Laboratory detection of intestinal carriage of carbapenemase-producing Enterobacteriaceae - A comparison of algorithms using the Carba NP test

•This study compared 4 testing protocols for detection of CPE from spiked stool specimens.•Algorithms using Brilliance ESBL media significantly outperformed chromID CARBA algorithms.•The difficulty detecting OXA-48-producers with the Carba NP test was a weakness of the testing.•This shortcoming was partly overcome with the addition of chromID OXA-48 media to the algorithms.•Selection of screening algorithms for CPE requires careful consideration of the local context.

Stool specimens spiked with a panel of 46 carbapenemase-producing Enterobacteriaceae (CPE) and 59 non-carbapenemase producers were used to compare the diagnostic accuracy of 4 testing algorithms for the detection of intestinal carriage of CPE: (1) culture on Brilliance ESBL agar followed by the Carba NP test; (2) Brilliance ESBL followed by the Carba NP test, plus chromID OXA-48 agar with no Carba NP test; (3) chromID CARBA agar followed by the Carba NP test; (4) chromID CARBA followed by the Carba NP test, plus chromID OXA-48 with no Carba NP test. All algorithms were 100% specific. When comparing algorithms (1) and (3), Brilliance ESBL agar followed by the Carba NP test was significantly more sensitive than the equivalent chromID CARBA algorithm at the lower of 2 inoculum strengths tested (84.8% versus 63.0%, respectively [P < 0.02]). With the addition of chromID OXA-48 agar, the sensitivity of these algorithms was marginally increased.