Transcript edited for clarity.
Hello, and welcome to the Carlat Psychiatry webinar, where we’re going to talk about the treatment of depression in pregnancy.
After taking this webinar, clinicians should understand how to:
- Identify risk factors for depression in pregnancy
- Reliably screen for depression among pregnant women
- Weigh the risks and benefits of various treatment options for depressed pregnant women
- Know how to select the most appropriate treatment.
Hello. My name is Dr. Victoria Hendrick. I will be talking about the treatment of depression during pregnancy. I’m a clinical professor at the department of psychiatry at Olive View UCLA Medical Center and I’m the Editor-in-Chief of The Carlat Hospital Psychiatry Report. I have no conflicts to disclose.
Today’s learning objectives are to identify risk factors for depression in pregnancy reliably screen for depression among pregnant women, weigh the risks and benefits of various treatment options for depressed pregnant women, and know how to select the most appropriate treatment.
To begin with, it’s important to know that pregnant women experience rates of depression, at least as high as non-pregnant women. About 10-14% of pregnant women experience depression. There’s often a notion that pregnancy is a time of quiescence of psychiatric illnesses and particularly of depression, from the belief that a pregnant woman must be so happy that she’s pregnant. But unfortunately, the rates of depression are, are not lower in pregnant women. Some women are at particularly high risk; risk factors include a poor relationship with her partner and intimate partner violence. Prior history of depression is a big risk factor. Lower socioeconomic status, lack of social support, and having an unintended pregnancy are additional risk factors.
What are the risks of untreated depression on pregnancy outcomes? There may be inadequate follow up with prenatal care. These women might be less motivated to care for themselves and their pregnancies. Depressed pregnant women are more likely to use drugs and to smoke, which we know might adversely affect the pregnancy. A colleague did a study a few years ago where she found that depressed women were more likely to terminate an otherwise healthy pregnancy that they would have otherwise wanted because of their depression.
Depressed women are also more likely to have low birth weight babies and babies that are preterm birth, especially when the depression is combined with anxiety. There is also a higher risk of suicide, although rates of suicide and self-harm generally drop in pregnancy to a little over half the rates of nonpregnant women. So, we do see that pregnancy can be somewhat protective against suicidal thoughts and self-injurious behaviors.
Older non-Hispanic white women appear to be at greater risk for suicide than other groups. A recent study found that of nearly 12,000 pregnancy-associated deaths, over 5% were due to suicide. That might not sound like a lot overall, but that still translates to hundreds of cases of suicide. And unfortunately, perinatal depression and even suicidality are frequently missed. The American College of Obstetrics and Gynecology, the American Academy of Pediatrics, and the American Psychiatric Association recommend screening for depression for all pregnant and postpartum women.
One of the simplest and most reliable screening tools is the Edinburgh Postnatal Depression Scale. It is cross-culturally validated across a variety of cultures and groups from around the world. It’s brief at only 10 questions and it is very easy to use. And it’s twice as effective as a clinician’s interview in detecting depression. You can easily find this scale on online.
Keep in mind when you screen for perinatal mood changes, many symptoms like fatigue, changes in sleep, and appetite are present normally in most pregnant women.
Also keep in mind that new mothers may be embarrassed to admit that they’re depressed because of a societal stigma against the notion that a pregnant mom might be depressed. I find that it helps to normalize mood changes in pregnancy and let patients know that it’s not at all unusual for pregnant women to experience some mood changes, some depression. I find that that helps women feel a lot more comfortable, sharing how they feel when they realize that it’s not an unusual situation to be in.
It’s also important to ask about domestic violence. Ask open-ended questions, like, “Tell me about your relationship.” It’s important to maintain a nonjudgmental manner when asking these questions, make it clear that these are routine questions you ask all your patients. And like I said earlier, normalization is important, reassuring women that their condition is common, that there are effective treatments, and they are no less of a mother for experiencing depression is one of the most important things that we can do when we’re working with pregnant women.
