In this report, we describe the case of a woman with recurrent major depression who wished to get pregnant. She was taking a daily antidepressant and sought information on the safety of the medication after conception. This case illustrates 1) the dilemmas that arise in the psychiatric treatment of pregnant and nursing women, 2) the risk-benefit assessments that are involved in the pharmacological treatment of depression during pregnancy and lactation, and 3) approaches to minimize the risk of such treatments.
Ms. A was a 36-year-old married woman with a history of recurrent major depression who sought consultation regarding the safety of antidepressant medications during pregnancy. She and her husband hoped to conceive in the near future but were concerned about Ms. A’s use of fluoxetine, 20 mg/day. The husband came to the appointment with his wife.
Ms. A had a history of three episodes of major depression between ages 24 and 33. She was treated with psychotherapy for her first episode and found it helpful in relieving her symptoms. She continued, however, with recurrent premenstrual dysphoria. After experiencing her second episode of major depression, on the advice of her psychotherapist, she sought psychopharmacological consultation for both the depression and the premenstrual mood symptoms. She started taking fluoxetine, 20 mg/day, which was highly effective in treating both conditions. After a year of treatment, she switched to luteal-phase use of fluoxetine. However, she experienced a relapse of major depression within 8 months. Her depression symptoms included low energy, anhedonia, irritability, poor appetite, insomnia, and low self-esteem. Prompt resumption of fluoxetine, 20 mg/day, produced a full and sustained remission of symptoms, with no recurrence up to the present time. She had no history of suicidality or psychiatric hospitalizations.
Ms. A’s medical history was unremarkable. She was a healthy Caucasian woman with regular menstrual cycles and no prior pregnancies. She used a diaphragm for contraception. She was a nonsmoker and did not use illicit drugs. She used alcohol socially.
The initial consultation included a detailed review of psychiatric treatment options during pregnancy and a review of the risks and benefits associated with the use of antidepressant medications. Approaches to antidepressant medication management during pregnancy were discussed, including 1) tapering and discontinuing the antidepressant before attempting conception, 2) continuing to take the antidepressant until she had a positive pregnancy test, followed by immediate discontinuation, and 3) continuing to take the antidepressant throughout pregnancy. After listening to her options, Ms. A expressed interest in discontinuing the antidepressant once she had a positive pregnancy test. She was reluctant to stop taking the medication before conception out of concern that premenstrual dysphoric disorder and/or major depression would recur.
Since the fluoxetine metabolite (norfluoxetine) has an extremely long elimination half-life (7–10 days), the developing fetus would continue to be exposed to this drug for several weeks after medication discontinuation. The plan was therefore made to switch to a shorter-acting medication. Alternative medications were discussed, including side effect profiles and family histories of response. Ms. A mentioned that her sister had responded well to sertraline and that she would like to try this agent. A switch was therefore initiated from fluoxetine, 20 mg/day, to sertraline, 75 mg/day, over the next 10 days.
Ms. A was encouraged to begin individual psychotherapy, and she agreed to this plan. She was given a list of therapists, including several with experience in working with pregnant women and new mothers. She chose a therapist with an eclectic approach that combined psychodynamic and supportive modalities.
At a follow-up visit 6 weeks later, Ms. A remained euthymic and reported no relapse of premenstrual symptoms. She and her husband continued to agree with the plans that were discussed at the first visit and felt that they were ready to stop using contraception.
After 4 months of attempting to conceive, Ms. A called to report that a pregnancy test that morning had given positive results. She discontinued the sertraline as of that day. At her visits 2 and 4 weeks later, she remained euthymic. Follow-up visits were then scheduled monthly. She continued to be well until her sixth month of pregnancy, when she reported that her mood had deteriorated. Ms. A increased her individual therapy to twice-weekly meetings. Two weeks later, she reported that her depression and hopelessness hampered her motivation for therapy. She added that her appetite was poor and that she had ceased to gain weight. She reported no suicidality but felt apathetic about the pregnancy and had lost her ability to take pleasure in activities she had previously enjoyed, such as attending the theatre or meeting with friends. The potential risks and benefits associated with the prenatal use of antidepressants in the second and third trimesters were reviewed. After considering the information, Ms. A chose to initiate fluoxetine, 20 mg/day. Within 2 weeks, Ms. A felt substantially better, and by 1 month she was euthymic.
