Emily S. Winn-Deen, PhD, is vice president, strategic planning and business development, at Cepheid (Sunnyvale, CA).
She can be reached at
Over the past year, there has been real progress in advancing “personalized medicine” from a concept to a standard part of medical care. Convincing genotype-phenotype correlation data have been developed for a few clinical applications, and corresponding tests and drug labeling have been approved by FDA. In addition, FDA has clarified its data requirements for studies to support clinical utility claims on such tests, and has published a special controls guidance document for tests designed to genotype drug-metabolizing enzymes.
Despite this progress, there are still a significant number of barriers that need to be overcome. Practicing physicians, and those who are still in training, will need to be educated on how to properly use evolving genetic knowledge in patient management. This need will be addressed primarily through the standard mechanisms for delivering continuing medical education. Once trained, physicians will also have to integrate this new information into their practice, a process that can often be slow.
Patients, who are becoming ever more involved in their medical care choices, will also need to be educated on these advances. As personalized medicine becomes more widespread, it will become increasingly important for patients to feel protected from any real or perceived potential for job or healthcare discrimination based on genetic test results.
Finally, the reimbursement system, already in an era of scarce healthcare dollars, will need to balance the costs of these new approaches with their benefits. Public policy will require ongoing attention to assure equal access to personalized medicine for all patients.
Establishing Clinical Utility
One of the most difficult scientific challenges to the adoption of personalized medicine is the need to develop strong data sets to support the predictive value of genetic tests. This requires understanding the role of both genetics and common environmental factors such as diet, as well as concomitant drug therapy. The completion of the international HapMap project last year will make it easier to tease out the genetic components of disease. The human genome has been further characterized by defining haplotype blocks, and the selection of markers for genotype-phenotype studies can now be made in a much more scientific manner. This will likely have the most immediate impact on genetics studies with a statistically powered sample collection, primarily for traits with high penetrance and a relatively strong component of genetic risk.
For the more common, complex diseases, for which genetic factors complement environmental factors to explain an individual's susceptibility, the study cohorts will need to be much larger to wield the statistical power to bring to light these more subtle effects. To address this need, NIH is now discussing the creation of large population studies. These types of studies are expected to take several years to design and decades to yield useful results.
The study cohorts would be similar to the longitudinal Framingham heart study, and each would require careful study design and the education of potential participants. For example: Should the study cover all health issues, or should it focus on specific complex diseases such as diabetes and heart disease? How long does it need to run to reveal environmental effects? How young should people be when they are enrolled? Over the next few years, NIH will convene study groups to address each of these study design issues, as well as to receive input from the public on how volunteers would prefer to participate in such studies.
FDA has been working with the pharmaceutical and IVD communities to help formulate its policies on reviewing and regulating the incorporation of pharmacogenomic markers into drug trials and product labeling. In addition, the groups have been discussing how best to regulate genetic testing, specifically pharmacogenetic testing. The first draft of FDA's pharmacogenomics guidance document for the pharmaceutical industry was published in November 2003 and, after receiving industry comments, FDA released the finalized guidance document in March 2005. Also in March, FDA published a Class II special controls guidance for industry and FDA staff on drug-metabolizing enzyme genotyping systems. A companion diagnostics white paper was published the following month as a precursor to development of an FDA guidance document on this subject. In February 2006, FDA published draft guidance for industry and FDA staff on pharmaco-genetic tests and genetic tests for her-itable markers.
In 2005, a landmark test—the AmpliChip CYP450 assay by Roche Molecular Diagnostics (Pleasanton, CA) for polymorphisms found in the CYP2D6 and CYP2C19 genes—was approved by FDA through the de novo classification process. For this first genotyping microarray, FDA required each gene test to be submitted and cleared separately, even though they are sold as a single product. As part of the approval process, FDA published a new special controls guidance document. This paper described the agency's expectations for future 510(k) submissions for IVDs used to genotype drug-metabolizing enzymes, and for clinical multiplex test systems.
Roche's AmpliChip targets the most common variants of the cytochrome P450 oxidase, which metabolizes more than 60 drugs currently approved by FDA. While the product labeling for most of these drugs does not yet contain information about dose adjustment depending on genotype, a relatively new drug, Strattera, by Eli Lilly and Co., mentions CYP2D6 information in seven places on its label, including in the sections on pharmacokinetics, adverse events, drug-drug interactions, and laboratory testing.
Last August, Third Wave Technologies Inc. (Madison, WI) released an FDA-cleared test for a polymorphism in the UDP-glucuronosyltransferase 1A1 (UGT1A1) gene. This polymorphism is associated with the metabolism of the colorectal chemotherapy drug Camptosar (irinotecan), which is manufactured by Pfizer Inc. About 10% of Caucasian Americans carry two copies of the genetic variant that inhibits metabolism of the drug, and this results in a much higher risk of toxic reaction. FDA approved safety labeling revisions two months earlier, in June 2005, to advise reducing the dose of irinotecan in patients who are homozygous for the UGT1A1*28 allele.
In a more controversial action, an FDA advisory panel also recommended last June that BiDil, a heart failure drug by NitroMed Inc., should be approved specifically for use by African -Americans. The therapy, which combines two generic drugs, works by raising the body's level of nitric oxide, thus helping the heart function more efficiently. In its phase III trial, carried out with 1050 African-Americans, BiDil demonstrated significant improvements in survival and a reduction in rehospitalization rates for treatment of congestive heart failure.
