Comparison of Two Commercial PCR Methods for Methicillin-Resistant Staphylococcus aureus (MRSA) Screening in a Tertiary Care Hospital

Nose/throat-swabs from 1049 patients were screened for MRSA using CHROMagar MRSA, LightCycler Advanced MRSA, and Detect-Ready MRSA. Results were compared to the CHROMagar MRSA results, which was set as reference system. MRSA was detected in 3.05% of the patients with CHROMagar MRSA. LightCycler MRSA Advanced showed a higher clinical sensitivity (84.38%) than Detect-Ready MRSA (57.69%).The negative predictive values were high for both tests (>98%). The specificity and the positive predictive value were higher for the Detect-Ready MRSA test than for the LightCycler MRSA test (99.59% and 78.95% versus 98.52% and 64.29%). For routine screening LightCycler MRSA Advanced proved to be more efficient in our clinical setting as the clinical sensitivity was much higher than the sensitivity of Detect-Ready MRSA. CHROMagar MRSA detected more MRSA positive samples than both PCR methods, leading to the conclusion that the combination of PCR with cultural screening is still the most reliable way for the detection of MRSA. LightCycler MRSA Advanced was faster and needed less hands-on time. The advantage of Detect-Ready MRSA was the additional identification of methicillin-sensitive S.aureus (here in 34.63% of the samples), an information which can be possibly used for reducing the risk of postoperative infections in surgical patients in future.

However, screening methodology is crucial for detecting MRSA and no method is tested the most sensitive, specific or cost effective, so far. Inexpensive culture methods have the disadvantage of a swab-to-result time of up to 48 hours and different commercially available media show even differing accuracy values [35].
More expensive molecular-based test systems show more interesting turn-over-times but only few studies estimating accuracy values in a routine diagnostic setting are available so far [36]. Data of the Detect-ReadyH MRSA Kit concerning these questions are still lacking.
In our study we compared the Detect-ReadyH MRSA Kit (MDI, Kent, UK) with the standard methods used in our laboratory, the LightCyclerH MRSA Advanced Test (Roche, Mannheim, Germany) and the CHROMagar MRSA II (BD, Heidelberg, Germany). Our screening comprised all patients at admission to our emergency department or two intensive care units in accordance with our routine MRSA screening management.

Settings and specimen collection
The study was conducted at the Kliniken der Stadt Köln, a 1500 bed tertiary care facility and university affiliated teaching hospital in the City of Cologne, Germany. During a four month period from August to November 2011 all patients admitted to two intensive-care units or the emergency department of the hospital were screened for MRSA and included into the study, leading to 1049 patients included. Specimens were collected by the nursing staff using double-headed swabs with amies gel (Copan, Italy). In each patient a combined swab was taken from the throat and both nares, the most important sites for MRSA colonization [15,16,17,18]. Technically, the throat was swabbed rotating against the mucosa using a double-headed swab which was then used for swabbing both nares in the same way. The aim of the study was to evaluate two diagnostic assays for potential routine use, thus no ethical vote was required. The LightCyclerH Advanced MRSA Assay (Roche) was performed as the routine diagnostic assay in our laboratory, whilst the Detect-ReadyH Assay (Molecular Detection Inc) was performed in addition during the study period. Therefore verbal informed consent was sufficient. All procedures were performed according to the declaration of Helsinki in its present form.

Laboratory processing and culture methods
The BBL CHROMagar MRSA II (BD, Heidelberg, Germany) was used as screening agar for MRSA [19]. Additionally, Columbia-Agar with 5% sheep blood (BD) was used as control for bacterial growth and for identification of methicillin-sensitive S. aureus (MSSA) (Figure 1). In the laboratory both swabs of the double-headed swabs were firstly separated and each part was used for inoculating one MRSA selective agar and one Columbia agar consecutively. The two parts of the swab were then used for DNA-extraction for the one or the other PCR method, one for the LightCyclerH MRSA Advanced Test, one for the Detect-ReadyH MRSA Kit. The culture plates were incubated at 36uC and evaluated after 24 and 48 hours. Mauve-coloured colonies onto the selective chromogenic agar were confirmed to be S. aureus using an agglutination test for the simultaneous detection of the fibrinogen affinity antigen (clumping factor), protein A, and the capsular polysaccharides of S. aureus (Pastorex Staph-Plus, Bio-Rad, Munich, Germany). In case of agglutination the culture was sent to the routine diagnostics microbiology laboratory serving the Kliniken der Stadt Köln for species identification and antibioticresistance testing via the Vitek2 system (bioMerieux).
In case of the absence of typical (mauve coloured) colonies on the chromogenic medium and colonies suspected to be S.aureus on the Columbia agar, these colonies were analysed in the same way.

