Multicenter collaborative evaluation of a standardized serum bactericidal test as a prognostic indicator in infective endocarditis.

One hundred twenty-nine patients with bacterial endocarditis were evaluated in a multicenter collaborative study to determine whether a standardized serum bactericidal test could predict the outcome of the infection. All centers used a microdilution test method that defined all known test variables, including inoculum size, culture medium, dilution technique, incubation time, method of subculture, and bactericidal endpoint. Peak serum bactericidal titers of 1:64 or more and trough serum bactericidal titers of 1:32 or more predicted bacteriologic cure in all patients. The traditionally recommended serum bactericidal titer of 1:8 had statistically significant predictive accuracy at trough antibiotic levels only. The serum bactericidal test was a poor predictor of bacteriologic failure and ultimate clinical outcome, which depends on many factors. Wider recognition by physicians and clinical microbiologists that this in vitro test of antimicrobial activity can accurately predict bacteriologic success but cannot accurately predict either bacteriologic failure or clinical outcome could lead to a better consensus about its appropriate use. On the basis of the results of this study, peak serum bactericidal titers of 1:64 or more and trough serum bactericidal titers of 1:32 or more are recommended to provide optimal medical therapy for infective endocarditis.