Hello and welcome. I'm Dr. Benjamin Dress, Professor at Rosalind Carter Chair in Mental Health at the Rollins School of Public Health at Emory University and health systems expert for SMI Advisor. I'm happy that you're joining us for today's SMI Advisor webinar, Fallacies and Coups of Psychopharmacology, focused on lithium. SMI Advisor, known as the Clinical Support System for Serious Mental Illness, is an APA and SAMHSA initiative devoted to helping clinicians implement evidence-based care for those living with serious mental illness. Working with experts from across the SMI clinician community, our interdisciplinary effort has been designed to help you get the answers you need to care for your patients. Today's webinar has been designated for one AMA PRA Category 1 Credit for Physicians, one Nursing Continuing Professional Development Psychopharmacology Hour. And credit for participating in today's webinar will be available until April 11th of this year. Slides from the presentation today are available in the handouts area found in the lower portion of your control panel. We'll reserve 10 to 15 minutes at the end of the presentation for Q&A. Now I'd like to introduce you to the faculty for today's webinar, Dr. Naseer Ghami. Naseer Ghami is the Director of the Mood Disorder Program and the Psychopharmacology Consultation Clinic at Tufts Medical Center and a professor at Tufts University School of Medicine. Dr. Ghami's clinical work and research has focused on depression and manic depressive illness, particularly in exploring the role of antidepressants in the treatment of bipolar disorder. Thank you so much, Dr. Ghami, for leading today's webinar. Well, thank you, Ben. And thank you for your invitation to join the webinar and to give this talk. I appreciate it very much. As you said, the focus of the lecture will be on lithium and the topic is fallacies and truths of psychopharmacology. And I'm going to move my slides ahead and you can see here how to download handouts, participate in Q&A. Okay. And my disclosure is that I worked at Novartis in research until June 2021 and no longer with them. And the off-label concepts that will be mentioned here are some of the indications of lithium outside of bipolar illness, like prevention of suicide, dementia, other things, and antidepressants and bipolar depression. And I will move on to the talk. The basic concepts that I want to share with you are that there are two kinds of drugs, two kinds of thinking about drugs that Hippocratic tradition needs to be properly understood and that the biological role of DSM in using drugs needs to be, I think, critically analyzed. And let me spell out these ideas up front and then we'll apply them to lithium. So the two types of drugs are symptomatic drugs and disease modifying drugs. Symptomatic drugs are like aspirin, Tylenol for a headache, steroids for short-term symptoms of various inflammatory diseases. These drugs improve symptoms, but they are not long-term solutions. They don't get at the underlying causes of the symptoms. And disease modifying drugs in contrast are drugs that affect the underlying pathophysiology of the disease that causes the symptoms. Sometimes they get at the actual cause, like penicillin for antibiotics for pneumonia. Sometimes they get at the pathogenesis, like the disease modifying therapies for multiple sclerosis, statins, and antihypertensive drugs for prevention of stroke and heart attacks. These are long-term treatments and they do not affect symptoms. They do not improve symptoms. They're not for the symptoms. They're for the underlying cause of the symptoms and the pathogenesis. And now my claim is that most of our psychotropic drugs are symptomatic drugs, like antidepressants, antipsychotics, etc. And one of the few disease modifying drugs in psychiatry is lithium and perhaps a few of the other mood stabilizers, but lithium most clearly. Therefore, it has a unique role in psychiatry, in my view. It is not just one of a list of drugs. It's in a class of its own. It's the only clear disease modifying drug. And we'll talk about how that plays out clinically. Now, the second concept is the Hippocratic tradition. The Hippocratic tradition is not what you are usually taught, the so-called first do no harm rule. Hippocrates actually never said that. And that's a statement that was falsely attributed to him 2,000 years later. In fact, in Hippocratic writings, what you find is the phrase as to diseases, try to help or at least not harm. This is a very different concept. It's not about just being conservative about drugs or not liking to give people treatments. It's about identifying diseases as the central role of the doctor and the clinician, and then to try to treat those diseases if you can. And if you can't, that's when you become more conservative and minimize your treatments. So here's a picture of the way this philosophy, I think, can be understood, which is that the patient presents to you with symptoms. The job of the physician is not simply to treat those symptoms, although that's basically what most people do in psychiatry today. Rather, the job of the physician is to decide if those symptoms reflect a disease or not to determine that, and then to determine what that disease is and to see if that disease has proven treatments and then to give them. And that sounds very logical, but we don't actually do that frequently. Many people don't identify diseases or don't even believe in them, or there's controversy about what the diseases are, or the evidence for treatment may actually not be there or may be misinterpreted. But anyway, that's the way it should go. Now, you may find that there legitimately is no effective treatment, that the evidence isn't there, so it's untreatable, or that the disease is self-limiting, that it goes away by itself in a short amount of time. In those cases, you identify a disease, but you don't treat the patient because you don't either need to or you can't. In the other case, you identify that there's no disease underlying the symptoms, and you don't treat the patient. Now, by don't treat, I don't mean you ignore the patient, you don't do anything to them. You might treat them with symptomatic drugs for their symptoms, short-term, low-dose, but you don't do it for years and decades, and you don't pretend that you're doing anything profound. You're not going to cure them. You're not even going to help their symptoms much usually, but you may help some. But you don't want to overdo that. That's how you cause more harm. So the Hippocratic concept of not causing harm has to do with not treating symptoms aggressively and to focusing more on treating diseases. That's the basic philosophy that was proposed by that school of medicine that has not actually been played out in medicine, or wasn't played out for centuries, millennia, actually, after Hippocrates, but is probably, in some ways, the core of scientific medicine today, and yet is not really the way psychopharmacology is practiced. So you have to remember that the natural course of the illness will lead to recovery. And for instance, antidepressants and bipolar depression are given, and patients think they get better, but it's really because the depressive episode is brief. It's not the antidepressant that's making them better. We've shown that over and over again in studies which show that the placebo effect is equal to the drug effect. The placebo effect is the natural history. It's the course of the illness, and we always have to keep that in mind. One