April 16 (Bloomberg) -- Companies that sell DNA testing to pinpoint individual risks for common diseases provide little real information, says one of four commentaries featured in the New England Journal of Medicine.

David Goldstein, a Duke University researcher, wrote that most common disorders, including cancer and diabetes, involve hundreds of genes, and many mutations may have the capacity to raise a person’s risk. That suggests the testing offered by companies now, which focuses on the few variants that have already been identified, isn’t likely to spot people at highest risk, he said.

“In pointing at everything, genetics would point at nothing,” Goldstein wrote yesterday in the journal.

DeCode Genetics Inc., based in Reykjavik, Navigenics Inc., of Foster City, California, and 23andme Inc., of Mountain View, California, are among companies that test the DNA of individuals to pinpoint variants that may identify disease risk. The journal commentaries are the first public forum featuring leading genetic researchers discussing the usefulness of so-called genome-wide association studies, research that scans the DNA of thousands of people to identify mutations that may be common to certain ailments.

The headline on the lead commentary asks, “Are we there yet?” The answer suggested by the authors is not quite. Goldstein refers to commercially available genetic tests as “recreational.”

Still Learning

“We agree for the most part that we’re still really early in the process of learning the genetic programming for risk factors,” said Linda Avey, co-founder of 23andme. “We are always reaching out to the research community to communicate with them and this type of commentary is very frustrating for us.”

Researchers hoped the studies, made possible by completion of the Human Genome Project in 2003, would open new windows on why ailments such as cancer and diabetes occur, and allow doctors to offer highly personalized treatment regimes that would focus on the underlying cause.

At the same time, companies such as DeCode, Navigenics and 23andme have designed tests, costing as much as $1,000 apiece, that search through the genomes of individuals for hints disease may be hiding based on findings from the larger research.

The association studies helped make it “routine to identify common, low-risk variants” present in less than 5 percent of the population that confer only “small” risks of disease, wrote John Hardy, a researcher at the Institute of Neurology at the University College London, and Andrew Singleton, of the Laboratory of Neurogenetics in Bethesda, Maryland, in one of the four commentaries.

Flaw Cited

“There is a flaw in these conclusions because they don’t compare preventive genomic medicine to standards today,” said Vance Vanier, the chief medical officer of Navigenics in a telephone interview. “This is akin to saying that if the authors were told they had high cholesterol they’d have to wait until all environmental protocols were found before they were given treatment.”

Joel Hirschhorn, an associate professor of genetics at Harvard Medical School and a member of the Broad Institute in Boston, wrote in a separate report that much of the DNA identified as being more common in diseased patients involve genes that have no known functions.

New Paths

More work on these genes could lead to new paths in studying disease, Hirschhorn wrote.

“It is already clear that the genes being identified expose relevant biology,” he said. “The main goal of these studies is not prediction of individual risk.”

Though genetic variants are important to basic research, commercially available tests are likely to misrepresent the danger of these variants to consumers, Peter Kraft and David Hunter, epidemiologists from the Harvard School of Public Health, wrote separately.

This is perhaps because the genes that are highest risk are “almost certainly overrepresented in the first wave of findings,” they said.

It is still too early for tests to provide stable estimates of disease risks, Kraft and Hunter wrote.

Genome-wide association studies haven’t explained as much of the genetic components of disease as anticipated, wrote Goldstein, director of the Center for Human Genome Variation at Duke University in Durham, North Carolina.

Rare Variants

For that reason, scientists ought to spend more time looking at rare variants that could lead to new drugs or suggest the designs for personalized prevention programs, he said. Beyond the variants for Alzheimer’s disease, glaucoma and macular degeneration, “there are probably either no more common variants to discover, or no more that are worth discovering,” Goldstein said.

The genome-wide association approach is in contrast to another genetic approach, where researchers hypothesize a specific gene causes a certain illness, and check the patient population to see if it occurs in the sick people.

To contact the reporters on this story: Elizabeth Lopatto in New York at elopatto@bloomberg.net; John Lauerman in Boston at jlauerman@bloomberg.net.

Last Updated: April 16, 2009 16:34 EDT