Methicillin-resistant Staphylococcus aureus (MRSA) is a significant nosocomial pathogen. In Europe, only a small number of countries have succeeded in controlling MRSA (Tiemersma et al. ²⁰⁰⁴). Epidemiological monitoring is essential for limiting the occurrence and spread of epidemic clones within and between hospitals. In the study of strain origin, adequate typing of MRSA has become an important tool. Several methods have been used to distinguish MRSA strains (Tenover et al. ¹⁹⁹⁴). The “gold standard” is pulsed-field gel electrophoresis (PFGE) (Murchan et al. ²⁰⁰³). However, while PFGE has excellent discriminatory power, it is labor-intensive and difficult to standardize among different laboratories (van Belkum et al. ¹⁹⁹⁸). The method which has become increasingly popular during recent years is spa typing (Strommenger et al. ²⁰⁰⁸). It has been shown that spa typing, in contrast to PFGE, can be used to study both the molecular evolution as well as hospital outbreaks of MRSA (Tang et al. ²⁰⁰⁰). Moreover, the implementation of software algorithms for the grouping of related sequence types enables the classification of isolates into particular clonal lineages (Rupptisch et al. ²⁰⁰⁶). A good correlation between the clonal groupings of MRSA obtained by spa typing and those obtained by other methods has been described (Strommenger et al. ²⁰⁰⁸; Tang et al. ²⁰⁰⁰).

Phage typing has been used for epidemiological surveillance of S. aureus strains for many years (Blair and Williams ¹⁹⁶¹; Widemauwe et al. ²⁰⁰⁴). Despite of being an internationally recognized method, the large number of non-typable strains limits its use, particularly with respect to MRSA. Thus, for typing of MRSA, a selected set of ten phages has been recommended (Richardson et al. ¹⁹⁹⁹). Our earlier studies, based on phage typing, hypothesized that the predominant epidemic MRSA strains were periodically disseminated in various hospitals in Gdańsk region (Wiśniewska et al. ²⁰⁰⁰; Wiśniewska et al. ²⁰⁰⁷). In the present study, we investigated the emergence and diversity of spa–clonal complexes among single-patient MRSA isolates obtained for phage typing from the University Clinical Center (UCC) in Gdańsk between 2008 and 2009. We used phage typing as the initial screening in this study. Furthermore, we compared the results with those of the previous study and provided more information about predominant clones.

The susceptibility of the MRSA isolates to antimicrobial agents was determined by the disk diffusion method according to the guidelines of CLSI ²⁰⁰⁶ and Hryniewicz et al. ²⁰⁰⁵. The following drugs were used for test: erythromycin, clindamycin, ciprofloxacin, co-trimoxazole, tetracycline, gentamicin, vancomycin, teicoplanin, fusidic acid, rifampicin, linezolid, quinupristin-dalfopristin, chloramphenicol (Becton Dickinson, Germany), and mupirocin (Oxoid, UK). For isolates identified as resistant to erythromycin, but susceptible to clindamycin, D test was performed to detect inducible clindamycin resistance (Fiebelkorn et al. ²⁰⁰³). For vancomycin, additionally, the minimal inhibitory concentration (MIC) was determined by E tests as described by the manufacturer (AB Biodisc, Sweden).

According to the classification using the basic set of phages, most of the isolates (56 %) belonged to the phage group III (Table 1). Twelve percent of the isolates was in the phage group II, and none in group I. Mixed groups or inhibition reactions were observed occasionally, and 30 % of the isolates were resistant to phages of the basic set. With the additional set of phages, eight types were identified. In 50 %, it was phage type MR8, the type MR25 in 8 %, and the type 622 in 6 %. The remaining types were found occasionally. The non-typable isolates accounted 26 %.

Looking at the frequency of resistance to antimicrobial agents of the strains (Table 2) revealed that the resistances to ciprofloxacin, erythromycin, and clindamycin were most frequent (86, 76, and 84 %, respectively). Inducible resistance to clindamycin was limited to one isolate (not shown), 42 % were resistant to tetracycline and 32 % to chloramphenicol (14 % were resistant to both antibiotics; not shown). Resistance to gentamicin, co-trimoxazole, and rifampicin were found in 16, 10, and 2 %, respectively. All isolates were sensitive to vancomycin, teicoplanin, fusidic acid, linezolid, mupirocin, and quinupristin-dalfopristin. The MIC_s vancomycin ranged between 0.75 and 1 μg/mL (not shown).

