Lee P. Topping and Dr. Nicholas T. Longford
De Montfort University, Leicester, England
Two formulations of a drug are declared bioequivalent if their bioavailabilities are similar. Under bioequivalence the two formulations are considered interchangeable because their effects with respect to efficacy and safety differ insubstantially. Research into the statistical methodology of bioequivalence started in the early 1970s [1] and has gradually gathered momentum. This has culminated in the current US FDA guidelines (the 80-125 Rule) issued in 1992 [2]. However this rule has been criticised on several counts. The “one size fits all” implication of the rule is unsatisfactory, particularly for inherently highly variable drugs. Also, showing that on average a new formulation is similar to an old formulation is inadequate as this gives no assurance that the new is not more variable than the old. Consideration of bioequivalence in terms of both average effects and variability of formulations is referred to as population bioequivalence. This still leaves the possibility of major differences between the formulations for many subjects. This leads to a new concept, individual bioequivalence, which was first considered in 1990 [3]. Several definitions of individual bioequivalence which attempt to adapt the criterion of average bioequivalence have appeared since. In some of the proposed definitions, relatively large differences in means can be compensated by differences in measurement error variances [4]. We propose a definition based on a latent variable model, which overcomes this anomaly and does not involve the 80-125 Rule. The approach is illustrated on data from a bioequivalence study. The issue of uncertainty about estimates of treatment and subject variances is also addressed.
References
[1] Westlake W.J. (1972) Use of Confidence Intervals in Analysis of Comparative Bioavailability Trials. Journal of Pharmaceutical Sciences. 61(8) 1340-1341.
[2] FDA. (1992) Guidance on Statistical Procedures for Bioequivalence Studies Using a Standard Two-Treatment Crossover Design. Division of Bioequivalence, Office of Generic Drugs, Center for Drug Evaluation Research, Food and Drug Administration, Rockville, Maryland.
[3] Anderson S, Hauck W.W. (1990) Consideration of Individual Bioequivalence. Journal of Pharmacokinetics and Biopharmaceutics. 18 259-274.
[4] Hauck W.W. et al. (1996) Mean Difference vs. Variability Reduction: Tradeoffs in Aggregate Measures for Individual Bioequivalence. International Journal of Clinical Pharmacology and Therapeutics. 34(12) 535-541.
Reference: PAGE 6 (1997) Abstr 668 [www.page-meeting.org/?abstract=668]
Poster: poster