Marlene P. Freeman, MD

In an age when psychiatry strives to identify the biologic causes of disease, studying endocrine-related mood disorders is particularly intriguing. DSM-5 defines premenstrual dysphoric disorder (PMDD) as a depressive disorder, with a 12-month prevalence ranging from 1.8% to 5.8% among women who menstruate.^1-3 Factors that differentiate PMDD from other affective disorders include etiology, duration, and temporal relationship with the menstrual cycle.

PMDD is a disorder of consistent yet intermittent change in mental health and functionality. Therefore, it may be underdiagnosed and consequently undertreated if a psychiatric evaluation does not coincide with symptom occurrence or if patients do not understand that intermittent symptoms are treatable.

This article summarizes what is known about the etiology of PMDD. Although there are several treatments for PMDD, many women experience adverse effects or incomplete effectiveness. Further understanding of this disorder may lead to more efficacious treatments. Additionally, understanding the pathophysiology of PMDD might shed a light on the etiology of other disorders that are temporally related to reproductive life changes, such as pregnancy-, postpartum-, or menopause-related affective dysregulation.

Making the diagnosis

The diagnosis of PMDD is made when a patient has at least 5 of 11 specific symptoms that occur during the week before onset of menses, improve within a few days after the onset of menses (shown as the “PMDD Hazard Zone” in Figure 1), and are minimal or absent post-menses.³ Symptoms should be tracked prospectively for at least 2 menstrual cycles in order to confirm the diagnosis (one must be an affective symptom and another must be a behavioral/cognitive symptom).³

The affective symptoms are:

• lability of affect (eg, sudden sadness, tearfulness, or sensitivity to rejection)
• irritability, anger, or increased inter­personal conflicts
• depressed mood, hopelessness, or self- deprecating thoughts
• anxiety or tension, feeling “keyed up” or “on edge.”

The behavioral/cognitive symptoms are:

• decreased interest in usual activities (eg, work, hobbies, friends, school)
• difficulty concentrating
• lethargy, low energy, easy fatigability
• change in appetite, overeating, food cravings
• hypersomnia or insomnia
• feeling overwhelmed or out of control
• physical symptoms (breast tenderness or swelling, headache, joint or muscle pain, bloating, weight gain).

Ruling out premenstrual exacerbation (PME). Perhaps the most common cause for misdiagnosis of PMDD is failing to rule out PME of another underlying or comorbid condition (Figure 2). In many women who have a primary mood or anxiety disorder, the late luteal phase is a vulnerable time. A patient might be coping with untreated anxiety, for example, but the symptoms become unbearable the week before menstruation begins, which is likely when she seeks help. At this stage, a diagnosis of PMDD should be provisional at best. Often, PME is treated by treating the underlying condition. Therefore, a full diagnostic psychiatric interview is important to first rule out other underlying psychiatric disorders. PMDD is diagnosed if the premenstrual symptoms persist for 2 consecutive months after treating the suspected mood or anxiety disorder. Patients can use one of many PMDD daily symptom charts available online. Alternatively, they can use a cycle-tracking mobile phone application to correlate their symptoms with their cycle and share this information with their providers.

Consider these 5 interwoven pieces

The many variables that contribute to the pathophysiology of PMDD overlap and should be considered connecting pieces in the puzzle that is the etiology of this disorder (Figure 3). In reviewing the literature, we have identified 5 topics likely to be major contributors to this disorder:

1. genetic susceptibility
2. progesterone and allopregnanolone (ALLO)
3. estrogen, serotonin, and brain-derived neurotrophic factor (BDNF)
4. putative brain structural and functional differences
5. further involvement of the hypothalamic–pituitary–adrenal (HPA) axis and hypothalamic–pituitary–gonadal (HPG) axis: trauma, resiliency, and inflammation.

Genetic susceptibility. PMDD is thought to have a heritability range between 30% to 80%.³ This is demonstrated by family and twin studies^4-7 and specific genetic studies.⁸ The involvement of genetics means an underlying neurobiologic pathophysiology is in place.

Estrogen receptor alpha (ESR1) gene. Huo et al⁸ found an associated variation in ESR1 in women with PMDD compared with controls. They speculated that because ESR1 is important for arousal, if dysfunctional, this gene could be implicated in somatic as well as affective and cognitive deficits in PMDD patients. In another study, investigators reported a relationship between PMDD and heritable personality traits, as well as a link between these traits and ESR1 polymorphic variants.¹ They suggested that personality traits (independent of affective state) might be used to distinguish patients with PMDD from controls.¹

Studies on serotonin gene polymorphism and serotonin transporter genotype. Although a study of serotonin gene polymorphism did not find an association between serotonin1A gene polymorphism and PMDD, it did show that the presence of at least 1 C allele was associated with a 2.5-fold increased risk of PMDD.⁹ Another study did not find an association between the serotonin transporter genotype 5-HTTLPR and PMDD.¹⁰ However, it showed lower frontocingulate cortex activation during the luteal phase of PMDD patients compared with controls, suggesting that PMDD is linked to impaired frontocingulate cortex activation induced by emotions during the luteal phase.¹⁰ 

Seasonal affective disorder (SAD) and PMDD have shared clinical features. A polymorphism in the serotonin transporter promoter gene 5-HTTLPR has been associated with SAD. One study found that patients with comorbid SAD and PMDD are genetically more vulnerable to comorbid affective disorders compared with patients who have SAD only.¹¹

Progesterone and ALLO. Chronic exposure to progesterone and ALLO (a main progesterone metabolite) and rapid withdrawal from ovarian hormones may play a role in the etiology of PMDD. Much like alcohol or benzodiazepines, ALLO is a potent positive allosteric modulator of GABA_A receptors and has sedative, anesthetic, and anxiolytic properties. In times of acute stress, increased ALLO is known to provide relief.^12,13 However, in women with PMDD, this typical ALLO increase might not occur.¹⁴

Patients with PMDD have been reported to have decreased levels of ALLO in the luteal phase.^15-17 In one study, women with highly symptomatic PMDD had lower levels of ALLO compared with women with less symptomatic PMDD.¹⁴ A gonadotropin-releasing hormone challenge study showed the increase in ALLO response was less in PMDD patients compared with controls.¹⁷ Luteal-phase ALLO concentrations are reported to be lower in women with premenstrual syndrome (PMS), a milder form of PMDD.^14,17

The efficacy of selective serotonin reuptake inhibitors (SSRIs) for treating PMDD could be the result of the interaction of these medications with neuroactive steroids,¹⁸ possibly because SSRIs enhance the sensitivity of GABA_A receptors or promote the formation of more ALLO (Figure 4).^19-21

Estrogen, serotonin, and BDNF. Estrogen affects multiple neurotransmitter systems that regulate mood, cognition, sleep, and eating.²² Studying estrogen in context of PMDD is important because women with PMDD can have low mood, specific food cravings, and impaired cognitive function.

Estrogen–serotonin interactions are thought to be involved in hormone-related mood disorders such as perimenopausal depression and PMDD.^23,24 However, the nature of their relationship is not yet fully understood. Ovariectomized animals have shown estrogen-induced changes related to serotonin metabolism, binding, and transmission in the regions of the brain involved in regulation of affect and cognition. Research in menopausal women also has provided some support for this interaction.²⁴

Positron emission tomography studies in humans have found increased cortical serotonin binding modulated by levels of estrogen, similar to those previously seen in rat studies.^24-27 One study showed an increased binding potential of serotonin in the cerebral cortex with estrogen treatment. This study further showed an even greater binding potential with estrogen plus progesterone, signaling a synergistic effect of the 2 hormones.²⁸

SSRIs are an effective treatment for the irritability, anxiety, and mood swings of PMDD.^29-30 Although the exact mechanism of action is unknown, the serotonergic properties are certainly of primary attention. For some PMDD patients, SSRIs work within hours to days, as opposed to days or weeks for patients with depression or anxiety, which suggests a separate or co-occurring mechanism of action is in place. In a double-blind, placebo-controlled crossover study, researchers administered the serotonin receptor antagonist metergoline to women with PMDD whose symptoms had remitted during treatment with fluoxetine and a group of healthy controls who were not receiving any medication.³¹ The women with PMDD experienced a return of symptoms 24 hours after treatment with metergoline but not with placebo; the controls experienced no mood changes.³¹

BDNF is a neurotransmitter linked to estrogen and likely related to PMDD. BDNF is critical for neurogenesis and is expressed in brain regions involved in learning and memory and also affects regulation.³² BDNF levels are increased by serotonergic antidepressants, affected by estradiol, and have cyclicity throughout the menstrual cycle.^33-35

Putative brain structural and functional differences. Imaging studies have suggested differences in brain structure in women with PMDD, with a focus on the amygdala and the prefrontal cortex. Women with PMDD have greater gray matter volume in the posterior cerebellum,³⁶ greater gray matter density of hippocampal cortex, and lower gray matter density in the parahippocampal cortex.³⁷

Some studies have shown a functional variability of the amygdala’s response to stress in women with PMDD vs healthy controls.^38,39 A proton magnetic resonance spectroscopy (1H-MRS) study of the displays the possibility of an altered GABAergic function in patients with PMDD.⁴⁰

Patients will PMDD have enhanced dorsolateral prefrontal cortex reactivity when anticipating negative stimuli (but not to the actual exposure) during the luteal phase. A positive correlation between this reactivity and progesterone levels also was observed.⁴¹ Some researchers have suggested that prefrontal cortex dysfunction may be a risk factor for PMDD.⁴²

HPA axis and HPG axis: Trauma, resiliency, inflammation. Altered cortisol levels (higher during the luteal phase⁴³ and lower during times of stress^14,44) suggest a possibly altered HPA axis in some women with PMDD. However, studies on this topic have been few and inconsistent.

Dysregulation of the HPG axis could cause vasomotor symptoms, sleep dysregulation, and mood symptoms during menopause; women with PMDD can also experience these symptoms. The influence of estrogen and progesterone on mood is also highly dependent on this axis.

Ultimately, the interplay between the HPA axis and the HPG axis is important. One study found that women with PMDD who had high serum ALLO levels (HPG-related) had blunted nocturnal cortisol levels (HPA-related) compared with healthy controls who had low ALLO levels.⁴⁵

Significant stress and trauma exposure have been associated with PMDD. A study of 3,968 women found a history of trauma and PTSD were independently associated with PMDD.⁴⁶ Another study of approximately 3,000 women found a strong correlation between abuse and PMS.⁴⁷ However, a third study found no correlations between PMDD and trauma.⁴⁸

Patients with a predisposition to PMDD may be more vulnerable to develop a posttraumatic stress-related disorder, perhaps due to decreased biologic resiliency. For example, the startle response (hyper­vigilance) has been shown to be different in women with PMDD. One study suggested that suboptimal production of premenstrual ALLO may lead to increased arousal and increased stress reactivity to psycho­social or environmental triggers.⁴⁹

The possible role of inflammation in PMDD deserves further investigation. The luteal phase entails an increase in the production of proinflammatory markers.^50,51 A 10-fold increase in progesterone is correlated with a 20% to 23% increase in C-reactive protein levels.^52,53 Women with inflammatory diseases (eg, gingivitis or irritable bowel syndrome) show worsening of symptoms prior to menstruation.^54-56 One study found increased levels of proinflammatory markers in women with PMDD compared with controls.⁵⁷

Because PMDD is heritable, it must have an underlying neurobiologic pathophysiology. Brain imaging studies show differences in structure and function in women with PMDD across the menstrual cycle. Conversion of progesterone to ALLO and the GABAergic influence of this metabolite is a topic of interest in current research. Similarly, the role of estrogen and its connection to serotonin and other neurotransmitters such as BDNF have been implicated.

