OVERVIEW

Psychiatric consultation to obstetric patients typically involves the evaluation and treatment of an array of psychopathology. Once thought to be a time of emotional well-being for women,¹ studies now suggest that pregnancy is not protective with respect to the emergence or persistence of psychiatric disorders.^2–7 Given the prevalence of mood and anxiety disorders in women during the childbearing years^8,9 and the number of women who receive treatment for these disorders, it is apparent that many women become pregnant either after the recent discontinuation of psychotropic medications or following a decision to maintain treatment during efforts to conceive. With increasing evidence of high rates of relapse following discontinuation of psychotropic medications (e.g., antidepressants,¹⁰ mood stabilizers,¹¹ antipsychotics,¹² and benzodiazepines¹³) and other data describing new-onset psychiatric illness during pregnancy,^2,6,14 the value of psychiatric consultation during pregnancy and after delivery is evident.

Psychiatric evaluation of pregnant women requires careful assessment of symptoms such as anxiety or depression and decisions about the nature of those symptoms, such as (1) normative or pathological, (2) manifestations of a new-onset psychiatric disorder, or (3) an exacerbation of a previously diagnosed or undiagnosed psychiatric disorder. Unfortunately, screening for a psychiatric disorder during pregnancy or the puerperium is uncommon. Even when depressed pregnant women are identified, definitive treatment is frequently lacking.¹⁵ Screening for depression during pregnancy followed by thoughtful treatment can minimize maternal morbidity, as well as the potential impact of an untreated psychiatric disorder on infant development and family functioning. Pregnancy is an emotionally laden experience that evokes a spectrum of normal reactions, including heightened anxiety and increased reactivity of mood.

Normative experience needs to be distinguished from the manifestations of psychiatric disorders. Treatment of psychiatric disorders during pregnancy involves a thoughtful weighing of the risks and benefits of proposed interventions, such as pharmacological treatment and the documented^16,17 and theoretical risks associated with untreated psychiatric disorders. In contrast to many other clinical conditions, treatment of psychiatric disorders during pregnancy is typically reserved for situations in which the disorder interferes significantly with maternal and fetal well-being; the threshold for the treatment of psychiatric disorders during pregnancy tends to be higher than with other conditions. Moreover, women with similar illness histories often make very different decisions about their care in collaboration with their physicians during pregnancy.

DIAGNOSIS AND TREATMENT OF MOOD DISORDERS DURING PREGNANCY

Although some reports describe pregnancy as a time of affective well-being^1,18–20 that confers “protection” against psychiatric disorders, at least one prospective study describes equal rates of major and minor depression (approximating 10%) in gravid and nongravid women. Several other recent studies also note clinically significant depressive symptoms during pregnancy (antenatal depression).^2,6,21,22 Furthermore, women with histories of major depression appear to be at high risk for recurrent depression during pregnancy, particularly in the setting of antidepressant discontinuation.^5,23

Diagnosis of depression during pregnancy can be difficult because disturbances in sleep and appetite, symptoms of fatigue, and changes in libido do not necessarily suggest an evolving affective disorder. Clinical features that may support the diagnosis of major depressive disorder (MDD) include anhedonia, feelings of guilt and hopelessness, and thoughts of suicide. Suicidal ideation is not uncommon^24–26; however, risk of frank self-injurious or suicidal behaviors appears to be relatively low in women who develop depression during pregnancy.^25,26

Treatment for depression during pregnancy is determined by the severity of the underlying disorder. Nonetheless, neurovegetative symptoms that interfere with maternal well-being require treatment. Women with mild to moderate depressive symptoms may benefit from nonpharmacological treatments that include supportive psychotherapy, cognitive therapy,²⁷ or interpersonal therapy (IPT),²⁸ all of which may ameliorate depressive symptoms. Given the importance of interpersonal relationships in couples who are expecting a child and the significant role transitions that take place during pregnancy and after delivery, IPT is ideally suited for the treatment of depressed pregnant women; preliminary but encouraging data support the efficacy of this intervention.²⁹

ANTIDEPRESSANT USE DURING PREGNANCY

Multiple reviews have been published over the last decade that describe available data (from anecdotal case reports and larger samples of patients) regarding risks associated with fetal exposure to antidepressants.^23,30–34 Although accumulated data over the last 30 years suggest that some antidepressants may be used safely during pregnancy,^35–37 information regarding the full spectrum and relative severity of attendant risks of prenatal exposure to psychotropic medications is still incomplete.

As is the case with other medications, four types of risk are typically cited with respect to potential use of antidepressants during pregnancy: (1) risk of pregnancy loss or miscarriage, (2) risk of organ malformation or teratogenesis, (3) risk of neonatal toxicity or withdrawal syndromes during the acute neonatal period, and (4) risk of long-term neurobehavioral sequelae.³⁶ To provide guidance to physicians seeking information on the reproductive safety of various prescription medications, the Food and Drug Administration (FDA) has established a system that classifies medications into five risk categories (A, B, C, D, and X) based on data derived from human and animal studies. Medications in category A are designated as safe for use during pregnancy, whereas category X drugs are contraindicated and are known to have risks to the fetus that outweigh any benefit to the patient. Most psychotropic medications are classified as category C agents, for which human studies are lacking and for which “risk cannot be ruled out.” No psychotropic drugs are classified as safe for use during pregnancy (category A).

Unfortunately, this system of classification is often ambiguous and may lead some to make conclusions that are not warranted. For example, certain tricyclic antidepressants (TCAs) have been labeled as category D agents, indicating “positive evidence of risk,” although the pooled available data do not support this assertion and, in fact, suggest that these drugs are safe for use during pregnancy.^38,39 Therefore, the physician must also rely on other sources of information when counseling patients about the potential use of psychotropic medications during pregnancy. Randomized, placebo-controlled studies examining the effects of medication use on pregnant populations are unethical for obvious reasons. Therefore, much of the data related to the profile of reproductive safety for a medication is derived from retrospective studies and case reports. More recently, studies that have evaluated the reproductive safety of antidepressants have used a more rigorous prospective design,^38–43 or have relied on large administrative databases or multicenter birth defect surveillance programs.^44,45

To date, studies have not demonstrated a statistically increased risk of spontaneous miscarriage or congenital malformations associated with prenatal exposure to antidepressants.

