I'm Nitin Gokde, I'm Chief of Research with the American Psychiatric Association and a Deputy Medical Director. It's my honor to welcome and introduce Dr. Mueller for this Marazek Award lecture this morning. Dr. Daniel Mueller is a senior scientist with molecular science and head of pharmacogenetics research clinic at the Center of Addiction and Mental Health, CAMH, and a full professor in the Department of Psychiatry at the University of Toronto. His overarching goal in his research is to improve drug treatment of psychiatric disorders. He feels pharmacogenetics holds the promise to identify gene variants that are associated with response and side effects. Once validated, this approach, he believes, will allow for precision medicine, avoiding long trial and error strategies, before the right drug for the right patient is identified. A particular focus of Dr. Mueller's research is to target genetic markers that predict medication side effects, such as antipsychotic-induced weight gain. Dr. Mueller's research has revealed significant associations between antipsychotic weight gain and the cannabinoid 1-receptor gene, the melanocortin 4-receptor or MC4R gene, the dopamine D2 receptor gene, and the neuropeptide Y gene. Dr. Mueller's research group is developing an algorithm that will incorporate these genes, along with clinical and dopaminergic, clinical and demographic risk factors, in order to develop a genetic risk model for antipsychotic-induced weight gain for clinical application. Dr. Mueller is a member of Clinical Pharmacogenetics Implementation Consortium, which is developing guidelines to use genetic information for psychiatric drug treatment. On behalf of APA, it's my honor and privilege to welcome Dr. Mueller. Dr. Mueller, the podium, please. Hello, good morning, everyone. I'm trying to go to the distance of the microphone. If it's not well heard in the back, please let me know. A bit of echoing, maybe, but hopefully not too bad. So thank you for this nice introduction, Nitin, and I'm very honored to be here, and I'm very honored to receive this award prize to celebrate and, you know, in memory of Dr. David Mrazek. Here are my disclosures at this point in time. I would then like to come back to David Mrazek, and if you might have noticed, today is exactly his 11th anniversary of his death. It was a tragic death. David Mrazek was a pioneer in his field. He worked at the Mayo Clinic in Rochester, Minnesota, and was really the first, I think, for Ward-Wild to broaden, you know, pharmacogenetic testing to the clinical practice. My first encounter with David was at a course in Rochester that he gave every year, an annual course. So that was, we're now going back to 2008, and it was a wonderful four or five day course, and you had, you know, these brochures about what you need to know about 2d6 and 2c19, and what the test would tell you. Very, again, pioneer work, and very clear, and I would say logical and compelling to anyone attending that, you know, it makes sense to know your CYP and, you know, the CYP genotypes metabolizer status if you operate with medications which are metabolized by those cytochromes. However, I was also, as an academic, you know, the usual skepticism and whatnot, but I was still new and, you know, I'm not a native speaker, so I kind of, I think, asked him at the end of the course in a clumsy way. I said, well, this is all very interesting, but, you know, how do you deal with the criticism that might be around you? Because, you know, not everyone might agree that this is ready for primetime. And David replied to me famously, if you don't like criticism, you better play golf. And I think that was him, very much, and I liked that the audience laughed, and, you know, I felt a bit silly, but, you know, the other thing was, yeah, if you're not, you know, if you're not willing to try where things might go, you know, if you are pioneering, you know, if you're too hesitant, you will also probably not do it better. So I really liked that interaction, and since then, after that, we have met at other places, I have read his book, and at one occasion, I said, hey, David, could you sign it, maybe? And he made me this nice note, dear Daniel, I wrote this book for our patients, which I also like very much, because I think he was about to say, this is not about talking just about populations, you know, study populations, this is really talking about people, talking about, you know, what if someone is, you know, such and such metabolizer has such and such gene variant, what are we going to do about that? And how might the patient feel about that? And how would it affect his outcome, his or her outcome? So that's, I think, what he intended here. So therefore, we're going to, you know, today, also in memory of David Razek, we're going to talk a bit about, you know, the updates since that book was published, if you wish, but also what the genetic testing in psychiatry would currently look like. Here are my learning objectives, one, two, and three. Basically, we're going to talk a little bit of the concept of personalized medicine. I'm sure many of you are familiar, but, you know, there are some, maybe some small aspects here and there that might be of interest to keep in mind. What are the currently, you know, evidence, what are the current evidence? And how do we measure evidence? And how are we, you know, how are we navigating through the complex data sets that are out there? And then, however, you know, what do expert panels say? What do they recommend? And we're seeing stands. And ultimately, well, a little bit about, you know, other, I would say, important aspects of the application of everything. So I'd like to start off with this quote from a famous Canadian doctor, actually, Sir William Osler, who then ended up, I think, working at Johns Hopkins. But what he said was, you know, if this were not for the great variability among individuals, medicine might be a science and not an art. And of course, I believe that Dr. Osler would have said medicine is an art, but it's a bit, you know, it's a bit challenging. It's a bit playing the advocatus diaboli, the devil's advocate, to say that, you know, we still see with same symptoms and the same treatment variations in outcomes. And so the physician remains basically often on his own or on their own and will need to find creative ways, right, to maneuver through things. And this was way before most of the medications that we currently know have been ever discovered. So I would say this was probably the first call for precision medicine, maybe, that, you know, we can identify. And since then, a lot of good things have happened. But as we will see, you know, they take a lot of mechanisms that needs to work, you know, nicely into each other. And I have depicted the most important here on this next slide. So this is, if you wish, you can call it a road map. And, well, again, there are eight different, you know, eight different, if you wish, slices of the pie that all need to be addressed before you implement pharmacogenetics. So it's not just about, have we done enough research, or have we had enough experts talking about that, or what are the regulatory agency talking or saying? You know, the question is what genetic, how genetic companies will also operate, how insurance companies or reimbursement will occur, then the question how to bring this into the electronical health records, and if, you know, how to train and educate not just healthcare providers, of course, but also the public, and so on. And ultimately, you know, what is the ethics about all that? So it is a long way, and I will talk about all of these eight slices in that presentation in the time that I have been given. And so let's go. First of all, however, you know, variability in medication is a challenge, but it's not like we have no clue at all when we do see our patients in our offices. I think, you know, we make a judgment call, first of all, of course, about the clinical symptom profile that we're seeing, and then we might make a decision about the pharmacodynamic component of the medication. For example, an agitated depression will probably get a sedative, antidepressant, or, you know, someone with lack of motivation might get bupropion, for example, right? So then we look at other things, of course. We look at age, we look at sex or gender, we look at ancestry, we look at, hopefully, we look at drug-drug interactions, because there can be, of course, detrimental if we don't do that. And then we have a toolbox here now, you know, that we would like to bring in. And you can imagine that this is a bit of a disruptive process now, you know, because everything which I just mentioned happened basically between the physician and the patient in the office, and it does not require any external connection, right? Now, if the patient doesn't have a genetic test, for example, you need to send them somewhere and get the test result, you know, and so on. And here, people say, well, this causes maybe delay, you know, this could potentially make life difficult for implementation, because now you have this delay. Normally, you would not, you would, without this toolbox, whatever it is, could be imaging, could be EEG, you know what I mean? Could be anything. In our case here, yeah, okay, it's drug-gene interactions. But to keep that in mind, that this is also one of the challenges in implementation, is to change practice. And change practice is something which always has, often enough, at least some disruptiveness component. And again, some potential delay. Now, we can argue, a test can be done in a couple of days, you know, the delays would not be immense, the delays can also be counteracted if you choose a drug where you think, you know, there's no genetic or little genetic risk, as opposed to other medications, you can play around with that. But that's just a principle, I think, we should keep in mind. Okay, then, as you all will know, probably from pharmacological perspectives, we differentiate between two compartments in our body, maybe three. The one of them is