Well, I want to thank people for tuning into this particular session. My name is Kimberly Yonkers, I'm from the Yale School of Medicine, and this session is How to Care for Pregnant Women with Psychiatric Illness. Let me start off with my disclosures. I receive royalties from UpToDate, and I've also received grant support from the National Institute of Drug Abuse, the National Institute of Alcoholism and Alcohol Abuse, the Centers for Disease Control, and the Doris Duke Foundation. So what we're going to be talking about today are strategies to manage pregnant women with psychiatric conditions. And one of the most common psychiatric disorders that women experience are major depressive disorders, or depressive disorders, and pregnancy is no different. We find that about 3 to 9% of women will experience an episode of major depression. So that's the full major depressive disorder, not simply depressive symptoms. So as you can see, this is not a rare occurrence. An additional 5% will develop a minor depressive disorder, which is very similar in terms of the duration and density to a major depressive episode, but has fewer symptoms associated with it. In addition to that, about 6 to 9% of women are depressed after they deliver. Interestingly, about one half of the postpartum episodes actually begin before a woman delivers. So clinically, what we see is we see women entering pregnancy as depressed. We see women becoming depressed during pregnancy. We see women improving during pregnancy in terms of their depression, and we see depression continuing to the point of delivery and starting after delivery. So it's not simply segregated to depression during pregnancy or depression after delivery. So you're a psychiatrist, you're a mental health professional, you've been following a patient, and they come in and they tell you you're pregnant. What do you do? Now, I do want to make a call out to a subspecialty which has emerged of psychiatrists who specialize in treating pregnant women with psychiatric conditions, but there are also a lot of things that general psychiatrists can do in terms of safely managing patients with common psychiatric conditions who become pregnant. Now, many mental health professionals feel like it can be a dilemma in terms of how optimally to manage a patient with a psychiatric condition who's pregnant, but it's also a dilemma for moms. You know, here they may have concerns about how their illness, the illness of depression or perhaps even another illness such as an anxiety disorder, can influence their baby and their baby's well-being on the one hand, and then they also have concerns about the treatments if they're pharmacological and what the pharmacological effect can have on their baby. The media can be informative at times but can also push out messages that can be problematic such as depression can harm your baby or antidepressants can harm your baby, which is why it's really helpful for all of us to have a sense of what some of the risks and what some of the benefits are because we'll need to answer those questions for many of our patients. Now, even though patients will ask us what they should do in terms of continuing their medications or stopping their medications, it's also important to know that many people are going to decide for themselves. And that's what you see in this particular slide. So here you see antidepressants and you see women who conceived right here, they were taking their antidepressant medications and then they stopped. And this is the rate of, here you see the total antidepressant treatment per week. So this curve reflects the fact that women have discontinued or in the process of discontinuing their medication. It's a bit different for antidepressants as you can see than for other conditions. So for example, antipsychotics we see on this and that has a lower rate in terms of discontinuation and antiepileptics are here. You see that women with epilepsy don't tend to discontinue their medication, although women who are on antiepileptics, such as women who have bipolar disorder, may be more likely to discontinue their medications. So the first thing to ask patients is if they're still continuing their medication. I actually said this is first things first, but perhaps it's really second. The other thing I think we need to get a handle on for many of our patients is what about the other habits that they have. People will come to see us and they'll be concerned about the medications they're prescribed, but they don't necessarily consider a lot of the other exposures that they have, which can be equally and sometimes even more problematic. And that may include over-the-counter treatments, but certainly it includes things such as smoking, which is very highly associated with delivery of a low birth weight infant, drinking, which can cause fetal alcohol effects, as well as fetal alcohol syndrome and a variety of perinatal complications such as preterm birth. And then other substance use, such as cocaine, stimulants, which are associated with intrauterine growth restriction, as well as preterm birth and placental abruption. Opioids are also associated with preterm birth. So we need to ask people about over-the-counter medications. We need to ask them about their health habits to really understand their overall risk. Now over half of pregnant women will take at least one prescription medication in pregnancy, and even more women will take an over-the-counter treatment. And that's why it's important to take a thorough history. The one medication that appears to generate most calls to teratogen information services, though, are the antidepressant medications. And so that's what I'd like to talk to or talk about first. We think about 5% of pregnant women receive antidepressant treatment during pregnancy in North America. As you saw from a previous slide, there are some variations across pregnancy. So the highest rate of antidepressant use is during that first trimester. Often people don't even know they're pregnant. Sometimes when the line turns blue, they discontinue it. But there certainly are women who need to take medication well into their pregnancy, and they do that. So what are the consequences of antidepressant medication use in pregnancy? And here, I'm focusing on the adverse consequences. Of course, the positive consequences are mood stability, as well as the ability to feel capable of even carrying a pregnancy to term. So even though I'm going to be talking about some of the adverse consequences, I think the flip side of the story that we also need to appreciate as mental health professionals is the fact that antidepressants work in pregnancy. They don't just stop working in pregnancy, and they can be very beneficial in pregnancy. They