Major Depressive Disorder Podcast

The Role of Sleep in Major Depressive Disorder

Madhukar Trivedi, MD; Andrew Krystal, MD


August 18, 2022

This transcript has been edited for clarity.

Madhukar Trivedi, MD: Let me introduce Dr Andrew Krystal, who is at the University of California, San Francisco (UCSF), and is one of my colleagues, friends, and a great inspiration for me to understand many things. One of the major things I can learn from Andy is the role of sleep related to depression. Thank you so very much, Andy, for joining us.

Andrew Krystal, MD: My pleasure.

Trivedi: Let me give you a minute to describe the work that you do.

Krystal: Thank you, Madhukar. I appreciate the opportunity to be part of this effort. The primary areas that I work in are mood disorders and sleep disorders. The way I ended up doing those two areas is that they have a lot of overlap, in terms of people suffering from sleep disorders who have a lot of trouble with mood difficulties and people with mood disorders who suffer with sleep difficulties. It was natural to embrace those two areas for me.

Trivedi: Andy, people come to you and say all the time, "If only I could take care of the sleep, I would be okay." There is a close tie between different kinds of insomnia and depression. How do you disentangle them? How do you help patients and providers understand how best to think about that?

Krystal: That is an important and challenging issue when managing people in either spectrum, people who primarily present with a mood problem or with the sleep problem, because there's often ambiguity about which is the primary or driving condition. We used to think that insomnia was a consequence of a problem like depression or anxiety. In this case, we're focusing on depression. But the more time has passed, the more we appreciate that insomnia is sometimes a driving force for depression and that the relationships between the two are somewhat complicated. My primary approach, in trying to disentangle, [whether] is there a primary driving factor among the two is looking at timing of onset. In people who have years or months of insomnia and then develop depression afterward, that's often an indicator that the insomnia is independent or an important driving factor and not more of a symptom consequence. In some people, insomnia is a symptom where you develop depression, and it has onset at the same time as all the other symptoms they experience of anhedonia and sadness and so on.

Trivedi: This is your area of work, and this is the body of work that you built. So I'm going to push you a little so that the audience can understand this issue. Is this like diabetes and hypertension? Is it different? Is it sleep and depression, interwoven? Or are they two different things that are happening together?

Krystal: That's a hard question. I'll try to answer it with a little bit of discussion of a few important considerations. One is that I believe that there are fundamental links between mood problems and sleep/wake function, and we have some evidence that that's because that the circuits that modulate those functions are interrelated. They're either the same in some cases or overlapping or interconnected. We've been doing a study of deep brain stimulation where we have taken patients and implanted a number of electrodes in their brain and cortical limbic structures, and then did stimulus response mapping to look at which areas among about 100 that we look at are affecting depression — it's primarily depression study. What we find is that when we stimulate some areas and they improve mood, they reliably change arousal level. You stimulate, let's say, the ventral capsule area, sometimes the anterior limb of the internal capsule, and you get an enhancement in mood, but you also get an enhancement in energy and arousal. If you give that too close to bedtime, people can't sleep. If you stimulate regions of orbitofrontal cortex, you get an improvement in mood and anxiolytic effect, but it's sedating for them. If you give that to them in the middle of the day, they get sleepy or they get tired. There's some very clear, direct interlocking of these of these circuits.

The other thing we see is that separate from mood disorders, sleep disorders have symptoms that overlap with mood disorders in their presentation. Let's take obstructive sleep apnea, for example. These are people who have trouble breathing during sleep that causes arousal that happens all night long and hypoxia. They have not only daytime sleepiness, but they also have loss of concentration, loss of energy, and loss of interest in things because they're sort of tired and fatigued. You get a lot of points on depression scales from that, and for the patients with insomnia, among the most common complaints is loss of ability to concentrate and low energy, so there's a lot of overlap in symptoms of the condition. To me, there's some fundamental circuitry overlap, and then there is just some symptom overlap that makes it very complicated to tease the two out.

