A Rapid and High-Throughput Screening Approach for Methicillin-Resistant Staphylococcus aureus Based on the Combination of Two Different Real-Time PCR Assays

Abstract

Methicillin-resistant Staphylococcus aureus (MRSA) is an important pathogen that has been responsible for major nosocomial epidemics worldwide. For infection control programs, rapid and adequate detection of MRSA is of great importance. We developed a rapid and high-throughput molecular screening approach that consists of an overnight selective broth enrichment, followed by mecA, mecC, and S. aureus-specific (SA442 gene) real-time PCR assays, with subsequent confirmation using a staphylococcal cassette chromosome mec element (SCCmec)-orfX-based real-time PCR assay (GeneOhm MRSA assay) and culture. Here, the results of the screening approach over a 2-year period are presented. During this period, a total of 13,387 samples were analyzed for the presence of MRSA, 2.6% of which were reported as MRSA positive. No MRSA isolates carrying the mecC gene were detected during this study. Based on the results of the real-time PCR assays only, 95.2% of the samples could be reported as negative within 24 h. Furthermore, the performance of these real-time PCR assays was evaluated using a set of 104 assorted MRSA isolates, which demonstrated high sensitivity for both the combination of mecA and mecC with SA442 and the BD GeneOhm MRSA assay (98.1% and 97.1%, respectively). This molecular screening approach proved to be an accurate method for obtaining reliable negative results within 24 h after arrival at the laboratory and contributes to improvement of infection control programs, especially in areas with a low MRSA prevalence.

INTRODUCTION

Methicillin-resistant Staphylococcus aureus (MRSA) is an important pathogen that has been detected throughout the world (1–3). MRSA has been responsible for major hospital epidemics, both regionally and globally (3, 4). The introduction of the so-called “search and destroy” (S&D) policy in countries such as the Netherlands and Norway has proven to be of great importance in preventing the spread of MRSA, since the prevalence of MRSA in these countries is very low (5, 6). Moreover, this strategy has been shown to save money and lives, which illustrates even more the impact of the S&D policy (7). Rapid diagnostic testing for MRSA is helpful for the S&D policy and will reduce the number of isolation days.

The Dutch guideline on the detection of MRSA recommends the use of a MRSA screening agar after overnight incubation with a nonselective broth, followed by molecular confirmation (8). With this method, it takes at least 72 h to obtain a (negative) result. Although culture remains the gold standard for detection of MRSA, several other approaches have been developed to detect the presence of MRSA. As early as 1991, the evaluation of a PCR targeting the mecA gene was published (9). In combination with a species-specific identification test for S. aureus, MRSA can be distinguished from methicillin-resistant coagulase-negative staphylococci (MRCNS), which also may contain the mecA gene. Such a species-specific PCR assay for S. aureus might be, for instance, a PCR targeting the nuc gene or the SA442 gene (10, 11). However, the combination of a mecA PCR and an S. aureus-specific PCR may result in false-positive results when used in nonsterile samples that can contain both methicillin-susceptible S. aureus (MSSA) and MRCNS. Therefore, Huletsky et al. developed a real-time PCR assay that is based on the integration site of the staphylococcal cassette chromosome mec element (SCCmec) cassette, containing the mecA gene, and the S. aureus-specific orfX gene (SCCmec-orfX) (12). Rapid diagnostic assays, such as the Xpert MRSA assay (Cepheid, Sunnyvale, CA, USA) and the GeneOhm MRSA assay (BD Diagnostics, San Diego, CA, USA), are based on the principles of Huletsky et al. and are widely used for rapid detection of MRSA (13, 14). Unfortunately, the SCCmec-orfX-based assays have proved not to be able to detect all MRSA isolates correctly. In the past, mecA dropouts have been reported, resulting in false-positive MRSA results (12, 15). Moreover, because of the high diversity in SCCmec-orfX sequences, correct detection remains a challenge (16, 17). The recent discovery of the mecC gene, which cannot be detected by both the mecA assays and the SCCmec-orfX-based assays, makes the detection of MRSA even more difficult (18). Since both the combination of mecA-mecC together with an S. aureus-specific real-time PCR assay and the SCCmec-orfX-based approach have their shortcomings when used as a single assay to detect MRSA in clinical samples, we decided to combine the two assays to acquire a rapid and high-throughput molecular MRSA screening tool.

