Depression is a common complication of pregnancy and the postpartum period, with significant implications for a woman, her fetus, and her family. The biological, psychological, and social changes inherent in this time all contribute to the propensity toward mood disorder during this critical period. Women with mood disorders during the perinatal period are faced with challenging decisions when considering treatment. Mental health and obstetrical providers have important roles in helping women and their families weigh the impact of untreated mental health symptoms against the impact of potential treatments.

Case illustration: Shanté, a 23-year-old single woman, gave birth to her first child 6 weeks ago. Labor and delivery had been painful but uncomplicated, and her baby, a son she named Ryan, was healthy. Shanté was happy, although tired. She began to feel vulnerable and upset when she tried breastfeeding Ryan and he had difficulty latching on. Although the hospital nurse assured her this happens often, and a lactation consultant gave her tips about helping Ryan to latch, Shanté began to wonder if she would be able to parent Ryan successfully. After taking Ryan home 2 days later, she began to feel overwhelmed. She became afraid to fall asleep in case she would miss Ryan’s crying or be unprepared to feed him. She felt exhausted. By 3 weeks postpartum, she felt emotionally numb and lost interest in food. She stopped showering, getting dressed or answering her phone, instead sitting leadenly in a chair near Ryan’s crib as he napped. One night Ryan slept longer than usual, and Shanté became convinced he was dead. Even after seeing that he was breathing comfortably, she had the thought that unlike other mothers, she would be unable to keep her baby alive, so did not deserve to be alive herself. She had fleeting thoughts of overdosing on over-the-counter pills from the medicine cabinet, but did not act on these because she felt she could not burden anyone else with taking care of Ryan.

EPIDEMIOLOGY

Depression is one of the most common disorders in the perinatal period, with higher prevalence than gestational hypertension or diabetes mellitus (Fig. 9-1). The risk of depression is elevated during pregnancy and the first year postpartum, particularly during the first 3 months postpartum. Meta-analysis shows a 7.5% incidence of major depression during pregnancy and 6.5% in the first 3 months of postpartum¹ with point prevalence of major and minor depression of 11.0% in the first trimester and 8.5% later in pregnancy. In the first postpartum year, 13.9% of mothers and 3.6% of fathers develop unipolar, nonpsychotic major depression.² Point prevalence rises postpartum to a high of 12.9% in the 3rd month, then drops to 9.9% to 10.6% from the 4th to 12th postpartum months, declining after that to 6.5%.¹ The incidence of major depression was higher for mothers in the first year postpartum than in subsequent years.

Perinatal depression prevalence varies widely among subsets of women. Comparing women from 40 countries,³ prevalence of postpartum major depression using DSM criteria ranged from 3.4% in a Swedish sample to 34.7% in a South African sample. Similarly, studies using Edinburgh Postnatal Depression Scale scores to measure depression show a range of 0.5% in Singapore to 57.0% in Guyana.³ In the United States, ethnicity, socioeconomic status, age, and trauma history serve as both independent and interactive risk factors. The odds of developing perinatal depressive symptoms are higher in African American women and Hispanic women than in non-Hispanic Caucasian women, even after adjusting for other demographic variables.^4,5 Although the cause of these differences is unknown, possible factors include differences in levels of exposure to environmental stressors, availability of protective resources, stigma, biological vulnerabilities, interpretations and conceptualizations of distress, and communication of symptoms. In studies using structured clinical interviews rather than self-reports, prevalence of major depression is comparable among different socioeconomic groups, while the prevalence of sub-syndromal depressive symptoms is higher among women of lower socioeconomic position.¹ Adolescents have a higher risk for perinatal depression than adults, especially in the context of early life trauma.^6,7

Women with pre-existing bipolar disorder have significantly more episodes during pregnancy and postpartum than women with unipolar depression, and significantly more episodes postpartum than during pregnancy.⁸ Among untreated women with bipolar disorder, the rate of recurrence during pregnancy is doubled and tripled during the postpartum period.^9,10 Most postpartum bipolar mood episodes occur in the first month postpartum.¹¹ In addition, while postpartum psychotic episodes are relatively rare (1 in 1,000 women who give birth),¹² women with pre-existing bipolar disorder are at substantially higher risk of psychotic and nonpsychotic episodes.

