Overblog Suivre ce blog
Editer l'article Administration Créer mon blog
4 septembre 2011 7 04 /09 /septembre /2011 20:50

Nature Reviews Drug Discovery 10, 641-642 (September 2011) | doi:10.1038/nrd3534

A call to reform the taxonomy of human disease
Ismail Kola1 & John Bell2

top of page
A coordinated effort to incorporate advances in the understanding of the molecular and genomic variations in common diseases, such as hypertension, into their diagnosis and treatment could transform drug development and medicine.

Many common human diseases are still diagnosed as if they were homogenous entities, using criteria that have hardly changed for more than a century. For example, a person with a systolic blood pressure of 140 mm Hg or greater and a diastolic blood pressure of 90 mm Hg or greater is diagnosed with hypertension, irrespective of the heterogeneous underlying molecular mechanisms in different individuals. Furthermore, the treatment approach for diseases that are diagnosed in this way is generic, with empiricism as its cornerstone. Continuing with the example of hypertension, the standard initial treatment is dietary changes and exercise, and if these do not lower blood pressure sufficiently, pharmacotherapy will usually be initiated with thiazide diuretics. Thereafter, a beta blocker will be tried, followed by angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, and then calcium channel blockers.

Similarly, disease heterogeneity is not accounted for in the design and conduct of clinical trials in drug development; in general, patients are enrolled in trials for common diseases on the basis of their classical diagnosis, and knowledge that might be used to select patients on the basis of molecular characteristics or genetic aetiology is not available or not applied. This is even largely the case in oncology, an area in which major expansions in the knowledge of the molecular basis of cancer during the past two decades have revealed that it involves the aberrant expression and/or function of various genes, resulting in the deregulation of several pathways or networks that initiate and promote tumours. Indeed, common pathways may be deregulated in tumours that have different anatomical locations, and the identity of these pathways may be more important to disease progression and treatment than the location of a tumour. Although a few trials in oncology have enrolled patients on the basis of genetic and/or molecular characteristics — such as the presence of the BRAFV600E mutation in trials of BRAF kinase inhibitors in patients with melanoma — such trials are still a small proportion of the total. More ambitiously, a pivotal trial in which patients are enrolled based on the presence of a shared mutation and/or a deregulated pathway, rather than on tumour location, has not yet been initiated to our knowledge, but is an approach that regulatory agencies may be comfortable with in the future.

The lack of recognition of disease heterogeneity in clinical development and medical practice has a number of well-known consequences. First, it will probably reduce the likelihood of success of clinical trials, perhaps more so for targeted therapies that have been pursued in recent years. Indeed, if the pathway that is being targeted is not responsible for disease in an unknown proportion of patients enrolled in the trial of such a therapy, potentially effective drugs may be mistakenly abandoned. In this respect, it is noteworthy that therapeutic areas that are characterized by poorly understood disease heterogeneity, such as central nervous system disorders and cancer, have historically had the lowest success rates, which may even be decreasing further1, 2. Furthermore, one approach that has been used to counteract disease heterogeneity is to increase the number of patients enrolled in clinical trials to improve the 'signal to noise' ratio. So, overall the influence of disease heterogeneity on both the failure rates and the size of clinical trials is a key component in the spiralling costs of drug development.

Second, the clustering of patients with multiple disease subtypes into a single group for empirical treatment is reflected in the standard 'blockbuster' model for pharmaceutical products that became established in the 1990s. In this model, the role of marketing in influencing prescribers and patients played a major part in the success of products, some of which might only have substantially benefited an undefined subgroup of the many patients who received them. It is now increasingly recognized that clearly demonstrating the medical value of new products — for example, through more precise definition of the subgroups that benefit substantially — rather than marketing power, will be crucial in gaining reimbursement from health-care payers and thereby achieving a return on investment in pharmaceutical research and development (R&D).