At 39½ weeks’ gestation, Ms. A delivered a healthy baby girl with 1- and 5-minute APGAR scores of 7 and 9, respectively. She took the same dose of fluoxetine after delivery and, in a phone appointment 2 weeks after the birth, reported feeling tired but generally euthymic. However, at a psychiatric visit 3 weeks after delivery, Ms. A reported difficulty sleeping even when the infant was asleep. She also described a growing apathy toward her parenting responsibilities, which produced a great deal of guilt. Furthermore, she reported a lack of confidence in herself as a mother and wept as she said this. A decision was made to add nortriptyline, 50 mg, at bedtime. This medication was chosen not only to augment fluoxetine but also to help with sleep. At a visit the next week, Ms. A reported that her sleep had improved. Her serum concentration of nortriptyline at that time was 70 ng/ml. Ms. A added that the infant was feeding well and appeared healthy. At a psychiatric visit 6 weeks after delivery, Ms. A reported feeling much better. She held her daughter tenderly and spoke about her with pride. Ms. A continued individual psychotherapy but reduced the frequency to weekly visits.
In visits at 3 and 6 months postpartum, Ms. A remained well. At 6 months postpartum, she weaned her daughter. She also discontinued the antidepressants since she felt she no longer needed them. Furthermore, she believed that the nortriptyline had impeded her efforts to lose the weight she had gained during her pregnancy. Ms. A’s menstrual cycles returned at 8 months postpartum. She experienced a recurrence of premenstrual mood changes, and the decision was made to resume taking fluoxetine, which had helped her premenstrual symptoms before her pregnancy. She reported no further premenstrual mood changes in subsequent months of follow-up.
This case illustrates the issues that commonly arise in the treatment of depression in women who are pregnant or trying to conceive. The patient wisely sought consultation before pregnancy, thus allowing time for a preconception treatment plan to be reviewed and implemented. Preconception evaluations provide an opportunity to discuss data on exposures to antidepressant medications during pregnancy and lactation, to review a woman’s likelihood of a depressive relapse during pregnancy and the postpartum period, and to institute supportive and prophylactic interventions that may reduce the risk of relapse. Ideal management that minimizes risk to both mother and infant remains uncertain since data on the impact of different treatment approaches for pregnant patients are incomplete. Even experts in the field express differing opinions on the optimal treatment of depression in pregnant women (1). The information to date will be reviewed.
The largest amount of information on prenatal antidepressant exposures involves fluoxetine. Neither retrospective nor prospective studies have found a greater risk of miscarriage or major congenital malformations with antidepressant treatment (2–8). Third-trimester use of fluoxetine has been linked with higher rates of perinatal complications (e.g., tachypnea, jitteriness, premature delivery) in some (3, 4), but not all (5–8), studies. A study of 226 prenatal exposures to sertraline, paroxetine, and fluvoxamine (9) found the rates of major malformations and preterm labor were no higher than those of nonexposed comparison subjects. A prospective study that used the Swedish Birth Registry (6) reported 969 cases of prenatal exposure to antidepressants, including citalopram (375 cases), paroxetine (122 cases), sertraline (33 cases), and fluoxetine (16 cases) and reported rates of perinatal complications and congenital malformations comparable to population norms. Among 150 women who took venlafaxine during pregnancy, the incidence of major malformations was no higher than the expected rate of 1%–3% (10).
