'Robert Koch Institute, Wernigerode Branch, Wernigerode, and 2Rhone Poulenc Rorer GmbH, Cologne, Germany

ORIGINAL ARTICLE Increasing incidence and widespread dissemination of methicillin-resistant Staphylococcus aureus (MRSA) in hospitals in central Europe, with special reference to German hospitals Wo&ang Witte ', Michael Kresken', Christine Braulke' and Christa Cuny ' 'Robert Koch Institute, Wernigerode Branch, Wernigerode, and 2Rhone Poulenc Rorer GmbH, Cologne, Germany Objective: to present data on prevalence and interregional spread of methicillin-resistant Staphylococcus aureus (MRSA) in Germany. Methods: A nationwide collection of MRSA isolates from nosocomial infections in 143 hospitals was established from isolates (n=4368) sent to a microbiological reference center during 1993-95. As chosen by distinguishable resistance phenotypes at each time of occurrence during the study period, 1830 isolates were subjected to molecular typing by means of Smal macrorestriction patterns, PCR for RNA gene spacer patterns, and PCR for patterns of DNA stretches flanked by the ERIC-2 sequence and flanked by Tn916 and ribosomal binding site. In addition, data from a multicenter study on the incidence of antibiotic resistance have been analyzed (32 centers, 637 S. aureus isolates). Results: In 1995 the prevalence of MRSA among S. aureus isolates was 8.7% overall in central Europe (including Germany), in comparison to 1.7% in 1990. From 1993 until now, a continuous interregional dissemination of six epidemic strains, which were identified by molecular typing, was recorded. Besides these epidemic strains, 15 MRSA strains were identified which could not be allocated to the epidemic MRSA or to the known clonal groups of the species S. aureus. MRSA from three cases of sporadic nosocomial infections exhibited characteristics of the clonal group of S. aureus with the capacity for toxic shock syndrome formation. The pattern of one MRSA corresponded to those of the S. aureus group exhibiting phage pattern 94,96. Conclusions: The prevalence of MRSA has increased in central Europe (and Germany) during the last 5 years, to 8.7%. The main source of infection with MRSA is obviously interregional dissemination of epidemic strains. At the same time, the meca gene has been acquired by strains previously sensitive to methicillin. Key words: methicillin-resistant Staphylococcus aureus, molecular epidemiology, nationwide spread INTRODUCTION Methicillin-resistant Staphylococcus aureus (MRSA) presents a major problem in hospital infection and infection control because of resistance to multiple antibiotics. Furthermore, there are particular strains with an evident capacity for spread once introduced Corresponding author and reprint requests: Wolfgang Witte, Robert Koch Institute, Wernigerode Branch, Burgstraae 37, G-38855 Wernigerode, Germany Tel: 49 3943 6790 Fax: 49 3943 679207 Accepted 25 April 1997 into a hospital. As already indicated by conventional typing, interregional dissemination has been observed with distinct MRSA strains (EMRSA) in British hospitals [l]. Although a multicenter study in 1990 revealed a relatively low incidence of MRSA in German hospitals [2], we have previously observed interregional clonal dmemination of two clearly distinguishable strains of MRSA in Germany [3], one mainly in the south and the other mainly in the north. The related isolates were identified by SmaI macrorestriction patterns. At the beginning of 1994, six different MRSA strains were observed [4]. This paper describes the current prevalence of MRSA among S. aureus isolates in central Europe (and Germany), 414