Trivedi: It's exciting that you're beginning to point out that it's the brain circuits that actually differentiate some of this, and we have to do the work in figuring this out, right?

Krystal: Exactly. So really teasing them apart, is it possible to separate them and to what degree? There are places that we do direct brain electrical stimulation, intracranially, and we don't see changes in arousal level. But very few of those have effects on mood. But there are some.

Trivedi: One of the most exciting things in having you here, Andy, is that you not only understand sleep and depression and understand the relationship, but you also have a huge program that is trying to actually stimulate different components of the brain circuits by doing deep brain stimulation. You just started by saying you stimulate one part, you get hypersomnia, you stimulate another part, you get insomnia. How does that really begin to get us to a point where we could actually find treatments?

Krystal: The purpose of this study we're doing is actually to find treatments for depression. But it's helping me now to think about treatments for sleep problems. The way we're doing this study is, we are recruiting people with very severe treatment-resistant depression. They undergo a neurosurgical procedure, and they have 10 depth electrodes implanted, each with 16 contacts. And then we do essentially trial-and-error stimulus response mapping to identify the very best site for treatment for that individual. So that means we stimulate them. We say, let's see what happens. And we get them to rate their mood. They get them the right level of alertness. We do depression scales, and we're finding that for individuals, there are different places we stimulate that are helpful. There's not just one, but the goal is that that kind of strategy can help us figure out a very personalized best treatment for an individual which. In this case, it has been done with deep brain stimulation. But our hope is to bridge on this methodology and transition it to noninvasive methods.

So far, we have been working with three people. We found three different sites that are the best. But each of the sites that we've been stimulating has tended to lead to some degree of mood improvement in the other people. It's just, for example, in one patient, the ventral capsule stimulation is the best treatment, but in another one, it's a good treatment but it triggers anxiety also, so they can't tolerate it.

One fundamental principle that this work has identified for me is that stimulating brain circuitry related to mood is ultimately a lot like stimulating the motor cortex. We think of the motor cortex as a deterministic sort of structure where we know what each part does, we stimulate it, we always get the same result. The left hand moves, every time. We're seeing the same thing with mood circuitry: You stimulate there, you get the same result every time. Whether it's beneficial or not tends to be a little more complicated because it seems to depend on patient state. Like if you're low energy and drowsy and you stimulate the orbitofrontal cortex, it's calming and it improves mood, but it makes people too slowed down and sluggish. But if they're energetic, when you do it, then they feel really good and they say, "This is an antidepressant for me, I love it."

Trivedi: I'm going to stop and interrupt you because I want to tell the audience how exciting this is and why Dr Krystal is the right speaker here. What you're identifying is the idea that no two brains are totally, completely the same. Maybe there are groups of people with brains that are similar, but that is really the critical thing. To be blunt about it, we in the whole mental health field have not invested in this idea as cancer has, as heart disease has. We are now getting to the point where people like Dr Krystal are helping us understand that if the National Institutes of Health (NIH) and people around the world don't invest in this idea, then they are not being very fair or even smart because this is the area where we need to invest.

Krystal: Well, I know you've always appreciated this point, that different people will need different treatments and work to try to understand which people to give what treatments to. But we've had limited ways to do it. This work I'm involved in now has convinced me 100% that that's true. In some ways it's sort of obvious, because no two people's faces are the same. We all have noses and eyes and mouths, but the shapes and their orientation are different. The brains are similar. We're finding that you need to be in exactly the right spot in the brain, because if you're not, it's just like if you were in the motor cortex area, it moves the hand instead of the arm or something. Because the anatomy is different, it's very hard to do things in a one-size-fits-all way. I'm more and more impressed at how important that personalization is based on this experience.

Trivedi: Thank you, Andy, for joining us, because I think what you're doing for us and our audience is elevating the idea that the brain is not the toenail.

Krystal: Right.