Prenatal use of tricyclic antidepressants similarly does not appear to increase the risk of congenital anomalies (5, 6), although these medications may produce transient neonatal toxicity or withdrawal symptoms when used near the time of delivery. These symptoms have included lethargy, hypotonia, and anticholinergic effects, such as constipation, tachycardia, and urinary retention (5, 11, 12). Nortriptyline and desipramine are preferable to other tricyclic antidepressants during pregnancy because of their lower likelihood of anticholinergic and hypotensive side effects. Their doses may need to be adjusted over the course of pregnancy since serum concentrations may fall below the therapeutic window (13). A small case series (N=7) of mirtazapine use in pregnancy (14) found no perinatal complications or congenital malformations in the infants.
A study of 21 prenatal exposures to monoamine oxidase inhibitors (MAOIs) (15) found a relative risk for congenital malformations of 3.4. In contrast, a more recent case report of phenelzine use throughout pregnancy (16) described a healthy outcome for the patient and her infant. Nevertheless, MAOIs are best avoided in pregnant women because of the risk of hypertensive crisis. Additionally, MAOIs contraindicate the use of tocolytic agents (e.g., terbutaline), which may be necessary to prevent premature labor. At present, to our knowledge, no published studies exist on bupropion or nefazodone in pregnancy.
When selecting an antidepressant medication for a pregnant woman, clinicians should be cautious in their use of the Food and Drug Administration (FDA) use-in-pregnancy ratings. These ratings categorize medications into one of five groups on the basis of known risk to the fetus: an "A" rating indicates that controlled studies of the medication show no risk to pregnant women; a "B" rating indicates that no evidence of risk exists to date in humans; a "C" rating is used when risk cannot be ruled out; a "D" rating indicates positive evidence of risk; and an "X" rating indicates that the medication is contraindicated in pregnancy. A medication carrying an FDA pregnancy category B rating is not necessarily safer to use in pregnancy than a medication with a pregnancy category C labeling. Medications may receive the category B rating in the absence of human data, provided that animal reproduction studies have not demonstrated a fetal risk. A category C medication for which there are human data showing low risk may be preferable to a category B medication for which there are no human studies, as is the case for bupropion. An additional concern regarding fetal exposure to bupropion arises from the medication’s seizure-inducing potential.
In contrast to morphological teratogenesis, behavioral teratogenesis has received minimal attention. Behavioral teratogenesis refers to the neurological or behavioral impact that may result in children from prenatal exposure to centrally acting medications. Use of antidepressant medications during pregnancy exposes the child’s brain to psychoactive agents at a time of maximal development of the CNS (17, 18). Furthermore, the relatively small proportion of myelin in fetal brains allows medications to accumulate in the CNS (19). Therefore, the safety of antidepressant use in pregnant women cannot be ensured until possible behavioral consequences are studied. Data on the developmental outcomes in children exposed to antidepressants during pregnancy are reassuring but limited. Two prospective studies (8, 20) and one case series (21) have reported developmental outcomes in children prenatally exposed to antidepressants. The first study evaluated 55 children who were prenatally exposed to fluoxetine, 80 children who were prenatally exposed to tricyclic antidepressants, and 84 unexposed children (8) and found no significant differences in global IQ, language development, temperament, mood, distractibility, or behavior in children up to the age of 7. The second study (20), which included 86 infants prenatally exposed to citalopram, fluoxetine, paroxetine, or sertraline, evaluated the children at age 18 months with the Bayley Scales of Infant Development and found scores in the age-appropriate range. In the case series (21), developmental milestones among nine children aged 4 months to 3 years were normal for chronological age after prenatal exposure to tricyclic antidepressants.
While these data are reassuring, the total number of reported exposures remains small. Therefore, discontinuation of an antidepressant before conception is a reasonable approach for women who are stable and, by history, appear likely to remain well for at least several months while not taking the medication. Continuation of the antidepressant until conception is another reasonable option. If the antidepressant is discontinued at the time of a positive pregnancy test (i.e., approximately 12–14 days postconception), the developing embryo will receive minimal medication exposure since the uteroplacental circulation does not form until 10–12 days postconception (22). It is important to keep in mind, however, that short-term discontinuation of an antidepressant may produce withdrawal symptoms (23) and is associated with a risk of relapse to depression over the course of the pregnancy of up to 75% (24, 25). For women like our patient who seek counseling before conception and obtain supportive and prophylactic interventions, it is possible that the relapse risk is lower. Each case must be evaluated on an individual basis by taking into account the number of previous episodes of major depression and the time to relapse after previous attempts at medication discontinuation.
