Dr. Morisada is going to be speaking about pharmacogenetics and pain control. So, I'm going to be talking about pharmacogenetic testing and how it applies to pain and really focusing on opiates, opioids for the purpose of this talk. I drew this cartoon in here because the point is, it sounds like I'm talking to the choir a little bit. But by the end of this talk, or even in the middle of the talk, you'll quickly realize that there's a medically veritable explanation for what this patient's complaining about and this sort of smug answer is not appropriate. This is totally a legitimate complaint that is sort of being brushed off. And pharmacogenomics will sort of illustrate why. So, pharmacogenomics or pharmacogenetic testing has lots of these different names. It's really the pillar of this personalized medicine movement, which says that if we better understand the genetic makeup of our patients, that will allow us to select and dose drugs more effectively to get better efficacy without toxicity. Because how I respond to a drug is different than how you respond to a drug and that will vary based on disease state. So in the world of cancer now, it's completely indispensable. You can't think about any type of cancer or the frontier of any type of cancer therapy at this point without knowing about the pharmacogenomics of that cancer. So cancers are now treated based on their molecular phenotype. In fact, pathologists doing H&E staining are sort of out the door because that's not really how treatment's guided anymore. It's based on the molecular phenotype of the cancer. So in pain, you know, we have sort of a similar thing where we have to understand the genetics that go into taking the drug and the action of the drug at its receptor and then the pain relief or the toxicity that it produces. You know, fortunately, compared to oncologists, it's a little bit simpler because we don't have to worry about 10 different, or I should say hundreds of different genes with lots of different mutations. We have to know a few genes that are involved in the metabolism of the drugs we commonly prescribe or our patients are commonly taking anyway. So you know, this sort of very simplistic idea that you give an opioid and the patient gets pain relief is obviously not, you know, correct. Many times, the drugs that we give, hydrocodone, oxycodone, codeine, they're in a prodrug form that actually has to be activated in the body to an active drug. That active drug then has to actually avoid being inactivated by an inactivating enzyme. So there's this gauntlet of enzymes that the chemical you actually put in your mouth goes through before then it gets to its receptor. And then even at the receptor level, the interaction with the receptor and how it activates downstream signaling to have pain relief varies based on the genetic makeup of the patient and that particular receptor. And then you get pain relief. So these three, you know, molecules or the sort of concepts of these molecules are where you can have genetic variability between patients. And if you understand that about your patient or at least have a sense that that exists in the population, you might better understand sort of the cartoon in the beginning. So and this is sort of just to illustrate that we can't just keep thinking about it as this simple, you know, give an opiate, expect to get a benefit. And if you don't get a benefit, then it doesn't make sense. So now this, obviously, fatal adverse drug reactions being the fourth leading cause of death is not specific to opiates or any pain medicine. This is all adverse drug reactions. But this, I think, makes a lot of sense in our field, which is just a trial and error approach of let's try this, let's try this. There might be more effective ways of selecting drugs and dosing them if we have some sense of the genetic makeup. So this is a more complete, a little bit more complete picture from that cartoon. And essentially, in red and white, you see various opioids. And what you can see is from the sort of standard drugs that we may prescribe or not really prescribe anymore, there's lots of proliferation in terms of the metabolites in the body. So there's a lot more going on in the body than the drug you're giving. And a lot of that depends on these various enzymes and the way they act on those drugs. In particular, the cytochrome P450 enzymes in the liver, part of the microsomal chemical modification system of the liver, is largely what acts on opioids to metabolize them, both activating enzymes, in the case of this CYP2D6, or enzymes that tend to be inactivating, like the 3A4. So I'm just going to go through a couple of these just to kind of hammer home the importance of these enzymes. So morphine, which is active, and it's sort of native as a natural opioid, it's metabolized to a much more potent analgesic, morphine 6-glucuronide, by this UGT enzyme, which is analgesic, or to an inactive form, M3G, which some think is actually involved in opioid-induced hyperalgesia. So the genetic makeup of this UGT enzyme can determine the relative ratio you have of this M6G or M3G, and perhaps predict the risk of opioid-induced hyperalgesia. Oxycodone is metabolized by 2D6 into the much more active oxymorphone. It's inactivated by the 3A4 enzyme to noroxycodone, which has like 1% the activity of oxycodone, which itself doesn't have much activity compared to oxymorphone. Hydrocodone similarly is activated into Dilaudid by the 2D6 enzyme. So Dilaudid is the active form, hydrocodone's basically a much less potent form of the drug, and then it's inactivated to the NOR compound by 3A4 again. Tramadol also metabolized through this 2D6 enzyme to a much more potent form. And then finally, fentanyl inactivated through the 3A4 system. So that's just to make the point that the enzymes and the metabolites are important. So I've kind of loosely introduced these metabolizing enzymes. I'm