U. Fluckiger et al., SIMULATION OF AMOXICILLIN PHARMACOKINETICS IN HUMANS FOR THE PREVENTION OF STREPTOCOCCAL ENDOCARDITIS IN RATS, Antimicrobial agents and chemotherapy, 38(12), 1994, pp. 2846-2849
The pharmacokinetic determinants of successful antibiotic prophylaxis
of endocarditis are not precisely known. Differences in half-lives of
antibiotics between animals and humans preclude extrapolation of anima
l results to human situations. To overcome this limitation, we have mi
micked in rats the amoxicillin kinetics in humans following a 3-g oral
dose (as often used for prophylaxis of endocarditis) by delivering th
e drug through a computerized pump. Rats with catheter-induced vegetat
ions were challenged with either of two strains of antibiotic-tolerant
viridans group streptococci. Antibiotics were given either through th
e pump (to simulate the whole kinetic profile during prophylaxis in hu
mans) or as an intravenous bolus which imitated only the peak level of
amoxicillin (18 mg/liter) in human serum. Prophylaxis by intravenous
bolus was inoculum dependent and afforded a limited protection only in
rats challenged with the minimum inoculum size infecting greater than
or equal to 90% of untreated controls. In contrast, simulation of kin
etics in humans significantly protected animals challenged with 10 to
100 times the inoculum of either of the test organisms infecting great
er than or equal to 90% of untreated controls. Thus, simulation of the
profiles of amoxicillin prophylaxis in human serum was more efficacio
us than mere imitation of the transient peak level in rats. This confi
rms previous studies suggesting that the duration for which the serum
amoxicillin level remained detectable (not only the magnitude of the p
eak) was an important parameter in successful prophylaxis of endocardi
tis. The results also suggest that single-dose prophylaxis with 3 g of
amoxicillin in humans might be more effective than predicted by conve
ntional animal models in which only peak levels of antibiotic in human
serum were simulated.