Performance and reliability of VITEK 2-Compact GN13 method and its Advanced Expert System validation for testing imipenem susceptibility of Klebsiella pneumoniae / 中国感染与化疗杂志

ABSTRACT

Objective To evaluate the performance of VITEK 2-Compact GN13 methods for testing imipenem susceptibility of Klebsiella pneumoniae and assess the reliability of its Advanced Expert System (AES).Methods A retrospective study was conducted with a total of 157K. pneumoniae strains, which were isolated from blood and intra-abdominal infections in the First Affiliated Hospital of Nanchang University from 2014 to 2015. Thein vitro minimum inhibitory concentration (MIC) values of imipenem were determined by disc diffusion, VITEK 2-Compact GN13 and broth microdilution methods, respectively. Categorical agreement (CA) rates of disc diffusion and VITEK 2-Compact GN13 methods were determined using broth microdilution as reference method. The genes encoding ESBLs and carbapenemase were screened by PCR and sequencing analysis. The phenotypic confirmatory tests such as modified Hodge test, PCR and DNA sequencing were used to confirm the resistance mechanism and evaluate the reliability of AES in interpreting the imipenem susceptibility of K. pneumoniae.Results Among the 157 isolates, 64 and 8 were identified as resistant and intermediate strains by broth microdilution method, respectively; 52 and 10 were tested as resistant and intermediate strains by disc diffusion method, respectively; 54 and 13 were determined as resistant and intermediate strains by VITEK 2-Compact GN13 method, respectively, while 70 and 3 were judged as resistant and intermediate strains by VITEK 2-Compact GN13 method plus AES validation. The CA of disc diffusion and VITEK 2-Compact GN13 methods compared with broth microdilution method were all higher than 90 %. However, the major error (ME) rate was 3.8 % and very major error (VME) rates were all 0.6 % in imipenem susceptibility testing by VITEK 2-Compact GN13 and disc diffusion. The imipenem susceptibility of 16 strains were modified by the AES, which eliminated 0.6 % VME, but increased major error by 1.3 % and minor error by 1.9 %. Phenotypic confirmatory tests showed that 75 % (12/16) of these strains were validated as producers of both ESBLs and carbapenemase, which was consistent with the result of AES validation. PCR and DNA sequencing analysis proved that 62.5 % (10/16) of these strains produce IMP-4/KPC-2 /NDM-1 and ESBLs.Conclusions Both disc diffusion and VITEK 2-Compact GN13 methods can be used for testing the imipenem susceptibility of K. pneumoniae isolates with reliable and accurate results. Attention should be paid to the possibility of ME and VME when testing imipenem susceptibility. The VME can be avoided by the AES mechanism. However, AES intervention will increase ME and minor error, which may be associated with decreased expression of carbapenemase.