Epidemiology of methicillin-resistant Staphylococcus aureus : results of a nation-wide survey in Switzerland

a Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland b Hôpitaux Universitaires de Genève, Switzerland c UniversitätsSpital, Zürich, Switzerland d UniversitätsSpital, Basel, Switzerland e Institut für Medizinische Mikrobiologie, Bern, Switzerland f Institut für Mikrobiologie, Zürich, Switzerland g Istituto Cantonale Batteriosierologico, Lugano, Switzerland h Ospedale Regionale di Lugano, Lugano, Switzerland

Methicillin-resistant Staphylococcus aureus (MRSA) is often resistant to most other antibiotics, frequently causes nosocomial infections, and is one of the greatest challenges for modern antimicrobial therapy, particularly since the emergence of S. aureus with intermediate susceptibility to gly-copeptides [1].In the early 1990s, a European survey indicated that MRSA tended to be more frequent in southern than in northern Europe.In countries such as France, Spain, and Italy more than 30% of S. aureus isolates were resistant to methicillin [2].This was further confirmed in a more Objective: To assess the epidemiology of methicillin-resistant Staphylococcus aureus (MRSA) in Switzerland.
Material and methods: One-year national survey of all MRSA cases detected in a large sample of Swiss healthcare institutions (HCI).Analysis of epidemiological and molecular typing data (PFGE) of MRSA strains.
Results: During 1997, 385 cases of MRSA were recorded in the 5 university hospitals, in 33 acute care community hospitals, and 14 rehabilitation or long-term care institutions.Half of the cases were found at the University of Geneva Hospitals where MRSA was already known to be endemic (41.1 cases/10,000 admissions).The remaining cases (200) were distributed throughout Switzerland.The highest rates (>100 cases/10,000 admissions) were reported from non-acute care institutions.Rates ranged from 3.3 to 41.1 cases/10,000 admissions for university hospitals (mean 15.5); 0.67 to 90.4 for community hospitals (mean 4.8), and 28.2 to 315 for non-acute care institutions reporting MRSA (mean 85.7).Forty percent of MRSA patients were infected, while 60% were only colonised.The leading infection sites were skin and soft tissue (21%), surgical site (15%), and the urinary tract (26%).Whereas in Eastern Swiss HCI most MRSA cases occurred in acute care hospitals (n = 47, 98%), rehabilitation and long-term care institutions accounted for an important number of the identified cases (n = 107, 38%) in Western Switzerland.
Conclusion: Low rates of MRSA were still observed in Swiss HCI, despite one outlying acute care centre with endemic MRSA and some nonacute care institutions with epidemic MRSA.Rehabilitation and long-term care institutions contributed to a substantial proportion of cases in Western Switzerland and may constitute a significant reservoir.Overall, a national approach to surveillance and control of MRSA is mandatory in order to preserve a still favourable situation, and to decrease the risk of epidemic MRSA dissemination.

Introduction
recent survey [3].However, within a country, MRSA prevalence may vary substantially from one hospital to another [4].
Although the reported proportion of MRSA among all S. aureus isolates in countries surrounding Switzerland is high, low rates (0 to 4%) have been reported in Swiss hospitals [5,6], with the exception of the University of Geneva Hospitals (HUG) (20%) [7].The objective of the present study was to establish a comprehensive picture of the epidemiology of MRSA in Switzerland by performing an extensive one-year survey.

General setting and organisation of the survey
In Switzerland, there are 5 university hospitals for a country of approximately 7 million inhabitants.In 1997, these centres organised a national survey of MRSA and collected data and isolates not only from their own hospital but, also, from other Health Care Institutions (HCI) in their region that agreed to participate.In 1997, 445 HCI were members of the H+ organisation (H+ The Swiss Hospitals, Aarau, Switzerland), which represent 80-90% of all Swiss HCI, without the nursing homes.

MRSA detection and control measures at the hospital level
Surveillance screening and control measures used in the participating institutions were standardised for the purpose of the study [6].In brief, patients with MRSA were identified by surveillance of microbiological laboratory data from clinical specimens as well as by cultures obtained from roommates of patients infected or colonised with MRSA.When a cluster was suspected, patients of the concerned ward were screened.In addition, surveillance cultures were performed on readmission of patients previously known to have been positive for MRSA and on patients transferred from foreign hospitals.At HUG, an automatic alert system allowed immediate recognition and admission screening of patients previously know as MRSA carriers [8].Cultured sites included anterior nares, perineal (or perirectal), and any infected sites such as open wounds.
A standard isolation procedure was applied to all patients positive for MRSA.The decolonisation procedure consisted of the application of nasal mupirocin (Bactroban ® ) twice daily for 5 days, associated with a daily body wash with a chlorhexidine-containing soap for 7 to 10 days.Isolation was stopped when decolonisation was documented by two sets of negative surveillance cultures obtained at least 48 h apart.Patients with infections received appropriate antibiotic therapy.