can be necessary in pregnancy. And in fact, when I do a consultation for somebody who is depressed or actually depressed and planning on becoming pregnant, one of the things I'll ask is what their history of medication discontinuation is. If they've tried to discontinue medication and they've relapsed, that is very useful information because it tells us as a mental health professional that this is somebody who maybe cannot stop their medication. I'll also ask them if they've tried other medication, and that's useful because some people can do well with a particular kind of medication, but they may not do well with another. And even though that other medication is less favorable in terms of the data we have for use in pregnancy, it may be that switching them to a medication that we like better or has more data or more favorable data isn't doing them any service because it doesn't work for them. So we also need to appreciate that medication can be very useful in pregnancy. We also can't forget that non-pharmacological treatments are very useful in pregnancy. And one of the things we also need to ask people is what they've tried in terms of cognitive behavioral therapy, behavioral activation treatment, all of the other psychotherapies that we have in our armamentarium that may be useful for pregnant women and we can encourage them to use during this particular time. And that's not just for major depressive disorder, it's for other conditions. So a balanced approach that discusses some of the benefits of treatment as well as the possible risks associated with treatment are very, very important for practitioners to emphasize. So let me move on to some of the adverse consequences that are reported in the literature in association with antidepressant use in pregnancy. So spontaneous miscarriage has been evaluated and the results are very mixed. Spontaneous miscarriage can be very difficult to get a handle on though because sometimes women will spontaneously miscarry before they even know that they were pregnant or very, very early on. So the reports that we have about spontaneous miscarriage tend to be the later spontaneous miscarriage. So even though they can occur in the first trimester, they're not before six weeks pregnancy, maybe they're eight weeks pregnant or 10 weeks pregnant, so they're later on. And that's just a subset of individuals who have spontaneous miscarriage. The data for fetal demise strongly suggests that there is no association between that outcome and antidepressant use in pregnancy. Preterm birth has been evaluated in a number of studies and I'll show you some of those. The studies are highly replicated and they do find an association between a variety of antidepressants in preterm birth, although the effects tend to be quite small and I'm going to review that. Use of antidepressants in small for gestational age or low birth weight, the results for an association are quite mixed. The positive studies still find very small effects, small associations. The better controlled studies tend to be negative and again, I'll show you some data about that. A great concern are major congenital anomalies, particularly these are anomalies that occur by virtue of first trimester exposure. Again, the results are quite mixed and I'll show you some of these results. There's greater consistency with paroxetine than with other antidepressants, but even with paroxetine, there's some inconsistencies. Persistent pulmonary hypertension is a disorder that happens in babies and can lead to right heart failure and even death, although there is a range of severity for persistent pulmonary hypertension, which can go from respiratory difficulties, but no demise to more dire outcomes. The data for this particular outcome are mixed, but they're better replicated for exposures, particularly to serotonin reuptake inhibitors later in pregnancy. Congenital adaptation syndrome is a syndrome where babies will have difficulty with respiration. They may have jitteriness, they have sleep-wake disruptions, and these findings are highly replicated in association with antidepressants. It's kind of not terribly dissimilar in some ways to the sort of discontinuation syndrome that adults experience if they abruptly stop taking some of the serotonin reuptake inhibitor antidepressants. It happens in adults, and I think it's fair enough to say that it happens as well in babies who may be exposed to abrupt discontinuation of an SRI. Autism is an outcome that's been explored. Again, the results are mixed, but weak. Some of the better controlled studies are negative. Other birth outcomes include preeclampsia and gestational hypertension. These are generally hypertensive diseases of pregnancy, and they have been replicated and found to be associated with use of antidepressants in pregnancy. Preeclampsia is an uncommon outcome, and although there may be a slight increase among women who use serotonin reuptake inhibitors in pregnancy, it's not a large increase. Attention deficit hyperactivity disorder is another condition, a later, not a perinatal condition, but a later condition that's been explored. Again, the results are mixed, but weak, and the better controlled studies are negative. Believe it or not, asthma in offspring has been explored and not found to be associated with antidepressant use in pregnancy. Now, these data are taken from a large registry study, from the Swedish Registry Study, and this is a very powerful registry. The way it works is that women who become pregnant book their prenatal care and their delivery. They're administered a very detailed questionnaire that asks about a variety of exposures and illnesses, as well as health habits. And then they are put into a database, and this database can be queried after delivery. And over time, the numbers of women in this large registry have accumulated, so when this particular paper was published in 2010, there were already 1.2 million women in the registry, and that included 15,000 women who had been exposed to antidepressants in pregnancy. In other words, moms took an antidepressant in pregnancy. Now, one of the issues with the registry is that the exposures really depend on how common a particular medication is being used at that time. So for example, you see in this particular table that the number of women who took a serotonin reuptake inhibitor in pregnancy was over 10,000, while only 1,600 women took a tricyclic antidepressant, because the field had moved to using more selective serotonin reuptake inhibitors rather than tricyclic antidepressants during that time. So numbers matter, and the number of exposures matter in terms of having sufficient power to detect an effect, and that is dependent on the number of people who are serendipitously