Limitations of Vitek GPS-418 Cards in Exact Detection of Vancomycin-Resistant Enterococci with the vanB Genotype

Abstract

The susceptibilities of 20 strains of vancomycin-resistant enterococci (VRE) with the vanB genotype obtained by using Vitek GPS-418 cards were compared with those obtained by the broth dilution method of the National Committee for Clinical Laboratory Standards (NCCLS) (approved standard M7-A4) and with those obtained by the agar screen method using bile esculin azide agar containing 6 μg of vancomycin per ml. Although both the broth dilution and agar screen methods disclosed no discordance, Vitek GPS-418 cards yielded a very major error compared with the results obtained by the reference broth dilution method of the NCCLS. Vitek GPS-418 cards were therefore found to have considerable room for improvement for the accurate detection of vanB VRE strains.

Since the late 1980s, strains of vancomycin-resistant enterococci (VRE) have emerged as significant causes of nosocomial infections and colonizations with increasing frequency in every part of the world (2, 4, 12, 14). Indeed, there is no proven chemotherapy for an infection, since VRE strains have been demonstrated to be resistant not only to vancomycin but also to multidrug therapy (11). Two genetically distinct forms of resistance, designated the vanA and vanB genotypes, are recognized to be clinically important, although intrinsic resistance occurs in some enterococcal species (vanC genotype) and a third form of resistance (vanD genotype) has emerged (10). Moreover, an additional fourth type of the resistance (vanE genotype) has recently been reported (1).

In fact, infectious diseases due to these VRE strains have become a serious problem (2, 4, 5). To control the transmission of VRE strains and to control successive outbreaks of their cross-infections in any medical facility (6, 7, 9), the necessity of evaluating the rapidity and reliability of the detection methods for VRE strains has been emphasized for clinical microbiology laboratories. Although genetic methods for the detection of the vanA and/or vanB gene by PCR have been evaluated with considerable certainty, routine tests, such as the E-test (AB Biodisk, Solna, Sweden), sensi-disk (Becton Dickinson Microbiology Systems, Sparks, Md.), and agar screen test using plates containing 6 μg of vancomycin per ml, are already commercially available for the confirmation of the phenotypes of VRE strains.

In Japan, many laboratories have adopted the Automicrobic System (Vitek Systems, Hazelwood, Mo.) for the detection of VRE strains in routine clinical microbiology. The Vitek GPS card has successively improved from the Vitek GPS-TA card with 30 wells released in July 1997 (used in Japan) to the Vitek GPS-101 card with 45 wells (used in the United States) to the Vitek GPS-418 card with 45 wells (used in Japan) released in February 1999. The abilities of both the Vitek GPS-101 and Vitek 418 cards to determine susceptibility to vancomycin were improved from a three-well method to a four-well method. Vitek GPS-101 cards used in the United States were the same as Vitek GPS-418 cards used in Japan except for a minor change in the antimicrobials other than vancomycin and have already been evaluated to have been significantly improved in the detection of VRE with the VanB phenotype (3). In fact, H. P. Entdz et al. emphasized that 15 investigated VRE strains with the vanB genotype were all correctly identified as VRE strains without exception by the Vitek GPS-101 card (3). Our report focused on the reevaluation, with 20 Japanese VRE strains with the vanB genotype, of the Vitek GPS-418 card (software configuration version VTK-R06.01) by comparing it with the cards provided both by the broth dilution method of the National Committee for Clinical Laboratory Standards (NCCLS) (approved standard M7-A4) (8) and by the sensi-disk method.

A total of 20 VRE strains with the vanB genotype from different patients were our stock cultures and were stored in Microbank vials (Pro-Lab Diagnostic, Ontario, Canada) at −83°C in a deep freezer. All of the strains examined were reidentified as Enterococcus faecalis by the Vitek GPI card, which also reconfirmed them to be VRE strains with the vanB genotype by detecting the vanB gene(s) by a PCR procedure (16) to coamplify vanB-specific genes of enterococci. Susceptibility tests were carried out in duplicate by the three methods, i.e., the Vitek GPS card method, the sensi-disk method, and the broth dilution method of NCCLS approved standard M7-A4 (8). The Vitek GPS card method was carried out using GPS-418 cards, as indicated in the instructions provided with the system. The sensi-disk method for susceptibility testing was also performed as indicated in the instructions provided by the manufacturer. The broth dilution method for the reference MICs of vancomycin for 20 VRE strains with the vanB genotype were done in accordance with the guidelines of NCCLS approved standard M7-A4 (8) on cation-adjusted Mueller-Hinton broth (Difco Laboratories, Detroit, Mich.). The results were read after 24 h of incubation at 37°C. VRE screen agars containing 6 μg of vancomycin (Nippon Becton Dickinson and Company, Ltd., Tokyo, Japan) per ml were used with an inoculum size of 10 μl of a McFarland 0.5 standard suspension as described by Tenover et al. (13). The NCCLS breakpoints were used for the interpretation of the MICs (8).

