In vitro systems are proving of increasing value in predictions of the metabolism of foreign compounds. A variety of such systems, including precision-cut liver slices, isolated hepatocytes, immortal cell lines, subcellular fractions and recombinant expressed enzymes, are now in regular use. When derived from human sources, such systems are often preferable to studies in animals. The assembly of panels of reagents (e.g. antibodies, substrates, inhibitors) specific for individual enzymes of drug metabolism is an important and essential development. Using such reagents, each system needs to be characterised according to an agreed protocol, with respect to the specificity and activity of the major enzymes of drug metabolism. Once characterised, it is important that systems are used within their limitations. Difficulties in maintaining the activity of drug metabolising enzymes long-term in hepatocytes mean that further improvements in such systems will be necessary before they can be used reliably for certain types of drug metabolism study. There is no doubt that in vitro systems are of considerable value in predictive drug metabolism and that they will be refined further. Such systems will play an essential role in product selection, safety evaluation and development, in many cases replacing animal studies.