Cumulative reports describing the reproductive safety of selective serotonin reuptake inhibitors (SSRIs) have been recently reviewed.^33,34 These reports provide some relative reassurance that as a group of medicines, SSRIs do not constitute major teratogens. However, some recent reports have suggested that first-trimester exposure to paroxetine is associated with an increased risk of cardiac defects, including atrial and ventricular septal defects.^46–49 While these findings prompted the FDA to change the category label of paroxetine from C to D, more recently published studies have not demonstrated increased teratogenicity of paroxetine.^44,45

Bupropion may be an attractive option for women who have not responded well to fluoxetine or TCAs. Data thus far have not indicated an increased risk of malformations associated with bupropion use during pregnancy.^50–52

There are limited data available on the use of the serotonin-norepinephrine reuptake inhibitor (SNRI) venlafaxine⁴³ during pregnancy. Nonetheless, given the frequency of use of these medicines and the frequency of unplanned pregnancy,⁵³ the data supporting the safety of bupropion and venlafaxine are increasingly reassuring.

Despite the growing literature supporting the relative safety of fetal exposure to SSRIs, multiple reports have described adverse perinatal outcomes, including decreased gestational age, low birth weight, and poor neonatal adaptation.^40,54,55 However, other investigators have failed to note these same associations.^38,56,57 Particular concern has been raised regarding the potential effects of late pregnancy exposure to SSRIs, with one recent report noting symptoms including jitteriness, tachypnea, and tremulousness,⁵⁸ and another study reporting an increased risk of persistent pulmonary hypertension of the newborn (PPHN).⁵⁹ Further investigation is warranted to clarify the association between SSRI use and PPHN, as previous studies investigating neonatal outcomes in infants exposed to antidepressants in utero have not described serious complications with significant morbidity and instead described neonatal distress syndromes that were generally mild, transient, and did not require specific medical intervention. Unfortunately, the vast majority of reports that have attempted to delineate the potential effects of peripartum exposure to SSRIs have been limited by small sample size, nonsystematic assessment of infant outcome, and frequent use of nonblinded raters. Moreover, these studies typically have failed to assess the impact of maternal depression, which in and of itself may be associated with compromised perinatal outcome.¹⁷

While much research has addressed the effect of antidepressant drugs on risk for congenital malformation, less research has focused on the long-term effects of prenatal antidepressant exposure. Studies thus far have indicated that there were no differences between the children exposed to fluoxetine or TCAs during pregnancy and children who were not exposed during pregnancy in terms of IQ, language, temperament, behavior, reactivity, mood, distractibility, and activity level.^41,56 In addition, there do not appear to be any differences in levels of internalizing behaviors in children exposed to SSRIs during pregnancy when compared to children of nondepressed mothers not on medication during pregnancy.⁶⁰ While the available data are reassuring, further investigation into the long-term neurobehavioral effects of prenantal exposure to antidepressants is warranted.

PHARMACOLOGICAL TREATMENT OF DEPRESSION DURING PREGNANCY: CLINICAL GUIDELINES

The last decade has brought increased attention to the question of how to best manage women who suffer from depression. Clinical lore previously suggested that women enjoyed positive mood during pregnancy, but more recent data suggest that subpopulations of patients may be at risk for recurrence or new onset of depression during pregnancy. There is also a greater appreciation that depression may exert an effect on fetal and neonatal well-being that needs to be taken into account in the risk-benefit decision-making process.^{23,30,35,37,61} The majority of women suffering from depression during pregnancy do not receive adequate treatment, even though this illness is relatively common.¹⁵ Despite the growing number of reviews on the subject, management of antenatal depression is still largely guided by practical experience, with few definitive data and no controlled treatment studies to inform treatment. The most appropriate treatment algorithm depends on the severity of the disorder and, ultimately, the patient’s wishes. Clinicians must work collaboratively with the patient to arrive at the safest decision based on available information. A patient’s psychiatric history, her current symptoms, and her attitude toward the use of psychiatric medications during pregnancy must be carefully assessed and factored into any decision.

In patients with less severe depression, it may be appropriate to consider discontinuation of pharmacological therapy during pregnancy. Though data on the use of IPT or cognitive-behavioral therapy (CBT) to facilitate antidepressant discontinuation before conception are not available, it makes clinical sense to pursue such treatment in women on maintenance antidepressant therapy who are planning to become pregnant. These modalities of treatment may reduce the risk of recurrent depressive symptoms during pregnancy, although as noted previously, this has not been studied systematically. Close monitoring of affective status during pregnancy is essential, even if all medications are discontinued and no need for reintroduction of antidepressant is apparent. Psychiatrically ill women are at high risk for relapse during pregnancy, and early detection and treatment of recurrent illness may significantly reduce the morbidity associated with having an antenatal affective disorder.

Many women who discontinue antidepressant treatment during pregnancy experience recurrent depressive symptoms.^62,63 In one recent study, women who discontinued their medications were five times more likely to relapse as compared to women who maintained their antidepressant treatment across pregnancy.⁵ Thus, women with recurrent or refractory depressive illness may decide in collaboration with their clinician that the safest option is to continue pharmacological treatment during pregnancy in order to minimize the risk for recurrent illness. In this setting, the clinician should attempt to select medications during pregnancy that have a well-characterized reproductive safety profile, which may necessitate switching from one psychotropic to another with a better reproductive safety profile. An example would be switching from duloxetine, a medication for which there is sparse data on reproductive safety, or paroxetine, which may be associated with cardiac malformations, to an agent such as fluoxetine or citalopram. In other situations, one may decide to use a medication for which information regarding reproductive safety is sparse. A scenario that highlights this decision is a woman with refractory depressive illness who has responded only to one particular antidepressant for which specific data on reproductive safety are limited (e.g., venlafaxine). She may choose to continue this medication during pregnancy rather than risk (1) potential relapse associated with antidepressant discontinuation or (2) a switch to another antidepressant to which such a patient has no prior history of response.

Women may also experience the new onset of depressive symptoms during pregnancy. For women with minor depressive symptoms, nonpharmacological treatment strategies should be explored first. IPT or CBT may be beneficial for reducing the severity of depressive symptoms and may either limit or obviate the need for medications.^27–29 In general, pharmacological treatment is pursued when nonpharmacological strategies have failed or when it is thought that the risks associated with psychiatric illness during pregnancy outweigh the risks of fetal exposure to a particular medication.

