Medicines, Genomics and Regulation

Margaret Cone reports on a workshop on pharmacogenomics and pharmacogenetics in drug development.

The pharmacogenomic era has begun in earnest and it is no longer a question of whether it will deliver a new generation of medicines, but when¹.

Genetic drugs research is simply putting profits into the hands of pharmaceutical firms and public health education could save more lives for less cost².

Such differing perspectives serve to illustrate the divergent views that can be expected as progress is made on the integration of pharmacogenomics and pharmacogenetics (see box) into the discovery and development of new medicines and the potential for such medicines to be `personalised' or tailored for optimal efficacy and minimal risk in the individual patient. The first of these viewpoints is from a workshop on Regulating Personalised Medicines that was convened by the CMR International Institute for Regulatory Science in April 2003, the second was a view expressed to a BBC radio programme by the eminent geneticist, Professor Sydney Brenner, in September 2003.

Whilst pharmacogenomics may initially be used to increase knowledge of disease aetiology, the `ultimate' personalised medicine is one that is marketed in combination with a genetic diagnostic test, or with labelling that requires such testing to be undertaken. The many regulatory issues that this new generation of medicines raises were the primary focus of the CMR International Institute workshop but the speakers from regulatory agencies and industry that took the platform, raised many broader considerations including:

• For the purpose of the workshop the following definitions were used:

Pharmacogenetics (PGt):

the study of interindividual variations in DNA sequence related to drug response.

Pharmacogenomics (PGx):

the study of the variability of the expression of individual genes relevant to disease susceptibility as well as drug response at cellular, tissue, individual or population level. The term is broadly applicable to drug design, discovery, and clinical development.

CPMP Position Paper on Terminology in Pharmacogenetics, 21 November 2002, EMEA/CPMP/3070/01.

the economics of developing medicines for targeted subsets of the patient population rather than the traditional `one product fits all' approach;

• the implications for the `payers' who will need to be convinced of the cost benefits of medicines that will almost inevitably be higher priced, and will carry the cost of additional diagnostic testing;
• the practical as well as economic issues for healthcare delivery infrastructures in moving towards an era where diagnostic genetic testing, as a prerequisite to prescribing medicines, could become the norm, rather than the exception; and
• the greater involvement of patients in decisions about their treatment.

A view that was re-iterated throughout the workshop was that, whilst pharmacogenomic research must be conducted within an appropriate regulatory framework, premature and inappropriate additional regulation could hamper development of the science. Dialogue between industry and regulators and a willingness to share experience of the issues were identified as critical factors in moving the science forward.

Integrating genetics into R&D

Pharmacogenomic and pharmacogenetic techniques are opening up fascinating new possibilities in drug discovery and development, with the ability to generate information for target identification and target validation and to explain variations in responses to drugs, adverse events and pharmacokinetics. There is now the realistic prospect that patient populations can be prospectively characterised according to likelihood of response and propensity to adverse reactions.

Margaret Cone is the Director of Regulatory Science for the CMR International Institute for Regulatory Science. This article quotes extensively (with permission) from the report of the Institute Workshop on Regulating Personalised Medicine but the commentary reflects the personal views of the author.

Although genotyping of the `target' for new drugs is now carried out routinely by some companies as part of drug discovery and development this does not mean that truly `personalised' medicines are the primary aim. Such medicines might have unrealistically small patient populations. More generally the aim is to ensure that the new candidate drug acts on the most frequent allelic variant of the target, and that this variant represents the majority of patients suffering from the disease.

Exceptions to this would be cases where there are sufficiently compelling efficacy drivers to justify the development of a medicine for a relatively limited patient population. An example would be serious progressive and/or life-threatening diseases where a population subgroup can be identified prospectively for whom the product is a clearly superior treatment. This is illustrated by the interest in using pharmacogenomic techniques to understand the aetiology of different types of cancer and to develop new medicines in the field of oncology. Other examples that would justify the deliberate search for medicines for a limited population can be found where there are serious adverse reactions that would normally result in the termination of development or marketing but the ability to identify susceptible individuals prospectively makes the product viable.

Such strategies must, however, be put into perspective since the truly targeted medicine involves the need for diagnostic testing, with all its accompanying drawbacks. The development of such medicines would obviously not be worth pursuing if there were equally useful products on the market without the need for a diagnostic test.

