Is postpartum depression a clinically useful concept?

KEYWORDS:

• Antidepressants
• classification
• depression
• hypomania
• mania
• peripartum
• perinatal
• postpartum
• postnatal
• treatment

1. Introduction

Postpartum psychiatric disorders have traditionally been classified as baby blues, postpartum depression (PPD), and puerperal psychosis. The baby blues is not a psychiatric illness but rather a constellation of transitory symptoms for which no treatment is required. Similarly, puerperal psychosis is a misnomer because it is generally a mood episode with psychotic features in women with bipolar disorder [1]. Postpartum depression is commonly defined as occurrence of a major depressive episode in the first few weeks after delivery although some women experience onset of symptoms during pregnancy. Affecting approximately 14% of women within the first 4–6 weeks of delivery [2], PPD is one of the most common medical complications of childbirth. Untreated or inappropriately treated PPD is associated with multiple adverse maternal and infant outcomes [3]. Thus, correct diagnosis in a timely manner is critically important for optimal treatment of women with depression and related disorders. The term PPD has been in use for half a century both as lay term and a diagnostic entity. Due to the increased public awareness of depression in the postpartum period and consequently reduced stigma, women are more likely to undergo screening and seek professional help [4]. For some women, a diagnosis of PPD may be more acceptable than a diagnosis of a mood disorder due to the purported role of hormonal imbalance in the former.

In this Editorial, the limitations of using PPD as a diagnostic category are discussed. It is argued that due to its limited diagnostic and therapeutic utility, preference should be given to more specific terminology to capture the heterogeneity of depression in the postpartum period.

2. Evolution of the concept

Whether or not PPD is a distinct illness has been debated for more than 150 years [5]. Jean-Etienne Esquirol was one of the first physicians to provide descriptions of postpartum mental illness among women admitted to Salpêtrière Hospital in Paris [6]. Esquirol was of the opinion that the prevalence of puerperal mental illness was much greater in the community than the data from mental hospitals would indicate. In his influential monograph Treatise on Insanity in Pregnant, Postpartum, and Lactating Women published in 1858, Louis-Victor Marcé first provided a systematic account of psychiatric illness during and after pregnancy [7]. He argued that PPD should be accorded a distinct diagnostic status even though most symptoms of the syndrome could be found in cases of non-perperperal depression.

With the advent of antidepressant drugs in the 1950s, there was renewed interest in the nosology of PPD, particularly among women who were not ‘ill enough to be admitted to hospital’ [8]. In a seminal paper published in 1968, Bryce Pitt suggested that PPD should be viewed on a continuum between severe puerperal depression on the one hand and the commonly occurring baby blues on the other. He found that a large number of cases of PPD were characterized by ‘atypical’ symptoms, such as anxiety, phobias, and irritability or had symptoms that were the opposite of classical depression, e.g. reversed diurnality and early rather than late insomnia. According to Pitt, the atypical depression is more likely to be encountered in outpatient settings compared to the ‘severe puerperal depression.’ Over time, Pitt’s severity-based classification was neglected and the atypical variant per se became symbolic of PPD. With the publication of its fourth edition in 1996, the DSM first acknowledged that PPD can also occur as an episode of bipolar I disorder. The DSM-5 expanded the scope of the specifier to characterize cases of bipolar II depression. And for the first time the specifier could be used for mood episodes occurring during pregnancy. In spite of the ‘official’ recognition that a postpartum depressive episode can occur as part of major depressive disorder or bipolar disorder, the term PPD does not make a distinction between the two main subtypes of depression in the postpartum period.

3. Comorbidities and associated features

There is emerging evidence that major depressive disorder is just one of several psychiatric disorders triggered by childbirth. Hypo/manic symptoms are common in the postpartum period and can occur as part of bipolar disorder or major depressive disorder with mixed features. Subthreshold episodes of hypomania are ubiquitous after delivery and are associated with increased risk of depression in the postpartum period [9].

Postpartum depression is invariably accompanied by anxiety and related disorders including generalized anxiety disorder, panic disorder, obsessive-compulsive disorder and post-traumatic stress disorder. There is emerging evidence that women may be at a particularly increased risk of first onset or recurrence of obsessive-compulsive disorder in the postpartum period [10]. A fear of having or acquiring grave physical illness as in illness anxiety disorder also appears common; however, women are reluctant to share physical health concerns in psychiatric settings. Unusual irritability is common but often neglected. Singular focus on depression may affect the identification and diagnosis of its common comorbidities.

