Predicting a New Disease: Pathological Consumption of Genetic Information

The current issue of the New England Journal of Medicine (10 January 2008; PMID: 18184955) contains a cleverly titled article “Letting the Genome out of the Bottle – Will We Get Our Wish?" The article, which explores the new phenomena of commercialized personal genome analysis through genetic profile tests capable of identifying several hundred thousand variations in any one genome, has drawn the attention of several astute bloggers, including: Myles Axton of Free Association, Blaine Bettinger of The Genetic Genealogist, Hsien-Hsien Lei of Eye on DNA, and Steve Murphy of Gene Sherpas. The general conclusion of the NEJM authors is that perhaps a person may get their wish, but the clinician certainly won’t. Why is that? Because the predictive value of many of the variations identified by the tests is questioned as is the utility of the information itself, i.e. what will the patient do with the information? While both of these arguments are accurate, they need to be tempered with common sense and a degree of humility rarely found when scientists address general society.

Beginning with the first point, the predictive value of these tests is limited. Lead author Dr. David Hunter develops this point in a US News and World Report interview with Nancy Shute (Why Not to Buy a Scan of Your Genome, 9 January 2008). Hunter notes that the predictive value of many of the genes or single nucleotide polymorphisms (SNP’s) found in these genetic profiles pales in comparison to the predictive value of tried and true genetic tests for specific genes like BRCA 1 and 2. To be sure, he is correct, but interestingly the predictive value of the BRCA gene is being called into question this month as well. A large population-based case-control study published in the Journal of the American Medical Association (Begg CB, et al. Variation of Breast Cancer Risk Among BRCA1/2 Carriers. JAMA. 2008;299(2):194-201. PMID: 18182601) suggests the BRCA 1 and/or 2 gene alone may not be as predictive as once thought, and that a variety of other genes may explain the strong familial clustering of breast cancer. What this means is that the more we know about genetics, the more we recognize the limits of our knowledge. How then can adding personal genome profiles adversely affect this pool of knowledge?

Moving now to the second point: the clinical utility of general genome profiles is questionable. Certainly this is the case. Still, the authors need to keep "personal" aspects of these profiles in mind. This is not a test marketed to health professionals to guide treatment; it is a test marketed to guide lifestyle, a type of guidance which most physicians are admittedly poor at providing for their patients anyway. Further still, the lifestyle changes dictated by carriers of the BRCA gene may be very dramatic—possible removal of both breasts and ovaries; contrast these to the lifestyle changes encouraged by a personal genome scan indicating an increased risk for heart disease—increased exercise and proper diet.

In sum, Dr. Hunter and his colleagues are right to raise their concerns about these tests: the tests have limited capability, unknown utility, and are expensive. Still, the capability of any medical test can only be magnified by an increase in data. Furthermore, an unknown clinical utility does not mean that a person cannot derive some utility from knowing their own genome profiles. If this were the case, why would they purchase the test at all? Perhaps Dr. Hunter’s lament is in part that the human genome, once the bastion of modern orthodox medical science, now will be shared with alternative medicine in a very real and technical way. Making this point clear is what is necessary, not admonishing patients on how to spend their money. - Patrick Barrett