Methicillin-resistant Staphylococcus aureus (MRSA) has become a persistent problem worldwide. MRSA has been established as a major hospital pathogen but it has also been found increasingly in long-term facilities, and in communities from persons having no connection to the health-care setting 1. It is generally accepted that hospital-acquired MRSA isolates spread to the long-term facilities as well as to the community. However, the community-based isolates may migrate into the health-care settings, and thus create a two-way flow of MRSA 2.

Of the genotyping methods, pulsed-field gel electrophoresis (PFGE) has been suggested to be the golden standard for local outbreak investigations 3, and multi locus sequence typing (MLST) and the mobile genetic element, staphylococcal cassette chromosome mec (SCCmec)-typing 45 for understanding the evolution of different MRSA clones 6. SCCmec types I-III, and IVA have been associated with hospital acquisition, and types IV and V are mostly found in strains circulating in the community 7. Several major different pandemic MRSA clones have been identified mainly by PFGE, i.e. Iberian, Brazilian, Hungarian, New York/Japan, Pediatric, and UK EMRSA-15 and -16 clones 8. MLST has, however, revealed that several MRSA clones, considered to be distinct by using PFGE and other molecular typing techniques, were indistinguishable 9. Using MLST and eBURST, the five most common pandemic MRSA lineages can be grouped as clonal complexes: CC5, CC8, CC22, CC30, and CC45. Several nationwide studies on molecular epidemiology of MRSA have also been performed 1011121314. However, the extensiveness of the MRSA isolate collections used in those studies varied, and most of the studies were based on selected isolates as opposed to all MRSA isolates being identified on the national level.

The purpose of our study was to examine the molecular epidemiology of MRSA in Finland during 1997–2004 in comparison to European and pandemic MRSA clones by using the nationwide collection including all MRSA isolates, thus giving an insight into the trends of MRSA evolution at the nationwide level over a certain period of time. We also performed a survey which focused on the background information of the MRSA-positive persons in 2001–2003.

Our study is based on comprehensive collection of MRSA isolates obtained nationwide, and during 1997–2004, 4026 MRSA notifications were received by the NIDR (120–1458 per year). The annual incidence of MRSA rose over ten-fold, from 2.3 notifications per 100,000 population in 1997 to 27.9 in 2004 (Figure 1). Among the 4026 MRSA notifications, 71 specimens were obtained from blood and 4 from CSF; 30 of the blood isolates and 3 of the CSF isolates were from 2004. The proportion of MRSA among blood isolates of S. aureus rose from under 1% to 2.8% in 2004, although it is still low compared with many other countries in Europe or worldwide 1516. During 1997–2004, the numbers of MRSA notifications were strikingly high in three HDs and in six others there was an increasing trend. In 1997, the incidence of MRSA was slightly over 10 notifications per 100,000 population only in the Helsinki metropolitan area, whereas the incidence rate rose to over 20 notifications per 100,000 population in eight HDs by 2004.

During the same period of time a total of 4091 MRSA isolates were studied in the reference laboratory. PFGE identified 253 different strains of which 38 were outbreak/epidemic strains, and 215 sporadic (Figure 2). Twenty-four new outbreak/epidemic strains appeared during 1997–2004. The proportion of sporadic strains varied between 3–13% during the study period. There is a slight difference in the numbers of MRSA notifications and MRSA isolates (65), this is partly due to the differences in the case definitions (a time interval of 12 months versus 36 months); also if the same person had two different types of isolates, both of them were included (a total of 27 persons carried two different strains). There may also exist some MRSA notifications of which the MRSA isolate was not sent to the reference laboratory and vice versa.

At the beginning of the study period (1997–2001), the distribution of different MRSA strains was more diverse, but since 2002, five strains (FIN-16 (ST125:IA), FIN-4 (ST375:IV), FIN-21 (ST228:I), FIN-7 (ST8:IV), and FIN-10 (ST45:IV and V) have been most predominant (Table 1). The number of internationally recognized epidemic strains, FIN-16 (ST125:IA), a single-locus variant of ST5 (New York/Japanese clone) – named previously Bel EC-3 – and FIN-21 (ST228:I), has strongly increased since 2003. In 2004 more than half of the isolates (53%) were composed of FIN-16 and FIN-21. The most common MRSA isolates found from blood and from CSF in 2004 were FIN-16 (12/34), of FIN-21 (9/34), and of FIN-4 (4/34); no significant associations were found between the type of specimen (blood and CSF vs. other) and these strains in 2004 nor during the whole study period.