Duplicate results obtained with each procedure were then compared, and any strain with an aberrant or questionable reaction was retested by the three methods for the susceptibility properties in discrepancy. A very major error was defined as an isolate which was resistant by the reference dilution method but susceptible by the test method. A major error was defined as an isolate which was susceptible by the reference dilution method but resistant by the test method. A minor error was defined as a discrepancy between the results of the reference dilution method and the test method corresponding to one interpretation category.

Table 1 demonstrates the distribution of MICs of vancomycin for the 20 strains with the vanB genotype tested. Shown are the MICs obtained by the reference broth dilution method of NCCLS approved standard M7-A4 (8). The vast majority (17 out of 20 strains) of MICs fell below 32 μg/ml, ranging from 8 to 32 μg/ml. On the other hand, the Vitek GPS-418 card method yielded an apparently discrepant MIC of less than 0.5 μg/ml, although it was observed with only a single strain (Table 1). This discrepant finding was always consistent even after the trials were repeated five times with Vitek GPS-418 cards. The MIC for this strain was 16 μ/ml by the reference broth dilution method of NCCLS approved standard M7-A4 (8).

TABLE 1.

MICs of vancomycin for 20 vanB VRE strains

Method	No. of strains for which the MIC (μg/ml) was:
	≤0.5	0.5	1	2	4	8	16	≥32
Broth dilution	0	0	0	0	0	1	2	17
Vitek GPS-418 cards	1a	0	0	0	0	0	0	19

^aThe MIC for this strain was 16 μg/ml by the broth dilution method. 

Table 2 represents the incidence of very major, major, and minor errors of the separate methods in comparison with results of the reference broth dilution method of NCCLS approved standard M7-A4 (8). Both the sensi-disk and agar screen methods were 100% sensitive with each other for the detection of vanB genes in the 20 VRE strains examined. As shown in Table 2, the results obtained by the sensi-disk method corresponded well to those obtained by the broth dilution method of NCCLS approved standard M7-A4 (8), and even a minor error was not observed. In addition, the agar screen method using VRE screen agar supplemented with vancomycin (Nippon Becton Dickinson and Company, Ltd.) at a final concentration of 6 μg/ml exceptionally allowed the growth of all the VRE strains investigated. It is very encouraging that both the sensi-disk method and the agar screen method correctly identified all of the 20 VRE strains examined, thus confirming the results of some recent evaluation studies (3, 15, 16), but these tests were time-consuming nevertheless.

TABLE 2.

Errors among the tested methods for the detection of 20 vanB VRE strains

Method	No. of strains showing an error that wasa:
	Very major	Major	Minor
Agar screenb	0	0
Sensi-disk	0	0	0
Vitek GPS-418 cards	1	0	0

^aSee the definitions of very major, major, and minor in the text. 

^bVRE screen agar containing 6 μg of vancomycin per ml. 

Indeed, Vitek systems have been widely adopted by many microbiology laboratories for the rapid identification and determination of susceptibilities of pathogens, including VRE strains. In order not only to treat VRE-infected patients but also to implement appropriate control measures to prevent the spread of VRE strains, rapid and accurate identification of VRE-colonized patients is essential. N. Yamane et al. (16) evaluated several automated methods, such as the Microscan conventional- and rapid-panel methods and Vitek GPS-TA cards, and found that every automated method tested was shown to yield a high error rate. Recently, H. P. Entdz et al. (3) pointed out that newly developed Vitek GPS-101 cards with 45 wells have been significantly improved and that they successfully detected all of the VRE strains with the vanB genotype tested (15 out of 15 strains).

During the course of this reevaluation study, however, the very major error (only 1 out of 20 strains detected successfully) that only the Vitek GPS-418 cards revealed was noteworthy. It should be emphasized that there still exists much room for improvement in Vitek GPS-418 cards for the accurate detection of VRE strains with the vanB genotype. Significant improvement of Vitek GPS cards for the exact detection of VRE strains is urgently desired.