In this context, participants at the meeting were also reminded of some medicines that, despite variability of action and adverse effects, have become widely used and well established. Subsequent investigation has pointed to genetically related differences in enzyme expression. Examples include: haemorrhage and warfarin (allele - CYP2C9); toxicity/lack of effect and antidepressants (allele - CYP2D6); lack of analgesic effect and codeine (allele - CYP2D6); and prolonged sedation with diazepam (allele - CYP2C19).

The integration of pharmacogenetics into pharmaceutical R&D should therefore be regarded as `evolution not revolution'.

Regulation

Both industry and regulatory participants at the CMR workshop were concerned to avoid premature regulation of pharmacogenomics in drug development. Too much regulatory oversight could hinder the development of the new sciences, too little could increase the risks inherent in the technology and this could also hamper development. These concerns were less about the lengthy process of changing actual regulation and more about avoiding hastily produced `reactive' guidelines and guidance documents in response to the perceived challenges of the new science. This is not only an issue between regulators and industry, but could also result from the action by advocacy groups which is already having an impact in this field, on such issues as:

• patent law and controversies over patenting in relation to genetic material;
• privacy regulations and confidentiality of patient information; and
• concerns about general genetic screening and the wider implications of misuse of information on disease susceptibility, e.g. by employers and insurance companies.

Not unexpectedly, and with echoes of the early days of biotechnologically developed medicines, the workshop advocated a `case-by-case' approach to regulation as being the only feasible approach for a science in its infancy. There was, however, a potential inconsistency between industry's desire for guidance from regulators, through the scientific advice procedure and through more informal channels, and resistance to the development of specific guidance. Regulators were also concerned about being faced with an increasing number of applications that are heavily dependent on pharmacogenomic data, in the absence of agreed guidance on their evaluation and assessment.

Of particular concern is the fact that ethics committees, faced with clinical trial protocols that include genetic testing of subjects, will be asked to make decisions in the absence of general and specific guidance and may be delivering advice that is not necessarily consistent from one case to the next.

Although many aspects of pharmacogenomics are new, variability in drug response and the expectation of underlying genetic factors are not. Dr Marisa Papaluca Amati of the EMEA, pointed out to the workshop that the need to take account of such variability is already incorporated into existing CPMP guidance documents. These include guidelines on: pharmacokinetic studies in man; bioavailability and bioequivalence; and drug interactions. The concept is also integrated into the ICH guidelines on dose-response information to support drug registration (ICH E4 ) and ethnic factors in the acceptability of foreign clinical data (ICH E5 ). Dr Papaluca Amati noted that one of the tasks of the CPMP working party would be to revisit the relevant guidance documents and decide on the need for updates and amendments.

Specific regulatory issues

One of the most pressing issues is the need for co-ordination between the regulation of new medicines that depend upon a diagnostic test for compliance with the labelled conditions for use, and the procedures for registering the genetic diagnostic tests themselves. In many regulatory systems, notably those of the US and many European countries, medicinal products and in vitro diagnostics (IVDs) not only fall under separate legislation, but are regulated through separate administrative bodies.

The workshop felt that there was a significant `gap' in communication not only between regulatory authorities but also between companies involved with pharmaceuticals and those developing IVDs. There appeared to be considerable and mutual ignorance about the regulatory procedures for the different sectors. International harmonisation was cited as an example where opportunities for closer interaction have been missed. The Global Harmonization Task Force (GHTF), for example, is working on initiatives to prepare harmonised guidance and regulation of medical devices and IVDs, but the discussions do not involve those concerned with the pharmaceutical products that might be affected.

Requirements for IVDs

The application of pharmacogenetics during the clinical development of a new medicine does not automatically mean that a test will need to be developed for marketing with the product. There are many valid applications of pharmacogenetics, where the science gives insight into the variability in drug response, but does not signal the ultimate need for a test. There are therefore different requirements for testing and validating the performance of IVDs. For those used during drug development it is necessary to ensure that the tests are relevant in terms of the genetic polymorphism that is tagged. In this context, the validation of so-called `home brew' tests was raised as a potential cause for concern.

