Postpartum Mental Disorders

The spectrum of postpartum mood disorders has come into clearer focus in recent years.^1,2 Postpartum psychosis, a relatively infrequent but serious outcome of childbirth, has long been studied.^3,4 Postpartum depression and blues are less serious than psychosis; however, they also are potentially difficult problems for women who experience them.⁵ The purpose of this chapter is to describe each of these three conditions, present data regarding their incidence and prevalence, identify potential causal factors and consequences, and discuss issues related to diagnosis and treatment.

Postpartum blues, depression, and psychosis are nowhere listed in the Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV),⁶ although there is a “postpartum onset specifier.” The postpartum onset specificer can be applied to the current or most recent major depressive, manic, or mixed episode of major depressive disorder, bipolar I disorder, bipolar II disorder, or brief psychotic disorder if the onset of the episode is within 4 weeks of childbirth. There are clear differences in the severity of the typical cases within each of these disorders. However, it is unclear whether these disorders are actually three distinct entities, or whether they all fall under the heading of postpartum mood disorders, in which symptoms range along a continuum of severity from little or no disturbance to severe disturbance.⁷ There are no natural dividing points to separate one of these disorders from the other, and the symptoms of women at the boundaries will be very similar to each other. Nevertheless, with the exception of the causal factors and treatment sections, each of these conditions in this review is discussed separately in accordance with the current literature.

Blues

Women commonly experience mild dysphoria during the first week after delivery.^8,9 Symptoms such as depressed mood, crying spells, irritability, anxiety, mood lability, confusion, and sleep and appetite disturbance are often reported.^8,10 Despite the connotation of the “blues,” women may experience symptoms such as crying, confusion, anxiety, and mood lability more so than sadness or depression.¹¹ Symptoms typically begin within a few days of delivery and may last from a few hours to several days. Various criteria have been used to diagnose the blues. Pitt¹⁰ diagnosed women as having the blues if they “felt tearful and depressed in the puerperium” (defined as within the first week to 10 days after delivery). Stein¹² reported that a score of 8 on his blues questionnaire (range, 0 to 26) usually indicated that a mood swing has occurred (on that day). O'Hara and associates,⁸ building on the work of Handley and colleagues¹³ specified seven blues symptoms: dysphoric mood, mood lability, crying, anxiety, insomnia, loss of appetite, and irritability. At least four of the seven symptoms had to be present and definitely noticeable to the patient or others (assessed in the context of a diagnostic interview) for a diagnosis of the blues to be made.⁸

Depression

Postpartum depression refers to those nonpsychotic mood disorders that meet the criteria for major or minor depression.^6,14 It falls in the middle range of severity between the blues and psychosis. The concept of postpartum depression per se is of relatively recent origin. For example, Winn¹⁵ only alluded to postpartum depression when he said, “There are at least two separate syndromes, with a possible phase that would appear to be a transitional one between the milder conditions and the more serious and disruptive one. The first illness is a very mild depression, or the ‘postpartum blues,’ and the second is the postpartum psychosis.” Prospective studies of childbearing women beginning in the late 1960s¹⁶ began to identify large numbers of women who experienced significant depressive symptoms after delivery but who tended not to be in treatment. Pitt¹⁶ referred to these depressions as “atypical” because they did not fit the typical endogenous pattern, which includes symptoms such as suicidal ideation, worsening of depression in the morning, and early-morning awakening. Although most postpartum depressions are probably nonendogenous, it has become increasingly common to use standard diagnostic criteria for postpartum depression (e.g., DSM-IV⁶). With respect to symptom characteristics specific to postpartum episodes, the DSM-IV notes that a fluctuating course and mood lability may be more common in postpartum than in non-postpartum episodes.⁶

Psychosis

The most severe form of psychiatric disturbance in the puerperium is postpartum psychosis. Women experiencing postpartum psychosis are grossly impaired in their ability to function, usually because of hallucinations or delusions. In other cases, women may be disturbed by severely depressed mood or confusion. Although postpartum psychosis was once believed to be a distinct diagnostic category, and this position is still maintained by some,¹⁷ the current opinion is that psychoses occurring in the puerperium are no different from psychoses occurring at other times in a woman's life.^6,18,19 The major reason for this change in perspective is that several investigations have revealed little difference between puerperal and nonpuerperal psychoses.^18,20

Although the major psychiatric disorders are recognized in the puerperium (e.g., major depression, mania, schizophrenia), studies have found that puerperal psychotic patients show some differences from nonpuerperal psychotic patients in the frequency with which certain symptoms occur. For example, Brockington and colleagues²¹ compared 58 puerperal and 52 nonpuerperal psychotic women and found that the puerperal psychotic patients experienced higher levels of euphoria, activity, incompetence, and confusion than nonpuerperal psychotic women. Nonpuerperal psychotic patients experienced higher levels of symptoms typical of schizophrenia than puerperal psychotics (e.g., odd affect, paranoia, systematization of delusions, social withdrawal, and hostility). These findings are understandable given that the puerperal group was composed primarily of depressed and manic women and the nonpuerperal group had a high proportion of schizophrenic and schizoaffective women. Nevertheless, Brockington and co-workers argued that postpartum psychosis may well have a single etiologic factor that results in an acute affective psychosis with florid psychotic symptoms. A later study²² with a better balance in the proportion of affective and nonaffective psychoses in the childbearing and non-childbearing subjects obtained results similar to those of Brockington and colleagues.²¹

Dean and Kendell²³ compared two groups of women in an inpatient setting receiving treatment, one for postpartum mental illness and the other for non-postpartum disorders. Subjects were matched by age, year of admission, and research diagnostic criteria (RDC) diagnosis. Among the depressed patients, the puerperal depressives were more psychotic, disoriented, agitated, and emotionally labile than the nonpuerperal depressives. The authors noted, however, that these differences were accounted for by a small subgroup of puerperal patients with “organic” symptoms. In a similar study, Katona²⁴ found that more puerperal depressives than nonpuerperal depressives experienced delusions and mood lability. Few other differences between the samples were noted.

