Formation of a carbapenemase resistance detection algorithm for use in the routine laboratory

The increasing prevalence of carbapenem non-susceptible Gram negative organisms demands prompt and accurate identification of resistance mechanisms to limit their transmission. The aim of this study is to evaluate the rapid CARB screen (Rosco Diagnostica, Denmark), the 'KPC/MBL in P aeruginosa/Acinetobacter Confirm Kit' (Rosco Diagnostica), Check-MDR Carba and Check-Direct CPE kits (Check-Points, The Netherlands). The purpose of this study is the formation of a carbapenemase resistance detection algorithm that can be used in the routine laboratory Results of the rapid CARB screen kit were improved when isolates were tested from Muller Hinton agar with a meropenem 10 mu g disc instead of blood agar. The rapid CARB screen (performed in 2-3 h) demonstrated overall 98.7% sensitivity and 87.7% specificity (n=133). The KPC/MBL in P aeruginosa/Acinetobacter Confirm Kit (which requires overnight incubation) demonstrated a high number of false-positive results giving 38.6% specificity and 100% sensitivity (n=44). The Check-MDR Carba (performed in 5 h), detecting carbapenemase presence, provides a cabapenemase-positive or -negative result demonstrated 96.7% specificity and 98.6% sensitivity (n=132). The Check-Direct CPE (performed in 3 h), which identifies KPC, NDM/VIM or OXA-48 type carbapenemases, demonstrated 96.5% specificity and 97.1% sensitivity (n=97). The Check-Direct CPE, however, failed to detect dual carbapenemase genes present in two out of four isolates. The principal conclusion is the recommendation of the rapid CARB screen and Check-MDR Carba for incorporation into a carbapenemase detection algorithm which, when used in combination, will yield results with 97.3% sensitivity and 99.6% specificity.