Grading the Germ Tests: An Update
Research Models
William Shek, Christine Farrance, PhD

Grading the Germ Tests: An Update

Several months ago, the Research and Animal Diagnostics Services division at Charles River embarked on an internal study to see how the standard VITEK® 2 system held up against an alternative microbial identification platform called MALDI-TOF MS that Charles River’s Delaware-based endotoxin testing division Accugenix markets widely to an array of biopharmaceutical, nutraceutical and medical device manufacturers around the globe (see Eureka blog, March 10, 2014).

The VITEK® 2 system, which has been streamlined over time, has served animal laboratories well, but is not perfect. This is particularly true for Pasteurella pneumotropica, a Gram-negative bacteria primarily carried by rats and mice that is common in animal laboratories. VITEK® 2 cards can’t accurately identify the biotypes so it becomes necessary to verify the two strains peculiar to mice using polymerase chain reaction (PCR) technology. PCR is also used routinely to verify Corynebacterium bovis in athymic nude mice.

So we decided to look at the Accugenix AccuPRO-ID service because in industry settings, at least, it can correctly identify bacteria approximately 98% of the time. In contrast, VITEK® 2 incorrectly identifies bacteria about a third of the time. We also wanted to assess any variations in the field isolates depending upon which system is used.

Here’s what our survey revealed. Of 28 bacterial species (but importantly excluding P. pneumotropica), identification by phenotypic methods matched those of MALDI-TOF MS for 22 (79%) of samples and of 16S rDNA sequencing for 18 (64%) of samples. Most mismatches were for the five isolates of salmonella, shigella or Escherichia. Excluding those, the correspondence of phenotypic IDs to those MALDI-TOF MS and 16S rDNA sequencing was 91% and 78%, respectively. Of 20 isolates phenotypically identified as P. pneumotropica, matching IDs were obtained for 17 (85%) by PCR, only slightly better than 16 (80%) by MALDI-TOF MS. The correspondence between PCR and MALDI-TOF MS IDs was 85%.

The data provide preliminary evidence that MALDI-TOF MS is a suitable alternative to phenotypic methods for identification of bacterial isolates, and may reduce the need for confirmatory PCR testing.