So how can we treat depression in pregnancy? There are four basic treatments: antidepressants, ECT, transcranial magnetic stimulation, and counseling.
Beginning with antidepressants, the use of antidepressants is surprisingly widespread in pregnancy. A study from a couple of years ago found that 6.6% of women used an antidepressant at some point in their pregnancy, and the number is expected to rise.
So, what do we know about antidepressants in pregnancy? Most of the data has been with serotonin reuptake inhibitors, and many studies have reported perinatal complications in newborns when they’ve been exposed to SSRIs in the third trimester—not the first trimester; some studies compared first trimester to third trimester exposure to serotonin reuptake inhibitors, and it was only the third trimester exposure that was linked with this poor neonatal adaptation syndrome. It seems to be something related to this late pregnancy exposure. The symptoms are typically respiratory distress, tremor, poor feeding, lower Apgar scores. You might remember Apgar scores are those assessments of newborns in the minutes after birth that look at things like skin color, muscle tone, and breathing. Exposure to SSRIs near term is also linked to a higher likelihood of a neonatal intensive care unit admission.
The good news is symptoms are short-lived. They last at most four days and in most cases, they only last a few hours. They seem to resolve fully with no sequelae, but if you work with pregnant women who are depressed and you prescribe serotonin retake inhibitors, you will see these symptoms in the baby because the incidence is high at 15-30%.
These symptoms might represent toxicity, ie, the baby is born with this medication that is no longer being metabolized by the mother, and now they have to metabolize it. It could be that, or it could be withdrawal now that they’re no longer exposed to the medication.
It’s not clear what it is, but the, the bottom line is it seems to, to resolve fully. One approach that I will tell you about, that some clinicians have used and that’s a little controversial, but just so you know all the options, is to reduce the dose of the antidepressant as the mother gets to term, as she gets close to her delivery date so that the baby is born with as little medication on board as possible.
And this does seem to help reduce the incidence of poor neonatal adaptation syndrome. The critique of this approach is that, Hey, you’re reducing the antidepressant right when the mom is going into the postpartum where she could be at risk of postpartum depression, and that’s a valid concern. On the other hand, postpartum depression doesn’t start right away after delivery.
It does not normally emerge until about 10 days to two weeks postpartum. New moms might have the baby blues. That’s very common in the first few days, postpartum where they’re a little moody. They describe symptoms similar to PMS. They feel a little irritable, a little jittery and tense.
We don’t want moms to have symptoms of any sort, but baby blues are very common and they’re short-lived. What we really want worry about is postpartum depression. And we want to avoid it as much as we can. So that is a risk about reducing the dose of medication as a mom gets close to term.
I like to present the choices to the pregnant moms, saying these are the pros and cons of this approach versus that approach. In my experience, women have generally preferred to reduce the dose of medication, knowing they’re slightly increasing their risk of postpartum depression, but also knowing that most likely, if we go back on her full dose, as soon as the baby is born, that will most likely protect her from postpartum depression.
But some women who have had postpartum depression in the past will say, “There’s no way I want to risk having another episode like that; I’m just going to stay on the same dose. If my baby has these symptoms, at least I know they’re going to resolve quickly.” So, all these decisions are basically you and your patient weighing the pros and cons of the various options because there’s no perfect solution.
You’re always weighing one set of risks and benefits versus another. Regarding antidepressants and birth defects, there have been tens of thousands of exposures and the rate of birth defects appears to be comparable among women who are depressed on serotonin retake inhibitors versus women who are depressed and unmedicated and women who are not depressed and not medicated. So, it doesn’t seem like most serotonin retake inhibitor antidepressants are linked with higher risks of birth defects with one exception, which is paroxetine.
Paroxetine has been linked with a higher incidence of cardiovascular defects, at nearly twice the rate of other SSRIs. There’s been some controversy about whether the rate is really that high, but nevertheless, the FDA has made, paroxetine the one SSRI that has a category D in pregnancy rating. And if you have other choices, why use paroxetine?
