Two‐stage designs versus scaled average designs in bioequivalence studies for highly variable drugs.

What is it about?

The development cost of a new drug is very high. Any drug goes through different development phases where thousands of patients are included in clinical trials before demonstrating safety and effectiveness (and only 1 out of 5,000 molecules reach the market). When a new molecule is investigated, it is patented for a period of 20 years. But a development period usually lasts more than 10 years before regulatory agencies allow its launch (the most common agencies are the EMA and FDA). When the commercialization of a new drug is authorized, and the company gives a brand name, negotiations to stablish pricing and reimbursement begins at country level, between sponsors and local governments. From this point and up to the patent expiration, sponsors have all rights and selling exclusivity. When the rights of a drugs expire, the market is spread out because the same or other sponsors choose to develop and commercialize generic drugs. The development cost of a generic drug is much lower because it is not necessary pass through all the development phases, and usually are based on bioequivalence studies which include some tens of healthy volunteers only. When generics enter the market, the supply extends and drug prices fall making easier access to consumers. Currently, generic drugs represent 70% of all the US medical prescriptions. Generic products must comply with bioequivalence criteria. They should contain identical amounts of the same active drug ingredient in the same dosage form and route of administration as the reference product, and the rate and extent of absorption must be similar. If bioequivalence is demonstrated, brand and generic drugs are interchangeable granting similar security and effectiveness. Usually, two drugs are claimed bioequivalent when its bioavailabilities are similar within a difference in means of ± 20% maximum. Sometimes, there are drugs whose rate and extent of absorption is highly variable dose to dose within the same patient. This kind of drugs are considered highly variable. In this case, following the classical bioequivalence approach, should be necessary to include many healthy volunteers to demonstrate bioequivalence, and because this is considered unacceptable, the EMA is more flexible and allows expanding the bioequivalence limits to a maximum difference in means of ± 30%. With this new criterion, these studies usually include only some tens of subjects, but more complex statistical designs are necessary requiring until two-fold laboratory extraction for each subject. Our publication explored the likelihood of declaring bioequivalence in highly variable drugs. As any other probabilistic study, it is subject to errors. The most common is the false positive rate (or consumer risk), i.e. the probability of declaring bioequivalence when it is not the case. Our article discussed and put on doubt the European rules where under some variability conditions this error is estimated at 8.5%. We proposed an adjustment which reduces this error below 5%. Secondly, we studied a methodology (also regulated) based on adaptive designs which uses simpler designs, and rationalize the number of healthy volunteers in two stages; in the first interim stage, only a few number of subjects are included (e.g. 12 subjects), and if it not possible to declare bioequivalence, but the results are promising, they allow to add some new subjects on a second stage depending on the variability observed in the interim, so finally increasing the likelihood of declaring bioequivalence. The advantage of this method is that the development time is usually reduced. In this case, we proposed a new approach to control false positive rate. This study was conducted through simulations based on assumptions about the drug variability and the number of subjects included in the first stage. Once we ensured that the false positive rate was controlled below 5%, we estimated the statistical power, running up to 1.000.000 simulations. We demonstrated that both methods show similar power, but the classical method requires less sample size, at the expense of exposing subjects twice as long as the adaptive method.

Featured Image

Why is it important?

To claim bioequivalence for generic drugs being highly variable, we show that both scaled and adaptive methodologies are useful, though the classical scaled methods requires less sample size, at the expense of exposing subjects twice as long as the adaptive method. The methods are reliable because we always ensured that the false positive rate (or consumer risk) is below 5%.

Perspectives