The link between a history of stress, trauma, and PMDD raises the question of biologic resiliency and illness in these patients, as it connects to the HPA and HPG axis and production of inflammatory stress hormones and steroid hormones and their metabolites. PMDD can be conceptualized as variable sensitivity to hormonal response to stress,⁵⁸ thus contextualizing biochemical and psychological resiliency.

Further research is needed to clarify the possibility of a shared pathophysiology between endocrine-related mood disorders such as postpartum depression (PPD) and PMDD because current research is controversial.^59,60 In PPD, women who are exposed to high levels of progesterone and estrogen during pregnancy (just like in the mid-luteal phase) have a sudden drop in these hormones postpartum.

The ‘withdrawal theory.’ The affective symptoms of PMDD resolve almost instantaneously after the start of menstruation. Perhaps this type of immediate relief is akin to substance use disorders and symptoms of withdrawal. It could be that reinstatement of a certain amount of gonadal steroids in the follicular phase of the cycle diminishes a withdrawal-like response to these steroids.

Currently, the main leading theory is that PMDD is a result of “an abnormal response to normal hormonal changes.”⁶¹ A new study also has shown that the change in estradiol/progesterone levels (vs the steady state) was associated with PMDD symptoms.⁶² Thinking of PMDD as a disorder of withdrawal offers an alternative (yet complementary) perspective to the current theory: PMDD may be caused by the absence or diminishing of the above-named hormones and their metabolites in the late luteal phase (in the context of developed “tolerance” during the early- to mid-luteal phase).

Considering the interplay between neuro­transmitters and neurosteroids, both a “serotonin withdrawal theory” (caused by a drop in steroid hormones) and a “GABAergic withdrawal theory” (due to the decline in progesterone) could be proposed. This theory would be supported by the fact that SSRIs seem to mitigate symptoms of PMDD as well as the genetic association between PMDD and ESR1. It is more than likely that the “withdrawal” is caused by the interactions between estrogen-serotonin, progesterone-ALLO, and GABA receptors, and the complementary fashion in which progesterone and estrogen influence each other.

In an age when psychiatry strives to identify the biologic causes of disease, studying endocrine-related mood disorders is particularly intriguing. DSM-5 defines premenstrual dysphoric disorder (PMDD) as a depressive disorder, with a 12-month prevalence ranging from 1.8% to 5.8% among women who menstruate.^1-3 Factors that differentiate PMDD from other affective disorders include etiology, duration, and temporal relationship with the menstrual cycle.

PMDD is a disorder of consistent yet intermittent change in mental health and functionality. Therefore, it may be underdiagnosed and consequently undertreated if a psychiatric evaluation does not coincide with symptom occurrence or if patients do not understand that intermittent symptoms are treatable.

This article summarizes what is known about the etiology of PMDD. Although there are several treatments for PMDD, many women experience adverse effects or incomplete effectiveness. Further understanding of this disorder may lead to more efficacious treatments. Additionally, understanding the pathophysiology of PMDD might shed a light on the etiology of other disorders that are temporally related to reproductive life changes, such as pregnancy-, postpartum-, or menopause-related affective dysregulation.

Making the diagnosis

The diagnosis of PMDD is made when a patient has at least 5 of 11 specific symptoms that occur during the week before onset of menses, improve within a few days after the onset of menses (shown as the “PMDD Hazard Zone” in Figure 1), and are minimal or absent post-menses.³ Symptoms should be tracked prospectively for at least 2 menstrual cycles in order to confirm the diagnosis (one must be an affective symptom and another must be a behavioral/cognitive symptom).³

The affective symptoms are:

• lability of affect (eg, sudden sadness, tearfulness, or sensitivity to rejection)
• irritability, anger, or increased inter­personal conflicts
• depressed mood, hopelessness, or self- deprecating thoughts
• anxiety or tension, feeling “keyed up” or “on edge.”

The behavioral/cognitive symptoms are:

• decreased interest in usual activities (eg, work, hobbies, friends, school)
• difficulty concentrating
• lethargy, low energy, easy fatigability
• change in appetite, overeating, food cravings
• hypersomnia or insomnia
• feeling overwhelmed or out of control
• physical symptoms (breast tenderness or swelling, headache, joint or muscle pain, bloating, weight gain).

Ruling out premenstrual exacerbation (PME). Perhaps the most common cause for misdiagnosis of PMDD is failing to rule out PME of another underlying or comorbid condition (Figure 2). In many women who have a primary mood or anxiety disorder, the late luteal phase is a vulnerable time. A patient might be coping with untreated anxiety, for example, but the symptoms become unbearable the week before menstruation begins, which is likely when she seeks help. At this stage, a diagnosis of PMDD should be provisional at best. Often, PME is treated by treating the underlying condition. Therefore, a full diagnostic psychiatric interview is important to first rule out other underlying psychiatric disorders. PMDD is diagnosed if the premenstrual symptoms persist for 2 consecutive months after treating the suspected mood or anxiety disorder. Patients can use one of many PMDD daily symptom charts available online. Alternatively, they can use a cycle-tracking mobile phone application to correlate their symptoms with their cycle and share this information with their providers.

Consider these 5 interwoven pieces

The many variables that contribute to the pathophysiology of PMDD overlap and should be considered connecting pieces in the puzzle that is the etiology of this disorder (Figure 3). In reviewing the literature, we have identified 5 topics likely to be major contributors to this disorder:

1. genetic susceptibility
2. progesterone and allopregnanolone (ALLO)
3. estrogen, serotonin, and brain-derived neurotrophic factor (BDNF)
4. putative brain structural and functional differences
5. further involvement of the hypothalamic–pituitary–adrenal (HPA) axis and hypothalamic–pituitary–gonadal (HPG) axis: trauma, resiliency, and inflammation.

Genetic susceptibility. PMDD is thought to have a heritability range between 30% to 80%.³ This is demonstrated by family and twin studies^4-7 and specific genetic studies.⁸ The involvement of genetics means an underlying neurobiologic pathophysiology is in place.

Estrogen receptor alpha (ESR1) gene. Huo et al⁸ found an associated variation in ESR1 in women with PMDD compared with controls. They speculated that because ESR1 is important for arousal, if dysfunctional, this gene could be implicated in somatic as well as affective and cognitive deficits in PMDD patients. In another study, investigators reported a relationship between PMDD and heritable personality traits, as well as a link between these traits and ESR1 polymorphic variants.¹ They suggested that personality traits (independent of affective state) might be used to distinguish patients with PMDD from controls.¹

Studies on serotonin gene polymorphism and serotonin transporter genotype. Although a study of serotonin gene polymorphism did not find an association between serotonin1A gene polymorphism and PMDD, it did show that the presence of at least 1 C allele was associated with a 2.5-fold increased risk of PMDD.⁹ Another study did not find an association between the serotonin transporter genotype 5-HTTLPR and PMDD.¹⁰ However, it showed lower frontocingulate cortex activation during the luteal phase of PMDD patients compared with controls, suggesting that PMDD is linked to impaired frontocingulate cortex activation induced by emotions during the luteal phase.¹⁰ 

Seasonal affective disorder (SAD) and PMDD have shared clinical features. A polymorphism in the serotonin transporter promoter gene 5-HTTLPR has been associated with SAD. One study found that patients with comorbid SAD and PMDD are genetically more vulnerable to comorbid affective disorders compared with patients who have SAD only.¹¹

Progesterone and ALLO. Chronic exposure to progesterone and ALLO (a main progesterone metabolite) and rapid withdrawal from ovarian hormones may play a role in the etiology of PMDD. Much like alcohol or benzodiazepines, ALLO is a potent positive allosteric modulator of GABA_A receptors and has sedative, anesthetic, and anxiolytic properties. In times of acute stress, increased ALLO is known to provide relief.^12,13 However, in women with PMDD, this typical ALLO increase might not occur.¹⁴

Patients with PMDD have been reported to have decreased levels of ALLO in the luteal phase.^15-17 In one study, women with highly symptomatic PMDD had lower levels of ALLO compared with women with less symptomatic PMDD.¹⁴ A gonadotropin-releasing hormone challenge study showed the increase in ALLO response was less in PMDD patients compared with controls.¹⁷ Luteal-phase ALLO concentrations are reported to be lower in women with premenstrual syndrome (PMS), a milder form of PMDD.^14,17

The efficacy of selective serotonin reuptake inhibitors (SSRIs) for treating PMDD could be the result of the interaction of these medications with neuroactive steroids,¹⁸ possibly because SSRIs enhance the sensitivity of GABA_A receptors or promote the formation of more ALLO (Figure 4).^19-21

Estrogen, serotonin, and BDNF. Estrogen affects multiple neurotransmitter systems that regulate mood, cognition, sleep, and eating.²² Studying estrogen in context of PMDD is important because women with PMDD can have low mood, specific food cravings, and impaired cognitive function.

