Choosing the Allometric Exponent in Covariate Model Building

Area of Science:

• Pharmacokinetics and Pharmacodynamics
• Biostatistics
• Drug Development

Background:

• Allometric scaling is a common method for modeling the relationship between total body weight (WT) and clearance (CL).
• There is a lack of consensus regarding the appropriate selection of the allometric exponent value in these models.

Purpose of the Study:

• To evaluate how between-subject variability (BSV) and study design influence the accuracy of selecting an a priori allometric exponent.
• To assess the impact of BSV and study design on obtaining unbiased estimates of the allometric exponent.

Main Methods:

• Stochastic simulation estimation was employed to assess the influence of WT distribution range, sample size (N=10 to 1000), and BSV (20% to 60%) on exponent selection and estimation.
• Simulations used a priori exponent values of 0.67, 0.75, and 1.

Main Results:

• Correctly selecting the allometric exponent from a priori values is difficult for drugs with normal to high BSV in studies with fewer than 200 adult participants.
• Regular study designs with < 200 participants are adequate for estimating the exponent, but designs with < 100 participants risk biased estimates.

Conclusions:

• Study designs with limited sample size and narrow WT range (<100 adults) risk inaccurate exponent selection or biased estimates.
• Bias in allometric exponent estimation is primarily a concern when extrapolating clearance values outside the studied population, such as from adults to children.