Background: Recent open label studies have suggested that dosing amphotericin B (AMB) by continuous infusion (CI) may reduce drug-associated infusion reactions and nephrotoxicity. In vitro and in vivo pharmacodynamic (PD) data, however, do not consistently support the concept of CI dosing based on the concentration-dependent activity of this agent and in vitro studies with AMB rarely account for the drug's high degree of protein binding. Therefore, we compared the PD activity of simulated continuous versus rapid infusion strategies of AMB in killing of AMB-susceptible and -resistant Candida species using an in vitro pharmacodynamic model.
Methods: Time-kill curves were performed with Candida albicans (Etest MIC 0.38 mg/L) and Candida lusitaniae (MIC 1.5 mg/L) at AMB concentrations between 0 and 16 mg/L in the absence and presence of 4 and 8% human serum albumin (HSA). A one-compartment in vitro pharmacodynamic model was used to simulate the steady-state PK parameters of bolus and CI AMB.
Results: The fungicidal activity of AMB was attenuated by the presence of HSA for both Candida species tested. The EC50 for each isolate significantly increased in the presence of 4% HSA (P<0.05), and fungicidal activity was completely abated for C. lusitaniae when HSA concentrations were increased to 8%. No substantial differences in the rate or extent of AMB killing were observed between rapid infusion or CI dosing and neither regimen produced fungicidal activity in the presence of HSA.
Conclusions: The presence of HSA changes the in vitro PD of AMB. In our model, CI and rapid infusion dosing of AMB exhibited similar activity when attempts were made to correct for protein binding that is likely to occur in vivo.