BACKGROUND: The existing modified carbapenem inactivation methods (mCIMs) recommended by the CLSI for detecting carbapenemase production have not been applicable for Acinetobacter baumannii. We evaluated the influence of matrices used in mCIMs and CIMTris on the stability of the disks for detecting carbapenemase producers and suggested optimal mCIM conditions for detecting carbapenemase-producing A. baumannii. METHODS: Seventy-three A. baumannii isolates characterized for antimicrobial susceptibility and carbapenemase encoding genes were tested for carbapenemase production using mCIM and CIMTris. The influence of the matrices (Tryptic soy broth [TSB] and Tris-HCl) used in these methods on the stability of the meropenem (MEM) disk was also evaluated. The mCIM conditions were adjusted to enhance screening sensitivity and specificity for detecting carbapenemase-producing A. baumannii. RESULTS: The matrices had an impact on the stability of the MEM disk after the incubation period (two or four hrs). TSB nutrient broth is an appropriate matrix for mCIM compared with Tris-HCl pH 7.6, which leads to the loss of MEM activity in CIMTris. The sensitivity and the specificity of the optimal mCIM were both 100%. CONCLUSIONS We established optimal mCIM conditions for simple, accurate, and reproducible detection of carbapenemase-producing A. baumannii.