Prospective allometric scaling: does the emperor have clothes?

Area of Science:

• Pharmacokinetics and Drug Development
• Translational Medicine
• Biostatistics

Background:

• Allometric scaling (AS) is frequently used to predict human pharmacokinetic (PK) parameters from animal data.
• However, the reliability and limitations of prospective AS in early-phase clinical trials require careful consideration.

Purpose of the Study:

• To evaluate the potential pitfalls and limitations of using prospective allometric scaling for dose selection in first-in-human (FTIM) studies.
• To highlight the impact of publication bias on the perceived success of AS.

Main Methods:

• The study critically examines the application of prospective AS in FTIM studies.
• It discusses common challenges including wide prediction intervals and prediction errors comparable to arbitrary constants.
• The authors address the difficulty in a priori identification of drugs likely to fail.

Main Results:

• Prospective AS may provide a misleading sense of security due to significant limitations.
• Prediction intervals are often too wide for practical application.
• Prediction error using AS is frequently not superior to using arbitrary constants.
• Identifying drugs likely to fail before FTIM studies remains challenging.

Conclusions:

• While an allometric relationship between animal and human PK exists, prospective AS for dose selection in FTIM studies has unrecognized limitations.
• Publication bias obscures the true failure rate of AS.
• Encouraging publication of failed AS studies is crucial for improving future prediction models and understanding drug-specific differences.