What can your DNA say about your risk of opioid addiction?

A new test suggests that, with a bit of your DNA, it can help identify whether you have a genetic susceptibility to opioid use disorder.

Aided by an influx of fentanyl and other substances, the number of drug overdose deaths continues to rise in the United States, a problem that only worsened during the COVID-19 pandemic. In 2022, data from the U.S. Centers for Disease Control and Prevention tallied 109,540 drug overdose deaths, most of which involved opioids.

In December 2023, the U.S. Food and Drug Administration approved AvertD (pronounced “averted”) for patients 18 and older who had never used opioids and were expecting to be treated for acute, not chronic, pain. It is the first genetic test to identify those at risk for opioid use to gain approval—and the only FDA-approved polygenic risk test for any psychiatric condition to date.

(Could monoclonal antibodies be the solution to the opioid crisis?)

“The opioid crisis, one of the most profound public health issues facing the United States, calls for innovative measures to prevent, diagnose and treat opioid use disorder, including to assess the risk of developing the disorder,” said the FDA in a statement announcing the decision. “This approval represents another step forward in the FDA’s efforts to prevent new cases of OUD.”

But some experts in the psychiatric community have expressed skepticism that any polygenic test—meaning a test that measures small inputs from many genes—can meaningfully identify people at high risk for psychiatric conditions like opioid addiction.

Psychiatric geneticist Arpana Agrawal from Washington University in St. Louis says that although genetics is important in understanding addiction, researchers still don’t know enough to predict who is at risk of addiction from genetics alone. Patrick Sullivan, a psychiatrist at the University of North Carolina, Chapel Hill, and principal investigator of the Psychiatric Genomics Consortium, agrees. DNA only explains a small piece of why someone becomes addicted to opioids or develops an illness like schizophrenia.

“It’s not as straightforward as we want to think,” Agrawal says.

Still, these and other experts National Geographic spoke to agree that polygenic risk tests offer promise for a range of other diseases from cardiovascular disease to Type 2 diabetes.

The promise of polygenic testing

Before the human genome was sequenced, geneticists focused much of their work on conditions caused by mutations in a single gene, such as cystic fibrosis and hemophilia. The work was groundbreaking, but it didn’t address the more common conditions like hypertension, high cholesterol, and diabetes.

Rather than a devastating impact from a single gene, many chronic diseases result from a complex interaction of environmental factors (such as pollution, childhood trauma, and food accessibility) and small inputs from hundreds, even thousands, of gene variants. Individually, the effect of each gene variant was insignificant. But scientists like Sekar Kathiresan, a cardiologist and geneticist who founded Verve Therapeutics, believed that all of these tiny influences could add up to something major.

(How personalized medicine is transforming your health care.)

His team’s first breakthrough came with a March 2008 publication in the New England Journal of Medicine, which combined the impacts of nine genetic variants into a unified genetic risk for cardiovascular disease. Kathiresan and colleagues could combine the tiny impacts of many genes across a person’s entire complement of DNA into a single assessment of genetic risk.

“It was a proof of concept that maybe this is a key part of how disease risk comes about,” Kathiresan says.

Scientists soon found an increasing number of genetic variants that influenced cardiovascular disease risk, making polygenic risk scores even better at identifying people whose genes predisposed them to heart disease.

Those people were treated more aggressively with statins, says Robert Green, a medical geneticist at Harvard Medical School, director of the Genomes2People Research Program at Mass General Brigham, and a paid consultant to Allelica, a company that sells polygenic risk score tests. Subsequent studies showed this likely helped lower their risk of heart attack and stroke.

The limitations of polygenic testing

Scientists began looking into other applications for polygenic risk scores. For some conditions, such as diabetes, Alzheimer’s disease, and breast and prostate cancers, scientists were able to calculate polygenic risk scores that were especially useful for screening and preventing disease, according to Kathiresan.

In other areas, however, researchers began running into problems—particularly as they attempted to identify the DNA variations that could contribute to conditions such as bipolar disorder, schizophrenia, and opioid use disorder.