Estrogen–serotonin interactions are thought to be involved in hormone-related mood disorders such as perimenopausal depression and PMDD.^23,24 However, the nature of their relationship is not yet fully understood. Ovariectomized animals have shown estrogen-induced changes related to serotonin metabolism, binding, and transmission in the regions of the brain involved in regulation of affect and cognition. Research in menopausal women also has provided some support for this interaction.²⁴

Positron emission tomography studies in humans have found increased cortical serotonin binding modulated by levels of estrogen, similar to those previously seen in rat studies.^24-27 One study showed an increased binding potential of serotonin in the cerebral cortex with estrogen treatment. This study further showed an even greater binding potential with estrogen plus progesterone, signaling a synergistic effect of the 2 hormones.²⁸

SSRIs are an effective treatment for the irritability, anxiety, and mood swings of PMDD.^29-30 Although the exact mechanism of action is unknown, the serotonergic properties are certainly of primary attention. For some PMDD patients, SSRIs work within hours to days, as opposed to days or weeks for patients with depression or anxiety, which suggests a separate or co-occurring mechanism of action is in place. In a double-blind, placebo-controlled crossover study, researchers administered the serotonin receptor antagonist metergoline to women with PMDD whose symptoms had remitted during treatment with fluoxetine and a group of healthy controls who were not receiving any medication.³¹ The women with PMDD experienced a return of symptoms 24 hours after treatment with metergoline but not with placebo; the controls experienced no mood changes.³¹

BDNF is a neurotransmitter linked to estrogen and likely related to PMDD. BDNF is critical for neurogenesis and is expressed in brain regions involved in learning and memory and also affects regulation.³² BDNF levels are increased by serotonergic antidepressants, affected by estradiol, and have cyclicity throughout the menstrual cycle.^33-35

Putative brain structural and functional differences. Imaging studies have suggested differences in brain structure in women with PMDD, with a focus on the amygdala and the prefrontal cortex. Women with PMDD have greater gray matter volume in the posterior cerebellum,³⁶ greater gray matter density of hippocampal cortex, and lower gray matter density in the parahippocampal cortex.³⁷

Some studies have shown a functional variability of the amygdala’s response to stress in women with PMDD vs healthy controls.^38,39 A proton magnetic resonance spectroscopy (1H-MRS) study of the displays the possibility of an altered GABAergic function in patients with PMDD.⁴⁰

Patients will PMDD have enhanced dorsolateral prefrontal cortex reactivity when anticipating negative stimuli (but not to the actual exposure) during the luteal phase. A positive correlation between this reactivity and progesterone levels also was observed.⁴¹ Some researchers have suggested that prefrontal cortex dysfunction may be a risk factor for PMDD.⁴²

HPA axis and HPG axis: Trauma, resiliency, inflammation. Altered cortisol levels (higher during the luteal phase⁴³ and lower during times of stress^14,44) suggest a possibly altered HPA axis in some women with PMDD. However, studies on this topic have been few and inconsistent.

Dysregulation of the HPG axis could cause vasomotor symptoms, sleep dysregulation, and mood symptoms during menopause; women with PMDD can also experience these symptoms. The influence of estrogen and progesterone on mood is also highly dependent on this axis.

Ultimately, the interplay between the HPA axis and the HPG axis is important. One study found that women with PMDD who had high serum ALLO levels (HPG-related) had blunted nocturnal cortisol levels (HPA-related) compared with healthy controls who had low ALLO levels.⁴⁵

Significant stress and trauma exposure have been associated with PMDD. A study of 3,968 women found a history of trauma and PTSD were independently associated with PMDD.⁴⁶ Another study of approximately 3,000 women found a strong correlation between abuse and PMS.⁴⁷ However, a third study found no correlations between PMDD and trauma.⁴⁸

Patients with a predisposition to PMDD may be more vulnerable to develop a posttraumatic stress-related disorder, perhaps due to decreased biologic resiliency. For example, the startle response (hyper­vigilance) has been shown to be different in women with PMDD. One study suggested that suboptimal production of premenstrual ALLO may lead to increased arousal and increased stress reactivity to psycho­social or environmental triggers.⁴⁹

The possible role of inflammation in PMDD deserves further investigation. The luteal phase entails an increase in the production of proinflammatory markers.^50,51 A 10-fold increase in progesterone is correlated with a 20% to 23% increase in C-reactive protein levels.^52,53 Women with inflammatory diseases (eg, gingivitis or irritable bowel syndrome) show worsening of symptoms prior to menstruation.^54-56 One study found increased levels of proinflammatory markers in women with PMDD compared with controls.⁵⁷

Because PMDD is heritable, it must have an underlying neurobiologic pathophysiology. Brain imaging studies show differences in structure and function in women with PMDD across the menstrual cycle. Conversion of progesterone to ALLO and the GABAergic influence of this metabolite is a topic of interest in current research. Similarly, the role of estrogen and its connection to serotonin and other neurotransmitters such as BDNF have been implicated.

The link between a history of stress, trauma, and PMDD raises the question of biologic resiliency and illness in these patients, as it connects to the HPA and HPG axis and production of inflammatory stress hormones and steroid hormones and their metabolites. PMDD can be conceptualized as variable sensitivity to hormonal response to stress,⁵⁸ thus contextualizing biochemical and psychological resiliency.

Further research is needed to clarify the possibility of a shared pathophysiology between endocrine-related mood disorders such as postpartum depression (PPD) and PMDD because current research is controversial.^59,60 In PPD, women who are exposed to high levels of progesterone and estrogen during pregnancy (just like in the mid-luteal phase) have a sudden drop in these hormones postpartum.

The ‘withdrawal theory.’ The affective symptoms of PMDD resolve almost instantaneously after the start of menstruation. Perhaps this type of immediate relief is akin to substance use disorders and symptoms of withdrawal. It could be that reinstatement of a certain amount of gonadal steroids in the follicular phase of the cycle diminishes a withdrawal-like response to these steroids.

Currently, the main leading theory is that PMDD is a result of “an abnormal response to normal hormonal changes.”⁶¹ A new study also has shown that the change in estradiol/progesterone levels (vs the steady state) was associated with PMDD symptoms.⁶² Thinking of PMDD as a disorder of withdrawal offers an alternative (yet complementary) perspective to the current theory: PMDD may be caused by the absence or diminishing of the above-named hormones and their metabolites in the late luteal phase (in the context of developed “tolerance” during the early- to mid-luteal phase).

Considering the interplay between neuro­transmitters and neurosteroids, both a “serotonin withdrawal theory” (caused by a drop in steroid hormones) and a “GABAergic withdrawal theory” (due to the decline in progesterone) could be proposed. This theory would be supported by the fact that SSRIs seem to mitigate symptoms of PMDD as well as the genetic association between PMDD and ESR1. It is more than likely that the “withdrawal” is caused by the interactions between estrogen-serotonin, progesterone-ALLO, and GABA receptors, and the complementary fashion in which progesterone and estrogen influence each other.

In an age when psychiatry strives to identify the biologic causes of disease, studying endocrine-related mood disorders is particularly intriguing. DSM-5 defines premenstrual dysphoric disorder (PMDD) as a depressive disorder, with a 12-month prevalence ranging from 1.8% to 5.8% among women who menstruate.^1-3 Factors that differentiate PMDD from other affective disorders include etiology, duration, and temporal relationship with the menstrual cycle.

PMDD is a disorder of consistent yet intermittent change in mental health and functionality. Therefore, it may be underdiagnosed and consequently undertreated if a psychiatric evaluation does not coincide with symptom occurrence or if patients do not understand that intermittent symptoms are treatable.

This article summarizes what is known about the etiology of PMDD. Although there are several treatments for PMDD, many women experience adverse effects or incomplete effectiveness. Further understanding of this disorder may lead to more efficacious treatments. Additionally, understanding the pathophysiology of PMDD might shed a light on the etiology of other disorders that are temporally related to reproductive life changes, such as pregnancy-, postpartum-, or menopause-related affective dysregulation.

Making the diagnosis

The diagnosis of PMDD is made when a patient has at least 5 of 11 specific symptoms that occur during the week before onset of menses, improve within a few days after the onset of menses (shown as the “PMDD Hazard Zone” in Figure 1), and are minimal or absent post-menses.³ Symptoms should be tracked prospectively for at least 2 menstrual cycles in order to confirm the diagnosis (one must be an affective symptom and another must be a behavioral/cognitive symptom).³

The affective symptoms are:

• lability of affect (eg, sudden sadness, tearfulness, or sensitivity to rejection)
• irritability, anger, or increased inter­personal conflicts
• depressed mood, hopelessness, or self- deprecating thoughts
• anxiety or tension, feeling “keyed up” or “on edge.”

The behavioral/cognitive symptoms are:

• decreased interest in usual activities (eg, work, hobbies, friends, school)
• difficulty concentrating
• lethargy, low energy, easy fatigability
• change in appetite, overeating, food cravings
• hypersomnia or insomnia
• feeling overwhelmed or out of control
• physical symptoms (breast tenderness or swelling, headache, joint or muscle pain, bloating, weight gain).

Ruling out premenstrual exacerbation (PME). Perhaps the most common cause for misdiagnosis of PMDD is failing to rule out PME of another underlying or comorbid condition (Figure 2). In many women who have a primary mood or anxiety disorder, the late luteal phase is a vulnerable time. A patient might be coping with untreated anxiety, for example, but the symptoms become unbearable the week before menstruation begins, which is likely when she seeks help. At this stage, a diagnosis of PMDD should be provisional at best. Often, PME is treated by treating the underlying condition. Therefore, a full diagnostic psychiatric interview is important to first rule out other underlying psychiatric disorders. PMDD is diagnosed if the premenstrual symptoms persist for 2 consecutive months after treating the suspected mood or anxiety disorder. Patients can use one of many PMDD daily symptom charts available online. Alternatively, they can use a cycle-tracking mobile phone application to correlate their symptoms with their cycle and share this information with their providers.

Consider these 5 interwoven pieces

The many variables that contribute to the pathophysiology of PMDD overlap and should be considered connecting pieces in the puzzle that is the etiology of this disorder (Figure 3). In reviewing the literature, we have identified 5 topics likely to be major contributors to this disorder:

1. genetic susceptibility
2. progesterone and allopregnanolone (ALLO)
3. estrogen, serotonin, and brain-derived neurotrophic factor (BDNF)
4. putative brain structural and functional differences
5. further involvement of the hypothalamic–pituitary–adrenal (HPA) axis and hypothalamic–pituitary–gonadal (HPG) axis: trauma, resiliency, and inflammation.

Genetic susceptibility. PMDD is thought to have a heritability range between 30% to 80%.³ This is demonstrated by family and twin studies^4-7 and specific genetic studies.⁸ The involvement of genetics means an underlying neurobiologic pathophysiology is in place.

Estrogen receptor alpha (ESR1) gene. Huo et al⁸ found an associated variation in ESR1 in women with PMDD compared with controls. They speculated that because ESR1 is important for arousal, if dysfunctional, this gene could be implicated in somatic as well as affective and cognitive deficits in PMDD patients. In another study, investigators reported a relationship between PMDD and heritable personality traits, as well as a link between these traits and ESR1 polymorphic variants.¹ They suggested that personality traits (independent of affective state) might be used to distinguish patients with PMDD from controls.¹

Studies on serotonin gene polymorphism and serotonin transporter genotype. Although a study of serotonin gene polymorphism did not find an association between serotonin1A gene polymorphism and PMDD, it did show that the presence of at least 1 C allele was associated with a 2.5-fold increased risk of PMDD.⁹ Another study did not find an association between the serotonin transporter genotype 5-HTTLPR and PMDD.¹⁰ However, it showed lower frontocingulate cortex activation during the luteal phase of PMDD patients compared with controls, suggesting that PMDD is linked to impaired frontocingulate cortex activation induced by emotions during the luteal phase.¹⁰ 

Seasonal affective disorder (SAD) and PMDD have shared clinical features. A polymorphism in the serotonin transporter promoter gene 5-HTTLPR has been associated with SAD. One study found that patients with comorbid SAD and PMDD are genetically more vulnerable to comorbid affective disorders compared with patients who have SAD only.¹¹

Progesterone and ALLO. Chronic exposure to progesterone and ALLO (a main progesterone metabolite) and rapid withdrawal from ovarian hormones may play a role in the etiology of PMDD. Much like alcohol or benzodiazepines, ALLO is a potent positive allosteric modulator of GABA_A receptors and has sedative, anesthetic, and anxiolytic properties. In times of acute stress, increased ALLO is known to provide relief.^12,13 However, in women with PMDD, this typical ALLO increase might not occur.¹⁴

Patients with PMDD have been reported to have decreased levels of ALLO in the luteal phase.^15-17 In one study, women with highly symptomatic PMDD had lower levels of ALLO compared with women with less symptomatic PMDD.¹⁴ A gonadotropin-releasing hormone challenge study showed the increase in ALLO response was less in PMDD patients compared with controls.¹⁷ Luteal-phase ALLO concentrations are reported to be lower in women with premenstrual syndrome (PMS), a milder form of PMDD.^14,17

The efficacy of selective serotonin reuptake inhibitors (SSRIs) for treating PMDD could be the result of the interaction of these medications with neuroactive steroids,¹⁸ possibly because SSRIs enhance the sensitivity of GABA_A receptors or promote the formation of more ALLO (Figure 4).^19-21

Estrogen, serotonin, and BDNF. Estrogen affects multiple neurotransmitter systems that regulate mood, cognition, sleep, and eating.²² Studying estrogen in context of PMDD is important because women with PMDD can have low mood, specific food cravings, and impaired cognitive function.