Here, we present the results of our molecular screening approach for the detection of MRSA, which consists of an overnight broth enrichment, followed by mecA, mecC, and an S. aureus-specific (SA442) real-time PCR assay, with subsequent confirmation using an SCCmec-orfX-based real-time PCR assay and culture. The present report describes the results of this approach over a period of 2 years (2012 and 2013).

MATERIALS AND METHODS

Samples and DNA extraction.

All samples for MRSA screening from 10 January 2012 to 31 December 2013 were inoculated into a phenyl mannitol broth (PMB) containing ceftizoxim (5 μg/ml) and aztreonam (75 μg/ml) for overnight incubation (35 ± 2°C) (19). Nose and throat swabs from each patient were pooled, while perineal swabs and all other possible samples (such as wound swabs, urine, or sputa) were processed separately. After at least 16 h of incubation, the broths were mixed thoroughly, and 100 μl of broth was used for DNA extraction. In short, 100 μl broth was mixed well with 100 μl (1 U/μl) achromopeptidase (Sigma-Aldrich Co., St. Louis, MO, USA) and 20 μl internal inhibition control (IC) (phocine herpesvirus) (20). After incubation for 15 min at 37°C, followed by 5 min at 95°C and centrifugation (1 min; 14,000 rpm), the supernatant was used for detection of MRSA.

Real-time PCR assays.

Two internally controlled real-time PCR assays were performed: the first assay detects the S. aureus-specific SA442 gene, while the second targets the mecA gene and the mecC gene (together forming the SMM panel). The sequences of all the primers and probes are shown in Table 1. Both assays were simultaneously performed in about 1 h 40 min on an ABI 7500 system (Applied Biosystems, Foster City, CA, USA) in a 30-μl volume containing 10 μl supernatant, primers, and probes as shown in Table 1 and 1× TaqMan Universal PCR Master Mix (Applied Biosystems, Foster City, CA, USA). The real-time PCR conditions were 50°C for 2 min, 95°C for 10 min, and 45 cycles of 95°C for 15s and 60°C for 60s.

TABLE 1.

Primer and probe sequences of the SMM panel

Assay	Oligonucleotide	Sequence (5′→3′)a	Final concn (nM)	Origin
PCR assay 1	SA442-forward	CAATCTTTGTCGGTACACGATATTCT	333	Modification of Martineau et al. (11)
	SA442-reverse	CAACGTAATGAGATTTCAGTAGATAATACAAC	333
	SA442-probe 1	FAM-CACGACTAAATAAACGCTCATTCGCGATTTT-BHQ1	150	Developed by E. J. van Hannen
	SA442-probe 2	FAM-CACGACTAAATAGACGCTCATTCGCAATTTT-BHQ1	150
	IC-forward	GGGCGAATCACAGATTGAATC	333	Van Doornum et al. (20)
	IC-reverse	GCGGTTCCAAACGTACCAA	333
	IC-probe	CY5-TTTTTATGTGTCCGCCACCATCTGGATC-BHQ2	150
PCR assay 2	MecA-forward	GATTATGGCTCAGGTACTGCTATCC	333	Developed by E. J. van Hannen
	MecA-reverse	TTCGTTACTCATGCCATACATAAATG	333
	MecA-probe	VIC-CCCTCAAACAGGTGAATTATTAGCACTTGTAAGCA-BHQ1	250
	MecC-forward	GCAAGCAATAGAATCATCAGACAAC	333	Developed by E. J. van Hannen
	MecC-reverse	TCTTGCATACCTTGCTCAAATTTT	333
	MecC-probe	NED-CCGCATTGCATTAGCATTAGGAGCCA-BHQ2	100
	IC-forward	GGGCGAATCACAGATTGAATC	333	Van Doornum et al. (20)
	IC-reverse	GCGGTTCCAAACGTACCAA	333
	IC-probe	CY5-TTTTTATGTGTCCGCCACCATCTGGATC-BHQ2	150

