Several recent reviews on the therapy of bacterial endocarditis have recommended that a serum inhibitory and/or bactericidal concentration (SIC/SBC) of 1:8 or more be achieved to ensure successful therapeutic outcome. We conducted a methodologic and statistical analysis of the available literature on endocarditis to determine the association between SIC/SBC titers of 1:8 or more and therapeutic outcome. We reviewed 17 studies published between 1948 and 1980 in which both SIC/SBC and therapeutic outcome were available. Factors that affect outcome, such as age, duration of symptoms, organism, and valve status, varied widely among the 226 patients. The methods used to measure SIC/SBC differed with respect to the time of obtaining the blood specimen relative to the antibiotic dose, size of the bacterial inoculum, type of broth, and definition of the bactericidal end-point. None of the 17 studies showed a significant association between SIC/SBC titers of 1:8 or more and survival or bacteriologic cure. Fifteen of the 17 also failed to demonstrate a significant association between SIC/SBC titers of 1:8 or more and medical cure. However, each of the studies that failed to demonstrate an association between SIC/SBC titers of 1:8 or more and improved therapeutic outcome had an insufficient sample size to confidently exclude a false-negative result. Analysis of the published data reveals insufficient evidence to demonstrate that SIC and SBC titers are of prognostic value in the therapy of patients with bacterial endocarditis.