Evaluation of three commercial assays for rapid detection of genes encoding clinically relevant carbapenemases in cultured bacteria

Objectives: To assess the performance of three commercial molecular assays for detecting major families of carbapenemases in pure bacterial isolates. Methods: A panel of 450 isolates with previously defined carbapenem resistance mechanisms was tested using the Check-Direct CPE kit, the eazyplex (R) SuperBug complete A kit and the Xpert (R) Carba-R kit. Isolates included 438 Enterobacteriaceae and 12 Pseudomonas spp. comprising 100 isolates each with KPC, NDM, VIM or OXA-48-like enzymes, two isolates producing both an NDM and an OXA-48-like enzyme, 24 IMP producers and 24 isolates without a known carbapenemase gene. Discordant results (commercial versus in-house) were investigated using in-house PCR and amplicons were sequenced to define the carbapenemase allele present. Results: All three commercial assays detected all isolates with KPC, VIM, NDM and classic OXA-48 carbapenemases (no false-negatives). Isolates producing the OXA-181 variant (n = 18) were not detected by the Xpert (R) Carba-R kit or the eazyplex (R) SuperBug complete A kit, but were subsequently detected with modified versions of these kits. Only the Xpert (R) Carba-R kit could detect IMP carbapenemases, although this was limited to the IMP-1 subgroup. Invalid or false-positive results were either not observed when following the manufacturer's protocols or were eliminated by making simple interpretative adjustments to allow use with bacterial isolates rather than clinical samples. Conclusions: Commercial assays offer a reliable means of detecting bacteria with clinically significant carbapenemases. Coverage of some assays required expansion to maximize the sensitivity for OXA-48-like carbapenemases. Choice will ultimately depend on preferred gene coverage, intended throughput, cost and ability to fit into local workflows.