Statistical aspects of bioequivalence--a review

Area of Science:

• Pharmacokinetics and Biopharmaceutics
• Regulatory Science
• Statistical Methods in Clinical Trials

Background:

• Bioequivalence testing is crucial for generic drug approval, ensuring therapeutic equivalence to reference products.
• Over the past two decades, numerous statistical methodologies have been developed for bioequivalence assessment.
• Regulatory authorities' current thinking on these methods is essential for study design and interpretation.

Purpose of the Study:

• To review and analyze statistical methods employed in bioequivalence testing.
• To reflect the current perspectives of regulatory authorities on these methodologies.
• To provide guidance on appropriate statistical approaches for bioequivalence studies.

Main Methods:

• Examination of statistical methods proposed for bioequivalence testing over the last 20 years.
• Analysis of standard bioequivalence study designs, typically single-dose, controlled crossover trials.
• Focus on pharmacokinetic parameters such as Cmax, tmax, and AUC derived from blood or urine samples.

Main Results:

• Classical hypothesis testing using the power approach is not directly applicable to bioequivalence trials.
• Logarithmic transformation is recommended for Cmax and AUC parameters before analysis.
• Non-parametric confidence intervals are advised for tmax when clinically relevant.

Conclusions:

• The 90% confidence intervals and two one-sided t-test approach are the preferred methods for assessing bioequivalence of Cmax and AUC.
• Standard bioequivalence studies involve crossover designs and pharmacokinetic parameter analysis.
• Appropriate statistical methods, including data transformation and choice of confidence intervals, are critical for accurate bioequivalence assessment.