Each of the studies just reviewed reported a high proportion of puerperal psychotic or hospitalized nonpsychotic women who experienced affective illness (depression or mania) relative to the women experiencing puerperal schizophrenia.^21,22,23,24 However, a review by Brockington and colleagues²⁵ of 20 studies conducted between 1911 and 1978 revealed that schizophrenia was reported to be more prevalent in 10 of the studies. Changing criteria for the diagnosis of depressive psychosis and schizophrenia over time, particularly among American psychiatrists, makes interpretation of older studies difficult.²⁵ Nevertheless, the most recent large-scale epidemiologic study found that 73% of psychiatric admissions in the puerperium were for affective disorders.²⁶

Postpartum psychoses occur in close proximity to childbirth. Several well-designed studies have now shown that the bulk of postpartum psychotic episodes occur within the first 2 to 4 weeks after delivery.^{19,26,27,28,29} Although there has been considerable debate regarding when after delivery this high-risk period ends,^21,25 disorders that require treatment up to 90 days after delivery are often considered postpartum disturbances.²³ Nevertheless, only episodes beginning within 4 weeks postpartum qualify as a DSM-IV postpartum onset specifier.⁶

Blues

Studies of the prevalence of the blues have reported a wide range (26% to 85%).^8,13,30,31 Studies reporting relatively high prevalence rates have used criteria such as the presence of crying at some time during the first week after delivery.³¹ Lower prevalence rates have been obtained in studies using more stringent criteria for the blues, such as high scores on standard depression or mood scales.^8,13

Depression

A recent meta-analysis³² based on 59 prevalence studies (N = 12,810 subjects) around the world found an average prevalence rate of 13%. Studies of postpartum depression in North America have obtained prevalence rates ranging between 6.8%³³ and 16.5%³⁴ for adult women. Many of these studies have been conducted in Iowa at the University of Iowa.^35,36,37,38 In our two studies, we obtained prevalence rates of 12%³⁷ and 10.4%³⁵ of major and minor depression during the first 9 weeks postpartum. Gotlib and colleagues³³ followed up one of the largest samples to date (N = 295) and obtained a prevalence rate of 6.8% for major and minor depression at 5 weeks postpartum. Cutrona,³⁶ using the more stringent Diagnostic and Statistical Manual of Mental Disorders, 3rd edition (DSM-III) criteria, found an 8.2% rate of major depression in the first 8 weeks postpartum. Whiffen's³⁴ higher rate of depression was obtained from a sample recruited from a facility near Vancouver, British Columbia. Troutman and Cutrona³⁸ studied a sample of 128 pregnant adolescents (age range, 14 to 18 years) and found a 26% prevalence rate of major and minor depression in the first 6 weeks postpartum.

Similar rates of depression were obtained in Great Britain when the diagnostic methods used were similar to those of North American studies. For example, Kumar and Robson³⁹ obtained a prevalence of 14.9% for major and minor depression at 12 weeks postpartum, similar to the 12% rate reported by Watson and colleagues⁴⁰ and the 10% reported by Wolkind and associates⁴¹ at 16 weeks postpartum. These studies were all conducted in London. However, Cox and co-workers⁴² reported a prevalence of 13% in the first 16 weeks postpartum in a study conducted in Edinburgh. Nott⁴³ obtained a prevalence rate of 18.5% at 12 weeks postpartum based on a sample from Southampton, and Cooper and colleagues⁴⁴ reported a prevalence of 8.7% in a sample from Cambridge. The criteria used in these studies were somewhat more variable than those used in American studies. British investigators often classify women as psychiatric “cases” based on their level of impairment and secondarily identify a specific syndrome. As a result, it is likely that some of their cases would not meet RDC or DSM-III criteria for depression because of, for example, a predominance of anxiety symptoms.

Most studies have found a somewhat higher risk for postpartum depression relative to comparable periods of time depressed during pregnancy. For example, Kumar and Robson³⁹ found an increased risk ranging from 1.3 (relative to first trimester) to 5.3 (relative to second and third trimesters). Similar findings were obtained in other studies as well.^36,40 Although Cox and associates⁴² reported a tripling of risk, the period under study during pregnancy was 1 week (around the 20th week of gestation) and the period under study after delivery was approximately 4 months. Moreover, O'Hara and co-workers³⁷ found a risk ratio of only 1.3 even though the period under study was longer after delivery (9 weeks) than during the second trimester (1 month). Finally, Gotlib and colleagues³³ found a greater risk for depression during the second trimester than at 5 weeks postpartum.

Recent studies have addressed the question as to whether the postpartum period is a time of greater risk for depression compared to non-childbearing periods. Watson and colleagues⁴⁰ noted that two studies of nonpuerperal women in nearby communities found, in one case, an equal prevalence of depression, and in the other case, a lower prevalence of depression compared with what they obtained in their study of postpartum depression. In another British study, Cooper and associates⁴⁴ compared rates of depression in a sample of childbearing subjects from Cambridge and a sample of sociodemographically similar subjects from Edinburgh and found no differences. A recent study conducted in Stoke-on-Trent⁴⁵ found no differences in the prevalence of depression over a 6-month period between postpartum and control women, but postpartum women had a higher rate of depression onset in the first 5 weeks after delivery compared to an equivalent period for control women. Finally, two studies at the University of Iowa comparing matched samples of adult³⁵ and adolescent³⁸ childbearing and non-childbearing subjects in a prospective design also found no differences in rates of depression, or in timing of onset.

Several of the more rigorous studies of postpartum depression have provided data regarding duration. For example, Kumar and Robson³⁹ reported that 50% of their depressed subjects had episodes lasting 6 months or more. Similarly, Watson and associates⁴⁰ reported that one quarter of their depressed subjects had episodes lasting 3 months or more and another quarter had episodes lasting 6 months or more. Cox and colleagues⁴⁶ reported that depressed women frequently had episodes lasting between 3 and 6 months, with many women reporting in retrospect that their episode had lasted 1 year or more. Wolkind and co-workers⁴¹ reported that 70% of their subjects who experienced a psychiatric disorder at 4 months postpartum were still experiencing a psychiatric disorder at 14 months postpartum. Pitt¹⁶ found that 40% of the depressed subjects whom he observed at 1 year postpartum had made little recovery. Finally, O'Hara and colleagues³⁷ reported an average length of a postpartum episode of 3.1 weeks; however, they monitored these women only until 9 weeks postpartum.

Psychosis

Recent studies of the incidence of postpartum psychosis within 90 days of delivery have reported rates ranging from 1.1 in 1000 deliveries³⁰ to 4 in 1000 deliveries.^26,27,28,47 These same studies have also demonstrated that pregnancy is a time of lowered risk for psychosis, whereas the first month postpartum, in particular, is a time of elevated risk (perhaps a 13-fold increase).²⁶

Many studies of causal factors in postpartum mental disorders report correlations between presumed causal factors and measures of depression symptom severity. In these studies, particularly those in which hormonal variables are examined, little distinction is made between symptoms of depression and symptoms of the blues. For this reason, studies of causal factors in the blues and depression have been combined in one section.

Background Factors

A large number of investigations have studied the relationship between sociodemographic variables and postpartum blues and depression. The findings of this body of research have been inconsistent across studies. Typical would be studies that have investigated the association between parity and postpartum depression. In 18 studies, no relationship between parity and postpartum depression was found. In 12 studies, higher levels of parity were associated with increased risk for postpartum depression. Finally, in three studies, lower levels of parity were associated with increased risk for postpartum depression.⁷ The results of a recent meta-analysis³² do suggest that lower family income and occupational status increase the risk of postpartum depression. In that same meta-analysis, sociodemographic variables such as mother's age, marital status, length of relationship, partner's level of education, number of children, and employment during pregnancy were not associated with risk for postpartum depression.³²

Biological Factors

The predominant explanation for the postpartum blues and depression is that levels of hormones such as estrogens, progesterone, prolactin, and cortisol are either too high or too low in the puerperium, or that changes in the levels of these hormones occur too quickly or not quickly enough.⁹ The hormonal dysfunction hypothesis is similar to that proposed for mood disturbances in the premenstruum and during menopause.^31,48,49 Although most of the hormones investigated in this area are ones that have primary relevance to postpartum disorders (e.g., estrogens, progesterone), others have relevance to mood disorders in general (e.g., cortisol, norepinephrine, tryptophan). George and Sandler⁵⁰ wrote a comprehensive review of the literature on this subject, and Harris⁵¹ reviewed the roles of steroids and thyroid hormones in postpartum disorders.

