A total of 194 new molecular entities approved by the Food and Drug Administration between 1992 and 1997 were surveyed to determine the role of in vitro metabolic interactions in the conduct of drug-drug interaction studies and to examine the methods used in these studies. Approximately 30% of the submissions were found to have in vitro metabolism-based interaction studies, most of which were inhibitory in nature. Chemical inhibition was the most commonly used approach in studying drug interactions in vitro. In this article, an attempt to assess the quality of the chemical inhibition approach was made. Four areas were found to be often overlooked: (1) incubation time and concentrations of the drug, (2) the difference between inhibition constant (k(i)) and 50% inhibitory concentration (IC50) values, (3) the substrate-dependent inhibition potential, and (4) the metabolic genotype or phenotype of the liver donor. We discuss the pitfalls in estimating drug interactions when these four areas are overlooked.