In situations in which pharmacological treatment is more clearly indicated, the clinician should attempt to select the safest medication regimen, using, if possible, medications with the safest reproductive profile. Fluoxetine and citalopram, with rather extensive data supporting their reproductive safety, can be considered first-line choices. The TCAs and bupropion have also been relatively well characterized and can be considered reasonable treatment options during pregnancy. Among the TCAs, desipramine and nortriptyline are preferred because they are less anticholinergic and less likely to exacerbate orthostatic hypotension during pregnancy. The amount of literature on the reproductive safety of the newer SSRIs is growing, and these agents may be useful in certain settings.^33,42,64 Paroxetine, however, may be best avoided until the conflicting data regarding the association between fetal exposure to this antidepressant and atrial and ventricular septal defects⁴⁸ are more clearly delineated.⁴⁹ When prescribing medications during pregnancy, every attempt should be made to simplify the medication regimen. For instance, one may select a more sedating TCA for a woman with depression and a sleep disturbance instead of using an SSRI in combination with trazodone or a benzodiazepine.

In addition, the clinician must use an adequate dosage of medication. Frequently, the dosage of a medication is reduced during pregnancy in an attempt to limit risk to the fetus; however, this type of modification in treatment may instead place the woman at greater risk for recurrent illness. During pregnancy, changes in plasma volume and increases in hepatic metabolism and renal clearance may significantly affect drug levels.^65,66 Several investigators have described a reduction (up to 65%) in serum levels of TCAs during pregnancy.^39,67 Subtherapeutic levels may be associated with depressive relapse³⁹; therefore, an increase in daily TCA or SSRI dosage may be required to obtain remission.⁶⁸

With multiple studies supporting the finding of transient jitteriness, tremulousness, and tachypnea associated with peripartum use of SSRIs,^58,69 some authors (as well as FDA-mandated labeling across the SSRIs) have suggested discontinuation of antidepressants just before delivery, presumably to minimize the risk of neonatal toxicity. Another potential rationale for antidepressant discontinuation proximate to delivery would derive from the assumption that this would attenuate the risk of PPHN, which has been associated with late-trimester exposure to SSRIs.⁵⁹ While the recommendation is intuitive, it is not data driven, and such a practice may actually carry significant risk as it withdraws treatment from patients precisely as they are about to enter the postpartum period, a time of heightened risk for affective illness.

Severely depressed patients who are acutely suicidal or psychotic require hospitalization, and electroconvulsive therapy (ECT) is frequently selected as the treatment of choice. Two reviews of ECT during pregnancy note the efficacy and safety of this procedure.^70,71 In a review of the 300 case reports of ECT during pregnancy published over the past 50 years, four premature labors have been reported and no premature ruptures of membranes have been reported. Given its relative safety, ECT may also be considered an alternative to conventional pharmacotherapy for women who wish to avoid extended exposure to psychotropic medications during pregnancy or for women who fail to respond to standard antidepressants.

BIPOLAR DISORDER DURING PREGNANCY

Women with bipolar disorder (BPD) have at times been counseled to defer pregnancy (given an apparent need for pharmacological therapy with mood stabilizers) or to terminate pregnancies following prenatal exposure to drugs such as lithium or valproic acid. Concerns regarding fetal exposure to lithium, for example, have typically been based on early reports of higher rates of cardiovascular malformations (e.g., Ebstein’s anomaly) following prenatal exposure to this drug.^72,73 More recent data suggest that the risk of cardiovascular malformations following prenatal exposure to lithium is smaller than previous estimates (1 per 2,000 versus 1 per 1,000).⁷⁴ Prenatal screening with a high-resolution ultrasound and fetal echocardiography is recommended at or about 16 to 18 weeks of gestation to screen for cardiac anomalies. Nonetheless, for the bipolar woman faced with a decision regarding use of lithium during pregnancy, it is approp-riate to counsel such a patient about the very small risk of organ dysgenesis associated with prenatal exposure to this medicine.

Lamotrigine is another mood stabilizer that is an option for pregnant women with BPD who demonstrate a clear need for prophylaxis with a mood stabilizer. While previous reports did not show an elevated risk of malformations associated with lamotrigine exposure,^75–77 data from the North American Anti-Epileptic Drug Registry indicates increased risk of oral cleft in infants exposed to lamotrigine during the first trimester; the prevalence rate was approximately 9 per 1000 births.⁷⁸

Compared with lithium and lamotrigine, prenatal exposure to some anticonvulsants is associated with a far greater risk for organ malformation. An association between prenatal exposure to mood stabilizers, including valproic acid and carbamazepine, and neural tube defects (3% to 8%) and spina bifida (1%) also has been observed.^79–82 Fetal exposure to anticonvulsants has been associated not only with relatively high rates of neural tube defects, such as spina bifida, but also with multiple anomalies (including midface hypoplasia [also known as the “anticonvulsant face”], congenital heart disease, cleft lip and/or palate, growth retardation, and microcephaly). Factors that may increase the risk for teratogenesis include high maternal serum anticonvulsant levels and exposure to more than one anticonvulsant. This finding of dose-dependent risk for teratogenesis is at variance with that for some other psychotropic medications (e.g., antidepressants). Thus, when using anticonvulsants during pregnancy, the lowest effective dose should be used, and anticonvulsant levels should be monitored closely and the dosage adjusted appropriately.

Information about the reproductive safety of newer anticonvulsants sometimes used to treat BPD, including gabapentin, oxcarbazepine, and topiramate, remains sparse.⁸³ Other efforts are underway to accumulate data from prospective registries regarding teratogenic risks across a broad range of anticonvulsants. The North American Antiepileptic Drug Pregnancy Registry recently was established as a way of collecting such information rapidly and efficiently (www.aedpregnancyregistry.org).