taking a medication, if you will. The other limitation is that much of this information, in terms of exposures, is dependent upon self-report. It's nice that most of the information was collected at baseline, before people had a particular birth outcome, so we're not seeing bias because somebody had a bad outcome, they're going to report that they took X, Y, or Z medication. But we are limited to self-report, and people may under-report certain habits, for example, smoking or drinking, or even the use of medication. So we have some difficulty, particularly with exposures that are not documented in any database or medical record. Having said that, what you see outlined here are the outcomes for women who took, for example, a tricyclic antidepressant, a selective serotonin reuptake inhibitor, a serotonin norepinephrine reuptake inhibitor, and then specifically fluoxetine, citalopram, and paroxetine. So you see that tricyclic antidepressants were associated with any type of severe malformation as was fluoxetine. The odds ratio for a tricyclic was 1.36, and for fluoxetine it was 1.29. So that generates an increase of 36% for a tricyclic and 29% for fluoxetine. Now I do want to mention that these are model data, so they did take into consideration possible confounders that could contribute to effects, although, as I just mentioned, self-report limits the number, the accuracy of some confounders. When we look at any cardiovascular effect for tricyclic antidepressants, we see this was based only on an exposure of 30 people, or the outcome was in 30 people of the 1,662 were exposed. That may seem like a lot, but for some of these outcomes, which are uncommon, it can be a relatively modest exposure. Nonetheless, the odds ratio was 1.63, so it increased the risk by 63%. Paroxetine, in this particular data set, increased the risk by 66%. For ventricular septal defects or other septal defects, the odds ratio was 1.84 for tricyclics, for hypospadias, a very unstable number, but it did appear to be increased in paroxetine among women whose babies were exposed to, male babies, exposed to paroxetine in utero, and then cystic kidney was also increased for any SRI. We used to think that bupropion would be an alternative that would be reasonable, but then some data that were published in 2014 suggested that even bupropion could be associated with cardiac defects, which is what you see here, so the adjusted odds ratio for any antidepressant showed an increase of 30%, and this was really a marginal significant finding because you see the confidence interval is 1.0 to 1.6, but for bupropion, the point estimate was 2.5, so that's two and a half fold increased risk, and for an SRI, it was 1.3. Now, bupropion is also used to treat smoking cessation. The data for smoking cessation in pregnancy are pretty problematic because showing a positive benefit has been really hard. The rate of smoking tends to decrease overall as you traverse the course of pregnancy, and that may be one of the reasons why it's been hard to show an effect, so some of these data may be confounded by the use of bupropion for smoking, smoking travels as well with other substance use, but these are the data that we have for bupropion to date. Now, I talked a little bit about this, so this slide says, what's the opposite of confounding? And the answer is, from the professorial pigeon, is enlighten, explain, explicate, calm, comfort, clear up, clarify, bore, which I've been known to do, help, as well as expect. So I want to talk a little bit about confounding, and I want to use this study that was published in the New England Journal of Medicine 2014, because I think, especially when we hear reports in the media about this particular medication causing this particular effect, and having a lot of questions from our patients, it's really important to discuss confounding and the complexity of what we can really say for sure, because these are not typically controlled data, in the sense that they're not placebo-controlled studies that we're relying on. These are really serendipitous exposures, and we try and make inferences from the information that we have, but we know that confounding can occur in these data sets. And I think this particular study sort of illustrates that effect quite nicely. You see in panel A, which says the unadjusted analysis, we have exposure groups, and you have a whole variety of antidepressants, and you have the odds ratio on the right side. And to orient you, if the point estimate, that little circle on the line, is far over to the right, that suggests that there's an elevated risk. And you see that written in text. So for example, any antidepressant shows a significant confidence interval, so you can hang your hat on that finding with a point estimate of 1.25. Now when you go to panel C, what you see is a fully controlled analysis. So what this particular research group did is they took a whole bunch of potential confounders and they built the cohort taking into consideration what some of these confounders could be. So they considered, for example, whether smoking could obscure the outcome, or drinking obscure the outcome, or the underlying illness obscure the outcome and show a false association. And when you look at this propensity score stratification, what you see is a shift of the lines that you saw above. And you see that even though a lot of the point estimates are still, those circles are still to the right part of the dotted curve, there's a lot more uncertainty. So that 1.25 that we saw in panel A now becomes 1.02 with a non-significant confidence interval. What this is suggesting is it's suggesting that it's entirely possible that the effect that we were looking at for congenital malformations is not attributable to the antidepressant, but may be attributable to some of the factors that we see among women who take antidepressants. So maybe they're taking another medication. Maybe they have a history of congenital malformations and they were so worried about it, or there's a family history of malformations that they started to take an antidepressant. We just don't know. And I'm not trying to paint a positive picture or to apologize, to be apologists for antidepressants, but I just wanna provide some of the data so that we can explain this to our patients who may come in and be worried about the possible impact of an exposure on their baby. Now, this particular study took a different approach. They looked at malformations among depressed women who did or did not take antidepressants. So the main feature that they controlled for was the exposure of depression in and of itself. And they compared malformations among women who were depressed and took antidepressants to those who were depressed and did