way of understanding what a disease-modifying drug is, is it's a drug that improves the course of the illness long term. And when you ask that question, you find that there's good evidence that lithium does it, but very little evidence that anything else does. The third basic concept is that the biological validity of DSM is questionable, and therefore it is not a reliable guide when you say, okay, I want a disease-modifying treatment. What's the disease that I should treat? It's not always helpful to say, well, okay, what's the DSM diagnosis for it? Now, that may be the case for schizophrenia, which is one of the best established diseases that has been also included in the DSM system in many ways along the lines of what's established scientifically. But it's not the case for many other diagnoses, among which I will show you is mood illness, both the diagnoses of major depressive disorder and even bipolar disorder. The way they're defined by DSM are not scientifically as valid as you might think, and therefore not biologically valid, and that leads to problems with making decisions regarding treatment. Now, why is this the case? I just need to give you a brief background to this before I show you how this plays out with mood illnesses. DSM-3, the third edition, which is the basis for the current system, as you know, was established in 1980. And the one basic idea behind it was to establish reliability, that is, to have us agree on definitions. That's like having a dictionary where you look something up and we all at least are using the same terms the same way. But the second step was validity, which was, okay, we might use the terms the same way, but is what we are describing real? You know, there are unicorns in the dictionary. That's not a real thing. You can define something very well, it doesn't mean it exists. The problem is that the dictionary became a Bible, as you know, a Bible of psychiatry, and when people worship things like the Bible, they don't tend to change them. And this led to a conservatism, whereby future revisions of DSM, the fourth edition in 1994, the fifth edition in 2013, made really minor changes on the third edition of 1980. And science doesn't work that way. Science isn't about keeping things the same, it's about changing them. And we are stuck with a diagnostic system that's based on what people knew in 1980, which in 2023, we know is wrong in many ways, but which has not been changed. If you go back to what people knew in 1980, the research diagnostic criteria, which were published then, identified about two dozen, a little less than two dozen diagnoses that were scientifically valid in their view. And we'll talk about what that means based on validation criteria. But DSM-III in 1980 had about 300 diagnoses, which means about 95% of the diagnoses were not scientifically valid at the time in 1980. They had not been validated. And we still don't have validation for most of those diagnoses. So what do we mean by validation? We mean that you take patients with certain symptoms and they ask the question, okay, here's a patient with a different symptom than another patient. Are those different diseases or those different diagnoses that may reflect different diseases? That's not necessarily the case because you can have different diseases with the same symptoms. And you can have the same disease with different, with a range of symptoms. So symptoms by themselves don't identify whether a disease is present or not, or whether two diagnoses are different. Just because you have different symptoms doesn't mean you have different diagnoses. So in medicine, we look at pathology. We look at autopsy to identify whether the clinical symptoms identified in the patient are reflective of a disease based on what you see in the organs, in the abnormalities in the body. In psychiatry, we don't have that. So what had been suggested by Robins and Gouzet in the 70s, which still is the basic way that diagnostic research is done, is to look at different diagnostic validators that are independent of each other. So one is family history or genetics, and that works for some conditions that are genetic. Another is the course of the illness, which is the most important because that tends to differentiate most conditions. Some are chronic, some are episodic, they have different ages of onset. Some get better by themselves, some don't. And biological markers, to the extent available. Treatment effects are nonspecific and it's not a good source of assessing this, although people often use that. With other biological markers, sometimes you can show out changes in MRI, differences in MRI, or other biological features. So this is the way we say that a diagnosis is valid, that it differentiates from other diagnoses based on genetics and course of illness and any other biological markers you might have. That kind of research has not shown that many of our current diagnoses are valid. And one of the reasons for that is because the definitions for the diagnoses has been pragmatic, as some of the DSM leaders have said in the past, in which the scientific research is not the main priority for making or changing criteria, but rather, what are the utilitarian benefits of a definition? What do clinicians think about it? How does it play out in practice? How are things reimbursed? What is the impact on legal risks? This is all fine for clinical work, but when you define diagnoses that way, it doesn't mean that the genetics and the pharmacology of those diagnoses are actually going to differentiate the way nature has it. In other words, when you define diagnoses for legal and economic purposes, you're not going to find a gene specific to that diagnosis. You're not going to find a receptor that is relevant to that diagnosis. And that's one reason why with DSM-5 in 2013, the NIMH put out a statement that it was no longer going to be using DSM criteria for its biological research. Now, the research diagnostic criteria, you know, you could criticize that also, which was the alternative they brought up. It's false in the sense that nothing is true until you prove it, and it hasn't been proven either. It's a hypothesis. But the question might be why we are okay with using some of these diagnostic definitions for practice when we aren't using them for research. And I think it becomes relevant as we think about understanding the research and applying it to practice. And so the example we're going to use is lithium, as I said. And what I want to show you today, the upshot of what I want you to understand, is that lithium is not a drug just for bipolar illness, but it's a drug for mood episode prevention, period, whether someone has the diagnosis of bipolar illness or so-called major depressive disorder, and that it is a drug that appears to prevent suicide. And I'll go through the details on the pros and cons of some of the evidence for this. I'm going to contrast it a little bit with the evidence on antidepressants and major depressive disorder. I won't focus on that too much, but just as a brief contrast, since that's a class of medication that's used so broadly, so extensively for this diagnosis of major depressive disorder. And as I said, it's also used in bipolar depression, where it's more clearly disproven. But in MDD, people think it's well proven. And I think there's questions about that. But the key point is that antidepressants, to the extent they're helpful, are symptomatic drugs. They are helpful short-term for the symptoms of MDD. They are not disease-modifying drugs. They