Witte et al: Prevalence and spread of MRSA among German hospitals 415 derived from the results of a surveillance study conducted in late 1995 by the Bacterial Resistance Study Group of the Paul Ehrlich Society for Chemotherapy. In addition, we have analyzed the interregional dissemination of MRSA in Germany on the basis of genomic typing. MATERIALS AND METHODS Surveillance study In November 1995, the Bacterial Resistance Study Group of the Paul Ehrlich Society for Chemotherapy conducted a surveillance study of the prevalence of antimicrobial resistance in central Europe. The study included 32 laboratories located in Austria (n=4), Germany (~2~21) and Switzerland (n=7). Each center was requested to collect 40 S. aureus isolates during October 1995, if available. No duplicate strains were allowed. All participating laboratories used the same standardized methods. Methods for identification of the S. aureus strains are described elsewhere [5]. Susceptibility testing, using microdilution, was performed according to the guidelines recommended by the National Committee of Clinical Laboratory Standards [6], with the exception that screening for S. aureus resistant to oxacillin was performed using Mueller- Hinton agar plates supplemented with 4% NaCl and 6 mg/l oxacillin. Bacterial isolates (n=4368) MRSA from sporadic infections and from outbreaks in hospitals were sent to a national reference center for staphylococci. Strains were defined as MRSA by MIC 24 mg/l for oxacillin and PCR demonstration of the meca gene [4]. Of these isolates, 1830 were subjected to molecular typing. If in one of the affected hospitals (n=143) several isolates were obtained at a given point of time during the study period, molecular typing was performed for each distinguishable resistance phenotype. S. aureus PS29 is the propagating strain for phage 29 of the international basic set for phage typing staphylococci. Genomic DNA fragment patterns SmaI digestion and subsequent pulsed-field gel electrophoresis were performed as previously described [3]. PCR typing The protocols previously described for rrna gene spacer patterns [7], for the length polymorphism of DNA sequences flanked by the Tn916 target region and ribosomal binding site (tar 916-shida PCR [S]), and for the length polymorphism of ERIC-2-flanked sequences [9], were followed. Documentation of macrorestriction patterns and similarity analysis Agarose gels were documented and image processed by the MWG-Biotech restriction fragment length polymorphism scan system. Cluster analysis was derived from patterns of molecular masses of DNA fragments for each of the investigated strains which were stored in the databank system. This database is implemented in MS-Access version 2.0. The cluster analysis is based on an optimizing similarity algorithm [lo]. Current prevalence of MRSA in central Europe (and Germany) The surveillance study conducted in November 1995 by the Bacterial Resistance Study Group of the Paul Ehrlich Society for Chemotherapy included 962 S. aureus strains. Based on the results of the MICs of oxacillin (resistant, MIC 24 mg/l) the overall prevalence of MRSA in central Europe was 8.7% (Figure 1). Of the 637 S. aureus strains collected at the 21 German laboratories participating in the study, 8.0% were MRSA. The agar screening test revealed slightly lower resistance rates, with 8.0% and 7.4% of strains found to be MRSA in central Europe and Germany, respectively. The study also investigated differences in the prevalence of MRSA between patient groups. Of 158 isolates obtained from outpatients, 