pharmacokinetics, or what the body does to the drug, metabolizing, excretion, and so on. The other one is pharmacodynamics, what is the drug doing to the body, and in our case, typically, the target organ would be the brain. And then we have these other factors as well, of course, you know, they're not really compartments, but they also interplay with the dynamics of everything here. And this could be lifestyle, diet, you know, potentially the gut microbiome could play a role, and whatnot. So it is complex, we are not claiming that one gene, or, I mean, we as pharmacogenic researchers, if I can say so, we wouldn't claim that one gene will now bring us all the answers, how naive would that be? But however, it is probably also, at this point of time, appropriate to consider that information when it's available. Similarly, like, you know, age, sex, gender, and whatnot. And that's, again, however, just a reminder how things are complex. So, because David Rezek was already into pharmacokinetics, and nowadays we're still seeing the best data, and the most data, probably, in pharmacokinetics, I will actually continue mostly now to speak about pharmacokinetics. And let's take a look at those enzymes. Those are the cytochromes, which are mostly involved in the degradation of psychiatric medication. And as you can see, there may be four major players, CYP2B6 is a smaller player, but CYP2C19, 1A2, 3A4, 2D6, are the ones which are frequently involved in drug metabolism of psychiatric medication. Now, two of them are highly genetically variable, and two of them are less genetically variable. Those who are high genetically variable are CYP2C19 and CYP2D6. CYP2D6 is extremely genetically variable, most, you know, one of the genes which is probably one of the most variable genes in our genome, you know, in our population, because it can either not exist, or basically be completely missing in terms of activity, or it can be duplicated. So, people have it multiple times on the genome, and that's when they get fast metabolizers. And that kind of variation from zero to superstar, super fast, you know, rising, or super fast metabolizer, that's very unique. We don't see that much of variation. Typically, we just talk about gene variants within the gene, around the gene, but here, again, we also have copies, and we also have people who have basically no functioning CYP2D6. Anyway, the other thing maybe to get, to keep in mind is 1A2 and 3A4, while they are genetically less variable, so they're not, they're not really, you know, recommended very much, but they are, of course, also inducible or inhibitable, depending on which, which you were talking about. So, these are the classic genes that would, or the classic enzymes that would matter for drug-drug interaction, or, for example, for grapefruit juice, 3A4 inhibitor, St. John's wort, 3A4 inducer, and those kind of things. So, but the 2D6 and 2C19 are actually not as much affected by food or ingredients or supplements as much as those two others. Anyway, if we look now at the distribution, how many of these CYP2D6 do we have in our liver? Interestingly, the CYP2D6, even though it's so important, is only available in a small amount, and I think now we have a bit of a problem here with 2D6. Anytime medication is metabolized by 2D6 and the affinity is high, however, the capacity is low and the genetic variability is high, you can imagine that this is a sensitive spot anytime that CYP2D6 medication has been prescribed. And if here, you know, something, something which is already sent, you know, already, let's say, critical, if you, if you then have phenoconversion, or if you have genetic variation, that will impact on this, on your metabolic activity, that can cause issues. And here, this is an older slide from 2006, just to also, you know, this is a time when Dr. Mrazek would have, would have, you know, designed his, his, his pharmacogenetics in Mayo Clinic. You know, here you can nicely see one thing, which is paroxetine. Paroxetine is highly metabolized, primarily metabolized by 2D6, and now if you give people different doses, 10, 20, 30, or 40 milligrams, you will see different plasma levels. And now the question is, why do we see those different plasma levels? And the answer is very easy. Yes, because it depends on the metabolizer status. So, the ultra-rapid does not build up any metabolize, any, any serum level, really, of paroxetine, whereas the poor metabolizer probably reaches now at 40 milligrams doses that are quite toxic, potentially. And now, if you don't know about pharmacogenetics, again, you see this variation, and, you know, let's say we're back at the times of Dr. Osler, you, you know, you may wonder, have, have some of the people overdosed or some of the people non-compliant, you know, you just would not know. With the genetic test, it gives an idea what's going on, in fact, and so this would be, you know, we could now go back to Dr. Osler and say, this is the explanation why, you know, people show so different, different serum levels. And at 40 milligrams, you can imagine that this will matter now, you know, on the experience of the patients and also on the outcomes. And I think, I think one thing is also important here to mention, maybe, that it looks very simple now to say, well, then I adjust the dose and things should be okay, right? As long as I stop early enough for the poor metabolizer, should be fine. As long as I keep titrating up the rapid metabolizer until I see an effect, things should be good, right? However, you, you also have, of course, to do with the metabolites. And a rapid metabolite will, will create metabolites down the road that might cause side effects or that might cause, you know, non-response or potentially, potentially other things. So we, this is very well, very much understudied, what really happens and what might be the, you know, what might need to be the best ratio. In other words, most medication trials are done for the normal metabolizer, and because the majority are normal metabolizers, then a medication might get approved, but ultimately those extreme metabolizers might not be considered, might be left out. So now we have genetic tools, and I'm not going to speak about in detail how we do that, how we are analyzing genetic variants, and, you know, and then to combine them, and we build haplotypes, and then we build star leads. This is all, I think, not a matter of big interest here at this point of time, but all you need to keep in mind now that we have the genetic tools available to very nicely screen the most common variants, or we can even do sequencing these days and get all the variations done. So these are things which are easily possible, easily doable, and which do not even need to cost much money to do. Okay, now let's look at a few of the studies, and I would like to show you some retrospective analysis, and the one nice thing about retrospective analysis is that people who are genotyped here had no clue about their metabolizer status at time when they took the medication, and the doctors neither. Only us, after the facts, we came in, we as researchers, again, these are not all my studies now, but, you know, you can imagine that if there is something like a placebo effect, you would not see any effects here if there was only placebo effect going on or no placebo effects because, again, people were not aware. So let's look at one study, a small study that we did, but the concept was interesting. The concept was such that we looked at a patient population treated by our colleague, Dr. Richter. She runs an OCD clinic in Toronto, and she has about 300 patients, although that she's monitoring on a follow-up basis, and ultimately we said, okay, why don't we look at their genotypes and see how many were poor and rapid metabolizers and whatnot, and how often they needed medication trials, what's the number of medication trials they needed to undergo until they finally found the right medication, so to speak, right? And so we saw that the numbers of, you know, super fast and super slow metabolizers were actually quite low in this population. Most people, the majority was normal metabolizers, but those who were extreme metabolizers underwent more clinical trials. On average, they needed to go through 4.2 medical trials as opposed to 2.7 for those who are normal or intermediate. And that was statistically significant. Now what does it mean? It means that people were probably in the past exposed to medication which was primarily metabolized by 2D6, and they have not been doing well, and that explains, most likely, the higher number of trials. And again, the fact that these two bars on the left and on the right, you know, this is an interesting concept here, that basically, you know, we probably need to better think of them as extreme metabolizers or atypical metabolizers, if you wish, and those are the ones who are probably most affected and vulnerable for medication, you know, non-response, potentially side effects too, because they are, you know, in one of the two categories, and that also helps us from a statistical point of view sometimes to merge the groups, because here we had very small numbers, right? So but again, let's go and take a look at this study, more recent. This is a nice Danish long, you know, cohort study where, again, people were genotyped only after the facts, only after they visited emergency rooms, and only after they've been prescribed medications like sertraline and acetalopram, but later on, you know, through this Danish cohort study, they could see that, ah, you know, CYP2C19 poor metabolizers who were on acetalopram were at an increased risk of switching medication. Obviously, they have not been doing so well. They were even at increased risk for suicide attempt or self-harm. In addition, young adults with CYP2C19 poor metabolizers using sertraline were also at risk of switching, were also at higher risk of switching, and ultimately, CYP26 poor metabolizers who used fluoxetine, primarily metabolized with fluoxetine, had an increased risk of what, you know, we normally wouldn't