don't prevent the episodes. They don't improve the course of illness, unlike lithium, which does that in MDD, not just bipolar. And the term MDD itself, I put in quotation marks because I don't like it. And the reason I don't like it is because I think it gives a false impression that it's a thing. It's a real thing. It's a reliably defined thing. But whether it's a valid thing is another question. And now I want to show you this background applying the discussion that we've just had about DSM. So in 1980, the third edition of DSM, this is applying it just to mood illnesses. What existed in terms of mood definitions were the phrases on the top of the slide, manic depressive illness, involutional melancholy, it was called, and neurotic depression. Manic depressive illness was a disease of severe recurrent mood episodes. They could be depressive only, which was called unipolar depression, or they could be depressive and manic, which was called bipolar. The term manic depressive illness did not require the presence of mania. You could have just repeated depression and you were defined as having, it was called manic depressive illness depressed type. So the first point to understand is that manic depressive illness in the older phrase is not the same thing as bipolar illness. It's what today you would call bipolar disorder, plus a lot of what you call major depressive disorder, MDD. That unipolar phrase was not included in DSM-3 for reasons that we won't get into. It has to do with some political haggling between the psychoanalytic community and others, where the term unipolar was seen as biological and thus would lead to medication treatment. Many in the psychotherapeutic community opposed that, so they did not allow the term to be included in DSM. The term major depression was then used instead, in contrast to minor depression. Again, I won't get into weeds on this, but the term minor depression was really replacing another kind of depression, which used to be called neurotic depression, which unlike manic depressive illness was not recurrent and episodic, but it was chronic. So it had a different course of illness, and it was mild, and it was associated with anxiety symptoms along with depression to a mild degree, whereas the manic depressive illness patients often did not have any anxiety at all. They could, but it wasn't required. And they differed then, and also the genetics differed. Manic depressive illness was a highly genetic disease. Neurotic depression was not a genetic condition. So these were two very different kinds of depressions, but they were going to be separated as major depression and minor depression. But again, because of political differences in the General Assembly of the APA and the voting for DSM-3, economic concerns came up that minor depression would lead to less insurance funding, and so the term minor was dropped, and that was included in major. The old concept of melancholia was that it was a chronic, severe condition with severe psychomotor retardation that lasted for years, sometimes decades. We won't get into that, but that was also included in the MDD definition. So the bottom line is MDD involves a range of different kinds of depression, and the way the term finally came about involved political and professional debates about reimbursement and about what kinds of treatments would be given, not scientific research on what was differentiating these things based on the validators we talked about. That's why the term MDD is of questionable validity. It includes all these different kinds of depressive states, which legitimately could be seen as different, but they're all lumped together in one diagnosis. And so the way I look at MDD, as is shown on this slide, is it has these different kinds of depressive conditions mixed within it. The neurotic depression on one pole, the melancholic psychomotor retarded depression on the other pole, mixed states, which are psychomotor agitated and are generally seen as different and involve manic symptoms, is also there. Now that's been pulled out, and the term MDD with mixed features at least captures it a little bit. But again, that mixed state and the melancholic states were all seen as part of manic depressive illness, which is a different illness than neurotic depression, where you have this other kind of milder anxious state. And more recently, vascular depression has been identified. That's still included in the definition. That's a very different thing with a different age of onset. So MDD is a mix of different biological types of depression. And it's therefore not surprising that if you take a drug with a simple single mechanism like serotonin reuptake inhibition, that it's not gonna improve all these different kinds of depression. It might improve some of them, like maybe a pure depression in the middle, but it worsens mixed states. It doesn't affect melancholia at all. And neurotic depression, it doesn't improve quickly because the natural course of that condition is that the depressive exacerbations are mild and quickly recover. And so you can't be placebo. And so you end up with results where SRIs are not as effective as you might think. They have a small effect size of benefit. And as we'll show, they don't actually have long-term benefits for prevention of episodes as much as you might think. And for suicide prevention, as much as might have been hoped. Very briefly, this is a meta-analysis of hundreds of acute depression studies in MDD. And what I wanna show you is to describe this picture a little bit for you. The red line is the antidepressant effect, which is large. So on the y-axis is what's called a Cohen's d, which is an effect size. One is considered a very large benefit. 0.5 is considered a moderate benefit. And the kind of the threshold for clinically meaningful benefit, less than 0.5 would be a very small benefit. Now, if you look at the overall benefit, you can see that above one, the antidepressant effect, which is the red line, is quite good. The x-axis is severity. So it works in milder depressive states and in more severe ones. The blue line here that's sloping downward is the placebo effect. And that's also quite good for the milder clinical depression, but it gradually decreases with more severe depression so that it's still clinically meaningful. It's above 0.5, but it's not as much as the antidepressant effect, which is stable. Now, the difference between antidepressant and placebo is what we talk about when we look at these clinical trials. And the difference between the two is very small for the milder depressions and only clinically meaningfully different for the larger depressions. Clinically meaningful means, again, a difference that's 0.5 or above, and you don't get that kind of difference until you get to severe depression. So what this is telling you is a few things. One is, imagine that that placebo line wasn't there. If the placebo line isn't there, you see the effect, the observations of clinical experience. Clinicians think antidepressants work. They think they work all the time, and they do. But the placebo effect is what God knows. It's what nature knows. Nature and God know that it's not the pharmacological effect of the antidepressant. It's not the serotonin reuptake that's making patients better. In many of these cases, it's the natural course of the illness. That's the case with the mild depressions. Now, in the severe depressions, it's true. There is a benefit that's pharmacological, but I'm gonna suggest to you that that's just