There is therefore a compelling case for reforming the taxonomy of human disease by moving away from traditional diagnostic criteria alone to ones that incorporate the scientific advances in molecular and genetic medicine over the past two decades with traditional diagnostic criteria. The key question is how can this be achieved?

From a technical perspective, many of the tools needed to reform taxonomy are becoming increasingly powerful, based on advances in molecular sciences: the array of 'omics' technologies, imaging technologies (especially for central nervous system disorders and cardiovascular disease), and pathway and network analyses. Access to tissues (dissected with techniques such as laser capture microscopy), the availability of patients' electronic medical records in many countries and informatics techniques are also facilitating the classification of more precisely phenotyped heterogeneous disease populations into specific disease subtypes. For example, despite the genetic complexity of disorders such as type 2 diabetes mellitus, tools such as high-density genome-wide association studies with very large sample sizes are providing increasingly in-depth information about genetic risk factors that could provide the basis for identifying disease subtypes and conducting pharmacogenetic clinical trials3.

However, from a logistical perspective, reforming disease taxonomy is clearly a massive challenge. The extent of this challenge is illustrated by progress in the efforts to update the diagnostic system in one therapeutic area: psychiatry. The Diagnostic and Statistical Manual of Mental Disorders (DSM) is a diagnostic system developed by the American Psychiatric Association that lists the currently recognized mental disorders and the criteria for diagnosing them. The diagnoses within the current version, DSM-IV, are based solely on clinical observation. There had been hope that DSM-V, which is due to be published in 2013 after more than a decade of planning and discussion, could improve on earlier versions by integrating new information from genetics, cognitive neuroscience and brain imaging, and thus contribute to a reconsideration of the boundaries and subtypes of disorders along the lines proposed in this article. However, on the basis of the draft version (available at http://www.dsm5.org/Pages/Default.aspx) it seems that we will need to wait for further updates before such an integration is achieved, even though our understanding of disorders such as schizophrenia and anxiety has already been substantially enriched by neuroimaging studies and genetic findings4.

Nevertheless, there is an urgent need to integrate the latest scientific advances into the taxonomy of human disease. Failure to do so perpetuates ineffective treatment in medicine and exacerbates the pressures on the current pharmaceutical business model, which is becoming increasingly unsustainable in the face of rising costs of pharmaceutical R&D and health-economic pressures. Given the enormity of the task, major collaborations — among academic institutions, pharmaceutical and biotechnology companies, regulatory agencies and funding agencies — need to be established to agree which diseases to focus on first, and to combine their respective expertises in a precompetitive manner. A promising first step would be to hold discussions between heads of R&D at major pharmaceutical companies, the US National Institutes of Health, the US Food and Drug Administration, the European Medicines Agency and the Innovative Medicines Initiative (IMI) on how to approach the task. Furthermore, these are precisely the type of projects that could be funded through initiatives such as the IMI and the new US National Center for Advancing Translational Sciences.

Competing interests statement

The authors declare no competing financial interests.

top of page

Kola, I. and Landis, J. Can the pharmaceutical industry reduce attrition rates? Nature Rev. Drug Discov. 3, 711–715 (2004).
Arrowsmith, J. Phase II failures: 2008–2010. Nature Rev. Drug Discov. 10, 328–329 (2011).
Billings, L. K. & Florez, J. C. The genetics of type 2 diabetes: what have we learned from GWAS? Ann. NY Acad. Sci. 1212, 59–77 (2010).
Hyman, S. E. Can neuroscience be integrated into the DSM-V? Nature Rev. Neurosci. 8, 725–732 (2007).
top of page
Author affiliations

Ismail Kola is Executive Vice President of UCB Pharma and President of New Medicines, UCB, Anderlicht, Brussels, Belgium.
Email: ismail.kola@ucb.com
Sir John Bell is Regius Professor, Oxford University, Oxford, United Kingdom.
Email: regius@medsci.ox.ac.uk
Published online 31 August 2011

top of page
Previous article
Next article

Partager cet article

Repost 0
Published by Chronimed
commenter cet article