measure so easily, but the Danish have those data readily available, emergency department contacts, and, you know, if you see a population coming up more often to the emergency room, you can bet there's something going on there that might be the reason, and, again, the metabolizer status could have been, again, one of the explanations. Perhaps they have had more side effects, and they run back to the emergency room, or they have not responded. We don't know exactly the differentiation about these two, but we know that they came more often to the emergency room now, and that was very interesting. So, and then we have another retrospective analysis, and I like it because here you can see one thing. This, again, tells you the risperidone levels on the X-axis, and you can see that people who were fast metabolizers, ultra-rapid metabolizers, received more risperidone, and those who were poor metabolizers received less risperidone. But the doctors did not know about the genotypes or the metabolizer status, which means they have compensated for it. They have done what we call, right, go slow and low. They have started to titrate the patient, you know, up to the first reactions, up to the first symptoms improvement, and if needed, they went higher. And so, basically, another way, nice way of saying it, you know, they could have known the genotypes before and could have done probably, you know, faster titration with ultra-rapid metabolizers maybe, and so on, and could maybe want the poor metabolizer, you know, to, you know, about things, but here, again, interestingly, they were just doing it without that. And then, however, coming back to this, you know, U-shaped curve, if you wish, again, those who were super fast or super slow metabolizer were the one who were more likely to be switched. So despite this compensation, again, despite this compensation of adjusting the medication dose, we've seen more failures, and I don't know why that was the case, to be honest. It could be, again, with the metabolite to compound ratio, it could have other things, it could be, you know, but one likely thing is that, yeah, that something is there that does not favor extreme metabolizers. And again, the numbers are smaller for, you know, in this large cohort, only 13 were ultra-rapid metabolizers, only 32 or, sorry, 23 of them were poor metabolizers, so these are also low numbers. Yeah, I would take the credit that, you know, this might be just a false result, but it's consistent with other studies, and there are more studies than that, but I will not show them all, but they all show the same pattern. And interestingly, two things, dose adjustment didn't work, and second, you know, if there was any placebo here, it is impossible that the placebo has influenced placebo response because no one knew about them, right? So that speaks definitely against the notion that PGX testing and, you know, your patients will do better because placebo, hooray. That's unlikely, at least in many cases. There are other cases where it might happen, but not in those studies that have been done on a retrospective way. And the same data, basically, the same pattern was seen for arepiprazole, again, for the sake of time, we will not show this, but you can read the study if you like. So, and now, because I talked about low numbers, I want to say some cautionary words about that, low numbers in rapid and poor metabolizers. Now, if we look at different populations in the world, and this is one way how anthropologists, you know, determine the major, you know, ancestry groups before large migrations hit 15,000 or 13,000 years ago, but these populations have stayed a little bit isolated on their own for some time, and now they have gene variations that are unique to these populations. That's what it means. And so, if we pick now a sample of Europeans, okay, here are the frequencies on average. You know, this is just a randomly picked sample of healthy individuals, and we see, okay, it's 2-3% fast metabolizers, 3-5% poor metabolizers. So, these are the people, these 8% maybe, you know, that we're really worried about, but 8% of the people, you know, that's for the white people still, but it's a lot if you wish, you know, it depends if you see them in your practice, you know, you may want to know, you know, they might want to know as well. But in reality, things are, of course, always a bit more complex, because even in Europe, you can see gradients, whereas these slow metabolizers are rather in the northwest, so in, you know, British Islands and Faroe Islands, a small archipelago northeast, I believe, of Britain has 15% poor metabolizers, it was 26, so five times more than, you know, a white person or so in Middle Europe. And then you have fast metabolizers predominantly in the southeast, and if you go to the Middle East, as you can see in the next here, in the Middle East, in Qatar, there is already 10% of fast metabolizers, so 10%, so 10% means every 10th person from Qatar or who has, you know, ancestry of Middle East will not respond to a standard dose of paroxetine, as we saw earlier, so that's a lot, that's much more, so it really depends what population you're looking at, if you judge, you know, frequencies. And so CYP2C19, we see a striking difference for poor metabolizers, three to five percent in Europe, maybe in 17 or more in East Asians and Oceania, so these are the islands between America and Asia, right? So here, the next slide is giving you a little bit colorful picture, and you can see the diversity very nicely. Here this is just an attempt, based on the data that we had available a few years ago, to make it colorful and to just show that no pies look the same, right? If you look at different populations, you know, and again, look at the Oceanian, how often there are poor metabolizers of 2C19 compared to the rest of the world, it's striking, right? And again, Europeans seem to have the highest rate for 2D6 poor metabolizers, but this is all preliminary. To be honest, there is not just enough studies out there that have really, you know, tested things to the ground, but basically, that currently shows you that it all depends also on which population you're looking at, and the gene variants might also still need to be clarified, might need to be also further studied that you need for other populations. So our impact study, which was directed by Jim Kennedy, who is here in the audience, that when we started to do large-scale genotyping in our patient populations, and we accepted referrals from our large mental health hospital, Center for Addiction and Mental Health, also named CAMH, and we accepted also referrals from community psychiatrists and from primary care physicians. And over the years, we got about 12,000 patients enrolled, and then if we took a look at the frequencies of metabolizers, we saw something very interesting now, which was we saw much higher prevalence of, again, of normal metabolizers, we can say, as opposed to what we would expect. Okay, so you can see the numbers here. If you compare them, we would expect 80% or so would probably be normal metabolizers for 2D6, and these are the Europeans, so we don't have any biases with ancestry, all right? But in these 1,000, 1,200 or so Europeans, we saw only 50%, 56% were normal metabolizers, whereas 15% were poor and 6% were ultra-rapid, so that's 21%. So we see that much higher, much higher percentage, and how is it, why is it happening? Why has it been such? Well, I mean, as tertiary care center, we get referrals of people who are difficult to treat, who might have, you know, been labeled with treatment-resistant depression or who have, you know, complained about side effects so often that they get ultimately referred to see a specialized center. So here we saw that in a specialized center like this, we have probably an enrichment of people who are poor and rapid metabolizers, so once again, the low numbers in the general population should not distract us that our clinical populations can show different numbers. And just a recent study, this was just published this month, a collaboration I'm doing with University Clinic of Würzburg in Germany, they have also, well, they have, first of all, they have a very nice TDM lab, they have always been titrating patients up and down by TDM until a few years ago, they started also to use genetics and now they have a genetic and TDM lab, and I think the combination of both is really a great thing to have and it's really a great and sensitive and reasonable thing to do to your patients, you know, to measure TDMs, you know, plus pharmacogenetics. And so what they saw also in their population, these are inpatients for depression, so inpatients, again, you need to be severely depressed, you have failed, you might go to ECT potentially and so on, and they found that, you know, only 20% were known metabolizers for both enzymes, 2D6 and 2C19, so the vast majority had at least a genetic variable or a genetic variant which was actionable as we define it. So again, inpatient populations, more severely ill, shows us, again, different numbers and shows us a bit, you know, that in this case here, what they're doing is probably makes sense to disentangle now by TDM and by pharmacogenetics what they can do and how they can best, you know, restart treatment, but this time aligning it to the genetic profile. And these were just 154 patients, again, I'm aware of small number biases and so on, but again, we see the pattern, right? And now, if you want really more patients, you know, this is another study that is just in press and I was invited kindly by this group of temples to join their study as scientific advisor and I have a slide here for you, again, these are 15,000 patients that they have genotyped, that they have included, but there were 4,000 or something patients which had data available on medications also. And now the question was, when are things actionable, even if you are just taking into account phenoconversion, right, including, well, you look at the genes, you look at phenoconversion and you look at a novel variant that is