Prenatal screening following anticonvulsant exposure for congenital malformations (including cardiac anomalies) with fetal ultrasound at 18 to 22 weeks of gestation is recommended. The possibility of fetal neural tube defects should be evaluated with maternal serum alpha-fetoprotein (MSAFP) and ultrasonography. In addition, 4 mg a day of folic acid before conception and in the first trimester for women receiving anticonvulsants is frequently recommended. However, it should be noted that the supplemental use of folic acid to attenuate the risk of neural tube defects in the setting of anticonvulsant exposure has not been systematically evaluated.

Whereas use of mood stabilizers (including lithium and some anticonvulsants) has become the mainstay of treatment for the management of both acute mania and the maintenance phase of BPD, the majority of patients with BPD are not treated with monotherapy. Rather, use of adjunctive conventional and newer antipsychotics has become common clinical practice for many bipolar patients. Moreover, with growing data supporting the use of atypical antipsychotics as monotherapy in the treatment of BPD, patients and clinicians will seek information regarding the reproductive safety of these newer agents. To date, abundant data exist that supports the reproductive safety of typical antipsychotics; these data have been reviewed extensively elsewhere.³⁵ However, despite their growing use in psychiatry, available reproductive safety data regarding the atypical antipsychotics are limited, but increasing. Some patients who benefit from treatment with antipsychotics may decide with their clinician to discontinue the atypical antipsychotic or switch to a typical antipsychotic with a better-characterized safety profile. Atypical antipsychotics are best avoided if possible, although they are not absolutely contraindicated during pregnancy. Atypical antipsychotics should be reserved for use in more challenging clinical situations where treatment with more conventional agents has not been helpful. Given the limited data supporting the use of typical antipsychotics as monotherapy for BPD, that course of therapy should not be pursued.

The most appropriate treatment algorithm for managing reproductive-age women who wish to conceive or who are pregnant depends on the severity of the individual patient’s illness. Patients with histories of a single episode of mania and prompt, full recovery followed by sustained well-being may tolerate discontinuation of mood stabilizer before an attempt to conceive.^74,84 Unfortunately, even among women with histories of prolonged well-being and sustained euthy-mia, discontinuation of prophylaxis for mania may be associated with subsequent relapse. For women with BPD and a history of multiple and frequent recurrences of mania or bipolar depression, several options can be considered. Some patients may choose to discontinue a mood stabilizer before conception as outlined previously. An alternative strategy for this high-risk group is to continue treatment until pregnancy is verified and then taper off of the mood stabilizer. Because the uteroplacental circulation is not established until approximately 2 weeks following conception, the risk of fetal exposure is minimal. Home pregnancy tests are reliable and can document pregnancy as early as 10 days following conception, and with a home ovulation predictor kit, a patient may be able to time her treatment discontinuation accurately. This strategy minimizes fetal exposure to drugs and extends the protective treatment up to the time of conception, which may be particularly prudent for older patients because the time required for them to conceive may be longer than for younger patients. However, a potential problem with this strategy is that it may lead to relatively abrupt discontinuation of treatment, thereby potentially placing the patient at increased risk for relapse. With close clinical follow-up, however, patients can be monitored for early signs of relapse, and medications may be reintroduced as needed. Another problem with the strategy of discontinuation of mood stabilizer when the patient is being treated with valproate is that the teratogenic effect of valproate occurs early in gestation, between weeks 4 and 5, often before patients even know they are pregnant. In such scenarios, any potential teratogenic insult from valproate may have already occurred by the time the patient actually documents the pregnancy.

For women who tolerate discontinuation of maintenance treatment, the decision of when to resume treatment is a matter for clinical judgment. Some patients and clinicians may prefer to await the initial appearance of symptoms before restarting medication; others may prefer to limit their risk of a major recurrence by restarting treatment after the first trimester of pregnancy. Preliminary data suggest that pregnant women with BPD who remain well throughout pregnancy may have a lower risk for postpartum relapse than those who become ill during pregnancy.⁸⁴

For women with particularly severe forms of BPD, such as with multiple severe episodes, and especially with psychosis and prominent thoughts of suicide, maintenance treatment with a mood stabilizer before and during pregnancy may be the safest option. If the patient decides to attempt conception, accepting the relatively small absolute increase in teratogenic risk with first-trimester exposure to lithium or lamotrigine with or without antipsychotic, for example, may be justified because such patients are at highest risk for clinical deterioration if pharmacological treatment is withdrawn. Many patients who are treated with sodium valproate or other newer anticonvulsants, such as gabapentin, for which there are particularly sparse reproductive safety data, may never have received a lithium trial before pregnancy. For such patients, a lithium trial before pregnancy may be a reasonable treatment option.

Even if all psychotropic medications have been safely discontinued, pregnancy in women with BPD should be considered as a high-risk pregnancy, because the risk of major psychiatric illness during pregnancy is increased in the absence of treatment with mood-stabilizing medication and is even higher postpartum. Extreme vigilance is required for early detection of an impending relapse of illness, and rapid intervention can significantly reduce morbidity and improve overall prognosis.

Although the impact of pregnancy on the natural course of BPD is not well described, studies suggest that any “protective” effects of pregnancy on risk for recurrence of mania or depression in women with BPD are limited,⁸⁴ and the risk for relapse and chronicity following discontinuation of mood stabilizers is high.^85–88 Given these data, clinicians and bipolar women who are either pregnant or who wish to conceive find themselves between a “teratologic rock and a clinical hard place.”⁸⁹

PSYCHOSIS DURING PREGNANCY

Although anecdotal reports describe improvement of symptoms in some chronically mentally ill women during pregnancy, as a group these patients are at increased risk for poor fetal outcome.^90,91 Acute psychosis during pregnancy is both an obstetric and psychiatric emergency. Similar to other psychiatric symptoms of new onset, first onset of psychosis during pregnancy cannot be presumed to be reactive; it requires a systematic diagnostic evaluation. Psychosis during pregnancy may inhibit a woman’s ability to obtain appropriate and necessary prenatal care or to cooperate with caregivers during delivery.^90–92

Treatment of psychosis during pregnancy may include use of high-potency neuroleptic medications, such as haloperidol or thiothixene, which have not been associated with an increased risk of congenital malformations when used in the first trimester of pregnancy.^93,94 Historically, lower-potency antipsychotics have typically been avoided because of data supporting an increased risk of congenital malformations associated with prenatal exposure to these compounds.^95,96 However, their use is not absolutely contraindicated.

