Kathy Archibald and colleagues' Correspondence prompts me to write. Suspected adverse drug reactions are unlikely to have recently become more frequent or severe, when the novelty and range of new medicines, biologics, and mechanisms of action are considered. Spontaneous reporting systems after a drug has reached the market suffer notoriously from underreporting (often only 2-5% of actual adverse drug reactions are reported), and the rate of under-reporting has remained relatively unchanged over many years. The limitations of animal studies are well known, and their main use for small molecules is early prediction of harms, such as teratogenicity and mutagenicity, as well as allowing some prediction of human pharma-cokinetics. The shortcomings of non-clinical studies of biologics is well documented in the Duff report, which introduced the concept of minimally acceptable levels of biological effect, in acknowledgment of the fact that the dose response from animal toxicokinetic studies has limited use in the prediction of first dose in man. There is consensus that greater use of in-vitro and clinical predictive models have potential, but these, like any new medicine, require careful testing and validation. A framework is required for more accurate and realistic measures of real-world safety and effectiveness of new drugs, which will allow an accurate assessment of their value, price, and safety.
展开▼
