Aim: While new tools have recently become available for laboratory detection of intestinal carriage of carbapenemase-producing Enterobacteriaceae (CPE), the optimal testing algorithm has not been firmly established. This study aims to determine the optimal approach by comparing the diagnostic accuracy of three media in combination with the Carba NP test.
Methods: A panel of Enterobacteriaceae including 46 CPE and 59 non-carbapenemase producers were each spiked into stool specimens at high and low inocula. These were plated onto each of the three test media - Brilliance ESBL (Oxoid, Thebarton, Australia), chromID CARBA (bioMérieux, Marcy l’Etoile, France) and chromID OXA-48 (bioMérieux). Colonies isolated were identified by MALDI-TOF MS (VITEK MS, bioMérieux) and subjected to the Carba NP test. Four algorithms were compared to identify the approach with the greatest sensitivity and specificity for the detection of CPE.
Results: At the high and low inocula, Brilliance ESBL followed by the Carba NP test had a sensitivity of 83-85% and a specificity of 100%. At the high inoculum, the addition of chromID OXA-48 (without the Carba NP test) to this algorithm improved the sensitivity to 87% and maintained a specificity of 100%. ChromID CARBA followed by the Carba NP test demonstrated a sensitivity of 67% at the high inoculum and 63% at the low inoculum. Specificity of this algorithm was 100% at both inoculum strengths. The Brilliance ESBL media followed by the Carba NP test was significantly more sensitive than the equivalent chromID CARBA algorithm at the low inoculum strength (p = 0.02).
Conclusion: Brilliance ESBL followed by the Carba NP test demonstrated superior detection of low-level carriage of CPE. The addition of chromID OXA-48 increased sensitivity and improved detection of this important group of strains.