short-term and symptomatic. It's not a long-term benefit. So here's what clinicians see. The antidepressants work no matter who you give it to. And here's the reality. They're not really working because of the pharmacology in the milder or moderate cases that much, but more in the severe cases. Now, what about the long-term? Why do I say it's symptomatic? Well, here is a meta-analysis from the FDA where they looked at all the maintenance studies, the long-term studies of antidepressants in MDD, comparing it to placebo. And basically, on the y-axis, you have the difference between antidepressant and placebo by each of the 15 studies or so that they have. And there are differences in favor of antidepressants where the relapse rate is lower than antidepressants. In the first couple of months, two months, four months, six months, six months and out, you see there's no difference. There's zero change. So you cannot say, based on the evidence that the FDA has, which is all of it, based on randomized data, that these antidepressants have been proven to prevent depressive episodes more than placebo six months or longer after the acute episode. Now, there's reasons for this that we can get into in terms of the research design of the study. There's so-called randomized discontinuation trials where patients initially are treated with the antidepressant and then only included in the study if they respond, and then they're randomized to stay on or come off. This produces a short-term withdrawal effect. In the interest of time, we won't get into that, but that's a critique that I make of these designs. I think that these maintenance designs actually do not effectively study and prove a long-term efficacy. But separate to that question, what they do show to the extent that you wanna believe them is a three to six month benefit. They do not show a longer benefit than that. And you don't get into the maintenance phase of so-called MDD until one year after the acute phase. Six months would be generous. And at six months to a year, you're seeing no difference at all. In other words, no further benefit with the antidepressant for prevention of mood episodes. This suggests it's a short-term benefit. I should add that the interpretation I'm giving you right now is not what the FDA writes in their paper, but they show the picture, which clearly can only be interpreted that way, it seems to me. Now, the other issue is suicide prevention. And this has been a topic of great debate for the last 10, 20 years since the FDA came out with its black box warning on increased suicide risk with antidepressants in young people, in young adults and adolescents and children. And that's what you see on this meta-analysis. So the meta-analysis figure, one means there's no difference between the groups. If you're to the left, there's some benefit. If you're to the right, there's some harm. And you can see in below age 25, there's clearly harm. There's a 62% increased odds of suicidal thinking or attempts. And then if you look at all the other age groups, there's a benefit, but it's a very small benefit. It's about a 17% when you add it up benefit. Maybe it's largest in older adults above age 65, but in middle-aged adults, 25 to 64, it's very small, about a 20% benefit of decreased suicidal attempts. That does not translate actually to prevention of suicide much because the concept of suicide is complex. But before I go there, I just wanna make this point that the benefits that you see with antidepressants is mild at best, but the harm in younger adults is actually quite large, quite notable. And again, the question is what that translates into in terms of actual suicide, which is not as bad as it might seem, and we'll get into that in a minute. Now, switching from this discussion briefly of the limitations of the benefits of antidepressants, I wanna show you kind of the opposite of how much stronger the evidence is for lithium, both for long-term prevention and also for suicide. So for long-term prevention of mood episodes with lithium, here you have, again, a meta-analysis figure, but here it's just written a little differently so that zero is no difference, anything to the rights of benefit. And you can see there's a threefold increased benefit in prevention of mood episodes with lithium in manic depressive illness, which means both unipolar depression and bipolaromas. Many of these are just so-called NDD studies, we would call them now, but they used to be called unipolar depression, and a threefold benefit, which is a huge benefit for prevention of mood episodes. And again, you can see these studies go from the 1960s. In this case, it stopped in the early 2000s, but they've continued. We have 50 years of research with dozens and dozens of studies. There's nothing better established in psychiatry in terms of replication and amount of benefit than the prevention benefit of lithium for mood episodes and for both depression and mania in both bipolar and unipolar depression. So you compare this to the studies I just showed you with, compare this to this figure, which is 15 studies of antidepressants, which show no benefit after six months. And here you have dozens of studies with lithium. Many of these studies are a year or two years long, and they do show continual benefit. Now, the suicidality issue. Here's a graphic picture that we need to understand. The studies that I showed you before reflected suicidal ideation and attempts. Now, that's not the same thing as completed suicide, as this picture shows. Many people who die by suicide don't make, have not made prior attempts. Actually, about half the people who die by suicide do so on their first attempt. So there's not a history of suicide attempts in half of those individuals. Therefore, even though suicide attempts predict suicide, they don't do so for half the people who die by it. One would presume that everybody has suicidal ideation who dies by suicide. But in fact, in many people, the suicidal attempt is very impulsive, and any ideation to the extent it exists is very brief. And it's not long-term. So the majority of people with suicidal ideation don't actually make suicide attempts. The majority of people with suicide attempts don't die by suicide. Half the people who die by suicide have never made attempts, and many of them have very limited suicidal ideation. So ideation and attempts is not a great proxy, is what I'm trying to say, for prevention of suicide. There's really no replacement for the actual death by suicide itself, which is hard to study, obviously, because it's infrequent, thankfully, but also there are ethical issues around that. Another comment just to make in general about this, about half the people who have a clinical depression have suicidal ideation at some point, but half never do. And at any point, it's actually much lower than that rate. And about 10 to 15% of them make suicide attempts. And about 10% of the people that make suicide attempts eventually die by suicide. So it's not a great predictor to say that most of the people with suicidal ideation and suicide attempts will not actually complete suicide. It's not a great predictor. And here are the numbers again. Half of the depression have suicidal ideation, 10% make attempts, 10% of them actually die by suicide. So we're really talking about a fraction of 1% or 2% of people with depression who eventually die by suicide. And that's over decades of follow-up. It's usually decades later. So if you start studying someone who's currently suicidal, the death