3.2% were MRSA; thus the predominant source of MRSA is the hospital. Furthermore, the prevalence of MRSA in intensive care units (13.5% of 148 isolates) is higher than that in other hospital units (9.0% of 654 isolates). Dissemination of MRSA Independently of the prevalence studies described above, after 1991 a network for surveillance of dissemination of MRSA in Germany was established based on continuous typing of MRSA from sporadic infections or from nosocomial outbreaks, which were isolated in designated local diagnostic laboratories (1 05 throughout Germany). Epidemic MRSA strains were named according to the geographic area of first recording, and include the northern German MRSA, the southern German MRSA, the Hannover area MRSA, the Berlin MRSA, the south eastern German western Austrian MRSA and the Vienna MRSA. For these strains, the possession of the meca gene was evidenced by PCR. The resistance phenotypes of these strains are shown in Table 1. Molecular typing of MRSA was based on SmaI macrorestriction patterns which were image processed, stored in the databank and analyzed for similarity. Isolates exhibiting more than four different bands were regarded as different. A difference of four or more bands

~ 416 Clinical Microbiology and Infection, Volume 3 Number 4, August 1997 corresponds to inore than one genetic event (mutation or rearrangements; for details see Tenover et a1 [l 11). The assignation of an isolate to one of the epidemic strains was confirmed by means of PCR for different and independent DNA polymorphisms. That the six epidemic strains represent different strains has already been shown by SmaI macrorestriction patterns [4] and has been confirmed by tar 916-shida patterns (Figure 2). The probable relatedness of the northern German MRSA and the Hannover area MRSA, the south eastern Germadwestern Austrian MRSA had been suggested previously by identical rrna gene spacer patterns [4], and this is confirmed by ERIC-2 patterns (Figure 2). Table 1 Resistance phenotypes of epidemic MRSA MRSA denomination Northern German MRSA Southern German MKSA Hannover area MRSA South eastern Germadwestern Austrian MRSA Vienna MRSA Berlin MRSA Antibiotics to which strain is resistance PEN, OA, GEN, ERY, CLI, OTE. ST, RAM (FUS in 4 of 16 outbreak?) PEN, OA, GEN, ERY, CLI, CIP, ST PEN, OA, GEN, ERY, CLI, OTE, CIP, ST PEN, OA, GEN, ERY, CLI, OTE, CIP, ST PEN, OA, GEN, ERY, CLI, OTE, CIP, ST PEN, OA, CIP (3 outbreaks) PEN, OA, GEN, ERY, CLI, CIP (6 outbreaks) PEN, OA, GEN, ST (9 outbreaks) PEN, OA, ERY, CLI, CIP (6 outbreaks) Abbreviations: PEN=petiicillin; OA=oxacillm; C:EN=gentanicin; ERY =erythrornycln, CLI=clindaniycin; OTE=oxytetracycIine. CIP=ciprofloxacin, ST=sulfonaniide-trimcthoprim, M=rifampicin, FUS= fusidic acid. Figure 1 Prevalence of MRSA among Staphylococcus aweux isolates in central Europe (and Germany) from 1976 to 1995. The stuhes performed between 1975 and 1984, and between 1990 and 1995, were conducted in cooperation with laboratories in Austria, Germany and Switzerland by the Bacterial Resistance Study Group of the Paul Ehrlich Society for Chemotherapy, and those performed in 1989 and 1992 were conducted with German hospitals by the study group of Knothe et a1 1121.

Witte et al: Prevalence and spread of MRSA among German hospitals 417 Figure 2 SmaI-macrorestriction patterns and PCR pattern of epidemic MRSA: A Berlin epidemic MRSA; *southern Gcrman epidemic MRSA; 0 northern German epidemic MRSA; 0 Hannover area epidemic MKSA; C3 southeastern Germadwestern Austrian epidemic MRSA; Vienna epidemic MKSA; S=S. a~reux 8325 as reference.

