Skip to Main Content

SAN FRANCISCO — It can be notoriously difficult for psychiatrists and patients to determine which antidepressant might be most effective, or which might cause side effects.

And so Color Genomics, a company that already sells genetic tests to determine someone’s risk of developing certain cancers, said this week that it will also begin to offer a DNA test to determine how well widely used antidepressants are likely to work for patients.

advertisement

With the new test (part of a $249 product), Color joins several dozen companies probing patients’ DNA in search of insights to help inform decisions about which psychiatry medications patients should take. They’re touting applications for depression, bipolar disorder, attention deficit hyperactivity disorder, and post-traumatic stress disorder.

But some top psychiatrists say the evidence doesn’t support the commercial rush.

In a review published this past April, a task force of the American Psychiatric Association’s research council concluded that such genetic testing is not ready for prime time in their field. “Although some of the preliminary published data sound promising,” the task force members wrote, “there is insufficient evidence to support widespread use of combinatorial pharmacogenetic decision support tools at this point in time.”

advertisement

Dr. Bruce Cohen, a Harvard Medical School psychiatrist who previously ran McLean Hospital and is now on staff there, told STAT he would not order Color’s test or any of the other comparable tests on the market for his patients. “They cost a lot of money, and they have absolutely no proven value,” he said.

Color’s test will initially look at two genes, CYP2D6 and CYP2C19, that the company says impact metabolism of 13 medications for mental health. Among the antidepressants it will report on are Zoloft, Paxil, and Lexapro.

A spokesperson for Color, based in the Bay Area, said the science behind pharmacogenetics for psychiatry “is sound and based on clear evidence and consensus statements.” The information Color provides is based on guidelines from the Food and Drug Administration and a consortium dedicated to implementing pharmacogenetics, the Color spokesperson said.

A pharmacogenetics initiative managed at Stanford also rated the two genes Color is analyzing as “very important” in how the body metabolizes and responds to certain drugs.

The psychiatry offerings from Color and other companies are part of a broader effort to try to use genetic testing to help inform medication choices for all sorts of different conditions. Leading hospitals including Mayo Clinic and Mount Sinai have programs to incorporate pharmacogenetics into clinical care. And the Cincinnati Children’s Hospital has been performing pharmacogenetic testing for neuropsychiatric medications for the last 14 years.

But compared with other fields, there’s relatively little evidence supporting pharmacogenetic testing in psychiatry.

Cohen drew a contrast between psychiatry and cancer. Oncology researchers have identified certain genes as major determinants in getting sick or responding to treatment. “In psychiatry, we don’t have that,” he said.

A 2013 meta-analysis looked at three studies that collectively gathered data on more than 2,200 patients, analyzing their genes and their response to psychiatric mediation. The result? “The authors identified no reliable predictors of antidepressant treatment outcome,” they wrote. They also ran an analysis that suggested that genetics may have accounted for between 1 and 2 percent of the differences in how patients responded. (Other research suggests the contribution is higher.)

The upshot: While genes seem to play some role in how well antidepressants work, factors like age, diet, and hormonal status may be far more influential.

Critics also say that research supporting the use of pharmacogenetics in psychiatry is flawed. Cohen co-authored an opinion article published last month in JAMA Psychiatry that looked at 10 published studies assessing the use of genetic testing to guide treatment for patients with major depressive disorder. They concluded that none of the studies “were adequately blinded and properly controlled” — features that are considered hallmarks of rigorous scientific research.

Another concern: Companies don’t share their secret sauce — how exactly they combine their analyses of different genetic variants to come up with an assessment of whether a drug is likely to work — so the formula can’t be independently reviewed.

Cohen said he understands that psychiatrists and patients are desperate for answers about which drug to try. But even basic questions about a patient’s symptoms, diet, and other medications will yield more insight than a genetic test, he said. “Why spend money on this when there’s a better way? Why not spend more time with the patient? Why not use the guidelines that are there? Why not get a consultation if you’re uncertain how to proceed?”

Just as with the spit kits marketed to people curious about their ancestry, Color’s customers mail in a saliva sample. But Color’s tests must be ordered by a physician — unlike genetic tests offered by direct-to-consumer companies like 23andMe — and factor genetic counseling into their price.

The venture-backed company launched in 2015 with a test analyzing 19 genes associated with common cancers that have a genetic component, including BRCA1 and BRCA2, the genes linked to elevated risk for breast and ovarian cancers. Color has since moved into genetic testing for cardiovascular risk: Last year it started testing for risk of high cholesterol, and earlier this year it expanded into testing for genetic predisposition to inherited heart conditions like cardiac arrhythmias.

Color announced on Thursday that it’s wrapping its cancer and heart tests into one product that also includes two new features: One is the 23andMe-style report on traits with no bearing on health, including whether you’re wired to like the taste of cilantro. The other is the report on what your genes say about how you’re likely to respond to medication; it will initially report on psychiatry drugs but expand in the coming months to include heart and pain medications.

The company is not accepting insurance for the new product.

Some psychiatrists may prove receptive to Color’s new pitch.

Dr. Erika Nurmi, a UCLA psychiatrist and neuroscientist who studies the genetic basis of psychiatric disorders and personalized medicine in psychiatry, has been an outspoken critic of companies that she says overhype the promise of pharmacogenetics in her field.

But Nurmi said she might consider ordering Color’s test if she suspected that a patient might have a drug metabolism problem, based on multiple medications not working or causing side effects. The goal, she said, would be to inform the dosing levels she prescribes. She added that she was pleased to see that Color is only basing its reports on genetic variants that she considers to be vetted.

“This is an example of where the industry needs to go to protect patient health,” she said.

STAT encourages you to share your voice. We welcome your commentary, criticism, and expertise on our subscriber-only platform, STAT+ Connect

To submit a correction request, please visit our Contact Us page.