also being now debated if that might also play a role in 2C19, we call it CYP2CTG, so if you would accept that this CYP2CTG is also a new variant on the block that might predispose normal metabolizers to become rapid or ultra-rapid metabolizers, you know, ultimately, they found in these 4,000 something people, 87% of them would have some way, you know, some of the components that would affect medication metabolism, so that's a lot, yes, that's a high number, again, if you just look at the genotypes, 66%, so this matches a little bit what was found in Germany, but here you also looked at phenoconversion and, again, the 2C19, 2CTG haplotype around 10%, and then there was also the notion that CYP2D6 is also complex sometimes to understand if people are ultra-rapid metabolizer, which of the alleles is actually duplicated, because if you are duplicated, if you have a duplication of an allele that is actually not really, you know, that is already, let's say, impaired in terms of enzymatic activity, right, which then, if that is the one that is duplicated, you know, you might not become a super-fast metabolizer, but maybe just a fast metabolizer, so they also showed here that, you know, this is important to use normal next-generation sequencing to really make sure that you have those, you know, those rapid metabolizer status. All right, so, but now, for the evidence, we need to look a bit further, and I know there's RCTs and meta-analysis that people would like to see at some point after, you know, talking a little bit about the general principles and the general interesting observations, right, so let's pick one RCT that we have been involved with in Toronto, and where the data has been extrapolated with data from a very similar trial in the U.S., which was called the GUIDED trial, and so, interestingly, when you extrapolate these two studies together, you actually see a very similar pattern, and that was, you know, at baseline, oops, excuse me, here at baseline, when you looked at patients who are coming to your study now, and they have not been doing well at least on one antidepressant, and now you do an analysis and say, okay, how many were actually already taking a medication that was aligned to the genotype, you know, congruent to the genotype, if you wish, this was around 80%. So basically because, again, not all the medications are metabolized by 2D6 and some of them might have gotten the right medications by chance or something, 80% was already a good enough number. But now when you start to say, okay, now we're going to give some patients pharmacogenetic testing and some of them will get the standard care that they usually get, you know, called treatment as usual, TAU, then you see that, interestingly, after a few weeks, there's a bit of a split now where there's about now 5 to 10% of people who will now get medications for the first time probably aligned to the genotypes. And after eight weeks, it's about 90%. So you increase it about 10% in this study, in both studies. And however, the question now, does it matter? Does it matter that, you know, you now have 10% more aligned to the genotypes, whereas the treatment as usual stayed on medications most of the time, which were not ideal, but of course, with some exceptions, I'm sure, but, you know, does it matter? And yes, it did matter for remission. It did matter in the sense that around 10% or so in the treatment as usual are improved or remitted, whereas 16% in that pharmacogenetic tested arm improved. So 5 to 6% of, if you wish, patients who are now remitted and would not have remitted maybe if there was no genetic testing guiding and helping it. Now, this was not a super blinded study, which I'll talk about blinding a bit later. Again, it is a study that was otherwise, I think, you know, well-powered in terms of people and outcomes so that we can, I think, say that there's a good chance here that pharmacogenetics testing has really made up for the difference. Then we also see some meta-analysis. We have seen now 5 or 6 in the last couple of years. They all show very similar results. And I mean, if we want to criticize biases because, you know, those studies are heterogeneous, I'm not going to say you better play golf or so, but I would say, you know, most studies show an effect towards superiority of testing. This can also unlikely be just or randomly for, you know, type 1 or false positives. We can leave it there, but I'm saying, because again, I don't want to argue statistically here if this meta-analysis should be done in other ways or if you can ever do meta-analysis with heterogeneous trials because they had maybe different outcomes, they had maybe different medications, they may have different durations, they might have had different genetic tests and so on. So we are aware of this heterogeneity and yet, you know, we can, I think we can carefully however conclude that the odds ratio is around 1.41 overall in randomized clinical trials and open-label trials and that this would indicate a modest effect of pharmacogenetic testing overall across studies. Again, leave it there for now, but again, you know, I'm open to discussion. Now expert groups and regulatory agencies have followed a little bit of a different pattern. They're not looking at gene combinations typically. They're not looking at, you know, what about if you mix, you know, if you put three, four genes together or if you are, you know, yeah, if you are, you know, doing an RCT with this or something. You know, they just sit down and review the literature in a very, I would say, in a very high standardized way and also with, you know, the best modern standards. And that group, CPIC, the Clinical Pharmacogenetics Implementation Consortium, is slowly progressing to walk through all the gene drug pairs relevant in medicine, not just in psychiatry. And luckily, I was involved in those ones here, which are targeting psychiatric medications. And so they would therefore also come to careful conclusions about what to do about a certain combination of genotypes or, you know, haplotypes, genotypes, whatever, and medications listed here. You know, for example, you take opioids, what about if you have 2d6 poor metabolizer, where it will affect codeine, it will affect tramadol, for example, right? That's what you can find there. For tamoxetine, primarily metabolized by 2d6, there's also indication that it will matter for children being taken a tamoxetine depending on the 2d6 metabolized status. But of course, we are mostly interested in antidepressants. And we have now high guidelines for SSRIs and for tricyclics and for antipsychotics. That's now just in preparation. But what this consortium will do, they will scrutinize all the literature, and then people are assigned to review the papers. And then in meetings, in regular meetings, over a year or two years almost, those will be discussed and consensus will be found if, you know, the relevance of a paper is weak, moderate, or strong. And then altogether, after assigning all the papers certain criteria, then there will be guidelines concluded out of that. And here's an example of how things can look like. So here, this is the SSRI guideline, one of the more recent ones. And interestingly, I would like to stress again the notion of extreme metabolizers, because they basically come to a very clear conclusion here about S-tetanoprime. So they would say, if you're ultra-rare metabolizers or if you're a poor metabolizer, you know, in both cases, you should probably better not use S-tetanoprime or, you know, do careful dose adjustments. But again, the strongest recommendations here for changing or for applying pharmacogenetics is again for the extreme metabolizers. That's what the literature tells us. So that's another way, I think, of seeing, you know, and talking about evidence that speaks, you know, that speaks towards those scenarios. And all that in clinical practice, what does that mean? So here, I'm just showing you an example of how we have built an integrated care pathway in Toronto to treat depression. And I'm sure, you know, clinics can have their own, they're different ones, everyone can have one on their own depending on what they want to admit or administer. You know, in this case, we have TMS on our place, we have ECT, but we also offer ketamine in some cases, now psilocybin studies and psychotherapy things as well. So but the point here is, you know, at some point when you go down these stages as a patient, right, at some point they will tell you probably no medication has ever worked for you or will ever work for you. Maybe we call you treatment resistant, and you should now consider TMS or ECT or something different, right? And here, I think that's the time when we should say, stop, let's do another careful evaluation of the medication. You know, if you have the information of your patient, you know, and it's all there, and you know, you can also, you know, use questionnaires like the antidepressant history form, and you can make a nice assessment. But if you have the suspicion that potentially there was maybe an extreme metabolizer role at play, that's a time point you should consider and say, maybe we should still, you know, do this one test before we go, before we go to ECT or something. People will not like ECT as much, but they will do it if needed. But you know, what is that compared to a genetic test? If you are now, you know, coming to a hospital for weeks and weeks and have, you know, relatively expensive interventions. So that's, I think, you know, the plea is to keep in mind before we label someone treatment resistant depression, or now we say difficult to treat depression, you know, to keep in mind to check for the genotypes. Okay, and then, now at the end of this first section, so to speak, let's just take a brief look at the gene drug pairs that people should ideally keep in mind when they are doing, when, you know, in clinical practice. Let's say you have a lab in your clinic, you would like to do your genetic testing on your own, maybe, and you want to know which genes do I need to look at, and