STEROID HORMONES.

Progesterone.

Several case reports have claimed that progesterone administered after childbirth is effective in treating postpartum psychosis and preventing postpartum depression.^49,52 These reports were consistent with the premise that progesterone deficiency may be causally related to postpartum mood disturbance.

In a sample of 27 women, Nott and colleagues⁵³ found a significant association between magnitude of decline in progesterone after delivery (between approximately 38 weeks' gestation and day 1 postpartum) and depressed mood in the first week postpartum. Several later studies of the postpartum blues failed to confirm this observation,^54,55,56 and one study found significantly higher levels of progesterone in five women experiencing the blues compared with five symptom-free women.⁵⁷ In our most recent study,⁵⁸ we found no differences in progesterone between depressed and nondepressed women after delivery, nor were antepartum or postpartum progesterone levels associated with the blues.⁸ Interestingly, based on assessments conducted at about 8 weeks postpartum, Harris and colleagues⁵⁹ reported lower levels of progesterone among depressed breastfeeding women than nondepressed breastfeeding women. Among the bottle-feeding women, just the opposite association was observed. In a more recent study, Harris and colleagues⁶⁰ found small but significant associations between the postpartum blues and lower levels of progesterone in the first 10 days postpartum. However, progesterone levels were not related to postpartum depression assessed at 5 weeks postpartum.⁶¹ Unlike earlier investigators, Harris and colleagues^60,61 measured the free (and biologically active) form of progesterone from saliva, rather than total progesterone, much of which is bound to binding globulin (and biologically inactive).

Estrogens.

Estrogen levels, much like progesterone levels, decrease markedly after childbirth. Several studies have investigated the relationship between estrogen levels (usually estradiol) and postpartum blues and depression. Most of these studies have produced negative findings.^55,56,59,62 Nott and colleagues⁵³ found lower levels of total estrogen before delivery among women who experienced the blues in the first week postpartum, and Feksi and associates⁵⁷ found higher levels of estradiol among women who experienced the blues in the first week postpartum. Finally, our group found significantly lower levels of estradiol at 36 weeks' gestation and at day 2 postpartum among women with postpartum depression.⁵⁸ We also found that higher levels of free estriol at 38 weeks' gestation and a greater decrease from mean antepartum levels to day 1 postpartum levels of free estriol were predictive of the postpartum blues in a sample of approximately 120 women.⁸

Cortisol.

Abnormalities in the hypothalamic-pituitary-adrenal system have been studied intensively as etiologic factors in major depression.⁶³ Both free and bound cortisol are elevated during late pregnancy, peak at high levels during labor, drop precipitously after delivery to about late pregnancy levels, and gradually decline over a period of time.⁶⁴ Fluctuation in corticosteroid levels (i.e., overproduction and sudden withdrawal) has been posited as a potential cause of postpartum mood disorders.^50,65 Nevertheless, the literature shows very little support for the relationship between cortisol levels and the postpartum blues and depression.^55,57,58,59 Handley and colleagues¹³ did find significantly higher levels of cortisol at 38 weeks' gestation in women who later experienced the postpartum blues than in women who did not; however, they were unable to replicate this finding in a later study.⁶² Ballinger and associates⁵⁴ reported significantly higher urinary levels of 11-hydroxycortisol steroids on day 1 after delivery in a group of women showing the greatest amount of increased positive mood between days 2 and 4 postpartum. Ehlert and colleagues⁶⁶ found that salivary cortisol was elevated on days when symptoms occurred in women who experienced the postpartum blues compared to those who did not. Harris and associates⁶⁰ found no evidence of a relation between cortisol levels and the occurrence of the blues, but they did find that lower levels of prenatal cortisol were associated with elevated depression scores at about 6 weeks postpartum.⁶¹ In our own work, we found no differences with respect to levels of urinary free cortisol or total plasma cortisol between women with and without postpartum depression.⁵⁸ Moreover, dexamethasone suppression does not distinguish women with and without postpartum depression in the early puerperium.⁵⁸ Nonsuppression rate during the first week postpartum is approximately 90%.⁵⁸

PEPTIDE HORMONES.

Prolactin.

Prolactin levels, which are very high by the end of pregnancy, decline more slowly after delivery than the gonadal hormones. Abnormally high levels of prolactin (hyperprolactinemia) in non-childbearing women are associated with depression, anxiety, and hostility.⁶⁷ With respect to postpartum mood disorders, George and colleagues⁶⁸ found that daily basal levels of plasma prolactin were positively correlated with daily measures of dysphoric mood on days 2, 4, and 6 postpartum. However, Harris and associates⁵⁹ found that lower levels of prolactin were associated with higher levels of depression in breastfeeding women at 8 weeks postpartum. Other studies have found no association between postpartum mood and prolactin levels.^53,58 It should be noted that prolactin is a difficult hormone to measure well in the puerperium. Breastfeeding stimulates prolactin secretion, and prolactin levels are significantly higher in breastfeeding women compared to non-breastfeeding women.⁶⁹ There has also been a suggestion in the literature that even slight ambulation stimulates prolactin secretion⁵⁰; however, in our own work we have found little evidence of this phenomenon.⁵⁸

b-Endorphin.

In the early 1980s, Halbreich and Endicott⁷⁰ suggested that endorphin withdrawal or imbalance might be an important etiologic factor underlying both premenstrual syndrome and postpartum depression. The small body of work that has been done on this subject has been largely disappointing. George and Wilson⁷¹ reported significant positive correlations between β-endorphin levels and measures of depression on days 2, 4, and 6 postpartum in a sample of 38 women. In a small sample, Newnham and colleagues⁷² found that lower levels of β-endorphins at 36 weeks' gestation were associated with higher degrees of blues symptoms in the first 3 days postpartum. Brinsmead and co-workers⁷³ found no association between β-endorphin levels and postpartum blues.

NEUROTRANSMISSION.

Abnormalities in neurotransmission are central to many theories of psychiatric disturbance. Most of the hormones that have been studied with respect to postpartum mood disorders affect various neurotransmitter systems in one way or another and are, in turn, affected by them.

Norepinephrine.