Estrogen–serotonin interactions are thought to be involved in hormone-related mood disorders such as perimenopausal depression and PMDD.^23,24 However, the nature of their relationship is not yet fully understood. Ovariectomized animals have shown estrogen-induced changes related to serotonin metabolism, binding, and transmission in the regions of the brain involved in regulation of affect and cognition. Research in menopausal women also has provided some support for this interaction.²⁴

Positron emission tomography studies in humans have found increased cortical serotonin binding modulated by levels of estrogen, similar to those previously seen in rat studies.^24-27 One study showed an increased binding potential of serotonin in the cerebral cortex with estrogen treatment. This study further showed an even greater binding potential with estrogen plus progesterone, signaling a synergistic effect of the 2 hormones.²⁸

SSRIs are an effective treatment for the irritability, anxiety, and mood swings of PMDD.^29-30 Although the exact mechanism of action is unknown, the serotonergic properties are certainly of primary attention. For some PMDD patients, SSRIs work within hours to days, as opposed to days or weeks for patients with depression or anxiety, which suggests a separate or co-occurring mechanism of action is in place. In a double-blind, placebo-controlled crossover study, researchers administered the serotonin receptor antagonist metergoline to women with PMDD whose symptoms had remitted during treatment with fluoxetine and a group of healthy controls who were not receiving any medication.³¹ The women with PMDD experienced a return of symptoms 24 hours after treatment with metergoline but not with placebo; the controls experienced no mood changes.³¹

BDNF is a neurotransmitter linked to estrogen and likely related to PMDD. BDNF is critical for neurogenesis and is expressed in brain regions involved in learning and memory and also affects regulation.³² BDNF levels are increased by serotonergic antidepressants, affected by estradiol, and have cyclicity throughout the menstrual cycle.^33-35

Putative brain structural and functional differences. Imaging studies have suggested differences in brain structure in women with PMDD, with a focus on the amygdala and the prefrontal cortex. Women with PMDD have greater gray matter volume in the posterior cerebellum,³⁶ greater gray matter density of hippocampal cortex, and lower gray matter density in the parahippocampal cortex.³⁷

Some studies have shown a functional variability of the amygdala’s response to stress in women with PMDD vs healthy controls.^38,39 A proton magnetic resonance spectroscopy (1H-MRS) study of the displays the possibility of an altered GABAergic function in patients with PMDD.⁴⁰

Patients will PMDD have enhanced dorsolateral prefrontal cortex reactivity when anticipating negative stimuli (but not to the actual exposure) during the luteal phase. A positive correlation between this reactivity and progesterone levels also was observed.⁴¹ Some researchers have suggested that prefrontal cortex dysfunction may be a risk factor for PMDD.⁴²

HPA axis and HPG axis: Trauma, resiliency, inflammation. Altered cortisol levels (higher during the luteal phase⁴³ and lower during times of stress^14,44) suggest a possibly altered HPA axis in some women with PMDD. However, studies on this topic have been few and inconsistent.

Dysregulation of the HPG axis could cause vasomotor symptoms, sleep dysregulation, and mood symptoms during menopause; women with PMDD can also experience these symptoms. The influence of estrogen and progesterone on mood is also highly dependent on this axis.

Ultimately, the interplay between the HPA axis and the HPG axis is important. One study found that women with PMDD who had high serum ALLO levels (HPG-related) had blunted nocturnal cortisol levels (HPA-related) compared with healthy controls who had low ALLO levels.⁴⁵

Significant stress and trauma exposure have been associated with PMDD. A study of 3,968 women found a history of trauma and PTSD were independently associated with PMDD.⁴⁶ Another study of approximately 3,000 women found a strong correlation between abuse and PMS.⁴⁷ However, a third study found no correlations between PMDD and trauma.⁴⁸

Patients with a predisposition to PMDD may be more vulnerable to develop a posttraumatic stress-related disorder, perhaps due to decreased biologic resiliency. For example, the startle response (hyper­vigilance) has been shown to be different in women with PMDD. One study suggested that suboptimal production of premenstrual ALLO may lead to increased arousal and increased stress reactivity to psycho­social or environmental triggers.⁴⁹

The possible role of inflammation in PMDD deserves further investigation. The luteal phase entails an increase in the production of proinflammatory markers.^50,51 A 10-fold increase in progesterone is correlated with a 20% to 23% increase in C-reactive protein levels.^52,53 Women with inflammatory diseases (eg, gingivitis or irritable bowel syndrome) show worsening of symptoms prior to menstruation.^54-56 One study found increased levels of proinflammatory markers in women with PMDD compared with controls.⁵⁷

Because PMDD is heritable, it must have an underlying neurobiologic pathophysiology. Brain imaging studies show differences in structure and function in women with PMDD across the menstrual cycle. Conversion of progesterone to ALLO and the GABAergic influence of this metabolite is a topic of interest in current research. Similarly, the role of estrogen and its connection to serotonin and other neurotransmitters such as BDNF have been implicated.

The link between a history of stress, trauma, and PMDD raises the question of biologic resiliency and illness in these patients, as it connects to the HPA and HPG axis and production of inflammatory stress hormones and steroid hormones and their metabolites. PMDD can be conceptualized as variable sensitivity to hormonal response to stress,⁵⁸ thus contextualizing biochemical and psychological resiliency.

Further research is needed to clarify the possibility of a shared pathophysiology between endocrine-related mood disorders such as postpartum depression (PPD) and PMDD because current research is controversial.^59,60 In PPD, women who are exposed to high levels of progesterone and estrogen during pregnancy (just like in the mid-luteal phase) have a sudden drop in these hormones postpartum.

The ‘withdrawal theory.’ The affective symptoms of PMDD resolve almost instantaneously after the start of menstruation. Perhaps this type of immediate relief is akin to substance use disorders and symptoms of withdrawal. It could be that reinstatement of a certain amount of gonadal steroids in the follicular phase of the cycle diminishes a withdrawal-like response to these steroids.

Currently, the main leading theory is that PMDD is a result of “an abnormal response to normal hormonal changes.”⁶¹ A new study also has shown that the change in estradiol/progesterone levels (vs the steady state) was associated with PMDD symptoms.⁶² Thinking of PMDD as a disorder of withdrawal offers an alternative (yet complementary) perspective to the current theory: PMDD may be caused by the absence or diminishing of the above-named hormones and their metabolites in the late luteal phase (in the context of developed “tolerance” during the early- to mid-luteal phase).

Considering the interplay between neuro­transmitters and neurosteroids, both a “serotonin withdrawal theory” (caused by a drop in steroid hormones) and a “GABAergic withdrawal theory” (due to the decline in progesterone) could be proposed. This theory would be supported by the fact that SSRIs seem to mitigate symptoms of PMDD as well as the genetic association between PMDD and ESR1. It is more than likely that the “withdrawal” is caused by the interactions between estrogen-serotonin, progesterone-ALLO, and GABA receptors, and the complementary fashion in which progesterone and estrogen influence each other.

In December 2014, the FDA issued draft guidance for sweeping changes to labeling of pharmaceutical treatments in regard to pregnancy and lactation information. These changes are now in effect for use in practice.¹ The undertaking has been years in the making, and is truly ambitious.

The outdated system of letter categories (A, B, C, D, X) falls short of clinical needs in several ways:

• the quality and volume of data can be lacking
• comparative risk is not described
• using letters can led to oversimplification or, in some cases, exaggeration of risk and safety (Box).

Other drawbacks include infrequent updating of information and omission of information about baseline rates of reproductive-related adverse events, to provide a more meaningful context for risk assessment.

A note before we continue discussion of labeling: Recognize that pregnancy itself is inherently risky; poor outcomes are, regrettably, not uncommon. The rate of birth defects in the United States is approximately 3%, and obstetric complications, such as prematurity, are common.^2,3

New system described
The new labeling content has been described in the FDA’s Pregnancy and Lactation Labeling Rule (also called the “final rule”), issued in December 2014. For each medication, there will be subsections in the labeling:

• Pregnancy
• Lactation
• Females and Males of Reproductive Potential.

In addition, FDA instructions now state that labeling:

• must be updated when new information becomes available
• needs to include evaluation of human data that becomes available mainly after the drug is approved
• needs to include information about the background rates of adverse events related to reproduction.

Labeling in pregnancy. As an example, the “Pregnancy” section of every label contains 3 subsections, all of great clinical importance. First is information about pregnancy exposure registries, with a listing of scientifically acceptable registries (if a registry is available for that drug) and contact information; this section focuses on the high value of data that are systematically and prospectively collected. The second summarizes risk associated with the drug during pregnancy, based on available human, animal, and pharmacologic data. Third is a discussion of clinical considerations.

Need for appropriate controls. Psychiatric disorders increase the risk of pregnancy complications, and often are associated with variables that might increase the risk of a poor pregnancy outcome. For example, a patient who has a psychiatric disorder might be less likely to seek prenatal care, take a prenatal vitamin, and sleep and eat well; she also might use alcohol, tobacco, or other substances of abuse.

The medical literature on the reproductive safety of psychotropic medications is fraught with confounding variables other than the medications themselves. These include variables that, taken alone, might confer a poorer outcome on the fetus or newborn of a pregnant or lactating woman who has a psychiatric illness (to the extent that she uses psychotropics during a pregnancy), compared with what would be seen in (1) a healthy woman who is not taking such medication or (2) the general population.

On the new labels, detailed statements on human data include information from clinical trials, pregnancy exposure registries, and epidemiologic studies. Labels are also to include:

• incidence of adverse events
• effect of dosage
• effect of duration of exposure
• effect of gestational timing of exposure.

