Lack of pharmacokinetic interaction as an equivalence problem.

The demonstration that concomitant administration of drug B does not affect the pharmacokinetics of drug A can be adequately handled as an equivalence problem. Administration of drug A alone serves as reference and simultaneous administration of drugs A and B as test situation. The range of clinically acceptable variation in the pharmacokinetic characteristics of drug A defines the equivalence range. This will usually correspond to the bioequivalence range accepted for the comparison of different formulations of drug A. Equivalence, i.e. lack of pharmacokinetic interaction, is concluded if the 90%-confidence interval for the ratio (difference) of the expected medians for test and reference is entirely within the equivalence range. This decision procedure ensures that the consumer risk of incorrectly concluding "lack of interaction" is limited to 5%. Moreover, the producer risk of incorrectly concluding "interaction" can be controlled by appropriate sample sizes.