Clinical importance of coagulase-negative staphylococci (CNS)

CNS are differentiated from the closely related but more virulent Staphylococcus aureus by their inability to produce free coagulase enzyme [Reference Bannerman, Peacock, Murray, Baron, Jorgensen, Pfaller and Landry1]. There are currently more than 40 recognized species in the group [Reference Euzeby2], many of which form part of the normal flora of healthy human skin and mucus membranes; those most frequently associated with infections in humans are S. epidermidis, S. haemolyticus and S. saprophyticus. Other species occasionally isolated from clinical specimens include S. hominis, S. warneri, S. capitis, S. simulans, S. cohnii, S. xylosus, S. saccharolyticus and S. lugdunensis [Reference Rogers, Fey and Rupp3–Reference McCann, Gilmore and Gorman5].

Within the last few years, CNS have been increasingly recognized to cause clinically significant infections and are particularly associated with the use of medical devices such as intravascular and peritoneal dialysis catheters, cerebrospinal fluid shunts, prosthetic heart valves, and other plastic indwelling devices. Most infections result from the inoculation of organisms on the skin at the time of device implantation in the hospital [Reference Rogers, Fey and Rupp3–Reference McCann, Gilmore and Gorman5]. The ability to infect patients carrying medical devices is primarily related to the capacity of some CNS to form biofilms where the bacteria adhere to the inert surface of the plastic devices, triggering a change in microbial behaviour such as by quorum sensing, in which the outer layer of cells protect the interior community [Reference Davey and O'Toole6] and confer increased resistance to antimicrobial agents [Reference Stewart and Costerton7].

S. epidermidis is the most commonly isolated CNS and was first identified in microbiological cultures in 1880. In contrast to S. aureus, which is often a cause of serious and fatal infections, S. epidermidis has a predisposition to cause chronic and recurring infections [Reference Rogers, Fey and Rupp3, Reference Huebner and Goldmann4, Reference Gill8, Reference von Eiff, Peters and Heilmann9]. The identification of this species as an aetiological agent of infection is sometimes difficult as most infections associated with it are characterized by non-specific clinical manifestations and as part of the normal skin flora it is frequently found as a contaminant in cultures. Nevertheless, S. epidermidis is today recognized as an important cause of nosocomial infections and has been ranked the third most common cause of nosocomial sepsis in some studies [Reference Pfaller10, Reference Emory and Gaynes11].

The second most commonly isolated species of CNS, S. haemolyticus is part of the resident flora in the axilla, perineum, and inguinal areas of humans. It is sometimes a cause of sepsis and has been repeatedly implicated in bloodstream infections, peritonitis, and infections of the urinary tract, wounds, bone, and joints [Reference Pfaller10, Reference Emory and Gaynes11]. Most strains of S. haemolyticus exhibit a highly antibiotic-resistant phenotype [Reference Diekema12].

Methicillin resistance and SCCmec

Methicillin-resistant S. aureus (MRSA) isolates were first reported in 1961 [Reference Barber13], and by the mid-1980s it was recognized as a significant problem in the USA, with several outbreaks documented in tertiary-care teaching hospitals [Reference Archer and Mayhall14, Reference Rhinehart15]. Today, MRSA strains are widespread in all parts of the world and represent a serious burden of infection in many countries, not only in hospitals but also in the wider community. Intrinsic methicillin resistance in staphylococci is due to the expression of a modified penicillin-binding protein PBP2a (PBP2′) encoded by the mecA gene and located on the mobile element staphylococcal cassette chromosome (SCCmec), a genomic island of variable size (range 21–67 kb) [Reference Katayama, Ito and Hiramatsu16] integrated at the 3′ end of the orfX [Reference Ito, Katayama and Hiramatsu17] gene located near the chromosomal origin of replication. For S. aureus, the element includes the repressor genes mec1 and mecR1 [Reference Katayama, Ito and Hiramatsu18–Reference Suzuki21].