not take antidepressants. Now, this was a heroic effort. I think it's informative and these are the results. They had about just under 20,000 pregnancies. Tricyclics were associated with a number of malformations, including eye, ear, face, and neck and digestive system. Benlofaxine was associated with malformations in the respiratory system. Citalopram was associated with malformations in the musculoskeletal system in the skull and paroxetine was associated with cardiac malformations. The caveat in this is that women who are depressed and take antidepressants tend to have more severe illness. And so to compare them to women who are depressed and not taking antidepressants is somewhat imperfect. The association between use of other medications as well as health habits is also stronger for women who take antidepressants compared to and are depressed and women who are depressed and don't take antidepressants. So it's imperfect in terms of controlling for confounding. As well, this particular study did not control for substance use, so that could be problematic. Again, I think what we need to do in terms of informing our patients is number one, explain that things like smoking and alcohol use or other substance use can obscure or can show associations that may not be there, but in any case should clearly be avoided in pregnancy. I wanna also provide some backdrop for this and just to put this in context. I wanted to provide the results of some exposures to medications which are clearly considered teratogens to the serotonin reuptake inhibitors. And here you see, and I wanna actually thank Christina Chambers for providing this slide to me. But here you see valproic acid, which is clearly a teratogen. The outcome you see here is a major defect, so a malformation. The estimated absolute risk is some sort of malformation could occur in 11 women out of 100. We're not in the offspring of 11 women who take a medication out of 100 of those who take the medication in pregnancy or 11 out of 100 babies who are exposed. So for the acne treatment, isotretinoin, I'm sorry, I'm mispronouncing. For malformations of the heart, CNS, ear, thymus, the estimated absolute risk is 18 for babies who are exposed out of 100. For alcohol, which I know I've been harping on, but it's a very important exposure. Craniofacial outcomes, brain growth defects, as well as fetal alcohol spectrum disorders, it's 25 out of 100. And reuptake inhibitors, even assuming the worst case scenario, as I've showed you some of the less controlled data, it's three out of 10,000. So what are some of the clinical implications? The effects of antidepressants on birth defects, whether antidepressants, including antidepressants such as peroxetine, increase the risk of malformations remains somewhat controversial. Most associations are with ventricular septal defects, which are relatively common malformations. If the risks are real, the absolute risk is low, and it must be viewed in this context, which I provided, of whether the medication is needed. Other exposures, such as alcohol, may confound results, particularly in some of the registry studies that we read that may have very limited information about these health habits that are provided by moms. Now I wanna focus on, drill down into some of the data that I showed you in terms of the big picture, with regard to preterm birth, low birth weight, and other birth outcomes. So I mentioned this at the beginning of the talk, then we talked about malformations, and now I'm gonna just focus on some of these perinatal effects. So these are baseline and adjusted sibling comparisons of first trimester antidepressant exposure and preterm birth small for gestational age. That's a mouthful. Why is this important? So this is a very clever study that was trying to get at what I had been talking about in terms of confounding. So we have registry data. It can be difficult to attribute a particular outcome to a particular exposure, such as antidepressant exposure, because there are all of these other factors that are going on. So in this particular study, what researchers did is they used siblings to compare, to just see if they could pull out some effects that may be attributable to the exposure of antidepressants. Why siblings? Well, it's not perfect, but siblings tend to have a lot of the health habits, as well as a lot of the genetic risks that their sibling pair has. So if you can do a sibling comparison, it sort of controls for a number of genetic and even non-genetic confounding factors. So here you have the data for any antidepressant agent with an outcome of preterm birth. And this is a baseline uncontrolled analysis, or it was controlled through parity in year. And you see that if you believed this baseline data without the adjustments that I'm gonna talk about, the risk for preterm birth with taking an antidepressant is increased by 47%. And for SSRIs specifically, it's increased by 38%. But if you add a number of adjustments, so for example, parity, year of birth, maternal paternal country, age, education, psychiatric illness, criminality, suicide attempts, then you get an adjusted risk of 35%. And if you use a sibling comparison, it's pretty close. Increased risk of 32%. So there's still an increased risk, but some of these adjustments kind of fine tune what that risk is for any antidepressant. And you see under SSRIs, the adjusted is an increased risk of 27%. Again, the sibling comparison is pretty close to 33% increased risk. In terms of small for gestational age, so these are babies, they may have been born early, but they're born early after correcting for their gestational. I mean, they're small after correcting for their gestational age and the fact that they were born early, or they maybe have been born at term, but they're just small for being born at term. So you see the baseline shows a significant risk associated with antidepressant use or SRI use in pregnancy. But when you use the adjusted and the sibling comparison, you see that risk is far attenuated and it even becomes non-significant with antidepressant exposures. So what does this mean if we're talking to our patients? It means that this is a very, very unstable association between antidepressant use or SSRI use in pregnancy. And the risk of having a baby that's born small for gestational age. This is a study that looked at hypertensive, the risk of hypertension or hypertensive disorders in pregnancy associated with late SSRI use or SNRI use. Now they took a different, a bit of a different approach. And what they did is they looked at adjusted odds ratios for women who did not, when they looked at women who did not use an antidepressant in pregnancy, women who discontinued and those who continued and the risk of these