Recently, the son of Gavril Pasternak, MD, underwent surgery to repair a torn anterior cruciate ligament in his knee. Following the repair, he was prescribed Percocet (Endo), which contains oxycodone and acetaminophen (APAP), and which worked well to control the pain. But when he went to refill the prescription, he was given Vicodin (Abbott) a combination of hydrocodone and APAP. Vicodin is almost identical in chemical structure to Percocet. Both drugs are combinations of an opioid and APAP, with a difference of just two atoms.

When Vicodin didn’t work to control the postoperative pain, Dr. Pasternak’s son returned to the physician and got the original drug, Percocet, which worked “like a charm,” said Dr. Pasternak, who heads the Laboratory of Molecular Neuropharmacology and also serves on the neurology and palliative care services at Memorial Sloan-Kettering Cancer Center in New York City.

“That was a difference of only a simple hydroxyl group in the whole molecule,” he said. “These compounds were extraordinarily similar, and yet one worked and one didn’t.”

The problem is familiar to any practicing physician who prescribes opioids for pain. One drug works well for a patient, but another from the same class doesn’t; or a “standard dose” of morphine controls pain well in one patient, but not in another patient.

In animal studies, the difference in responsiveness to morphine between individuals can be as high as 10-fold.Although it is difficult to get an exact figure, Dr. Pasternak said that a relatively similar number would not be unlikely in humans.

“I don’t think that people are any less variable than the animals,” he added. “Probably, we are going to have a broader range of sensitivities, so if you have a 10-fold increase in animals, it would not surprise me if we had a 10-fold [increase] in humans.”

A study that looked at 3,200 patients recovering from back surgery found that some patients required as much as 40 times more morphine than patients at the lowest end of the range.

The question is not whether opioids are effective—a high enough dose of opioids will overcome any pain, with consequent and proportional side effects.

“The issue is not efficacy,” said Charles Inturrisi, PhD, a professor of pharmacology at Weill Cornell Medical College in New York City. “We know opioids as a function of dose can relieve pain regardless of the source; rather, it is opioid responsiveness that exhibits wide variability. There is substantial variation in opioid responsiveness in pain patients.”

One of the unique features of opioid therapy is that, unlike most drugs, clinical experience far outweighs the understanding of how these drugs actually work. As an analgesic, opioid use dates back thousands of years.

“We don’t really appreciate the subtleties of these drug actions,” said Dr. Pasternak. “The pharmacology of the opiates is upside down. We are still using the same drugs we were using 100 years ago in some cases … we know more in the clinical setting than we know in the preclinical.”

A Boom Phase or a “Lot of Hype”?

Essentially working backward, investigators are beginning to understand some of the factors that contribute to how patients respond to opioids, and it is becoming increasingly clear that genetic variations may play a significant role.

Jeffrey Mogil, PhD, heads the Pain Genetics Lab at McGill University in Montreal, Canada, and after 20 years of looking specifically at genetics and pain, probably knows as much or more about the subject than anyone. “At the start of my career, people used to deny that there was any variability at all,” he said. “Now of course, everyone not only admits there’s variability, but actually thinks that variability is a problem.”

Of course, not all variability is genetic. Responsiveness to analgesics can be shaped by the source of pain and a patient’s psychological propensity to suffer, as much as by genetically based differences in drug metabolism. Nevertheless, genetic differences probably play a relatively significant role, according to Dr. Mogil. “I believe fully that the answer to a good portion of the variability we see is genetic,” he said.

Some of these differences have been well established. For example, differences in how the CYP2D6 enzyme is expressed can significantly affect how a patient metabolizes certain opioids. Thus, drugs that require bioactivation to increase their potency—like codeine, for example—can be rendered relatively useless if a person is a poor metabolizer.

At least four other genetic variations have been found to significantly alter the effectiveness of opioids.

One example is mutations that affect P-glycoprotein, a protein that carries molecules, including opioids, across cell membranes. When the enzyme is overexpressed, drugs may be under- absorbed and never reach therapeutic concentrations; conversely, if the enzyme is underexpressed, drug plasma levels could reach toxic levels.

Another is the COMT gene, which makes enzymes that degrade neurotransmitters like dopamine, epinephrine and norepinephrine and has been shown to affect μ-opioid receptor binding.