Bacterial isolates
MRSA isolates were collected by the laboratories of the university centres.Identification of S. aureus was confirmed by standard methods and susceptibility testing was performed by disk diffusion on Mueller-Hinton agar with 24-h incubation at 35 °C.Interpretation criteria were those of the National Committee for Clinical Laboratory Standards (NCCLS) [9].Resistance to oxacillin was confirmed by the screen agar test [10] and on random isolates by amplification of the mecA gene.

Epidemiological data and definitions
A case was defined as a hospitalised patient found to be infected or colonised with MRSA during the study period.For each case, the following characteristics and risk factors for MRSA colonisation/infection were recorded: demographic data, geographic origin, admission and discharge dates, prior hospitalisation during the preceding 3 years, past history of MRSA colonisation or infection, presence of comorbidities (Charlson score [11]), functional status (Karnofsky index [12]), underlying diagnosis, infection versus colonisation with MRSA, presence of indwelling urinary or vascular catheter, wound, drains, endotracheal tube, antibiotic treatment and operation.
Cases considered as having acquired MRSA before admission (pre-hospitalisation acquisition) were: (1) patients known to have been colonised/infected with MRSA during a prior hospitalisation; (2) cases with MRSA detected within 72 h of admission; and (3) cases where the isolate exhibited a PFGE pattern observed only once (unique) in a given hospital (thus presumed to have been introduced by the patient).Other cases were considered as newly hospital-acquired MRSA.
Standard definitions for nosocomial infections of the Centers for Diseases Control and Prevention (CDC) were used [13].A case was considered infected by MRSA (as opposed to colonised) when MRSA was isolated from a sterile site, or when signs of infection motivated systemic antibiotic therapy and/or surgical procedure.

General data
From January to December 1997, 385 cases of MRSA were identified and reported from the 5 university hospitals, 33 other acute care hospitals, and 14 rehabilitation or long-term care institutions throughout the country (table 2, figure 1).Ten additional HCI reported not to have observed any MRSA cases.Both epidemiological data and isolates were available for 225 (63%) of the cases, whereas only epidemiological data or isolates were available for 97 and 63 cases, respectively.Demographic and clinical characteristics of the MRSA cases are shown in table 1. Variations in the origin of the patients between institutions were observed (table 2).Although 52% of the patients were admitted from their home 92% had been hospitalised in the previous 3 years.Transfer from another hospital (within or outside Switzerland) was the next most frequent origin of cases.

MRSA rates
The proportions of MRSA among all S. aureus isolates were 4, 3, 23, 3, and 6% in Basel, Bern, Geneva, Lausanne, and Zurich university hospitals, respectively.The highest number of MRSA cases/10,000 admissions in a university hospital was found at HUG (41.1), whereas it was below 10/10,000 admissions in the other university hospitals (table 2).The mean rate was 4.8 cases/10,000 admissions in acute community hospitals (range 0.67 to 90.4) and 85.7/10,000 admissions in rehabilitation and long-term care institutions that reported MRSA cases (range 28.2 to 315).The incidence-density of all MRSA cases averaged 0.93/10,000 patient-days, and varied from 0.2 to 2.1 in university hospitals (mean 1.2), from 0.06 to 3.0 (mean 0.47) in other acute care hospitals, and from 0.6 to 13.5 (mean, 2.2) in rehabilitation or long-term care institutions.A. Geographical distribution of MRSA cases in Switzerland, 1997.Each circle represents one institution, the number in the circle is the number of MRSA cases reported for this institution.

University hospitals
Other 1 during the previous 3 years. 2for cases for whom data was known.n.a., not available

Table 2
Incidence and demographic characteristics of MRSA cases.

Hospitalisation ward
Table 3 shows the wards on which patients were hospitalised at time of detection.Most cases (72%) occurred in medical or surgical units.MRSA observed in intensive care units (ICUs) represented less than 8% of cases in Western Swiss hospitals whereas, in Eastern Switzerland, they accounted for 21%.Of note, among the 20 cases which occurred in Swiss ICUs, 9 (45%) were transferred from foreign hospitals.