For women with histories of rapid and severe relapse of major depressive episodes after medication discontinuation, the antidepressant may need to be maintained throughout the pregnancy. These recommendations are consistent with the Expert Consensus Guideline Series on Treatment of Depression in Women 2001 (1).
Women who are pregnant or attempting to conceive should be encouraged to initiate psychotherapy if they are not already obtaining it. Interpersonal psychotherapy—a form of therapy that examines role transitions and disputes and deficits in interpersonal interaction—has been evaluated for pregnant women and appears to be effective (26).
The risk-benefit discussion should be recorded in the patient’s file along with a statement indicating that the patient understands the information provided and is in agreement with the plan. Whenever possible, the partner and other health care providers should be included in the risk-benefit discussion.
Impact of Depression on Pregnancy
A number of studies have reported an association between maternal anxiety/stress during pregnancy and negative pregnancy outcomes, such as preterm labor and low- birth-weight infants (27–37), even after control for confounding variables such as socioeconomic status, maternal weight gain, and health habits during pregnancy. Fewer investigations have evaluated the impact of maternal major depression on pregnancy outcomes. A recent study of 623 women followed prospectively over pregnancy (38) reported a significant relationship between depression (as measured with the Beck Depression Inventory) and preeclampsia. In another prospective study of 389 women of low socioeconomic class (39), depression during the first trimester of pregnancy was associated with a greater risk of low-birth-weight infants and preterm delivery. In contrast, a prospective study that administered psychiatric interviews to 94 women across pregnancy (40) found no relationship between maternal mood during pregnancy and infant birth weights or gestational ages. This study also found no association between maternal depression during pregnancy and children’s cognitive scores on the McCarthy Scales at age 4 years.
In summary, maternal anxiety and stress during pregnancy appear to predict adverse pregnancy outcomes. The impact of maternal major depression on pregnancy outcomes is less clear at this time and requires further investigation.
Risk Factors for Postpartum Relapse
Our patient experienced a recurrence of major depression in the postpartum period, in which she had chosen to remain medication-free while nursing. A woman with a history of major depression, like Ms. A, is at approximately a 25% risk for relapse of depression after childbirth (postpartum depression). Depressive symptoms during pregnancy raise the risk even higher (41). Other risk factors include a conflictual relationship with the baby’s father (42), stressful life events (42), low socioeconomic status (43), and frequent health problems in the infant (44).
Postnatal Depression and Children’s Outcomes
Ms. A experienced postpartum depression with symptoms including apathy, low energy, guilt, and low self-confidence. Her decision to restart the antidepressant medication helped her significantly and may also have been beneficial for her infant. A substantial literature documents the potential negative impact of maternal postnatal depression on children. Most studies have included women with depression occurring within 3–12 months of delivery rather than the shorter time frame of 4 weeks that DSM-IV defines for "postpartum onset." Therefore, we will use the term "postnatal" rather than "postpartum" when referring to depression occurring within the first year after delivery.
Compared with children of nondepressed mothers, children of mothers who have experienced postnatal depression perform worse on cognitive and behavioral measures (45–51) and exhibit higher rates of insecure attachment (45, 46, 52). Children of mothers who experienced depression in the child’s first year of life appear to be at greater risk for adverse cognitive outcomes than children who were older when their mothers experienced depression. Children’s exposure to subsequent relapses of maternal depression also appears to increase their risk for poor cognitive outcomes.