The labels emphasize quantifying risk relative to the risk of the same outcome in infants born to women who have not been exposed to the particular drug, but who have the disease or condition for which the drug is indicated (ie, appropriate controls).

Clinical considerations are to include information on the following related to the specific medication (when that information is known):

• more information for prescribers, to further risk-benefit counseling
• disease-associated maternal-fetal risks
• dosage adjustments during pregnancy and postpartum
• maternal adverse reactions
• fetal and neonatal adverse reactions
• labor and delivery.

Clearly, this overdue shift in providing information regarding reproductive safety has the potential to inform clinicians and patients in a meaningful way about the risks and benefits of specific treatments during pregnancy and lactation. Translating that information into practice is daunting, however.

Important aspects of implementation
Pregnancy exposure registries will play a crucial role. For most medications, no systematic registry has been established; to do so, rigorous methodology is required to acquire prospective data and account for confounding variables.⁴ Appropriate control groups also are required to yield data that are useful and interpretable. Primary outcomes require verification, such as review of medical records. Last, registries must be well-conducted and therefore adequately funded, yet labeling changes have not been accompanied by funding requirements set forth by regulators to pharmaceutical manufacturers.

Labeling must be updated continually. Furthermore, it is unclear who will review data for precision and comprehensiveness.

Data need to be understandable to health care providers across disciplines and to patients with varying levels of education for the label to have a meaningful impact on clinical care.

As noted, there is no mandate for funding the meticulous pharmacovigilance required to provide definitive data for labeling. It is unclear if the potential benefits of the new labeling can be reaped without adequate financing of the pharmacovigilance mechanisms required to inform patients adequately.

Role of pregnancy registries
Over the past 2 decades, pregnancy registries have emerged as a rapid, systematic means of collecting important reproductive safety data on the risk for major malformations after prenatal exposure to a medication or a class of medications.^5,6 Such registries enhance the rigor of available cohort studies and other analyses of reproductive safety data that have been derived from large administrative databases.

NPRAA and NPRAD. Recently, the National Pregnancy Registry for Atypical Antipsychotics (NPRAA) and the National Pregnancy Registry for Antidepressants (NPRAD) were established in an effort to obtain reproductive safety data about fetal exposure to second-generation antipsychotics (SGAs) and to newer antidepressants.⁷ Based at Massachusetts General Hospital in Boston, NPRAA and NPRAD systematically and prospectively evaluate the risk of malformations among infants who have been exposed in utero to an SGA or an antidepressant.

The structure of both registries are the same, modeled after the North American Antiepileptic Drug Registry.^5,8 Data are collected prospectively from pregnant women, age 18 to 45, by means of 3 telephone interviews conducted proximate to enrollment, at 7 months’ gestation, and at 2 or 3 months’ postpartum.

Participants include (1) pregnant women who have a history of fetal exposure to an SGA or an antidepressant, or both, and (2) a comparison group of non-exposed pregnant women who have a history of a psychiatric illness. Authorization for release of medical records is obtained for obstetric care, labor and delivery, and neonatal care (≤6 months of age).

Information on the presence of major malformations is abstracted from the medical record, along with other data on neonatal and maternal health outcomes. Identified cases of a congenital malformation are sent to a dysmorphologist, who has been blinded to drug exposure, for final adjudication. Release of findings is dictated by a governing Scientific Advisory Board.

Results so far. Results are available from the NPRAA.⁹ As of December 2014, 487 women were enrolled: 353 who used an SGA and 134 comparison women. Medical records were obtained for 82.2% of participants. A total of 303 women completed the study and were eligible for inclusion in the analysis. Findings include:

• Of 214 live births with first-trimester exposure to an SGA, 3 major malformations were confirmed. In the control group (n = 89), 1 major malformation was confirmed
• The absolute risk of a major malformation was 1.4% for an exposed infant and 1.1% for an unexposed infant
• The odds ratio for a major malformation, comparing exposed infants with unexposed infants, was 1.25 (95% CI, 0.13–12.19).

It is reasonable, therefore, to conclude that, as a class, SGAs are not major teratogens. Although the confidence intervals around the odds ratio estimate remain wide, with the probability for change over the course of the study, it is unlikely that risk will rise to the level of known major teratogens, such as valproate and thalidomide.^10,11

Help with decision-making
Given recent FDA guidance about the importance of pregnancy registries (www.fda.gov/pregnancyregistries), such carefully collected data might help clinicians and patients make informed choices about treatment. Future efforts of NPRAA and NPRAD will focus on sustaining growth in enrollment of participants so that the reproductive safety of SGAs and newer antidepressants can be delineated more clearly.

Last, you can refer potential participants to NPRAA and NPRAD by calling 1-866-961-2388. More information is available at www.womensmentalhealth.org.

In December 2014, the FDA issued draft guidance for sweeping changes to labeling of pharmaceutical treatments in regard to pregnancy and lactation information. These changes are now in effect for use in practice.¹ The undertaking has been years in the making, and is truly ambitious.

The outdated system of letter categories (A, B, C, D, X) falls short of clinical needs in several ways:

• the quality and volume of data can be lacking
• comparative risk is not described
• using letters can led to oversimplification or, in some cases, exaggeration of risk and safety (Box).

Other drawbacks include infrequent updating of information and omission of information about baseline rates of reproductive-related adverse events, to provide a more meaningful context for risk assessment.

A note before we continue discussion of labeling: Recognize that pregnancy itself is inherently risky; poor outcomes are, regrettably, not uncommon. The rate of birth defects in the United States is approximately 3%, and obstetric complications, such as prematurity, are common.^2,3

New system described
The new labeling content has been described in the FDA’s Pregnancy and Lactation Labeling Rule (also called the “final rule”), issued in December 2014. For each medication, there will be subsections in the labeling:

• Pregnancy
• Lactation
• Females and Males of Reproductive Potential.

In addition, FDA instructions now state that labeling:

• must be updated when new information becomes available
• needs to include evaluation of human data that becomes available mainly after the drug is approved
• needs to include information about the background rates of adverse events related to reproduction.

Labeling in pregnancy. As an example, the “Pregnancy” section of every label contains 3 subsections, all of great clinical importance. First is information about pregnancy exposure registries, with a listing of scientifically acceptable registries (if a registry is available for that drug) and contact information; this section focuses on the high value of data that are systematically and prospectively collected. The second summarizes risk associated with the drug during pregnancy, based on available human, animal, and pharmacologic data. Third is a discussion of clinical considerations.

Need for appropriate controls. Psychiatric disorders increase the risk of pregnancy complications, and often are associated with variables that might increase the risk of a poor pregnancy outcome. For example, a patient who has a psychiatric disorder might be less likely to seek prenatal care, take a prenatal vitamin, and sleep and eat well; she also might use alcohol, tobacco, or other substances of abuse.

The medical literature on the reproductive safety of psychotropic medications is fraught with confounding variables other than the medications themselves. These include variables that, taken alone, might confer a poorer outcome on the fetus or newborn of a pregnant or lactating woman who has a psychiatric illness (to the extent that she uses psychotropics during a pregnancy), compared with what would be seen in (1) a healthy woman who is not taking such medication or (2) the general population.

On the new labels, detailed statements on human data include information from clinical trials, pregnancy exposure registries, and epidemiologic studies. Labels are also to include:

• incidence of adverse events
• effect of dosage
• effect of duration of exposure
• effect of gestational timing of exposure.

The labels emphasize quantifying risk relative to the risk of the same outcome in infants born to women who have not been exposed to the particular drug, but who have the disease or condition for which the drug is indicated (ie, appropriate controls).

Clinical considerations are to include information on the following related to the specific medication (when that information is known):

• more information for prescribers, to further risk-benefit counseling
• disease-associated maternal-fetal risks
• dosage adjustments during pregnancy and postpartum
• maternal adverse reactions
• fetal and neonatal adverse reactions
• labor and delivery.

Clearly, this overdue shift in providing information regarding reproductive safety has the potential to inform clinicians and patients in a meaningful way about the risks and benefits of specific treatments during pregnancy and lactation. Translating that information into practice is daunting, however.

Important aspects of implementation
Pregnancy exposure registries will play a crucial role. For most medications, no systematic registry has been established; to do so, rigorous methodology is required to acquire prospective data and account for confounding variables.⁴ Appropriate control groups also are required to yield data that are useful and interpretable. Primary outcomes require verification, such as review of medical records. Last, registries must be well-conducted and therefore adequately funded, yet labeling changes have not been accompanied by funding requirements set forth by regulators to pharmaceutical manufacturers.

Labeling must be updated continually. Furthermore, it is unclear who will review data for precision and comprehensiveness.

Data need to be understandable to health care providers across disciplines and to patients with varying levels of education for the label to have a meaningful impact on clinical care.

As noted, there is no mandate for funding the meticulous pharmacovigilance required to provide definitive data for labeling. It is unclear if the potential benefits of the new labeling can be reaped without adequate financing of the pharmacovigilance mechanisms required to inform patients adequately.

Role of pregnancy registries
Over the past 2 decades, pregnancy registries have emerged as a rapid, systematic means of collecting important reproductive safety data on the risk for major malformations after prenatal exposure to a medication or a class of medications.^5,6 Such registries enhance the rigor of available cohort studies and other analyses of reproductive safety data that have been derived from large administrative databases.

NPRAA and NPRAD. Recently, the National Pregnancy Registry for Atypical Antipsychotics (NPRAA) and the National Pregnancy Registry for Antidepressants (NPRAD) were established in an effort to obtain reproductive safety data about fetal exposure to second-generation antipsychotics (SGAs) and to newer antidepressants.⁷ Based at Massachusetts General Hospital in Boston, NPRAA and NPRAD systematically and prospectively evaluate the risk of malformations among infants who have been exposed in utero to an SGA or an antidepressant.

The structure of both registries are the same, modeled after the North American Antiepileptic Drug Registry.^5,8 Data are collected prospectively from pregnant women, age 18 to 45, by means of 3 telephone interviews conducted proximate to enrollment, at 7 months’ gestation, and at 2 or 3 months’ postpartum.

Participants include (1) pregnant women who have a history of fetal exposure to an SGA or an antidepressant, or both, and (2) a comparison group of non-exposed pregnant women who have a history of a psychiatric illness. Authorization for release of medical records is obtained for obstetric care, labor and delivery, and neonatal care (≤6 months of age).

Information on the presence of major malformations is abstracted from the medical record, along with other data on neonatal and maternal health outcomes. Identified cases of a congenital malformation are sent to a dysmorphologist, who has been blinded to drug exposure, for final adjudication. Release of findings is dictated by a governing Scientific Advisory Board.

Results so far. Results are available from the NPRAA.⁹ As of December 2014, 487 women were enrolled: 353 who used an SGA and 134 comparison women. Medical records were obtained for 82.2% of participants. A total of 303 women completed the study and were eligible for inclusion in the analysis. Findings include:

• Of 214 live births with first-trimester exposure to an SGA, 3 major malformations were confirmed. In the control group (n = 89), 1 major malformation was confirmed
• The absolute risk of a major malformation was 1.4% for an exposed infant and 1.1% for an unexposed infant
• The odds ratio for a major malformation, comparing exposed infants with unexposed infants, was 1.25 (95% CI, 0.13–12.19).

