Study Says DNA’s Power to Predict Illness Is Limited
http://www.nytimes.com/2012/04/03/health/research/dnas-power-to-predict-is-limited-study-finds.html
April 2, 2012
By GINA KOLATA
If every aspect of a person’s DNA is known, would it be possible to predict the diseases in that person’s future? And could that knowledge be used to forestall the otherwise inevitable?
The answer, according to a new study of twins, is, for the most part, “no.”
While sequencing the entire DNA of individuals is proving fantastically useful in understanding diseases and finding new treatments, it is not a method that will, for the most part, predict a person’s medical future.
So, the new study concludes, it is not going to be possible to say that, for example, Type 2 diabetes will occur with absolute certainty unless a person keeps a normal weight, or that colon cancer is a foregone conclusion without frequent screening and removal of polyps. Conversely, it will not be possible to tell some people that they can ignore all the advice about, for example, preventing a heart attack because they will never get one.
“The punch line is that this sort of personalized medicine will not in any way be the most important determinant of patient care,” said Dr. Bert Vogelstein of Johns Hopkins, who, with his colleagues and his son Joshua, analyzed the power of sequencing all of a person’s DNA to determine an individual’s risk of disease. The study, published online Monday in the journal Science Translational Medicine, involved data from 53,666 identical twins in registries from the United States, Sweden, Finland, Denmark and Norway. The registries included data on 24 diseases, telling how often one twin, both or neither got a disease.
Since identical twins share all of their genes, the investigators could ask to what extent genes predict an increased chance of getting a disease. Using a mathematical model, they reached an answer: not much. Most people will be at average risk for most of the 24 diseases.
They asked: Would those who ultimately got one of the 24 diseases have been forewarned by DNA sequencing? “Unfortunately, it tells them they are at roughly the same risk as the general population,” said Dr. Vogelstein.
The researchers also asked whether healthy people would learn by DNA sequencing that they were at low risk for a disease. Again, the results were disappointing. For example, more than 93 percent of women would learn they were at low risk for breast cancer and more than 97 percent of men and women would learn their risk for lung cancer was low. “But these negative tests do not mean they are at no risk for these cancers,” Dr. Vogelstein said. Their risk is more like that of the general population. And, Dr. Vogelstein says, even knowing you are at high risk for a disease may be less useful than it sounds. A woman who is at high risk for ovarian cancer might have a 10 percent risk, many times higher than average. That, Dr. Vogelstein said, “is unlikely to be the main determinant of her health.” But there was one positive finding — as many as 90 percent of people would learn that they are at high risk of getting at least one disease. And gene sequencing could, in theory at least, identify as many as 75 percent of those who will develop Alzheimer’s disease, autoimmune thyroid disease, Type 1 diabetes and, for men, heart disease.
However, with the exception of heart disease, there is as yet no way to prevent these diseases or slow their progress. And since high risk of an infrequent disease, like ovarian cancer, is far from a prediction that the disease is in the person’s future, the information might be valuable but would not necessarily make much difference in the end.
“The general point is absolutely correct,” said Dr. David Altshuler, professor of genetics and medicine at Harvard Medical School, who was not involved with the research. “Even if you know everything about genetics, prediction will remain probabilistic and not deterministic.”
The reason, he suspects, is that behavior, environment and random events tip the balance. “I am a big believer in randomness,” Dr. Altshuler said.
Dr. Vogelstein is too, but he had hoped the study might prove him wrong. He and his colleagues had studied a patient with pancreatic cancer. Several family members had also developed this rare disease, and so Dr. Vogelstein and his colleagues decided to determine the sequences of the patient’s genes, looking for a mutation.
“Indeed, we found the culprit,” Dr. Vogelstein said.
Several other research groups looked at families with other diseases and also found unexpected genetic culprits by sequencing all of a patient’s DNA.
“It occurred to us that maybe we could do this for everyone,” Dr. Vogelstein said. “Maybe we would find that most disease risk was concentrated in a relatively small number of people. That would have dramatic health policy implications. It would mean we could concentrate our surveillance on that proportion of the population that was at high genetic risk.”
The twins study let him see what might be possible. And, he says, “it puts limits on what people might expect with this sort of testing.”
Other experts pointed out different aspects of DNA sequencing that can improve health and medical care. Sequencing can, in some cases, aid in determining a patient’s prognosis. It can find the causes of mystery ailments in individuals, and it can find mutations that appear to be driving the growth of cancers in individual patients.
Sequencing also is starting to help doctors decide who should take drugs to prevent diseases, as is happening with heart disease.
In heart disease, one pressing problem is how to decide which young and middle-aged adults would benefit from cholesterol-lowering statins to reduce the risk of a first heart attack, said Dr. Sekar Kathiresan, a genetics researcher who is director of preventive cardiology at Massachusetts General Hospital. The drugs reduce the risk by 20 percent, but if your risk is low to start with, a 20 percent reduction does not mean much.
Now, Dr. Kathiresan said, by analyzing data from studies that sequenced entire genomes, researchers have found 30 gene variants that, taken together, can identify healthy people who have twice the average risk of heart disease. “There is a great attraction to using genetics in this way,” he says.
Dr. Robert Cook-Deegan, professor of law, ethics and policy at Duke, notes that every person whose DNA is sequenced will get information about whether he or she will respond to certain drugs and whether certain side effects will result from taking certain drugs. Vanderbilt University is already doing genetic analyses of patients to help in prescribing a short list of drugs, says Dr. William Schaffner, chairman of the department of preventive medicine at its medical school.
But the real benefit of studying the human genome, Dr. Altshuler said, is not to predict people’s medical futures but instead to understand how diseases occur and to use that knowledge to develop better therapies. Already this sort of work has succeeded with an entirely new type of drug to lower levels of LDL, or “bad” cholesterol, he said.
“The reason we do it is because we want to use genetics to pry open the black box of how disease works,” Dr. Altshuler said. “Not to personalize existing treatments, but to develop new treatments that are more effective.”
And that, he said is “a work in progress.”
Roberts NJ, Vogelstein JT, Parmigiani G, Kinzler KW, Vogelstein B, Velculescu VE. The Predictive Capacity of Personal Genome Sequencing. Science Translational Medicine. The Predictive Capacity of Personal Genome Sequencing
New DNA sequencing methods will soon make it possible to identify all germline variants in any individual at a reasonable cost. However, the ability of whole-genome sequencing to predict predisposition to common diseases in the general population is unknown. To estimate this predictive capacity, we use the concept of a "genometype". A specific genometype represents the genomes in the population conferring a specific level of genetic risk for a specified disease. Using this concept, we estimated the capacity of whole-genome sequencing to identify individuals at clinically significant risk for 24 different diseases. Our estimates were derived from the analysis of large numbers of monozygotic twin pairs; twins of a pair share the same genometype and therefore identical genetic risk factors. Our analyses indicate that: (i) for 23 of the 24 diseases, the majority of individuals will receive negative test results, (ii) these negative test results will, in general, not be very informative, as the risk of developing 19 of the 24 diseases in those who test negative will still be, at minimum, 50 - 80% of that in the general population, and (iii) on the positive side, in the best-case scenario more than 90% of tested individuals might be alerted to a clinically significant predisposition to at least one disease. These results have important implications for the valuation of genetic testing by industry, health insurance companies, public policy makers and consumers.