may happen 10 or 20 years down the road. Usually it's not gonna be next year or in that next month, but our studies are one or two months long in these cases. So we're not really following these patients in a way where we can assess it. So in the lithium studies, we have the benefit that we're not looking at just suicidal ideation and suicide attempts, we are looking at completed suicide. And we're doing this in a couple of different kinds of studies. There are geological studies, which I'll show you. And then there are some randomized trials with some recent meta-analyses. So here are the geological studies. Here's an example from Japan where you look at suicide rates in the general population and correlated with lithium levels in the drinking water and based on where people live, there's decreasing suicide rates correlating with increasing lithium levels in the water. This has been replicated numerous times, numerous places, different countries. Here's another example from Chile, which is an interesting big country because of the geography where it's a very narrow country. And at the top, there's the Atacama Desert, which is a lithium salt desert, has extremely high lithium rates. And in the bottom is a very, towards Patagonia, it's very green and there's no lithium there at all. The suicide rates are very high in Southern Chile and very low in Northern Chile. Now you might say there are lots of other reasons why people might have difference in suicides. Of course they do. That's why these geological studies are not definitive, but many of these analyses try to control for some of those things, such as urban versus rural location, as well as socioeconomic status and such. And still these differences persist. These are the differences that are shown with correction for some of those potential confounding factors. But one of the benefits here is you're looking at completed suicide, at death by suicide. So it does get at the definitive outcome. Now, what do we mean by high lithium levels in the drinking water? Well, as you know, lithium is an element. It was created in the Big Bang and no one can ever add to or subtract a single iota of the amount of lithium in the world. It was created by nature in the Big Bang. And it's in the world, it's in the rocks, it's in the water, it's in the environment. It gets into plants and vegetables and into meats and humans in their diet get about one milligram a day of elemental lithium. That corresponds to about five milligrams a day of lithium carbonate. Now, given that we prescribed 900 milligrams as a standard dose, that's obviously a very small amount of lithium, five milligrams of lithium carbonate, one milligram of elemental lithium. But even that normal amount, we don't always get. And in these areas where there's not much lithium in the water or depending on your diet, you could get up to near zero amounts of lithium in your diet. And those are the people that have multiple fold increased rates of suicide and even homicide in some of these studies. Whereas the rates are decreased two fold or greater by having more than a milligram, two or three milligrams of lithium in the diet, that's considered high. Well, that's still equivalent to only 10 or 15 milligrams a day of lithium carbonate. In fact, you can get five milligrams a day of lithium over the counter with lithium orotate, lithium aspartate. You can buy it on the internet and that's a high amount of lithium, which based on these geological studies would have a benefit for suicide prevention. Now, the only way you can really know, of course, if anything causes anything in medicine is doing a randomized trial. And in the case of suicide prevention, as I said, this issue has come up and this meta-analysis done in 2013 is probably the most definitive and valid one in my view. And what it showed was that indeed, lithium does have suicide prevention benefits. Now, if you look at the meta-analysis figure here, it's comparing lithium to different drugs. There, the differences are fewer, there are fewer studies, but more definitively, if you compare lithium to placebo, you see the benefit over placebo. There were no suicides with lithium, six in the placebo arm, and the difference is a odd ratio of 0.13, which is almost a 90% decreased risk of suicide with lithium. Remember we said with antidepressants, there was a questionable 20% decreased risk of suicidal ideation and attempts, which 98% of the time doesn't translate to the actual completed suicide. So it's really a minimal benefit if any, but here you have actual prevention of completed suicide with a 90% decrease with lithium, which is huge. Now, when they looked at not completed suicide, but self-harm, self-cutting and such, there was no differences between lithium and placebo or other options. And that's an important observation. Now, here's the picture of the suicide prevention, just making those couple studies bigger for you to see it. And here you can see it maybe more clearly. And when they looked at unipolar depression, they had the same benefit in those studies. Many of them had unipolar patients and they began manic depressions, same thing as unipolar depression, lithium prevented suicide, almost 90%. So it's not about bipolar illness, it's about unipolar or bipolar. Now, recently in the VA system, this major randomized trial came out and has gotten some attention, which claimed no benefit with lithium for suicidality. Their outcome was suicide attempts, not completed suicide. And in one year, as you can see in the picture, there wasn't much of a difference, but two points to make here. One is suicide attempts is not the same thing as completed suicide, and it doesn't invalidate the data on completed suicide, which I just showed you. Secondly, when you actually look at completed suicide rates in the study, lithium was three times better than the placebo. There was one case with lithium, three on placebo. Now, again, that's not statistically meaningful in a study like this, but it is when you add it to that meta-analysis, it just gives you pretty much the same effect, a multiple full decreased risk of suicide, actual completed suicide. Now, in this study, most of the patients did not have bipolar illness. They were diagnosed with so-called MDD and in the VA system, not surprisingly, 60% had PTSD. Now, remember, we've already discussed the fact that MDD is not a biologically valid condition. It's highly questionable to be a biologically valid condition. So it's not necessarily, this is not really a study necessarily of any specific diagnosis. It's a mixture of disease states, especially those, I'm sorry, of depressive states, especially those that happen with PTSD, really. And a very small percentage of people that we could say might be legitimately a mood disease like bipolar illness. If you take out that 15% in the bipolar group, there actually was a benefit of lithium in prevention of suicide attempts. And that's in the paper, actually, here's the table on the top level. You can see that there were 30 patients with bipolar, 10 of them had suicide attempts on lithium, 20 on placebo. It's a two-fold decreased risk, which was statistically significant. The authors of the paper never analyzed these data that way and never presented it that way. But I did, and there's a two-fold decreased risk. So lithium in this study actually did prevent suicide attempts in the bipolar subjects. And in the non-bipolar subjects, because of the high prevalence of PTSD, it's hard to say anything about it. And they were