Was the African-American-specific data from this trial a case of substituting race for a genetic trait associated with drug response that we have yet to uncover? Is race a suitable surrogate marker until the underlying genetics are better understood? The nine-member FDA panel voted unanimously to recom-mend the approval for BiDil. The vote for labeling indications limited to African-Americans was 7 in favor and 2 opposed.
Prevention of Genetic Discrimination
Protection against patient discrimination is still needed to overcome the public's fear of genetic information misuse. Instances of genetic discrimination by both insurance companies and employers have been documented during public testimony before the HHS Secretary's Advisory Committee on Genetics, Health, and Society (SACGHS).1 Fear of genetic discrimination is causing some people to forego potentially lifesaving genetic testing and is discouraging others from participating in clinical trials. A federal law could create a national uniform standard, ensuring that all Americans would receive the same minimum protections, without preempting existing or future state laws.
Where do we stand today? Current patient protection of genetic test results is at least as good as that for any other confidential medical information. Almost all states have laws regulating the use of genetic tests by insurers and employers. The Health Insurance Portability and Accountability Act of 1996 regulates how group insurance can and cannot use genetic information. Similarly, the U.S. Equal Employment Opportunity Commission (EEOC) has said that the Americans with Disabilities Act of 1990 applies to the use of genetic information in employment settings.
In February 2000, President Bill Clinton issued an executive order protecting federal workers from discrimination on the basis of genetic test results. This order prohibits federal employers from requiring or requesting genetic tests as a condition of being hired or receiving benefits. It also prohibits federal employers from using protected genetic information to classify employees in a manner that deprives them of advancement opportunities and provides strong privacy protections to any genetic information used for medical treatment and research. Various bills extending this protection to all Americans have been proposed by Congress, but none to date have been approved.
The latest attempt to pass federal legislation is the Genetic Information Nondiscrimination Act of 2005 (S. 306 and H.R.1227). On February 16, 2005, President George W. Bush offered his full support of the act in a Statement of Administrative Policy, and on the following day, the U.S. Senate version passed by a unanimous vote of 98–0. This bill would prohibit health insurers and employers from discriminating against individuals based on their genetic information. Specifically, the bill protects consumers by taking the following steps:
• Prohibiting insurers in both the group and individual health insurance markets from “requesting or requiring” genetic testing of an individual or family.
• Prohibiting insurers from using genetic information to determine eligibility or establish premiums.
• Prohibiting employers, including employment agencies and labor organizations, from “requesting or requiring” genetic testing of an individual or family.
• Prohibiting employers from using genetic information to make hiring or promotional decisions, or when determining eligibility for training programs.
In March 2005, the companion bill was introduced in the U.S. House of Representatives and was referred to the Committees on Energy and Commerce, Education and the Workforce, and Ways and Means, where it still remains. Unlike previous versions of the legislation, H.R.1227 narrows the definition of “genetic information,” specifically excluding protections for genetic tests related to “manifest disease,” requires claimants to exhaust administrative state and federal EEOC procedures before seeking court damages, and limits the amount of punitive damages that can be awarded. H.R.1227 now has over 180 co-sponsors, but still needs more votes to pass the House and become federal law.
Access to personalized medicine depends greatly on the coverage of this type of care by health insurance. SACGHS has spent considerable time over the past two years reviewing the issue for the secretary of HHS. In April 2005, SACGHS released a draft report on coverage and reimbursement of genetic tests and services for public comment. The report contained a series of recommendations for improvements in the federally controlled part of the healthcare system.2 These recommendations focused on the following areas:
• Obtaining coverage for preventive services such as screening for genetic risk factors.
• Developing a strong evidence base for establishing the clinical utility of genetic tests.
• Developing fair and reasonable fee schedules and payment rates for genetic tests.
• Recognizing the important role played by genetic counselors and covering their services.
• Training allied health professionals to provide services such as obtaining informed consent and family history, and integrating family histories into patient medical records.
The U.S. healthcare system is already under tremendous pressure to reduce costs and will remain so for the foreseeable future. For each instance in which the cost of a new test is added to the system, reduction in other costs may need to be demonstrated. For example, can getting a patient on the proper medication dose sooner save costs associated with longer hospital stays, number of doctor visits, or monitoring for drug efficacy? Will personalized medicine reduce the costs, both medical and legal, of dealing with adverse events?
Today, most patients receive the same basic wellness advice from family physicians—eat a balanced diet, don't smoke, exercise regularly, and get plenty of sleep. The concept of personalized medicine is centered on predicting which therapies we will most likely respond to once we get sick. Personalized medicine of the future will also need to address personalized wellness advice. This will require the development of much better data on the approaches that will benefit each patient. There is still plenty of work left before the full power of personalized medicine is realized, but progress is now quite visible.
1. “Public Perspectives on Genetic Discrimination: September 2004–November 2004,” National Institutes of Health Web site (Bethesda, MD; 2004 [cited 15 January 2006]); available from Internet: www4.od.nih.gov/oba/sacghs/reports/Public_Perspectives_GenDiscrim.pdf.
2. “Coverage and Reimbursement of Genetic Tests and Services,” Secretary's Advisory Committee on Genetics, Health, and Society Web site (Bethesda, MD; 2006 [cited 5 May 2006]); available from Internet: www4.od.nih.gov/oba/sacghs/reports/CR_report.pdf.