A Simulation Approach to Evaluate the Impact of Patterns of Bioanalytical Bias Difference on the Outcome of Pharmacokinetic Similarity Studies

Pharmacokinetic (PK) similarity studies are vital to assess the biosimilarity of a biosimilar to a reference product. Systematic bias in a bioanalytical method that quantify products could be a potential source of error affecting the variability of the data and influencing the outcome of a PK similarity study. We investigated the impact of six varying patterns of bioanalytical bias difference (biasdiff) between the similar products on the probability passing the PK similarity test. A population PK model was used to simulate concentration‐time profiles for a biosimilar and a reference product and added biasdiff ranging from 030%. The probability of achieving the PK similarity criteria (90% confidence interval between 0.8 and 1.25) for the maximum serum concentration (Cmax) and area under the curve (AUC) was assessed. The data indicate that an increase in absolute biasdiff between products of ≥ 10% would decrease the power to assess the similarity criteria for Cmax and AUC.