looking at suicide attempts. You might conclude that lithium may not prevent suicide attempts in patients with PTSD. We could potentially draw that conclusion from the study. But this doesn't mean that lithium in general does not have benefits for suicidality because it actually did prevent suicide attempts in patients with bipolar illness. And it did reduce the suicide completed suicide rate. Now there's one last misleading meta-analysis that was just recently published that's worth mentioning. This was one published in a journal where kind of pseudoscientifically, what they did was they included some of the studies where lithium prevents suicide, but they excluded some others just because they were older. And then they included many, many studies that were very short, that were just about three months long or three to six months long, but it's not long enough to show prevention of suicide. So essentially by increasing the denominator of negative studies, by including these very short studies, they claimed that lithium did not have a statistical benefit. But if you look at their analysis, here's suicide attempts on the left, there was not a benefit. And you look at suicide attempts on the right, there actually was a benefit. You see the overall lithium had two cases, there were five for control. The overall odds ratio was 0.42, which means almost a 60% reduction in completed suicide. But the reason they said there's no benefit is they included all these other studies where there was no suicides at all in either group because they were only two or three months long. And that's not a way to actually legitimately do this. That's a false negative. Because when you don't have any outcomes, you can't say anything about, you can't really even assess the outcome. You have to have one outcome in some group to be able to say whether one is better or worse than the other and as I said, you're not really assessing completed suicide if you're doing a three month trial, you have to do, it takes decades for the suicide to happen, you have to do trials that are years long. So without any outcomes, you can't say anything. So this is a pseudoscience false claim that lithium did not prevent suicide. Actually it did, the number is pretty much the same, 60% decrease rate versus close to 90%. It's still a massive decrease. It's just they have a false negative statistical non-significance by including a dozen studies with no outcomes at all, which are too short to assess the outcome. Now on suicide attempts, there is no difference. Okay, we can accept that, but that's already been proven. I showed you that in the first meta-analysis. Lithium is not about preventing suicide attempts per se. It's about preventing actual suicide. And half the people that die by suicide never make prior attempt. So it doesn't matter whether it prevents attempts or not. What matters is whether it prevents people from dying. Now, the last point I want to make after having proven that lithium prevents suicide, as I hope I have, is that it can prevent dementia. And there are studies, and I'll go through it briefly, where, as you can see here, the dementia rates are 22% of people with mood illnesses compared to four-fold lower rates by one's late 60s, 70s, in the general population. And lithium has been shown to reduce that rate back to the general population rate in these observational studies. Lithium is highly neuroprotective. It keeps neurons alive longer, as you can see in this picture. So on the bottom is a neuron that's dying by excitotoxicity. On the top, a neuron that's lithium-treated that stays well. And there's a Scandinavian study showing that lithium-treated patients have lower dementia rates than anticonvulsants. And I guess that's the last slide. So the point there is that when you're looking at disease-modifying effects, you're looking at also the mortality of the disease, like the suicide, but also the medical morbidity of the disease, which in this case is dementia. And here you have a benefit with lithium as well. So lithium's disease-modifying because it clearly prevents mood episodes, which means it improves the course of the illness. That's a definition of disease modification that I propose. It also decreases the mortality and medical morbidity of the disease, which is suicide and dementia in mood illnesses. And that's some of the evidence here I wanted to show you. Now, obviously we still need randomized data for dementia to be definitive on this. There are some studies that are coming out there, which do show some benefit with lithium, but they're not definitive yet, of course, dementia being a very difficult thing to study also. So let me finish on my fallacies and truths with all this background on the concepts and some of the data we've reviewed. I'm just gonna run through these. There's 10 in each group, and I'm gonna make a few comments in each one and then we'll stop. So the fallacies of psychopharmacology. One is that more is better. If you have a higher dose, you'll do better than a lower dose. Well, that may be for some symptomatic treatment symptom benefits, but not necessarily long-term. We didn't get into this, but I actually think that we use much too high doses of lithium and in the long run, doses that are lower than 900 milligrams do quite well in many people. The second fallacy is what gets you well, keeps you well. Excuse me. This can also be called the halfway ever after fallacy. This is the view that if drugs work symptomatically, acutely, then you should just stay on them long-term and they'll prevent the condition. That's not true. You know, you have a headache, you take Tylenol. You don't take Tylenol forever to prevent future headaches. Symptomatic drugs are not disease-modifying drugs. Treat to remission. If you aren't getting your symptoms with the drugs you're on now, add some more. Well, your side effects will increase, but you don't always actually improve your symptoms. And again, it's just symptomatic treatment. That's not a long-term solution to anything. Polypharmacy is good. Two drugs are better than one. Not necessarily, and clearly your side effects get worse. This is a variant on the first fallacy that more is better. Make one change at a time. Go slow, go slow. Those are two related ones. Well, these are opinions, but in my view, sometimes we go too slow and clinical practice isn't research. And it's not about doing one thing at a time just to make sure you know what works. You should just be trying to get the patient better. I've spent decades doing research. I have no problem with making one change at a time, but I think the clinicians who don't do research, I don't understand why they pretend, why they're using research methods and pretend that that's important. Patients suffer far too much sometimes by doing too few things. Now you can obviously do too many too, so you want to get it just right. Too slow or too fast are both a problem. Four to eight weeks is a sufficient trial of a drug. Well, not for long-term treatment, that's for sure. It's only for symptomatic and even symptomatically sometimes you need to go longer. Always taper a drug. Maybe if you have withdrawal symptoms, but not if you don't. Treat the most severe symptom. This is a very common fallacy that's based on this lack of understanding of the distinction between symptomatic and disease-modifying drugs. All we're doing is symptomatic treatment in psychiatry. So people focus on this. What's the most severe symptom? Let me treat that. That would be like if you have HIV