It is reasonable, therefore, to conclude that, as a class, SGAs are not major teratogens. Although the confidence intervals around the odds ratio estimate remain wide, with the probability for change over the course of the study, it is unlikely that risk will rise to the level of known major teratogens, such as valproate and thalidomide.^10,11

Help with decision-making
Given recent FDA guidance about the importance of pregnancy registries (www.fda.gov/pregnancyregistries), such carefully collected data might help clinicians and patients make informed choices about treatment. Future efforts of NPRAA and NPRAD will focus on sustaining growth in enrollment of participants so that the reproductive safety of SGAs and newer antidepressants can be delineated more clearly.

Last, you can refer potential participants to NPRAA and NPRAD by calling 1-866-961-2388. More information is available at www.womensmentalhealth.org.

In December 2014, the FDA issued draft guidance for sweeping changes to labeling of pharmaceutical treatments in regard to pregnancy and lactation information. These changes are now in effect for use in practice.¹ The undertaking has been years in the making, and is truly ambitious.

The outdated system of letter categories (A, B, C, D, X) falls short of clinical needs in several ways:

• the quality and volume of data can be lacking
• comparative risk is not described
• using letters can led to oversimplification or, in some cases, exaggeration of risk and safety (Box).

Other drawbacks include infrequent updating of information and omission of information about baseline rates of reproductive-related adverse events, to provide a more meaningful context for risk assessment.

A note before we continue discussion of labeling: Recognize that pregnancy itself is inherently risky; poor outcomes are, regrettably, not uncommon. The rate of birth defects in the United States is approximately 3%, and obstetric complications, such as prematurity, are common.^2,3

New system described
The new labeling content has been described in the FDA’s Pregnancy and Lactation Labeling Rule (also called the “final rule”), issued in December 2014. For each medication, there will be subsections in the labeling:

• Pregnancy
• Lactation
• Females and Males of Reproductive Potential.

In addition, FDA instructions now state that labeling:

• must be updated when new information becomes available
• needs to include evaluation of human data that becomes available mainly after the drug is approved
• needs to include information about the background rates of adverse events related to reproduction.

Labeling in pregnancy. As an example, the “Pregnancy” section of every label contains 3 subsections, all of great clinical importance. First is information about pregnancy exposure registries, with a listing of scientifically acceptable registries (if a registry is available for that drug) and contact information; this section focuses on the high value of data that are systematically and prospectively collected. The second summarizes risk associated with the drug during pregnancy, based on available human, animal, and pharmacologic data. Third is a discussion of clinical considerations.

Need for appropriate controls. Psychiatric disorders increase the risk of pregnancy complications, and often are associated with variables that might increase the risk of a poor pregnancy outcome. For example, a patient who has a psychiatric disorder might be less likely to seek prenatal care, take a prenatal vitamin, and sleep and eat well; she also might use alcohol, tobacco, or other substances of abuse.

The medical literature on the reproductive safety of psychotropic medications is fraught with confounding variables other than the medications themselves. These include variables that, taken alone, might confer a poorer outcome on the fetus or newborn of a pregnant or lactating woman who has a psychiatric illness (to the extent that she uses psychotropics during a pregnancy), compared with what would be seen in (1) a healthy woman who is not taking such medication or (2) the general population.

On the new labels, detailed statements on human data include information from clinical trials, pregnancy exposure registries, and epidemiologic studies. Labels are also to include:

• incidence of adverse events
• effect of dosage
• effect of duration of exposure
• effect of gestational timing of exposure.

The labels emphasize quantifying risk relative to the risk of the same outcome in infants born to women who have not been exposed to the particular drug, but who have the disease or condition for which the drug is indicated (ie, appropriate controls).

Clinical considerations are to include information on the following related to the specific medication (when that information is known):

• more information for prescribers, to further risk-benefit counseling
• disease-associated maternal-fetal risks
• dosage adjustments during pregnancy and postpartum
• maternal adverse reactions
• fetal and neonatal adverse reactions
• labor and delivery.

Clearly, this overdue shift in providing information regarding reproductive safety has the potential to inform clinicians and patients in a meaningful way about the risks and benefits of specific treatments during pregnancy and lactation. Translating that information into practice is daunting, however.

Important aspects of implementation
Pregnancy exposure registries will play a crucial role. For most medications, no systematic registry has been established; to do so, rigorous methodology is required to acquire prospective data and account for confounding variables.⁴ Appropriate control groups also are required to yield data that are useful and interpretable. Primary outcomes require verification, such as review of medical records. Last, registries must be well-conducted and therefore adequately funded, yet labeling changes have not been accompanied by funding requirements set forth by regulators to pharmaceutical manufacturers.

Labeling must be updated continually. Furthermore, it is unclear who will review data for precision and comprehensiveness.

Data need to be understandable to health care providers across disciplines and to patients with varying levels of education for the label to have a meaningful impact on clinical care.

As noted, there is no mandate for funding the meticulous pharmacovigilance required to provide definitive data for labeling. It is unclear if the potential benefits of the new labeling can be reaped without adequate financing of the pharmacovigilance mechanisms required to inform patients adequately.

Role of pregnancy registries
Over the past 2 decades, pregnancy registries have emerged as a rapid, systematic means of collecting important reproductive safety data on the risk for major malformations after prenatal exposure to a medication or a class of medications.^5,6 Such registries enhance the rigor of available cohort studies and other analyses of reproductive safety data that have been derived from large administrative databases.

NPRAA and NPRAD. Recently, the National Pregnancy Registry for Atypical Antipsychotics (NPRAA) and the National Pregnancy Registry for Antidepressants (NPRAD) were established in an effort to obtain reproductive safety data about fetal exposure to second-generation antipsychotics (SGAs) and to newer antidepressants.⁷ Based at Massachusetts General Hospital in Boston, NPRAA and NPRAD systematically and prospectively evaluate the risk of malformations among infants who have been exposed in utero to an SGA or an antidepressant.

The structure of both registries are the same, modeled after the North American Antiepileptic Drug Registry.^5,8 Data are collected prospectively from pregnant women, age 18 to 45, by means of 3 telephone interviews conducted proximate to enrollment, at 7 months’ gestation, and at 2 or 3 months’ postpartum.

Participants include (1) pregnant women who have a history of fetal exposure to an SGA or an antidepressant, or both, and (2) a comparison group of non-exposed pregnant women who have a history of a psychiatric illness. Authorization for release of medical records is obtained for obstetric care, labor and delivery, and neonatal care (≤6 months of age).

Information on the presence of major malformations is abstracted from the medical record, along with other data on neonatal and maternal health outcomes. Identified cases of a congenital malformation are sent to a dysmorphologist, who has been blinded to drug exposure, for final adjudication. Release of findings is dictated by a governing Scientific Advisory Board.

Results so far. Results are available from the NPRAA.⁹ As of December 2014, 487 women were enrolled: 353 who used an SGA and 134 comparison women. Medical records were obtained for 82.2% of participants. A total of 303 women completed the study and were eligible for inclusion in the analysis. Findings include:

• Of 214 live births with first-trimester exposure to an SGA, 3 major malformations were confirmed. In the control group (n = 89), 1 major malformation was confirmed
• The absolute risk of a major malformation was 1.4% for an exposed infant and 1.1% for an unexposed infant
• The odds ratio for a major malformation, comparing exposed infants with unexposed infants, was 1.25 (95% CI, 0.13–12.19).

It is reasonable, therefore, to conclude that, as a class, SGAs are not major teratogens. Although the confidence intervals around the odds ratio estimate remain wide, with the probability for change over the course of the study, it is unlikely that risk will rise to the level of known major teratogens, such as valproate and thalidomide.^10,11

Help with decision-making
Given recent FDA guidance about the importance of pregnancy registries (www.fda.gov/pregnancyregistries), such carefully collected data might help clinicians and patients make informed choices about treatment. Future efforts of NPRAA and NPRAD will focus on sustaining growth in enrollment of participants so that the reproductive safety of SGAs and newer antidepressants can be delineated more clearly.

Last, you can refer potential participants to NPRAA and NPRAD by calling 1-866-961-2388. More information is available at www.womensmentalhealth.org.

When a patient who has a preexisting medical illness seeks prenatal care, the obstetrician asks herself (himself) 2 questions:
   • What impact will the illness have on the pregnancy?
   • What impact will the pregnancy have on the illness?

Depression is both a pregnancy-associated and pregnancy­-independent illness, which, in the setting of a pregnant woman who has a depressive disorder, makes these questions particu­larly difficult to answer. In such a case, coordination of care with a mental health provider is essential.

Awareness of the obstetrical complications associated with depression during pregnancy, as well as their implications for the future health of the mother–infant dyad, is important for the entire care team. This article reviews the associations and interconnectedness of depression with complications of preg­nancy, childbirth, and the neonatal period.

Diagnosis of depression during prenatal care
The American College of Obstetricians and Gynecologists (ACOG) states that evidence is insufficient to support a rec­ommendation for universal screening for depression among prenatal patients, although such screening should be considered.¹ There is considerable variability among obstetrical pro­viders regarding the practice of depression screening; tools to be used if such screening is done; and screening frequency through the pregnancy.

Discernment of depression is difficult. Many somatic symptoms of depression overlap with common prenatal complaints and, consequentially, can be overlooked. Among a sample of 700 pregnant women, for example, 56% complained of lack of energy; 19%, of insomnia; and 19%, of appe­tite changes.² Weight change, of course, is universal.

The 10-question self-rating Edinburgh Postnatal Depression Scale has been vali­dated for use during pregnancy and post­natally. This screening instrument can be helpful for differentiating purely physical complaints from mental distress due to depressive symptoms.^2,3

When an obstetrical provider suspects a depressive disorder, or one has been diagnosed, she (he) faces the problem of what to do with that information. Women of low socioeconomic status and victims of domestic violence are at increased risk of depression during pregnancy, but barri­ers to appropriate referral can seem nearly insurmountable because they lack insur­ance and social support.^4-9

In addition, within the setting of numer­ous tasks that need attending during the relatively short prenatal period, it is com­mon for women newly given a diagnosis of depression to fail to follow up on a refer­ral to a mental health provider.

Although most providers will “check in” with a depressed or at-risk patient at each prenatal visit about her mood, any effort at follow-up can be overshadowed by tangible physical concerns, such as pre­term contractions, fetal growth restriction, and coordination of routine testing that has been delayed because of scant prena­tal care. All these physical concerns and circumstances of care are associated with maternal depression, as we will discuss.

