Evaluation of the carcinogenic potential of pharmaceuticals. Opportunities arising from the International Conference on Harmonisation.

The evaluation of the carcinogenic potential of pharmaceuticals is currently undergoing dramatic changes. For the past 25 years the regulatory expectation for agents intended for long term use has been that lifespan studies (usually lasting 2 years) in 2 rodent species be conducted. These studies take at least 3 years to plan, execute and interpret, and use over 1200 animals. It is now recognised that the quality of the information obtained from these studies is unreliable for prediction of carcinogenic risk to humans. Over the past 4 years, the International Conference on Harmonisation (ICH) has recommended changes in approaches to assessing the carcinogenic potential of pharmaceuticals. In future, only one long term rodent study will be routinely required (usually in rats), provided this is complemented with a short or medium term test in one of the emerging new models for carcinogenicity, such as transgenic mice or newborn mice. However, the relevance of these new models to human cancer and their use in risk assessment is still largely unknown and this situation must be kept under review as knowledge accumulates. A long term study in a second rodent species is still an option. Dose selection has also been improved inasmuch as there are now several alternatives to the use of the maximum tolerated dose (MTD). In the past, the use of the MTD, when the normal homeostasis of the test animals is disturbed, has been considered one of the major problems with the rodent carcinogenicity bioassay. However, one of the alternative end-points to the use of the MTD, i.e. the comparison of plasma concentrations in rodents and humans, must be viewed with caution. While this may contribute to limiting the high dose level for agents of very low toxicity, the concept should not be interpreted as signifying that plasma concentrations provide a sound basis for comparing the carcinogenic activity of agents in different species. Recognition of the 4 properties (genotoxicity, immunosuppression, steroid hormonal activity and long term tissue damage), at least one of which is associated with each of the pharmaceuticals known to be carcinogenic to humans, should focus more attention on a search for these properties in patients. Absence of these properties at clinically relevant dose levels indicates that a pharmaceutical is highly unlikely to be carcinogenic to humans.