(What exactly is DNA? Here’s what you need to know.)

Study after study has shown that risk for many mental health conditions is highly heritable, and people diagnosed with them have an array of genetic variants that differ significantly from those without.

But what’s often more challenging than calculating a specific risk score is determining how useful it will be in medical practice. Take height, says Sullivan. The average man is significantly taller than the average woman, yet there are lots of short men and tall women. As a result, you can’t accurately guess a person’s sex based on their height, Sullivan says. Similarly, scientists might be able to detect genetic differences between those who have a psychiatric condition and those who don’t, but there’s too much overlap between the two groups to tell them apart.

Scientists are still not sure why. One possibility is that they don’t yet know enough about the underlying genetics of psychiatric conditions to be able to use a test to identify high and low genetic risk. The other possibility is that the genetic differences between people at high and low risk just aren’t quite different enough to be medically useful.

Additionally, individuals of white European ancestry are disproportionately represented in most genetics studies, which means researchers know less—often much less—about the genes of people with other ethnicities, according to a 2019 study. This makes it harder to calculate a polygenic risk score for these populations, says Sullivan.

Nor do polygenic risk scores measure the full breadth of a person’s risk of developing a disease. A polygenic risk score can’t account for the myriad environmental factors that often have a greater impact than genetics on who gets sick. And a person’s environment is particularly malleable, which provides more opportunities to shift the needle in the direction of health.

“DNA is not destiny,” says Kathiresan. “It’s one component, and it’s not deterministic.”

Genetics and opioid use disorder

Still, the promise of a test that could prevent someone from getting addicted to opioids has been alluring to many scientists.

Keri Donaldson—founder and CEO of Solvd Health, the manufacturer of AvertD—felt that artificial intelligence algorithms were up to the task of identifying individuals at high genetic risk of developing opioid use disorder.

Instead of isolating specific genetic variants to calculate someone’s risk, Donaldson asked a computer to determine the genetic differences between individuals who misused opiates and those who didn’t. The study analyzed millions of points along the genome and identified 15 genetic variants that could help distinguish between these two groups.

The collective impact of these variants is calculated using a mathematical model that gives each test-taker a score between zero and 1. A score greater than 0.33 indicates an elevated genetic risk for opioid misuse. The higher the score, the greater the risk. That information forms the basis of AvertD.

But when Agrawal’s team tried to use artificial intelligence to predict opioid use disorder risk, they were unable to replicate the results. She says researchers need more diverse, larger studies to be able to identify it with a test.

“Anytime we look at these polygenic factors, we have to consider that they’re only ever going to be part of the picture. Environmental factors are going to be the other half,” Agrawal says.

Donaldson agrees that while AvertD isn’t a crystal ball, “not providing that information is not the answer.”

He defends the science behind the test, citing one study that shows AvertD’s algorithms were able to distinguish between individuals with and without opioid use disorder over 80 percent of the time. “Opioid use disorder is a complex trait, both nature and nurture. We’re informing on the genetic portion of it,” he says.

Understanding your own risk

As these tests gain popularity, clinicians will have to learn how to interpret these results with their patients. Donaldson says that prescriber education is key.

“Can we help individual patients, as well as prescribers, understand risk differently? That was the question we started on,” Donaldson says.

The scores can’t be read like a weather forecast, he says. An AvertD score of 0.5, for example, doesn’t mean you have a 50/50 chance of becoming addicted to opioids. Rather, it means you have an above average risk and might want to pursue pain management options that don’t involve opiates.

“With AvertD risk scores in hand, physicians and patients can engage in informed conversations regarding opioid sparing techniques or alternative pain management options,” he says.

Yet Agrawal worries about what this might mean for people whose tests suggest they are at high genetic risk of opioid addiction. She argues that many doctors still don’t understand how to manage pain without opioids—which could lead to needless suffering.

Instead of identifying opioid addiction, she prefers to focus on “how we can make an easier pathway to recovery or safety,” she says.