Genetic Snake Oil?

Kathy Hudson is the founder and director of the Genetics and Public Policy Center at Johns Hopkins University. In "A Case Study of Personalized Medicine," published today in the journal Science, Hudson and her colleagues argue that the rapidly expanding world of genetic testing needs far greater oversight. Hudson spoke with NEWSWEEK's Claudia Kalb. Excerpts:

NEWSWEEK: Your center reports that there are now genetic tests for more than 1,500 diseases and conditions. Examples?
Kathy Hudson:
Most of them are for rare diseases that are hard to pronounce. They range from tests for birth defects and mental retardation to tests for various types of cancers, like retinoblastoma, breast cancer and colon cancer. And there are some tests for diseases that used to be considered quite fatal but are getting better with new therapies. Cystic fibrosis. Sickle cell anemia. Tay-Sachs. These are mostly genetic tests for single gene disorders, which means that if you have mutation in that gene, you will, or you're very likely to, have or develop that disease.

What else is out there?
There's another category of tests, which is what we wrote our paper about, an area where there's an enormous amount of enthusiasm. And there should be. This is the whole area of personalized medicine or pharmacogenetics. At the end of television ads for drugs, you hear somebody in a low voice saying, "This drug may cause dizziness or death." If you could figure out the genetic markers associated with how well you'll respond and whether or not you're at risk for adverse effects, that could be a phenomenally wonderful thing and it could save a bazillion dollars in health-care costs.

Tell us about your case study. What did you find?
We did a case study of genetic tests for a family of enzymes called CYP450. CYP450 genes are involved in the metabolism of SSRIs, which are drugs prescribed to treat depression. While these genes clearly play a role, there are not yet clinical studies showing that CYP450 genetic testing is beneficial for SSRI selection and dosing. Despite this lack of scientific evidence, what we found was that companies were making very specific claims to consumers that were either incomplete or blatantly false and misleading. They were claiming that CYP450 genetic testing was required in order to select the right SSRI and to determine the appropriate dose. This is unproven. Some of these companies are selling the tests directly to consumers over the Internet, so there's no health-care intermediary, no gatekeeper to prevent inappropriate testing or misinterpretation of test results.

Where's the oversight? Is the FDA involved?
There is an FDA-cleared test for CYP450. In this particular case the FDA verified that the test did what it said it did: it tested for CYP450. But the FDA didn't require that the manufacturer show any evidence of the clinical usefulness of the test. As a consequence other companies are making these claims, and the FDA doesn't really have the stick to beat them up with.

What should the FDA do?
In the ideal scenario, genetic tests that are linked to drug selection or dosing would be under the purview of the FDA. That's one recommendation. The second is that there be more government attention to the claims that are being made about genetic tests and that the government actually take action against those who make false and misleading claims. At the moment they're not doing that. The FTC could step in. It has authority over all consumer product advertising claims. But to date the FTC has not taken any enforcement action against genetic test manufacturers or laboratories.

How many of the 1,500 disease tests have actually been approved by the FDA?
About a dozen.

What are the implications of that?
Because most of the existing tests are for rare diseases, the public health implications are not very widespread. But as we start to test for complex diseases and have tests that are going to dictate treatment choices, those are going to have larger public health consequences. So looking to the future, it's going to be increasingly critical that we have a regulatory system in place, both to reassure investors that the regulatory climate is stable and to give doctors and patients the confidence and information they need to make really wise health-care decisions.

Unlike rare diseases, complex diseases are not one-gene disorders, right?
These are places in the genome where you tend to find a higher incidence of fill-in-the-blank disease. These "association" studies are really, really important because they are helping to identify all of the molecular actors that have a role in the play of, say, diabetes. Right now that's the big challenge for the research enterprise. To figure out which genetic molecular players are involved in the development of complex diseases.

Right now, though, scientists don't know all the genes involved in these complex diseases. So you might get tested and get falsely reassured, correct?
False reassurance is a bad thing. Let's say there's a lot of diabetes in my family. I get this test and I find out I don't have an increased risk over the general population. I'm going to be less worried, even if the test did not screen for all relevant mutations or take into account my diet and lifestyle.

You could also test positive for a complex disease and worry about a risk that's really pretty low.
If you've got an odds ratio of 1.3 and the actual risk of one person is something like one in 24,000, then your risk is insignificant. It's an insignificant finding. So what are people going to do with that? Say people get tested and they find they have a modest increased risk of developing type 2 diabetes. Let's say they fall into the same personality trait that I have, which is I worry a lot. They go dashing off to the doctor, saying, "I had my genome done and I'm at increased risk for diabetes." The doctor isn't going to know what they're talking about. And second, the doctor is going to feel he needs to do a whole lot of screening tests he wouldn't normally do. We're going to be sucking resources out of the health-care system for wealthy, worried, well people.

Are any of these tests making a positive impact on treatment yet?
It used to be you could get a test, you'd know you're at risk, and there wasn't a thing you could do about it except get depressed or go on a cruise. Now, increasingly, there are things you can do to reduce your risk. For a long time there was no good evidence to say there was anything you could do to lower your risk if you tested positive for the BRCA 1 or 2 mutation. Now it's clear that surgical removal of breasts and ovaries reduces risk. The Alzheimer's clinical trials for the drug Avandia are incredibly exciting. In an ideal world the drug will prevent Alzheimer's progression and also restore function. If you can test for Alzheimer's and know you're at increased risk and you're starting to develop symptoms, maybe one day you can take this drug.

So there's a clear upside to genetic testing.
I think genetics has a huge amount to offer in making really important real-time health-care decisions. I think increasingly we're going to see genetic testing as being really important in both prescribing decisions and in treatment decisions. For example, will you respond to chemotherapy or not? Will you keel over and die if I prescribe this antiviral versus that antiviral?

What are you hoping your paper will do?
Our hope is that our paper and other research will be a catalyst for policy change that is sorely needed. [Department of Health and Human Services] Secretary Mike Leavitt put together a personalized health-care initiative over a year ago, which had multiple components. One was ensuring the quality of genetic tests. We provide strong evidence for why he should enhance regulation for genetic tests. His time is waning. He has the opportunity for some leadership here in creating a good, nurturing, safe environment for the future of personalized medicine.