here they are. You will find genes that we have not talked much about. That's because, for example, SRIP3A4, there's just one rare variant at this moment of time which was reported to affect quetiapine levels, but that variant is relatively rare. We're not generally worried about that too much, but it is there. And then we have HLA-B and HLA-A, which are two interesting genes. For the sake of time, we will not mention them, but they are interesting for, you know, potentially severe side effect, Stephen Johnson syndrome, with oxcarbazepine and carbamazepine, which we don't also use very much. Same for CYP2C9. Yes, there is a relationship to phenytoin, but I don't think that most of us will use phenytoin. These are a bit become obsolete in psychiatry, and yet that would be, based on the recommendations given out by CPEC, given out by labeling as the FDA or Health Canada or the European Medicine Agencies or the Japanese recommendation group, that would be what is currently the best, so to speak, evidence-based-driven panel. However, could it not be that if I test someone for 2D6, for example, or 2C19, could it not be that there is medication out there for which this information could also be used? And of course it does, right? In other words, medication is, you know, we're talking here about psychiatric medication, but let's just step back a moment and say, okay, but what about if people take other meds, and how can we maybe, you know, use information across pharmacological agents? Because CYP2C19, of course, also does metabolize or is involved in other medication. Tropidogrel, proton pump inhibitors, for example, are often metabolized by 2C19. And so I brought you this picture here, which gives you a bit of general idea about, you know, about the full, most complete picture, if you wish. But now what does that mean? Well, one thing that it means is maybe people are more likely at some point to go and get a genetic testing done for all these genes. So they're not coming to say, I have a depression, and now I'm going to see Dr. Mueller, and now he's going to tell me what to do, and maybe he's going to tell me to do this and that, and maybe this includes pharmacogenetics. No, people might go at some point and go to the family doctor and say, I have this and that, I have coughs, nausea, whatever, and the doctor says, okay, you know what, for $300, $400, $500, we can now test all your genes at the same time, and we can get a bit of better understanding what's going on. And then these people might come to you in your psychiatric office and already bring you the genotype information because they've done a test elsewhere. And this is probably what's likely going to happen because it's also much more economical to do those tests at once rather than to send a patient each time that they need medications to different labs because now they're doing psychiatric medication, now they go to a psychiatric lab, now next morning they may have to deal with GERD, and now they go to see another lab for the proton pump inhibitor. That doesn't make sense, right? So genomic medicine will at some point, you know, be, I guess, a common standard like electric cars will, and then we will have this information there no matter how much we like, you know, to think of it as beneficial, maybe not, or if we're on the fence, I think that's where things will be going. And I give you one example because I think we should also not forget our medications are relatively safe. They don't kill anyone. They might cause side effects, of course, but they don't kill anyone. But what about an example here, I picked out 5-fluorouracil. So this is a case report of a doctor who tragically passed away last year close to Toronto, and he was diagnosed with stage 4 colon cancer. But what killed him was actually the drug, the medication, 5-fluorouracil. And what happened? So 5-fluorouracil is a anti-cancer drug, as I said, and, you know, typically if you have a gene called DPD and you take the standard dose of 5-FU, enough of this 5-FU will be metabolized in a way that you are only getting a certain amount of cytotoxic metabolites, which are those who are supposed to attack the cancer. Now if you are, unfortunately, a poor metabolizer and you will not degrade enough of these 5-FU, now you get an overdose reaction and you might die of the overdose. Now ironically, that doctor has done a genetic test and it proved that it was normal. And so he was given the standard dose based on this relief. However, again, severe intolerance, basically gastrointestinal inflammations and bleedings were such that ultimately the patient died because of that. And so the tragedy was that this patient was from Southern Asia and carried a gene variant that was not well studied and was not part of the general panel. And so that also raised the question about, you know, how are we really taking care of different, of diversity, of diverse populations. Now because of this case, I'm sure in Ontario this variant will now be on the panel, but it shows us, you know, in other disciplines in medicine, medications can be deadly, can be toxic, and they want this genetic test to be done. And by this door, by this entrance, you know, it's likely that they will also genotype for all other medication at some point and that we will just be faced with the genotypes even if we like it or not, or if we have sent patients out or not. I think that's the story I wanted to tell here. Next about regulatory agencies. So we are just about to bring out a new guideline in Canada. The Canadian Network of Mood and Anxiety Treatment, CanMED, has received quite a lot of attention internationally, so an update is a very nice thing to do. And I remember the last one in 2016 was still very hesitant about pharmacogenetics and said it doesn't really make sense, you know, would not really be helpful at this point of time. They always talk about routine testing also, I should say, but now I see a nice little improvement in a way that it's encouraged to maybe be used, you know, in some clinical situations, you know, when there are several severe adverse effects with low doses or when there's poor response of therapeutic doses. So you know, even a conservative guideline in the past has come to say a few good things about, you know, keep it in mind. In some cases, again, it's not about routine testing, you know, at this point of time testing everyone would be expensive, who's going to pay for this, and so on. I mean, if funds were available, why not? But of course, we need to be careful with our few health care dollars that we have at our disposition. But again, they're also mentioned nicely this time, and again, I was involved, I have advertised for that, to also keep in mind that certain populations might have different gene variants we just not know about, or maybe will know about in the near future, so that things are dynamic. So this at least gives me some personal pleasure that in Canada we have now this guideline that at least, you know, is advocating and calling out for PGX testing in situations as I said earlier, for example, when you are about to decide treatment-resistant depression or not. So now, commercial labs, a few words about them. The general trend is people like to know the genotype, they like to know the genetic vulnerability, they like to know the ancestry, and so on. So if you take it from a recreational level, but also from a, you know, I think a well-educated level people typically want to know the genotypes. I have never seen a patient who says, you know, well, there is, you know, your drug is metabolized, we know exactly how it's metabolized, there is a chance that how it's metabolized might affect your outcome, would you like to know it or not? I've never heard, no, I don't want to know it, why should I? Most people will say, yes, of course, I, you know, thank you for telling me, how can I do it? Does it really matter? And then I give a careful explanation and I would not oversell things, but typically people want to know, and that's where things are going. So people are more likely to also come up with their own tests and, you know, bring them to the clinic and whatnot. However, one caution I would like to make is that these tests are very different. So the last time we checked in what tests are available in Canada, and many of them are available in the U.S. as well, yes, there is a good amount of difference between them in prices, in terms of turnaround time, in terms of which alleles are looked at, and if they are physician-gated, which means a physician is involved in the discussion or not. And, of course, the alleles tested as well, unfortunately, means that, you know, sometimes there might be a bias towards European populations and non-Europeans are at greater risk for maybe false negative to PGX testing results. So it's not that, you know, it's an easy way, it's an easy market to look through. And yet, again, if done correctly, if you find yourself the company, and I'm not going to talk about for the company names, but if you find yourself the right company, you know, who does follow the principle, yeah, then you should be good, at least, you know, for what you want to know. You know, those important evidence-based gene variants that we talked about. Another tool I like to use in my own studies, in the studies that we just did with autism, is a sequence to script. So we do the genotype in our lab. We have our own genetic lab. Jim Kennedy is basically, you know, the director of the unit, and so we do our genotypes, and then we use the software in collaboration with Dr. Bausman, and that tells us, it gives us a nice report about what should be, what would be recommended based on evidence-based, based on the evidence-based recommendations given out by CPIC, WP, DPWG, which is a Dutch group, and, or FDA, and this is how a report would look like, so it's not exciting, it's not, you know, in any kind of way exciting here. In this case, it's, but easily understandable. I mean, you know, it's nothing fancy. There's no, I don't know, there's no, like, like pictures or