PCR methods
The LightCyclerH MRSA Advanced Test targets the integration site of the SCCmec cassette into the S. aureus chromosome (SCCmec:orfX junction). The Detect-ReadyH MRSA Panel Kit simultaneously targets three MRSA DNA regions, namely the mecA gene, the nuc gene and SCCmec:orfX junction. In this way the assay is able to identify MRSA, MSSA and Coagulase negative staphylococci (CoNS) with the help of integrated software considering the different amplification rates. CoNS are assumed if the mecA gene is detected whereas the nuc gene and the SCCmec:orfX junction remain undetectable. Both, the Light-CyclerH MRSA Advanced Test and the Detect-Ready MRSA Panel Kit (Rotor-Gene compatible version, MDI, Kent, UK) were performed according to the manufacturer's instructions. DNA was extracted using the LightCycler Advanced Lysis Kit (Roche) for the LightCycler MRSA Advanced Test and the Detect-Ready MRSA Swab Lysis Kit (MDI) for the Detect-Ready MRSA Panel Kit according to the manufacturer's instructions.
Each PCR method contains an Internal Control (IC) to exclude or detect PCR inhibition and to monitor reagent integrity. The IC must be detected positive in all MRSA negative specimens otherwise the specimen is automatically set as ''invalid'' by the program software.

Result definitions
A true positive (TP) result was defined as positive result in the PCR and the corresponding culture. A true negative (TN) result was defined as a negative result in the PCR and the corresponding culture. A sample positive in the PCR and negative in the corresponding culture was defined as false positive (FP), a sample negative in the PCR and positive in the corresponding culture was defined as false negative (FN).

Strain typing
Cultured MRSA isolates from samples with negative PCR results were reanalyzed by PCR. Isolates with repeated negative result were genotyped with Identibac StaphyType (Clondiag, Jena, Germany). The analysis was conducted by Alere Technologies GmbH (Jena, Germany). In addition the SCCmec type was determined according to the method of Boye et al. [20].

Data analysis
The results of the LightCyclerH MRSA Advanced test and the Detect-ReadyH MRSA test were compared to the corresponding CHROMagar MRSA plates. The clinical sensitivity, specificity, and the negative and positive predictive values (NPV and PPV) were calculated for each PCR method. Samples with invalid PCR results were excluded from the calculations.  (Table 1).