Psychiatric consultation may be requested regarding treatment options for mild or intermittent symptoms of psychosis or for pregnant women with chronic mental illness, such as schizophrenia, who have discontinued therapy with neuroleptic medications. Although as-needed high-potency neuroleptic medications are appropriate for treatment of more mild symptoms of psychosis, consideration should be given to introduction or reintroduction of maintenance high-potency antipsychotics in schizophrenic women who have new-onset illness or a recurrent disorder. This approach can potentially limit overall exposure to these drugs by reducing the need for treatment with higher doses of drug during relapse allowed by drug discontinuation. Patients with florid psychosis during labor and delivery may benefit from intravenous haloperidol, which may permit cooperation between the patient and the obstetrician, thereby enhancing the overall safety of the delivery.^97,98

Fewer reproductive safety data on the newer atypical antipsychotic medications are available as compared to the conventional antipsychotics. Thus far, most of the data on the reproductive safely of atypical agents have often been limited to manufacturers’ accumulated case series, and some teratovigilance data reflecting a small number of total drug exposures and spontaneous reports. Data available on olanzapine, clozapine, quetiapine, and risperidone are increasingly reported but still remain relatively sparse.^99,100

Decisions regarding the use of these and other psychotropic medications must be made on a case-by-case basis. Given the limited data regarding the reproductive safety of atypical agents, patients taking an antipsychotic drug may choose to discontinue their medication or switch to a better-characterized conventional antipsychotic, such as perphenazine or haloperidol. However, many women do not respond as well to the typical agents or have such severe illness that making any change in their regimen may place them at significant risk. Thus, women and their clinicians may choose to use an atypical agent during pregnancy in order to sustain functioning, while acknowledging that information regarding their reproductive safety remains incomplete.

ANXIETY DISORDERS DURING PREGNANCY

Although modest to moderate levels of anxiety during pregnancy are common, pathological anxiety (including panic attacks) has been associated with a variety of poor obstetric outcomes, including increased rates of premature labor, low Apgar scores, and placental abruption.^101–104 Unfortunately, the course of panic disorder in pregnancy is variable. Pregnancy may ameliorate symptoms of panic in some patients and may provide an opportunity to discontinue medication.^105–108 Other studies have noted the persistence or worsening of panic-related symptoms during pregnancy.^4,109,110

Consultation requests regarding appropriate manage-ment of anxiety symptoms during pregnancy are common. The use of nonpharmacological treatment, such as cognitive-behavioral strategies and supportive psychotherapy, may be of great value in attenuating symptoms of anxiety in some cases.^111,112 For other patients, especially those who experience panic attacks associated with new-onset or recurrent panic disorder or those with severe generalized anxiety, pharmacological intervention may be necessary. Concerns regarding the potential association between first-trimester exposure to benzodiazepines, such as diazepam, and increased risk for oral clefts have been noted in some older studies,^113–115 although other studies do not support this association.^116,117 One meta-analysis that pooled data from multiple samples of patients exposed to different types and doses of benzodiazepines for variable durations of time supported an increased risk of oral clefts following first-trimester exposure to these drugs³⁵; this risk for oral clefts was approximately 0.6% following first-trimester exposure. However, a more recent meta-analysis evaluating this potential association did not support the increased risk.¹¹⁸

For patients with panic disorder who wish to conceive, slow tapering of antipanic medications is recommended. Adjunctive CBT may benefit in helping patients discontinue antipanic agents and may increase the time to a relapse.¹¹² Some patients may conceive inadvertently on antipanic medications and may see their clinician for emergent consultation. Abrupt discontinuation of antipanic maintenance medication is not recommended given the risk for rebound panic symptoms or a potentially serious withdrawal syndrome. However, gradual taper of benzodiazepines (longer than 2 weeks) with adjunctive CBT may be pursued in an effort to minimize fetal exposure to medication.

If tapering medication is unsuccessful or if symptoms recur during pregnancy, reinstitution of pharmacotherapy may be considered. For patients with severe panic disorder, maintenance medication may be a clinical necessity. Use of TCAs or SSRIs is a reasonable option for the management of panic disorder during pregnancy.³⁵ If patients do not respond to these antidepressants, benzodiazepines may be considered.¹¹⁹ Although some patients may choose to avoid first-trimester exposure to benzodiazepines given the data on the risk for cleft lip and palate, benzodiazepines may be used without significant risk during the second and third trimesters and may offer some advantage over antidepressant treatment because they may be used as needed.

Pharmacotherapy of severe anxiety during pregnancy may include treatment with benzodiazepines, TCAs, SSRIs, or SNRIs. These classes of drugs have all demonstrated efficacy in the management of either generalized anxiety disorder (GAD)^120,121 or panic disorder.^122–124 Pharmacological treatment of severe anxiety during pregnancy may include the use of TCAs or SSRIs as a non-benzodiazepine alternative for panic attacks. Nonetheless, patients treated with antidepressant alone for management of anxiety symptoms may not respond optimally. For these patients, benzodiazepines represent a reasonable alternative.

With respect to the peripartum use of benzodiazepines, reports of hypotonia, neonatal apnea, neonatal withdrawal syndromes, and temperature dysregulation^125–131 have prompted recommendations to taper and discontinue benzodiazepines at the time of parturition. The rationale for this course is suspect for several reasons. First, given data suggesting a risk for puerperal worsening of anxiety disorders in women with histories of panic disorder and obsessive-compulsive disorder (OCD),^109,132,133 discontinuation of a drug at or about the time of delivery places women at risk for postpartum worsening of these disorders. Second, data describe the use of clonazepam during labor and delivery at doses of 0.5 to 3.5 mg per day in a group of women with panic disorder without evidence of perinatal sequelae.¹¹⁹

ELECTROCONVULSIVE THERAPY DURING PREGNANCY

The use of ECT during pregnancy typically raises considerable anxiety among clinicians and patients. Its safety record has been well documented over the last 50 years.^134–136 Requests for psychiatric consultation on pregnant patients requiring ECT tend to be emergent and dramatic. For example, expeditious treatment is imperative in instances of mania during pregnancy or psychotic depression with suicidal thoughts and disorganized thinking. Such clinical situations are associated with a danger from impulsivity or self-harm. The safety and efficacy of ECT in such settings are well described, particularly when instituted in collaboration with a multidisciplinary treatment team, including an anesthesiologist, psychiatrist, and obstetrician.^71,136–138 A limited course of treatment may be sufficient followed by institution of treatment with one or a combination of agents, such as antidepressants, neuroleptics, benzodiazepines, or mood stabilizers.

