Is pharmacokinetic variability in microdosing trials comparable to variability following therapeutic doses?
Ivelina Gueorguieva1 and Malcolm Rowland2
1 Lilly Research Laboratories, Department of Drug Disposition, Lilly Research Laboratories, Sunninghill Road, Windlesham, Surrey, UK and 2 Centre for Applied Pharmacokinetic Research, School of Pharmacy and Pharmaceutical Sciences, The University of Manchester, Manchester, UK
Objectives: Following a single ascending dose trial, knowledge on within and between subject variability around Cmax and AUC0-inf is often required to adhere to a certain margin of safety and when necessary to design drug-drug interaction trials. Accepting that microdosing can be used to understand mean pharmacokinetic behaviour (Lappin et al., 2006), there is currently no investigation on whether variability following microdosing is comparable to that after therapeutic doses. The objective of this analysis is to quantify observed pharmacokinetic variability in microdosing trials (CREAM). This is compared to variability from both similar small sample size trials in the same subjects as well as larger trials following therapeutic doses.
Methods: Data from the CREAM trial (Lappin et al., 2006) was used, where warfarin, diazepam and midazolam were administered as microdoses and pharmacologic doses to healthy volunteers in a cross over trial. Additionally, literature and in-house data following therapeutic dose administration to a larger size healthy volunteer population were available. Population PK compartment models were fitted using nonlinear mixed effects models as well as naïve pooling of data in NONMEM, version V. Between and within subject variability were quantified and confidence intervals around the mean population estimates as well as prediction intervals were projected.
Results: A two-compartment model with iv infusion best described diazepam concentration-time data following microdose and therapeutic doses. The predicted mean CL and Vss from the microdose and therapeutic doses were comparable. The estimated, albeit from differing sample sizes, between-subject variability were also consistent at approximately 26% for exposure. Prediction intervals from micro, therapeutic doses and large dataset (with therapeutic dose) were very close and were plotted together with dose normalized data. Similar analyses were performed for warfarin and midazolam.
Conclusions: It was previously shown that for 5 drugs, two of which not examined here, mean pharmacokinetic behaviour (Lappin et al., 2006) was comparable following therapeutic and micro doses. Additionally a demonstration of consistent variability for a large number of compounds will lead to acceptance of microdosing as a Phase 0 trial to help (1) decide whether a compound is likely to have the desired pharmacokinetic behaviour and (2) plan the ascending dose ranging study. This is undoubtedly desirable with Phase 0 trials requiring a much reduced safety testing package by regulatory agencies and with that quicker access to human testing.
 Lappin G et al., Use of microdosing to predict pharmacokinetics at the therapeutic dose: Experience with 5 drugs. Clinical Pharmacology and Therapeutics 80 (3):203-215, 2006.
Acknowledgement We acknowledge CREAM and EUMAPP consortia for providing the data.