Preterm labor and birth
Preterm labor is defined as uterine contrac­tions that lead to cervical change before 37 weeks gestational age. Preterm labor increases the risk of preterm birth; pre­term labor precedes 50% of preterm births. Preterm birth is the leading cause of neona­tal mortality in the United States, and rates of morbidity and mortality increase as ges­tational age decreases.¹⁰ Common neonatal complications related to prematurity are shown in the Figure.¹¹

Women who suffer from depression have an increased risk of preterm labor and preterm birth, as many studies of treated and untreated depressed pregnant women have shown.^12-20 The causative mechanism is unknown; it has been proposed that the increase in maternal cortisol production asso­ciated with depression and distress triggers overproduction of placental cortisol releasing hormone, which is thought to be involved in initiation of parturition.^21,22 Depression also is associated with other risk factors for preterm birth, such as low socioeconomic status, sub­stance use, and smoking.

Intrauterine growth restriction
Women who have depression during preg­nancy have an increased risk of intrauterine growth restriction (IUGR), which leads to delivery of an infant who is small for ges­tational age (SGA) or of low birth weight (LBW) (weighing <2,500 g at birth), or both.²³ Again, the basis of the association between depression and IUGR and SGA is unknown; it is theorized that increased levels of cortisol and catecholamines associated with maternal distress might, by increasing blood pressure and inducing vasoconstriction, cause placen­tal hypoperfusion.^24,25

It also is possible that the association of depression with other risk factors for IUGR, such as smoking, substance use, obesity, and poor prenatal care, puts the infants of depressed women at risk of growth restric­tion.²⁶ Several large-scale studies showed that the association between LBW and depres­sion is lost when smoking and substance use are accounted for; other studies, however, found a persistent association in untreated depressed women when smokers, substance users, and drinkers were excluded.^17,26,27

IUGR infants are at increased risk of iat­rogenic prematurity and stillbirth. Fetuses that weigh <10th percentile for their ges­tational age are delivered no later than 40 weeks; delivery can be indicated as early as 32 weeks, depending on the results of other antenatal tests. Women who have a growth-restricted infant have a higher risk of cesarean delivery because growth-restricted infants often have less reserve and poorer tolerance of labor.

Preeclampsia and eclampsia
Preeclampsia is defined as blood pressure >140/90 mm HG on at least 2 occasions, with proteinuria, that occurs later than the twentieth week of pregnancy in women who did not have hypertension or renal dysfunction at baseline. Preeclampsia is a progressive disease that can cause severe maternal morbidity, including renal failure, stroke, hepatic rupture, pulmonary edema, and heart failure.

Eclampsia refers to onset of seizures in the setting of preeclampsia. These 2 hyperten­sive disorders are the third leading world wide cause of maternal mortality.²⁸

Depressed women have an elevated risk of preeclampsia. The association between preeclampsia and depression might be caused by the presence of increased levels of inflammatory mediators^29,30; other comorbidities, such as increased body mass index, also might be involved, but the risk for preeclampsia in depressed women still is increased after controlling for obesity.³¹

The presence of preeclampsia is respon­sible for a high percentage of iatrogenic pre­term births, because the cure for the disorder is delivery—even at early or previable gesta­tional age. Complication rates for mother and infant are high.

The presence of preeclampsia is a sig­nificant risk factor for intrauterine fetal demise. Treating the mother after delivery involves administration of IV magnesium for 24 hours; often, the mother is separated from her infant for a day after birth.

Impact on prenatal care
Depression increases odds that women will have fewer prenatal visits.³² During pregnancy, women typically initiate pre­natal care during the first trimester, when pregnancy-dating ultrasonography and early screening tests for chromosomal abnormalities are performed. Prenatal vis­its occur monthly until the third trimes­ter, then every 2 weeks between 32 and 36 weeks’ gestation, increasing to weekly after 36 weeks’ gestation.

The increased number of visits in late pregnancy allows for early detection and treatment of hypertensive disorders; assesses fetal well-being; and decreases the risks of morbidity and mortality for mother and fetus.³³ Because women who suffer from depression are at increased risk of an array of adverse pregnancy outcomes, the impor­tance of regular and timely prenatal care cannot be understated.

In addition, the prenatal visit gives the obstetrician the opportunity to connect women with other specialists for manage­ment of any unmet medical needs. One study showed that, when women have adequate prenatal care (measured by the number of visits), the association between preterm birth and self-reported maternal depression was eliminated.³⁴

Substance use
Substance use and depression often co-exist.^35,36 Unlike screening for depression, screening for substance use is universal dur­ing prenatal care. Studies have shown that women who screen positive for depression are at higher risk of a number of comorbidi­ties, including substance use.^37,38 Conversely, women who use substances are more likely to screen positive for depression.

Evidence suggests that best practice might be to screen for depression in any woman who has a positive drug screen, if a provider is not routinely screening their general patient population.³⁹ Substance use in pregnancy is associated with a number of poor outcomes, including placental abrup­tion (cocaine use); dysmorphic facies and congenital anomalies (alcohol); and neonatal abstinence syndrome (heroin).

Antidepressants in pregnancy
A full discussion of the risks and ben­efits associated with pharmacotherapy for depression in pregnancy is beyond the scope of this article. Generally, antidepressant use is fraught with concerns over teratogenic­ity and adverse fetal outcomes. Although ACOG states that (1) pharmacotherapy for depression should be individualized and (2) most selective serotonin reuptake inhibi­tors (SSRIs) are not considered major terato­genic agents, many obstetricians and patients feel uncomfortable using these medications in pregnancy.⁴⁰ Often, pre-pregnancy antide­pressants are discontinued in the first trimes­ter; one large population-based study found that only 0.9% of women who had depres­sion filled their antidepressant prescription consistently throughout their pregnancy.⁴¹

It is unclear whether antidepressant use in pregnancy contributes to the risk of preterm birth seen in women who have depression. In a large population-based study, use of antidepressants in the second trimester was associated with preterm delivery but severe depression was not.¹⁸ A recent meta-analysis revealed an increased risk of preterm birth in women who used an antidepressant, com­pared with healthy women and untreated depressed women.⁴²

Research limits, unanswered questions. Regrettably, it is difficult to untangle risk factors for preterm birth among depressed women without randomized controlled studies that are not ethically feasible. It can­not be said with certainty whether antide­pressant pharmacotherapy is associated with a higher risk of preterm birth than depression alone.

Likewise, it is difficult to clarify the extent to which antidepressants contribute to infant growth restriction, if at all. Two recent meta-analyses concluded that exposure to antidepressants is associated with a statisti­cally significant risk of LBW.^42,43 However, increased severity of depressive symptoms generally is associated with exposure to anti­depressants during pregnancy, and a ran­domized controlled trial is, again, impossible to conduct for ethical reasons.

Whereas a plausible biological mecha­nism associating IUGR, SGA, and LBW with depression exists, the same cannot be said for antidepressants. In one study, exposure to maternal depression altered the expres­sion of certain placental genes but exposure to SSRIs did not cause further changes. This suggests that, on a cellular level, placental function might differ in depressed women.⁴⁴ Although antidepressants do cross the pla­centa, it remains to be seen whether fetal growth is impacted as a result. One study found decreased fetal head circumference in infants who had been exposed to antidepres­sants during pregnancy, but no increased risk for having a SGA or LWB infant.⁴⁵

Obstetrical management and mental health implications
Treated or not, women who suffer depres­sion are a high-risk group when it comes to preterm birth and a host of other pregnancy comorbidities. Women with serious compli­cations of pregnancy often are hospitalized for observation, and can undergo a pro­longed stay when close proximity to medical services or a surgical suite is required.

For example, hospitalization until deliv­ery is the standard of care for women who have preterm premature rupture of mem­branes or preeclampsia before 34 weeks’ gestation. Prolonged inpatient admis­sions and associated restriction of activity is profoundly deleterious on mood, with depression and anxiety significantly cor­related with length of stay.^46,47 Given the associations between depression and pre­term birth, it might be reasonable to con­sider screening antenatal inpatients at risk of preterm birth for depression on a regu­lar basis, so that treatment can be initiated if needed.

Depression during pregnancy is rela­tively common; an estimated 12.7% of pregnant women are affected at some time between conception and birth.⁴⁸ Not only does depression appear to have deleterious effects on pregnancy outcomes, it also plays a pivotal role in the qualitative experience of pregnancy for the mother.

Bottom Line
Awareness of obstetrical complications associated with depression in pregnancy is important for the entire care team, including the psychiatrist and obstetrician. Depression not only appears to have deleterious effects on pregnancy outcomes, it also plays a pivotal role in the qualitative experience of pregnancy for the mother. Antidepressant use generally is fraught with concerns over teratogenicity and adverse fetal outcomes.

Related Resources
• Freeman MP. Some SSRIs are better than others for pregnant women (audio interview). Current Psychiatry. 2014;13(7). http://www.currentpsychiatry.com/specialty-focus/practice-trends/article/some-ssris-are-better-thanothers-for-pregnant-women/e3adb4704e25492f3e15331fc1cc058d.html.
• Freeman MP, Joffe H, Cohen LS. Postpartum depression: Help patients find the right treatment. Current Psychiatry. 2012;11(11):14-16,19-21.

Disclosures
Dr. Habecker reports no financial relationships with any company whose products are mentioned in this article or with manufacturers of competing products.

Dr. Freeman is a member of the advisory board of JDS Therapeutics, Sunovion Pharmaceuticals, Inc., and Takeda Pharmaceutical Co. She receives research grant support from Takeda Pharmaceutical Co.

When a patient who has a preexisting medical illness seeks prenatal care, the obstetrician asks herself (himself) 2 questions:
   • What impact will the illness have on the pregnancy?
   • What impact will the pregnancy have on the illness?

Depression is both a pregnancy-associated and pregnancy­-independent illness, which, in the setting of a pregnant woman who has a depressive disorder, makes these questions particu­larly difficult to answer. In such a case, coordination of care with a mental health provider is essential.

Awareness of the obstetrical complications associated with depression during pregnancy, as well as their implications for the future health of the mother–infant dyad, is important for the entire care team. This article reviews the associations and interconnectedness of depression with complications of preg­nancy, childbirth, and the neonatal period.

Diagnosis of depression during prenatal care
The American College of Obstetricians and Gynecologists (ACOG) states that evidence is insufficient to support a rec­ommendation for universal screening for depression among prenatal patients, although such screening should be considered.¹ There is considerable variability among obstetrical pro­viders regarding the practice of depression screening; tools to be used if such screening is done; and screening frequency through the pregnancy.

Discernment of depression is difficult. Many somatic symptoms of depression overlap with common prenatal complaints and, consequentially, can be overlooked. Among a sample of 700 pregnant women, for example, 56% complained of lack of energy; 19%, of insomnia; and 19%, of appe­tite changes.² Weight change, of course, is universal.

The 10-question self-rating Edinburgh Postnatal Depression Scale has been vali­dated for use during pregnancy and post­natally. This screening instrument can be helpful for differentiating purely physical complaints from mental distress due to depressive symptoms.^2,3

When an obstetrical provider suspects a depressive disorder, or one has been diagnosed, she (he) faces the problem of what to do with that information. Women of low socioeconomic status and victims of domestic violence are at increased risk of depression during pregnancy, but barri­ers to appropriate referral can seem nearly insurmountable because they lack insur­ance and social support.^4-9

In addition, within the setting of numer­ous tasks that need attending during the relatively short prenatal period, it is com­mon for women newly given a diagnosis of depression to fail to follow up on a refer­ral to a mental health provider.