418 Clinical Microbiology and Infection, Volume 3 Number 4, August 1997 Figure 3 Interregional dissemination of multiply resistant and oxacillin-resistant S. aureus (MRSA) in Germany in 1992 and 1993: Onorthern German epidemic strain; *southern German epidemic strain; A Berlin epidemic strain; 0 Hannover area epidemic strain; 0 southeastern Germadwestern Austrian epidenlic strain; I Vienna epidemic strain. Figure 4 Interregional dmemination of niultiply resistant and oxacillin-resistant S. aureus (MRSA) in Germany in 1995 and 1996. ABerlin epidemic MRSA; *southern German epidenlic MRSA; 0 northern German epidemic MRSA; 0 Hannover area epidemic MRSA; 0 southeastern Gernxdwestern Austrian epidemic MRSA; Vienna epidemic MRSA. Interregional spread of epidemic MRSA Figure 3 shows the distribution of occurrence of the northern German MRSA and the southern German MRSA by the end of 1993. At the end of 1995 (Figure 4), the southern German MRSA had been found also in hospitals in the north of Germany and in Thuringia, and the northern German MRSA had been disseminated to the east. The Hannover area MRSA, which was observed in hospitals of Hannover and its vicinity in 1992, was also seen in hospitals in the western half of Germany. Of special interest was the demonstration of the spread of the Berlin MRSA, which at the end of 1993 was present in six Berlin hospitals, but by the end of 1995 was present in 17 Berlin hospitals, five hospitals in Brandenburg county, one hospital in Lower Saxony county and one hospital in Nordrhein-Westfalen county. The Vienna MRSA, which was isolated in nosocomial outbreaks in a number of Viennese hospitals in 1991-92, was seen at the end of 1995 in outbreaks ofinfection in six German hospitals. MRSA different from epidemic strains Besides outbreaks of infections by interregionally disseminated epidemic strains, MRSA with SmuI restriction patterns which were different from those of the epidemic strains by more than four fragments were observed in connection with infection in 15 German hospitals. SmuI macrorestriction patterns of these isolates (Figure 5) were compared to those of the epidemic strains described above and of reference strains of the major clonal groups within the species S. aureus by means of similarity analysis. The corresponding dendrogram is shown in Figure 6. This dendrogram also includes the patterns of reference strains for the major clonal groups within the species S. aureus as grouped by phage-typing patterns for historical reasons. According to this dendrogram, nine of these MRSA strains belong to the cluster of phage group 111 strains, which also includes the northern German MKSA, the Hannover area epidemic MRSA, the south eastern Germadwestern Austrian MRSA and the Vienna MRSA. The pattern of one strain was similar to those

Witte et al: Prevalence and spread of MRSA among German hospitals 419 of strains with phage pattern 94,96; and that of three strains resembled those of the clonal group of S. aimus with the capacity for toxic shock syndrome toxin formation. That strains falling into the phage group I11 cluster are probably related to the MRSA within this cluster is also indicated by rkna gene spacer patterns, which were identical or similar to those of these strains (Figure 5). The MRSA belonging to the cluster of phage pattern 29 strains exhibited the same rrgs patterns as reference strain PS 29 (Figure 5). Smal macrorestriction pattern S=8325 as standard rrna gene spacer pattern S=mol. mass standard Figure 5 SrtzaI nlacrorestrlctlon and rrna gene spacer patterns of MKSA whlch are different from epidenuc MRSA.

