Correlation of antibacterial activities of antibiotics in vitro and in animal models of infection.

Laboratory tests of antibiotic activity in vitro, ranging from simple MIC determinations to sophisticated, computerized models for studying the effects of simulated plasma concentration-time curves of antibiotics in broth, differ inherently from tests in vivo in both the general and specific variables. There are therefore likely to be discrepancies between the results obtained. Apart from the obvious lack of host-defence mechanisms in the in-vitro tests systems, the reasons for these discrepancies have still not been fully elucidated. For the time being, it seems unrealistic to expect that in-vitro tests could be developed that would make it possible to predict the efficacy of any antibiotic against any specific infection in vivo.