Detecting carbapenemase-producing Enterobacteriaceae (CPE) has become increasingly difficult due to the emergence of diverse enzymes. The aim of the study was to evaluate an agar plate-based modified carbapenem inactivation method (p-mCIM) for detection of CPE. Stock strains and clinical isolates of CPE were used to evaluate the p-mCIM. The p-mCIM was performed as described for the mCIM, except that meropenem disks were placed on the lawn of test organisms on Mueller-Hinton agar (MHA) plates. Among 17 stock strains of CPE, six of eight KPC-2-like- and all six NDM-1-like carbapenemase-producing strains were positive by the p-mCIM without incubation in the carbapenem inactivation (CI) step. Among 380 CPE clinical isolates detected, 308 and 38 were KPC-2-like and NDM-1-like enzyme producers, respectively. The required incubation time in the CI step to show all isolates were positive by p-mCIM was 3 h for isolates with KPC-2-like enzyme and 1 h for isolates with metallo-beta-lactamases. Twenty-eight of 30 isolates with OXA-48-like enzymes were p-mCIM positive. Sensitivities of both the p-mCIM and the mCIM (based on inhibition zone of <= 15 mm) for detection of CPE were 100%. All 70 ertapenem-nonsusceptible, but carbapenemase gene-negative isolates tested were both p-mCIM (based on inhibition zone of >= 21 mm) and mCIM negative. In conclusion, performance of the p-mCIM, which uses a lawn of bacterial colonies on MHA plate instead of a bacteria-suspended Tryptic soy broth tube in the CI step, is essentially identical to that of the CLSI-recommended mCIM in the detection of clinical isolates of Enterobacteriaceae producing carbapenemases including difficult to detect bla(OXA-48)-like enzymes.