outcomes. So you see that antidepressant, people who continued antidepressant until delivery had a higher risk for gestational hypertension or either gestational hypertension or preeclampsia. Preeclampsia is a very uncommon outcome and you need really, you need large sample sizes to actually show an effect. And this may have been the, any associations with preeclampsia may have been difficult to establish because of the sample size in this study, even though it's large, but anyway. So it appears that the association is stronger for people who stay on the antidepressant throughout pregnancy. What does that tell us? Well, one of the potential confounders is women who are sicker and need to be treated throughout pregnancy are gonna stay on it throughout pregnancy. So it's a little bit difficult to compare the discontinuers and the people who stay on it during pregnancy because they're kind of two different groups. Nonetheless, it does bring up the possibility that there may be an effect of the antidepressant on risk of gestational hypertension or even preeclampsia that is attributable to the pharmacology of the antidepressants. We don't know for sure, but it does appear that for whatever reason, whether they're ill, more ill, or they have had a higher load of antidepressant exposure that women who take antidepressants in pregnancy are greater risk of gestational hypertension, maybe even preeclampsia. So this is a maternal outcome that we're looking at. In terms of neonatal outcomes, the effects of an SSRI use, an SNRI use, or tricyclic antidepressant use increases the likelihood that somebody will be admitted to a NICU. It also appears to increase the likelihood that a baby will have some degree of respiratory distress or persistent pulmonary hypertension. And again, what we see here is there is a late versus an early effect. So those people who stay on it longer, these outcomes appear to be more likely. And here you see the 95% confidence intervals for these outcomes. They're all significant. So a 50% increased risk in the likelihood that somebody will have a NICU admission if they're taking a serotonin reuptake inhibitor, that's twofold, 2.2 fold higher for moms, for babies whose moms took an SNRI, and a 70% increased risk for moms who took, for babies whose mothers took a tricyclic antidepressant in pregnancy. Threefold higher risk for babies whose moms took an antidepressant in pregnancy with regard to persistent pulmonary hypertension and respiratory distress is 80% more likely in babies whose moms took an antidepressant in pregnancy. This again, we talked about this earlier, the poor neonatal adaptation syndrome. I actually warn moms who come in and are taking an antidepressant in pregnancy about this. And what I tell them is I tell them just as adults could be at risk of having some discontinuation symptoms, so can babies. And that they should A, be prepared that their baby may be monitored in a step-down unit, and that their baby may have some complications, particularly things like respiratory distress and tremors. But you see that they can also have jaundice or hypoglycemia, and that this could almost be expected. And this just prepares moms and dads for the fact that this could happen. I also try and reassure them that in most instances, these are benign and convulsions are not so benign, but they're also very, very uncommon. But the tremors or the respiratory distress are easily treatable, they're time limited, and there aren't long-term consequences to most of these. Now, I mentioned in my overview that people have looked at autism spectrum disorder and antidepressant exposure, and there was a lot of interest in the media about this outcome. And part of that, I think, is because we've seen an increased recognition of autism, or autism spectrum disorder in kids, and people have tried to understand what may be driving that. And it was hypothesized that perhaps the use of antidepressants could be contributing to that. This is a result of a, this is a meta-analysis that looked at a variety of studies, and you can see that what they also tried to do in this meta-analysis is address that confounding that I've mentioned over and over again. So let me just walk you through this. So the studies that just looked at maternal exposure during pregnancy and compared that to unexposed women, the adjusted risk ratio is 1.53, so that would suggest a 53% increased risk. But if you drill down and try and control for some of these confounding factors that I mentioned, the risk really is attenuated. So here you see maternal exposure pre-pregnancy versus unexposed women. Let me just reiterate, this is maternal exposure pre-pregnancy. So this is when moms took the medication before pregnancy compared to the unexposed women. The adjusted risk ratio was 1.48. So that's a 48% increased risk. Hard to attribute that to an antidepressant when the exposure was before pregnancy. This is maternal exposure during pregnancy versus unexposed women with a history of affective disorder. Again, you see this attenuated risk ratio. So the increased risk is 18%. If somebody, among the studies that use this sort of sibling design that I showed you, there's no risk. And this is even paternal exposure during the maternal pregnancy versus unexposed women. And you see an elevated risk ratio. So this is supposed to inform us about antidepressant exposure and autism risk. It does appear that there are some exposures that really are untenable, not biologically possible. Unless maternal exposure pre-pregnancy was confounded by not accurately reporting continuing medication in pregnancy. But this just shows you that the literature can be pretty messy. And you really need to have a sense of what's going on before we become alarmist for our patients. And this really does, I think, throw a lot of water on a potential exposure between antidepressant risk or antidepressant use in pregnancy and the risk of autism spectrum disorder in offspring. It's a very similar study with the outcome in this particular instance of attention deficit disorder. Again, what we see is this high risk when it's an uncontrolled maternal exposure during pregnancy versus unexposed. But then when you start to look at maternal exposure during pregnancy versus unexposed women with a history of affective disorder and the sibling design, you start to see that effect decrease. The paternal exposure is still there, but it's not due to moms taking a medication. It's due to paternal factors. So I've covered a lot. Let me just summarize the SRI exposure and effects of children. Exposed children are more likely to be admitted to a neonatal ICU or step down. Part of this may be the result of observation bias, i.e. docs know that mom was taking the