Enterococci are normal commensals of the human and animal gut that are particularly associated with nosocomial infections, as well as community carriage. Vancomycin-resistant enterococci (VRE) express high-level resistance to glycopeptide and aminoglycosides, resistance having been promoted via (i) the extensive use of vancomycin in healthcare environments, and (ii) the use of the animal growth promoter avoparcin. Several different VRE phenotypes have been described in the literature, including VanA, VanB, VanC, VanD, VanE, VanG, VanL and VanM, of which vanA, vanB, vanD and vanM genes code for a d-alanyl- d-lactate ligase, and vanC, vanE, vanG and vanL genes for a d-alanyl-d-serine ligase. Recently a new transferable VanN-type resistance, with homology to VanL-type vancomycin resistance has been reported [Reference Lebretin1].

In the nosocomial situation, the VanA phenotype is most prevalent and isolates are characterized by the expression of both high-level and inducible resistance to vancomycin [minimum inhibitory concentration (MIC) 64–1024 mg/l], and teicoplanin (MIC 16–512 mg/l), a phenomenon facilitated by the carriage of the transposon Tn1546. Surveys have shown the worldwide distribution of this phenotype [Reference Khan2–Reference Khan4]. The VanB phenotype is also common in nosocomial infections and confers lower levels of acquired resistance to vancomycin, but not to teicoplanin. To date, few reports have indicated the presence of clinically relevant vanA/B Enterococcus faecium isolates [Reference Dendle5]. The prevalence and genotypic relationships of VRE isolates have been documented in many parts of the world but relatively few studies have investigated the Middle East further and characterized VRE from hospitals in Saudi Arabia [Reference Khan4]. Therefore, the aim of the study was to expand the phenotypic and molecular characterization of Saudi Arabian VRE isolates, by investigating all vancomycin-resistant E. faecium isolates recovered from the King Fahad Specialist Hospital, Damman, Saudi Arabia between February 2006 and December 2007, and to characterize any vanA Tn1546 transposons present. The hospital is a tertiary-care centre with 230 beds which treats about 6000 patients annually and is situated in the eastern province of Saudi Arabia, about 375 km from the capital city of Riyadh.

During 2006–2007, 26 vancomycin-resistant E. faecium isolates were cultured from clinical specimens from hospitalized patients (Table 1). Isolates were confirmed as E. faecium species by polymerase chain reaction (PCR) [Reference Dutka-Malen, Evers and Courvalin6]. Antibiotic susceptibility testing for ampicillin, ciprofloxacin, tetracycline, quinupristin/dalfopristin and linezolid was performed using the VITEK 2 automated identification and susceptibility system (bioMérieux, France), and E tests (AB Biodisk, Sweden) were used for vancomycin and teicoplanin.

Table 1. Medical history of patients ranked according to their E. faecium PFGE genotypes

Unk., Unknown; n.a., not applicable; ARDS, acute respiratory distress syndrome.

Wards: AICU, adult intensive care unit; HBW, hepatobiliary ward, MM, male medical ward, MS, male surgical ward.

Specimens: RS, rectal swab; BLD, blood; PUS, pus; ETT, endotracheal tube; TLC, triple lumen catheter; SAC, sacral sore; CBD, common bile duct stent; URI, urine; BED, bed sore; TTD, T-tube drain; ABD, abdominal drain; TRA, tracheal aspirate.

Genotyping of isolates was performed using pulsed- field gel electrophoresis (PFGE) of Sma I digests of chromosomal DNA. Fragments were separated in 0·8% agarose gel at a constant voltage of 6 V/cm at 14 °C with pulse times of 3·5–25 s increased linearly over 12 h, and 1–5 s over 8 h. Gel profiles (48·5– 339·5 kb) were analysed using BioNumerics v. 6·0 software (Applied Maths, Belgium) with gel lanes normalized against a lambda DNA ladder (BioRad, USA) with band tolerance of 1·0%. Multi-locus sequence typing (MLST) was performed on 10 randomly selected isolates representative of PFGE types according to the protocol described at http://efaecium.mlst.net/misc/info.asp. The presence of vanA or vanB genes, and the characterization of Tn1546 elements was determined by PCR with published primer pairs [Reference Miele, Bandera and Goldstein7], using a touchdown PCR protocol comprising an initial annealing temperature of 70 °C, which was reduced by 1 °C per cycle over 15 cycles and followed by 20 cycles of amplification at 55 °C [Reference Khan4]. Characterization of Tn1546 elements was further achieved using primer pairs primers vanS1.seq (5′-TGCTCGTTCTTCCGATAC-3′), H1.seq-i (5′-AGCGCAAGAAGAATAGAGGC-3′); vanX1. seq (5′-ACGTTTGCGCTCCATCAT-3′); vanY1. seq-i (5′-CACACTTTCTTGGCGAACAG-3′); IS1216VD (5′-TGGGATTCCCAATAATACC-3′); IS1216VA (5′-GGAAAGCAATTTCAGCAG-3′); IS1251.f (5′-GC CCACTTGAATGTCTGA-3′); and IS1251.r (5′-GGATGCATTCTTGGGCTGTA-3′). PCR amplicon sizes were compared with those calculated theoretically using the Tn1546 reference sequence BM4147 (GenBank accession no. M97297) and the location of insertion sequences within the transposons was determined by DNA sequencing of relevant PCR products using the primers above.

