30 years ago an anthracycline antibiotic was shown to have antineoplastic activity. This led to the development of well over 1000 analogues with a vast spectrum of biochemical characteristics. Many biological actions have been described. The original anthracyclines are active against many types of cancer and are an integral part of several curative combinations. They are ineffective against other tumours. Although some analogues show an altered spectrum of activity or an improved therapeutic index relative to the older agents, it is not clear that cardiotoxicity can be totally avoided with these agents. Primary and secondary resistance to anthracyclines remain major clinical problems. Pharmacokinetic studies have been of limited help in explaining this. Overexpression of a surface-membrane permeability glycoprotein (Pgp) was identified in ovarian cancer of patients who had clinical multidrug resistance in 1985. This led the way for the discovery of a number of resistance mechanisms in vitro. Some of these have been found in more than 1 type of cell line, and more than 1 mechanism may exist in a single cell. Additional resistance proteins have been identified, qualitative and quantitative alterations of topoisomerase II have been described, and some mechanisms in other systems have not yet been identified. Some of these may prove to be important in clinical drug resistance. Drugs such as calcium antagonists and cyclosporin, studied initially for their ability to block the Pgp pump, appear to be heterogeneous in this capacity and may have additional sites of action. It will be critical for clinical studies to define the precise resistance mechanism(s) that must be reversed. To date this has been difficult, even in trials ostensibly dealing with the original Pgp. Liposomes can potentially alter toxicity and target drug delivery to specific sites. In addition, they may permit the use of lipophilic drugs that would otherwise be difficult to administer systemically. Resistant tumours may be sensitive to anthracyclines delivered by liposomes. To reduce cardiac toxicity, administering doxorubicin (adriamycin) by slow infusion through a central-venous line should be considered whenever feasible. Monitoring of cardiac ejection fraction and the use of endomyocardial biopsy will permit patients to be treated safely after they reach the dose threshold at which heart failure begins to be a potential risk. A number of structurally modified anthracyclines with the potential advantages of decreased cardiotoxicity and avoidance of multidrug resistance mechanisms are entering clinical trials. Meanwhile, the vast weight of clinical experience leaves doxorubicin as a well tolerated and effective choice for most potentially anthracycline-sensitive tumours.