Early biological theories implicated lowered levels of norepinephrine in the etiology of affective disorders. Only two studies have investigated norepinephrine with respect to postpartum mood disorders. Treadway and colleagues⁷⁴ found a significant association between reduced norepinephrine levels and increased depressed mood in women at about 34 weeks' gestation but found no significant associations between norepinephrine or epinephrine levels and postpartum mood disturbance. Kuevi and associates⁵⁵ found that norepinephrine and epinephrine levels were significantly lower on the day of peak mood disturbance after delivery compared with the preceding and following days; however, there were no differences in norepinephrine and epinephrine between women who experienced the blues and those who did not.

a2-Adrenoceptors.

Only one study has investigated α₂-adrenoceptors in postpartum mood disorders. Metz and colleagues⁵⁶ found significantly greater numbers of α₂-adrenoceptors in women experiencing postpartum blues than in those who were not. These findings are consistent with studies of non-childbearing endogenous depressives.⁵⁰

Plasma Tryptophan.

Tryptophan is a precursor of serotonin (5-HT) and has been studied in postpartum mood disorders on the assumption that an inadequate supply of the 5-HT precursor tryptophan might result in inadequate release of 5-HT. Stein and co-workers⁷⁵ reported a significant negative correlation between free tryptophan levels and postpartum depression at 6 days postpartum. A similar finding was also reported by Handley and colleagues,⁷⁶ although there was some question about the appropriateness of the data analysis. In a much larger study, Handley and associates¹³ found that total tryptophan recovered to its normal, higher level more quickly after delivery in women who were not experiencing the blues than in those who were. There were no differences in free tryptophan between the groups. Harris⁷⁷ conducted a double-blind, placebo-controlled study of the effects of 3 g of L-tryptophan per day for 10 days in a sample of 55 women. Supplementary tryptophan did not reduce the incidence of the blues.

Monoamine Oxidase.

There has been one study on the role of monoamine oxidase activity in postpartum blues. George and Wilson⁷⁸ found significant positive correlations between monoamine oxidase activity and depression on days 2, 4, and 6 postpartum.

Cyclic Adenosine Monophosphate.

Ballinger and colleagues⁷⁹ found that nine women who were rated as having the most mood disturbance on day 3 postpartum had significantly higher levels of cyclic adenosine monophosphate (cAMP) than the rest of the sample (92 women), a finding opposite to what might have been predicted⁸⁰ and at odds with another finding that the 13 women who showed the greatest shift toward elated mood between days 2 and 4 postpartum had significantly higher levels of cAMP compared with the rest of the sample. In a later study, Ballinger and associates⁵⁴ again found a significant association between positive mood change after delivery and higher levels of cAMP.

ELECTROLYTES.

A few investigations of electrolyte levels and excretion have been conducted. Stein and colleagues⁸¹ found no association between postpartum blues and body weight change, sodium excretion, and calcium excretion. They did note, however, a coincidence in timing between the onset of blues and the start of rapid weight change and rise in sodium excretion. Riley and Watt⁸² reported on a study of 53 women hospitalized for severe puerperal psychiatric illness. They found significantly higher levels of serum calcium in 15 women who had postpartum psychiatric illness and no personal or family history of psychiatric illness compared with 38 women who had postpartum psychiatric illness and a personal or family history of psychiatric illness, 49 postpartum women who were not ill, and 35 non-postpartum women who were psychiatric patients. They suggested that hypercalcemia may be one cause of postpartum psychiatric illness in women not otherwise vulnerable to postpartum psychiatric illness.

VITAMINS.

Livingston and colleagues⁸³ found no evidence for vitamin B₆ deficiency in women with postpartum depression. However, Pulkkinen and associates⁸⁴ found a significant negative correlation between vitamin B₆ levels and level of depression among 15 depressed women during the second trimester of pregnancy. Also, Riley⁸⁵ administered either 100 mg of pyrodoxine hydrochloride or placebo (double-blind) to 102 women beginning on day 1 postpartum and extending through day 28 postpartum. She found that although the women receiving pyrodoxine showed significant improvement in their mood during the 28 days, the women receiving the placebo did not.

THYROID SYSTEM.

Hamilton⁸⁶ argued that hypothyroidism may be a cause of postpartum depression, particularly when it begins more than 2 weeks after delivery. However, neither Grimmell and Larsen⁸⁷ nor George and Wilson⁷¹ found evidence of any association between thyroid function or thyroid levels and postpartum mood disturbance. More recently, Harris and colleagues⁸⁸ reported an association between depression meeting DSM-III criteria and thyroid dysfunction (either hyperthyroidism or hypothyroidism) measured at 6 weeks postpartum. A similar finding was reported by Pop and colleagues.⁸⁹ In an even more recent study, Harris and co-workers⁹⁰ found that positive thyroid antibody status was predictive of postpartum depression, a finding not replicated by Pop and associates.⁹¹ In summary, thyroid dysfunction may account for a small percentage of postpartum depression cases. Pop and colleagues⁸⁹ estimated the prevalence rate of postpartum major depression associated with thyroid dysfunction to be 1%. With respect to long-term consequences of thyroid dysfunction, Pop and co-workers⁹² did find that positive thyroid antibody status during pregnancy was related to lower levels of cognitive ability in the offspring at 5 years of age.

Although there has been little consistency in the findings of studies of hormonal variables, the recent work of Harris and colleagues⁵¹ has highlighted the potential contributions of progesterone withdrawal and lower levels of prenatal cortisol to the development of the postpartum blues and depression, respectively. These investigators argue that it is critical to measure the biologically active fraction of the steroid hormones because so much of the hormone is inactive and bound to binding globulin. Even given some of these newer findings, the simple hormonal models are unable to account for much variability in postpartum mood. There may be many reasons for this state of affairs. For example, the wrong hormones may have been studied, or the complex interrelationships among hormones, neurotransmitters, and other biological factors may not have been adequately captured. Finally, hormonal factors may only be important for women who are otherwise vulnerable to affective disorder. This vulnerability could be caused by biological/genetic (e.g., reflected in personal or family history of affective disorders), psychological (e.g., poor cognitive/social coping with stress), or social/environmental factors (e.g., poor marital relationship, adverse social circumstances). Obviously, all three of these possibilities may be true; the potential roles of these other factors are reviewed in the next several sections.

Obstetric Factors

Obstetric and delivery complications have been inconsistently related to postpartum depression. Some studies have found less mood disturbance in women with more stress or delivery complications,^93,94 others have found the expected relationship between obstetric complications and postpartum mood disturbance,^37,95,96 and still others have found no association.^42,97,98,99 Regarding cesarean section specifically, a recent meta-analysis reported a small but nonsignificant relationship between cesarean section and postpartum depression.¹⁰⁰

Stressful Life Events

Negative life events (e.g., loss of family member, unemployment) that have been found to increase the likelihood of depression in nonpuerperal women may also occur during pregnancy and after delivery.¹⁰¹ Several studies have found that higher levels of stressful life events during pregnancy and after delivery were associated with higher levels of postpartum depression symptoms^{36,37,58,94,102,103} and an increased probability of a clinical postpartum depression.^93,104,105 Three studies failed to find any association between stressful events and postpartum depression.^16,39,106

Relationship with Partner

Probably no adult relationship is more important to a woman during the puerperium than that with her husband. Many studies have found that women with postpartum depression report poor marital relationships after delivery,^{42,93,105,107,108} although one study found no differences between depressed and nondepressed women.¹⁰⁶ More importantly, several studies assessed the marital relationship during pregnancy and found that the poor marital relationship preceded the postpartum depression.^{39,40,104,109} However, at least two studies have found that marital distress during pregnancy was not predictive of postpartum depression.^105,110

Relationship with Parents

A few investigators have studied the effect of women's conflict with parents and loss of parents in childhood on the likelihood of postpartum depression. Several studies found that a poor relationship with their mother was associated with some women's postpartum depression.^{39,108,109,111} However, Paykel and colleagues⁹³ found no association between postpartum depression and parental conflict or childhood parental loss. Similarly, Watson and co-workers⁴⁰ found no association between childhood separation from parents and postpartum depression.