and you say, okay, well, let me focus on the fever because that's the most severe symptom. Oftentimes the severe symptoms have no relation to the underlying disease and we need to change our focus to the disease, not the symptom. Always incorporate the patient's preference. Well, this is always preferable, but sometimes the patient's preference is about, for instance, the most severe symptom and they don't understand the link, say, to the underlying disease and there's some important education that needs to be done there. Well, what are the 10 truths of psychopharmacology? Well, the truths really are flip sides of the fallacies. Your treatment is only as good as your diagnosis as the former chair of Mass General, Ned Kassin, used to say. This gets back to the concept of disease-focused treatment, this Hippocratic tradition of treating diseases, not symptoms, as William Osler always taught. That's the focus we should take, not just focusing on the symptoms. All drugs are guilty until proven innocent, as the great 19th century physician, Alfred Wendell Holmes said. You shouldn't give drugs and just wait until they're proven harmful. You should not give drugs until they're proven effective for a disease that you can treat. Again, back to that Hippocratic philosophy. All drugs are toxic. Only the dosing and the indication makes them therapeutic from the medieval founder of modern pharmacology, Paracelsus. All drugs are toxic. Only the dosing and indication makes them therapeutic. All drugs have harms, biological harms, because if they didn't, they wouldn't have any biological benefits either. If there are no biological harms, then there are no biological effects at all and therefore no benefits. But if you dose it low enough and you give it for a disease, the indication, then the benefits will outweigh the harms. If you're giving it for symptoms and you're dosing at high doses, often the benefits don't outweigh the harms. Always have an exit strategy, as General Colin Powell used to teach from his experience in the Vietnam War. You don't start a war until you know how you plan to end it. Don't start treatment unless you know how long you're going to treat. Don't assume that you can just give a drug, see what happens, and then let them stay on it indefinitely. That kind of treatment is anti-apocratic and often leads to more harm than good. Most current psychotropic drugs are symptomatic, not disease-modifying. The exception is lithium. Well, that's been the focus of this talk and the important take-home point I want you to have. Seven and eight, older drugs are more effective than newer drugs. Newer drugs are more tolerable than older drugs. Those are related. And lithium's a great example of this. A lot of people say, oh, it's so old. It's been around for 50 years. People don't want to use it. They think the newer drugs are better. Well, the newer drugs have less side effects often. That's how they're developed by the pharmaceutical companies. And I observed that myself and my own personal experience working inside Novartis for a few years. But they're not more effective. And if you look at it, clearly lithium's more effective than antidepressants, for instance, we've discussed now. But even like antipsychotics, the older ones like clozapine are the most effective ones. Number nine, treatment resistance usually reflects either misdiagnosis or an invalid diagnosis. And that's not relevant so much to the topic today, but it's a general point about the importance of diagnosis. When we often say someone's treatment resistant depression, we act as if, well, it's just that their depression is really terrible, difficult to treat. The problem may actually be that the diagnosis of quote unquote depression, quote unquote major depressive disorder is not biologically valid enough to be able to get a effective treatment to them. And the last thought, course not symptoms reveals the diagnosis. This comes from Emil Kreplin, great 19th century German psychiatrist who laid down the basic features of the scientifically valid diagnostic systems that we can define these days still. I, in my opinion, he famously said, prognosis is diagnosis. It's the course that tells you what the diagnosis is. It's not just the symptoms. The symptoms are really nonspecific and you need to focus on the course of illness to know what the diagnosis is, but also, and I've been emphasizing today, focus on the course of illness as a major focus of your treatment. That's the disease modifying effect. That's what you should be trying to do to try to help patients in the long run. And that's what lithium does the best. And so here are the summary of what we've discussed. And the take home point, like I said, is to focus on disease modifying, not just symptomatic treatment. With lithium as excellent, perhaps the best example of that in psychiatry. Thanks for your attention. Thanks so much, Dr. Ghami. Really interesting, thought-provoking presentation. So just quickly, before we shift into Q&A, I want to take a moment to let you know that SMI Advisor is accessible from your mobile device. Use the SMI Advisor app to access resources, education, and upcoming events, complete mental health rating scales, and even submit questions directly to our team of SMI experts. Download the app now at smiadvisor.org. Forward slash app. So we have several questions kind of building on your talk about lithium. So one, Luke Mesquita asks, regarding more not necessarily being better, can you speak more about lithium levels and how to determine a target level? Yes, so I think that we should, my view, this is my view, is that we should use lithium by targeting the clinical effect, not the blood level. This is my view, and it's something that's opposite of what I've been taught and practiced myself for many years. The blood level of 0.8, if you dig into the literature and look for the basis for it, is mostly for acute mania. It's not actually for acute bipolar depression. It's not for acute unipolar depression. It's not for maintenance prevention of mood episodes. It's for acute mania. And usually we're not using lithium for acute mania. It's also for bipolar type one studies. It's not in bipolar type two illness. So basically my view is, unless you're in the acute phase of type one bipolar, you should be much more flexible with how much lithium you're using, flexible in the downward direction. So bipolar type two patients, 0.6 to 0.8, even bipolar type one patients, 0.6 to 0.8 for maintenance is probably fine. And then for unipolar depressed patients, same. And then mood temperaments and other people also with formal depression, 0.4 to 0.6 can be fine. So my general view is that you should just dose it at 300, 450, 600, wait and see if the patient gets better. And if they get better, stop right there. Don't keep increasing the dose to a theoretical level. Dose it to the clinical effect. And the research is reasonably supportive of taking that view. Thanks. We have a question from Gulam Sumro. Can you comment on nephrotoxicity from lithium in the long term? Yeah, and we didn't get into side effects today, which of course is the next question you'd want to ask. 1% at 20 years is the number you should tell your patients. And that's what's been shown. The likelihood of getting chronic renal insufficiency with lithium is 1% at 20 years of treatment. It can be up to 5%, depending on the study. One to 5% would be the range. So the way I describe it to people is 95 to 99% of people can take lithium without any long-term kidney