Variability and Highly Variable Drugs in Bioequivalence Studies
When we look at the result of a bioequivalence study, we often focus on a single question: are the average values of the two products close to each other? Yet the average is only part of the story. Two products with the same average value may behave entirely differently in terms of the spread around that value. In this article we look at why variability matters so much in bioequivalence studies and why some active ingredients require a special approach.
Two kinds of variability
Variability appears in two distinct forms in bioequivalence studies. The first is between-subject variability: it describes how much different people taking the same product differ from one another. The second is within-subject variability: it shows how much the response of the same person to the same product fluctuates at different times.
This distinction matters, because what truly governs the assessment of bioequivalence is within-subject variability. This is also the purpose of the crossover design described in the previous article: by making each participant their own control, it sets aside between-subject differences and bases the comparison on within-subject variability.
Low bioavailability, high variability
Variability is not distributed at random; it is related to the properties of the active ingredient. Investigations have shown an inverse relationship between the bioavailability of a substance and the variability between individuals. In other words, the lower the absorbed fraction of a substance, the higher the variability observed from one person to another tends to be.
The logic behind this is intuitive. For a substance that is largely broken down after entering the body, or of which only a small portion is absorbed, even small differences lead to large proportional deviations in the result. Substances that undergo an extensive first-pass effect are a typical example. Interestingly, this relationship has been found to be largely independent of the sex or type of the participants; what determines variability is primarily the behavior of the substance itself.
What is a highly variable drug?
Some active ingredients show high variability by their very nature. If the response a substance produces in the same person, under the same conditions, at different times fluctuates markedly, that substance is considered highly variable. By the common definition, substances whose within-subject variability is thirty percent or above fall into this class.
This variability often arises from the absorption stage of the substance. The same substance may show low variability when given intravenously, yet become highly variable when taken by mouth. This indicates that the variability largely originates in the absorption process.
Why the standard approach struggles
The classical bioequivalence assessment expects the ninety percent confidence interval calculated for the ratio of the geometric means of two products to fall between eighty percent and one hundred twenty-five percent. This criterion works well for most substances. However, in a substance with high within-subject variability the confidence interval naturally widens, and the likelihood of crossing these limits increases.
The result is this: even if two products are in fact identical, the study may “fail” because of the substance’s own high variability. In that case the number of participants required grows to an impractical size. The problem, then, is not that the products differ, but the variability inherent in the substance.
The solution: a scaled approach
To remove this unfairness, the scaled average bioequivalence approach was developed. The core idea here is to adjust the acceptance limits according to the substance’s own variability, rather than keeping them fixed. If the within-subject variability of the reference product is high, the acceptance limits are widened in proportion to that variability.
This approach makes it possible to assess whether two products are truly equivalent without penalizing the variability inherent in the substance. For the scaled approach to be applied, the variability of the reference must be measured; this in turn requires replicate designs in which each product is given more than once to the same person.
Measuring variability accurately
The precondition for all of this assessment is that variability is measured accurately. To capture the true variability of a substance, the concentrations in the blood must be measured consistently and reproducibly even at low levels. If the measurement method itself becomes a source of variability, the substance’s true variability and the noise coming from the method become mixed, and the result loses its reliability.
At this point the reproducibility of the measurement becomes decisive. As an authorized bioequivalence center, NanoTox conducts its LC-MS/MS measurements with validated methods, minimizing the variability that comes from the method, so that only the substance’s true variability remains. In highly variable substances in particular, this consistency of the measurement directly determines the reliability of the result.
In the next article of this series, we will examine the regulatory framework of bioequivalence and how the approaches of different countries come together.