PATHOPHYSIOLOGY

The perinatal period is a time of substantial biological and social flux. Hormonal shifts, heightened inflammatory responses, and changes in circadian rhythm occur throughout pregnancy and in the postpartum period (Box 9-1 and Box 9-2). Genetic susceptibility intensifies mood changes in response to these perturbations. Table 9-1 summarizes factors which have been shown to increase vulnerability to perinatal depression.^13–20

Postpartum psychotic mood episodes are not associated with stress, lack of social support, or negative attitudes toward pregnancy.¹² There is a strong familial aggregation of postpartum psychotic mood episodes. In a genetic study of bipolar disorder, women with a family history of puerperal psychosis were six times more likely to have a psychotic episode within 6 weeks of giving birth.²¹

CLINICAL PRESENTATION

Major depressive episodes present differently in the antepartum, postpartum, and nonperinatal periods (Box 9-3).²² When compared to women with onset of major depression postpartum, women with new episodes of major depression during pregnancy are more likely to have had unplanned pregnancies, inadequate social support, a history of physical and/or sexual abuse, and a history of major depression, as well as more complicated pregnancies and/or neonatal medical problems. In contrast, postpartum depressive episodes are more likely to include obsessive symptoms. Ego-dystonic, intrusive, violent thoughts, especially about harming the baby, are relatively common postpartum. In one study, 41% of mothers with postpartum depression and 7% of nondepressed control mothers had such thoughts.²³ The risk of women acting on such thoughts is low. By contrast, the delusional, ego-syntonic thoughts of harming babies which sometimes occur during episodes of postpartum psychosis pose substantially greater risk of being acted upon.

There are phenomenologic differences between postpartum and nonpostpartum psychotic mood episodes as well. Postpartum psychotic episodes are more often characterized by a subjective sense of confusion, disorganized thinking, lability, and homicidal ideation.²⁴ The waxing and waning nature of postpartum psychotic symptoms can contribute to underrecognition or underestimation of risk; patients can manifest severe symptoms at times and can appear much better hours later, with symptoms intensifying again shortly thereafter.

COURSE AND NATURAL HISTORY

Women with postpartum depressive symptoms or episodes have a high likelihood of subsequent depression, both during postpartum periods and in general, which may be as much as four times the risk for mothers without perinatal depression.²⁵ On the other hand, women whose first major depressive episode occurred within the first 4 weeks postpartum have a lower risk of recurrence than women those whose first episode is unrelated to pregnancy.²⁶ Effective treatment reduces recurrence rates.²⁷ Women with postpartum bipolar mood episodes have a particularly high likelihood of subsequent episodes.²⁸

Perinatal depression has intergenerational effects. Thus maternal antenatal depression is associated with an increased risk of preterm birth²⁹ and long-term risks to offspring. These risks may stem in part from depression-linked maternal behavioral changes during pregnancy, such as reduced prenatal care, worse nutrition, reduced physical activity, and increased use of alcohol and tobacco. Some effects of maternal depression may be mediated via fetal programming, a term for epigenetic influences on fetal development that have enduring effects on emotions, behavior or cognitive abilities. Antenatal maternal depression is associated with increased methylation affecting glucocorticoid receptor gene expression, potentially influencing fetal hypothalamic pituitary adrenal (HPA) axis programming.³⁰ Resultant increased risks for offspring include poorer growth, increased risk of infection, and more difficult temperaments.^31–33

Postpartum depression can adversely affect parenting capability, although it does not do so consistently. Mothers with postpartum depression may have impaired ability to read infant cues, reduced responsiveness to infant distress, constricted ability to communicate a full range of emotions, provide fewer enrichment activities for babies, and maintain less healthy feeding and sleeping practices.³⁴ Children and adolescents whose mothers were depressed postpartum experience poorer health-related quality of life,³⁵ higher rates of HPA abnormalities and depression,³⁶ higher rates of behavioral disturbance,³⁷ and lower cognitive abilities.³⁸