^aFAM, 6-carboxyfluorescein; BHQ, black hole quencher; NED, 2′-chloro-5′-fluoro-7′,8′-fused phenyl-1,4-dichloro-6-carboxyfluorescein; VIC, 6-carboxyrhodamine.

A sample was suspected to be MRSA positive when the SA442 PCR resulted in a threshold cycle (C_T) value of 39 or lower in combination with a positive mecA or mecC PCR (C_T value, <45). Samples with a C_T value of >39 for the SA442 PCR were interpreted as negative, since an in-house validation showed that no S. aureus isolate was cultured when a C_T value of >38 was found (data not shown). Samples were interpreted as inhibited when the C_T value of the IC was >2 C_T values higher than the IC of a negative isolation control. For these samples, DNA extraction and the SMM panel were repeated with a 10-fold dilution of the broth. Interpretation of these diluted samples was performed as described above. Samples showing inhibition after dilution were cultured.

As illustrated in Fig. 1, in case of a suspected MRSA-positive PCR result in the SMM panel, a second real-time PCR assay (GeneOhm MRSA; BD Diagnostics, San Diego, CA, USA) was performed on the vast majority of samples. Only when a mecC signal in combination with the SA442 signal was detected or when the suspected MRSA-positive sample (a positive result in the SMM panel) was sent by a general practitioner, that is, when rapid results of MRSA detection were less urgent and costs could be reduced, was the broth subcultured directly without the second PCR step, as described below. The GeneOhm MRSA assay was performed on the Smartcycler system (BD Diagnostics, San Diego, CA, USA) by adding 3 μl supernatant to 25 μl of master mix provided with the kit, as instructed by the manufacturer. After a run of about 1 h, interpretation was performed by the software, and the results were reported as positive, negative, or unresolved (inhibition). Each sample that was either positive or unresolved was presented for culture. Moreover, samples negative in the GeneOhm MRSA assay that were suspected to contain livestock-associated MRSA (LA-MRSA), as stated by the applicant, were cultured.

FIG 1.

Schematic overview of the MRSA screening approach. GP, samples obtained from patients of general practitioners.

Culture and spa typing.

The broth was subcultured using a 1-μl loop onto a Brilliance MRSA2 (Oxoid Ltd., Basingstoke, United Kingdom) and 5% sheep blood agar (Oxoid) and analyzed after overnight incubation (35 ± 2°C). Suspected colonies were captured and identified as S. aureus using matrix-assisted laser desorption–ionization time of flight mass spectrometry (MALDI-TOF MS) (Bruker Daltonik GmbH, Bremen, Germany). Methicillin resistance was confirmed with a cefoxitin disc (30 μg) and Vitek2 (bioMérieux, Marcy l'Etoile, France). Each new MRSA isolate was sent to the Dutch National Reference Center (RIVM, Bilthoven, the Netherlands) to determine the spa type of the isolate.

Evaluation of the real-time PCR assays.

In order to assess the performance of both the SMM panel and the GeneOhm MRSA assay, a total of 101 assorted MRSA isolates of unknown origin and 3 reference strains were tested (Table 2). All isolates were suspended in 0.45% saline (NaCl) up to a density of approximately 0.7 McFarland standard. After a short centrifugation step, the supernatant was tested using the different real-time PCR assays.

TABLE 2.