Social Support

Social support from spouse, family, and friends during times of stress (e.g., helping in household tasks, acting as a confidant) is thought to reduce the likelihood of depression.¹¹² Lack of support from the spouse has been found to be associated with increased levels of postpartum depression,^{104,105,111,113} although Hopkins and colleagues¹⁰⁶ failed to replicate this finding. Other studies have found that lack of an adequate confidant or inadequate levels of support from a confidant is associated with postpartum depression.^93,102,105 Cutrona¹¹⁴ reported that lower levels of perceived social support during pregnancy was associated with postpartum depression symptoms. Finally, in a study of largely low-income women, Collins and associates¹¹⁵ found that women who reported lower quality support and who had fewer network resources during pregnancy had higher levels of postpartum depressive symptomatology.

Personality

Several studies have found that women with postpartum depression have higher levels of “neurotic” personality characteristics than nondepressed mothers.^16,40,116 But two studies did not.^39,103 A number of investigators have also reported that a higher level of anxiety during pregnancy is predictive of postpartum depression,^40,115 although Pitt¹⁶ did not find such an association. Finally, several studies have found significant correlations between levels of depression/emotional distress during pregnancy and those levels after delivery.^36,94 Moreover, levels of depressive symptoms during pregnancy have also been found to predict the diagnosis of postpartum depression in at least one study,⁴⁰ but not in one other.³⁷

Two constructs derived from cognitive-behavioral theories of depression have been tested as predictors of postpartum depression.^118,119 Several investigators^36,94,103 have found that attributional style (the types of causes women identify for good and bad events) measured during pregnancy predicted level of postpartum depression¹¹⁸; however, two other studies failed to replicate these results.^37,120 O'Hara and colleagues⁹⁴ found that subjects' attitudes about self-control, measured during pregnancy, were significantly correlated with postpartum depression level but were not significant in a regression equation predicting postpartum depression level.¹¹⁹ In a later study, self-control attitudes were found to be significant in a regression equation predicting postpartum depression level but not postpartum depression diagnosis.³⁷

Personal and Family Psychopathology

Women who have experienced a previous psychiatric disorder unrelated to pregnancy would appear to be at risk for postpartum depression. Several studies have obtained data regarding previous psychiatric disturbance and found an association with postpartum depression.^{40,58,93,104,105} Other studies, however, have not found the expected association between psychiatric history and postpartum depression.^{16,39,108,110}

Mood disturbance during pregnancy does seem to be related to the subsequent development of the blues.^8,53,77 This finding is consistent with reports indicating that measures of depression taken during pregnancy are significantly correlated with measures of depression taken during the puerperium.^36,37,94 It may reflect vulnerability in some women to mood disturbance or an ongoing mood disturbance that is briefly suppressed immediately after delivery.³⁰

Family history of psychopathology has also been studied as a predictor of postpartum depression in a few studies. Studies are split between those that found a significant association between family history of psychopathology and postpartum depression^37,40 and those that did not.^39,58,108

None of the potential social and psychological causal factors in postpartum depression have been supported unambiguously in the literature. The pattern of results reviewed here would suggest that gynecologic and obstetric variables are not specifically related to postpartum depression. There is good evidence that a woman's psychological adjustment before and during pregnancy are associated with the development of postpartum depression. Moreover, women who experience high levels of stress during pregnancy and after delivery and those who lack a supportive spouse would appear to be particularly prone to postpartum depression. These are psychosocial variables that figure very prominently in models of non-postpartum depression.¹²¹ In fact, many investigators have assumed that the postpartum period is simply a high-risk time for depression and that it is an appropriate context in which to test etiologic models developed to account for non-postpartum depression.^36,94,122

Because of the relative rarity of postpartum psychosis, few prospective studies have been carried out.^19,123 These studies have followed pregnant women who had a previous history of nonorganic psychosis. Most other studies have been retrospective chart reviews in which cases of postpartum psychosis that have been assembled over several years are compared with samples of women not experiencing postpartum psychosis. Because of this, variables that have been studied are those commonly available in psychiatric or obstetric charts.

Although parity does not seem to be related to postpartum depression, there is a general consensus that primiparous women are most vulnerable to postpartum psychosis.^25,26,29 The risk for a primiparous woman is about twice as high as for a multiparous woman, and this figure is too high to be accounted for by avoidance of future pregnancies by primiparous women who become psychotic.¹²⁴ The only other characteristic of the mother that has been found to be associated with postpartum psychosis is unmarried status.^27,124 However, at least one major study found no evidence of increased risk for unmarried women.¹²⁵

A few studies have found that women who have been diagnosed with manic-depressive disorder are at high risk for postpartum psychosis. For example, Reich and Winokur¹²⁶ reported that 50% of all deliveries following diagnosis of manic-depressive disorder (with clear evidence of mania) resulted in a postpartum mental illness (both depression and mania). Bratfos and Haug¹²⁷ reported that in six manic-depressive women (most were depressed) whose first severe episode was nonpuerperal, 11 of 12 deliveries were followed by puerperal depression. In McNeil's¹⁹ prospective study, there were 25 cases of psychosis following 88 deliveries (in 83 women) in women with a history of nonorganic psychosis. Similarly, in the prospective study of Wieck and colleagues,¹²³ there were six cases of psychosis following deliveries in 15 women with a history of bipolar or schizoaffective psychosis.

Family history of psychopathology has been found to be related to puerperal psychosis.^25,124 Whalley and colleagues¹²⁸ reported on the relative risk for postpartum psychosis in the first-degree relatives of puerperal psychotics (22.2 in 1000 births) and nonpuerperal psychotics (33.3 in 1000 births), both much higher than the expected rate from the general population (1 in 1000 births). Several other studies, although not using nonpsychotic control groups, have found high rates of mental illness in the family members of puerperal psychotics.¹²⁹ It is evident from the study by Whalley and colleagues,¹²⁸ however, that postpartum psychosis is no more prevalent in family members of puerperal psychotics than in family members of nonpuerperal psychotics.