Trivedi: The second idea is that solutions are within our arm's reach. We have to do the work. Let's spend the next 10-15 minutes trying to use this information.

Your excitement — and you can see my excitement — in trying to figure out practically, how do psychiatrists, psychologists, and primary care providers figure out how to handle people who come to them with problems? They have problems that are related to sleep, they have problems that are related to sadness, anhedonia, et cetera. How do you clinically help people think about how best to approach that patient? I know it's an unfair question because there's a lot of uncertainty, but I'd love to hear your way of doing it.

Krystal: It's a great question, and I'm going to try to focus on sleep problems and mood problems as examples of this. When we're thinking about people with depression, I think the lesson that I've learned is take the time to identify the specific features of the depression in each person and try to make sure that you know what's unique about them. About 70% of people who are depressed report insomnia, and about 15% report hypersomnia.

Trivedi: I'm going to stop you there. As a sleep expert and with all the other expertise you have, it seems to me that in the 70% with insomnia that only 20%-30% of them need special attention for insomnia and the others don't. How do you think about that?

Krystal: Those people who come in who've had insomnia predating their depression — which many do, and often clinicians don't take the time to figure out whether this insomnia that emerged right at the time when all these other symptoms emerged or whether it was preexisting — those who had it as preexisting need special attention. That insomnia has a certain degree of independence. If you don't address it, the data suggest that it will impede the antidepressant response. A very similar story is if people come in and they are sleepy or with insomnia (but sleepy is particularly common), they may have obstructive sleep apnea that wasn't detected; that needs to be treated because if you don't treat it, it will prevent your response to antidepressant medications. There's about three studies now that show this. So it's worth it to take the time to figure out what was preexisting before, let's say, a mood episode started. There was a desire in the sleep community to be able to develop a story that suggests that we have to treat insomnia in people with depression, because it won't get better on its own.

I've done a bunch of studies treating insomnia in people with depression, and it's very clear to me that antidepressants fix a great deal of insomnia in people with depression. But it doesn't fix the people who have the independent significant insomnia or who have obstructive sleep apnea or something like that going on. It's important to take the time to figure out, is this an independent insomnia or did it emerge at the time of symptoms?

The other question is, does a person have a breathing-related sleep disorder? This needs to be treated for a whole variety of reasons for people's health, but it will also impede antidepressant response if it's not addressed.

The other part of this is that people often don't take the time to do careful individual assessment of things like, just what is the nature of a person's insomnia? We have a very common experience where I will talk to people and they'll say, "What is your go-to drug to treat insomnia?" And I say, "What do you mean? I don't have a go-to drug." Do we have a go-to antibiotic? It depends on what the infection is. So if a person has trouble falling asleep, there are some kinds of drugs that are helpful. If a person has trouble waking up in the middle of the night, there are certain things that are helpful. If they have a trouble with insomnia because they have a circadian rhythm problem where they tend to stay up late and sleep late, but they're trying to go to bed earlier than what their best biological time to do so is because they need to get up for school or work, that needs a whole other treatment. It really is important to take the time to figure out what the nature of the problem is, and then you can attack it with a treatment that's targeted to the specific kind of problem they have.

One of the most common errors I see is people saying, "Oh, I use Ambien [zolpidem] to treat the problem." Ambien is approved for helping people fall asleep, but it doesn't help people sleep toward the end of the night. It just doesn't last long enough and it's not going to. You really can't use a one-size-fits-all approach for this.

An important lesson in this to me is figuring out what is this particular patient's nature of their problem. In general, if you think about treating insomnia in people with depression, you have three choices. One is you can institute an antidepressant, a standard first-line antidepressant selective serotonin reuptake inhibitor (SSRI) or a serotonin and norepinephrine reuptake inhibitor (SNRI). And if you believe that the insomnia is a symptom of the depression, which is sometimes the case, it'll probably get better with the drug. But otherwise you're going to need to institute therapy for the insomnia, and you have the choice of either using a sedating antidepressant or using insomnia therapy along with a nonsedating antidepressant. And really, at this point in history, the only one that's in common use that's a sedating antidepressant is mirtazapine. It's a sleep agent, basically, in its antidepressant dosage — typically in the 15- to 30-mg range. As you get higher, it may lose some of its sleep-enhancing effects.