graphs or anything. Here you see, you know, an example of someone who's 2C19 and 2D6, poor, poor metabolizer, and the recommendations that are listed on the left side, and an explanation on the right button, why this, you know, what would exactly be recommended for estitalopram, for example, why the decision, estitalopram, is to probably not use it, right? So that's a little bit where things are in terms of, you know, going out to commercial labs and commercial groups, but again, keep in mind, many, many clinics are doing the testing in their own house. Finally, two more things, maybe to say some cost considerations. The famous quote, it's the economy, ta-ta-ta, right? Reminds us, at the end, it's also about money, and here, we just have a recent study conducted in the province of British Columbia, so this is the north, the northwest in Canada, or the southwest, to be fair, but for you, it would be the northwest. There, they have done a very elegant cost-effectiveness study estimating the savings for genetic testing in depression, and come to the conclusion that within two years, all the costs for genotype would have been offset, and over 20 years, you would even save a billion dollars, potentially, but they also nicely say that cost savings are mainly driven by, you know, avoiding, avoiding the, if you wish, you know, to classify someone as pseudo-resistant, so to avoid, you know, not labeling someone treatment-resistant, whereas, you know, there is still a chance that medication could help if a genetic test is done at that time. Good, so then, nothing is perfect until maybe everything is also available in the integration in the electronic health records, right? So that's another challenge ahead, and if we have, if your institution has the option, the possibility to include things in the electronic health records, well, that's great, but many times, people will still come around with flyers, and then, where are they when you need them? So I think, in the interim, what we always recommend is to carry at least wallet cards so that people can, you know, quickly take them out when they go to a pharmacist or someone, and say, this is what was tested, maybe it's relevant or not, and until we might find better ways of, you know, doing integration of electronic health records, which is underway in many clinical settings, but so far, not largely implemented. Good, second last thing, I think, is about the ethics. There's not much to say, I think, about the ethics, but this quote here was actually a quote that we came up once we did a retreat. Now, Victoria Marsh, she kindly quoted me on that, but it was really the group who came up with this kind of summary about what we're doing, and Victoria's now a postdoc here at Columbia University, and has been doing a tremendous amount of work while she was a PhD student, but the ethics, again, it's interesting, because, you know, we have to ask ourselves, of course, are we doing any harm, maybe, with pharmacogenetic testing, right? So one way how we looked at it was to do a survey after pharmacogenetic testing was done in Toronto, and then we had those, you know, 382 physicians who replied, and we wanted to know, basically, well, if we have given you a recommendation for a treatment change, and you have used it, have you seen your patients later maybe doing worse, or have the patient maybe complained, or have the patient maybe felt misled in any kind of way? And we saw, relatively surprisingly, a clear non-harmful intervention here, because there were only two people out of these 380 or so where the physicians said maybe they have minimally, you know, got minimally worse, but we had had a big portion, most of the people benefited. So in other words, that survey, and again, it's not like a proof of anything, but it has not shown any time like frustrations or misleading or going the wrong alley. I think it's because people are reasonable, ultimately. In practice, they will not typically oversell the results. They will use it as an additional piece of information. They will discuss it, consult with the patients, and then they will make a decision based on that. And ultimately, I heard cases in children and adolescent psychiatry where this way of maybe giving out reports was overly interpreted that this would be the only medication that you can take, and then maybe things did not work out, and maybe they left a little bit the treatment protocols, which would recommend SSRIs in most cases in children, but then maybe they would get something else because it was not listed. So there are these cases out there which are sad, but definitely, however, we have very little evidence to believe that there is any harm done at this point of time either based on our own data or on our own experiences. The other thing is now people want blinded studies, right? People said, well, these studies that you did or that are shown are all not really blinded. We need more blinding, we need better blinding, and so on. And that's true. It would be great to have the blinding. It's technically not easy to do, but there's one study I'm aware of where they would give you a report which is either the one which includes the standard recommendations, or you would receive a report which is standard recommendations plus pharmacogenetic, guidance included, but they would not tell the physician and the patient which one they receive, right? And so that's an elegant study. There are two also out on the way, one in children and adolescent psychiatry and one, I think, in adults in Australia. Let's see what they do. But I can tell you so much. When I submitted a grant proposal about a year ago or two years ago, and I said I would like to blind my patients to see what the situation is, to better understand pharmacogenetics, the reviewer came back and said, this seems ethically problematic, because you're withholding useful information, right? And yeah, well, I said, true. I mean, thank you for being so, believing into PGX, but you're also now telling me I'm not getting the grant because I'm not doing ethical work here. That's tragic. So we changed the protocol and now we don't have those blinded arm anymore and we will maybe do another study with blinding. But those studies, those clinical trials are very, very expensive. I will just let you know. We have spent millions of dollars for the GAP studies. Those million of dollars, not easy to get because again, the reviewer will say, but the evidence is already there. What are you gonna prove? So it's hard really to imagine that we will see new large trials of that kind, hopefully, but I'm a bit skeptical. Finally, last point, education and training. So I showed you David Razek's book and I think it's still a great book, even though it's now 12 years old. We need to educate the patients, the public, the healthcare providers who are in training, residents, pharmacists, and so on, and of course, of psychiatrists. So for that book, we hope to be able to edit an update, right, Jim, which we will hopefully get some of you maybe also involved in writing, helping writing chapters or so. Please join us if you wish. For psychiatrists, I can offer a clinical fellowship to be considered in Toronto. We are seeing patients and we'll do pharmacogenetic testing included. So this is a pharmacogenomic fellowship if you wish. And we are, however, we have many other interesting domains. I'm just highlighting mine here. But interestingly, again, if you're a physician, psychiatrist, you have finished your residency, you can come as an international medical graduate from all over the world. You will not need to do any written exam to do a clinical fellowship and see patients. So I'm coming to an end. And again, if you have any questions about anything I said, please join me later. But here I wanted to let you know I've covered all these domains, if you wish, of how to get pharmacogenetics implemented and how to make best use of it. I still believe that there is a bit more to say, perhaps, as of society. I think we are now seeing a model of medicine where we are offering equality. Everyone gets treatment. I hope at least that everyone who comes to see a doctor in a clinic, in a walk-in clinic, will get some standard treatment. But that's not really equity. That's not really what we probably are doing here if we give the same tool, the same medication, the same intervention to everyone without taking into consideration their own individual dispositions and maybe vulnerabilities, right? So this is where I think the equity comes in also. And we're talking really about, again, individuals and not populations that we are seeing in practice. So when I got these nice testimonials of patients that we've seen in our own practice and they tell you about their frustrations with their system and with the medications and how, for them, it seemed that the test has really made a difference in their lives because finally they found a medication and they might have given up to take any medication. So this was also, for some patients, a restart to say, okay, maybe I can try again. And that was a very, I think, interesting observation that, again, it is, yes, sometimes, often, it can help to explain past non-responses and whatnot. And ultimately, again, very rewarding to see your own patients doing so well if you used a tool that would normally not be available. Summary and take-home. I would say that for, as we see, pharmacogenetic testing for actual gene-drug pairs and if using expert guidelines, so if you follow the principles of good medical practice, finding suggested individuals who are extreme metabolizers, as I called them, have a mild to moderate to higher risk of failing antidepressant medications. I think that is really the way it is. We cannot debate it away, I think. The next thing is, think of your clinical population, who you see. There seems to be an enrichment of extreme metabolizers in patient populations who fail past treatments, right? We've seen this all the time here. That's another interesting component. And, again, as such, it seems reasonable to consider PGx testing