420 Clinical Microbiology and Infection, Volume 3 Number 4, August 1997 111 t 0 0 0 207194 206194 462193 A 994193 1000193 19 t I I 5 I I V A II IV t * 354196 353196 537196 4 251194 306194 1150/93 20 17 18 34 32 1446195 33 37 3 t 35 1886/95-1 538195 0.0 % 35.6 % Heterogeneity Figure 6 Relatedness of SmaI macrorestriction patterns of MRSA which are different from those of epidemic strains:, MRSA hfferent from epidemic strains; f, reference strains for clonal groups within the species S. aureus, (for symbols of epidemic MRSA, see Figure 2); 1, group of S. aureus exhibiting phage pattern 29, 52, ; 11, group of S. aureus exhibiting phage group I1 patterns; 111, group of S. aureus exhibiting phage group 111 patterns; IV, group of S. aureus exhibiting phage pattern 94, 96; V, group of S. aweus exhibiting phage pattern 95. DISCUSSION In this surveillance study conducted in late 1995 by the Bacterial Resistance Study Group of the Paul Ehrlich Society for Chemotherapy, the prevalence of MRSA among S. uureus isolates was 8.7% in central Europe (and Germany). This rate was substantially higher than that recorded for MRSA in the previous surveillance study conducted in 1990 (1.7%) [2] and those found in two German multicenter studies in 1989 (4.6%) and 1992 (6.0%) by another study group [12]. Reviewing the data of all surveillance studies performed since 1976

Witte et al: Prevalence and spread oi MRSA among German hospitals 421 by the Bacterial Kesistance Study Group, after 1984 the prevalence of MRSA has increased from 1-2%) to 8.7% in central Europe (and Germany), although the prevalence in 1990 was lower (Figure 1). The rate of 8.7% MRSA, however, is still relatively low compared to the much higher resistance rates observed in southern Europe, where the proportion of MRSA has already exceeded 30% in some countries [2,13,14]. In recent studies, the proportion of MRSA among S. au~cus from nosoconiial infections ranges from lower than 1%) in Denmark [I 51 and in The Netherlands [16] to 30% and even higher in Spain, Italy and France 19,131. In the USA, the average frequency is 29% 1171. In parallel with the increasing incidence of MRSA up to 1995 in Germany, continuous interregional dissemination of epidemic MKSA was observed. The different strains exhibited characteristic patterns of DNA polyniorphisms with regard to.!ha1 fragments, rrna gene spacers, DNA stretches flanked by ERIC- 2 sequence or transposon 916 targ:et region and ribosomal binding site. MRSA strains from 15 hospitals were clearly different from the epidemic strains by SmaI fragment patterns. Their allocation to distinct clonal groups of the species S. au~eus was confirmed by rrna gene spacer patterns (for rrna gene spacer patterns of representative strains of S. auyeus see Cuny et a1 [71). Of special interest was the occurrence of inethicillin resistance in groups of S. aureus which are also disseminated among the community and up to now have been sensitive to methicillin, such as those with phage pattern 29, phage pattern 94,96 and phage pattern 95. Recent data from the USA [l8,19] and from Canada [20] suggest that circulation of MKSA in the community could be another source of interregional spread. As is evident from data on the origin of MRSA strains sent for typing to the authors laboratory in 1995 (in which only three among 2500 isolates came from patients who had not been admitted to a hospital or nursing facility) and from a study in the Saarland federal county with the demonstration of five MRSA in nasal swabs from 1000 non-hospitalized patients (P. Rohr and W. Witte, unpublished), MRSA strains are obviously still rare among staphylococcal carriers outside hospitals in Germany. Nevertheless, this aspect requires further attention, especially with regard to the Berlin EMRSA, which has most probably originated from the clonal group of S. aweus with phage pattern 95 [4]. Recent data on the incidence of MRSA in Danish hospitals indicate that inter- and intrahospital spread of MRSA can, in principle, be prevented suces:ifully [14]. The results from Germany described in this communication suggest that besides the control measures for preventing the dissemination of MRSA (for details see Boyce et a1 [21]), iniprovenient in the information given to destination hospitals before transfer of patients with MRSA is necessary. Acknowledgments We gratefully acknowledge the cooperation of local centers participating in the multicenter study of the Paul Ehrlich Society on prevalence of bacterial resistance, and we thank Mrs B. Pasemann, P. Vielbrandt and E. Baier for technical assistance. References 1. Marples RR, Cooke EM. Current problems with methicillin resistant Staphyloroccus aweus. J Hosp Infect 1988; 11: 381-92. 