Trivedi: In your remaining time, I'd like to address two things. One is easy and one hard. Okay? I don't always get a chance to have a Dr Krystal on my podcast, so thank you for being here.

Krystal: Of course.

Trivedi: The first one is if a patient whom you see has depression and is using an antidepressant but has significant insomnia that is probably not responding to the antidepressant or is independent. How do you address this?

Krystal: The first question is, how do I manage insomnia in people with depression? I just want to share some findings in the literature that are useful for guidance in terms of important principles.

Trivedi: By the way, the findings in the literature pale in comparison to the findings you have published.

Krystal: Well, some of these things that I have actually published are in the literature! I've done three studies of cognitive-behavioral therapy for insomnia along with antidepressant medications. I had great hopes that that was going to be a combination that improved depression outcomes, and it doesn't. But we now have three recent studies that are a very exciting body of literature growing from a number of labs showing that if you have a person with insomnia and you treat their insomnia with cognitive-behavioral therapy, it prevents future emergence of depression. So it does have some efficacy. One of the studies where I did cognitive-behavioral therapy for insomnia along with an antidepressant had an arm in it where cognitive behavioral therapy alone was given for people with depression and insomnia. And it worked just as well as SSRI medication for depression. So the bottom-line take-home [message] is, if you give a person an SSRI or an SNRI, giving cognitive-behavioral therapy on top of it will improve sleep some, but it doesn't improve depression outcomes. But it itself is an antidepressant that seems to have effects around the same degree and perhaps mechanistically overlapping with an SSRI. So it's a viable approach, but it doesn't add to SSRI.

But there is evidence that there are some medications that augment SSRI response that are sleep-enhancing. The strongest data for this are with the medication eszopiclone. Eszopiclone is one of a series of drugs that are GABAA-positive allosteric modulators like zolpidem, zaleplon, and all the benzodiazepines, and is US Food and Drug Administration (FDA)–approved for treatment of insomnia. But its pharmacology is a little different than zolpidem and zaleplon. Eszopiclone enhances a subset of GABAA receptors that seem to probably have some independent antidepressant effects. We did studies with people with depression and insomnia who were given an SSRI, and we added eszopiclone and we got a better antidepressant response. The same thing happened with general anxiety disorder (GAD). The exact same studies were done with zolpidem, and no benefit [was seen] in depression or anxiety. Its pharmacology is different, and my guess is it's probably got something in common with brexanolone, which is also a GABAA- positive allosteric modulator and has antidepressant effects. It's a different subset of GABAA receptors that it's hitting.

Trivedi: In the interest of time, let me just finish with the following two things. One is a very difficult question, but the reason it is something I would like you to comment on is that anxiety, suicidal ideation, and depression in teenagers in this country has been consistently going up for the past 10-12 years. That is unlike what we see in the rest of the westernized world. If you look at the data in Europe, they use treatments for insomnia, for depression, and for anxiety for these kids, and we don't. We don't use benzodiazepines; we don't use the sedative-hypnotics for teenagers. So this is a difficult question, Andy, but I'd love for your idea about how you would want the world to change. How do you think we should deal with the 16-year-old who comes in with some anxiety and some insomnia who doesn't have an anxiety disorder, doesn't have depression, and doesn't actually benefit from an antidepressant. How would you treat that, and what are the risks of using something like a benzodiazepine?