antidepressant and they wanna be careful and just monitor the baby. But part of it could be this neonatal adaptation syndrome that I've mentioned. Tends to be more benign in terms of course, a few days and characterized by some difficulty with tremors and breathing, time limited. So I think this needs to be a discussion that we have with our patients and there can be some problematic outcomes and they need to be informed of that. Probably the most problematic would be something like persistent pulmonary hypertension. The longer term risks for kids in some of these outcomes such as ADHD and autism appears to be related to the population rather than the actual exposure to antidepressants. So if you have a patient who comes in and she's on an antidepressant and she found a media report or a study of this association, you can talk with her in an informed manner or her and her spouse or partner about what some of the literature shows with regard to these outcomes. Now I focused a lot on antidepressant medications but I wanna talk about some of the travelers to antidepressants that we often use. And some of these exposures happen before people even become pregnant. And probably one of the biggies is benzodiazepines. We don't use them as much as perhaps we did 10 years ago but women will not uncommonly be taking an antidepressant for depression and a benzodiazepine for sleep or they'll be taking both an antidepressant and a benzodiazepine to control an anxiety disorder. So these medications in and of themselves are associated with a number of adverse birth outcomes. Again, they could be effective at controlling symptoms and they may be needed in pregnancy but there can be problems with the use of some of them, particularly if they're used at high doses and daily throughout pregnancy. And these are some of the outcomes that have been shown to have increased risks. So cesarean delivery, low birth weight, neonatal ventilatory support, preterm delivery and miscarriage. So you see cesarean delivery is very common and that may sort of be under the radar screen because it's so uncommon. Low birth weight is not something that we find to be particularly highly associated with antidepressant use but there are some data including ours to show that low birth weight associates with benzodiazepine exposure. And then this neonatal ventilatory support, and this could be something as benign as CPAP or DLE. Rarely it's a stronger ventilatory support, but it can happen and has been shown to happen in association with benzodiazepine use. I wanna talk a little bit about the treatment of postpartum depression. First, I wanna talk about breastfeeding. Breastfeeding is recommended even in the setting of antidepressant use because it's such a great treatment for babies. It's such great nutrition for babies and it's natural. It fits where the baby is in terms of how far it is from delivery. Breast milk changes from one day after delivery to one week after delivery to months after delivery. So it's strongly recommended in its use for even moms who take antidepressants in pregnancy or after delivery. Now, not surprisingly, a lot of moms and dads are concerned about possible exposure to their baby when they're breastfeeding. And we can talk with them about the benefits of breastfeeding. We can also offer some suggestions if they're really, really concerned about breastfeeding their baby when they're taking an antidepressant. They can take their antidepressant and then pump and not use that feed, although that can be very difficult to ask moms to do, especially if they're anxious and depressed, but it is an option for people. Rather than monitor, you can monitor levels in breast milk. They tend to be rather low for most antidepressants, but rather than monitor levels, it's probably better to monitor the baby. And we're learning that now for a variety of exposures such as opioids in babies who are born to mom, who are taking opioids in pregnancy or exposed postnatally, that the baby should be monitored for problems in, we talk about eating, sleeping, and consoling, but feeding or sleeping. So if they're having trouble with feeding or sleeping or being consoled, then maybe it's time to think about whether the antidepressants have a contribution to that. Now they may not, but it might be worth erring on the side of safety if that's the case. Now there is a new sort of treatment frontier, if you will, for postpartum depression, and it may not be limited to postpartum depression. It may also be effective for depression outside of the postpartum period. And what we're talking about here are metabolites of progesterone. In particular, the one that's generated the most interest is what you see here, allopregnanolone, or 3-alpha, 5-alpha tetrahydroprogesterone. This you can see is a derivative initially of cholesterol, which is then metabolized to pregnenolone, the predecessor of progesterone, and then undergoes reduction and eventually to become allopregnanolone. These compounds are very interesting because they bind to the GABA receptor, the gamma-aminobutyric acid receptor. We call them neuroactive steroids, and they bind at this unique site here on the receptor. And what they do is they increase channel opening, and that can help modulate the amount of chloride that goes through the receptor and stabilize these neurons. Now, people probably have heard a lot about brexanolone, which is an analog of allopregnanolone, and it's been used in a number of studies now, several studies to treat postpartum depression. So I wanna talk about just a couple of the pivotal studies. Here you see a pivotal study that was published in The Lancet in 2018. This particular study enrolled women who had an onset of depression either in the third trimester or in the month after delivery. They had a minimal Hamilton rating scale of depression of 26, which is quite depressed. They underwent a 60-hour infusion of this analog of allopregnanolone, and then they were followed up to 30 days. And you can see the efficacy here. So brexanolone, 60 milligrams. This is the change from baseline in the Hamilton depression scale on this panel. So lower shows a greater degree of improvement, and you see that the least improvement was in the placebo group. The greatest improvement was in the brexanolone 60 group, and it was quite remarkable, the degree of improvement and the effect size from this small study. The caveats are there were some problems with presyncope in about 4% of the population, and what this meant is that the rollout of this infusion has to be monitored. So this has to be done in a monitored setting, kind of like ketamine infusions. So they have to be done in a monitored setting, and it can only be done in centers that are certified to provide