Sixteen distinct spa types were observed (Table 3). The most frequent were t003 (22 %), t151 (16 %), and t008 (12 %). The remaining types accounted for between 10 and 2 % each. BURP analyses of the spa types identified two clusters and eight singletons. The predominant was CC010, which harbored 50 % of MRSA isolates. It consisted of the strains associated with six types: t003, t151, t002, t2065, t045, and t010. The most common within CC010 were t003 (11/30 = 36.7 %) and t151 (8/30 = 26.7 %). Most of the strains of CC010 (80 %) were identified as identical SCCmec type II with five bands of 162, 209, 284, 311, and 342 bp (the results of the PCR not shown). The remaining strains of three types of this complex harboring SCCmec type IV and were positive for 343 bp alone (t2065 and t010) or additionally for 311 bp (t045). The second cluster identified has no founder (12 % of the isolates). Only two spa types: t037 (all harboring SCCmec type III with 209 and 243 bp) and t029 (SCCmec IV type) were assigned to this cluster. Among eight singletons identified, the most frequent was represented by 12 % of the isolates and was consisted of only spa type t008. Six singletons were unique to one isolate. The strains from all singletons were identified as SCCmec type IV as to be positive for 311 bp alone or additionally for 343 bp.

The isolates belonging to the predominant cluster CC010 was distributed in 13 wards or clinical departments (Table 4) with the most frequency in dermatology, venerology, and allergology department (6/30 = 20 %) and neurology ward (5/30 = 16.7 %). The prevalent cluster was disseminated in different clinical settings and was constantly present throughout the observed period. Most of the isolates of CC010 (16/30 = 53.3 %) were resistant to erythromycin, clindamycin, and ciprofloxacin, other twelve (40 %) were resistant to chloramphenicol and six (20 %) to tetracycline. The isolates resistant to tetracycline were sensitive to macrolides. Eighty-three percent (25/30) isolates of CC010 belonged to the phage group III. The most common phage types were 47/54/75/83A and 54/75/83A (12/30 = 40 %). Four of the isolates of CC010 were non-typable with basic set of phages (4/30 = 13.3 %), and one belonged to the phage group II. The additional set of phages was unable to type only three isolates (10 %), including two isolates non-typable with the basic set. Irrespective of spa type, the isolates belonged to the predominant type which was MR8 (23/30 = 80 %). The isolates non-typable by additional set of phages and three isolates identified as type 622, in contrast to phage type MR8, were sensitive to erythromycin and clindamycin, but generally resistant to ciprofloxacin, tetracycline, and chloramphenicol. According to MLST, two related sequence types were identified within CC10. The isolate representing the spa type t003 was classified as ST225 and the remaining ones as ST5 (not shown in the table).

We used phage typing as the first step in monitoring the epidemiology of MRSA infection in the main clinical center in North of Poland. The initial Polish investigations with the additional set of phages were carried out on 150 clinical MRSA isolates collected by the Microbiology Department of the Medical University of Gdańsk between 1990 and 1998 (Piechowicz et al. ¹⁹⁹⁹). In contrast to the results of the present study, the initial survey showed 96.7 % of the isolates as typable. Moreover, in 1997–1998, in most of hospitals of Gdańsk region, including the UCC, predominated the MRSA strains belonging to the new phage type MR25/M5, resistant to erythromycin, clindamycin, ciprofloxacin, doxycycline, gentamicin, and rifampicin and resistant or medium susceptible to fusidic acid (Wiśniewska et al. ²⁰⁰⁰). The study performed more recently in five selected hospitals in Gdańsk area showed that in 2005–2007, the phage type MR25/M5 disappeared but the most frequent was the type 56B, resistant to erythromycin, clindamycin, ciprofloxacin, tetracycline, gentamicin, and co-trimoxazole (Wiśniewska et al. ²⁰⁰⁷). The present study showed that a significant shift in the prevalence of phage types was achieved as over 10 years later; the most common phage type in the UCC was MR8. Moreover, if we accept the guideline of Blair and Williams ¹⁹⁶¹, that isolates may be considered to be different if they differed in strong reactions to more than one phage, we found only 6 % of typable isolates from this center to be quite different.

Our study confirmed the general trends that most of typable MRSA strains belong to the phage group III. However, compared to the previous study we performed, this study revealed a remarkable increase in the number of isolates non-typable by the additional set of phages (Piechowicz et al. ¹⁹⁹⁹; Wiśniewska et al. ²⁰⁰⁷). With respect to the basic set of phages, a non-typability was comparably high. The survey performed on Belgian MRSA isolates collected from 93 different hospitals between 1992 and 2001 showed that typability of these strains by basic set of phages was at least 95 %, except for the year 1997 where 9 % of the studied population were non-typable, but it was mostly due to 26 % non-typability in one hospital (Widemauwe et al. ²⁰⁰⁴). In other countries, problems due to non-typable MRSA have also been encountered (Samra et al. ²⁰⁰¹).