Overview of results obtained evaluating the real-time PCR assays using test strains

spa type	N	Result
		SA442	mecA	mecC	BD GeneOhm
t002	3	3	3	0	3
t003	2	2	2	0	2
t008	14	14	14	0	14
t011a	26	26	26	0	26
t015	1	1	1	0	1
t019	8	8	8	0	8
t022	1	1	1	0	1
t034a	7	7	7	0	7
t044	1	1	1	0	1
t064	8	8	8	0	8
t105	1	1	1	0	1
t108a	6	6	6	0	6
t189	2	0	2	0	1
t202	1	1	1	0	1
t223	2	2	2	0	2
t267	1	1	1	0	1
t296	1	1	1	0	1
t311	1	1	1	0	1
t316	1	1	1	0	1
t318	1	1	1	0	1
t359	2	2	2	0	2
t437	1	1	1	0	1
t548	1	1	1	0	1
t571a	1	1	1	0	1
t688	1	1	1	0	1
t786	1	1	1	0	1
t791	1	1	1	0	1
t899a	1	1	1	0	0
t1456a	1	1	1	0	1
t1814	1	1	1	0	1
t3932	1	1	1	0	1
t9633	1	1	1	0	1
ATCC 43300 (t007)	1	1	1	0	1
NCTC 10442 (t008)	1	1	1	0	1
mecC (t7603)a	1	1	0	1	0
Total	104	102 (98.1%)	103 (100%)	1 (100%)	101 (97.1%)

^aLivestock-associated MRSA (ST398).

RESULTS

Real-time PCR assays.

During a 2-year period, a total of 13,387 samples were presented for MRSA screening and tested using the SMM panel (Table 3). Of these samples, 2,112 (15.8%) were defined as suspected MRSA positive based on the presence of the SA442 gene (C_T value, ≤39) in combination with the presence of the mecA gene. In none of the samples was the mecC gene detected. Of all samples tested with the SMM panel, only 6 (0.04%) were inhibited after dilution.

TABLE 3.

Results of the screening approach

Yr	Results
	SMM panel			GeneOhm MRSA assay			Culture
							GeneOhm MRSA assay positive		GPa
	No. testedb	No. positive	No. unresolved	No. tested	No. positive	No. unresolved	No. tested	No. positive	No. tested	No. positive
2012	4,117 (2,040)	557	3	430 (306)	76	5	76 (39)	50	127 (87)	31
2013	9,270 (5,708)	1,555	3	1,428 (1,166)	289	12	288 (184)c	239	127 (95)	30
Total	13,387 (7,748)	2,112	6	1,858 (1,472)	365	17	364 (223)	289	254 (182)	61

^aGP, general practitioner.

^bIn parentheses, the number of pooled samples tested.

^cOne GeneOhm MRSA assay-positive sample could not be cultured.

Of the 2,112 suspected MRSA-positive samples, 1,858 (88%) were subsequently tested with the GeneOhm MRSA assay, of which 365 samples were positive and 17 had an unresolved result. The remaining 254 samples (12.0%) were sent by a general practitioner and were therefore subcultured directly. Overall, based on the results of both real-time PCR assays, a total of 12,745 (95.2%) samples could be reported as negative within 24 h.

MRSA culture and spa typing.

Of the 13,387 samples received for MRSA screening by real-time PCR, 642 samples (4.8%) were eventually cultured. This was either due to a positive result in both the SMM panel and the GeneOhm MRSA assay (n = 365), due to a positive result in the SMM panel in combination with the fact that they were sent in by a general practitioner (n = 254), or due to inhibition in the real-time PCR assays (n = 23). Seven additional samples had to be cultured because of suspicion of LA-MRSA, for which the SMM panel was positive and the GeneOhm MRSA assay remained negative.

Out of the 365 samples that were presented for culture because of the positive interpretation of the GeneOhm MRSA assay, one sample was not available for culture. Of the remaining 364 samples, 289 (79.4%), 121 of which eventually proved to be LA-MRSA, from 213 different patients were confirmed MRSA positive by culture.

