The multi-resistant nature of most MRSA clones found in hospitals represents a therapeutical challenge for treating serious MRSA infections. The burden that the Iberian clone posed in Spanish hospitals in the early 90 s [3, 28], shifted to other clones susceptible to more antibiotics, which have been dominant in recent years [8, 29]. In this paper, we described the emergence and spread of a MRSA clone resistant to clindamycin, erythromycin, gentamicin, tobramycin, ciprofloxacin and rifampicin which has reduced substantially the number of effective antibiotics for treatment of serious MRSA infections.
Rifampicin is an antibiotic of substantial interest in the rise of MRSA infections, but cannot be used as a single agent to treat such infections because rapid emergence of resistance can occur, even during therapy [16, 30, 31]. Although the role of rifampicin as adjunctive therapy is controversial , the combined therapy seems beneficial as long as the bacteria exhibit susceptibility to the antibiotics combined . The distinctive phenotypic feature in the particular clone of ST-228 described here was the borderline resistance to rifampicin that could be missed by some methods of antimicrobial susceptibility testing (i.e. disk diffusion or E-test). Hence our interest in studying whether this low-level RIF-R was an adaptive phenomenon or to the contrary, known rpoB mutations underlay such phenotype.
Almost all isolates belonging to this multi-resistant MRSA clone (104/108) showed a low-level rifampicin resistance (MICs, 1 to 4 mg/L) and carried the amino acid substitution 481His/Asn in the RNA polymerase. Only 4 isolates showed additional substitutions known to be involved in a high-level rifampicin resistance: two isolates (MICs, 128 mg/L) carried mutational change 477Ala/Thr, and one isolate (MIC, ≥ 256 mg/L) 468Gln/Lys [13, 17, 27, 32]. The fourth isolate (MIC, ≥ 256 mg/L) showed substitution 527Ile/Leu, the only one which mutation was found in the rifampicin resistance-determining cluster II, described recently among Japanese MRSA isolates . It is noteworthy that 20 isolates (19%) of the RIF-R isolates, carrying rpoB mutation resulting in amino acid substitution in position 481, were detected as rifampicin susceptible by the disk-diffusion test. However, the inhibition zones of these strains were between 20 and 23 mm, closer to the susceptibility breakpoint established by CLSI (susceptibility ≥ 20 mm) than inhibition zones among RIF-S MRSA isolates that were usually ≥ 30 mm. Therefore, if screening for rifampicin resistance is made only by disk diffusion, special attention needs to be paid to strains borderline to the CLSI susceptibility breakpoint to avoid reporting false susceptibility results. MICs by E-test failed to detect rifampicin resistance following CLSI guidelines  in a group of 12 strains (MICs, 0.75-1 mg/L). These isolates showed MICs by microdilution of 2 mg/L and carried the rpoB mutation responsible for amino acid substitution in position 481. Thus, and according to other authors, it would be advisable to apply ≤ 0.5 and ≥ 8 mg/L as new breakpoints to classify rifampicin susceptibility or resistance in S. aureus clinical isolates [13, 17].
High-level rifampicin resistance could be attributable to double mutations within rpoB, as previously described . We did not find in this particular clone that the presence of a prior mutational change (481His/Asn) increased the frequency of acquisition of additional mutations responsible for a higher level of rifampicin resistance, when compared to a reference strain.
The multi-resistance pattern exhibited by the Iberian clone, dominant lineage in our hospital during the 90's, also included resistance to tetracycline. The new RIF-R MRSA isolates were resistant to clindamycin, erythromycin, gentamicin, tobramycin, ciprofloxacin and susceptible to tetracycline. However, molecular typing showed that the Iberian clone and the new RIF-R MRSA clone had different genetic backgrounds represented by ST-247 and ST-228, respectively, with only a single locus in common. Although both clones carried a SCCmec element type I, PFGE patterns and spa-types were clearly different.
All strains with the multi-resistant phenotype described in this work, showing resistance or decreased susceptibility to rifampicin, belonged to ST-228, carried a SCCmec element type I and were spa-type t041. This clone seems to be related to the Southern Germany clone (ST-228, SCCmec type I, spa-type t001 or spa-type t041) reported in Germany in 1997-98 [21, 33]. In the same period, strains of ST-228 and SCCmec type I were reported at several hospitals located in seven Italian cities , although these isolates also showed resistance to multiple antibiotics, rifampicin resistance was not stated. Recently, strains of ST-228 have spread epidemically in Finland in 2002-2004 and in Hungary in 2003-2004 [35, 36]. Also, ST-228 has been reported in other European countries: Belgium, Slovenia or Switzerland . The first isolate ST-228, SCCmec type I was isolated in our hospital in September 2003, from a patient admitted to the ICU. However, it was not until March 2004 that this clone spread epidemically in our hospital and currently represents one third of all clinical MRSA isolates in our institution. Strains belonging to ST-228 have been reported in other hospitals in Spain since 1996 [9, 29, 38]. However, none of these reports (from Spain or other countries) analysed the decreased susceptibility to rifampicin among representative strains of ST-228. During the 2004-2007 period, we did not find significant changes in the rifampin consumption in our institution, which was on average 0.5 DDD/100 patients-days for intravenous and 1.0 DDD/100 patients-days for oral administration.
A set of 5 strains resistant to clindamycin, erythromycin, gentamicin, tobramycin, ciprofloxacin, but fully susceptible to rifampicin with MICs of 0.012 mg/L were included in this study. On average, this RIF-S pattern represented 4% of all MRSA isolated between 2004 and 2006, however this resistance phenotype can be traced back to 1999 in our hospital. The RIF-S isolates were classified as ST-228, the same as the RIF-R MRSA. Isolates of ST-228 (MLST, arcc 1, aroe 4, glpf 1, gmk 4, pta 12, tpi 24, and yqi 29) belong to the Clonal Complex 5, as well as isolates of ST-125 (MLST, arcc 1, aroe 4, glpf 1, gmk 4, pta 12, tpi 1, and yqi 54) which was the dominant MRSA clone in Hospital Universitari de Bellvitge from 1996 to 2003. Consequently, an alternative hypothesis to explain the emergence of multiresistant clone ST-228, SCCmec type I would be the SCCmec type I transfer from ST-247 to the ST-125 background, with an intermediate stage where RIF-S isolates could be found belonging to ST-228.