Regulations are no barriers for molecular medicine
Regulatory authorities only grant a Marketing Authorisation to medicinal products which have demonstrated a positive benefit/risk ratio in a specific population suffering from a specific disease. Acceptable risks may be very high if the benefits are very significant. For example, in clinical studies intravenous infusion of Interleukin 2 (Aldesleuline® , Proleukine®) lead directly to death in 4% of patients with metastatic renal carcinoma but prolonged survival for 15% of these patients in a disease which is universally and rapidly fatal. In contrast, a number of medicinal products with good safety profiles did not receive a Marketing Authorisation because they did not demonstrate clinical benefit. If the benefit is nil, the benefit/risk ratio will always be unacceptable!
Following approvals of medicinal products, regulatory authorities regularly review the safety information collected post-approval to improve the benefit/risk ratio of each medicinal product. This can be done by issuing specific warnings (“Black Box”), restricting the target patient population, mandating monitoring of specific physiological functions (renal, cardiac, coagulation, etc.), changing the dosing regimen, etc. In addition, most marketing authorisations are now conditional on performing clinical studies aimed at better evaluating adverse events and developing mitigating strategies for the risks identified during the development of the medicinal product [Risk Management Plan (RMP) in Europe, Risk Evaluation and Mitigation Strategy (REMS) in the USA].Personalised medicines: an old and a new idea
Adapting medicines to the specific characteristics of a patient is not a new idea. The fact that whole classes of “similar” drugs (anti-hypertensives, antibiotics, etc.) exist to satisfy specific medical needs is testimony to this assertion. Since several decades antibiograms have been used to select the most effective antibiotics to treat the infection of a specific patient. Polyclonal anti-Rhesus D sera are used to prevent foeto-maternal allo-immunisation of Rhesus negative women having a Rhesus positive child which occurs in approximately 15% of pregnancies. As this medicinal product is deleterious for Rhesus positive women, its indication is limited to pregnancies of Rhesus negative women. Furthermore foetal blood testing is preformed to determine if the Rhesus negative mother is carrying a Rhesus positive foetus so being at risk of potential allo-immunisation. Prevention is performed only if this risk is established.
However, this old concept got a new life at the beginning of this millennium. As underlined in a reflection paper by the European Medicines Agency (EMA) “the last decade has seen rapid progress and development in the understanding of genetic influences that underlie inter-individual differences in drug action and drug response. This includes areas of pharmacogenomics that study genetic variability in relation to drug transport, drug product metabolism, enzymes responsible for metabolism, and drug response.” The aim of this research was to identify subsets of patients identifiable by a common marker or characteristic that may determine either increased efficacy or increased risk of adverse events, i.e., improved benefit/risk analysis.
Between 2000 and 2007 the EMA reviewed data on 26 new medicinal products for the treatment of cancers. Most of these datasets included attempts at identifying specific target populations using specific biomarkers. For six products, these data were sufficient to allow inclusion of important pharmacogenomic information in their labelling. The relation between the epidermal derived growth factors, their receptors and tumour response to treatment in cases of breast, lung or other cancers have been well recognised both by the scientific and regulatory communities. The discovery of targeted monoclonal antibodies to specific antigens has had a major impact on diagnostic and prognostic aspects of pharmacogenomics in oncology. Regulatory perspectives
Regulatory agencies always look for the most precise definition of the patients who may benefit of a specific therapy. Among others, they use the concept of number of patients treated for each patient who benefits of a drug. A decade ago, haemoglobin substitution therapies were in full development. One of them showed that when surgical patients required blood transfusion, 20% of them could be treated with a new medicinal product and would not need any administration of heterologous blood. The number to treat for this product was five, as for five patients who were treated one would really benefit by completely avoiding heterologous blood transfusion. However, four patients who had received the new medicinal product would still require some blood transfusion even if at a lower dose. This ratio was not considered as an acceptable benefit/risk and the medicinal product was not approved.
Clearly the recent advances in pharmacogenomics will help the pharmaceutical industry in better pre-identifying the subpopulation that will benefit from the new drugs in development. This will help getting the needed medicinal products to the populations of patients who might benefit from them by improving their benefit/risk ratio.
The biomarker used to select a specific population should meet a number of characteristics: 1/ be measurable and quantifiable with a certain degree of consistency, 2/ demonstrate a relation between the marker and the disease mechanism as well as with the pharmacological actions of a medicinal product; these are usually obtained from pre-clinical or early clinical studies (phase I-II), 3/ be validated in relation to the parameters discussed above in Phase II and Phase III clinical trials or by epidemiological methods.
From a regulatory perspective there is not much difference between a “normal” and a “personalised” medicinal product, except the need of a companion diagnostic assay to select the target patient population. This assay should be fully validated and approved. In Europe this is done in accordance with the in vitro diagnostic medical devices directive. Of course, this assay should be commercially available in the countries where the medicinal product will be marketed. Otherwise a “personalised” medicinal product is like all other medicinal products. Its positive benefit/risk ratio should be demonstrated in controlled clinical studies. However, appropriate selection of the target population for “personalised” medicines should lead to easier and higher benefit/risk estimates as patients who cannot benefit from the drug or who are at increased risk of adverse events would be excluded.