ECT during pregnancy tends to be underused because of concerns that treatment will harm the fetus. Despite one report of placental abruption associated with the use of ECT during pregnancy,¹³⁹ considerable experience supports its safe use in severely ill gravid women. Thus, it becomes the task of the psychiatric consultant to facilitate the most clinically appropriate intervention in the face of partially informed concerns or objections.

BREAST-FEEDING AND PSYCHOTROPIC DRUG USE

The emotional and medical benefits of breast-feeding to mother and infant are clear. Given the prevalence of psychiatric illness during the postpartum period, a significant number of women may require pharmacological treatment while nursing. Appropriate concern is raised, however, regarding the safety of psychotropic drug use in women who choose to breast-feed while using these medications. Efforts to quantify psychotropic medications and their metabolites in the breast milk of mothers have been reported. Serum of infants can also be assayed to assess more accurately actual neonatal exposure to medications. The data indicate that all psychotropic medications, including antidepressants, antipsychotic agents, lithium carbonate, and benzodiazepines, are secreted into breast milk. However, concentrations of these agents in breast milk vary considerably. The amount of medication to which an infant is exposed depends on several factors: the maternal dosage of medication, frequency of dosing, and rate of maternal drug metabolism.¹⁴⁰ Typically, peak concentrations in the breast milk are attained approximately 6 to 8 hours after the medication is ingested. Thus, the frequency of feedings and the timing of the feedings can influence the amount of drug to which the nursing infant is exposed. By restricting breast-feeding to times during which breast milk drug concentrations would be at their lowest (either shortly before or immediately after dosing medication), exposure may be reduced; however, this approach may not be practical for newborns, who typically feed every 2 to 3 hours.

The nursing infant’s chances of experiencing toxicity are dependent not only on the amount of medication ingested but also on how well the ingested medication is metabolized. Most psychotropic medications are metabolized by the liver. During the first few weeks of a full-term infant’s life, there is a lower capacity for hepatic drug metabolism, which is about one-third to one-fifth of the adult capacity. Over the next few months, the capacity for hepatic metabolism increases significantly and, by about 2 to 3 months of age, it surpasses that of adults. In premature infants or in infants with signs of compromised hepatic metabolism (e.g., hyperbilirubinemia), breast-feeding typically is deferred because these infants are less able to metabolize drugs and are thus more likely to experience toxicity.

Over the past 5 years, data have accumulated regarding the use of various psychotropic medications during breast-feeding.^140–143 Much data have accumulated on the use of antidepressants in nursing women. The available data particularly on the TCAs, fluoxetine, paroxetine, and sertraline during breast-feeding have been encouraging and suggest that the amount of drug to which the nursing infant is exposed is low and that significant complications related to neonatal exposure to psychotropic medications in breast milk appear to be rare.^144–151 Typically, very low or nondetectable levels of drug have been detected in the infant serum, and one report indicates that exposure during nursing does not result in clinically significant blockade of serotonin (5-HT) reuptake in infants.¹⁵¹ Although less information is available on other antidepressants, serious adverse events related to exposure to these medications have not been reported.^141,142

Given the prevalence of anxiety symptoms during the postpartum period, anxiolytic agents often are used in this setting. Data regarding the use of benzodiazepines have been limited; however, the available data suggest that amounts of medication to which the nursing infant is exposed are low.¹⁴⁴ Case reports of sedation, poor feeding, and respiratory distressin nursing infants have been published^127,152; however, the data, when pooled, suggest a relatively low incidence of adverse events.^141,144

For women with BPD, breast-feeding may pose more significant challenges. First, on-demand breast-feeding may significantly disrupt the mother’s sleep and thus may increase her vulnerability to relapse during the acute postpartum period. Second, there have been reports of toxicity in nursing infants related to exposure to various mood stabilizers, including lithium and carbamazepine, in breast milk. Lithium is excreted at high levels in the mother’s milk, and infant serum levels are relatively high, about one-third to one-half of the mother’s serum levels,^153–155 thereby increasing the risk of neonatal toxicity. Reported signs of toxicity include cyanosis, hypotonia, and hypothermia.¹⁵⁶ Although breast-feeding typically is avoided in women taking lithium, the lowest possible effective dosage should be used and both maternal and infant serum lithium levels should be followed in mothers who breast-feed. In collaboration with the pediatrician, the child should be monitored closely for signs of lithium toxicity, and lithium levels, thyroid-stimulating hormone (TSH), blood urea nitrogen (BUN), and creatinine should be monitored every 6 to 8 weeks while the child is nursing.

Several studies have suggested that lamotrigine reaches infants through breast milk in relatively high doses, ranging from 20% to 50% of the mother’s serum concentrations,^157,158 which may be explained by poor neonatal metabolism of lamotrigine. In addition, maternal serum levels of lamotrigine increased significantly after delivery, which may also have contributed to the high levels found in nursing infants. None of these studies has reported any adverse events in breast-feeding newborns. One worry shared by clinicians and new mothers is the risk for Stevens-Johnson syndrome (SJS). This is a severe and potentially life-threatening rash, most commonly resulting from a hypersensitivity reaction to a medication, which occurs in about 0.1% of bipolar patients treated with lamotrigine.¹⁵⁹ Thus far, there have been no reports of SJS in infants associated with exposure to lamotrigine. In fact, it appears that cases of drug-induced SJS are extremely rare in newborns. In a single case report, authors described a neonate developing the syndrome after exposure to the anticonvulsant phenobarbital.¹⁶⁰

Similarly, concerns have arisen regarding the use of carbamazepine and valproic acid. Both of these mood stabilizers have been associated in adults with abnormalities in liver function and fatal hepatotoxicity. Hepatic dysfunction associated with carbamazepine exposure in breast milk has been reported several times.^161,162 Most concerning is that the risk for hepatotoxicity appears to be greatest in children younger than 2 years old; thus, nursing infants exposed to these agents may be particularly vulnerable to serious adverse events. Although the American Academy of Pediatrics has deemed both carbamazepine and valproic acid to be appropriate for use in breast-feeding mothers, few studies have assessed the impact of these agents on fetal well-being, particularly in nonepileptic mothers. In those women who choose to use valproic acid or carbamazepine while nursing, routine monitoring of drug levels and liver function tests is recommended. In this setting, ongoing collaboration with the child’s pediatrician is crucial.

