Surveillance and Typing Scheme of MRSA

Finnish microbiology laboratories report (generally electronically) all MRSA isolates to the National Infectious Disease Register at the National Public Health Institute (KTL). The KTL records the date, source of specimen, and the patient’s birth date, sex, and place of treatment. Using this information and a time interval of 36 months, multiple isolations from the same person are deleted from the database. The microbiology laboratories also send the MRSA isolates to the Laboratory of Hospital Bacteriology at KTL for further analysis. Phage typing, PFGE, and antimicrobial drug susceptibility testing were performed as described (11). In brief, phage typing was performed with the universal set of phages (13) at 1x and 100x routine test dilutions, both with and without heat treatment of bacteria (14). Antimicrobial drug susceptibilities were tested by the disk diffusion method according to guidelines recommended by the National Committee for Clinical Laboratory Standards (NCCLS). The antibiotics tested were oxacillin, ampicillin, penicillin, cephalexin, cefuroxime, gentamicin, tobramycin, erythromycin, clindamycin, chloramphenicol, ciprofloxacin, rifampicin, fusidic acid, mupirocin, and vancomycin. MICs of oxacillin were determined by E-test (AB Biodisk, Solna, Sweden) according to manufacturer’s instructions. If the oxacillin MIC was <64 μg/mL, methicillin resistance was verified with the MRSA screen test (Denka Seiken, Japan) or mecA-polymerase chain reaction (15). For PFGE, genomic DNA prepared in agarose blocks was digested with SmaI restriction endonuclease, and chromosomal fragments were separated with a Chef DR III (Bio-Rad Laboratories, Hercules, CA) for 24 hours with initial and final switching times of 10 seconds and 60 seconds, respectively. PFGE profiles differing by fewer than four bands were interpreted as identical or closely related (16). Bionumerics software, with Dice coefficient and unweighted pair group matching average for clustering, and optimization and tolerance values of 1% and 1–2%, respectively, were used to verify the relatedness of PFGE profiles. MRSA isolates sharing identical or closely related PFGE profiles and phage types were considered to be the same strain type. If the PFGE profiles were related, but the phage types were different, ribotyping with one to three restriction enzymes (HindIII, EcoRI, ClaI) was performed to verify the relatedness of isolates (11). Riboprofiles with fewer than four bands’ difference were regarded as identical or closely related. The genomic variation within the mec determinant hypervariable region (HVR) was analyzed for a subset of MRSA isolates as described (12). In brief, genomic DNA digested with EcoRI and BglII restriction endonucleases was hybridized with two probes prepared from plasmid pBBB30 (17) and was recognized as the mec hypervariable region.

Study Population, MRSA Isolates, and Definitions

All MRSA-positive persons in Finland from 1997 to 1999 and their MRSA isolates (one from each) were included in the study. A sporadic strain was defined as a strain type isolated from one person only. An MRSA isolate was defined as hospital acquired if the MRSA-positive specimen was obtained 2 days after hospital admission, and for discharged patients, if the patient had been hospitalized within 2 years before the date of MRSA isolation.

Previous Contacts with a Health-Care Facility

National identity codes for each person with an MRSA-positive culture were obtained either through the primary diagnostic laboratory or from the infection control nurse of the health-care facility. Based on the first isolation date of MRSA and the national identity code, data on previous hospitalizations within 2 years before the MRSA isolation date were retrieved from the National Hospital Discharge Register (HILMO). The HILMO is a civil register comprising comprehensive health-care records, provided by all hospitals and health-care centers in Finland, including outpatient surgery. Each report to the register includes patient identity information, admission and discharge dates, a code of the health-care provider, type of service, specialty, the place (home or institution) from which the patient came to the institution, and data on surgical procedures.

Additional background information was collected for persons for whom no HILMO reports could be found by sending questionnaires to infection-control nurses at relevant health-care facilities. The information collected included: 1) whether the MRSA-positive person was a patient or a staff member, 2) whether the specimen was taken on clinical or screening basis, and 3) whether the screening sample was taken because of a hospital contact abroad or because of an epidemic situation.

Statistical Analysis

For categorical variables, proportions were compared by the chi-square test with Yates correction or Fisher’s exact test, as appropriate. The means and medians of the continuous variables were compared by the Student’s t test or Mann-Whitney U test, depending on the sample distribution.

With approval from the Ministry of Social Affairs and Health and the Finnish data protection authority, the National Research and Development Centre for Welfare and Health gave permission to use the data from the HILMO register.

Contacts with Health-Care Facilities

Of the 526 MRSA-positive persons, 108 (21%) did not have any verified link to health-care facilities 2 years before the MRSA isolation date, including 17 persons whose MRSA was isolated within 2 days of hospital admission. Their MRSA isolates were classified as community acquired. Specimens were taken from 69 persons; 37 persons were specifically screened for MRSA. Of those 37, 21 had a known MRSA contact in their family, 13 were otherwise exposed to a known MRSA carrier, and 3 were known to have carried MRSA previously.

The HILMO register and the questionnaire survey showed 418 (79%) persons who had at least one connection to a hospital, and their MRSA isolates were classified as hospital acquired. According to the HILMO register, 376 persons were hospitalized in Finnish hospitals within 2 years before the MRSA isolation date: 156 hospitalized patients with MRSA isolated 2 days after hospital admission and 220 discharged patients. The time frame between the MRSA isolation date and the previous hospitalization for the discharged patients was <6 months for 156 (71%) patients, 6 to 12 months for 36 (16%), and >12 months for 28 (13%) patients. The questionnaire survey identified 42 additional persons who had an obvious contact with a hospital, including 23 staff members (12 working in a Finnish health-care facility and 11 who had recently worked abroad) and 19 patients, 15 of whom had recently been hospitalized outside Finland.

The median age of the persons who did not have a contact with a health-care facility was lower than that of persons who had contact (34 vs. 58 years, p<0.01). The proportion of children <15 years of age was higher in community-acquired than in hospital-acquired MRSA strains (27 [25%] of 108 vs. 31 [7%] of 418, p<0.01).

Strain Types

Among the 526 MRSA isolates, our typing scheme showed 84 strain types, 56 (67%) of which were sporadic and 28 (33%) shared by at least two persons. The distribution of sporadic (in total 56 [11%] of 526) and shared (in total 470 [89%] of 526) strain types was similar in persons with and without connections to health-care facilities (Table 1).

Figure

Figure. Pulsed-field gel electrophoresis (PFGE) profiles of the 14 most common methicillin-resistant Staphylococcus aureus (MRSA) strain types identified, Finland, 1997–1999. Lanes 1, 9, 17: S. aureus NCTC 8325 (molecular weight marker); lanes...

Fourteen strain types, each of which were isolated from >10 persons, represented 421 (80%) of 526 MRSA isolations (Table 2, Figure). Three of the 14 most common strain types were more likely to be found in persons who did not have a contact with a health-care facility than in those who had such a contact: Mikkeli clone (41 [38%] of 108 vs. 75 [18%] of 418, p<0.01), E31 (10 [9%] of 108 vs. 6 [1%] of 418, p<0.01), and E22 (8 [7%] of 108 vs. 7 [2%] of 418, p<0.01).

Of all strains isolated from persons who had no hospital contact, 94% were nonmultiresistant. In addition, of the 14 most common strain types, all 7 nonmultiresistant strains, but none of the multiresistant strains, showed HVR type A (Table 2). Of the 56 sporadic MRSA strains, 41% were nonmultiresistant, including all but one of the strains isolated from persons without connections to hospitals (Table 1).