Variables of the effect of penicillin G on the numbers of viable group B streptococci in broth cultures were studied. One-fourth of the MIC was the lowest concentration that reduced the viable count compared with antibiotic-free controls. The rate of killing increased with the concentration of penicillin up to 4x MIC, but no further. During the first 2 or 3 h, the bactericidal activity was more rapid than later on. The MIC and supraoptimal concentrations of penicillin killed an inoculum of 10' organisms more rapidly than an inoculum of 104 organisms. The MIC was not inoculum dependent. The reverse inoculum effect was revealed by the killing curves but not by the MBC. There were reproducible differences among strains as to the rate of killing by penicillin; these did not correlate with the rate of multiplication, which also varied among strains. Among the 11 strains tested, there were no tolerant ones.Huge numbers of antimicrobial susceptibility tests are performed daily to predict therapeutic efficacy. Such predictive value is known in only few instances (10). Most tests detect growth inhibitory activity, and dissociation of in vitro and in vivo results is to be expected when bactericidal activity is needed for cure, as is believed to be the case with infections of immunocompromised patients or with infections, like endocarditis, of immunologically privileged sites. Group B streptococci are susceptible to penicillin in the conventional sense, but neonatal infections have turned out to be difficult to treat, and failures have been ascribed to penicillin tolerance, i.e., relative resistance to the bactericidal effect (4, 23).The description of bactericidal activity has remained a problem. The most widely used parameter is the MBC, and we have characterized several variables of the effect of penicillin on the numbers of viable group B streptococci to see to what extent the MBC reveals all relevant information about bactericidal activity. The model combination was chosen, because it has been studied extensively before, because it is one of the combinations with which phenomena like paradoxically reduced killing at high antibiotic concentrations (7) and antibiotic tolerance have been described, and because it has been associated with therapeutic failures.
MATERIALS AND METHODSBacteria. Group B streptococci were isolated from clinical specimens from various male and female genital, throat, nose, ear, and skin infections as the predominant probable pathogen. They were identified by three presumptive tests, hippurate hydrolysis, cAMP reaction, and pigment production, as well as serologically, as described earlier (14). Thirty-five isolates were selected randomly for these studies from a collection maintained lyophilized and by serial passage on blood agar plates.Inoculum. In preliminary studies of 18-to 21-h growth curves, supplementing Todd-Hewitt broth (BBL Microbiology Systems, Cockeysville, Md.) with 0.8% glucose reduced the growth of group B streptococci, and 10% horse serum had essentially no effect. Thus, plai...