Eight different SCCmec types (I–VIII) have been identified and are classified according to different sets of ccr (chromosome cassette recombinase) (ccrAB1, ccrAB2, ccrAB3, ccrAB4, ccrC), which are genes that are responsible for both its chromosomal integration and excision [Reference Katayama, Ito and Hiramatsu16, Reference Zhang22]. Several SCCmec subtypes IIA–E [Reference Shore23], IVa–g [Reference Ito24–Reference Ma26], and SCC non-mec types have been reported [Reference Ito27–Reference Mongkolrattanothai31]. In addition, differences in the mec gene complex (classes A–E) [Reference Ito27, Reference Chongtrakool32–Reference Okuma35] have been described. It should be noted that the majority of the work on the characterization of mecA and its vector, SCC, has centred on MRSA, but this element is not exclusive to S. aureus as it has been found in other species of the genus.

The prevalence of methicillin resistance has been reported to be higher in CNS than in S. aureus, with rates ranging globally from 75% to 90% during the 1990s [Reference Diekema12]. S. epidermidis is the most frequently isolated of the methicillin-resistant (MR)-CNS [Reference Machado36–Reference Hanssen, Kjeldsen and Sollid40], which mirrors the frequency of both methicillin susceptible and resistant isolates in CNS species. Among human isolates, SCCmec has been described in S. epidermidis, S. haemolyticus, S. hominis, S. capitis, S. sciuri, S. warneri, and S. saprophyticus [Reference Machado36–Reference Higashide45] (Table 1).

Table 1. Distribution of SCCmec types found in coagulase-negative staphylococci, both in animals and in humans

NT, Non-typable; SCCmec types in parentheses amplified for both elements.

It soon became evident that novel classes of the SCCmec element may exist in CNS as many reports described strains lacking the elements present in S. aureus or reported strains that amplified two elements, suggesting the presence of different elements of SCCmec [Reference Machado36, Reference Ruppé37, Reference Ibrahem39, Reference Hanssen, Kjeldsen and Sollid40]. As reported for MRSA, high levels of resistance to oxacillin or cefoxitin were linked to the presence of SCCmec type III in CNS [Reference Machado36, Reference Ruppé37]. Nevertheless, additionally to SCCmec type III, other types and subtypes have been described for CNS. For example, for S. epidermidis, types I, II (subtypes a, b), III, IV, (subtypes a, b, c, d) and V have been described as well as other putative different SCCmec elements characterized by amplification for two elements: for I and III, for III and IV, for II and V, and for III and V (Table 1).

For S. haemolyticus SCCmec types I, II, III, IV, V and, as described for S. epidermidis, new proposed elements (III variant, amplification for II and V). Additionally to S. epidermidis and S. haemolyticus, SCCmec types have been described for S. warneri, S. lentus, S. sciuri, S. xylosus, S. saprophyticus, S. hominis and S. capitis (Table 1).

MR-CNS in animals

MR-CNS has been reported in several species of animals, such as cattle, sheep, goat, pigs, chickens, dogs, cats, horses and Cope's grey treefrogs where a prevalence as high as 59% has been found [Reference Yasuda46–Reference Slaughter50]. In several of these studies, some SCCmec elements have been reported (Table 1). It has been suggested that the diversity of CNS species and the SCCmec elements found in isolates from several species of animals, some of them domestic or so-called food animals, and the demonstrated capacity of these bacterial species to cause suppurative disease in these animals (which undoubtedly increases the shed of microorganisms) render them a potential threat to humans and constitutes a potential risk from the consumption of foods of animal origin [Reference Lee51, Reference Burriel52].

Clonal diversity of MR-CNS

The classification of CNS in relation to the SCCmec elements has been achieved primarily by the use of the same strategies used for MRSA. However, as CNS can harbour SCCmec elements different from those in MRSA, the high diversity of these mobile elements poses a significant challenge for researchers. CNS are highly clonally diverse, the most noteworthy example being S. epidermidis, in which a high degree of genetic diversity has been described for the SCCmec element IVa [Reference Jamaluddin38]. Moreover, epidemiological studies using multilocus sequence typing (MLST) suggest that S. epidermidis isolates prevalent in the hospital environment differ from those causing community-acquired disease [Reference Ziebuhr53]. Lancastre et al. proposed that S. epidermidis clones could be defined by the combination of a carefully standardized pulsed-field gel electrophoresis (PFGE) protocol and identification of the SCCmec type [Reference Miragaia54].

Another study analysed representative isolates of S. epidermidis from diverse geographic and clinical origins, characterized by SCCmec and MLST, and reported the finding that nine epidemic clonal lineages are disseminated worldwide with one single clonal lineage comprising 74% of the isolates. They concluded that S. epidermidis has a population with an epidemic structure, in which clones have emerged as a result of high levels of recombination and evolved through the transfer of genetic mobile elements, including SCCmec [Reference Miragaia55].