Many other factors associated with childbearing have been studied as potentially associated with postpartum psychosis. For example, Paffenbarger¹²⁵ found that stillbirth and perinatal death, which are associated with advanced maternal age, short gestation, and difficult labor, were related to increased risk for postpartum psychosis. Other studies have not replicated these results.⁴⁷

McNeil¹⁹ monitored 88 women with a previous history of nonaffective psychosis and classified them as having an early onset if their illness (predominantly affective disorder) began within 3 weeks after delivery, and as having a later onset if their illness (predominantly schizophrenia) began later than 3 weeks after delivery. Lower parity and younger age were related to risk of psychosis for the group with early onset but not for the group with later onset.¹³⁰ Both early and later onset were related to total length of previous hospitalizations and severity of disturbance in the 6 months before the index pregnancy.¹³⁰ There were few other social or obstetric factors that were related to risk for postpartum psychosis in this high-risk group.¹³¹ Moreover, several more recent studies have found little evidence of a causal role for stressful life events.^132,133 Interestingly, Marks and colleagues¹³² did find that a relatively poor marital relationship and a relatively low level of interaction with the partner were associated with relapse in women with a previous history of psychosis.

Dopamine supersensitivity is one biological variable that has been studied prospectively in women at high risk for postpartum psychosis by virtue of having had a previous psychotic episode.¹²³ Wieck and colleagues^123,134 have reported that women who relapse with an affective psychosis have an enhanced growth hormone response to the dopamine agonist apomorphine on day 4 after delivery and before the relapse. Wieck and associates¹²³ suggested that there may be a link between estrogen “withdrawal” and this increased sensitivity. However, in a very small study, Meakin and colleagues¹³⁵ failed to confirm these findings.

Blues

The major negative consequence of the blues is that they may herald a later, more serious postpartum depression.^8,43 Even this consequence was not replicated in a recent study.⁶² There is no evidence that any real impairment results from the blues.³⁰ There have been no studies of the consequences of the blues on the child or family; however, the transient nature of the blues would suggest that even short-term negative consequences would be few.

Depression

Several long-term follow-up studies of women participating in postpartum depression projects have been reported.^{39,136,137,138} Wolkind and colleagues^41,137 reported on a 42-month follow-up of 108 of 131 mothers who had participated in a longitudinal study of pregnant and puerperal women and their children.¹³⁹ They also reported on follow-ups conducted at 14 and 27 months postpartum. They found that very high percentages of women (range, 44% to 78%) who experienced postpartum depression experienced later depressions at 14, 27, and 42 months after delivery. They also reported that maternal depression at 4 months postpartum was not related to mothers' reports of behavior problems in their children at 14 or 27 months, but was related to behavior problems at 42 months. The authors argued that postpartum depression per se is not directly associated with later child behavior problems; however, it may be indirectly related to child behavior problems through its association with later maternal depression.

Kumar and Robson³⁹ followed up on 99 of 119 women 4 years after delivery. They found that 6 of 14 women who had experienced a new episode of depression at 3 months postpartum continued to seek psychiatric assistance in the ensuing 4 years. They also reported that one woman who had become depressed shortly after 3 months postpartum and one other woman who could not be located for follow-up also were in need of continuing psychiatric care beyond the postpartum period. Using the McCarthy scales of children's (cognitive) abilities, Cogill and associates¹³⁶ demonstrated that children of women who experienced a postpartum depression performed significantly worse than children of those who did not; however, in a later study they did not find differences between the two groups of children with respect to their social and emotional development.¹⁴⁰

Wrate and colleagues¹⁴¹ reported on a 3-year follow-up of 91 of 103 women who had participated in a prospective study of postpartum depression.⁴² Surprisingly, the children of women who had experienced brief postpartum depressions were rated as experiencing significantly more behavior disturbance than children of mothers who experienced more severe postpartum depressive illness and mothers who were not depressed after delivery. The only explanation that the authors had for these findings was that the women experiencing briefer postpartum depressions compared with the women experiencing more significant postpartum depressions were more anxiously preoccupied with their baby and were also more reluctant to continue breastfeeding. Moreover, those who experienced briefer postpartum depressions were also more likely to experience later depression (10:16) than women who experienced prolonged postpartum depressions (2:11) and women who had not experienced a postpartum depression (15:63). Thus, it may have been the more chronic long-term course of depression experienced by the women with the milder postpartum depression that was responsible for increased problems in their children.

Philipps and O'Hara¹³⁸ conducted a 4¹/₂ -year follow-up study of 70 of 99 women and children who had participated in an earlier prospective study.³⁷ They found evidence that postpartum depression was associated with risk for further episodes of depression during the follow-up period; however, there was no association between postpartum depression and mothers' reports of behavior problems in their children at age 4¹/₂ . In a later study, O'Hara's group¹⁴² followed 124 of 182 women and children who were recruited in the context of an earlier prospective study³⁵ for 3¹/₂ years. Postpartum depression was directly associated with increased risk for later episodes of maternal depression and mothers' reports of behavior problems in their children.

Psychosis

Postpartum psychosis is a severe illness, and the short-term consequence is usually social incapacitation and the need for hospitalization. Since the advent of psychotropic medication and electroconvulsive therapy (ECT), the length of hospitalization for postpartum psychotic women has decreased to less than 3 months on the average.^25,129 Several follow-up studies have been conducted and have found that although women usually recover from the postpartum psychosis itself, puerperal and nonpuerperal recurrences are common whether or not the woman had experienced a psychotic episode before her first postpartum psychosis. For example, Protheroe¹²⁹ found that 49 of 104 women discharged from a psychiatric hospital after experiencing a postpartum psychosis required further inpatient or outpatient treatment for other episodes of psychosis. Protheroe also reported that about one in six subsequent pregnancies was followed by puerperal psychosis. Davidson and Robertson¹⁴³ reported that 51% of their sample, whose postpartum episode was their first, experienced another episode not associated with childbearing, and 28% of subsequent pregnancies in this group were followed by a postpartum episode. Similarly, Rohde and Marneros¹⁴⁴ found that of 61 women who experienced a postpartum psychosis over a 12- to 41-year period, 64% of women had a recurrence; among the 31 of these women who had further pregnancies (range 1–6) resulting in 46 deliveries, the recurrence rate was 17%. Brockington and colleagues,²⁵ in a review of several studies, estimated that 21% of pregnancies subsequent to a postpartum psychosis are followed by another postpartum psychosis.

Having a psychotic mother would seem to pose obvious problems to an infant. However, special psychiatric units, found frequently in the United Kingdom, hospitalize the mother and baby together and give the mother as much responsibility for the baby's care as possible.¹⁴⁵ In their review, Margison and Brockington¹⁴⁵ suggested that children who are hospitalized with their psychotic mothers are not at great risk for physical injury. Moreover, Margison¹⁴⁶ reported that psychotic mothers were less likely than nonpsychotic depressed mothers to report feelings of rejection or hostility toward their baby.