Patient profiles and culture results
With the Detect-ReadyH MRSA Kit only 1028 samples could be processed. The test identified 19 samples as MRSA positive, 15 of those samples were positive on CHROMagar and therefore classified as true positive, the other four samples were classified as false positive. The other 989 samples were tested as MRSA negative; in detail 173 were identified as MSSA, 183 as a mixture of MSSA and coagulase negative staphylococci (CoNS), 324 as CoNS, and 309 as negative ( Table 2). Ten of the MSSA and CoNS positive and one of the CoNS positive samples showed MRSA growth on CHROMagar and were classified as false negative in the PCR (Table 1). Samples with MRSA growth on CHROMagar and negative PCR results were reanalyzed. The cultured MRSA isolates retrieved from the CHROMagar plates were used as templates for the respective PCR. In this approach three of five isolates from samples with false negative results in the LightCyclerH Advanced MRSA-PCR were tested positive, one negative. One cultured sample was missing and could not be reanalyzed. The isolate tested negative was MRSA positive in the than performed Detect-ReadyH MRSA Test. Concerning the Detect-ReadyH MRSA PCR, the analysis of the isolates from 11 false negative PCR samples resulted in eight MRSA positive and three negative PCRs (one negative, one CoNS, one MSSA+CoNS). All three negative Figure 1. Clinical experimental design. Nasal/throat specimens were tested by Detect-ReadyH MRSA Kit and by the LightCyclerH MRSA Advanced Test. For testing, the double-headed swab was separated. One swab head was processed for directly plated culture on BD BBL TM CHROMagar TM MRSA Medium II, BD Columbia Agar with 5% sheep blood plates, and the LightCyclerH MRSA Advanced Test. The other swab was used for the assay with Detect-ReadyH MRSA Kit and directly plated culture on BD BBL TM CHROMagar TM MRSA Medium II, and BD Columbia Agar with 5% sheep blood plates. MRSA positive colonies onto the selective chromogenic agar were confirmed to be S. aureus using an agglutination test. doi:10.1371/journal.pone.0043935.g001 samples were tested positive in the LightCyclerH MRSA Advanced Test. The four samples that failed to be detected in one of the PCR assays were sent to Alere Technologies GmbH for genotyping. All four MRSA strains belonged to the ST5/ST225-MRSA-II strain. SCCmec typing revealed that all had the SCCmeccassette type II.

