Highly variable drugs (HVDs) are a category within pharmaceuticals that exhibit significant intra-subject variability in pharmacokinetic parameters such as peak concentration (Cmax) and the area under the concentration-time curve (AUC).
Specifically, a drug is classified as highly variable when the intra-subject variability, represented by the coefficient of variation (CV), exceeds 30%. This variability indicates that when the same individual takes the same drug under similar conditions at different times, the absorption rate and extent can differ substantially.
The presence of high variability in these drugs presents substantial challenges in demonstrating bioequivalence, which is a prerequisite for the approval of generic drugs. In the conventional Average Bioequivalence (ABE) framework, the 90% confidence interval for the ratio of the geometric means (test or reference) for AUC and Cmax must fall within the preset limits of 80% to 125%. However, for HVDs, achieving this standard can be particularly challenging.
The high intra-subject variability inherent to HVDs means that there is a significant chance that the measured pharmacokinetic parameters will fluctuate widely across different administrations in the same individual. This fluctuation can lead to a broad spread of data, making it difficult to conclusively prove that the test and reference drugs are equivalent within the stringent ABE limits. Conventional ABE studies with typical sample sizes may not have sufficient statistical power to consistently fall within these narrow confidence intervals, leading to a higher likelihood of Type II errors—incorrectly concluding that two equivalent drugs are not equivalent.
Furthermore, conventional ABE methods, which generally involve a fixed sample size and do not account for the variability specific to the drug being tested, are often inadequate for HVDs because they are not designed to handle the extremes of pharmacokinetic variability. As a result, generic versions of highly variable drugs may fail to show bioequivalence using standard ABE testing, not due to actual differences in formulation effectiveness but because of the statistical complications posed by high variability.
This inadequacy has prompted the development of alternative methodologies like Reference-Scaled Average Bioequivalence (RSABE), which are specifically tailored to address the challenges posed by HVDs. The RSABE methods adjust the bioequivalence criteria based on the within-subject variability of the reference product. This scaling allows for wider bioequivalence limits if the reference drug exhibits high variability, thus providing a more realistic and achievable framework for demonstrating bioequivalence in HVDs.
Reference-Scaled Average Bioequivalence (RSABE)
RSABE is an advanced statistical methodology designed to address the unique challenges posed by HVDs in demonstrating bioequivalence. RSABE adjusts the bioequivalence criteria based on the within-subject variability of the reference drug, offering a more flexible and accurate approach for assessing the equivalence of generic drugs where traditional average bioequivalence (ABE) methods may not be suitable.
RSABE primarily differs from traditional ABE in that it allows for the scaling of bioequivalence limits according to the variability of the reference drug. This scaling is crucial for drugs where intra-subject variability is high, typically greater than 30% CV (coefficient of variation). Under RSABE, the bioequivalence window is widened based on a scaling factor derived from the within-subject standard deviation of the reference drug (SWR). The fundamental concept here is that as the variability of the reference drug increases, the acceptable range for demonstrating bioequivalence also expands, allowing for a fair assessment of the generic drug’s performance relative to the reference.
The RSABE approach often involves a replicated crossover design where subjects receive the reference drug more than once. This design enables the accurate estimation of within-subject variability, which is critical for scaling the bioequivalence limits. The statistical model used in RSABE incorporates this variability to adjust the confidence interval for the ratio of the geometric means of the test and reference drugs’ pharmacokinetic parameters (mainly AUC and Cmax).
Read also:
Resource Person: Chandramouli R