going to focus a little bit on 2D6 right now, because it's one that comes up over and over again and has a fairly high population frequency of having variants in it. So these variants, the way we describe them is as gene polymorphism. So polymorphism in a gene is a variation in the DNA sequence that produces a different allele. So you get one allele from your dad, one from your mom. That's your complement for that particular gene. And if you get two, and depending on what allele combination you get, that will determine your phenotype. So in the case of these metabolizing enzymes, most of the population, although not by much for some of them, most of the population are what we call extensive metabolizers, where they have two normal copies of the gene. If you have one kind of deficient copy, you're an intermediate metabolizer. If you have two deficient copies, you're a poor metabolizer. If you have extra good copies, then you're an ultra-rapid metabolizer. So on the other end of the spectrum, where you can run into, for example, toxicity. So the prevalence of these variations is really going to determine whether or not this is of any significance, right? It's how common are these variations. So for 2D6, you know, I think 10% is like, it's starting to matter. You know, I would say it's not so much that it doesn't matter. And again, this is from, you know, a few papers. This is from a review. The exact population you look at, this is going to change a little bit. And especially if you're in a pain practice, you're probably looking at much higher rates of the variants, because there are people that aren't responding kind of the way that might have been expected. So if you look at this 10% poor metabolizers, and you break it down by ethnicity, you see Caucasians have relatively higher rates of being poor metabolizers compared to some of the other ethnic groups. So again, this goes with the genetic variability. So with 2D6, there's at least 80 identified alleles at this point. And the range of activity you get can be anywhere from 1% to 200% of the normal activity. And as far as we have in pain, you know, in comparison to the cancer studies I was telling you about, there's no prospective randomized data that would tell us how to use these pharmacogenetic tests. We don't have the effect of using these tests on the outcomes of our patients. We just don't have that. So what we have is these observational retrospective studies that in the end, I think, just sort of make the point that what we think should happen rationally based on the science I just presented is what happens. So here, this group studied 300 post-abdominal surgery patients. 30 of those patients were in the poor metabolizer group. So that's about 10%. They had a higher percentage of tramadol non-responders that required rescue medication in the poor metabolizer group, you know, about double what they had in their extensive metabolizer group. And as you can see, their poor metabolizer group had higher cumulative dosing of tramadol through their PCA. So, again, this is just observational study that is like empiric evidence of the genetic point I was making before. Here's 45 post-obstetrician section patients. They only had two in this group that were poor metabolizers. Both of them reported no analgesia with codeine. They had three ultra-metabolizers, and two of those reported immediate relief and actually stopped taking it from medication side effects. And then finally, here's a retrospective view, 230 female patients coming to the recovery room, various different types of surgery. And there's a, you know, high correlation between the phenotype of the patient in terms of their ability to metabolize opioids and the likelihood that they were going to have acute severe pain episodes. So then, just to switch gears for a second, OPRM1, so I talked about some metabolizing enzymes. This is now the mu-opioid receptor. This is the gene that encodes the mu-opioid receptor. It can have variable action based on a couple of different things that happen at the molecular level. So, generally, that includes polymorphisms, which are, you know, again, changes in the DNA of this gene. This specific polymorphism, where one nucleotide gets switched from an A to a G, is the one that's most common and, again, you know, becomes clinically relevant based on some of the studies I'm going to show you. So, again, for this one, the allele frequency is somewhere between 10 to 48 percent, depending on patient population. So, it's relevant. It's a variant that happens enough where it's relevant. Again, observational retrospective studies, nothing randomized at this point in pharmacogenetics and pain. So, this group looked at 207 patients in the palliative care setting and showed that those that had the GG phenotype, so two abnormal alleles, had basically double the dose of drugs than the normal genotype, and there's actually a nice sort of dose response, although not linear, curve with, you know, the dose of drug you need based on how many of the G alleles you have. And then, this is another one. This is looking at patients with GI adenocarcinoma and chemotherapy-induced neuropathy. Here's the prevalence they found. Here you see, in this population, about 10 percent were homozygous for the, you know, the bad gene. And again, they found that the AA genotype had much better analgesia than the genotypes that had either of the G alleles. So, how do you kind of use this information in practice? What are the tests you order? So, one way to derive this information indirectly is by doing urine drug toxicology, not just a screen, but an actual quantitative test. And another way is to do the formal pharmacogenetic testing with a buccal swab, which I can answer questions about related to the economics of that, but I'm not going to get into it. So, the urine drug toxicology, so this is sort of one example of a report, and you see a