this particular treatment. And then, of course, the infusion is 60 hours, so that's, they have to, moms have to be able to be separated for 60 hours from their neonate. I'm gonna move on to talk about bipolar disorder in pregnancy and after delivery. There's a lot of concern about how women who have bipolar disorder can be managed in pregnancy. They have lower rates of discontinuation of medication in pregnancy, and the medications to which they're exposed tend to have higher, or have less reassuring data in terms of their risks on babies. Now, many people probably heard about the lithium registry, which is very old now. It created a lot of concern about the possible impact of lithium on neonates. Then, the lithium registry was thought to be problematic because they kind of selected babies, or they were more likely to enroll babies who had bad outcomes. So it looked like the effects of lithium on, maternal lithium use on babies was exaggerated. And now, you know, there are some studies, and I'm showing you this really pivotal study from the New England Journal of Medicine that looked at the risk of adverse outcomes to babies who were exposed in utero to lithium. And really, and one of the things I like about this particular study is, not only did it look at lithium exposure, but it also looked at lamotrigine exposure to really try and get a handle on, is it the illness, or is it the medication itself? That could be contributing to some of the adverse outcomes. The adverse outcome in this particular study were cardiac malformations, as you can see, although there's some non-cardiac malformations that are mentioned as well. There were no effect for the non-cardiac ones. But you see here that this particular team also used that matching, propensity score matching that I talked about before to control for confounders, because women with bipolar disorder as a group tend to have higher rates of using cigarettes and alcohol, as well as other substances. So this is something that we need to control for, as well as other medications, I might wanna mention. So what you see here in this slide is a 2.25 increase risk. So over a twofold increased risk of cardiac malformations if baby was exposed to lithium during the first trimester of pregnancy. And for comparison, the risk for moms, for babies whose moms took lamotrigine was used as a reference and it wasn't elevated. This did not appear to greatly increase the risk of non-cardiac malformations, so it does appear to be something associated with cardiac malformations. What this translates into is an increased likelihood of one to two cases per 100 live births if this exposure is considered to be causal. And here you see the absolute relative risk of cardiac malformations among the lithium-exposed and lamotrigine-exposed infants as compared to unexposed. And this was stratified by mother's dose. And so I think this is also really important to point out that it was the high doses for which the risk was the greatest. So what we're seeing is no appreciable risk associated with lamotrigine, but when the doses were greater than 900 milligrams per day, this study would suggest the risk is increased. If the dose is 600 milligrams per day or less, the risk was not elevated. This is another study that was a meta-analysis just showing the overall risk among various studies that explored associations between cardiac anomalies and lithium use in pregnancy. And again, you see that there is overall an increased risk associated with lithium use in pregnancy. And for cardiac anomalies, the point estimate is about an 86% increased risk, somewhat lower than the study I just showed you, which is the study I just showed you is here. This is the overall risk. But I think that there's a certain level of consistency that we need to be mindful of. And that really merits some discussion with your patients. And it probably merits discussion with your patients before they even get pregnant. So women of childbearing age who are on lithium need to be, I think they need to be told that they have to be careful about contraception and they need to be mindful that there are some risks associated with the medications if they need to take them. They need to not stop their medications, but they just need to be cautious and plan ahead. These are some data with regard to attention deficit hyperactivity disorder in offspring who are exposed to anti-epileptic drugs in pregnancy. And it appears that here the major culprit is valproate. Although carbamazepine is pretty close, you see the adjusted hazard ratios here. So this is something else, of course, anti-epileptic drugs have also been associated with malformations. And that's a discussion that needs to occur with patients who have bipolar disorder may be taking, for example, valproate or carbamazepine less commonly. But there may be some longer term effects to these agents as well. This is a study that looked at postpartum relapse in women with bipolar disorder. And I think it really just underscores the notion that for our patients who have bipolar disorder, it's unacceptable to just discontinue their medication because they became pregnant and say, okay, bye, because they are at risk of relapse. They are very high risk of relapse. They're at high risk of relapse, whether or not they take medication and they're particularly high risk if they discontinue their medication. So we need to be mindful in this particular scenario that we're treating for two. This particular study showed that relapse occurred in 23% of women with bipolar illness who took medication and 66% of those who did not. So we need strategies to manage their illness and we need to inform them. We can't just say they can't be treated, but we need to come up with treatments that are as reassuring as possible. Mood stabilizers can be used in certain instances in breastfeeding. It's recommended for women or it's allowed, it's recommended if women are taking valproate or carbamazepine or lamotrigine, but not in lithium. And we have less information on some of the second generation antipsychotics, but they've been associated with some problems in babies like being tremulous and having difficulty feeding. Lithium equilibrates very, very well in placenta and breast milk. So babies will get high levels of almost maternal blood levels of lithium. So that's one of the reasons why the recommendation is to avoid that if possible. Now, I said we needed to come up with other strategies that we may be able to use for patients with bipolar disorder who either have conceived or would like to conceive. And part of that strategy may include antipsychotic medications. As well, there are patients with other psychotic disorders like schizophrenia who may become pregnant. So I wanna