Several researchers have evaluated the mechanisms mediating the negative effect of parental depression on children’s development and have identified negative parenting behaviors that occur commonly among depressed parents. For example, depressed mothers have been observed to be intrusive or withdrawn and disengaged when interacting with their infants (53, 54) and less sensitively attuned to their infants than healthy women (55). The disturbances in mother-infant interactions observed among depressed mothers were found to highly predict poor infant cognitive outcomes at 18 months (55). A recent study of 2,017 parents of children ages 0–3 years (56) identified parental depressive symptoms as the most consistent predictors of negative parenting behaviors (e.g., yelling, hitting, shaking) after controlling for socioeconomic status, ethnic group, parental years of education, parental age, and parental employment status. For each depressive symptom, the odds that the parent would engage in at least one form of negative interaction rose by 25% for mothers and 43% for fathers (56). These studies indicate that children’s exposure to parental depression represents a substantial risk for poor developmental outcomes. Underscoring the importance of intervention, research has found that if maternal depression is prevented, infant problems associated with maternal depression, such as behavioral problems and insecure attachment (57) and decline in IQ (46), can also be prevented.
Treatment of Postpartum Depression
A number of treatment interventions are helpful for postpartum depression and include psychotherapy, support groups, and referrals to self-help and national organizations such as Postpartum Support International (805-967-7636) or Depression After Delivery (800-944-4773). Interpersonal and cognitive behavior therapy are forms of individual psychotherapy that have been found effective for treatment of postpartum depression (58, 59).
Antidepressants are also helpful in relieving postpartum depression (59–61). The choice of antidepressant should take into account whether or not the patient is nursing since antidepressant medications traverse readily into breast milk. To date, the largest literature on the use of antidepressants by nursing women involves the selective serotonin reuptake inhibitors (SSRIs). A study compared 26 nursing infants whose mothers took fluoxetine with 28 nursing infants whose mothers did not take the medication (62) and reported a significantly lower growth rate (although within the normal range) among the fluoxetine-exposed group over the first 6 months of age. The study was retrospective and did not control for the impact of maternal depression on breast-feeding behavior. Also, two case reports (63, 64) attributed uneasy sleep and colic to breast-feeding infants exposed to citalopram and fluoxetine, respectively, through breast milk, with resolution after discontinuation or diminution of breast-feeding. However, several studies totaling more than 210 mother-infant pairs (12, 21, 65–77) observed no adverse effects of SSRI exposure in nursing infants. The majority of these studies involved fluoxetine, sertraline, and paroxetine. These data show that in many cases women can safely continue breast-feeding while taking an antidepressant medication.
A consideration for new mothers who have recently weaned their infants is the potential occurrence of premenstrual mood changes once menstrual cycles resume (78). This is of particular concern for women with histories of premenstrual dysphoric disorder. As in this case, Ms. A’s postpartum major depression returned soon after antidepressant discontinuation and remitted with reinstitution of treatment. High relapse rates have been reported for postpartum major depression after treatment discontinuation (79), supporting the need for ongoing treatment for women with this condition.
Women with histories of major depression may relapse during pregnancy or the postpartum period. Depression during pregnancy can present risks to the mother and fetus, such as inadequate maternal weight gain and, in the extreme, suicidality. After childbirth, maternal depression may interfere with mother-child bonding and may be detrimental to the infant’s development. Psychotherapeutic approaches, including individual psychotherapy, couples counseling, and support groups, can help treat major depression during pregnancy and the postpartum period. If a woman remains depressed despite nonpharmacologic interventions or if her depression is associated with significant morbidity, antidepressant medications should be considered. Given the data on birth outcomes after prenatal exposure to several antidepressant medications and the accumulating data on the use of antidepressant medications by breast-feeding women, the relative benefits of a medication to treat depression in pregnant or nursing women may outweigh the theoretical risk of its use.
Received Feb. 20, 2002; revision received May 29, 2002; accepted June 6, 2002. From the Mood Disorders Research Program, UCLA Neuropsychiatric Institute and Hospital; and the West Los Angeles Veterans Administration Medical Center. Address reprint requests to Dr. Hendrick, Mood Disorders Research Program, UCLA Neuropsychiatric Institute and Hospital, UCLA Medical Plaza, Bldg. 300, Suite 2345, Los Angeles, CA 90095; firstname.lastname@example.org (e-mail).