relatively normal patient. Here's the, you know, oxycodone level, and you can see it's being metabolized into the inactive metabolite and the more active metabolite at a reasonable rate. Here's a patient now that notes poor relief with oxycodone, and you see buildup of the oxycodone. It's being metabolized heavily into the inactive form by that other enzyme, the 2D6, which is supposed to metabolize oxymorphone, is probably deficient in this patient. So, this is a sign of a deficiency in that metabolizing enzyme. This is where I was saying there's indirect evidence where you can then confirm it with a pharmacogenetic test. Here's a patient who's a rapid metabolizer. You see their oxycodone levels are far lower. They're producing lots and lots of oxymorphone, so every drug, you know, every molecular drug that's getting into their system is rapidly converted to oxymorphone. So here's what the pharmacogenetic testing looks like, and obviously the exact interface is going to look a little different based on the company you use, but you can screen for, you know, various genes that are involved in the metabolism of different drugs. The ones of interest, as I've been talking about, are the 2D6 metabolizer and the mu opioid receptor. You see, for example, this patient has normal OPRM1 function, but is a rapid metabolizer. They have an extra allele. This one, normal metabolizer for 2D6, but altered OPRM1. They have the AGE mutation. And this one has both, poor metabolizer and altered OPRM1. So if this patient gets, you know, upset not getting opioid relief, like, you know why. It's not a mystery. So what are the barriers to widespread use of this kind of testing? Well, it's new technology. Physicians are unfamiliar with it, don't know how to use the test, don't know when to order it, don't know how to interpret it, don't know how to change their decision making based on it. Patients don't like DNA testing, although this is not the kind of classical thinking about DNA testing, where we're not looking for disease risk genes. We're just, this is what you have, and we're going to, you know, treat you according to what you have now. It's not some future risk DNA type test. There's lack of data, especially in the pain world for randomized, you know, lack of randomized trials to show any impact or what is the scale of that impact that can then guide, well, how often should I be doing this? And lack of guidelines, like I said, for how to use it. So one study design that we're sort of tentatively planning, and, you know, maybe if it gets off the ground I'll be able to tell you about another time, is to enroll patients from a, you know, fairly homogeneous, you know, let's say lumbar spine practice, because we're going to be able to treat acute pain in that setting, and do profiling of every patient in part one of the study, and just do some correlations to figure out what's the rate of, you know, getting those positive hits for those variants in this population, and then what's the correlation between those that have the variation and have, then, you know, poor pain outcomes postoperatively. And then also, importantly, with the aim in this study to determine if there's any preoperative clinical predictors to enrich that 10% population, because the more we can enrich the pretest probability and get, you know, more identified variants, the more economical this strategy becomes. So, because you're not having to do these tests, which are getting actually a lot cheaper on everybody just to identify the 10%. And then you can design a randomized study where you say, okay, we're going to do this preoperative screen, those patients that meet such and such, you know, survey questions will get testing, and then based on that testing we're going to randomize, and some of them will get, you know, postoperative pain management guided by the results, and some of them are just going to still get standard of care. And here's sort of like a, I wish this was us, this is a potential study outcome in a nursing home setting of chronically ill patients over 50 years of age with polypharmacy, so they used pharmacogenetic testing to try to maximize the efficacy of the various drugs that they were on and reduce polypharmacy. And you see in the people that got tested, this is rehospitalizations and ED visits, and I think both, and both with deaths, so, but the people that are getting tested are doing better than the people that weren't tested, so it makes an impact in this population. And then, you know, the take home, I sort of already alluded to this, but we want to end up with something like this, where you have screening questions, you know, these are pretty good screening questions, I think, as it is, you know, based on kind of what I've been reading about this, which is essentially if you just get a good history from the patient, fairly open-ended, do opioids work for you or do they not work for you? And also family history, because this is familial, obviously. And then you order these tests, whether they're, you know, if they're on opiates, the toxicology makes sense, it's cheap, or you can get the pharmacogenetic testing, and then based on, you know, these various things, that can guide your management. So, if they're 2D6 deficient, don't give them drugs that are metabolized by 2D6. They can get dilaudid, because it doesn't need the 2D6. And if they're OPR1 deficient, then you want to give them an agonist for the kappa receptor, if you're going to use an opioid at all. Anyway, so those are some of the considerations. I wanted to say thanks to the joint pain section. This is actually a work that's in collaboration with Bill Rosenberg and Andrea Trescott, who she is particularly interested in, devoted a lot of her energy into this field. And then Ashwin Viswanathan, who I also spent some time with.

pharmacogenetic testing

pain control

opioids

pharmacogenomics

personalized medicine