Origin of methicillin resistance

Evidence suggests that acquisition of SCCmec elements in S. aureus by susceptible ancestors has taken place at different times and at different locations. The first MRSA strain emerged when a SCCmec element was integrated into the chromosome of a susceptible S. aureus strain, but the donor remains unidentified. The mechanisms may involve the action of recombinases, which are capable of striking and incorporating the element into the bacterial chromosome [Reference Katayama, Ito and Hiramatsu16]. Although the mec origin remains unknown, it has been suggested that SCCmec can shift between both coagulase-positive and coagulase-negative staphylococci and mecA-positive CNS may act as potential SCCmec donors accounting for the rise in new MRSA clones [Reference Musser and Kapur56, Reference Hanssen and Ericson Sollid57].

In general, CNS are thought to comprise a reservoir of resistance genes for S. aureus and some features that support this hypothesis are:

1. (i) the higher frequency of SCCmec elements in CNS than in S. aureus;

2. (ii) the evidenced horizontal transfer of resistance genes from CNS to S. aureus [Reference Archer and Scott58–Reference Berglund and Söderquist64];

3. (iii) genome flexibility in S. epidermidis which may lead to the acquisition of some transferable resistance elements [Reference Kozitskaya65];

4. (iv) the species-independent conservation of some ccr elements [Reference Hanssen and Sollid66];

5. (v) CNS are more likely to contain different ccr complexes [Reference Hanssen and Sollid66];

6. (vi) the great diversity of SCC elements found in CNS (Table 1),

7. (vii) the high variation in the prevalence of MRSA according to geographic region and the high prevalence of MR-CNS regardless of the geographic area [Reference Diekema12].

New SCCmec types or elements are being continuously discovered around the world in different species and it will be necessary to determine if these diverse elements have been present for a long time or if new clones continue to arise in a microevolutionary process.

Detection of methicillin resistance

Detection of methicillin resistance in CNS is problematic, as expression of the mec gene is heterotypic, being constitutive in many strains, whereas in others it is inducible. Phenotypic methods have been used to detect MR-CNS because they can be readily performed in most laboratories. According to the Clinical and Laboratory Standards Institute (CLSI), the cefoxitin diffusion disk is the preferred method for detection of mecA-mediated resistance in CNS (except for S. lugdunensis); cefoxitin is used as a surrogate for detection of oxacillin resistance [67]. When compared to the molecular detection of methicillin resistance, the sensitivity and specificity of the phenotypic tests for cefoxitin disk diffusion is 94·9% and 97%, respectively, with detection of resistance dependent on species [Reference John68].

The performance of the automated Vitek 2 system for identifying mecA-positive staphylococci was comparable to PCR and the CLSI disk diffusion method (sensitivity and specificity of 94·6% and 93·5%, respectively). However, its performance was species dependent as results were poor in tests with S. cohnii, S. hominis and S. saprophyticus [Reference John68].

Identification of CNS

Several commercial kits have been used in diagnostic laboratories with varying success for the identification of CNS species including API Staph ID 32 and Vitek 2 (bioMérieux, France); Staph-Zym test (Rosco, Denmark); Staphylo LA test (Wako, Japan); Sensititre (TEK Diagnostic Systems Inc., USA); MicroScan (Dade Behring, USA); Phoenix and Crystal GP systems (Becton Dickinson, USA). The sensitivity of these systems is often not better than fair, particularly for species other than S. epidermidis and S. haemolyticus. For example, MicroScan, Vitek 2, and Crystal GP correctly identified 82·5%, 87·5% and 67·5% of CNS, respectively in human clinical isolates in which S. epidermidis and S. haemolyticus predominated [Reference Kim69]. Furthermore, the new low-inoculum mode of the Phoenix system correctly identified 90·5% of isolates and accuracy was satisfactory for S. epidermidis, S. saprophyticus and S. haemolyticus, but was notably lower for other species such as S. hominis (69·6%) [Reference Brigante70]. In general, for most kit systems a lower sensitivity has been reported for the identification of CNS species when S. epidermidis and S. haemolyticus are infrequent in the sample. For example, API Staph ID 32 correctly identified 41% of the CNS isolates from veterinary sources and showed complete (100%) sensitivity for S. epidermidis, but lower values were observed for S. xylosus (87%), S. chromogenes (37%) and S. warneri (15%) [Reference Sampimon71]. It should be noted that the prevalence and diversity of the SCCmec in CNS has been studied mainly by the use of phenotypic systems, mainly by API ID 32 Staph (Table 2). The results obtained using such systems should be interpreted with caution in the light of the sometimes low specificity for species other than S. epidermidis and S. haemolyticus. As a result many laboratories are increasingly relying on DNA sequencing for species identification [72], while for identification of CNS species, sequence data of housekeeping genes such as rpoB [Reference Drancourt and Raoult73, Reference Mellmann74], hsp60 [Reference Kwok75], dnaJ [Reference Shah76], tuf [Reference Capurro77], sodA [Reference Poyart78] and 16S RNAr [Reference Becker79] is recommended.

