When doctors try to figure out whether a patient might become addicted to opioid painkillers, they rely on clinical risk factors like family history, medical history, and other social and environmental clues.
Now, two companies -- and Canterbury Healthcare's (IMT) -- want to add genetic information to that picture in order to improve risk prediction.
But academic geneticists say they may be getting ahead of the science.
Both tests are already being used in clinics around the U.S. – even though experts interviewed by MedPage Today say there are no data on exactly how well they predict risk.
"Any expert in the field would tell you that genetic vulnerability to addiction would involve dozens of SNPs," or genetic variants, said , an expert in addiction genetics at Rockefeller University in New York City. "The idea that anyone would say they are currently able to definitively evaluate an individual's genetic vulnerability to addiction testing by 1, 2, or 20 variants is quite frankly absurd."
The tests are allowed on the market because the FDA classifies them as "laboratory developed tests," a category that includes screens as simple as vitamin D assays. Most genetic tests fall into this category, and have enjoyed lax oversight and freedom in making marketing claims.
Getting the nation's opioid addiction epidemic under control is a top priority for U.S. public health officials as overdose deaths continue to climb -- with nearly 19,000 deaths related to prescription opioid painkillers and more than 10,000 deaths related to heroin in 2014 alone.
But the question remains as to whether the companies making these tests are advancing scientific knowledge rapidly, or if they are profiting off patients who may be put at risk.
The Genetics of Addiction
Addiction is estimated to be about 40% to 60% heritable, meaning genetics likely account for about half of a person's risk.
That means the other half of addiction risk comes from environmental, social, and cultural factors, like "being exposed to peer pressure, individual coping skills, chronic pain, depression, the properties of drugs themselves," said , of McMaster University in Canada, who studies the genetics of addiction.
Over the years, scientists have studied many likely candidates involved in genetic risk of opioid addiction, such as variants in genes that encode for the mu and kappa opioid receptors, for dopaminergic and serotonergic receptors, and for the enzyme catechol-O-methyltransferase (COMT), which helps break down neurotransmitters.
But the science is far from settled, Samaan said.
"We have some good ideas about the genes controlling opioid receptors, and some ideas about genes controlling impulsive behaviors, but these are not the only genes involved in addiction risk," she told MedPage Today.
Samaan added that there may be additional genes involved in addiction risk that haven't yet been discovered.
And therein lies the danger: patients who are reassured by a false low risk may not be as cautious about using opioids, and could end up getting addicted, she added.
Proove relies on a panel of 12 genetic variants, including those in opioid receptors, COMT, and serotonergic and dopaminergic receptors, to assess genetic risk.
In a phone interview, Proove Biosciences CEO Brian Meshkin told MedPage Today the panel was derived from reviewing the existing literature, and was validated in studies conducted by the company.
That genetic information is then combined with more traditional information to make a prediction about risk.
Much of that information is drawn from the , one of the most commonly used addiction screens, which was developed by pain specialist . Webster sits on the board of Proove.
But those screens have inherent flaws: "We're asking you to tell the truth about whether you're going to lie about your prescription," Meshkin said of devices like the ORT. "Genetic information allows us to pull objective information about someone."
Proove's test combines both types of information to get a weighted risk score that's designed to help a physician decide whether patients should be on opioids. Generally, 50% of patients are typically low risk, 45% have a moderate risk, and 5% are high risk and should not get opioids, Meshkin said.
"Instead of being restrictive for everyone when it's only a small percentage of the population, let's identify who those people are and protect them," Meshkin said.
The '' test is being deployed at several clinics around the country, including , the largest chain of pain medicine facilities in the mid-Atlantic region. It is also in use at the , and other pain management and orthopedic centers in the U.S.
The company has long experience in the pain therapeutics area, and offers other genetic risk assessments for pain.
Proove's competition, Canterbury Healthcare's Innovative Medical Testing subsidiary, would not disclose any information about its genetic test to MedPage Today.
Such testing fits into the company's broader goals, as Canterbury is focused on reducing costs in the Workers' Compensation system, and already offers drug testing to help payors determine whether patients are taking their drugs correctly or if they are using other illicit drugs.
Adding genetic testing would help payors determine whether a worker "already has a proclivity to becoming addicted to prescribed medications," according to a company press release.
Yet, the company doesn't disclose any information on the genes that the tests screens for, nor does it state if it combines the genetic information with environmental factors to assess risk.
Through a spokesperson, the company declined to answer questions from MedPage Today, and calls to Dean Clifton, whose phone number is listed as the main contact on the website, were not returned.
Defining Risk
The main criticism of these genetic tests is that there are no data on the tests' predictive value – at least not in the public domain, said , of Rockefeller University in New York.
"There are no data on how you can build a risk score by combining all of those [SNPs] together," he told MedPage Today. "And no one has shown mathematically how you can combine that genetic information with environmental information."