harm. Now that's good, but obviously the one to 5% is also a concern. That's been shown to be decreaseable in a couple of ways. One is to keep the dose low, which gets back to what we just said about no need for high levels. Keep the doses in the low, medium level if someone responds. Secondly, and very, very importantly, lithium has a 24-hour half-life and should only be dosed once a day. Once a day dosing with lithium has been shown to decrease the kidney effects long-term. Multiple daily dosing is very harmful, is more harmful. So you can reduce that one to 5% range to even less than that by giving once a day dosing and keeping the dose as low as you can. Thanks. That's very helpful. We have a question from Danielle Feagly. Can you please speak to the use and efficacy of lithium in children with mood episodes? Yes. Lithium has been studied in adolescents in particular and somewhat in pre-adolescent children. And the efficacy is pretty much the same. The levels are reported to be pretty much the same. So standard level 0.8 or so for acute mania and maintenance has been studied somewhat, though less. Most of the studies are for acute mania. Less studies for acute depression as well. My view is the same as an adult's. I think 0.8 is too high for most people, to be honest. And I think if you dose it at 300 milligrams, 450 milligrams at the level of 0.2 to 0.4 to a child with depressive states, for instance, or mixed states who has a parent who has bipolar illness or who have other reasons to think that they might be in the manic depressive spectrum, you'll be surprised how well they do. I think they do quite well with low dose of lithium in my clinical experience. But the research, like I said, is there more for the higher levels and in the standard kind of way. Great, thank you. Question, you had talked about having an exit strategy. So do you think of that as applying to lithium? Do you see lithium as something that you think of in the same way as what you were discussing as something that would be short-term or is that something also that you would have an exit strategy for? Well, yes and no. So the exit strategy concept is really to, again, emphasize the distinction between symptomatic and disease-modifying drugs. You need an exit strategy for symptomatic drugs. The disease-modifying drugs is, you talk about the exit strategy in the sense of saying, since this is a disease-modifying drug and it improves the long-term course of your illness, et cetera, you should stay on it long-term. So you're explaining why you don't have an exit then. You do have a strategy, the strategy is to continue it, but you're explaining why and you're giving a rationale for it. What I'm really arguing against is just the assumption that it's okay to leave things on without having a rationale or with just having this happily ever after fallacy that because you get better short-term, you should stay on long. But of course, the exit strategy that you should of course bring up is if they do develop the one to 5% kidney impairment in 20 years of treatment, that's when you start talking about cutting off. Thank you for clarifying. And then there's a question about suicide and if there is a response, given that there is a response, difference in response to lithium between attempters and completers, do you think that says anything about kind of the underlying psychopathology or problem in the two populations? Yeah, it's interesting. I'm not convinced there's a difference between attempters and completers if they have a mood disease. So for instance, even in the VA study, as I showed you, the bipolar patients had a decrease in suicide attempts. And I think if you could tease out the unipolar depression patients from the MDD patients, and I hope it might be clear now based on our discussion that I distinguish those two things. You might've been able to show some benefit there too, but it wasn't done because of the way MDD is defined. I think what the study suggests though is if you just take current DSM definitions and apply it to a high PTSD-like population, you're not gonna see a major decrease in suicide attempts. I'm not convinced you still won't prevent completed suicide if you were able to measure it. And that study actually showed some decrease in completed suicide. I think there is a difference. I think that even the 2013 BMJ study showed that this parasuicidality did not decrease with lithium. Lithium is not the kind of drug you give to somebody with borderline personality and they'll cut themselves less. It's not the kind of drug you give someone who makes many suicide attempts and they never have any others. But it is the kind of drug that if you give it to someone who's seriously thinking about dying and makes a few attempts, some of which are near fatal, it can keep that from actually proceeding to death. Now they might make more attempts, but maybe it will be less fatal. Maybe they'll intervene more quickly. I think it prevents death. It doesn't do it in a straightforward way, but I don't think suicide is a straightforward line. The suicidality isn't a straightforward thing. Yeah, thanks so much. Well, I mean, we still have questions popping up and I wish we had more time, but I think we're just about out of it. But I think that that reflects people's kind of real engagement around this issue. And just to paraphrase one, thank you from one of the presenters, to thank you for spreading the word about this relatively underutilized and talked about medication. So we just have a few more slides housekeeping. If there are topics that are covered in this webinar that you would like to discuss with colleagues in the mental health field, you can post a question or comment on SMI Advisor's webinar Roundtable Topics Discussion Board. It's a way to network and share ideas with other clinicians who are participating in this webinar. That sounds like it might be very good for folks. If there are questions about this webinar or any other topic related to evidence-based care for SMI, you can get an answer within one business day from one of SMI Advisor's national experts on SMI. The service is available to all mental health clinicians, peer support specialists, administrators, and anyone else in the mental health field who works with individuals who have SMI. It is a completely free and confidential service. SMI Advisor is just one of many SAMHSA initiatives that are designed to help clinicians implement evidence-based care. We'd encourage you to explore the resources available on the Mental Health Addiction and Prevention CTCs, as well as the National Center for Excellence for Eating Disorders and the Suicide Prevention Resource Center. These initiatives cover a broad range of topics from school-based mental health through the opioid epidemic. To claim credit for participating in today's webinar, you'll need to have met the requisite attendance threshold for your profession. Verification of attendance may take up to five minutes. You'll then be able to select Next to advance and complete the program evaluation before claiming your credit. Please join us on February 23rd as Dr. Susanna Evry-Palmer presents Clozapine and the Gastrointestinal Tract. Again, this free webinar will be February 23rd from 3 to 4 p.m. Eastern time on Thursday. Thank you so much for joining us. Until next time, take care.

psychopharmacology

fallacies

truths

lithium

symptomatic drugs

disease-modifying drugs

psychotropic drugs

preventing mood episodes

reducing suicide risk

limitations of antidepressants