The proportion of FIN-16 (ST125:IA) strains increased from <1% in 1997 to 25% in 2004. The majority of all FIN-16 strains (554/742) were found in one HD in central Finland, and it has dominated in that area since 2001, but it also extended to 14 other HDs. Pérez-Roth and others have reported a similar increase of ST125:IV strains in Spanish hospitals 17. Our survey (2001–2003) showed that FIN-16 strains were more likely to be found from clinical specimens than from those obtained on screening and other basis (199 [25%] of 805 vs. 160 [18%] of 902; p < 0.05) (Table 2), and it was the most common strain found from invasive specimens in 2004 indicating that FIN-16 might be of epidemic potential and virulent. Since FIN-16 has dominated in one HD for four years it can be considered an endemic strain in that area, especially in long-term facilities. Thus, we can equally assume that its capacity to cause clinical infections may occur because FIN-16 strain has reached a sufficiently high level in the population, especially among the oldest and most seriously ill patients who are being transferred between long-term facilities and acute care hospitals.

The FIN-21 (ST228:I) strain was found in Finland for the first time in September 2002 18. Its proportion increased from 6% in 2002 to 28% in 2004, and it occurred mostly in the Helsinki metropolitan area (98%; 572/581) where it caused many epidemic clusters during 2002–2004. The strain has a similar PFGE pattern to the Italian clone 12, expresses MLST allelic profile ST 228, a double-locus variant of ST5, and SCCmec type I. MRSA strains with the same ST228 have also been reported as the "Southern Germany" epidemic clone 19. The strain tended to be found more often from screening and other specimens than from those taken on clinical basis (58 [7%] of 805 vs. 87 [10%] of 902; p = 0.086) (Table 2), as opposed to FIN-16. This may be due to differences in the MRSA screening activity between the HDs, or the later appearance of the strain in Finland. We have previously shown that during 2001, the screening activity varied widely between the HDs (0–2768 per 100,000 population), but it was not related to the rate of MRSA 20.

UK EMRSA-16, currently named as FIN-6 (ST36:II), decreased from 13% in 1997 to <1% in 2003, and FIN-3 (ST5:II) from 18% to 3%, respectively. UK EMRSA-15, now FIN-12 (ST22:IV) varied between 1–6%. This is in contrast to the situation in the UK where the same strains (UK EMRSA-15 and -16) are predominant 21. In Spain a rapid nosocomial dissemination of an identical strain (UK EMRSA-16) has been reported 17, and also in Sweden it caused a large outbreak in hospital setting 22. In our material, FIN-12 was the most commonly found strain, 14/19 (74%), among staff members screened because of hospital contact abroad (Table 2).

More than half of the MRSA strains carried SCCmec IV or V (Figure 2), which are often seen in community-acquired MRSA (CA-MRSA) 7. One of the three strains which was previously associated with community-acquisition 23, FIN-4 (ST375:IV, a single locus variant of ST59), was the second most common strain (Table 1). FIN-4 was more likely to be found among isolates obtained from patients and staff members on clinical basis than in screening (167 [21%] of 805 vs. 81 [9%] of 902; p < 0.05) (Table 2) in addition to FIN-16. It has been found in every HD during the study period (range of annual numbers in HDs 7–17). MRSA isolates of ST59 are considered as one of the most divergent MRSA sequence types 9. ST59 has been reported in a few MSSA isolates from the United States 24, and in a few MRSA isolates in Asian countries 2526. ST375, a single locus variant of ST59, has recently been shown to spread as CA-MRSA in Denmark 27. Unfortunately, no health-care records were available in the present study, and we cannot assess whether acquisition of FIN-4 strain is still associated with community. The two other prevalent SCCmec IV expressing strains, FIN-7 (ST8:IV) and FIN-10 (ST45:IV) were more likely to be found in screening specimens of patients and staff members than in their clinical specimens (70 [9%] of 805 vs. 194 [22%] of 902; p < 0.05, and 89 [11%] of 805 vs. 149 [17%] of 902; p < 0.05) (Table 2). FIN-15 (ST8:IV) caused an epidemic in a health-care facility in Eastern Finland in 2000, and since then, the strain has been found occasionally during 2000–2004.

In addition to the presence of SCCmec type IV or V, the presence of lukS-PV and lukF-PV-genes along with epidemiological data are generally used in the definition of CA-MRSA. Of the representative strains, FIN-11 (ST80:IV) and FIN-25 (ST8:IV) were PVL positive (Figure 2). In addition, some strains of FIN-5 (ST30:IV) and FIN-19 (ST1:IV) possessed lukS-PV – lukF-PV genes encoding PVL. The majority of all MRSA isolates which were previously associated with community-acquisition possessed type IV SCCmec, but only a minority (12%) of them were PVL positive 28. Most of these PVL-positive isolates were of ST80 as reported elsewhere 2729. However, although many isolates found in the community harbor these markers, transmission of "CA-MRSA" strains between health-care settings and community may occur 30.

