New Vitek 2 Colorimetric GN Card for Identification of Gram-Negative Nonfermentative Bacilli

Abstract

The recent publication of Zbinden et al. (4) regarding the Vitek 2 colorimetric GN card for the identification of gram-negative nonfermentative bacilli contains data inconsistencies which the authors should have addressed before making recommendations for the use of this product. The authors stated that the 90 isolates tested were “well identified” but that 20/90 (22.2%) were not identified to the species level. Their “gold standard” reference method (partial 16S rRNA gene sequencing) failed to characterize 10 strains beyond the genus level, and the remaining 10 strains could not be resolved between two species. The authors recommended the Vitek 2 GN card for only five nonfermentative taxa—Achromobacter xylosoxidans, Acinetobacter spp., Burkholderia cepacia complex, Pseudomonas aeruginosa, and Stenotrophomonas maltophilia—but the authors lacked sufficient data to contraindicate the use of the card for other species claimed (identified) by this product. While the Vitek 2 GN card claims over 60 nonfermenter species, the authors tested only 35 taxa, of which at least 14 represented unclaimed taxa; all things considered, less than one-third of the claimed species were actually tested. Of the species whose claims were questioned, over 10 were represented by testing only a single strain. These results appear to lack statistical significance. As mentioned above, at least 14/90 (15.6%) study isolates represented species for which there is no product claim. From the data shown in Table 2 of the article (4), one cannot determine the actual number of unclaimed taxa that were tested, but it appears to be at least 15 (16.7%), if one excludes the three unclaimed Acinetobacter spp. There is no commercial identification product that can give the correct result for an unclaimed taxon. With this in mind, it is purely a subjective choice by the authors as to whether unclaimed taxa are included in their study and, if so, to what extent. Clearly, that choice can bias and affect the overall product performance. In the article by Zbinden et al. (4), the totals shown in Table 1 are not an obvious translation of the individual data shown in Table 2. Additionally, there was no protocol in place to resolve discrepancies and verify whether the correct culture was tested. The “not identified” category members (n = 9 isolates) in Table 1 were erroneously grouped with and considered part of the “misidentified” category in the abstract. Additionally, the authors categorized low-selectivity results (n = 5 isolates) in which more than one genus was listed as “not identified” in Table 1. However, these results represent correct identifications to the species level once supplemental testing is performed. The authors recommended the use of the GN card with five of the most commonly encountered nonfermenters in a clinical lab, but they also made recommendations regarding the qualitative level of identification that were unfounded based on their data. Based on their recommendation, one would discount the reliability of the “acceptable” identifications by the Vitek 2 fluorescent card that proved to be 100% (3/3) correct to the species level in Table 3 of the article (4). A review of the literature shows that other investigators (1-3) found the Vitek 2 GN card to be very reliable for the identification of gram-negative nonfermentative bacilli. Correct results were obtained for 98.7%, 100.0%, and 92.1% of 144, 95, and 88 isolates studied, respectively. It is, therefore, curious how disparate the results of this study were, unless something went awry with the authors' experimental design.