Other antidepressants that have been studied include duloxetine; it seems safe, although there’s not a ton of data. Venlafaxine has been found in two studies to be linked with a higher rate of cardiovascular defects and other miscellaneous birth defects, but the absolute risk was low. Bupropion has been linked with an elevated risk of miscarriage or spontaneous abortion and maybe a higher risk of cardiovascular defects. But the data is mixed. It’s important to keep in mind that bupropion is also marketed as Zyban for smoking cessation. And we know that women who smoke are at higher rate of spontaneous abortion and of having babies with cardiovascular defects. So that really muddies the data on bupropion. I, I think bupropion is going to turn out to be safer than what these data show, but until we have more data, it’s best to stick with the medications where we have more reassuring data, like the serotonin reup reuptake inhibitors with the exception of paroxetine.
Regarding mirtazapine and trazodone, we have too little data to make any recommendations.
A topic of great interest to me that I did research in for many years was what about prenatal serotonin reuptake inhibitor exposure and child development. Overall, there’s very little data. When we look at 4-year-olds who were exposed to SSRIs prenatally and compared them to children without this exposure, there have been some interesting findings that externalizing and internalizing behaviors do not differ between the groups, externalizing behaviors you might remember include acting out, being aggressive and impulsive, while internalizing behaviors include being shy and withdrawn. The factors that were associated in the children with these, behavior were maternal depression and anxiety at the four-year follow-up and parental stress, underscoring the importance of ensuring that our patients are as stable psychiatrically as we can help them be, especially if they’re going to be going home to take care of children because of the potential adverse impact, not only on the mother, but also on her child, if she’s experiencing depression and anxiety.
Some take home point—there are now tens of thousands of published exposures, mostly of serotonin retake inhibitors, and there is no consistent evidence that prenatal use of these medications increases the risk of major malformations except for paroxetine and possibly venlafaxine and bupropion.
Third trimester use is associated with a 15-30% incidence of neonatal complications, particularly respiratory distress that appears to be short lived and to have no long-term sequelae.
It’s currently uncertain what the impact of antidepressants is on newborns’ birth weight and gestational age. Some studies have found that women on antidepressants have babies with lower birth weight and briefer gestational age, but most studies have not found these adverse outcomes when they control for depression.
Until more data are available for newer antidepressants, it’s best not to use them for pregnant women, and whenever possible, prescribe the lowest antidepressant dose for the least time necessary. You’re not aiming for perfection in pregnancy. You’re just aiming for your patient to be able to function; to not be suicidal, of course; to be able to eat adequately and take care of herself; to make it to her prenatal appointments; and to be getting along with her partner. Sometimes one thing we’ve seen with depression is it can lead to higher rates of conflict within a relationship. We want to make sure that baby is born into as stable and happy a household as possible. But I aim to prescribe the lowest possible dose, at least until the baby is born.
Talking about doses, it’s surprising how little we know about the relationship of antidepressant dose with adverse outcomes in the infant. You would think a higher dose would be linked with more side effects, but that’s never been studied to my knowledge. But it has been studied with lithium. So, I’m going to tell you about the data with lithium. There are a couple of studies and I’ll show you data on the most recent study, which found that the rate of cardiovascular malformations—which we know is a potential risk with lithium exposure prenatally—tripled with doses above 900 milligrams a day, when compared with doses under 600 milligrams a day and maternal lithium levels of under 0.64 mEq per liter. At these lower doses, the risk of cardiovascular malformations was comparable to that of unexposed newborns.
I try to keep the dose as low as possible while making sure the mother is feeling well, and of course always discussing treatment decisions with the mother and getting her input in the process.
Another key point when using medications in pregnancy is to emphasize how little is known about developmental outcomes following prenatal exposures. We still don’t have information about formal assessments of infant and child development. It’s tricky to assess developmental outcomes in children exposed to anything in pregnancy because development in young children is so varied. A young child might not be speaking much, or walking compared to his or her peers. And we worry, and then a few months later they’re totally caught up developmentally. Children develop at their own speed for their first three or so years of life. Testing children at five years gives more reliable information because at that point, if a child is delayed, then there is reason to be concerned.
