Indirect bioequivalence assessment using network meta-analyses.

Area of Science:

• Pharmacokinetics and Drug Development
• Biostatistics
• Pharmaceutical Sciences

Background:

• Market approval requires demonstrating bioequivalence (BE) between new and reference drug formulations.
• Assessing BE becomes complex when multiple reference or test formulations are involved.
• Existing methods may not efficiently handle indirect comparisons across different trials.

Purpose of the Study:

• To present an enhanced statistical model for indirect bioequivalence assessment using network meta-analysis.
• To derive and compare the statistical properties and efficiency of parallel and bridging approaches.
• To illustrate the methodology with real-world pharmacokinetic data.

Main Methods:

• Development of an enhanced statistical model for network meta-analysis.
• Derivation of statistical properties for parallel and bridging estimation approaches.
• Application to individual subject data from two 3x3 crossover trials of metformin formulations.

Main Results:

• The bridging approach avoids confounding with between-trial differences, unlike the parallel approach.
• Bridging approach yields smaller standard errors when within-subject correlation is >0.5, indicating higher efficiency.
• This efficiency condition is typically met in pharmacokinetic crossover trials.

Conclusions:

• Indirect bioequivalence assessment via network meta-analysis efficiently determines relative bioavailability for formulations not directly compared.
• The developed methodology enables robust estimation of formulation differences within larger networks.
• This approach enhances the ability to compare multiple drug formulations indirectly.