The majority (23/26) of the isolates were recovered from patients in the adult intensive care unit (AICU), whose underlying conditions included cancer (n=7) and septic shock (n=2) among others. Five (19%) of the 26 VRE isolates were of the VanA phenotype and vanA gene-positive, with 20 (77%) isolates of the VanB phenotype and vanB gene-positive, a single VanA phenotype isolate was positive for both vanA and vanB genes by PCR. All 26 isolates were resistant to vancomycin, ampicillin and ciprofloxacin, and all but one (isolate 37) were susceptible to linezolid and quinupristin/dalfopristin (Fig. 1). From Table 1, it can be seen that nine rectal swabs were positive for vancomycin-resistant E. faecium, and were identified as part of the ICU screening process for all patients upon admission. The remaining 17 isolates were obtained from clinical samples as part of work-up for patients with fever and clinical signs of sepsis. In these cases, treatment targeting VRE was initiated accordingly. Both routine screening and clinical samples contained both vanA- and vanB-positive E. faecium isolates from both of the major genetic clusters identified. This finding points to a possible source for the vancomycin-resistant E. faecium infections, i.e. isolates were most probably brought into the hospital by patients, particularly as the King Fahad Specialist Hospital in Dammam is a referral hospital, with patients admitted to the ICU usually arriving from other ancillary hospitals and where patients have usually been exposed to antibiotics for at least 1 week prior to admission. Further, environmental samples obtained as part of outbreak investigations identified no environmental source of these resistant E. faecium. Microbial analysis of drains showed that the most common co-pathogens were E. coli and Klebsiella spp., although in 2007 these organisms [including extended spectrum β-lactamase (ESBL)-producing variants] were not a big problem.

Fig. 1. PFGE patterns obtained from 26 vancomycin-resistant E. faecium isolates originating from the King Fahad Specialist Hospital, Damman, Saudi Arabia between 2006 and 2007. Superscript letters indicate Tn1546 lineage (based on reference 9). AMP, Ampicillin; CIP, ciprofloxacin; TET, tetracycline; Q/D, quinupristin/dalfopristin; LZD, linezolid; R, resistant; S, susceptible. Numbers in circles represent the two major genetic clusters observed.

PFGE of the 26 E. faecium isolates (Fig. 1) showed the presence of two major clusters (isolates 42–35 and 11–22), based on an approximate similarity cut-off value of 90%, as well as a single unrelated genotype (isolate 37). Resistance to tetracycline was a distinguishing feature between the two clusters in the majority of isolates except for one (isolate 30). The clusters were not associated with Van A or Van B phenotypes or year of isolation. However, all representatives of cluster 2 originated from the AICU whereas cluster 1 was more widespread in the hospital. Figure 1 shows that eight of the 10 isolates typed by MLST were of sequence type (ST) 78 and seven of these grouped in cluster 1. Single isolates of ST290 and a novel type ST727 were also identified; the latter sequence type showed a single allelic difference from previously described types (ST16 and ST182). The three sequence types found here are associated with the global clonal cluster (CC) 17 which is often associated with outbreaks of hospital infections worldwide [Reference Leavis, Bonten and Willems8].

Of the six isolates positive for vanA, Tn1546 PCR amplicon size analysis showed that two (isolates 26 and 29) contained no major insertions or deletions compared to the reference Tn1546 BM4147 sequence. However, in four vanA isolates (30, 34, 37, 39), both insertion elements IS1251 and IS1216V were found (Fig. 2), with the IS1251 element being associated with a duplication of an ATAATTTT motif, as described previously [Reference Willems9]. Moreover, although related to a previously described group F Tn1546 lineage [Reference Khan2, Reference Willems9], the resultant Tn1546 variant had not been previously described in the literature. The presence of two new Tn1546 lineage types in E. faecium isolates from Riyadh isolated between 2000 and 2003 has previously been noted [Reference Khan4]. Interestingly, these workers also found a new Tn1546 type vanA E. faecium associated with nosocomial infections, which suggests that there may be yet more vanA Tn1546 variants circulating in Saudi Arabian hospitals. The question remains as to how important these different lineages are in epidemiologically linking endemic and epidemic nosocomial infections of VRE in Saudi hospitals, as well as the relationship between VanA Tn1546 variants and clinical resolution of infection with successful antibiotic therapy in Saudi Arabia.

Fig. 2. Genetic map of Tn1546 types associated with six VanA E. faecium isolates originating from Dammam, Saudi Arabia. The unbroken black lines represent reference Tn1546 sequence BM4147. Unfilled block arrows represent the positions of genes and open reading frames (orf1 and orf2). Filled boxes represent IS elements and arrows their direction of transcription. The positions of the first nucleotide upstream and downstream of the insertion sequence are shown. Group definitions (right) are based on those published by Willems et al. [Reference Willems9]. Numbers in parentheses indicate isolates belonging to each Tn1546 grouping. * The new Tn1546 lineage type F3 is closely related to the previously described Group F1 Tn1546 lineage.

Previous reports have shown that nosocomial infections due to enterococci are not new to Saudi Arabian hospitals, for example Tayfour et al. [Reference Tayfour, Al-Ghamdi and Al-Ghamdi10] reported finding two E. faecium, four E. faecalis and 13 Enterococcus spp. in a total of 366 surgical site infections over 1 year (2003–2004) at the King Fahad Hospital, Al-Baha. Further, a more recent publication from a hospital in the Riyadh area in 2009–2010 also documented the emergence of VRE (3·9% in 206 enterococcal isolates) and increased rates of multidrug resistance [Reference Salem-Bekhit11]. Consistent with our observations, Al-Otaibi et al. [Reference Al-Otaibi, Kambal and Baabbad12] found that a significant number of patients presenting with enterococcal bacteraemia between 2001 and 2002 at the King Khalid University Hospital, Riyadh (n=60), had previously stayed in an ICU, and our results also show the significance of these units in Saudi nosocomial enterococcal infections. Indeed at that time, vancomycin and imipenem consumption was not controlled, but as a result of the increased cases of VRE observed, an antibiotic stewardship programme was initiated and guidelines for treatment were adopted within the King Fahad Specialist Hospital, Dammam.