Of FIN-2 strains, FIN-2e (ST241:IIIB) subtype caused an outbreak in Southern Finland in 1999, and only a few strains have been found since then. FIN-2e is a single locus variant of ST239, which is the main ST of FIN-2 strain, and thus it is possible that FIN-2e may have evolved from the FIN-2 by a point mutation. FIN-2 strain includes isolates possessing three different SCCmec types (SCCmec III, IIIA, and IIIB; Figure 2), indicating that this strain may have received the mecA gene several times or SCCmec may have been modified. This may be true also with both of FIN-1 and FIN-10 strains which possess two different SCCmec types (SCCmec IA and IVA, and SCCmec IV and V, respectively).

The annual proportion of sporadic strains decreased during the study period, although the total number increased slightly, indicating that new strains appear constantly. Among the patient findings, more than half of the sporadic strains were derived from clinical specimens, 23% from screening specimens based on exposure to MRSA, 20% from screening specimens based on hospital contact abroad, and 2% were other indications (Table 2). The sporadic strains were found most often from specimens taken because of a hospital contact abroad (30%; 19/64). Blanc and co-workers have reported that at least one-third of the sporadic isolates found in Switzerland were due to an ongoing introduction of new isolates from abroad 10. They have also shown that most of the sporadic Swiss isolates were related to other European epidemic MRSA clones 31. We cannot be confident of the origin of our sporadic strains because our routine molecular typing scheme is primarily based on the highly discriminatory PFGE, which allows detection of genomic microevolution only. A thorough MLST and eBURST analysis as well as spa typing might have revealed whether our sporadic strains have descended from a few ancestors.

In Finnish acute care hospitals the control of MRSA is based on active culture screening and contact isolation. The increase in the incidence of MRSA remained one of the lowest in Southwestern Finland in comparison with the other regions. In the beginning of the 90's, the first three major MRSA outbreaks in Southwestern Finland and the Helsinki metropolitan area were successfully controlled 323334. Later, molecular typing revealed numerous smaller outbreaks 13. In 2001 the number of MRSA cases rose in elderly persons outside the Helsinki metropolitan area, suggesting an emerging problem in long-term care 20. Consequently, the national guidelines for control of MRSA were updated to cover also long-term facilities.

Regardless of a strict control policy, the most recent MRSA manifestations in Central and Southern Finland were related to the spread of two internationally recognized MRSA strains, FIN-16 and FIN-21, and these two strains also caused the striking increase in blood and CSF findings. The outbreaks were, however, different in their epidemiology: FIN-16 prevailed in long-term facilities and FIN-21 in acute care hospitals. During the survey period (2001–2003), the proportion of screening specimens increased among the patients and staff members, and almost half of the MRSA isolations were recovered from patients' screening specimens (Table 2), reflecting the active screening policy. The increase in the screening activity may partly be related to the increase in the MRSA cases we detected, especially in cases with FIN-21 18, but it is unlikely that it had an influence on the worrying trend in blood and CSF findings. In addition, the implementation of control measures may not have been in accordance with the presumptions, which may be due to the lack of resources in some hospital districts or in some settings. Although we had only limited data on clinical infections caused by MRSA, and could not differentiate screening specimens from those taken on clinical basis over the whole study period, we believe that the spread of MRSA has increased, as indicated by each new MRSA case. To guarantee appropriate resources in infection control in all HDs the Ministry of Social Affairs and Health allocated additional funds to MRSA for the years 2005–6. Thus far, this special investment may have paid off since the rising incidence of new MRSA cases seems to have slowed down (data not shown). However, a permanent trend in the near future remains to be seen.

The predominant MRSA strains change over time, but two internationally spread epidemic strains of MRSA were related to the increase of MRSA incidents detected most recently, and those strains were also the cause of the striking increase of invasive MRSA findings, thus showing the real worsening of the situation in Finland. In addition, the rise in MRSA strains with SCCmec types IV or V, possible CA-MRSA, was similar to other European countries 2729. The future will show if active control measures will still function in controlling MRSA in this new setting.

The authors declare that they have no competing interests.

A-MK carried the main responsibility of writing the manuscript. She also collected and analyzed the data of the questionnaire survey, analyzed the typing scheme data, and participated in confirmation and typing of MRSA, and in PFGE profile analysis. OL conceived the study and participated in its design and coordination, and participated in writing the manuscript. MK-L participated in the confirmation and typing of MRSA, in PFGE profile analysis, and especially in MLST-typing and analysis. SI participated in the confirmation and typing of MRSA, in PFGE profile analysis, and carried out SCCmec-typing. SS participated in the confirmation and typing of MRSA, especially in the analysis of PFGE profiles, was involved in drafting the questionnaire survey, and revision of the manuscript. AV participated in the study design and in writing the manuscript JV-V conceived the study and participated in its design and coordination, and participated in writing the manuscript.