How Are Bioequivalence Studies Designed?
The value of a bioequivalence study lies in its design even before its result. To compare two products fairly, every factor that might hide the difference between them must be accounted for in advance. A well-designed study makes a real difference between two products visible if one exists, and allows us to say with confidence that there is none if there isn’t. In this article we look at how a bioequivalence study is built and why each decision is made.
Why a crossover design?
People differ from one another. In two people taking the same medicine, absorption, distribution, and elimination can vary considerably. If we give the test product to one group and the reference product to another and then compare the results, it becomes difficult to tell whether the difference that emerges comes from the products or from the natural variation between individuals.
To solve this problem, bioequivalence studies generally use a crossover design. In a crossover design each participant receives both the test product and the reference product, in separate periods. Each person thereby becomes their own control, and the risk that differences between individuals will blur the result is largely prevented.
The most widely used form is the two-sequence, two-period design. Participants are randomly divided into two groups. One group first receives the test product and then the reference product; the other group reverses the order. Determining the order at random prevents the sequence of periods from influencing the result.
The washout period
For a crossover design to work, a participant must be completely cleared of the first product before receiving the second. Otherwise the part of the first product remaining in the blood distorts the second measurement. For this reason a washout period is placed between the two periods.
The length of the washout period is determined by how quickly the active ingredient is eliminated from the body. For a substance with a long half-life, this period may extend over several weeks. Choosing this period too short is one of the most common design errors in such studies.
When a crossover design is not suitable
Not every study is suited to a crossover design. If the half-life of the active ingredient is very long, the required washout period makes the study impractical. In such cases a parallel design is used, in which each participant receives only one product. Because a parallel design cannot eliminate variation between individuals, it requires more participants and more careful statistics.
Other decisions also shape the design. Whether the study is conducted with a single dose or with repeated doses, and whether it is performed on an empty stomach or together with food, is determined by the properties of the active ingredient and the way the product is used. For active ingredients with high within-subject variability, replicate designs in which each product is given more than once to the same person may be preferred; this allows the variability to be measured directly.
Who takes part?
Bioequivalence studies are most often conducted with healthy volunteers. The aim is not to treat a disease but to compare the behavior of two products in the body; because variability related to illness is removed in a healthy group, the comparison is cleaner.
There are exceptions to this. Some classes of active ingredient cannot be given to healthy volunteers for safety reasons; in these cases the study is planned with patients who already use the medicine. The decision is always made with participant safety in mind.
How many participants are needed?
The right number of participants is determined by calculation, not by guesswork. The deciding factor here is the variability of the active ingredient: the more a person’s response to the same product varies at different times, the more participants are needed to capture a real difference with confidence.
This calculation is called sample size determination and is directly related to the statistical power of the study. A study conducted with too few participants may fail to distinguish two products that are in fact bioequivalent, or conversely may produce a misleading difference. For this reason the number of participants is calculated carefully before the study begins.
How is the result evaluated?
When the study is complete, the area under the curve and the maximum concentration values for both products are obtained for each participant. These values are compared not directly but after a logarithmic transformation is applied, because such quantities are more meaningfully interpreted in proportional terms.
At the heart of the evaluation lies the ratio of the geometric means of the two products. A ninety percent confidence interval is calculated for this ratio, and this interval is expected to fall entirely between eighty percent and one hundred twenty-five percent. Widely adopted by international regulatory authorities, this range defines the bounds of a difference between two products that can be regarded as clinically negligible. If the confidence interval stays within these limits, the products are considered bioequivalent.
A good design is completed by accurate measurement
A good design gains meaning only if the concentration of the active ingredient in the collected samples is measured accurately; the rigor brought by the design and the accuracy brought by the measurement complete each other. As a center authorized to conduct these studies, NanoTox provides, with validated LC-MS/MS methods, the measurement precision that a good design requires.
In the next article of this series, we will examine why the variability of study results matters and how active ingredients with high variability are handled.