Of the suspected MRSA-positive samples based on the SMM panel, 254 were sent by a general practitioner and were presented for culture directly. Sixty-one (24.0%) of these samples from 45 different patients were MRSA positive in culture.

Among the 23 samples that had to be cultured because of inhibition in the real-time PCR assays, 20 samples were available for culture, and all were determined to be MRSA negative.

Six out of the seven samples that were negative in the GeneOhm MRSA assay were cultured because of suspicion of LA-MRSA. In none of the cultured samples was MRSA found. Due to an error, the remaining sample could not be cultured.

In total, 350/13,387 samples (2.6%) were reported MRSA positive, 150 of which (42.9%) proved to be LA-MRSA, since the isolates belonged to sequence type 398 (ST398). The 350 MRSA-positive samples were obtained from 256 individuals.

Each new MRSA isolate was sent to the RIVM for spa typing. As shown in Table 4, spa typing of 258 MRSA isolates showed 45 different types, with t011 the most prevalent, followed by t064. Two individuals had two different MRSA isolates, with spa types t011 and t899 found in the nose/throat samples and spa types t4119 and t4571 in the perineal samples.

TABLE 4.

spa types detected during the study

spa typea	No. detected
	2012	2013	Total
t011b	22	36	58
t064	2	33	35
t002	2	23	25
t008	5	19	24
t034b	3	14	17
t899b	5	7	12
t1081		12	12
t108b	1	9	10
t032		9	9
t223		4	4
t1457b		3	3
t179	2	1	3
t359		3	3
t105	2		2
t003	1	1	2
t044	1	1	2
t067		2	2
t1474		2	2
t2922b	1	1	2
t437	2	1	3
t447		2	2
t571b		2	2
t791	2		2
t019, t084, t211, t311, t324, t441, t786, t965, t1430, t1451b, t1930, t2011b, t2330b, t4119b, t4132b		All 1	15
t121, t127, t148, t304, t321, t443, t4571b	All 1		7
Not determined	1		1
Total	59	200	259

^aAll single spa types found per patient.

^bLivestock-associated MRSA (ST398).

Evaluation of the real-time PCR assays.

In general, real-time PCR assays may result in false-negative results due to primer-probe mismatches. This is especially the case for SCCmec-orfX-based assays, where large variations in the primer-probe region are described. In order to assess the correct performance of the real-time PCR assays used in this method, a total of 101 assorted MRSA isolates of unknown origin detected at three different laboratories for medical microbiology, along with 3 reference isolates, were subjected to these PCR assays. All the MRSA isolates were correctly detected by both the mecA and the mecC real-time PCR assays. The real-time PCR targeting the S. aureus-specific gene SA442 was able to detect 102 (sensitivity, 98.1%) of the MRSA isolates. The two SA442-negative isolates both belonged to spa type t189. Moreover, one of these samples was not detected by the GeneOhm MRSA assay. Furthermore, the GeneOhm MRSA assay was not able to detect an isolate with spa type t899, as well as the isolate with the mecC gene (Table 2).

DISCUSSION

In this study, we evaluated our molecular screening approach for the detection of MRSA over a 2-year period. During this period, 13,387 samples were inoculated into a selective broth and tested in the SMM panel after an overnight incubation. Of these, 2,112 (15.8%) were defined as suspected MRSA positive, since the SA442 PCR showed a C_T value of 39 or lower in combination with a positive mecA PCR. In the vast majority of cases (n = 1,858), the GeneOhm MRSA assay was used as a confirmation for the samples that were suspected MRSA positive in the SMM panel. Culture was performed from a total of 618 suspected MRSA-positive samples. Of these samples, 364 had a positive result using the GeneOhm MRSA assay, while the remaining 254 were subcultured directly following the SMM panel, since the samples originated from a general practitioner. Eventually, 350 samples (2.6%) were MRSA positive.