In our earlier studies with phage typing, no molecular data were included. Phage typing has been the classical method for detecting epidemic and pandemic strains, although confirmation of selected isolates by molecular methods is now needed. Spa typing appears to have significant advantages over many currently available methods. The main advantage is the portability of sequence data which greatly simplifies the sharing of information between laboratories and facilitates the creation of the large-scale database for the study of global as well as local epidemiology (Harmsen et al. ²⁰⁰³; Strommenger et al. ²⁰⁰⁸). The spa types among MRSA strains are area specific (Fenner et al. ²⁰⁰⁸; Fridrichova et al. ²⁰¹⁰; Marchese et al. ²⁰⁰⁹). The ability of spa typing to distinguish strains makes it particularly well suited for initial screening that may be used for further identification of the strains (Rupptisch et al. ²⁰⁰⁶). Looking at sequence database for spa types, specific repeats seem to be associated with particular MLST sequence types. From the sixteen spa types observed in our study, six constituted one predominant clonal lineage with the most frequent t003/ST225. Moreover, all of the remaining spa types clustered into the CC010 belonged to the worldwide widespread MLST type ST5. Spa type 003 deposited in the spa server corresponds to the epidemic clone isolated in the USA and in many European countries, including the neighboring countries of Poland (Strommenger et al. ²⁰⁰⁸). The spa type t002, which in our study also belonged to the CC010, is the fourth most frequent spa type worldwide among MSSA and MRSA isolates (Fenner et al. ²⁰⁰⁸). Similarly, spa type t008, in our study clustered in the most frequent singleton #1, belonged to the top ten spa types in this database. If only predominant spa CC010 data were analyzed, the concordance between spa clone and phage type was very high, but the dominant phage type MR8 was observed in the isolates belonging to the different spa types. A promising fact is that only 10 % of the isolates of CC010 are non-typable by the additional set of phages. Thus, the high number of non-typable isolates observed in this study was caused by the presence of MRSA clusters which occasionally occurred in the center.

In comparison with the results of our previous study, in the present study, multi-resistant strains were found occasionally (Wiśniewska et al. ²⁰⁰⁰) and the most prevalent strains were resistant only to macrolides and ciprofloxacin. In other countries in Europe, a similar trend in levels of antibiotic resistance can be observed (Marchese et al. ²⁰⁰⁹; Kluytmans-Vandenbergh and Kluytmans ²⁰⁰⁶). Since the 1990s, community-associated MRSA (CA-MRSA) has emerged worldwide (Bassetti et al. ²⁰⁰⁹). This strains, unlike to hospital acquired MRSA (HA-MRSA) are generally more sensitive to non-β-lactam antibiotics (Naimi et al. ²⁰⁰³). The potential risk of its introduction into hospitals are currently matters of great concern. HA-MRSA and CA-MRSA can be distinguished on the basis of the phenotype and genetic profile in SCCmec typing (Naimi et al. ²⁰⁰³). SCCmec types I–III carry additional genes that provide resistance to antibiotics other than ß-lactams and are most often found in HA-MRSA. SCCmec types IV and V usually carry no additional drug resistance genes and are mainly associated with CA-MRSA. In our study, almost half of the MRSA population showed a pattern band characteristic for SCCmec type II. Moreover, this type was associated exclusively with the spa-clonal complex predominantly found in the hospital and corresponded with one of the most frequent hospital clone ST5 or its single-locus variant Rhine-Hesse clone ST225. Both of these STs recently have been observed as an epidemic hospital clones in Poland and in neighboring countries (Łuczak-Kadłubowska et al. ²⁰⁰⁸). SCCmec type III was sporadically occurring and related to the second cluster identified. As it was shown, irrespective of the spa type and spa clusters, that SCCmec type IV circulated in the hospital. The fact that the amplification patterns of such strains differed only in one band showed that they occurred in limited genetic variants of the mec cassette. Most of the strains of SCCmec type IV, including predominant spa-CC, as expected, were sensitive to more antibiotics than the strains belonging to SCCmec II and III. However, within other spa types, in single cases, resistance patterns were similar, especially with respect to fluoroquinolones, clindamycin, tetracyclines, and gentamicin. This observation suggests appearance of unique variant subtypes.

In conclusion, by this study, we demonstrate that phage typing is still an internationally recognized method and, with addition of spa typing, may be a useful tool for rapid research of correlation between MRSA isolates. This is the first attempt at using these methods in combination, with respect to MRSA population of one of the main clinical center in the North of Poland. Although, our investigation may be considered to be an initial screening for studies providing full genetic characteristic of MRSA strains isolated in the UCC. Further study should focus on monitoring of changes in the resistance profile of MRSA and search for emergence of CA-MRSA in this hospital. According to the report from a countrywide survey, MRSA clone, possibly representing community-acquired strain, are circulating in Polish hospitals (Łuczak-Kadłubowska et al. ²⁰⁰⁸). One of the distinguishing features of CA-MRSA is that high percentage of such strains are positive for Panton–Valentine leukocidin gene (Naimi et al. ²⁰⁰³). Thus, such investigation would be on interest.