ABSTRACT
Bloodstream infections are life-threatening conditions which require timely initiation of appropriate antimicrobial therapy. The accuracy of direct disk diffusion susceptibility testing of positive blood cultures was investigated, including for the first time beta-lactam/beta-lactam-inhibitor combination antibiotics. Results of direct testing, following the guidelines of the Clinical and Laboratory Standards Institute, were compared to standard microtitre broth dilution susceptibility testing of the subcultured isolate on the Merlin MICRONAUT system. Altogether, 758 isolates and 4930 organism/antibiotic combinations from 590 patients were evaluated. With regard to Gram-positive cocci (n=532), agreement between both methods was found in 93.9% of cases, with 1.6% very major, 1.1% major and 2.6% minor errors. For Gram-negative rods (n=226), agreement was found in 91.9% of cases, with 1.2% very major, 0.7% major and 6.3% minor errors. When applying the breakpoints of the Deutsches Institut für Normung for interpretation of MICRONAUT tests, agreement of direct disk diffusion with standard testing decreased to 82.4% in Gram-negative rods, with 3.6% very major, 0.5% major and 13.4% minor errors. A high rate of disagreement was observed with oxacillin and gentamicin in Gram-positive cocci, and with cefuroxime, amoxycillin/clavulanate and piperacillin/tazobactam in Gram-negative rods. In conclusion, the limitations of direct disk diffusion testing of positive blood cultures must be kept in mind by the clinical microbiologist and should, where necessary, be communicated to the clinician to ensure adequate treatment of severely ill patients.
