ABSTRACT
One hundred ninety eight clinical isolates, including Enterobacteriaceae (70 strains), Pseudomonas aeruginosa (20 strains), Acinetobacter baumannii (10 strains), staphylococci (50 strains), enterococci (20 strains), Streptococcus pneumoniae (15 strains) and Haemophilus influenzae (13 strains) were tested for their antimicrobial susceptibilities by using the Rapid Lumi 'Eiken' (RL). As a reference method, broth microdilution method according to the National Committee for Clinical Laboratory Standards was used. Then, each MIC obtained by both of these methods was compared. In order to improve the discrepancy between MICs obtained by both methods, modification of the RL method was studied. All MICs using the RL method were obtained with an incubation period of 4 hours. The essential agreement (to within one twofold dilution) between MICs obtained by both methods was 82%. The false susceptible in the RL method test results (very major error) and the false resistant in the RL tests (major error) were 0.9% and 2.3%, respectively. The agreement of interpretive category (that is, when the categories obtained by both methods are in perfect agreement) was 89%. Proteus spp. and A. baumannii showed low essential agreements, 59% and 46% respectively. The differences were resulted from the RL method's MICs being higher than the reference method. In order to improve the difference between both methods, the RL method's procedure was modified in the inoculum size (10(6) CFU/mL to 10(5) CFU/mL), the menadione concentration (5 mg/L to 25 mg/L) and the interpretive criteria for Enterobacteriaceae and A. baumannii. As the results of the modification, the essential agreement in Proteus spp. and A. baumannii increased to 82% and 81%, respectively, and there was no significant change in the other species of Enterobacteriaceae. In the case of the modified RL method to Enterobacteriaceae and A. baumannii, the essential agreement, the very major error, the major error and the agreement of interpretive category of all 198 strains were 87%, 1.4%, 1.5% and 90%, respectively. In conclusion, with only 4-hour incubation period, the RL method based on chemiluminescent assay gave reliable susceptibility testing among the most clinically important bacteria. Although several tests showed low essential agreement, it was possible to improve by use of the modified RL method. The Rapid Lumi 'Eiken' will provide useful information for choosing the most effective antibiotic for primary treatment to bacterial infections.
