Comparison of the Vitek Gram-Positive Susceptibility 106 Card, the MRSA-Screen Latex Agglutination Test, and mecA Analysis for Detecting Oxacillin Resistance in a Geographically Diverse Collection of Clinical Isolates of Coagulase-Negative Staphylococci

Abstract

The Vitek automated susceptibility testing system with a modified gram-positive susceptibility (GPS) 106 card (bioMerieux Vitek, Inc., Hazelwood. Mo.) and a rapid slide latex agglutination test (MRSA-Screen test; Denka Seiken Co., Ltd., Tokyo, Japan) were evaluated for their abilities to detect oxacillin resistance in coagulase-negative staphylococci (CoNS). The reference broth microdilution method and the detection of the mecA gene by PCR (“gold standard” reference result) were used to compare the results obtained with the commercial products. A total of 123 clinical isolates consisting of eight species were selected from U.S. surveillance collections. Among the mecA-positive isolates (95 strains), 30 isolates were initially negative on the MRSA-Screen test read at 3 min. When the agglutination reaction was extended for 10 min, 26 of the 30 isolates became positive. For a different four isolates, the oxacillin MIC was ≤0.25 μg/ml on the Vitek GPS 106 card. Among the mecA-negative isolates (28 strains), for two Staphylococcus warneri, two S. lugdunensis, and two S. saprophyticus strains MICs were ≥0.5 μg/ml by the reference broth microdilution method. Four of these isolates were also categorized as resistant with the Vitek GPS 106 card and two isolates were positive by the MRSA-Screen test. Overall, the MRSA-Screen test, GPS 106 card, and reference broth microdilution method had sensitivities of 95.7 (result at 10 min), 95.7, and 100%, respectively, and specificities of 92.8, 85.7, and 78.5%, respectively. Although the MRSA-Screen test required a slight procedural modification, both commercial methods achieved a sensitivity and specificity at detecting oxacillin resistance in CoNS at a level that was acceptable for clinical laboratory use.

The coagulase-negative staphylococci (CoNS) are a common cause of nosocomial bacteremia, particularly in patients with prosthetics and indwelling vascular devices. The accurate detection of oxacillin resistance among staphylococcal isolates in the clinical laboratory is important as a guide to therapy and for the prudent use of vancomycin. Oxacillin-susceptible staphylococcal infections are more effectively treated with β-lactam antimicrobials because of their higher intrinsic activities, better concentrations in tissue, low incidence of side effects, more rapid bactericidal action, and lower therapeutic costs (11, 21).

Recently, several investigators have reported disagreement between the oxacillin MICs for CoNS and the presence or absence of mecA (11, 12, 17, 21). Accordingly, the National Committee for Clinical Laboratory Standards (NCCLS) (15, 16) has modified the oxacillin interpretive criteria for this organism, applying an MIC breakpoint of ≤0.25 μg/ml for susceptibility for suspectibility and an MIC breakpoint of ≥0.5 μg/ml for resistance (16). The accurate and timely detection of oxacillin-resistant CoNS is directly dependent upon the quality of the manual or automated susceptibility testing system used in the clinical laboratory.

The present study was performed to assess the abilities of two newly available tests: (i) an automated susceptibility testing system with a gram-positive susceptibility (GPS) card (GPS 106 card; bioMerieux Vitek, Inc., Hazelwood, Mo.), which was modified to address new NCCLS oxacillin breakpoint criteria, and (ii) a rapid slide latex agglutination test, (MRSA-Screen test; Denka-Seiken Co., Ltd., Tokyo, Japan) (14) to accurately classify CoNS as oxacillin susceptible or as having resistance.

MATERIALS AND METHODS

Bacterial isolates.

