Yanna Lee

When prescribing opioid painkillers, physicians run the risk that their patients develop an opioid dependence. After new research from Yale scientists, doctors might be one step closer to predicting which patients are at risk for opioid use disorder — before patients first try the drugs.

Researchers at the School of Medicine and the University of Pennsylvania fine-tuned genetic prediction models for opioid dependence by considering a combination of factors, including genetics, socioeconomic level and mental health history. 

Their latest study, based on 1,958 volunteers and published last week in the journal Psychological Medicine, found that genes alone were weak predictors of opioid use disorder. Instead, they improved their ability to predict volunteers’ susceptibility to opioid dependence by factoring in personal experiences, such as a prior diagnosis of post-traumatic stress disorder, or PTSD. 

Their study helps clarify a larger dilemma facing physicians as they attempt to treat pain with opioids: How do doctors know which patients who need painkillers are at risk of developing opioid dependence? 

“Some people may blame the genes, some people might blame the social environment, but I would say both are important,” said Peter Na, one of the investigators and an assistant professor of psychiatry at the School of Medicine. 

This study is a part of a larger effort by the Yale-Penn research team that has recruited more than 14,000 participants to investigate substance use disorders. Since 1999, they have interviewed and collected genetic samples from these volunteers, building a repository of biological and background data on substance use disorders. 

In previous research, Joseph Deak, a member of the Yale-Penn team and an associate research scientist in psychiatry at the School of Medicine, analyzed whether any common genetic differences between individuals carried an increased risk for opioid use disorder. Looking at 7.5 million sites in the human genome, Deak identified 19 high-risk genetic locations that were common across individuals with opioid use disorder.

Still, Deak found that the genetic differences at those locations explained only a small fraction of different individuals’ risks for opioid use disorder: in other words, genes alone were not an effective predictor of opioid misuse, he said. 

Na said that experts consider opioid use disorder through a “biopsychosocial model,” considering how biology, psychology and social circumstances converge to explain addiction.

But scientists do not know how much of the vulnerability to opioid use disorder is due to genetic risks compared to other social and psychological factors, like income level, psychiatric diagnoses or childhood trauma. 

In their most recent study, the scientists re-ran Deak’s prior genetic analysis while factoring in participants’ non-genetic information — including partnered status, household income, sex and exposure to trauma — to see how non-genetic risk factors might explain the development of opioid use disorder. To do so, they examined interview data and genetic information from 1,958 patients who had enrolled in a prior Yale-Penn study. 

After correcting for non-genetic differences, the researchers found that genes, on their own, continued to be a small yet informative predictor of opioid use disorder diagnosis. 

But participants’ education levels and mental health histories were more closely tied to opioid dependence risk, they found. Patients with less than a bachelor’s degree and a previous PTSD diagnosis had a higher risk of developing opioid dependence, regardless of their genes, according to their research. 

“Not only does genetic risk of opioid use disorder increase the risk of developing the disorder, as one would expect, but it interacts with other features, such as educational level and post-traumatic stress disorder,” said Henry Kranzler, a professor of psychiatry and the director of the Center for Studies of Addiction at the University of Pennsylvania’s Perelman School of Medicine, who co-founded the Yale-Penn collaboration. “It’s a gene-by-environment kind of interaction.” 

The finding that non-genetic risks to opioid use disorder can overwrite genetic predispositions is a new insight. A better understanding of risk factors for opioid dependence could create a future where physicians can better weigh the risks of prescribing opioids. 

As it stands, however, the model is far from accurate enough to be a clinically useful tool, Na said. 

Kristen Brennand, professor of psychiatry and genetics at the School of Medicine who was not involved in the research, suggested that the study’s data was not large enough to draw conclusions from. In the future, larger genetic studies with more target gene locations will be key, she said. 

Brennand also highlighted the limitations of the researchers’ risk assessment — a method called polygenic risk score. The scores can be a useful tool to understand how common gene variants are tied to the probability of disease, but they are not a rock-solid diagnostic, she said.

“It’s just a math problem — polygenic risk scores are mediocre at predicting disease,” Brennand said. “You would never use them in a fertility clinic. You could barely use it in an adolescent high-risk clinic.”

In the paper, the scientists caution against kits that claim to predict opioid dependence risk based on genetics alone. The FDA, for instance, approved a genetic test in December though researchers found that its analysis was inaccurate.

“The level of genetic prediction is so low at this point that we can’t really use it for identifying people at risk in the population,” said Dr. Kranzler. “But it’s only a matter of time — using things like the environmental factors we looked at — before you can come up with a profile of risk.” 

Other experts also pointed out what they said was a limited diversity of the study’s participants. The researchers only had enough data to look at individuals from European ancestry. It’s a shortcoming that led Sterling Elliott, a pharmacist and assistant professor of orthopedic surgery at the Feinberg School of Medicine at Northwestern University who specializes in post-surgery pain management, to question whether these findings were “generalizable enough to the American population.”

But Brennand agreed that further research on the links between genetics and substance use disorders, like this one, will be critical for doctors to better understand how addiction can pass through generations.

“So many people just don’t even think about the heritability of addiction,” Brennand said. “And what we actually know from genetic studies is not only is addiction heritable, but the type of drug of choice actually seems to be in the genetics too.” 

According to the Centers for Disease Control and Prevention, more than 1 million people have died since 1999 from a drug overdose.