But the problem with testing a child at five years is that there are so many exposures a child has had in all those years: how many books are there in the home? How many other siblings? There are so many other factors that affect development so how can you definitively know it was because of an exposure that happened years earlier in pregnancy versus something else that happened in the child’s life.
So, you can see how it’s tricky to make determinations about child development following prenatal exposure to any medications, not just antidepressants.
Some more key points in pregnancy are that you want to use pharmacologic options in pregnancy only once you’ve made a good attempt at nonpharmacologic options; you’ve tried counseling, you’ve tried stress reduction.
Some patients may be able to stop their medications prior to conceptions. If your patient wants to try to get pregnant, you could try to reduce and then discontinue the medication while she’s trying to conceive. And then if she gets pregnant, hopefully easily and quickly, then the time she would’ve been off medications would hopefully be fairly brief. And then she can see if she can make it through the nine months of pregnancy medication-free. If not, you can always reinstate the antidepressants, but if you can at least get through the first trimester, that would be worth trying.
But in my experience here in Southern California, the patients we often see who want to get pregnant are often professional women in their late 30s, early 40s, and a concern with stopping medications prior to conception is unfortunately for this population of women, it can take them a long time to get pregnant.
Many are unlikely to get pregnant within a month or two of trying. So, if you have them off antidepressants for months, then you’re putting them at risk of relapse all that time.
An approach that is important to know about is that the placental circulation and embryological exposure does not begin immediately upon conception. When the woman conceives and that fertilized egg is traveling up the fallopian tubes and implanting in the uterine lining, early on there’s basically no medication exposure. The fetal placental circulation takes some time to get established. The very initial formation of the connecting stock, which later becomes the umbilical cord happens about 16-18 days post-conception. So, there’s minimal embryologic exposure if the medications are discontinued around the time when a pregnancy test turns positive, which even the over-the-counter pregnancy tests are so sensitive that by 12-14 days post-conception, a woman will know she’s pregnant. And if she were to stop or quickly taper her medications at that point, there would be minimal exposure to that embryo.
This is a helpful technique to use for women who you worry might not get pregnant quickly, so you can keep them on the antidepressant as long as possible, then try to get them off the antidepressant while they’re in the first trimester. And in my experience, this works very often that we’re able to get women through the first few weeks of pregnancy.
And that’s all we really want. Just get through those first few weeks. And then if the depression starts to come back, then we don’t worry quite as much once she gets to the 10th or 12th week of pregnancy.
Keep in mind, if you stop antidepressant medications in a pregnant woman, woman, there is going to be a risk of relapse. I was part of a study that came out several years ago that found there was a 68% relapse rate among women who stopped antidepressants versus 26% among women who remained on the antidepressants.
But was a naturalistic study that followed women who decided on their own to stop the antidepressants. No other treatments were implemented. The women who stopped the antidepressants weren’t being given extra counseling or stress reduction techniques to help minimize the risk of relapse.
It’s important to warn your patients that if they stop the antidepressant, they are facing a higher risk of relapse. And if they’re going to choose to go that route, try to have as many other interventions as possible in place to help minimize the risk of a relapse.
Now, what about benzodiazepines in pregnancy? Many of us who work with pregnant women know that a lot of them are going to experience anxiety and insomnia, and about 15% will experience a full-blown anxiety disorder during pregnancy. What are the risks of using benzodiazepines in pregnancy? There are some risks, namely neonatal sedation and hypotonia where the babies are born sleepy, and floppy infant syndrome and withdrawal can happen in the newborn. We recommend tapering these medications prior to delivery whenever possible, so that the baby is born with as little benzodiazepine on board.
What about birth defects? There have been some reports of elevated risk of cleft palate, but the data are inconsistent. The absolute risk, if it exists, appears low. For example, a recent study found the risk ratio was 1.1 for diazepam, for orofacial defects like cleft palate.