A total of 123 isolates were collected and selected from geographically diverse hospitals in the United States. Fifty isolates were chosen from a challenge collection for which oxacillin MICs were previously demonstrated to range from 0.25 to 4 μg/ml (18). The other 73 isolates were chosen from the collection of surveillance isolates obtained from 1995 to 1996 (11). All of the isolates were identified to the species level by standard biochemical methods (8), by automated ribotyping (RiboPrinter; Dupont Qualicon Inc., Wilmington, Del.), or with the Vitek system. The following species were tested: Staphylococcus epidermidis (78 strains), S. haemolyticus (17 strains), S. hominis (16 strains), S. warneri (6 strains), S. lugdunensis (2 strains), S. saprophyticus (2 strains), and S. capitis and S. simulans (1 strain each). All isolates were frozen at −70°C until they were processed. testing, each isolate was subcultured at least twice on blood agar plates (Remel, Lenexa, Kans.) to ensure purity and optimal growth characteristics. All isolates were subjected to blind testing with the MRSA-Screen (Denka Seiken Co., Ltd.) test and by determination of oxacillin MICs with the GPS 106 card and by the reference NCCLS broth microdilution test (15, 16). Control strains used for all assays included oxacillin-resistant strain S. aureus ATCC 43300 and a well-characterized clinical CoNS strain from the previously described challenge collection (18), as well as quality control strain S. aureus ATCC 29213.

Susceptibility testing methods.

The MRSA-Screen test was performed according to the manufacturer's instructions, with slight modification. Briefly, approximately 5 μl of bacterial cells was taken from a fresh subculture and was then suspended in 4 drops of extraction reagent 1 and boiled for 3 min. After the suspension was allowed to cool to room temperature, 1 drop of extraction reagent 2 was added and the mixture was vortexed thoroughly. The suspension was centrifuged at 1,500 × g for 5 min. A 50-μl aliquot of the supernatant was mixed with 1 drop of anti-PBP 2a monoclonal antibody-sensitized latex beads. A negative control test was performed by using 50 μl of supernatant mixed with 1 drop of negative control latex. The samples were then placed on a shaker and gently mixed for up to 15 mins. The resulting agglutination pattern was read at 3, 6, and 10 min.

Automated susceptibility testing was performed with the GPS 106 card (software configuration version R07.01) with the Vitek system according to the manufacturer's recommendation. Antimicrobial susceptibility testing of isolates with broth microdilution trays was performed in accordance with NCCLS guidelines (15, 16). The interpretation for oxacillin followed the breakpoint recommendation (resistant, ≥0.5 μg/ml; susceptible, ≤0.25 μg/ml) of NCCLS (16).

PCR was performed for the detection of mecA as described previously (3, 5). When isolates of CoNS yielded discrepant result among the tests used, the tests were repeated in triplicate from the same organism culture source. The replicate results, if reproducible (three of three results), were used as the value for analyses, regardless of whether they remained discrepant.

RESULTS

Table 1 summarizes the results obtained by the MRSA-Screen test, with the Vitek GPS 106 card, by the reference broth microdilution method, and by mecA PCR detection. Among the 123 isolates tested, 95 isolates were positive for mecA and 28 isolates were negative for mecA by PCR. Of the mecA-positive isolates, 30 isolates were initially negative by the MRSA-Screen test read at 3 min (data not shown). When the agglutination reaction was extended for 10 min, 26 of the 30 isolates became positive. Oxacillin MICs for another four isolates were ≤0.25 μg/ml with the Vitek GPS 106 card. For all mecA-positive isolates, MICs were ≥0.5 μg/ml by the reference broth microdilution method.

TABLE 1.

Evaluation of MRSA-Screen test, Vitek GPS 106 card, and broth microdilution method for detection of oxacillin resistance in 123 CoNS

PCR mecA detection	No. of isolates	No. of isolates with the indicated resulta
		MRSA-Screen testb		Vitek GPS 106 card MIC (μg/ml) of:		Microdilution MIC (μg/ml) of:
		Negative	Positive	≤0.25	≥0.5	≤0.25	≥0.5
Negative	28	26	2	24	4	22	6
Positive	95	4	91	4	91	0	95

^aThe sensitivities of the MRSA-Screen test, Vitek GPS 106 card, and the microdilution test were 95.7, 95.7, and 100%, respectively, and the specificities were 92.8, 85.7, and 78.5%, respectively. 