Discussion
MRSA is still a growing problem in health care settings leading to increased costs and patient risks. Identification of colonized patients is the first step in the containment of MRSA spreading. Different diagnostic tests are available for the identification of colonized patients so far.
Factors to consider for the choice of a MRSA screening platform include sensitivities, specificities, turnaround time, costs, and ease of interpretation, which had been shown to vary considerably [35]. MRSA screening methodology is already widely in use even though accuracy values of the respective tests are not all sufficiently tested in a clinical routine setting or even in comparison to already better evaluated tests.
Here, we compared the Detect-ReadyH MRSA Kit (MDI) with the methods used routinely for MRSA screening in our hospital (LightCyclerH MRSA Advanced Test (Roche) and CHROMagar MRSA (BD)).
Costs and turnaround time play an important role in the decision which assay is the test of choice. Reagent and instrument costs are much higher and turnaround times shorter for the PCR assays. LightCyclerH MRSA Advanced Test had the shortest turnaround time with less than 2 hours. Processing of the Detect-Ready MRSA Test was finished within 5 hours requiring a considerable hand-on time. Results from cultures were available after 24 hours of incubation at the earliest (94.78% of MRSA positive cultures could be identified after 24 hours (Table 1)). Interpretation of positive results was easy to perform for both PCR platforms, with software programs providing clearly arranged result lists after every PCR run Both test methods are based on the detection of the SCCmec-orfX junction. The staphylococcal cassette chromosome mec (SCCmec) is the carrier of the resistance-gene mecA. It integrates into the S. aureus genome in the region of the open reading frame orfX which is specific for S. aureus [21]. Therefore, an amplification product is obtained in SCCmec carrying S. aureus but not in SCCmec carrying CoNS. SCCmec PCRs have been reported to have high sensitivities and specificities and are used in several commercial available MRSA detection kits [22,23,24,25,27].
In this study both tests had high specificities (LightCyclerH MRSA Advanced Test: 98.52%, Detect-ReadyH MRSA Kit: 99.59%). LightCyclerH MRSA Advanced Test was more sensitive for the detection of MRSA (84.38%) than Detect-ReadyH MRSA Kit, but both tests had poorer sensitivities in our real-life study setting than reported in previous studies [36]. Especially the Detect-ReadyH MRSA Kit had a surprisingly low clinical sensitivity (57.69%). This might be due to a software/cut off problem of the Detect-Ready MRSA test, as the software calculates automatically if MRSA or a mixed population of staphylococci is present in the samples (10 of the culture positive samples were identified as MSSA and CoNS and one as CoNS). After re-analysis of the false negative results using the cultured strains as PCR template, three of the five false negative Light-CyclerH Advanced MRSA Test results and eight of the 11 false negative Detect-ReadyH MRSA results tested positive, which would increase the sensitivities so 93.75% and 88.46%, respectively, if the samples were tested positive in the first PCR.
The detection of false positive and false negative results could have had several reasons. Recent studies reported a S. aureus strain (LGA251) resistant to methicillin but negative for mecA [28,29]. This strain harbours a divergent mecA homologue with a different organization than other SCC elements leading to a false negative result. Another study from Denmark [30] revealed that a specific common SCCmec clone was frequently not detected in a commercial MRSA assay leading to the conclusion that local diversities play an important role in the performance of MRSA assays, as undetectable low prevalence strains could become widespread among S. aureus. In our study, most of the samples which were false negative in the PCR proved to be positive in a second PCR approach using the isolated MRSA cultures. In those cases most probably an inoculum effect (plates were inoculated prior to MRSA PCRs), low concentration on the epithelium of MRSA, or inhibitory PCR effects seem to be the cause for retrieving false negative results in the PCR assays. Four samples were tested negative again in the second PCR approach, but all samples were MRSA positive in the conquering assay. Genotyping revealed that all strains belonged to the ST5/ST225-MRSA strain with a SCCmec-cassette type II, a common type in this region. Future sequence analysis of these four strains would be interesting to determine why they could not be detected by the PCR assays. Minor changes in the sequence of the primer binding sites could be the reason. Maybe there is a MRSA subpopulation in our region which is not detectable with certain commercial assays comparable to the MRSA clone in the above mentioned Danish study [30].
The reason for false positive PCR results can be orfX genes in CoNS homologue to the S. aureus variant or SCCmec cassettes lacking mecA [31,23,32,33]. To avoid false positive results in this case, the Detect-ReadyH MRSA Test detects a second S. aureus specific marker (nuc) and the mecA gene additionally to the SCCmec-orfX amplicon. Overall, this approach enables the differentiation between MRSA and MSSA in addition. This may be important in surgical patients, as preoperative detection of MSSA nasal carriage and decolonization of the patient could reduces the risk of surgical site infections due to S. aureus [34]. 95% of the cultured MSSA were detected by the Detect-Ready MRSA Assay. The assay detected MSSA in nearly twice the number of samples than the culture method. This may either be due to false positive PCRs, or to a lack of selectivity of the here uses culture method for S. aureus. Columbia agar enables the growth of numerous bacteria which can overgrow small amounts of MSSA which are then not detected. A comparison of a selective MSSA agar with the Detect-Ready Assay would be an interesting approach for the future but could not be conducted in scope of this study.
In this study PCR based methods were compared to the direct plating on chromogenic agar. Since not performed under routine conditions neither, no broth-enriched culture was used in the study. Thus, our study is limited by the possibility that brothenrichement would have lead to more culture positive results. PCR results defined here as false positive could also reflect a detection of non viable MRSA. The reason for choosing the routine culture method as standard was our demand to focus on patients colonised with a considerable amount of still viable MRSA, which determines its transmissibility.
Our study is further limited by the fact, that the false positive samples (amplicons) could not be characterized by molecular methods to examine whether other genetic abnormalities lead to a positive PCR result. No additional data concerning former antibiotic treatment was retrieved for the patients, so it could not be excluded that the detection of false positive samples was due to prior antibiotic treatment, nor to what extent. In contrast to our results, Peterson et al [26] showed, a higher sensitivity of 95% of the the LightCyclerH MRSA Advanced Test in comparison to direct culture, possibly due to an exclusion of patients with antibiotic therapy active against MRSA.
In conclusion, our data show that the LightCyclerH MRSA Advanced Test demonstrated a better clinical sensitivity compared to the Detect-ReadyH MRSA Kit. We would recommend additional cultural testing in a clinical setting to close the diagnostic gap and to avoid false results. With the CHROMagar MRSA the majority of the positive results (95%) were achieved already after 24 hours, demonstrating this culture based test as a relatively fast, cheep and reliable screening method in situations where no immediate results are needed.,.