Women with bipolar disorder are at greater risk during pregnancy for addictive substance use, being overweight, and needing instrumental delivery or Cesarean section.³⁹ Effects on offspring are not well studied, though there may be increased likelihood of low birth weight (OR 1.66), premature birth (OR 2.08), and being small for gestational age (OR 1.47).⁴⁰ In a controlled study of mother–infant dyadic interactions, mothers with bipolar disorder were less expressive and showed more negative communication styles.⁴¹

ASSESSMENT AND DIFFERENTIAL DIAGNOSIS

Normal somatic changes of pregnancy, such as increased appetite, weight gain, sleep disturbances, and reduced energy, overlap with the diagnostic criteria for depression (Box 9-4). The Edinburgh perinatal depression scale (EPDS) is a valid and sensitive tool that is less likely to be influenced by these potential confounds.^42,43 It is a 10-item self-report tool available in over 20 languages. The EPDS has a sensitivity of 0.88 to 1.00 and a specificity from 0.71 to 0.79 using a cutoff score ranging from 10 to 12.⁴³ Several general depression screening tools, including the Beck depression inventory (BDI), the Center for epidemiologic studies depression (CES-D) Scale, and the patient health questionnaire (PHQ), have also been validated for perinatal use, and compare favorably with the EPDS. The postpartum depression screening scale is specifically for postpartum use, taking into account the context of new motherhood,⁴⁴ and though less well studied, has psychometric properties that compare favorably to the EPDS.

Normal postpartum mood changes known as “baby blues” may be confused with postpartum depression. Although the term “blues” suggests a mild form of depression, normal postpartum “blues” are a qualitatively different experience in which the predominant mood is happiness⁴⁵ that is interspersed with intense reactivity of mood and heightened emotional responses to stimuli, resulting in frequent tearfulness and mood lability. For example, a woman with postpartum “blues” might be happy most of the time, but cry upon hearing a love song, feel highly anxious when reading about a train accident, and become irritable and overwhelmed from everyday stresses. These mood changes occur in approximately 50% of postpartum women, peaking from days 3 to 5 after delivery and lasting several days to weeks.⁴⁶

Medical conditions that can present with perinatal depressive symptoms include:

• 
  

  Thyroid disease: Increased thyroid synthesis is required to meet fetal needs during pregnancy, and subclinical hypothyroidism in the antenatal period may confer a higher risk of postpartum depressive symptoms.^47,48 The postpartum period also confers a heightened risk of thyroiditis.

  
  
• 
  

  Anemia: Repleting iron in pregnant women with iron-deficiency anemia has been shown to reduce rates of clinically significant depression.⁴⁹

  
  
• 
  

  Vitamin D deficiency: Low vitamin D levels are associated with increased perinatal depressive symptoms,⁵⁰ however, it is unknown whether vitamin D supplementation reverses depressive symptoms in the absence of other treatment for depression.

TREATMENT

Treatments with well-established efficacy for perinatal depression include interpersonal psychotherapy (IPT), cognitive-behavioral therapy (CBT), electroconvulsive therapy (ECT), and antidepressant medication (Box 9-5). Estrogen therapy, phototherapy, transcranial magnetic stimulation, and omega-3 fatty acid supplementation also appear to have promising efficacy though further study of these interventions is needed.

ANTIDEPRESSANTS

Many conflicting findings have emerged in studies about the risks of antidepressant medications during pregnancy. This is in part because there are no randomized, double-blind, placebo-controlled trials. Retrospective database studies track prescriptions rather than actual medication exposure and chart-recorded diagnoses rather than active symptoms. Many such studies test correlations among numerous congenital anomalies and multiple medications, often without statistical correction for multiple queries. In addition, correcting for the influence of major confounds is difficult. The most reliable data for the risks of antidepressant medications come from adequately powered, prospective, controlled studies and meta-analyses of these studies, summarized in Table 9-2.^51–58

For women with moderate to severe and/or recurrent depressive episodes, decisions about the use of antidepressant medication during pregnancy must take into account the likelihood of relapse and the risks of untreated symptoms. A study of pregnant psychiatric patients suggested an increased incidence of depressive relapse following discontinuation of treatment during pregnancy,⁵⁹ whereas an obstetrical sample was not found to be at increased risk of relapse following discontinuation of antidepressant medication.⁶⁰ This conflicting evidence may relate to differences in severity of disease and/or higher dose/duration of therapy.