Only a few studies have examined the long-term effects of postpartum psychosis on the child. For example, Cohler and colleagues¹⁴⁷ found that the children (ages 5 to 6) of women with affective psychosis after delivery were more impaired on the Weschler Preschool and Primary Scale of Intelligence than children of women who were diagnosed as schizophrenic after delivery and children of well mothers. Children of depressed mothers also made more errors on an embedded figures task than children of schizophrenic and well mothers. In contrast, in a 6-year follow-up by McNeil and colleagues¹⁴⁸ of 17 children whose mothers had experienced a postpartum psychosis, there were no negative developmental consequences.

Many other studies have been conducted examining severe mental illness in the mother (schizophrenia and affective psychosis) and its relation to problems in the child. These studies, although not including women with postpartum psychosis per se, have found that children of psychotic mothers are at risk for attentional problems,¹⁴⁹ impaired communication skills,¹⁵⁰ deviant behavior in the school setting,¹⁵¹ and childhood depression.¹⁵² Whether these problems would accrue to children of mothers who experienced a short-term psychotic episode after delivery with no further recurrences is unclear; however, as noted earlier, a large proportion of women who experience postpartum psychosis are vulnerable to additional episodes.

In this section, depression and psychosis prevention and treatment studies are reviewed. In the next section, issues in the clinical management of postpartum mood disorders are discussed.

Psychotherapy

TREATMENT.

In recent years, there have been several controlled trials evaluating psychological treatments for postpartum depression. Holden and colleagues¹⁵³ evaluated the effects of a brief version of client-centered therapy for postpartum depression using health visitors (in Scotland) over a period of 8 weeks. Depression remitted in a significantly greater proportion of the women exposed to the health visitor intervention (69%) than in those who received no treatment (38%). These findings were replicated in two additional studies conducted in Sweden¹⁵⁴ and England.¹⁵⁵ Both studies found evidence of a significant effect of counseling relative to no treatment for postpartum depression, and one study found evidence of a significant effect of cognitive-behavioral therapy relative to no treatment¹⁵⁵; however, the average clinical response to treatment was relatively small. A recent study in the UK evaluated one or six sessions of “cognitive-behavioral” counseling provided by health visitors combined with fluoxetine or placebo over 12 weeks.¹⁵⁶ The authors reported that fluoxetine was superior to placebo and that six sessions of counseling was more effective than one session of counseling. Combining fluoxetine and counseling had no additive effect. Finally, Stuart and O'Hara¹⁵⁷ reported on a series of six cases using interpersonal psychotherapy¹⁵⁸ for postpartum depression. Patients showed major improvements in depressive symptomatology with four of six patients evidencing complete recovery.¹⁵⁷ Interpersonal psychotherapy has also been shown to be effective in women with antepartum depression.¹⁵⁹

PREVENTION.

A number of studies have also been conducted to evaluate the efficacy of psychological interventions to prevent postpartum depression. In a seminal study, Gordon and Gordon¹⁶⁰ evaluated a two-session prenatal group intervention aimed at providing women (or couples) with practical advice to reduce postpartum emotional distress. The following is an example of the advice given in the group sessions: “The responsibilities of being a mother (and not a martyr) are learned, hence get help and advice.” The authors found that women who attended these groups, particularly those who attended with their spouse, experienced significantly less emotional distress after delivery than women who did not attend these group sessions.

Halonen and Passman¹⁶¹ reported on a study of prenatal preparation in which they assigned 48 pregnant women, who had received labor-specific relaxation training, to one of four groups: (1) additional relaxation training, (2) extended relaxation training that emphasized possible postpartum stressors, (3) discussion of postpartum stressors, and (4) a control discussion about their awareness of postpartum stress. They found that the groups receiving relaxation training were significantly less distressed than the nonrelaxation training groups during the first 9 weeks postpartum. They also found that the groups that discussed possible postpartum stressors were less elated after delivery than the groups not exposed. The authors recommended the use of extended, nonspecific relaxation training both before and after delivery as a way of reducing postpartum emotional distress.

Elliott and co-workers¹⁶² evaluated the efficacy of prenatal groups for women at risk for postpartum depression on the basis of having been treated for a prior depression, having a poor marital relationship, lacking a confidant, or having a high level of anxiety during pregnancy. These women, who were considered to be at high risk, were expected to have about a threefold increased risk for an episode of postpartum depression. The group sessions were psychoeducational, and to a large degree the participants were free to influence the course of the meetings, five of which occurred during pregnancy and six of which occurred after delivery on a monthly basis. The findings were that women who participated in these groups were significantly less likely than the women who did not receive the treatment to experience a depression at 3 months postpartum.

In a more recent study, Wolman and colleagues¹⁶³ evaluated the efficacy of providing companionship during labor to a group of women who had no companions of their own. The authors argued that labor was a time when women were especially vulnerable to losing confidence in their competence as a mother, and that feelings of incompetence as a mother was a causal factor in the development of postpartum depression. The provision of support during labor was hypothesized to increase women's confidence in their competence and to reduce depressive and anxious symptomatology during the postpartum period. This rather brief intervention during labor resulted in significantly greater levels of confidence and significantly lower levels of postpartum depressive and anxious symptomatology in the treated relative to the untreated group.

Pharmacotherapy

TREATMENT.

Gregoire and associates¹⁶⁴ evaluated transdermal estrogen as a treatment for postpartum depression. Women whose depression began within 3 months of childbirth and persisted for up to 18 months postpartum were assigned to either an estrogen regimen or a placebo control. Women received 3 months of transdermal 17β-estradiol 200 mg daily and then 3 months with added dydrogesterone 10 mg daily for 12 days each month. The placebo group used patches and took tablets on the same schedule. The investigators found that women receiving estrogen improved to a significantly greater extent than the women receiving the placebo. It should be noted, however, that women on antidepressant medication were not excluded from this study.

Appleby and colleagues¹⁵⁶ compared fluoxetine plus counseling to placebo plus counseling in a 12-week trial in a group of postpartum depressed women recruited from the community. Fluoxetine proved superior to placebo; however, counseling did not add to the effects of fluoxetine. Stowe and associates¹⁶⁵ evaluated sertraline in a open trial; 21 women experienced a postpartum major depression; 14 of the 21 women showed complete recovery by the end of the 8-week treatment trial.

Only a few studies have systematically evaluated interventions for postpartum psychosis. In a controlled study, Baker and colleagues¹⁶⁶ found ECT to be more effective than chlorpromazine. Steiner and associates¹⁶⁷ compared propranolol and chlorpromazine in 10 patients with “psychosis associated with childbearing” and found some advantage for the propranolol.

PREVENTION.

Dalton,⁴⁹ in an uncontrolled study, treated 27 women who had had at least one prior episode of postpartum depression with 100 mg of progesterone once daily for 7 days, administered intramuscularly, followed by 400 mg of progesterone suppositories twice daily for 60 days. She reported no cases of postpartum depression within 6 months of delivery; however, she did not include a control group of untreated women. Wisner and Wheeler¹⁶⁸ recruited a sample of 23 pregnant women who had previously experienced a major episode of postpartum depression and gave them a choice of beginning an antidepressant immediately after delivery or being monitored by the clinical staff. Of the 15 women who chose to receive antidepressant medication (mostly nortriptyline and fluoxetine), 1 experienced a recurrence (compared to 5 of the remaining 8 women who chose postpartum monitoring).