^bThese results were read at 10 min. See the text for further details. 

Among the mecA-negative isolates (28 strains), oxacillin MICs for 6 isolates consisting of 2 isolates each of S. warneri, S. lugdunensis, and S. saprophyticus were 0.5 to 2 μg/ml by the reference broth microdilution method (Table 2). Four isolates (S. lugdunensis and S. saprophyticus), were also categorized as resistant with the Vitek GPS 106 card, and two isolates (S. warneri) were positive by the MRSA-Screen test.

TABLE 2.

Characteristics of isolates of CoNS which had discrepant results by mecA PCR detection, by the MRSA-Screen test, with the Vitek GPS 106 card, and by the broth microdilution method

Isolate	Species	Reproducible resultsa of the following test:
		mecA PCR	MRSA- Screen testb	Vitek GPS 106 card	Broth microdilution method
024	S. warnei	−	+	≤0.25	0.5
037	S. warnei	−	+	0.5	0.5
036	S. lugdunensis	−	−	≤0.25	2
001	S. lugdunensis	−	−	0.5	1
048	S. saprophyticus	−	−	0.5	1
017	S. saprophyticus	−	−	0.5	2
060	S. hominis	+	+	≤0.25	>8
112	S. hominis	+	+	≤0.25	1
002	S. hominis	+	+	≤0.25	4
007	S. epidermidis	+	+	≤0.25	1
084	S. epidermidis	+	−	0.5	>8
086	S. epidermidis	+	−	4	>8
097	S. epidermidis	+	−	0.5	2
005	S. haemolyticus	+	−	2	4

^a+, positive; −, negative; the results for Vitek GPS 106 card and broth microdilution tests are given as MICs (in micrograms per milliliter). 

^bThese results were read at 10 min. See the text for further explanations. 

Overall, the MRSA-Screen test, the GPS 106 card, and the broth microdilution method had sensitivities of 95.7, 95.7, and 100%, respectively, and specificities of 92.8, 85.7, and 78.5%, respectively. The calculations of sensitivities and specificities were based on the total number of isolates listed in Table 1, with the mecA PCR results used as the reference value.

DISCUSSION

Testing for oxacillin resistance in staphylococci has been problematic for clinical laboratories for more than 15 years (13). The difficulty in the detection of phenotypic methicillin (oxacillin) resistance is due to the heterogeneous expression of the mecA gene by strains of staphylococci (2, 4). Each cell in the population may carry the genetic information of resistance, but only 1 in 10⁵ to 10⁷ organisms expresses the resistance phenotypically (4). This heterogeneity is more common in CoNS than in S. aureus (1) and makes the phenotypic detection of resistance in CoNS difficult and potentially time-consuming (≥24 h).

The Vitek system with the modified GPS 106 card was observed to be easy to use and provided results within 8 h. The Vitek GPS 106 card was revised so that three wells contained oxacillin, and the updated software (Vitek software configuration version R07.01) was developed to analyze this modified number of concentrations. This card was further adjusted to meet revised NCCLS oxacillin breakpoints for CoNS (15, 16). As a result, Vitek system cards now possesses a high degree of sensitivity (95.7%), similar to the results of a preliminary assessment during Food and Drug Administration clinical trials (10). Only 4 isolates among 95 mecA-positive isolates were misclassified as oxacillin susceptible by phenotypic MIC criteria.