Type of institution
The distribution of cases according to the type of institution (acute hospitals of various sizes, rehabilitation, and long-term care facilities) is shown in table 4. In Western Swiss HCI, between 24 and 39% of cases were observed in non-acute care institutions (rehabilitation and long-term care facilities) whereas this proportion was below 2% in Eastern and Southern Switzerland.Analysis of molecular data showed that 86% (48/56) of cases in non-acute care wards of Western Switzerland were due to an epidemic clone.By comparison, this clone accounted for only 40% (19/47) of the cases observed in acute care hospitals.

Prior versus new acquisition of MRSA
For the university hospitals, the epidemiological data permitted determination of whether MRSA had most likely been acquired before (= prior acquisition) or during the hospitalisation considered in the study (= new acquisition) (table 5).Overall, 86/309 (28%) cases were considered to have been newly-acquired MRSA.The rates varied from 8.3 to 29% between the different university hospitals.Based on this data, the attack rate of MRSA, defined as the number of newly-acquired MRSA while admitted to the different university hospitals, averaged 2.8 ranging from 0.3 to 8.1 cases/10,000 admissions.
Among the 144 cases with prior acquisition of MRSA and for which data were available, 94 (65%) came from home, 36 (25%) were transferred from Swiss (7%) or foreign (18%) hospitals, and 9 (6%) from non-acute care institutions.Among patients admitted directly from home, 92% of cases had been hospitalised previously (table 2).

MRSA infections
Overall, 116/291 (40%) of cases were considered as having clinical infections with a total of 138 infections.Infection rates ranged from 0.  (table 2).Data were not available for most other HCI.The leading infection sites were the urinary tract (25%), skin and soft tissue (20%) and the surgical site (15%) (figure 2).No significant difference in the distribution of the sites of infection was observed between the different hospitals.