The risk I worry about is when the baby is born, because of the baby’s been exposed to benzodiazepines close to the delivery date, then the baby might be born sleepy. If you use a benzodiazepine, use something medium-acting like lorazepam rather than long acting like clonazepam. Lorazepam also has the advantage that it has no active metabolites.
What about any long-term behavioral sequelae in children? There were some mixed data from many years ago that maybe exposed children had lower IQs, but that has been found to not be the case, there’s no compelling evidence from the data we have. Again, we have very little data on long term neurobehavioral sequelae with any psychiatric medication, but what data we have shows no compelling evidence of behavioral sequelae in exposed children.
Some more key points: when you work with pregnant patients, your documentation must include that you’ve reviewed risks and benefits with the patient and anyone else involved in this pregnancy, the father, any other family members that the patient might want to get involved. You also want to document the patient’s symptoms are improving with the medication. Otherwise, why are you exposing her and her baby to a medication? You also want to document that the patient is competent to consent to treatment. It can be helpful to write down verbatim comments and questions that the patient makes to show that she’s grasping the information you’re providing.
And you also want to document that you’ve informed her of the baseline rate of birth defects of 2-4%, so no one should expect a perfect baby every single time, because that’s not the way nature works. Sometimes babies are born with miscellaneous birth defects.
So these are four very important documentation points that you should never forget to include.
And one more point: don’t choose medications only on the basis of FDA use in pregnancy ratings. I want to emphasize this point. The FDA use in pregnancy ratings used to be A, B, C, D, and X. They recently changed, but many of our medications that we use every day have been grandfathered into this system, so you will still see these use in pregnancy ratings. There’s been a lot of confusion with these pregnancy ratings, with the B category in particular. So “A” is straightforward: control studies show no risk. Synthroid is a typical example.
But the B category gets tricky: animal studies show no risk, but there are no control studies in humans or animal studies show adverse effects that has not been confirmed in human studies. We don’t want to prescribe medications to our patients because of animal studies where there’s no control studies in humans.
For Category C: animal studies show risk, but there are no control studies in humans or studies in animals in humans are not available. This category is where most psychiatric medications fall. And it’s also a messy category. Category D is for when there is evidence of risk in humans, but the drug may have benefits that outweigh the risk. An example is lithium. And then the X category is used when the risk outweighs any benefit, like Coumadin, the blood thinner, falls under category X.
The problem with these categories is that sometimes medications get placed in category B, just because there are no controlled studies in humans. And then as the studies come out in humans, they get bumped down to category C. I’ve seen that some pharmaceutical companies promote their B category medication as if it’s safer than medications that have been around for decades that are in a C category. In fact, just today, one of the residents at the hospital suggested we switch a patient who’s on olanzapine because she’s pregnant to lurasidone because lurasidone has a category B use-in-pregnancy rating.
But we have very little data on lurasidone. Just the fact that it’s in category B is not a reason to switch a patient from a medication where we do have data on humans that looks reassuring.
Several other medications are labeled as category B, despite having much less human data. Another example is clozapine. We don’t want to switch a patient from say olanzapine or risperidone to clozapine, but I have seen that happen.
Better choices during pregnancy are well studied medications with category C levels like haloperidol, olanzapine, fluoxetine citalopram. The FDA created new use in pregnancy ratings because they realized how confusing these old ratings were. These new ratings came out about seven years ago, in 2015, and there are no more letter grades. Now manufacturers provide a narrative description of the risks.
And this new system applies for all medications approved after June 2015. Medications approved between 2001 and 2015 must create revisions; and medications approved before 2001, which include many of the medications we use, are basically grandfathered into the old system.
One more point is that some stages of fetal development are more sensitive than others, so we can prescribe medications to our pregnant patients judiciously, if we know what organ system could be most at risk. For example, we know that lithium affects heart development during the first 10 weeks of gestation. If we, whenever possible, can avoid using lithium during those first 10 weeks at gestation, that’s preferable because then we avoid exposing the fetus to lithium when the heart is in its most sensitive period of development.