In the postpartum period, the decision to use antidepressants must consider effects on sleep, weight, and sexual function. Sedating agents, even if they do not cause daytime drowsiness, may interfere with a woman’s ability to hear and respond safely to her baby during the night. Agents that interfere with sleep may compound postpartum sleep deprivation. Medications that cause weight gain may interfere with a woman’s ability to lose pregnancy weight. Adverse effects on sexual function can add to postpartum stresses as partners attempt to maintain intimate relationships as they transition to parenthood.

For women who are breastfeeding, the effects of medications on their breastfeeding babies are also key factors in choice of agent. Table 9-3 summarizes breastfeeding exposure (as estimated weight-adjusted percent of maternal dose received by a breastfeeding baby) and reported effects on breastfeeding babies for commonly used antidepressants.⁶¹

LITHIUM

Treatment of bipolar depression during pregnancy can be especially challenging. Lithium fully equilibrates across the placenta,⁶² and may be a weak cardiovascular teratogen, although findings are equivocal. Although individual studies have reported an increased risk of cardiovascular defects, especially right-sided defects, such as Ebstein’s anomaly, prospective studies, and meta-analysis, have not.⁶³ In utero lithium, exposure has not been found to affect subsequent growth or neurodevelopment.⁶⁴ Lithium can cause transient fetal nephrogenic diabetes insipidus (DI) by impairing the concentrating ability of the fetal kidney,⁶⁵ which can present with polyhydramnios and resultant uterine enlargement. As this progresses, it can cause maternal shortness of breath if the diaphragm has insufficient room to fully expand. DI can last a week or more after birth. Lithium is also associated with premature labor⁶⁶ and transient neonatal hypothyroidism.⁶⁷ Newborns can develop toxicity, which is greater at higher serum levels and associated with lower Apgar scores and longer infant hospital stays.⁶⁸ Signs of neonatal toxicity include flaccidity, hypotonia, hyporeflexia (decreased Moro reflex, poor suck), cyanosis, lethargy, poor feeding, abnormal breathing, cardiac arrhythmias, and poor myocardial contractility. Perinatal women are also at heightened risk of lithium toxicity under certain circumstances, summarized in Table 9-4.⁶²

Guidelines for prescribing lithium during pregnancy include:

• 
  

  Avoid lithium during the period of cardiac formation when possible. Consider ultrasound evaluation for cardiac anomalies during weeks 16 to 18 when there has been early exposure.

  
  
• 
  

  Monitor for signs of fetal DI (maternal shortness of breath, uterine enlargement, and/or polyhydramnios).

  
  
• 
  

  Monitor lithium serum level. Dose increases are often required as pregnancy progresses because pregnancy-linked pharmacokinetic changes, including increased glomerular filtration rate and increased total body water, may reduce serum lithium level at a given dose.

  
  
• 
  

  Consider divided doses near the end of pregnancy to reduce toxic peaks for the newborn.

  
  
• 
  

  Consider cutting the dose in half at labor onset to reduce the risks of maternal and neonatal toxicity, then resuming the pre-pregnancy dose postpartum.

Breastfeeding babies whose mothers are taking lithium have serum lithium levels that average about 25% of maternal levels, but can be higher. Infants are vulnerable to lithium toxicity if they become dehydrated. Elevations of TSH, BUN, and creatinine are seen in infants exposed to lithium in breast milk. Monitoring infant serum lithium level, TSH, BUN, and creatinine is warranted.⁶⁹

ANTIEPILEPTIC DRUGS

Table 9-5 summarizes the rates of birth defects associated with valproate, carbamazepine, and lamotrigine.^70,71–73