The need to ensure validity and consistency of IVDs that are destined for the commercial market, in conjunction with specific medicinal products, is better defined. One of the presentations at the meeting was made by Dr David Feigal, Center for Devices and Radiological Health (CDRH), FDA. Admitting that the US regulatory system for established devices and diagnostics relies on somewhat `arcane' rules based on precedent, Dr Feigal made it clear that the new generation of IVDs for genetic testing would need to meet rigorous standards and levels of evidence appropriate to the novelty of the claims made for the product. The statutes specify that `valid scientific evidence' is required to support claims and reflects the fact that the majority of `traditional' devices and diagnostics do not require human testing to determine safety and effectiveness or compliance with performance standards. There will, however, be many examples of genetic tests that do require human trials and studies in clinical settings. Decisions on the safety of a diagnostic reflect a risk and benefit analysis where the primary risk is the consequence of a wrong diagnosis leading to a wrong decision.

With the attendance of Dr David Jefferys, head of the devices sector of the MHRA, at the workshop, participants (who were predominantly from pharmaceutical companies) were also updated on the implementation of the European IVD Directive , which comes into force on 7 December 2003³. This will require not only that tests are validated for sensitivity and specificity but also that the laboratories and services undertaking the tests meet the required standards. The requirements for registration and CE marking for laboratories and services applies equally to commercial enterprises and to those operating within the health services.

As regulation moves forward in Europe, genetic tests may need to be classified under Annex 2 to the IVD Directive which means that common technical standards will need to be defined. This in turn, highlights the need for international harmonisation and reference was again made to the role of the GHTF.

Samples

Several speakers at the workshop identified concerns about the handling, coding and storage of genetic samples collected during clinical trials. Ethics committees insist upon measures to safeguard confidentiality of collected samples but there is debate as to whether stored samples should be anonymised or coded in a way that allows subsequent identification, bearing in mind requirements for auditing.

Issues were also raised relating to informed consent and the ability to carry out future analyses on stored samples. There was discussion on whether this is covered implicitly in patient consent and, if not, how to obtain explicit consent to all pharmacogenetic tests that the company may wish to perform at a later stage.

The need for education and change

The public perception of all matters connected with `genes' and `genetics' has doubtless been coloured by sensationalised press coverage of genetically modified foods, cloning and `designer babies'. Whilst the publicity surrounding the unravelling of the human genome in June 2000 was generally positive and seen as holding out the hope of dramatic medical progress in the twenty-first century, there is little doubt that the science surrounding genetics is widely regarded with deep suspicion by the lay public.

There was consensus at the workshop that education is the key to `demystifying' the nature of genetic science in general and pharmacogenomics in particular. The public is well aware, through public debate, of the potential hazards of the misuse of genetic information. They are aware that genetic data can provide information about an individual's susceptibility to disease and there are concerns about such information becoming known to employers or insurance companies. The need to redress the balance through positive, factual information and balanced arguments was clear although the views expressed that `politicians and policy makers, regulators, industry and scientists all have a role' may ring of wishful thinking.

New terminology needed?

One solution proposed was the need for a new, non-threatening vocabulary and simplified terminology to convey information to the public and politicians about a new generation of medicines `tailored' to the needs of individuals rather than the general population. Not, perhaps, as far-fetched as it may seem was the proposal for a simplified phraseology that would explain the role of the new diagnostic tests in terms of:

• health profiles to determine the best type of therapy for the individual; and
• treatment guidance to ensure the best effectiveness and safety in the use of a particular medicine in the individual.

Whilst explaining the nature of these tests openly, they should not be portrayed as significantly different from other diagnostic tests on blood and urine samples that are routinely used as guides in diagnosis and the prescription of medicines.

Educating the professionals

The need for education and a change of attitude is not confined to the public. A significant move towards the use of medicines that require patients to be selected on the basis of routine genetic testing would require fundamental changes in medical practice. Physicians would need to be trained in the use of such tests and the implications of not carrying out testing, as well as in the interpretation of results. `Personalised' medicines, by definition require greater involvement of the individual and his or her personal choice and perception of risk. Such counselling will also be a matter of education and may raise questions of the extent to which, for example, the pharmacist should be involved in the new procedures.

