfindlichkeitspriihng von bakteriellen Krankheitserregern (auber Mykobakterien) gegen Chemotherapeutika. Teil 8: Mikrodilution. In: Deutsches Institut &r Normung ev, ed. DIN-Taschenbuch Memzinische Mikrobiologie und Immunologie-Normen und weitere Unterlagen. Berlin: Beuth-Verlag, 1992: 381-4. 13. Endtz H, van den Braak N, van Belkum A, et al. Comparison of eight methods to detect vancomycin resistance in enterococci. J Clin Microbiol 1998; 36: 592-4. 14. Rosenberg J, Tenover FC, Wong J, Jarvis W, Vugia DJ. Are clinical laboratories in Cahfornia accurately reporting vancomycinresistant enterococci? J Clin Microbiol 1997; 35: 2526-30. 15. Sahm DF, Olsen F. In-uifro detection of enterococcal vancomycin resistance. Antimicrob Agents Chemother 1990; 34: 1846-8. 16. Schulz JE, Sahm DE Reliability of the Etest for detection of ampicillin, vancomycin, and high-level aminoglycoside resistance in Enterococcus spp. J Clin Microbiol 1993; 31: 3336-9. 17. Klare I, Reissbrodt R. Glycopeptide-resistant enterococci (GRE). Biotest Bull 1998; 6: 5944. 18. Patel R, Uhl JR, Kohner P, Hopkins MK, Cockerill I11 FR. Multiplex PCR detection of uana, uanb, uunc-f, and vanc-2/3 genes in enterococci. J Clin Microbiol 1997; 35: 703-7. 19. Jett B, Free L, Sahm D: Factors iduencing the Vitek Gran- Positive Susceptibility System s detection of vanbencoded vancomycin resistance among enterococci. J Clin Microbiol 1996; 34: 701-6. 20. Devriese LA, Pot B, Collins MD. Phenotypic identification of the genus Enterococcus and differentiation of phylogenetically distinct enterococcal species and species groups. J Appl Bacteriol 1993; 75: 399408.