These mutations were discovered using gene association studies, which puts the field of pain and opioid genetics about where genetic research in psychiatry was 15 years ago—in a boom phase.

“There is a lot of hype, a lot of people doing these studies and a lot of expectation that these genes are going to pan out,” said Dr. Mogil.

But like gene association studies in other fields, early, promising results in pain genetics haven’t necessarily borne fruit over the longer term.

“One person finds an association, publishes in a big [journal] and then over the next few years, papers start trickling out that say ‘well, we tried it too and didn’t find that association,’” Dr. Mogil said.

Genome-wide association studies, which are still prohibitively expensive, would solve the replicability problem, but they do not explain much of the variance seen between individual patients.

One other major advantage with a genome-wide study is the possibility of discovering new receptors involved in pain and opioid metabolism, because it has become increasingly clear that the biology of opioid analgesia is far more complicated than researchers previously have thought.

The Tip of the Genetic Iceberg

The μ-opioid receptor gene is a case in point. It was not only “biologically reasonable,” as one study put it, that polymorphisms of the gene would significantly affect opioid response, but it was also well established in both experimental and clinical studies.

Jörn Lötsch, Dr.med, a professor at the Institut für Klinische Pharmakologie in Frankfurt, Germany, is one of the few physicians in the world to study these genetic variations at the clinical level. In a recent study, Dr. Lötsch genotyped more than 350 patients in several outpatient pain centers; he then followed their opioid dosing regimens and pain scores. Although not significant, patients with variants of the mu receptor gene showed different pain scores and opioid dosages.

These early findings, in part, prompted Dr. Lötsch to do a meta-analysis of the available studies. The results were less than promising. “We found no consistent association between OPRM1 118A>G genotypes and most of the phenotypes in a heterogeneous set of eight clinical studies,” he wrote at the time.

When asked which area of genetic research holds the most promise to change how clinicians treat pain, Dr. Lötsch said that a year ago, he would have said the μ-opioid receptor gene. “Today,” he said, “I have to answer, unfortunately, currently none.”

“There is currently no prospective study that has shown that a genetic-guided pain therapy provided a real benefit for the patients,” he added.

Genetic variation does likely play a critical role in the large differences in patients’ pain and their opioid needs; it’s just that the μ-opioid receptor gene—or any of the others that have already been identified—aren’t responsible for all of it.

“I’m afraid we’re not as close to figuring this out as we thought we were, probably because there are a lot more genes involved than we expected. We were being incredibly naïve to think that the genetic variability of morphine analgesia would all boil down to the mu receptor gene,” Dr. Mogil said. “That would be a nice, tidy little biological universe, but that’s not the way biology works.

“We are at the tip of the iceberg of finding relevant genes. We’ve found a tiny handful of them. But we have reason to believe that there are 20 or 50 or 180 more to find,” Dr. Mogil said. “And only when we have found a much larger number and have convinced ourselves that we have a reasonable proportion of the variance explained, then we can start looking at the big picture.”

The clearest clinical application of this emerging genetic knowledge would be to personalize a patient’s pain treatments based on his or her genetic makeup. “The ideal would be basing a therapy plan—dose and choice of analgesic—on genotyping information,” Dr. Lötsch said. “This works in cancer therapy and for warfarin therapy, but not yet for pain.”

For example, a patient might be genotyped before an operation and the postoperative pain interventions might be based in part on the sum total of the pain-related genetic variations.

Further down the road, a better understanding of the genetics of pain and opioid metabolism could move into other areas, like predicting a patient’s risk for developing painful disease, the clinical course it might run or even a patient’s risk for developing drug addiction.

“I’m simultaneously optimistic and pessimistic,” said Dr. Mogil. “I’m still a believer that in theory, given enough time, we’ll be able to get this information.”

But even if genetic research doesn’t immediately translate into genotyping in the clinic, the explanation of why some patients respond differently to analgesics may still be clinically important.

“The undertreatment of pain, to some degree, is due to the fact that people aren’t educated,” said Dr. Pasternak. “What happens is that somebody will say, ‘Well, a dose of morphine is 15 mg and this person is asking for more, obviously they’re abusing. They’re not using it right, they’re not responding the way they are supposed to.’”

The genetics themselves aren’t going to give us all the answers, he added, “but what they do is enable us to tell people that are skeptical [of variability], this is why.”