Whilst there is a tendency to assume that physicians are, in general, resistant to change, and would not want to invest the necessary time and effort to educate themselves and apply the necessary tests, a report given by Ann Raven, of a Cambridge University project funded by the Wellcome Trust Bioethics Programme appeared to contradict this⁴. This found little resistance from physicians who deal with patients with life-threatening or crippling diseases where therapies under development have the potential to halt the disease process.

Generation gap

Throughout the discussions, as with many other debates on the introduction of new technology, there was a suggestion that resistance to, and coming to terms with, change may only be a problem for the `in-between' generation. This applies equally to the public and professions. Whilst the concept of genetic testing might raise alarm and concern today, the next generation might accept as normal, that a DNA swab should be taken from a newborn baby and stored in a database to provide a reference source throughout its life.

Ethical dilemmas

The ethical issues raised by the investigation and recording of genetic profiles and information were an integral part of the workshop discussions. Participants were cautioned against allowing the ethical dilemmas to grow out of proportion since they are not in fact new. Sensitive health-related and financial information on individuals is already handled within existing privacy and human rights laws as well as ethical codes and confidentiality agreements. Handling genetic information about individuals does not raise significant new issues that cannot be accommodated within existing safeguards.

There was discussion of whether the selection of patients for treatment on the basis of their genotype could give rise to accusations of discrimination against minorities, or possibly even racial discrimination. It was, however, pointed out that currently, it is common for medicines to be tested in one ethnic population without bridging studies being undertaken to determine their suitability for the treatment of other populations. Furthermore there are many examples where patients are excluded from particular therapies on the grounds of age and pre-existing medical conditions, although it was acknowledged that perceived `discrimination' because of genotype is likely to be a more sensitive issue.

Another scenario raised was the situation where future research reveals a matter which is of significant medical importance to patients from earlier clinical trials who have specific genetic characteristics that can be identified from stored samples (which have not been anonymised). The question was whether the company has an ethical obligation to follow up such patients. It could be argued that this confused clinical research with clinical care and that such retrospective follow-up was beyond the boundaries set by participation in clinical research. There was also a view that such boundaries were incompatible with the ethical standards set for research.

It was apparent that there are many ethical issues yet to be resolved and no doubt many more yet to be revealed. One of the recommendations from the workshop was that, whilst specific regulatory guidance for the development of medicines may be premature, there is a need to develop a `regulatory framework' for establishing good practices in the field of genetic testing. It was suggested that a code of `Good Gene Practice' might be appropriate.

Cost and returns: the economic issues

Whatever the medical advantages, any move towards the development of medicines that treat a smaller rather than larger section of the patient population is going to be met with concern by the business sector. It would appear that the development of medicines for smaller markets is only economically viable if the new technologies result in lower R&D costs, a shorter development time or, preferably, both. The presentations at the workshop, however, gave little prospect that, in the short term, this would be achieved, leaving the spectre of the `worst case' scenario where development costs and times remain as at present but the market is considerably reduced.

It was pointed out that some so-called blockbuster drugs are only used in a relatively small proportion of the totality of patients that suffer from a particular disease and it could therefore be argued that, where the disease is widespread, a targeted medicine could have the same economic profile as a conventional blockbuster. The more pragmatic view was that, in the majority of cases, the economic equation would only be balanced if higher prices could be negotiated on the grounds of improved efficacy and safety leading to greater cost efficiency.

The various economic scenarios were put to the workshop by Mr Adrian Towse, UK Office of Health Economics (OHE). He cited the extreme example of gene therapy which, if successful, would mean that a once-only treatment provided a life-time of health gain. In the virtual monopoly situation of the UK, the National Health Service (NHS) would carry the cost of that patient's medication and care but it would be almost impossible to persuade the NHS to pay a price for a one-off `cure' that would recoup the value of a medicine that would otherwise need to be taken for a life-time. There is a resistance to absolute price levels for new drugs and a reluctance to link price to the potential value. This resistance would be even stronger in a competitive health insurance market where the payer that carries the initial cost of treatment may subsequently lose the patient to a cheaper insurance scheme.

Mr Towse suggested, however, that the economics of medicines requiring genetic testing are not quite so bleak. He cited examples where the cost of combining a diagnostic test with treatment could be justified in terms of the savings through targeted use of an expensive but effective treatment.