4. Diagnostic and treatment implications

A diagnosis of PPD is generally based on a cross-sectional assessment of symptoms. A postpartum episode can be the first occurrence of depression, but the majority of women have a history of depression. First onset of depression after delivery is associated with increased risk of future hypo/manic episodes. Awareness of clinical features, such as timing of episode onset vis-à-vis delivery (pregnancy vs. postpartum), type of onset (rapid vs. gradual), polarity [11], and type of symptoms (typical vs. atypical) is important for diagnostic and treatment purposes [12]. Equally important is an understanding of the longitudinal illness course.

Due to increased awareness of PPD, women are commonly screened and assessed for depression but not for other commonly occurring psychiatric disorders. The Edinburgh Postnatal Depression Scale (EPDS) – perhaps the most commonly used screening instrument for PPD – has three items to screen for anxiety. However, the EPDS may not reliably discriminate between depression and anxiety. In spite of their common occurrence and associated emotional distress and disability, women are generally not screened or assessed for co-morbid psychiatric disorders.

Failure to screen for hypo/mania may lead to misdiagnosis of bipolar disorder as major depressive disorder. Use of antidepressants in these patients may induce hypo/mania and increase risk of self-harm and psychiatric hospitalization. A population-based cohort study from Denmark found a previous diagnosis of bipolar disorder was the strongest predictor of rehospitalization 10–19 days postpartum. In total approximately 27% of women with this diagnosis required rehospitalisation within the first postpartum year [13].

5. Reconceptualization of postpartum depression

In light of its limited clinical utility, the term PPD should be abandoned in favor of more specific terminology, such as major depressive disorder or bipolar disorder with peripartum onset. In addition to clarifying the polarity, the application of DSM-5 specifiers, such as peripartum onset, anxious distress, or mixed features should provide clinically useful characterization of the index depressive episode. Positioning PPD in the context of an existing or developing mood disorder is similar to the approach used for physical illnesses with high risk of postpartum recurrence. For example, we do not use the term postpartum multiple sclerosis even though up to 30% of women with multiple sclerosis have a postpartum recurrence. In other instances, a specific term (such as gestational diabetes) is used when a woman without past history develops the disease for the first time during pregnancy.

The recommended approach should facilitate rigorous follow-up of women who are at risk of having postpartum occurrence of depression due to the personal history of depression, anxiety disorders or obsessive-compulsive disorder. Since more than half of women with PPD have prepartum onset, early intervention during pregnancy may attenuate or prevent depression in the postpartum period. A similar approach could be taken in the care of primiparous women who are at risk of developing depression due to the history of mood disorders in first-degree relatives. Prospective monitoring of at-risk women should allow for easier identification of hypomanic symptoms that closely follow delivery. Reconceptualizing PPD as an episode of major depressive disorder or bipolar disorder would prompt clinicians to carry out comprehensive assessment to clarify the diagnosis before embarking upon treatment. Diagnosing PPD using the diagnostic criteria should also facilitate studies on the impact of maternal depression on infant attachment [14]. Use of safe and disorder-specific treatments could improve the outcome of depression. Antidepressants are the mainstay of drug treatment of unipolar PPD and are recommended for moderate to severe PPD, or for women who fail to respond to or refuse psychological interventions. However, two systematic reviews concluded that there is insufficient evidence from clinical trials to recommend treatment with antidepressants. It is possible that failure to exclude women with bipolar disorder (particularly those with bipolar II or other specified bipolar and related disorder) may have affected the results of these studies. In contrast to the substantial literature (six open-label studies and nine randomized controlled trials of antidepressants) on the drug treatment of unipolar PPD, the literature on bipolar PPD is sparse and consists of only two open trials of quetiapine [15,16].

In conclusion, it is proposed that episodes of depression that are now subsumed under the PPD umbrella should be repositioned within the existing DSM classification framework for mood disorders. Although there might be some benefit in retaining the term for lay descriptions or explanations (e.g. treatment seeking, stigma reduction), there would be greater benefit to clinical practice and research if such episodes were classified within the current framework for episodes of mood disorders.

Declaration of interest

The author reports grants from Otsuka and the Academic Medical Organization of Southwestern Ontario and participating on advisory boards for Sunovion Pharmaceuticals and Otsuka, outside the submitted work. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or conflict with the subject matter or materials discussed in this manuscript apart from those disclosed.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This paper was not funded.