Among 28 mecA-negative isolates, for 4 isolates Vitek MICs were 0.5 μg/ml (specificity, 85.7%). Those isolates were also classified as oxacillin resistant by the reference broth microdilution method. In a recent report (6), the new oxacillin MIC breakpoints were found to be less accurate when they were applied to some species of CoNS such as S. cohnii, S. saprophyticus, S. warneri, S. lugdunensis, and S. xylosus (6), and our findings confirm this observation (Table 2). The new MIC breakpoint (≤0.25 μg/ml) was selected to be the best choice for maximizing the sensitivity for detection of mecA-positive S. epidermidis isolates (18), realizing that S. epidermidis was the major species of CoNS tested by clinical laboratories (11, 18). In the usual clinical situation, infections caused by CoNS other than S. epidermidis may be serious, but less common, therefore, false-resistance errors should be relatively rare. The use of the new recommendations appears to be practical, and the Vitek GPS 106 card should be useful for oxacillin susceptibility testing of CoNS in clinical laboratories.

The MRSA-Screen test was very rapid (<1 h) and easy to perform. It has been extensively evaluated mainly for S. aureus and has been found to be very sensitive and specific for the detection of oxacillin resistance (9, 19, 20). Hussain et al. (7) described its reliability in a variety of species of CoNS. The MRSA-Screen test was able to accurately detect oxacillin-resistant CoNS when it was performed with isolates in which oxacillin resistance had been induced. In contrast, only 57.6% of the isolates tested gave a positive reaction without induction (7). Although our experiments were performed without induction, the test sensitivity was improved compared to that of the study by Hussain et al. (7) simply due to the use of a greater inoculum concentration. While Hussain et al. (7) used only a loopful of cells for inoculation, we used approximately fivefold more cells. In a recent study with S. aureus, it was found that the use of a larger inoculum also resulted in an improved sensitivity without a loss of specificity (9). Since heterogeneous expression of resistance is more common in CoNS than in S. aureus (1), a larger inoculum would be required to enhance the detection of PBP 2a. Our findings also demonstrate that the MRSA-Screen test can function well even without induction, leading us to consider this test for the direct detection of oxacillin resistance in organisms from blood cultures. This application has the potential to greatly reduce the overall time from specimen collection to the time of the initial report of susceptibility to oxacillin.

In applying the MRSA-Screen test to CoNS, we also slightly extended the reaction time of the agglutination reaction (to 10 min). When the test was read at 3 min, only 68% of mecA-positive isolates showed a positive reaction, but the sensitivity was increased to 95.7%. The weakly positive reactions by some isolates at 3 min became stronger by extension of the test interval. Similar results have been reported with methicillin-resistant S. aureus (19, 20), and the manufacturer recently changed its package insert recommendations, indicating the need for an extended time (10 min) for agglutination.

Overall, the MRSA-Screen test had better specificity (92.8%) than both MIC methods tested (85.7 and 78.5%). This higher degree of specificity and the equally high degree of sensitivity were favorable because of the desire to avoid the use of vancomycin for treatment of infections caused by oxacillin-susceptible isolates. Two isolates of S. warneri showed false-positive reactions on the MRSA-Screen test. One isolate showed agglutination at 1 min, and the other showed a positive reaction at 6 min. The former isolates were categorized as resistant by the Vitek and microdilution methods, so that a false-negative result by PCR cannot be entirely ruled out. Nevertheless, because we modified the MRSA-Screen test procedure, the possibility of false-positive results because of this procedural change needs to be evaluated in tests with larger numbers of isolates and species of worldwide distribution.

In conclusion, the MRSA-Screen test, was observed to be a very rapid and simple test for the detection of oxacillin-resistant CoNS. The Vitek GPS 106 cards also provided accurate phenotypic results in a relatively short time frame. Although a slight modification of the MRSA-Screen test procedure was required, both methods achieved sensitivities and specificities acceptable levels for the detection of oxacillin-resistant CoNS. The present study demonstrates the potential usefulness of both methods in the clinical microbiology laboratory, but suggests that molecular methods may be the preferred procedures for the identification oxacillin resistance in CoNS.