The design, performance and evaluation of bioequivalence studies have received major attention from academia, the pharmaceutical industry and health authorities over the past decade. Despite the efforts of various regional and national bodies in setting up guidelines and recommendations (APV 1987; CPMP 1991; FDA 1977 FDA , 1988 Nordic Council of Medicines 1987;Skelly 1984), a universal standard is still lacking. Such a standard is becoming ever more necessary, in order to avoid costly and unethical replications of largely identical studies. Although the requests by the various health authorities may appear to differ only marginally, they may lead to different conclusions with regard to approval or rejection of bioequivalence_ Among the issues discussed during a recent workshop of the Drug Information Association (DIA) on 'Bioavailability/Bioequivalence: Pharmacokinetic and Statistical Considerations' (DIA 1991) were the logarithmic transformation of bioequivalence characteristics and the corresponding bioequivalence range. Other controversies concerning the design of bioequivalence studies are: selection of the study population (healthy subjects/ patients, smokers/nonsmokers, normal metabolisers/fast and poor metabolisers); posture (supine/ mobile); fasting or fed conditions and time between drug administration and meal intake; volume of fluid given after drug administration; the selection of the appropriate reference formulation; single vs multiple doses; and the replicated administration of test and reference formulations in the same study participant for 2 periods each. This last would allow investigation of the reproducibility of test and reference formulation, i.e. a comparison of the within-subject variability of test and reference. This has to be seen in connection with the most recent controversy in bioequivalence assessment, namely, whether average bioequivalence is a sufficient criterion to allow interchangeability of generic drugs, or whether the more stringent individual bioequivalence should become a necessary condition (Anderson & Hauck 1990).The option to choose between pharmacokinetic characteristics for rate and extent of absorption has attracted considerable attention from pharmacokineticists, particularly with regard to controlled released formulations (e.g. Steinijans 1990). During the recent DIA workshop it appeared, however, that speakers ofthe US Food and Drug Administration (FDA) still consider area under the plasma concentration-time curve (AUC) and peak concentration (Cmax) as the primary characteristics of extent and rate of absorption, although the limitations of Cmax as a rate characteristic were admitted, and despite other rate characteristics such as the peaktrough fluctuation having been recommended in earlier FDA guidance (Skelly 1984). If pharmacokinetic characteristics cannot be determined, e.g. due to lack of assay sensitivity, pharmacodynamic characteristics may be used; but possible implications for the design of the study (e.g. parallel groups vs crossover) should also b...