Premium
Optimal sparse pharmacokinetic (PK) sampling strategies (OSS) for multi‐drug antiretroviral (ARV) regimens
Author(s) -
Zack J. Z.,
Forrest A.,
Ma Q.,
Rosenkranz S.,
Para M. F.,
Adams E.,
Reichman R. C.,
Yaresheski K.,
Morse G. D.
Publication year - 2005
Publication title -
clinical pharmacology and therapeutics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.941
H-Index - 188
eISSN - 1532-6535
pISSN - 0009-9236
DOI - 10.1016/j.clpt.2004.12.236
Subject(s) - efavirenz , saquinavir , pharmacology , clinical pharmacology , medicine , human immunodeficiency virus (hiv) , drug , population , nonmem , pharmacokinetics , amprenavir , sampling (signal processing) , antiretroviral therapy , viral load , computer science , virology , chemistry , environmental health , filter (signal processing) , computer vision , biochemistry , hiv 1 protease , protease , enzyme
Aims ARV therapy is a rapidly evolving area that requires an excellent understanding of drug interactions & metabolism as well as adaptive feedback control (PK/PD individualization). Clinicians & patients are confronted with increasingly complex regimens in order to optimize treatment & to prevent emergence of resistance, failure or toxicity. ARV regimens often include 3 or 4 drugs, making population PK/PD analysis based on traditional sampling strategies difficult. Our goal was to develop a sparse efficient multi‐drug OSS for the evaluation of PK in patients receiving ARV regimens. Methods Saquinavir (SQV) & efavirenz (EFV) data from 10 healthy subjects also on amprenavir, in AACTG A5043, were fit to compartmental models (maximum likelihood, ADAPT II). Optimal Sampling Theory (ADAPT II, D‐optimality) was used to determine sparse, efficient study designs (best 6‐, 4‐, 3‐ & 2‐ sample OSS) constrained to 8h post‐dose, at steady state. Bias & precision of clearance (CL) estimates were evaluated. Results OSS3 included samples at 0, 4 & 8h post‐dose; OSS4 added a 6h & OSS6 added 1& 2h samples. CL estimates for all OSSs were unbiased & even OSS3 provided acceptable precision; mean absolute errors, with OSS6, for SQV & EFV were 14.9 & 4.1% respectively & with OSS3, for SQV & EFV were 16.9 & 5.4%. No OSS with <3 samples was adequate for SQV. Conclusions CL for both SQV & EFV were reasonably estimated using ≥ 3 sample strategy. This approach will facilitate population PK/PD analyses of combined ARV regimens. Clinical Pharmacology & Therapeutics (2005) 77 , P90–P90; doi: 10.1016/j.clpt.2004.12.236