Using this molecular screening approach, 95.2% of the samples could be reported negative based on the real-time PCR assays. Both assays had shown good sensitivities and negative predictive values in earlier studies, making this screening approach very useful for detection of MRSA-negative samples (15, 21).

As this approach is performed on a daily basis, results can be available within 24 h. Moreover, since the SMM panel has a capacity of 44 samples per real-time PCR run of 1 h 40 min, 176 samples (a total of 4 runs) can be analyzed per real-time PCR system within one working day. This high throughput makes the approach an extremely valuable tool during outbreak situations to monitor and adjust the infection control interventions. By skipping the SMM panel and performing the GeneOhm MRSA assay directly on the broth, the time to result would decrease from 2 h 40 min (combining the SMM panel and the GeneOhm MRSA assay) to 1 h. However, since the Smartcycler has a capacity of only 14 samples per real-time PCR run, only 98 samples (a total of 7 runs) can be analyzed per Smartcycler system within one working day. Moreover, the price per sample is approximately 8 times higher when the GeneOhm MRSA assay is used instead of the in-house SMM panel. Therefore, the strength of this approach is the combination of both real-time PCR assays, in which the SMM panel is associated with a high throughput and low cost. The surplus value of the GeneOhm MRSA assay was demonstrated because MRSA was cultured in 79.4% of all samples that tested positive by the GeneOhm MRSA assay, while the percentage was much lower when samples were cultured directly following the SMM panel (24%). Nevertheless, it should be stressed that positive results of both real-time PCR assays should always be confirmed using culture, since false-positive interpretations can be observed in the SMM panel and the GeneOhm MRSA assay because of a mixture of MRCNS and MSSA or a SCCmec-cassette with a mecA dropout, respectively (21–23).

For nonhospitalized patients, however, there is usually less need for rapid MRSA results than for hospitalized patients. To reduce costs, the GeneOhm MRSA assay was not applied for samples supplied by general practitioners; if the SA442 assay showed C_T values of 39 or lower in combination with a positive mecA or mecC PCR, broth samples were subcultured on Brilliance MRSA2 and blood agar plates immediately, as shown in Fig. 1.

Introduction of a ΔC_T-based strategy has been shown to improve the sensitivity and specificity of MRSA detection using real-time PCR assays targeting an S. aureus species-specific gene and the mecA gene (24, 25). Bode et al. applied this strategy in their method and focused on the difference in C_T values between SA442 and mecA (dC_T, the C_T value of SA442 minus the C_T value of mecA) (21). By introducing this dC_T, they showed 97.5% sensitivity (39/40) and 96.6% specificity (1,701/1,760) when a dC_T between −2 and +8 was assessed. Although the dC_T can, for instance, be influenced by the use of different broths (with different [concentrations of] antibiotics), we did compare our data with the study of Bode et al., since the same selective broth and target genes were used. In our study, the dC_T varied from −10.13 to 14.9, with a median of −0.51. In our hands, the sensitivity would be only 86.3% (302/350) when the dC_T between −2 and +8 was used for the results. Despite some minor differences between the two real-time PCR assays, we discourage the use of such a dC_T for the interpretation of the SMM panel, since this may lead to false-negative results.

In general, false-negative results may occur using real-time PCR assays because of primer-probe mismatches. In order to assess the performance of the real-time PCR assays used in this method, a panel of 104 assorted MRSA isolates, consisting of 101 MRSA isolates of unknown origin and 3 reference strains, were subjected to the real-time PCR assays. This panel of isolates included the 5 most prevalent detected spa types in the Netherlands during 2008 and 2009 (t011 and t108 [both LA-MRSA] and t008, t002, and t064 [not LA-MRSA]) (26). While the sensitivity of the mecA-mecC PCR was 100%, the SA442 assay tested negative for 2 of the isolates (sensitivity, 98.1%). This is comparable to the findings of two studies that showed 99.7% and 99.9% sensitivity in panels of 2,600 and 4,500 S. aureus isolates, respectively (27, 28). Although the sensitivity is very high, these results still might lead to undetected MRSA. The introduction of a second target gene for the detection of S. aureus, such as the nuc gene, might further improve the sensitivity. Although nuc-deficient isolates have sporadically been reported, combining the two genes for detection of S. aureus will be important in order to approach correct detection of all possible S. aureus strains (29).

