OBJECTIVES MALDI-TOF MS has been successfully used for empirical antibiotic selection. However, limited data are available regarding the usefulness of MALDI-TOF MS in common resistant organisms compared with rapid antimicrobial susceptibility testing (AST). We prospectively evaluated the usefulness of rapid AST, compared with MALDI-TOF MS, for optimal antibiotic selection by infectious disease (ID) physicians in patients with bacteraemia including polymicrobial infection.

METHODS Three hundred and fifty-nine patients with positive blood culture were included for analysis. ID physicians prospectively decided on antibiotic regimens with consensus at each timepoint of receiving results of Gram stain, MALDI-TOF MS and rapid AST, the last of which was performed using QMAC-dRAST.

RESULTS ID physicians with MALDI-TOF MS results chose optimal targeted antibiotics in 255 (71.0%) cases, with appropriate antibiotic selection in 303 (84.4%) cases. The proportion of optimal targeted antibiotic selection and appropriate antibiotic selection was significantly lower for resistant strains than for susceptible strains [62.5% versus 79.2% (P < 0.001) and 68.2% versus 100% (P < 0.001), respectively]. QMAC-dRAST results led to optimal antibiotic treatment in 95 (91.3%) of the 104 cases receiving non-optimal targeted antibiotics. Optimal targeted treatments based on QMAC-dRAST results were possible in 322 (98.2%) of the 328 cases with monobacterial infection and in 345 (96.1%) of the 359 cases with monobacterial and polymicrobial infection.

CONCLUSIONS MALDI-TOF MS has a high chance of failure in guiding ID physicians to optimal antibiotics, especially against resistant organisms. With increasingly common resistant organisms, rapid AST is needed to identify optimal targeted antibiotics early in bacteraemia.