2. Kresken M. Pravalenz der Resistenz bei klinisch wichtigen Bakterieiispecies gegeniiber alteren und neueren Antibiotika in Europa. Bundesgesundhbl 1995; 38: 170-8. 3. Witte W, Cuny C, Braulke C, Heuck D. Clonal dissemination of two MKSA strains in Germany. Epidemiol Infect 1993; 113: 67-73. 4. Witte W, Cuny C, Halle E, Wagner J. Methicillin resistance in an epidemic Staphylococcus fliircw strain with genomic fingerprints corresponding to thox of a rensitive strain in the community. Med Microbiol Lett 1994; 3: 388-395. 5. Kresken M, Wiedemann B. Die Epidcniiologie der I<es~stenz bei Bakterien und ihre Bedeutung fur die Wirksamkeit von Chemotherapeutika. Fortschr Antimikrob Antincoplast Chernother 1987; 6 (6): 869-1063. 6. National Committee for Clinical Laboratory Standards. Performance standards for antimicrobial susceptibility testing. NCLS document MI OO-Sb/M7-A3. Methods for dilution antimicrobial susceptibility tests for bacteri.1 that grow aerobically, 3rd edn. Approved stmdard, 1995. 7. Cuny C, Claus H, Witte W. Di~rimination of S. m ~ m s strains by PCR for r-rna gene spacer size polymorphism and comparison to SmaI-niacrorestrictioii patterns. Zentralbl Bakteriol 1996; 283: 466-76. 8. Cuny C, Witte W. Typing of Staphylococcus uiww by PCR for DNA sequences flanked by transposon Tn 916 target region and ribosomal binding site. J Clin Microbiol 1906; 34: 1502-5. 9 Van Brlkum A, Bax R, Prevost G. Comparison of four genotyping assays for epidemiological study of iiiethicilli~iresistant Staphylococcus aureiis. Eur J Clin Microbiol Infect 1 )is 1994; 13: 420-4. 10 Claus H, Cuny C, Pasemann B, Witte W. Ein Datenbanksystem fur Fragment-muster der genomischen IINS von Staphyhiroctr~s nureus. Bundesgesundhbl 1996; 39: 68-74. 11 Tenover F, Arbeit I), Goering RV, et al. Interpreting chroniosonial DNA restriction patterns produced by pulsedfield gel electrophoresis: criteria for bacterial strain typing. J Clin Microbiol 1995; 33: 2233-0. 12 Knothe H, Hapken E, Machka K. In vitro activity of antibiotics-results of two German niulticentre in vitro studies in 1989 and 1992 [abstract 801. Poster presented at

422 Clinical Microbiology and Infection, Volume 3 Number 4, August 1997 the 6th International Congress for Infectious Diseases, Prague, 26-30 April 1994: 176. 13. Voss A, Milatovic D, Wallrauch-Schwarz C, Rosdahl VT, Braveny I. Methicillin-resistant Staphylococcus uureus in Europe. Eur J Clin Microbiol Infect Dis 1994; 13: 50-5. 14. Struelens M, Mertens R. National survey of methicillinresistant Staphylococcus uureiis in Belgian hospitals: detection methods, prevalence trends and infection control measures. Eur J Clin Microbiol Infect lh 1994; 13: 56-63. 15. Rosdahl VT, Esperson J, Frimodt-Moller N, Skinkoj F! Changing Staphylococcus aureus epidemiology: 30 years experience. Zentralbl Bakteriol Suppl 1994; 26: 3-8. 16. Frenay HMF, Peerbooms PGH, Van Leeuwen VJ et al. Nursing home: not a source of niethicifin resistant Stuphylo- Coccux aurem in the Netherlands. Clin Infect Dis 1994; 19: 213-14. 17. Panlilio AL, Culver DH, Gaynes RF! Methicillin resistant Staphylococcus uureus in US hospitals 1975-1991. Infect Control Hosp Epidemiol 1992; 13: 582-6. 18. Layton MC, Hierholzer W Jr, Patterson JE. The evolving epidemiology of MRSA at a university hospital. Infect Control Hosp Epidemiol 1995; 16: 12-17. 19. Morena F, Crisp C, Jorgenson JH, Patterson JE. Methicillin resistant Staphylococcus aureus as a community organism. Clin Infect Dis 1995; 21: 1308-12. 20. Embil J, Ramotar K, Romance L, et al. MRSA in tertiary care institutions on the Canadian prairies 1990-1992. Infect Control Hosp Epidemiol 1994; 15: 646-651. 21. Boyce JM, Jackson MM, Pugliese G. et al. Methicillinresistant Staphylococcus uureus (MRSA): a briefing for acute care hospitals and nursing facilities. Infect Control Epidemiol 1994; 15: 105-15.