Consultation about the safety of breast-feeding among women treated with psychotropic medications should include a discussion of the known benefits of breast-feeding to mother and infant and the possibility that exposure to medications in the breast milk may occur. Although routine assay of infant serum drug levels was recommended in earlier treatment guidelines, this procedure is probably not warranted; in most instances low or nondetectable infant serum drug levels will be evident and, infrequently, serious adverse side effects have been reported. This testing is indicated, however, if neonatal toxicity related to drug exposure is suspected. Infant serum monitoring is also indicated when the mother is nursing while taking lithium, valproic acid, or carbamazepine.

PSYCHIATRIC CONSULTATION AND POSTPARTUM PSYCHIATRIC ILLNESS

The postpartum period has typically been considered a time of risk for the development of affective disorder.¹⁹ Although several studies suggest rates of depression during the postpartum period equal to those in nonpuerperal controls, other research has identified subgroups of women at particular risk for postpartum worsening of mood.^21,163–165 At highest risk are women with a history of postpartum psychosis; up to 70% of women who have had one episode of puerperal psychosis will experience another episode following a subsequent pregnancy.^19,165 Similarly, women with histories of postpartum depression are at significant risk, with rates of postpartum recurrence as high as 50%.¹⁶⁶ Women with BPD also appear to be particularly vulnerable during the postpartum period, with rates of postpartum relapse ranging from 30% to 50%.^88,164,167 The extent to which a history of MDD influences risk for postpartum illness is less clear. However, in all women (with or without histories of major depression), the emergence of depressive symptoms during pregnancy significantly increases the likelihood of postpartum depression.²¹

POSTPARTUM MOOD AND ANXIETY DISORDERS: DIAGNOSIS AND TREATMENT

During the postpartum period, about 85% of women experience some degree of mood disturbance. For most women the symptoms are mild; however, 10% to 15% of women experience clinically significant symptoms. Postpartum depressive disorders typically are divided into three categories: (1) postpartum blues, (2) nonpsychotic major depression, and (3) puerperal psychosis. Because these three diagnostic subtypes overlap significantly, it is not clear if they actually represent three distinct disorders. It may be more useful to conceptualize these subtypes as existing along a continuum, where postpartum blues is the mildest and postpartum psychosis the most severe form of puerperal psychiatric illness.

Postpartum blues does not indicate psychopathology, but it is common and occurs in approximately 50% to 85% of women following delivery.^168,169 Symptoms of reactivity of mood, tearfulness, and irritability are, by definition, time limited and typically remit by the tenth postpartum day. As postpartum blues is associated with no significant impairment of function and is time-limited, no specific treatment is indicated. Symptoms that persist beyond 2 weeks require further evaluation and may suggest an evolving depressive disorder. In women with histories of recurrent mood disorder, the blues may herald the onset of postpartum major depression.^21,170

Several studies describe a prevalence of postpartum major depression of between 10% and 15%.^22,168 The signs and symptoms of postpartum depression usually appear over the first 2 to 3 months following delivery and generally are indistinguishable from characteristics of MDD that occur at other times in a woman’s life. The symptoms of postpartum depression include depressed mood, irritability, and loss of interest in usual activities. Insomnia, fatigue, and loss of appetite are frequently described. Postpartum depressive symptoms also commingle with anxiety and obsessional symptoms, and women may have generalized anxiety, panic disorder, or hypochondriasis.^171,172 Although it may sometimes be difficult to diagnose depression in the acute puerperium given the normal occurrence of symptoms suggestive of depression (e.g., sleep and appetite disturbance, low libido), it is an error to dismiss neurovegetative symptoms, such as severe decreased energy, profound anhedonia, and guilty ruminations, as normal features of the puerperium. In its most severe form, postpartum depression may result in profound dysfunction. Risk factors for postpartum depression include prenatal depression, prenatal anxiety, and a history of previous depression.

A wealth of literature on this topic indicates that postpartum depression, especially when left untreated, may have a significant impact on the child’s well-being and development.^173,174 In addition, the syndrome demands aggressive treatment to avoid the sequelae of an untreated mood disorder, such as chronic depression and recurrent disease. Treatment should be guided by the type and severity of the symptoms and by the degree of functional impairment. However, before initiating psychiatric treatment, medical causes for mood disturbance (e.g., thyroid dysfunction, anemia) must be excluded. Initial evaluation should include a thorough history, physical examination, and routine laboratory tests.

Although postpartum depression is relatively common, few studies have systematically assessed the efficacy of nonpharmacological and pharmacological therapies in the treatment of this disorder. Nonpharmacological therapies are useful in the treatment of postpartum depression, and several preliminary studies have yielded encouraging results. Appleby and associates¹⁷⁵ have demonstrated in a randomized study that short-term CBT was as effective as treatment with fluoxetine in women with postpartum depression. IPT has also been shown to be effective for the treatment of women with mild to moderate postpartum depression.¹⁷⁶

These nonpharmacological interventions may be particularly attractive to those patients who are reluctant to use psychotropic medications (e.g., women who are breast-feeding) or for patients with milder forms of depressive illness. Further investigation is required to determine the efficacy of these treatments in women who suffer from more severe forms of postpartum mood disturbances. Women with more severe postpartum depression may choose to receive pharmacological treatment, either in addition to or instead of nonpharmacological therapies.

To date, only a few studies have systematically assessed the pharmacological treatment of postpartum depression. Conventional antidepressant medications (e.g., fluoxetine, sertraline, and venlafaxine) have shown efficacy in the treatment of postpartum depression.^{57,175,177–179} In all of these studies, standard antidepressant doses were effective and well tolerated. The choice of an antidepressant should be guided by the patient’s prior response to antidepressant medication and a given medication’s side-effect profile. SSRIs are ideal first-line agents because they are anxiolytic, nonsedating, and well tolerated, and bupropion is also another good option. TCAs are used frequently, and, because they tend to be more sedating, may be more appropriate for women who have prominent sleep disturbances. Given the prevalence of anxiety in women with postpartum depression, adjunctive use of a benzodiazepine (e.g., clonazepam or lorazepam) may be very helpful.