talk a little bit about antipsychotic agents. This was a meta-analysis that looked at the association between antipsychotic exposure during pregnancy and spontaneous abortion. And the good news is that there was no association. And I think others have similar findings. We do find that there is an association between the use of antipsychotic medication in pregnancy and preterm birth. This particular study, which I was part of, I would have to say is a strong study, but it doesn't have a lot of the controls for confounding that I've been talking about that we see with the antidepressant agent studies. So even though there is a risk, we can also say that there may be some confounding with this. I mentioned earlier that very, very valuable and large Swedish registry. And these data are from the Swedish registry looking at antipsychotic exposure, mostly the older antipsychotics. And you see that some of the outcomes were increases in gestational diabetes, low birth weight, and severe malformations. Although these findings tended to be significant, some of them are rather modest. And again, this is based upon self-report and it doesn't control for everything, although the models were adjusted for smoking and a history of miscarriage as well as age. Here we have a study that looked at absolute risk differences in child outcomes with antipsychotic use in pregnancy for other outcomes. So here we have preeclampsia, gestational diabetes, C-section, and perinatal death. The good news is that of these outcomes, the maternal outcomes, cesarean was the one outcome that was a standout. And in terms of major congenital malformations, there did not appear to be a risk, but again, the numbers may be too small to say. Preterm birth and low birth weight outcomes were found to be associated with antipsychotic use during pregnancy. Next slide. We do have some data now for some of the newer antipsychotics, which is really, really helpful. What's also nice is that we can use some of those same strategies to control for confounding to look at these data and some of the outcomes appear to be reassuring. So let me walk you through this again. So here we have some of the newer antipsychotic agents, and these are the comparisons of exposed versus not exposed. So babies who were exposed, babies who were not exposed. Here we have the same data set, but the outcomes were adjusted for psychiatric indications. So for example, the mom had schizophrenia, the mom had bipolar disorder, and we start to see, again, this shift in these curves toward no effect in a lot of these outcomes crossing this risk ratio of 1.0, saying that we can't hang our hat on these particular agents being associated with malformations. And then here's a fully adjusted model, really, and the only outlier in terms of an increased risk of malformations was with risperidone of these newer antipsychotics as well as the older typical antipsychotics. So the risperidone had a 26% increased risk of malformation in this particular large Medicaid database after controlling for a whole host of factors. There has been a risk of gestational diabetes associated with second-generation antipsychotic agents. And in this particular study, that's what was probed. But what you see here is here's the antipsychotic, atypical antipsychotic group, the reference group. Here we have the crude odds ratio and adjusted with that same stratification model that I just mentioned. And the good news is that there does not appear to be an effect of gestational, of antipsychotic, second-generation antipsychotic agent use in gestational diabetes. However, there may be a dose-related phenomenon, and that's what's shown here. So it may be that those individuals that are receiving the highest dose are at risk for gestational diabetes compared to those that are taking lower dosages. The implication is try and use the lowest dose if possible to try and minimize the likelihood that this could happen. I'm aware of other literature that suggests that women who enter pregnancy, who are taking antipsychotic agents, are at higher risk of developing gestational diabetes because they already have a number of the risk factors associated with gestational diabetes because they are already overweight from the use of these agents. So entering pregnancy with lower risk is also a strategy. So I just want to mention as I close out here, this has been a packed talk, and your head is probably spinning now, but I do want to mention some of the do's and don'ts when managing psychiatric illness in pregnancy. Some of the do's, ask about other substance use and encourage abstinence. Encourage prenatal vitamins. Some people will also encourage higher doses of folate. For example, the NICE recommendations from the UK encourage higher doses of folate. Encourage social support. People have a lot of worries about their illness, their medication. Make sure that they have the right information. Adjust medication for breakthrough symptoms. Does no good to have somebody on a medication if somebody is symptomatic. Adjust medication if needed in the postpartum interval. And then finally, some of the don'ts. It's best to avoid using lithium and peroxetine as well as anti-epileptic agents other than lamotrigine in the first trimester if possible. Try not to switch from one medication to another in the first trimester if possible. This can end up simply increasing the number of medications that a baby's exposed to. So for example, switching from fluoxetine to sertraline in the first trimester doesn't make a lot of sense. Possible exception is lithium as well as some of the anti-epileptic agents. They can also have effects that go into the early second trimester. Don't automatically stop all medication in pregnancy. Again, the possible exception is trying to find alternatives to lithiums and AEDs, but we don't wanna leave people who need medication in pregnancy without any protection. And certainly don't abandon a patient who comes in and tells you that they're pregnant just because you're uncomfortable treating them. At the very least, continue taking care of them until you can hook them up with somebody who can provide you a consultation or who is more comfortable treating them. So thank you very, very much for your attention. This has been a, I know it's been a long talk and it's been complex, and I sincerely hope that the information that I've provided to you is helpful in your practice with women who are pregnant and have a psychiatric condition. And with that, I'll stop. Thank you.

pregnant women

psychiatric illnesses

managing

major depressive disorder

bipolar disorder

treatment during pregnancy

antidepressant medications

antipsychotic medications

effects on the baby

mood stabilizers

breastfeeding