Krystal: You raise a question that has a lot of important implications. I'm going to use insomnia medications as an example for this. We tend not to use insomnia medications in children. In a way, it turns out that they're used quite a bit. They're just used without any data and with not such good logic. People tend to use atypical antipsychotics more than they use benzodiazepines in children, and the safety profile of the benzodiazepines is probably better. My sense is that what we need desperately is studies of benzodiazepines and other GABAA-positive allosteric modulators in children and other insomnia agents in children, so that we know what their risk-benefit is. We just don't know. My strong sense is that it's important to treat these symptoms.

Trivedi: But do you feel like it would be a worthwhile exercise to treat aggressively this insomnia and anxiety with whatever it takes? Or should we be cautious?

Krystal: There are now about 50 studies that show that having insomnia at one timepoint is a risk for suicidal ideation, suicide attempts, and completed suicide across different age ranges and populations controlling for depression, not controlling for depression, on all the continents on earth, except maybe for Antarctica. We now have some data in adults that treating the insomnia improves suicidality compared to just using an SSRI alone. The study has not been carried out in adolescents or kids, but I'm hoping to do that. We tried to put in NIH grants and so far haven't been successful. I feel strongly that we don't want to ignore it.

Suicide in teens is such a huge epidemiologic and societal problem that we need to really work very hard to address it. One of the most important ways to do it is to aggressively treat the symptoms that are associated with it and aggressively identify people at risk and intervene. You use the term "aggressive"; I would say it's worth making sure to be very vigilant for symptoms and to treat them effectively and in the safest way possible. I'm not advocating for irresponsible management, but very careful, responsible, evidence-based management. But then this comes back to the huge problem we have, which is that we don't have data in kids for lots of the important therapies that we have available that have shown benefit or have good risk-benefit profiles in adults.

Trivedi: Thank you very much. Final thoughts on how best to make sure that we do not ignore sleep problems in preference to anxiety or depression?

Krystal: It's important whenever you have a person who presents with anxiety or a mood disorder to take a sleep history. Everybody says that. And I know there's plenty of reasons why we don't do that in practice. One of them is that it sometimes takes a while to take the history, but there are relatively brief means to identify sleep and characterize sleep problems. You can use some scales that may be helpful with this, like the Insomnia Severity Index, a self-reported, very brief scale that can help detect when significant levels of insomnia are present. It's extremely important to do so and treating the insomnia has a likely benefit not only just for people's function and quality of life, but in many people improving their outcome for depression and hopefully decreasing the likelihood that they'll attempt to end their life or actually do so.

Trivedi: Thank you so much, Andy.


Biomarkers of Depression and Treatment Response (SUNSET)

Customized Brain Stimulation: New Hope for Severe Depression

Closed-Loop Deep Brain Stimulation for Major Depression (PReSiDio)

Insomnia in Patients With Depression: A STAR*D Report

Prevalence of Obstructive Sleep Apnea in Suicidal Patients With Major Depressive Disorder

Sleep Disorders as Core Symptoms of Depression

Treating Insomnia in Depression: Insomnia Related Factors Predict Long-term Depression Trajectories

Mirtazapine Tablet - Uses, Side Effects, and More

Cognitive Behavioral Insomnia Therapy for Those With Insomnia and Depression: A Randomized Controlled Clinical Trial

Efficacy of Cognitive-Behavioral Therapy for Insomnia Combined With Antidepressant Pharmacotherapy in Patients With Comorbid Depression and Insomnia: A Randomized Controlled Trial

Cognitive Behavioural Therapy for Insomnia (CBT-I) to Treat Depression: A Systematic Review

Eszopiclone - Uses, Side Effects, and More

Eszopiclone Co-Administered With Fluoxetine in Patients With Insomnia Coexisting With Major Depressive Disorder

brexanolone (Rx)

Sleep Disturbances as an Evidence-Based Suicide Risk Factor

Reducing Suicidal Ideation Through Insomnia Treatment (REST-IT): A Randomized Clinical Trial

Insomnia Severity Index (ISI)

Follow Medscape on Facebook, Twitter, Instagram, and YouTube


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.