Some investigators have also explored the role of hormonal manipulation in women who suffer from postpartum depression. The postpartum period is associated with rapid shifts in the reproductive hormonal environment, most notably a dramatic fall in estrogen and progesterone levels, and postpartum mood disturbance has been attributed to a deficiency (or change in the levels) in these gonadal steroids. Although early reports suggested that progesterone may be helpful,¹⁸⁰ no systematically derived data exist to support its use in this setting. Two studies have described the benefit of exogenous estrogen therapy, either alone or in conjunction with an antidepressant in women with postpartum depression.^181–183 Although these studies suggest a role for estrogen in the treatment of women with postpartum depression, these treatments remain experimental. Estrogen delivered during the acute postpartum period is not without risk and has been associated with changes in breast-milk production and more significant thromboembolic events. Antidepressants are safe, well tolerated, and highly effective; they remain the first choice for women with postpartum depression.

In cases of severe postpartum depression, inpatient hospitalization may be required, particularly for patients who are at risk for suicide. In Great Britain, innovative treatment programs involving joint hospitalization of the mother and the baby have been successful; however, mother-infant units are much less common in the United States. Women with severe postpartum illness should be considered candidates for ECT. The option should be considered early in treatment because it is safe and highly effective. In choosing any treatment strategy, it is important to consider the impact of prolonged hospitalization or treatment of the mother on infant development and attachment.

Although symptoms of postpartum panic attacks and OCD symptoms are frequently included in the description of postpartum mood disturbance, a growing literature supports the likelihood that postpartum anxiety disorders are discrete diagnostic entities.^109,172 Several investigators have described postpartum worsening of panic disorder in women with pregravid histories of this anxiety disorder but with an absence of co-morbid depressive illness.⁴ Postpartum OCD has also been described in the absence of co-morbid postpartum major depression. Symptoms often include intrusive obsessional thoughts to harm the newborn in the absence of psychosis. Treatment with antiobsessional agents, such as fluoxetine or clomipramine, has been effective.¹³³

Postpartum psychosis is a psychiatric emergency. The clinical picture is most frequently consistent with mania or a mixed state¹⁹ and may include symptoms of restlessness, agitation, sleep disturbance, paranoia, delusions, disorganized thinking, impulsivity, and behaviors that place mother and infant at risk. The typical onset is within the first 2 weeks after delivery, and symptoms may appear as early as the first 48 to 72 hours postpartum. Although investigators have debated whether postpartum psychosis is a discrete diagnostic entity or a manifestation of BPD, treatment should follow the same algorithm to treat acute manic psychosis, including hospitalization and potential use of mood stabilizers, antipsychotic medications, benzodiazepines, or ECT.

Although it is difficult to reliably predict which women will experience a postpartum mood disturbance, it is possible to identify certain subgroups of women (i.e., women with a history of mood disorder) who are more vulnerable to postpartum affective illness. Several investigators have explored the potential efficacy of prophylactic interventions in these women at risk.^184–187

Several studies demonstrate that women with histories of BPD or puerperal psychosis benefit from prophylactic treatment with lithium instituted either before delivery (at 36 weeks of gestation) or no later than the first 48 hours following delivery.^184–187 Prophylactic lithium appears to significantly reduce relapse rates and diminish the severity and duration of puerperal illness.

For women with histories of postpartum depression, Wisner and colleagues¹⁸⁸ have described a beneficial effect of prophylactic antidepressant (either a TCA or an SSRI) administered after delivery. However, a subsequent randomized, placebo-controlled study from the same group did not demonstrate a positive effect in women treated prophylactically with nortriptyline.¹⁸⁹ The authors have suggested that nortriptyline may be less effective than SSRIs for the treatment of postpartum depression. The efficacy of prophylactic treatment with SSRIs in this population is under investigation.

In summary, postpartum depressive illness may be conceptualized along a continuum, where some women are at lower risk for puerperal illness and others are at higher risk. Although a less aggressive, “wait-and-see” approach is appropriate for women with no history of postpartum psychiatric illness, women with BPD or histories of postpartum psychiatric illness deserve not only close monitoring but also specific prophylactic measures.

PERINATAL PSYCHIATRY: FROM SCREENING TO TREATMENT

Clinicians who manage the care of female psychiatric patients before, during, and after pregnancy may be called to evaluate women who experience a broad spectrum of difficulties. Symptoms may be mild, although the consultant is typically requested when symptoms become severe. Psychiatric disorders may emerge anew during pregnancy, although more often clinical presentations represent persistence or exacerbation of already existing illness. Physicians therefore should screen more aggressively for psychiatric disorders either before conception or during pregnancy, integrating questions about psychiatric symptoms and treatment into the obstetric history. Identification of “at-risk” women allows the most thoughtful, acute treatment before, during, and after pregnancy and signals the opportunity to institute prophylactic strategies that prevent psychiatric disturbances in women during the childbearing years.

One recent report has described the finding that even among women with identified psychiatric illness during pregnancy, definitive treatment is frequently lacking or incomplete.¹⁵ The extent to which women suffering from postpartum psychiatric illness are undertreated as a group is also very well described. Perhaps one of the reasons for failure to treat women with psychiatric disorders during pregnancy is the concern regarding fetal exposure to psychotropics. Many clinicians can conceptualize the need to weigh relative risks of fetal exposure on one hand versus the risk of withholding treatment on the other. However, given the inability to absolutely quantify these risks, clinicians often defer treatment entirely and consequently put patients at risk for the sequelae of untreated maternal psychiatric illness. Clinicians should realize that the process of managing psychiatric illness during pregnancy and the puerperium is not a process like threading a needle; it is not necessarily clear-cut, and much treatment described in the literature is not evidenced based. However, thoughtful decisions can still be made with these patients as clinicians review available information with them and as both clinician and patient realize that no decision is risk free and no decision is perfect.

Thoughtful treatment decisions can be made nonetheless, taking into account available information regarding relative risks of treatment options and respective patient wishes.