The GeneOhm MRSA assay detected 97.1% of the isolates correctly. Among the 3 isolates that were determined to be negative, one isolate contained the mecC gene, known to be negative in this assay. Since in our approach the GeneOhm MRSA assay was abolished for mecC-positive isolates and broths were subcultured directly, no mecC-containing MRSA isolates among our clinical specimens would have been missed (18). Apart from the mecC isolate, false-negative results can be produced by this assay, especially in LA-MRSA, because of the variability in the SCCmec and/or orfX sequences (30, 31). Therefore, samples in which LA-MRSA is suspected (as stated by the applicant) were subcultured even when a negative result was found using the GeneOhm MRSA assay.

Only 23 samples were inhibited in either the SMM panel or the GeneOhm MRSA assay. Three samples were lost; all other samples showing inhibition were further processed by culture and produced a negative result.

None of the samples contained the mecC gene, indicating a very low prevalence of MRSA containing mecC in the Netherlands, just as in, for instance, the United Kingdom, Germany, and Switzerland (32–35). In contrast to these findings, the prevalence of MRSA containing the mecC gene seems to be increasing in Denmark, indicating the need to screen for the mecC gene (36).

Due to several outbreaks in hospitals and nursing homes in 2013, an enormous increase in samples was observed, with 4,117 and 9,270 samples in 2012 and 2013, respectively. Hence, a high number of MRSA isolates with spa types t002, t008, t034, t064, and t1081 were found compared to 2012 (Table 4). Overall, the most prevalent spa type was t011, followed by t064, t002, and t008. This is in accordance with the epidemiological data for 2008 and 2009 published by Lekkerkerk et al. and with the data shown in a report on the surveillance of MRSA in the Netherlands in 2011 available on the website of the RIVM (Bilthoven, the Netherlands) (https://mrsa.rivm.nl/jaarverslagen_mrsa; 26).

The screening approach used differs from the Dutch guidelines on the laboratory detection of MRSA, which recommend the use of (relatively) nonselective broth enrichment in combination with subculture on a MRSA screening agar and nonselective blood agar plate for 48 h, yielding negative results no sooner than 72 h (8). Our approach consisted of an overnight incubation with PMB. To our knowledge, no head-to-head comparison between PMB and a nonselective broth has been reported, but Böcher et al. did show improvement of screening for MRSA when using a semiselective broth over a nonselective broth (37). Although this result cannot be directly extrapolated to the performance of the selective broth used in our study (PMB) in comparison with a nonselective broth, it suggests a comparable improvement in detecting MRSA. Following broth enrichment, culture using a MRSA screening agar instead of molecular detection methods is recommended by the Dutch guidelines, since the authors found that evidence was too limited to recommend the use of broth enrichment in combination with molecular methods (8). Recently however, multiple studies have compared the use of broth enrichment in combination with real-time PCR assays to the combination of broth enrichment followed by culture methods. These studies have shown high sensitivities when using broth enrichment and subsequent real-time PCR, showing the reliability of the combined methods (21, 24, 25). Based on the published data in combination with the data here, we suggest that recommendations on the use of broth enrichment in combination with real-time PCR assays be included in the Dutch guidelines on the laboratory detection of MRSA.

In conclusion, this molecular screening approach for MRSA based on the combination of two different real-time PCR assays has proven to be highly sensitive and specific. Because of the great sensitivity and negative predictive values of the real-time PCR assays, reliable negative results can be obtained within 24 h after the samples arrive at the laboratory. Especially during outbreaks of MRSA, this high-throughput screening approach yielding rapid negative results can be a major improvement in infection control programs.