Lithium prophylaxis for women with a history of bipolar affective disorder or postpartum affective psychosis appears to be an effective method of preventing postpartum psychosis.^169,170 For example, Cohen and associates¹⁷⁰ reported that only 1 of 14 high-risk patients beginning a course of lithium and/or carbamazepine soon after delivery suffered a postpartum relapse, whereas 8 of 13 patients who did not receive antimanic drugs experienced a manic or depressive relapse within the first 3 months postpartum.

In most cases, the obstetrician (or a staff member) will be responsible for assessing a woman's risk for postpartum mood disorders, preparing her for problems that may occur, assessing for the presence of mood disturbance after delivery, and if necessary making an appropriate referral for treatment. Unless the childbearing woman has had a previous psychiatric disorder, it is unlikely that she would come to the attention of a clinical psychologist or psychiatrist until a mood disorder manifested after delivery.

Prenatal Screening

A psychiatric history should be routinely obtained on a new obstetric patient. Patients should be asked about any previous treatment for psychological or emotional problems. Because problems such as major depression and anxiety disorders are often left untreated, obstetricians should inquire about times when the patient might have been out of her normal mood state for more than 2 weeks at a time. An affirmative answer would require follow-up questions regarding depressive symptoms consistent with DSM-IV criteria for the various mood disorders. Multiparous women should be asked specifically about postpartum adjustment after previous deliveries.

Prenatal Preparation

It is important to emphasize to pregnant women that although mild mood disturbances (i.e., the blues) are common after delivery, they are usually not very severe and are short-lived. More serious disorders such as postpartum depression and particularly postpartum psychosis are much less common. There are several preventive measures, such as those evaluated by Gordon and Gordon,¹⁶⁰ that an obstetrician can recommend to women and their partners to forestall emotional distress after delivery.^171,172 Moreover, patients can be told that the experience of the blues is rather normal and may reflect the effects of exhaustion, relief of tension built up before and immediately after delivery, or the initial process of coping with the demands of a new baby. Finally, a woman should know who to contact if she is having emotional problems after delivery.^171,172

Patients who have had previous psychiatric disturbances should be monitored more closely. For any woman with a history of psychosis, it is advisable to make arrangements during pregnancy for her to consult with a psychiatrist (preferably one whom she knows). Given the high risk of relapse after delivery for these women, intensive monitoring of the patient's emotional functioning for at least 30 days and then less intensive monitoring for an additional 60 days is appropriate. Ordinarily, prophylactic drug treatment for high-risk patients is not begun until after delivery to minimize risk to the fetus.^168,170

Postnatal Assessment

Obstetricians should inquire about a patient's mood before sending her home from the hospital. If a patient is doing well, she can be advised to call if problems develop. If a patient reports some mood disturbance, reassurance can be provided regarding the likelihood of its being short-lived. The obstetrician might also inquire about possible psychosocial sources of mood disturbance. For example, a woman may be returning to an unhappy or abusive marital relationship or she may have severe financial problems that will affect her ability to mother. Appropriate referrals to hospital social workers, clinical psychologists, or psychiatrists are indicated in these cases. It is wise to follow up any referral, because the woman may be overwhelmed with responsibilities when she arrives home and may not seek appropriate help.

With respect to postpartum depression, at the 6-week postpartum examination, it is useful to ask women to complete a questionnaire on which they rate the severity of a number of depression symptoms (e.g., dysphoric mood, guilt, problems in concentration, fatigue, appetite disturbance). Examples of appropriate measures include the Beck Depression Inventory¹⁷³ and the Edinburgh Postnatal Depression Scale.¹⁷⁴ Although these measures cannot yield a definitive diagnosis of depression, they can alert a physician to those women requiring more intensive assessment.

Postnatal Treatment

The treatment of postpartum (and pregnancy-related) mood disorders ordinarily may proceed along the same lines as the treatment of mood disorders occurring at other times. However, there are special problems in pharmacotherapy for the pregnant woman and the breastfeeding postpartum patient.^175,176 Innovative treatment programs involving joint hospitalization of mother and child and outpatient treatment of psychotic mothers have been employed in Great Britain.^177,178 Several psychological therapies have been validated for postpartum mood disorders^153,155,157; however, appropriately trained clinicians may not always be available for referral. In those cases, clinicians who are competent in brief psychotherapy for depression in women should be able to provide appropriate care.

Most antidepressants (e.g., amitriptyline, desipramine, doxepin, fluoxetine, sertraline) are appropriate for major depression after childbirth.^165,176 The American Academy of Pediatrics categorizes all psychotropics as drugs whose effect on nursing infants is unknown but may be of concern.¹⁷⁹ Although many clinicians recommend that women discontinue breastfeeding while taking antidepressant medications because of the unknown effects of tricyclics on the immature neonatal nervous system,¹⁸⁰ recent evidence suggests that many antidepressants are relatively safe for breastfeeding infants.¹⁷⁶ ECT is recommended when antidepressant medications are ineffective or if a patient is a danger to herself or others.

Lithium may be used prophylactically for a woman with a history of affective psychosis.¹⁷⁰ Ordinarily, it is started immediately after delivery at a dose of 300 mg three times a day.¹⁸⁰ Lithium therapy should be continued for at least 6 months. Lithium is contraindicated in patients with abnormal kidney function. It is also now considered contraindicated during breastfeeding¹⁷⁹ because of the high blood concentrations that accrue in infants.

Haloperidol (Haldol) may be used in the treatment of postpartum psychosis, starting at 5 mg/day and increasing to 30 to 40 mg/day.¹⁸⁰ Initial response to treatment may be expected after a few days to a few weeks. Once the patient has been symptom-free for a few weeks, the drug may be gradually tapered off. Antipsychotics are also in the category of drugs whose effect on nursing infants is unknown but may be of concern.¹⁷⁹

The postpartum period is a high-risk time for the least (the blues) and most severe (psychosis) mental disorders. Most postpartum psychotic disorders occur in the first month postpartum; nonpsychotic major depressions may occur anytime in the postpartum period. The major risk factor for psychotic and nonpsychotic postpartum depressions is a previous history of a psychiatric disorder. Psychosocial factors such as stressful life events and a poor marital relationship also increase the risk for nonpsychotic postpartum depression; only primiparity has been consistently associated with increased risk for postpartum psychosis. Although a specific biological cause for postpartum mental disorders may yet emerge, the findings to date have been inconsistent and often contradictory. Women who have experienced postpartum mental disorders are at increased risk for further psychiatric problems; the risk to the child of later social and cognitive deficits is less clear. Obstetricians have an important role to play in preparing women for potential postpartum disorders and in referring them for help should problems occur. Traditional treatments for psychiatric disorders are generally appropriate for women with postpartum mental disorders.