The International Conference on Harmonization has put forth a tripartite guideline addressing mechanisms by which regulatory agencies might be able to accept foreign clinical data. A major issue is the effect of ethnicity on the drug's pharmacokinetics, pharmacodynamics and/or safety. The purpose of this review was to determine whether the effects of ethnicity on pharmacokinetics are predictable. We also evaluated whether the premise that only pharmacokinetic processes which are biologically or biochemically mediated are likely to exhibit ethnic differences was supported by the literature. Bioavailability is determined by absorption, gut metabolism/transport and hepatic first pass metabolism. Absorption is usually a passive process and, as would be expected, no examples of ethnic differences in passive absorption were found in the literature. Direct evaluation of ethnic differences in gut metabolism/transport or hepatic first pass metabolism is largely lacking, although some studies suggest such differences exist. These differences would also be expected based on known ethnic differences in hepatic clearance of drugs. Ethnic differences in plasma protein-binding to the two major drug-binding proteins, alpha1-acid glycoprotein (AGP) and albumin, have been studied in several populations. Based on these studies, ethnic differences in plasma protein-binding for drugs which bind exclusively to albumin (e.g. acids) are uncommon. Conversely, ethnic differences in plasma protein-binding of drugs to AGP appear to be very common, with the studies consistently showing Caucasians have higher binding (lower plasma free fractions) than other ethnic groups. This difference appears, in all cases, to be explained by racial differences in plasma AGP concentration. Hepatic metabolism is the most common pharmacokinetic parameter for which there are ethnic differences. Differences have been documented in both oxidative and conjugative metabolism. Ethnic differences in hepatic metabolism are unpredictable by race (e.g. one racial group is not consistently higher or lower than another group) and specific enzyme (e.g. studies with different CYP3A4 substrates have yielded different results). Ethnic differences in renal tubular secretion have been documented, but also appear to be unpredictable, while differences in the passive processes of renal elimination, filtration and reabsorption, have not been observed. The literature supports the premise that pharmacokinetic processes which are active (e.g. involve a protein) are the ones with the potential for differences between ethnic groups while passive pharmacokinetic processes do not exhibit such differences. Based on the available literature, the drugs most likely to exhibit ethnic differences in their pharmacokinetics are those that undergo significant gut metabolism/transport and/or hepatic first pass metabolism; are highly bound to plasma proteins (especially AGP); or have hepatic metabolism as a major route of elimination.