S. Jönsson and M.O. Karlsson
Div. of Biopharmaceutics and Pharmacokinetics, Dept. of Pharmacy, Uppsala University, Sweden
Introduction: Drugs mainly eliminated by renal excretion are usually dose adjusted based on creatinine clearance (CLcr). The crucial questions in dose adjustment are at what CLcr to decrease the dose and what dose decrement(s) to use. We address these questions by estimation of optimal dosing strategies given the characteristics of the patient population, pharmacokinetic characteristics and therapeutic goal.
Purpose: To propose a method of dose algorithm selection in renal impairment that is minimising a therapeutically defined penalty function given known target population characteristics and drug pharmacokinetics.
Methods: The influence of a number of factors regarding renal function (median, SD and range of Clcr), pharmacokinetics (fraction excreted unchanged, unexplained interindividual variability in renal and non-renal CL) and penalty function (symmetric or non-symmetric, quadric or non-quadric penalty function for deviation from the target AUC) on the key features of a generic dosing algorithm were estimated.
Results and conclusions: The fraction excreted unchanged and the median CLcr in the target population are the two most important factors when defining the dosage regimen in terms of Clcr cut-off(s) and dose ratio(s).
Extending the number of Clcr cut-offs decreased the overall variability in AUC with the most significant effect when going from 1 to 2 cut-offs.
The use of different penalty functions had a slight effect on the choice of doses whereas the estimated cut-offs were mainly unaffected. Non-symmetric penalty functions i.e. assigning penalty to either high or low AUCs, resulted in higher overall variability in exposure but with shifted AUC distributions towards low or high AUCs, respectively.
The proposed approach may be used for generalisation providing that some basic information is available for the drug of interest.
Reference: PAGE 8 () Abstr 156 [www.page-meeting.org/?abstract=156]
Poster: poster