at this point of time. Mostly, the best bang for the buck, if you want, is to consider to rule out extreme metabolizers. I call it also ruling out. I'm not even saying it's going to give you complete new medications or something, but rule it out by that, that you don't have extreme metabolizers before you make further decisions. And now, I would just like to thank my group, who some are here, but without them, without the group, without the team, this work would not have been possible. This award would not have been, I think, been awarded. And so, I would like to thank them all, and I will all take them out for a good karaoke night, because we like to do that in Toronto. And I would like to thank you for your time and patience, for coming on an early morning on a Monday. And please let me know if you have any questions. Thank you very much for your attention. Thank you. Thank you for the excellent presentation. I had a question, I'm from Canada as well. Are these tests available on the insurance, private insurance? I have not fully understood, sorry. The PGX test, are they available on the insurance, private insurance in Canada? Are PGX tests available in Canada? Covered by insurance, sorry. There's still noise here all the time, sorry. Covered by insurance. Covered, no, they're not. No, well, we have a national or provincial healthcare system where typically the province would take care of all the expenses. And there, pharmacogenic testing is not yet covered, not for psychiatric medications. Is that the question? But there are private insurances that could cover them. You know, you can go and get yourself private insurance and they might cover it. But it's not covered for psychiatric medication on a large scale at this point of time. Not in Ontario. We have also provincial regulations, so, I mean, medicine or healthcare is regulated by provinces like here, probably the states are also having their individual rights. So it's not covered, I believe, in none of the provinces. The second part of the question was, do you need to have a mental health problem before you can ask for this test? Can we have tests done and have the results, a SIPP profile, for example, being available to me if I need medication in the future to see what my metabolism would be? Sorry, again, I might need to come closer or so. Can you speak up a little bit? Sorry. So the second part of the question was, do I need to have a mental health condition before I get a test to be covered for the insurance? Do you have to have a clinician? A mental health condition or to go on some medication before I can be covered for this insurance? Not sure. If you need to have a diagnosis to be covered, yes, you would need, or yes, maybe that's a question was, do you need to diagnose to be covered by the insurance companies? Yes, typically you would need to have, of course, a good argument would be a diagnosis or at least, you know, the intention to treat, right? The intention to treat and a diagnosis like depression would definitely be, I think, required, yes. Did that answer the question? It did. My next question was, shouldn't we have it as, for example, shouldn't I be carrying a card with me? My zip profile, for example, like if I go to see a doctor, a pharmacist, I want the interactions to be checked out, it should be readily available, wouldn't that be very helpful? Sorry, can we maybe talk about this later? Sorry, I just have a hard time. Maybe I have a hearing impairment, I don't know. Hi, thank you for the talk. Can you hear me okay? Yes, very good. Thank you. Okay, my name is Sonali Mahaj and thank you for your talk. I'm a child and adolescent psychiatrist and I was wondering if there's any evidence for using the testing to help us figure out what stimulant might be helpful or not or the profile for stimulants as well as the longer acting non-stimulants, guanfacine and clonidine. Okay, so the question is if there is genetic testing for stimulants, right, in stimulants. I'm not aware, really, of good data right now, convincingly arguing that, you know, it would make sense. I mean, there was a bit something for modafinil, I believe, but I'm not completely sure. If you want, however, I can follow up with you and see what the situation is. You know, there is a website called PharmGKB, Pharmacogenomics Journalist Base. That's the one I would look at, what is currently being collected as data. You know, they're basically starting to collect the data that is then ultimately making it to a guideline if there is enough data. So, but let me know. I'm happy to continue the conversation. Yes, please. Thanks for a great presentation. One of the problems I have with the pharmacogenomic testing is that it tends to push patients towards desvenlafaxine, as was the case that you presented. Right. And that's simply a drug that's renally metabolized. Right. So, we could just not do the pharmacogenomic testing and just put patients on renally metabolized drugs. I agree. Yeah. And that is interesting that, in fact, without maybe, yeah, with an absence of a pharmacogenetic you might be inclined to do that. But you might also say, you know, the likelihood of having a normal metabolizer is also high. So, you would have to balance it. And based on what you think the patient would benefit, if you think you would normally prescribe something else, but now you're going for desvenlafaxine, right, I think you would have to have an argument for that. But it's up to you, obviously, to make an informed decision and to discuss it with your patient. Right. Now, if in this case, the patient I've seen, I remember that she failed on various other antidepressants and only agomelatine worked for her. Problem was agomelatine was not available in Canada because she lived outside. And so, she couldn't get agomelatine. And then she came and said, I would like to do something on my depression, but I'm so afraid because I didn't tolerate desvenlafaxine, I didn't tolerate sertraline, and so on. And in that case, I said, well, okay, what you have never tried probably is desvenlafaxine. Let's just do it and also genetic test and see later on what's the situation. So we just prescribed desvenlafaxine and it proved out to be the only one left on the genetic test later on. And so, it worked for her. Luckily, she could have also not responded to desvenlafaxine. But again, I think it's from case to case where I would make a decision, you know. But I can tell you that, yes, you know, sertraline seems to be more, you know, more metabolized by various, I mean, by various cytochromes, right? Not just 2D6 or 2C19, but it's a bit of 2C19 and a bit of 2B6 and so on. So sertraline is my favorite choice now. Not so much because I'm thinking that it's doing much better, you know, antidepressant effects than maybe others, but because it's just a bit more inert to extreme metabolizers, right? So my first choice would be half-driven by pharmacogenetic reasoning without having the test, you know. And the next one, maybe, yes, if I can go to an SNRI, go to desvenlafaxine. So you can circumvent a little bit, as you say, these situations. And yes, if you, and then, you know, hopefully things will work. But again, it's all a matter of, I think, from case to case. Thank you very much. You're very welcome. Maybe it might be looked at in the other way, in terms of using electronic health records, for instance. If you maybe have, somebody has a side effect, say, to quetiapine, and then you might say, well, the people who have the side effect to quetiapine, maybe we should look at their genetics, because that would predict responses in other medications, and it would decide maybe an African population, it might be a case where we might look further for other problems. So as you might drive the investigation, look for investigation based on side effects that a person has experienced. Sorry, but what is the question, then, about this? The question is, you know, why look at genetics, you know, and when should you look at genetics? It might be guided by somebody has had a certain side effect, so maybe the genetic information might be useful. In other cases, it might be an enriched population for knowing what the genetic situation is. Yeah, well, I'm sorry, I'm still not sure. Well, anyway, maybe that's just a thought, maybe it's just irrelevant. Yeah, maybe it's a thought, yeah, or it's a comment, I don't know, but... Because you're going to start off with, we're going to start to look at genetics and go from there, and you might even go the other way and say, well, let's start with a side effect and see what the usefulness of the genetic testing would be in that particular population. Good, okay, yeah, sure, I mean, but for side effects, we don't have really that good amount of data, you know, to begin with, because it's not well monitored in many times, you know, like it's not standardized in many ways, it's not quantified, you know, it is a little bit then often debated whether it's part of the depressive symptoms that people might sleep poorly or have sexual side effects and so on, although I mean, I switch a lot of people away from SSRIs when they have sexual side effects, but I'm just saying there is not the data supporting pharmacogenetic decisions based on side effects alone. Please. Thanks again for that extremely informative and helpful review of the pharmacogenetic story, but I was struck by that review of a grant that concluded that it would be unethical maybe not to have your treatment as usual group not getting genetic testing, which is kind of a comment on what they think of treatment as usual, that it's not very good quality overall, and lots of other views have concluded similar things, so when you have that as your control versus this one test, the genetic testing, it's not surprising you find some improvement. Maybe you should do some studies where the treatment as usual group gets a little enhancement over what they normally do, such as, as was mentioned in the Canadian practice guidelines, checking a plasma level when somebody's not doing well, either in the too many side