Update on the statistical analysis of bioequivalence studies

Area of Science:

• Pharmacokinetics and Biopharmaceutics
• Statistical Methods in Drug Development

Background:

• Assessing bioequivalence between test and reference drug formulations is critical in pharmaceutical development.
• Statistical methods must balance consumer risk (accepting non-bioequivalent drugs) and producer risk (rejecting bioequivalent drugs).

Purpose of the Study:

• To review statistical methods for bioequivalence assessment.
• To select a method that minimizes both consumer and producer risks, not exceeding a 5% nominal consumer risk.

Main Methods:

• Evaluation of statistical procedures for bioequivalence, considering distribution of characteristics and associated risks.
• Selection of a preferred method based on risk-benefit analysis.
• Application of a 90%-confidence interval strategy for the ratio of expected medians.
• Consideration of nonparametric (distribution-free) methods for non-log-normal distributions.
• Discussion on modifying bioequivalence ranges for characteristics like Cmax.

Main Results:

• A recommended strategy uses the shortest 90%-confidence interval for the ratio of expected medians within the bioequivalence range (80-120%) for rate and extent of absorption.
• Nonparametric confidence intervals are recommended when log-normal distribution assumptions are violated.
• A decision rule for time to maximum concentration (tmax) is also presented.

Conclusions:

• The proposed statistical strategy ensures an acceptably low consumer risk for bioequivalence assessment.
• The method is robust, accommodating both normal and non-normal distribution assumptions.
• The approach provides a framework for evaluating bioequivalence of drug formulations effectively.