The conclusion was that, with appropriate pricing and appropriate targeting of health effects, personalised medicine is quite compatible with the concept of a `blockbuster' but this is crucially dependent on whether payers are willing to recognise the value of the health gain and accept prices adjusted accordingly. This does, however, leave a problem when a group of patients is identified for whom existing therapies or new therapies will not work and patient numbers are too small, even at very high prices, to justify the investment required to develop a new medicine.

CMR Institute survey

It is normal practice for the CMR Institute for Regulatory Science to carry out a survey and study as part of the preparations for its workshops. For this workshop, the study, reported by project leader, Carly Anderson, collected data and opinions from international pharmaceutical companies and regulatory authorities on the integration of pharmacogenomic and pharmacogenetic techniques into the drug development process.

The authorities' perception of the benefits of using the new technologies included; increased safety and efficacy, better response rates in defined sub-populations and a better understanding of therapies, leading to improved decision making and disease management. The industry saw benefits in being able to produce more competitive, differentiated products with reduced side effects and size and duration of clinical trials as well as improved benefit risk ratios and enhanced cost effectiveness.

Key issues of concern shared by both regulators and industry included the lack of experience and expertise on both sides and the need to provide reliable diagnostic tests that are validated, sensitive, specific and stable. There were also concerns about the interpretation of data, how to handle findings of uncertain significance and the implications for labelling.

Looking at the future impact of pharmacogenomics, the majority of authorities predicted that increased human and financial resources would be needed, with increased requirements for scientific advice. It was envisaged that there would be improved benefit/risk decision making. From an industry perspective some companies (but not the majority) predicted that financial cost would be reduced over the next ten years and that development times would shorten. There was, however, a majority view that the new technologies would lead to reduced attrition rates in Phases 2 and 3 and improved benefit/risk decisions.

Not unsurprisingly, the survey results supported the view of the workshop that there was very little support for definitive regulatory guidelines at this stage although the majority of both companies and authorities would welcome general guidance. The preferred approach, identified by both parties, was more for interaction amongst authorities and between industry and authorities.

In conclusion

Whether `personalised' is the right term for the new generation of medicines or whether it is better to speak in terms of `targeted' or `tailor-made' medicines there is no doubt that, from the perspective of the individual, the outcome of using a medicine is very much a matter of personal concern. Traditionally clinical trials have been designed to demonstrate the ratio of efficacy and adverse reactions in relation to the total patient population but the individual patient perceives the outcome not as a ratio but as a single, personal response: whether the medicine works, is ineffective or causes an adverse reaction⁵. The integration of pharmacogenomics into drug development may bring about significant changes in clinical study design with a shift in emphasis from the majority to the individual.

Parallel changes are already being seen in the shift in emphasis from `population health' to personalised health with the public actively seeking information, often from the Internet, on health issues, diseases and medicines. It is unusual, these days, to attend any function related to health issues without hearing of the `empowerment of patients'.

Does the move towards `personalised' medicine, represent revolution or, as suggested, merely `evolution'? Many of the discussions at the workshop reflected the different perceptions of pharmacogenomics that were encountered in the Cambridge University study⁶. There were those who saw this as a unique and clearly distinctive new branch of science that could transform the practice of medicine, but which also brought new and complex regulatory and ethical issues that need to be resolved. To others it was perceived as an important development, but in reality only another prescribing tool to be integrated into clinical pharmacology, health professionals' education, post-marketing surveillance and other elements of the existing infrastructure.

References

1 Workshop on `Regulating Personalised Medicines', CMR International Institute for Regulatory Science, Nutfield Priory, Surrey, UK, 14-15 April 2003, www.cmr.org/institute

2. Remarks made by Sydney Brenner, Adjunct Professor of Biology, The Salk Institute, to BBC Today Programme, reported 2 September 2003, BBC Radio 4 website www.bbc.co.uk/radio4/today

3. Directive 98/79/EC of the European Parliament and of the Council of 27 October 1998 on in vitro diagnostic medical devices, OJ , 1998, L331 , 1-37

4. My very own medicine: What must I know? Information Policy for Pharmacogenetic, published by the Department of Public Health and Primary Care, University of Cambridge, 2003, www.phpc.cam.ac.uk/epg/IPP.html

5. Workshop presentation `The impact of pharmacogenetics and pharmacogenomics on clinical trials' by Dr Duncan McHale, Pfizer, Inc.

6. See reference 4