effect direction or no response direction, because the blood level gives you a lot of additional information. You could get a really high or low level, which would reflect possible pharmacogenetic issues, but also you pick up on non-adherence from that, potentially, people with a ton of side effects, but with zero level, you're picking up on nocebo effect, which is common with many treatment-resistant cases, and furthermore, we could add to the treatment as usual group to make a special effort to look for bipolar depression, so to see if they've got some history of hypomania that you've missed, I'm involved with a group that's done almost 6,000 consults on treatment-resistant depression cases, and it's amazing how many of these people have missed a diagnosis of bipolar. For one reason or another, they don't find out about the hypomanias, and this can lead to, of course, not a poor response to antidepressants, so I think the treatment as usual group needs enhancement, but I don't know that this is the best enhancement to routinely give them unless you could show us that this testing improves on somewhat enhanced treatment as usual. Yeah, thank you, these are good thoughts, and again, we would like to see more, I would say, real-world data at this point of time where we would see how things really work out if people come to, for example, to an integrated care pathway, they get measurement-based care, and we have a nice history about them, how can we best include or use genetic information in that population, because, yeah, a clinical trial where we would have, again, to deal with the treatment as usual, and yeah, and as I said, treatment as usual and blinded protocol also has its own challenges, and you know, I would say we would have to still work on the ideal clinical trial, and then hopefully, you know, there will be the million dollars, maybe five million dollars that would maybe, you know, require us to do it, so yeah, if you want to, if you want to continue, we can continue the conversation if you want to, you know, if you have good ideas about that, sure. Thank you. Thank you very much for the lecture, my name is Dr. Moukas, I come from Greece, do you see a future in measuring plasma levels, combination with pharmacogenomics, in order to have the best result for the patients, after all, metabolism is about how fast and how much of plasma levels we're going to have, do you see a future in that? So the question, if plasma levels would also... So we measure plasma levels also in a future clinical practice, as we probably use pharmacogenomics? Yeah, I mean, personally, we do it on a case-to-case basis, and we would potentially do it, I do believe that plasma levels are important for many medications, there are guidelines also like in our case, TDM guidelines, if you have the opportunity, and if you have, you know, really, the doubt that maybe there's plasma levels are, could be altered, you know, they could also be used complementary, it's a bit of a different approach than pharmacogenetics, but they can give you, of course, also exact measures, and can be very nicely complemented to those PGX or pharmacogenetic analysis. So ideally, yeah, you have both, a lab which can do either or, and maybe you combine them, again, on a case-to-case basis, potentially. And it's also great for research, because ultimately, we would like to see more databases where we have plasma levels and pharmacogenetic data, so we can have a real-world sample collected here, you know, that would be great, and that's what we're trying to do more biobanking. So, all, you don't have any studies, all the studies is about pharmacogenetics and clinical outcomes, you don't have any studies to see whether a certain genetic variant would change the possibility of a plasma level? Oh yeah, yeah, those studies do exist, for sure, yeah, yeah, they exist where the, definitely, where the genetics and the plasma levels very nicely correlate between the metabolizer status, and so on, right, yeah, yeah, they have been done. Thank you very much. Sure. Hi, I have a couple of clinical questions. So first of all, what do you tell patients when you're doing the testing, and they have to pay for it, and yet you know there may be variants that aren't being tested for, which could clinically affect the response? So you're telling them, get the genetic testing, however, you know, there are other issues we're not testing for, so it may not translate into a response. Yeah, so I think the first part was, what do I tell patients before the testing, right? Yes, well, I would review their, typically, their history, you know, and then if I notice that, you know, they have really had, say, very bad experiences, or very long term, you know, very long trials at high doses, and so on, I just got the possibility that it could be related to genetic factors, and if they would be interested, you know, that we could look into those, you know, depending on if we have a study available right now, or if they have to go to a lab, you know, I would then give them the directions. But ultimately, again, I would do it on a case-to-case basis, as being, you know, reasonable and cautious, but if there are ways or studies that we've had, like in the past, where the test is readily available, you know, my experience was most people would then be happy to include the genetic information, because it just doesn't, if it doesn't cost them anything, just other than a saliva and a research consent, you know, the interest is very high. But again, it always depends a bit, you know, to use resources mindfully, right? And also, when I test my patients, I tell them, you know, if you're on another medication from another doctor, I can look up that medication's metabolism, and then look at your profile and kind of extrapolate what might happen with that drug when you take it, whether it's blood pressure or gastrointestinal. Is that legitimate that I'm doing that? Is it ethical? Do I really have that ability to extrapolate that information, even though it's not listed on the genetic testing? Because I'm just doing psych meds. Yeah, I think so. I think that information that, you know, the patients are receiving is relatively, you know, limited to the drug metabolism or, you know, or nothing really otherwise in terms of, you know, being associated with medical conditions or risk for other things, you know. So, it is not, I think, it's not critical information, you know, genetic information to just use, I mean, if you just have the metabolizer information. Okay, my last question. I do a lot of the MTHFR, the folate testing, and a lot of my people are in that intermediate zone, the 40 percent that aren't normal or poor. Would you be supplementing them, you know, with a Deplin or an L-methylfolate at that point, which is usually an out-of-pocket cost for them? I'm not sure. You're supplementing them with what? What was it? I do genetic testing for folate metabolism. Ah, folate, or the MTHFR? Yeah, so in those people in that middle zone that are partial metabolizers, do you encourage them to take the extra, the L-methylfolate as opposed to folate? See, I don't think we have convincing evidence, really, to test that folate genes, MTHFR or so, you know. I must say, it is a little bit, I think, has gone a bit off track, you know, this folate testing, so to speak, you know. I don't, I would not support it, to be honest, you know. I don't, I don't know what it means. I don't know what it tells people. I don't know if it does anything, you know, those variants, I'm afraid to say. Is it part of your testing in Canada, or you don't test for it? In general, I mean, but I would, I mean, I'm talking about general genetic testing, no matter where. I would generally not recommend including MTHFR or so, you know, the, you know, the gene panel that I listed earlier, if you were already here, you know, did not list MTHFR for that reason, right? We had CYP2D6, CYP2C19. We've had, we've had seen a few other genes, but not MTHFR for the folate, for the folate metabolism. Okay, thank you. Sure. Hi, I'm Gen Shinozaki from Stanford, and I had a pleasure mentored by David Mrazek when I was a resident at Mayo, like 15 years ago, and pharmacogenomics, my entry to my research field, so I'm so glad that you guys continue working for that, so thank you for that. You know, but back then, my favourite gene was a serotonin transporter gene, 5-HTT-LPR thing. Are you still using it in a gene site, or I didn't see that information for that aspect. Are we using what, sorry? Serotonin transporter gene. No, oh, the serotonin transporter, yeah, good point. Excellent. I'm not following the literature, so I'd like to know. Yeah, no, that's a great question. The serotonin transporter, then, would be part of the pharmacodynamic genes, right, as you know, and same as the norepinephrine transporter, by the way, and we've had a nice finding with the norepinephrine transporter gene, and lofaxine, and elderly patients. Really very nice finding. It was published in the American Journal of Psychiatry, but unfortunately, the amount of data and the conclusions to draw are relatively limited at this point of time, and someone, somewhat there seems also to be potentially ancestry-specific, you know, characteristics going on there that we have not fully understood, so I think we would not endorse serotonin transporter at this point of time as a standalone gene. You know, if people use it in a combinatorial way, and if they find, you know, and they can prove that it has done something in addition to not using it, I'm not against, I'm not against good science, so to speak, but I'm just saying that as a standalone gene, you know, it would not help us a lot at this point of time. Thank you very much. Sure. That was the last question, I think. Well, thank you very much.

pharmacogenetics

psychiatry

Dr. Daniel Mueller

antipsychotic side effects

genetic markers

precision medicine

treatment-resistant depression

genetic testing

cannabinoid 1-receptor

melanocortin 4-receptor

dopamine D2 receptor

neuropeptide Y

clinical integration