"It seems a little premature to use it for actual diagnostic purposes," he added.
, Proove Bioscience claims its test is 93% accurate. Full data on the sensitivity and specificity, as well as positive and negative predictive value, are confidential, Meshkin said. Those data were presented at the last year, although a search of PubMed could not turn up an abstract from that presentation.
Although the genes involved in the Proove panel are "reasonable genes," Butelman said, the problem with picking them from the literature is that various studies were done in different populations under various circumstances – and ethnicity plays a major role in genetic risk.
Another problem with gene variants for opioid risk has been replication, said , of the University of Vermont.
"Many of the positive genetic association reports were eventually not able to be replicated in other populations or within the same population but just by another research lab," Li said.
While there are good data on the genetics of opioid metabolism, that's easier than trying to quantify addiction risk, Samaan said.
"We are nowhere near having any diagnostic genetic test for a disease as complex as a psychiatric disorder," she told MedPage Today. "We don't have a single group of genes to say this is the culprit, that this is what's causing addiction."
In the Clinic
Some physicians who are deploying the genetic tests believe they are simultaneously helping patients and furthering science.
, who owns and operates a pain clinic and a surgical center in Roswell, Ga., in the Atlanta metropolitan area, recently gained an additional board certification in addiction medicine to better understand the challenges he's facing in what many are calling the state's "heroin triangle" due to a high prevalence of abuse inRoswell, Marietta, and Alpharetta.
He uses the Proove genetic test in order to assess addiction risk among patients.
"If your grandfather had an addiction, you may not know that," Kimes said. "That might have been kept secret."
The test involves a simple cheek swab at a clinic visit, along with the clinical information that can be elicited during an interview or from the chart review. Results come back within a week, he said.
He is also involved in Proove's clinical trials of the test, and plans to collect data on some 500 patients.
"Physicians are scientists, and we need to look at which data may benefit the patient and present it to them at least as an option," Kimes said.
But some researchers say those studies should be completed first, before the tests are made widely available for clinical use.
"First you need to do the study so you can identify the genetics, not the other way around," Samaan said.
Clinical trials run by academia or those that will be used for FDA approval first must pass the standards of an institutional review board, to make sure ethical and regulatory concerns are being met, Butelman said.
"The FDA should have strict guidance on these diagnostic tests because these are exploratory and might be overemphasizing an outcome that's just not there yet," Samaan said. "There should be better oversight."
And Now: The Regs
The FDA has acknowledged a need for better regulation of genetic tests that fall under its "laboratory developed test" (LDT) category, which was developed in 1976. Technological advances and newer business models have pushed the limits of that classification.
Surveys have indicated that there may be some 60,000 genetic tests on the market, accounting for about half of the entire LDT market.
In 2010, the agency it would develop a regulatory pathway, but that has not been finalized yet.
In 2014, FDA stopped 23andMe from marketing its genetic tests over concerns about interpreting the health data it offered -- but these were marketed directly to consumers, so tests that are managed by physicians may not be as much of a priority, Butelman said.
The FDA does regulate two types of genetic tests: those for that have a more definitive pattern of inheritance, and to tell how well patients will respond to certain drugs.
These have typically been approved via standard device pathways such as the 510(k) or premarket approval (PMA) pathways.
Yet for more genetically complicated conditions like heart disease and addiction, where multiple risk variants are suspected to be involved, there has been little regulatory oversight.
For instance, genetic testing company Celera, of Human Genome Project fame, had deployed its KIF6 heart disease risk test in clinics across the country. Primary care physicians were using the tests to make decisions about whether patients should be put on statins.
But a 2010 meta-analysis in the Journal of the American College of Cardiology found no relationship between the gene and risk of coronary artery disease.
The company subsequently applied for FDA premarket approval in 2011, but the test was rejected.
KIF6 testing, however, is .
FDA spokesperson Lyndsay Meyer said the agency has "long had regulatory authority over these tests, but we have been exercising enforcement discretion."
That means the agency is essentially hands-off unless the test is found to cause serious harm.
Meyer said the FDA acknowledges that there have been "increasing concerns about LDTs being marketed without clinical studies to support their use," and that it will likely deploy a risk-based framework for regulating these tests.
Under this system, there would be low-risk tests, LDTs for rare diseases, traditional LDTs, and LDTs for unmet needs.
The agency plans to issue a final guidance this year, Meyer said.
Ready or Not
Meshkin said his company's opioid risk test only requires at this point, but he's aware that FDA regulations may change and is preparing for that with ongoing studies.
"My hope would be if we have to go through FDA approval it won't slow down the pace of science and innovation. We're moving at such a fast pace now," Meshkin said. "When you control anything it tends to slow down its natural movement."
Academic researchers tend to see it differently: science should advance quickly, but not at the expense of patients, Samaan said.
"We have no commercial interests or patents," she told MedPage Today. "This is a premature use. It is not ready for prime time."