Benzodiazepines have been linked in some studies with cleft lip and palate. The lip and palate form around weeks six to nine. The risk appears low, if it is even a risk, but if we can try to help a patient get through those weeks without benzodiazepines, that can be worth trying. But it’s not always possible to avoid benzodiazepines; if a patient has severe panic attacks, we know anxiety attacks are not good for the pregnancy. There’s a fair amount of data that high levels of stress hormones, like corticotropin releasing hormone have been linked themselves to higher risks of preterm labor and low birth weight. So, we know that exposing a fetus to high levels of stress hormones is not a good thing either. We want to weigh that risk versus exposing a child to a benzodiazepine.
I’ll say a few things about ECT. It is probably underutilized for pregnant women, considering that it’s safe, rapid, effective, and helps minimize exposure to medications. A study of ECT in pregnancy reported a complete response to treatment in 84% of patients with major depression, 92% of patients with bipolar disorder, and 50% of patients with schizophrenia.
Another approach that will hopefully get used more often for pregnant patients is transcranial magnetic stimulation. It’s approved for the treatment of depression and also for migraines, and no adverse pregnancy or fetal outcomes have been seen from its use in pregnancy. It certainly seems like a promising treatment but unfortunately, it’s very expensive. If insurance covers it, then, it’s an option and hopefully, prices will come down over time.
I’ll say a few things about psychotherapy. There are two forms of therapy that have been specifically studied for perinatal patients: interpersonal therapy is brief—12 to 16 weeks—and focuses on resolving interpersonal problems, improving communication patterns, and establishing realistic expectations from a patient’s social support networks. Cognitive therapy is also a brief form of therapy, 12 to 16 weeks, in which patients are taught to challenge their negative thought patterns.
Lastly, I’ll say something about micronutrient implementation. Mood disorders have been associated with low levels of omega fatty acids and low levels of folate. There’s not a whole lot of compelling data on this topic, but it’s good for women to have more omega fatty acids in their diet during pregnancy and if it can help their mood, then all the better. We know that the fetus needs omega fatty acids for brain and eye development—all the more reason for women who are pregnant to consume foods with high levels of omega fatty acids, like salmon and other low mercury fish like cod, sole, sea bass.
Folate controls cofactors in the synthesis of serotonin, dopamine, and norepinephrine, so I encourage consumption of green leafy vegetables. The word folate comes from foliage.
In conclusion, depression in pregnancy is common, but frequently missed. Untreated depression is not only linked with maternal suffering, but also with adverse pregnancy outcomes. Several treatment options are available and when using antidepressants, choose medications with extensive data on human exposures and maximize nonpharmacologic treatments.
Ensure that you’ve had a proper discussion and documentation of risks and benefits and educate your patients about the baseline risk of two to 4% of birth defects, regardless of prenatal exposures. I hope this was helpful and thank you for listening.
Earn CME for watching the webinars with a Webinar CME Subscription.
Sayres Van Niel M and Payne J, Cleve Clin J Med 2020;875:273-277
Dagher RK et al, J Womens Health (Larchmt) 2021;302:154-159
Hope H et al, Br J Psychiatry 2022;1-7 (online ahead of print)
Roberts MH et al, Obstet Gynecol 2022;139(2):172-180
Wallace ME et al, Am J Obstet Gynecol 2016;215(3):364.e1-364.e10
Bandoli G.,Pediatrics 2020;146(1):e20192493
Ban L, BJOG 2014;121(12):1471-181
Kolding L et al.,BJOG 2021;128(12):1949-1957
Hendrick V et al, Prim Care Companion CNS Disord 2017;19(5):17r02160
Oberlander TF et al , Arch Pediatr Adolesc Med. 2007;161(1):22-29
Misri S et al, Am J Psychiatry 2006;163(6):1026-1032
Fornaro M et al. Am J Psychiatry 2020 Jan 1;177(1):76-92