The company 23andMe promises to "unlock the secrets of your own DNA." Navigenics wants you to be tested to "do everything you can to stay healthy." And deCODEme hopes that genetic testing will "prompt people to do the right thing."
It all sounds so good. If you have a couple of thousand dollars to part with (along with some saliva), why not have one of these companies scan your genome?
The primary caution about genetic testing has usually been that you will learn that you are destined to develop some dreadful disease (such as Huntington's disease, a degenerative neurological disorder) for which there is no known therapy. A positive test only allows you to start worrying about your demise earlier. Do you really want to know?
Then again, your genome includes lots of other information, which the scans are beginning to tap. So the more relevant question is: What is the point of knowing?
Some of what you can learn is how well your genotype (your DNA code) relates to your phenotype (your personal attributes). The testing companies tout their ability to see whether your DNA says that you like Brussels sprouts, can tolerate dairy or are plagued by ear wax.
This may be a novel hook for some, but it's really pretty frivolous stuff. Geneticists point out that "phenotype trumps genotype." In other words, the fact that you don't like Brussels sprouts trumps any associated DNA code that says you should.
You can also learn something far more serious from the tests information about your risk of developing a number of diseases: heart disease, diabetes and multiple sclerosis, as well as breast and prostate cancer. But the information is not as good as you might think.
It would be simpler if all genetic information were definitive: If you have the gene, you will get the disease; if you don't, you won't (as is the case with Huntington's disease). Most gene variants, however, only marginally increase or decrease the chance of getting cancer or other diseases. And because so many nongenetic factors contribute to illnesses, having a rough estimate of a small genetic effect doesn't help us predict future health with any certainty. And even if the information were perfect, it wouldn't tell you what to do.
Imagine you are a 40-year-old woman receiving the results of your genome scan. Your profile shows that you have four times the average risk for developing ovarian cancer but below-average risk for lung cancer. Your risk for breast cancer could be almost 50 percent increased, but it also might not; the research is divided. Your chances of developing heart disease turn out to be 25 percent higher than normal. (Not surprising, you think, because your mother had a heart attack a couple of years ago.) One factor that raises your risk of heart disease also lowers your risk for macular degeneration but you also have a second genetic variant that raises the risk of that disease.
Now what? Oops, the test doesn't answer that question. Some might argue that the first step would be to deal with the ovarian-cancer risk and remove your ovaries. Others might point out that heart disease is a greater concern and that taking out the ovaries, removing estrogen, would only increase your risk. Another doctor might suggest removing your ovaries and starting estrogen replacement. Someone else will point out that that will increase your risk for breast cancer. (These uncertainties, combined with the absence of increased lung-cancer risk, may tempt you to keep smoking.)
What's the right thing to do? With the exception of quitting smoking, the truth is: No one knows. Our ability to read the genome is well ahead of our ability to know whether medical intervention based on such a reading does more good than harm. But we can be sure that haphazard genetic testing will needlessly make well people worry about becoming sick.
H. Gilbert Welch is an internist with the Department of Veterans Affairs and professor at Dartmouth. Wylie Burke, a medical geneticist, is a professor at the University of Washington.