Table 2. Distribution of identification of coagulase-negative staphylococci species and the method used for identification

NT, Non-typable.

Two-year surveillance of MR-CNS in a tertiary-care hospital in Guadalajara, Mexico

To determine the distribution of the SCCmec in MR-CNS and its relation to antimicrobial resistance in a tertiary-care hospital in Guadalajara, Mexico, we studied all the MR-CNS collected from blood samples in 2007 and 2008 (n=45). CNS isolates were identified by conventional biochemical tests [Reference Bannerman, Peacock, Murray, Baron, Jorgensen, Pfaller and Landry1] and via the use of API strips. Resistance to methicillin was determined for all isolates by the cefoxitin disk test [67] and by PCR [Reference Zhang80]. MR-CNS with uncertain biochemical patterns, including species other than S. epidermidis and S. haemolyticus were further examined by amplification and sequencing of the 16S rRNA gene using primers F-5′-GGIACTGAGACACIGICCIIACTCCT-3′ and R-5′-TTCCIITACIGITACCTTGTTACGACTT-3′ [Reference Pei81].

All sequencing was performed at the Instituto de Biotecnología de la Universidad Nacional Autónoma de México.

Preparation of template DNA and SCCmec, ccr and mec class typing were performed as previously described [Reference Zhang80] and susceptibility testing was performed by broth microdilution using panels from Sensititre (TEK Diagnostic Systems Inc.) according to the manufacturer's instructions. For all assays, the control organism (S. aureus ATCC 29213) was used and susceptibility breakpoints were as recommended by CLSI [67].

As expected, the most frequently identified MR-CNS species was S. epidermidis (n=34) followed by S. haemolyticus (n=7) and S. hominis (n=4) (Table 3).

Table 3. Distribution of SSCmec cassettes among methicillin-resistant isolates from bloodstream infections within a 2-year period and minimal inhibitory concentrations (MIC) of antimicrobials (μg/ml)

AMC, Amoxicillin-clavulanic acid; Cfp, cefepime; Cxm, cefuroxime; Mer, meropenem; Imp, imipenem; Clin, clindamycin; Cip, ciprofloxacin; Nor, norfloxacin; Nit, nitrofurantoin; Sxt, trimethoprim–sulfamexazole; Clo, chloramphenicol; Rif, rifampicin; Tet, tetracycline; NT, non-typable.

All isolates had a MIC of ⩽4 for teicoplanin, ⩽2 for vancomycin, >32 for cefotaxime, >4 for erythromycin and >16 for cephalothin.

Of the methicillin-resistant S. epidermidis isolates examined, 11-harboured SCCmec type III, 12 SCCmec type IVa, three showed amplification products for II and V elements and one strain was positive for both V and III SCCmec elements. Seven strains were non-typable (Table 3).

Among the S. haemolyticus isolates, type III element was detected in two isolates and two others amplified both types II and V. For S. hominis, one strain typed for the III SCCmec element and three others were non-typable. Multiple ccr elements were found in all three species. MRSA clinical isolates that had been previously characterized in this hospital were SCCmec type II [Reference Echániz-Aviles82]. Interestingly, the SCCmec type II alone was not detected among MR-CNS strains. It should be noted that the MRSA study represented clinical isolates collected from 1999 to 2003.

An extremely high level of drug resistance was found for all isolates (Table 3). Lower resistance rates were observed for cefuroxime, gentamicin, nitrofurantoin and rifampicin; regardless of the SCCmec detected (the difference was not statistically significant).