Why Family Medical History Trumps DNA Testing

dna-ta03-hsmall
Andrew Brookes / Corbis

In this age of DNA, with for-profit companies scanning customers’ genomes for DNA variants associated with disease and the “$1,000 genome” within reach, it seems positively antediluvian to bother with family medical histories. Why ask aging relatives what Uncle Harry died of when you can scrape some cells from your cheek and ship them off to a sequencing company to learn which diseases your DNA—which you might or might not share with Uncle Harry—has in store for you?

For two important reasons. The simplest is that many of the studies linking a gene variant to one or another disease are incorrect at worst and useless at best. The more interesting reason to rely on your family medical history is that genes do not act in a vacuum. Those realizations have now prompted the American Society of Human Genetics (ASHG) to declare family medical history “the gold standard” for assessing what diseases you are at risk for—a conclusion as surprising, given the source, as if Exxon declared solar-powered cars superior to the gasoline variety.

An estimated two thirds of the early studies linking an aberrant form of a gene to a particular disease turned out to be mistaken. The latest to implode is the purported link between a variant of a gene called KIF6 and cardiovascular disease. A study of 57,000 people found that genetic status had essentially no predictive value for coronary-artery disease, as had been claimed (the test is sold commercially for about $100). “This study puts the nail in the coffin,” said Tom Quertermous of the Stanford University School of Medicine.

More interesting are correct but useless discoveries, in which a gene variant indeed raises your risk of some disease but knowing your genetic status doesn’t make predicting your medical future any more accurate. A study this year looked at four gene variants associated with Alzheimer’s. Adding them to traditional risk factors such as age and sex did nothing to improve prediction. A 2009 study focused on a genetic variant on chromosome nine linked to an elevated risk of cardiovascular disease (CVD). But when scientists added genetic status to traditional risk factors (such as high blood pressure, smoking, high cholesterol, and family history of heart attack at a youngish age), the crystal ball for who would develop CVD was no more accurate than with traditional factors alone. And not to beat a dead horse, but a 2008 study of 19,000 people examined whether including the status of 16 genes that raise the risk of type 2 diabetes better predicts who will develop the disease than family history, high body-mass index, elevated liver-enzyme levels, smoking, and reduced insulin secretion. It did—by a paltry 0.01 in the units geneticists use (from 0.74 to 0.75).

No one is arguing to ignore DNA. The question is whether to read genes directly and in a vacuum—as personal genome scans do—or indirectly but in context, which family medical history provides, points out Charis Eng of the Genomic Medicine Institute of the Cleveland Clinic. That’s because whether a gene variant causes a disease depends in part on genetic background—your other 24,000 or so genes. After all, you share with relatives not only high-profile disease genes, but also run-of-the-mill ones. Their ability to quash or amplify the effect of disease-related genes remains unknown. But we get a glimpse of that effect in our relatives’ medical histories, notes Muin Khoury of the U.S. Centers for Disease Control and Prevention: in them, genetic background plus disease-specific genes have or have not led to that disease.

In research Eng presented at ASHG’s annual meeting this month, she found a serious clash between the risk for cancer as assessed by commercial DNA testing and by family medical history. For instance, family history correctly classified eight women as being at high risk for breast cancer. But only one of the eight was so classified by a commercial genetic test. Many of the companies scan for only one variant of the BRCA breast-cancer gene and so miss other mutations, giving women a false sense of security. “Family health history is a better predictor of cancer risk than personal genomic testing,” Eng says.

Not that family medical history is perfect either. Just because a relative died of cancer or a heart attack doesn’t mean you will: most diseases have strong environmental components, and Grandpa’s cancer might reflect an environment he did not share with you rather than genes he does. But until genetic testing improves, using online tools that let you collect and record your family health history will give you a more accurate glimpse of your health future than a personal DNA scan. Make it a Thanksgiving project.

Join the Discussion