EXPERIMENTAL-MODEL OF STREPTOCOCCUS-PNEUM ONIAE INFECTIONS WITH DIMINISHED SENSITIVITY OF PENICILLIN-G - ANALYSIS OF AMOXICILLIN ACTIVITY

The number of strains with diminished sensitivity to penicillin G has gradually increased in France, along with the MIC values of all betala ctam agents. We analysed several studies of amoxicillin efficacy again st such strains in animal models. Coherent results were obtained despi te differences in disease severity and end points. In a murine model o f severe pneumonia used to study the bactericidal activity of a single injection of amoxicillin, the amoxicillin MIC was predictive of bacte rial killing in the lungs: increasing MIC values necessitated increasi ng doses to achieve 50% of the maximum bactericidal effect (P 50). The P 50/CMI curve for this model shows a kink between 2 and 4 mg/l, poss ibily owing to tolerance phenomena linked to the resistance phenotype. To obtain pulmonary bactericidal efficacy on strains with amoxicillin MICs of 4 mg/l or more similar to that observed with sensitive strain s, the P 50 value must be markedly increased. In models simulating hum an serum concentrations, an oral regimen of 500 mg three times daily i s effective on strains with MIC values of 2 mg/l or less, in both resp iratory tract and thigh infection models. In contrast, this regimen is ineffective against strains with amoxicillin MICs of 4 mg/l or more. The interval between two doses, during which amoxicillin concentration s are above the MIC for the infecting strain (Delta t > MIG) seems to be the pharmacodynamic parameter correlating most closely with amoxici llin efficacy in vivo. These results are coherent with published clini cal data and validate the conclusions of the 1991 consensus conference of the treatment of respiratory tract infections, which recommended a dose of 1 gram every 8 hours for the treatment of pneumococcal pneumo nia, given the emergence of strains with diminished sensitivity to pen icillin G in France.