Molecular epidemiology
Molecular typing (pulsed field gel electrophoresis) was performed on isolates of 288/385 of the cases (one isolate per case).Detailed results have been published elsewhere [25].Results showed that 65% of the isolates belonged to four predominant clones, three of which were mostly present in Geneva hospitals (figures 3 and 4).The remaining 35% of the isolates were clustered into 67 genotypes and they accounted for 1 to 5 patients per hospital (figure 3).This study is the first nationwide survey on MRSA carried out in Switzerland.The results confirm that Swiss HCI have a low prevalence of MRSA, with the exception of the HUG and some non-acute care institutions in the western part of the country [8,14].
Although a substantial number of institutions participated in the study, our results are derived from a convenience sample of hospitals that volunteered to participate, and cannot be extrapolated to the whole of Switzerland for two reasons.First, university hospitals which have all participated in the study, represent a non-negligible bias since risk factors for MRSA are much higher in this type of institution.Second, although we received data from 10 community hospitals reporting no MRSA during the period of the survey, we had no data from the other Swiss healthcare centres.However, regarding the relatively low number of MRSA cases recovered in most of the reporting institutions and taking into account the tight network among Swiss infection control practitioners, it is unlikely that non-reporting institutions have experienced a substantial MRSA burden.
Different methods of calculation can be used to measure the magnitude of MRSA occurrence at a particular institution.The least informative figure, but also the most frequently reported as it is the easiest to obtain and relies only on microbiology laboratory data, is the proportion of MRSA among all S. aureus isolates in an institution.This proportion allows a rough comparison between hospitals, although it also requires precise definitions (eg, one isolate per patient).Thus, the HUG, with 23%, have a rate comparable to hospitals of countries surrounding the southern part of Switzerland [2,[15][16][17].The other Swiss institutions showed proportions below 5%, figures which are comparable to northern European hospitals such as those in Germany [2], the Netherlands [2,18,19], Denmark [2] or in Sweden [2,20].In Germany, an increase of the proportion of S. aureus strains resistant to methicillin from 1.7% to 12.9% has been reported between 1990 and 1995 [21].
The number of infected or colonised MRSA cases/10,000 admissions, or the number of MRSA infections/10,000 admissions provide more appropriate figures for comparison between institutions or for longitudinal surveillance.With 0.4 to 2.7 MRSA infections/10,000 admissions, our study shows that university hospitals in Switzerland currently have a low incidence of MRSA, with the exception of HUG where a rate of 13.8 infections/10,000 admissions was reported.By comparison, MRSA-infected patients/10,000 admissions in 43 French hospitals were recently estimated to range between 18 to 158 [22].It was estimated at 165/10,000 admissions in Irish hospitals [23].The incidence density (number of cases per patientday) may be the best indicator for comparison be-tween institutions.Indeed, in the present study, the incidence seems quite high in long-term care institutions when using the number of cases/10'000 admissions, but is much more comparable to university hospitals when looking at the incidence density (patients stay longer in long-term care hospitals, therefore there are fewer admissions for more patient-days of stay).
The study also shows that the epidemiology of MRSA in Switzerland differs from one region to another.The highest reported incidences of MRSA cases (>100 cases/10,000 admissions) was reported from non-acute care institutions (rehabilitation and long-term care), in particular in the western part of the country.Importantly, whereas the high number of cases of Western Switzerland were reported from the non-acute institutions, these institutions did not seem to play a major role in Eastern and Southern Switzerland.A recruitment bias between Western and Eastern Switzerland can, nevertheless, not be excluded.As the majority of the cases in non-acute care institutions were due to an epidemic clone [24], these institutions probably played an important role in the dissemination of the clone, as transfers of patients from and to acute care hospitals are frequent.These institutions might serve as a reservoir for MRSA and infection control strategies are mandatory to limit the spread of the MRSA within these HCI and reduce MRSA transfer to others.
Owing to a higher prevalence of MRSA in HUG, one aim of this study was to establish to what extent this potential reservoir could spread to other western institutions in Switzerland.Part of the answer is given by the results of molecular typing of these isolates [25] which showed that: (1) 3 clones are responsible for the majority (73%) of the cases at HUG; (2) these clones were rarely observed in other hospitals; and (3) the majority of the cases in the western part of Switzerland, excluding Geneva, was due to a fourth epidemic clone [24].Thus, it seems that HUG faced mainly a geographically-limited problem which did not extend to the rest of Switzerland.As described for Western Switzerland, non-acute care beds were responsible for a large proportion of MRSA cases at HUG.
An important finding of the present study was that MRSA cases in ICUs accounted for less than 9% of all cases, as compared with other European hospitals where the majority of cases are observed in these units.However, the proportion was somewhat higher in the eastern part of Switzerland.Another important finding is the fact that in university hospitals, less than 30% of cases can be considered as having acquired their MRSA during the hospitalisation, the remaining 70% being probably acquired during previous hospital stays.Finally, the attack rate of MRSA transmission, which could only be assessed appropriately in university Discussion hospitals, remains extremely low in Switzerland with an average incidence of 2.8 cases/10,000 admissions.This illustrates the efficacy of infection control programs, even in hospitals with high endemicity or surrounded by known acute care facilities with extremely high numbers of MRSA patients, probably resulting from an ongoing outbreak.
In conclusion, the study confirms and extends previous data and shows that most Swiss hospitals have a low incidence of MRSA, despite some outlying institutions with endemic or highly epidemic MRSA conditions and neighbouring countries with a high prevalence.This might be the result of an "aggressive" policy of MRSA surveillance and control [6,14].Overall, a national approach to sur-veillance and control of MRSA is warranted in Switzerland in order to prevent conditions arising which favour its further development, thereby decreasing the risk of epidemic MRSA transmission.
A special thanks to all Swiss hospitals and laboratories which participated in the Study.

Figure 3 Distribution
Figure 3Distribution of genotypes in community and university (Basel, Bern, Geneva, Lausanne and Zurich) hospitals (one strain per case).The four predominant clones (WCH, GE1, GE2 and GE3) are represented by their major type (1, 9, 40 and 57, respectively) and their subtypes (letters).Reproduced with the permission of Clinical Microbiology and Infection[25].

Figure 4
Figure 4 Geographical distribution of the four predominant clones of MRSA in Switzerland in 1997 (one symbol per institution and per clone, the number in the symbol indicates the number of cases harbouring the clone in the institution).Reproduced with the permission of Clinical Microbiology and Infection [25].
: Dominique.Blanc@chuv.hospvd.chSwiss Medical Weekly: Call for papers Swiss Medical Weekly The many reasons why you should choose SMW to publish your research Official journal of the Swiss Society of Infectious disease the Swiss Society of Internal Medicine the Swiss Respiratory Society Med Wochenschr (1871-2000) Swiss Med Wkly (continues Schweiz Med Wochenschr from 2001) Editores Medicorum Helveticorum 9 • w w w. s m w. c h