Vol. 280, Issue 1, 310-315, 1997

Correlation Between In Vitro and In Vivo Activity of Amoxicillin Against Streptococcus pneumoniae in a Murine Pneumonia Model

Pierre Moine , Jean-Xavier Mazoit, Jean-Pierre Bédos, Éric Vallée and Esther Azoulay-Dupuis

Département d'Anesthésie-Réanimation Chirurgicale, Université Paris Sud, Centre Hospitalier de Bicêtre, Le Kremlin-Bicêtre, France (P.M., J.-X.M.), and Institut National de la Santé et de la Recherche Médicale U13, Hôpital Claude-Bernard-Bichât, Paris, France (P.M., J.-P.B., E.V., E.A.-D.)

We studied the relationship between in vitro bacteriological parameters [minimal inhibitory concentration (MIC), minimal bactericidal concentration (MBC) and killing rate, defined as the reduction in the inoculum within 1, 3 or 6 hr] and in vivo activity of amoxicillin against 12 strains of Streptococcus pneumoniae, with penicillin MICs of <0.01 to 16 µg/ml, in a cyclophosphamide-induced neutropenic murine pneumonia model. Dose-response curves were determined for amoxicillin against each strain, and three quantitative parameters of in vivo amoxicillin activity were defined, i.e., maximal attainable antimicrobial effect attributable to the drug [i.e., reduction in log colony-forming units (CFU) per lung, compared with untreated controls], dose required to reach 50% of maximal effect and dose required to achieve a reduction of 1 log CFU/lung. We demonstrated a highly significant correlation between the dose required to reach 50% of maximal effect and MIC (Spearman r = 0.98, P < .0001) or MBC (Spearman r = 0.95, P < .0001) for amoxicillin against strains of S. pneumoniae with a wide range of amoxicillin MICs (0.01-8 µg/ml). Significant correlations between the dose required to achieve a reduction of 1 log CFU/lung and MIC (Spearman r = 0.98, P < .0001) or MBC (Spearman r = 0.95, P < .0001) were also observed. In contrast, there were no significant correlations between the maximal attainable antimicrobial effect attributable to the drug and MIC, MBC or killing rate or between killing rate and the dose required to reach 50% of maximal effect or the dose required to achieve a reduction of 1 log CFU/lung. We conclude that in vitro susceptibility test results (MICs and MBCs) correlated well with in vivo amoxicillin activity against pneumococcal strains, including highly penicillin-resistant strains, in this animal model. Furthermore, these data suggest that the estimated MIC breakpoints for amoxicillin against S. pneumoniae would be 2 µg/ml for intermediate-resistant and 4 µg/ml for resistant, although this remains to be confirmed in clinical studies.