Although most providers will “check in” with a depressed or at-risk patient at each prenatal visit about her mood, any effort at follow-up can be overshadowed by tangible physical concerns, such as pre­term contractions, fetal growth restriction, and coordination of routine testing that has been delayed because of scant prena­tal care. All these physical concerns and circumstances of care are associated with maternal depression, as we will discuss.

Preterm labor and birth
Preterm labor is defined as uterine contrac­tions that lead to cervical change before 37 weeks gestational age. Preterm labor increases the risk of preterm birth; pre­term labor precedes 50% of preterm births. Preterm birth is the leading cause of neona­tal mortality in the United States, and rates of morbidity and mortality increase as ges­tational age decreases.¹⁰ Common neonatal complications related to prematurity are shown in the Figure.¹¹

Women who suffer from depression have an increased risk of preterm labor and preterm birth, as many studies of treated and untreated depressed pregnant women have shown.^12-20 The causative mechanism is unknown; it has been proposed that the increase in maternal cortisol production asso­ciated with depression and distress triggers overproduction of placental cortisol releasing hormone, which is thought to be involved in initiation of parturition.^21,22 Depression also is associated with other risk factors for preterm birth, such as low socioeconomic status, sub­stance use, and smoking.

Intrauterine growth restriction
Women who have depression during preg­nancy have an increased risk of intrauterine growth restriction (IUGR), which leads to delivery of an infant who is small for ges­tational age (SGA) or of low birth weight (LBW) (weighing <2,500 g at birth), or both.²³ Again, the basis of the association between depression and IUGR and SGA is unknown; it is theorized that increased levels of cortisol and catecholamines associated with maternal distress might, by increasing blood pressure and inducing vasoconstriction, cause placen­tal hypoperfusion.^24,25

It also is possible that the association of depression with other risk factors for IUGR, such as smoking, substance use, obesity, and poor prenatal care, puts the infants of depressed women at risk of growth restric­tion.²⁶ Several large-scale studies showed that the association between LBW and depres­sion is lost when smoking and substance use are accounted for; other studies, however, found a persistent association in untreated depressed women when smokers, substance users, and drinkers were excluded.^17,26,27

IUGR infants are at increased risk of iat­rogenic prematurity and stillbirth. Fetuses that weigh <10th percentile for their ges­tational age are delivered no later than 40 weeks; delivery can be indicated as early as 32 weeks, depending on the results of other antenatal tests. Women who have a growth-restricted infant have a higher risk of cesarean delivery because growth-restricted infants often have less reserve and poorer tolerance of labor.

Preeclampsia and eclampsia
Preeclampsia is defined as blood pressure >140/90 mm HG on at least 2 occasions, with proteinuria, that occurs later than the twentieth week of pregnancy in women who did not have hypertension or renal dysfunction at baseline. Preeclampsia is a progressive disease that can cause severe maternal morbidity, including renal failure, stroke, hepatic rupture, pulmonary edema, and heart failure.

Eclampsia refers to onset of seizures in the setting of preeclampsia. These 2 hyperten­sive disorders are the third leading world wide cause of maternal mortality.²⁸

Depressed women have an elevated risk of preeclampsia. The association between preeclampsia and depression might be caused by the presence of increased levels of inflammatory mediators^29,30; other comorbidities, such as increased body mass index, also might be involved, but the risk for preeclampsia in depressed women still is increased after controlling for obesity.³¹

The presence of preeclampsia is respon­sible for a high percentage of iatrogenic pre­term births, because the cure for the disorder is delivery—even at early or previable gesta­tional age. Complication rates for mother and infant are high.

The presence of preeclampsia is a sig­nificant risk factor for intrauterine fetal demise. Treating the mother after delivery involves administration of IV magnesium for 24 hours; often, the mother is separated from her infant for a day after birth.

Impact on prenatal care
Depression increases odds that women will have fewer prenatal visits.³² During pregnancy, women typically initiate pre­natal care during the first trimester, when pregnancy-dating ultrasonography and early screening tests for chromosomal abnormalities are performed. Prenatal vis­its occur monthly until the third trimes­ter, then every 2 weeks between 32 and 36 weeks’ gestation, increasing to weekly after 36 weeks’ gestation.

The increased number of visits in late pregnancy allows for early detection and treatment of hypertensive disorders; assesses fetal well-being; and decreases the risks of morbidity and mortality for mother and fetus.³³ Because women who suffer from depression are at increased risk of an array of adverse pregnancy outcomes, the impor­tance of regular and timely prenatal care cannot be understated.

In addition, the prenatal visit gives the obstetrician the opportunity to connect women with other specialists for manage­ment of any unmet medical needs. One study showed that, when women have adequate prenatal care (measured by the number of visits), the association between preterm birth and self-reported maternal depression was eliminated.³⁴

Substance use
Substance use and depression often co-exist.^35,36 Unlike screening for depression, screening for substance use is universal dur­ing prenatal care. Studies have shown that women who screen positive for depression are at higher risk of a number of comorbidi­ties, including substance use.^37,38 Conversely, women who use substances are more likely to screen positive for depression.

Evidence suggests that best practice might be to screen for depression in any woman who has a positive drug screen, if a provider is not routinely screening their general patient population.³⁹ Substance use in pregnancy is associated with a number of poor outcomes, including placental abrup­tion (cocaine use); dysmorphic facies and congenital anomalies (alcohol); and neonatal abstinence syndrome (heroin).

Antidepressants in pregnancy
A full discussion of the risks and ben­efits associated with pharmacotherapy for depression in pregnancy is beyond the scope of this article. Generally, antidepressant use is fraught with concerns over teratogenic­ity and adverse fetal outcomes. Although ACOG states that (1) pharmacotherapy for depression should be individualized and (2) most selective serotonin reuptake inhibi­tors (SSRIs) are not considered major terato­genic agents, many obstetricians and patients feel uncomfortable using these medications in pregnancy.⁴⁰ Often, pre-pregnancy antide­pressants are discontinued in the first trimes­ter; one large population-based study found that only 0.9% of women who had depres­sion filled their antidepressant prescription consistently throughout their pregnancy.⁴¹

It is unclear whether antidepressant use in pregnancy contributes to the risk of preterm birth seen in women who have depression. In a large population-based study, use of antidepressants in the second trimester was associated with preterm delivery but severe depression was not.¹⁸ A recent meta-analysis revealed an increased risk of preterm birth in women who used an antidepressant, com­pared with healthy women and untreated depressed women.⁴²

Research limits, unanswered questions. Regrettably, it is difficult to untangle risk factors for preterm birth among depressed women without randomized controlled studies that are not ethically feasible. It can­not be said with certainty whether antide­pressant pharmacotherapy is associated with a higher risk of preterm birth than depression alone.

Likewise, it is difficult to clarify the extent to which antidepressants contribute to infant growth restriction, if at all. Two recent meta-analyses concluded that exposure to antidepressants is associated with a statisti­cally significant risk of LBW.^42,43 However, increased severity of depressive symptoms generally is associated with exposure to anti­depressants during pregnancy, and a ran­domized controlled trial is, again, impossible to conduct for ethical reasons.

Whereas a plausible biological mecha­nism associating IUGR, SGA, and LBW with depression exists, the same cannot be said for antidepressants. In one study, exposure to maternal depression altered the expres­sion of certain placental genes but exposure to SSRIs did not cause further changes. This suggests that, on a cellular level, placental function might differ in depressed women.⁴⁴ Although antidepressants do cross the pla­centa, it remains to be seen whether fetal growth is impacted as a result. One study found decreased fetal head circumference in infants who had been exposed to antidepres­sants during pregnancy, but no increased risk for having a SGA or LWB infant.⁴⁵

Obstetrical management and mental health implications
Treated or not, women who suffer depres­sion are a high-risk group when it comes to preterm birth and a host of other pregnancy comorbidities. Women with serious compli­cations of pregnancy often are hospitalized for observation, and can undergo a pro­longed stay when close proximity to medical services or a surgical suite is required.

For example, hospitalization until deliv­ery is the standard of care for women who have preterm premature rupture of mem­branes or preeclampsia before 34 weeks’ gestation. Prolonged inpatient admis­sions and associated restriction of activity is profoundly deleterious on mood, with depression and anxiety significantly cor­related with length of stay.^46,47 Given the associations between depression and pre­term birth, it might be reasonable to con­sider screening antenatal inpatients at risk of preterm birth for depression on a regu­lar basis, so that treatment can be initiated if needed.

Depression during pregnancy is rela­tively common; an estimated 12.7% of pregnant women are affected at some time between conception and birth.⁴⁸ Not only does depression appear to have deleterious effects on pregnancy outcomes, it also plays a pivotal role in the qualitative experience of pregnancy for the mother.

Bottom Line
Awareness of obstetrical complications associated with depression in pregnancy is important for the entire care team, including the psychiatrist and obstetrician. Depression not only appears to have deleterious effects on pregnancy outcomes, it also plays a pivotal role in the qualitative experience of pregnancy for the mother. Antidepressant use generally is fraught with concerns over teratogenicity and adverse fetal outcomes.

Related Resources
• Freeman MP. Some SSRIs are better than others for pregnant women (audio interview). Current Psychiatry. 2014;13(7). http://www.currentpsychiatry.com/specialty-focus/practice-trends/article/some-ssris-are-better-thanothers-for-pregnant-women/e3adb4704e25492f3e15331fc1cc058d.html.
• Freeman MP, Joffe H, Cohen LS. Postpartum depression: Help patients find the right treatment. Current Psychiatry. 2012;11(11):14-16,19-21.

Disclosures
Dr. Habecker reports no financial relationships with any company whose products are mentioned in this article or with manufacturers of competing products.

Dr. Freeman is a member of the advisory board of JDS Therapeutics, Sunovion Pharmaceuticals, Inc., and Takeda Pharmaceutical Co. She receives research grant support from Takeda Pharmaceutical Co.

When a patient who has a preexisting medical illness seeks prenatal care, the obstetrician asks herself (himself) 2 questions:
   • What impact will the illness have on the pregnancy?
   • What impact will the pregnancy have on the illness?

Depression is both a pregnancy-associated and pregnancy­-independent illness, which, in the setting of a pregnant woman who has a depressive disorder, makes these questions particu­larly difficult to answer. In such a case, coordination of care with a mental health provider is essential.

Awareness of the obstetrical complications associated with depression during pregnancy, as well as their implications for the future health of the mother–infant dyad, is important for the entire care team. This article reviews the associations and interconnectedness of depression with complications of preg­nancy, childbirth, and the neonatal period.

Diagnosis of depression during prenatal care
The American College of Obstetricians and Gynecologists (ACOG) states that evidence is insufficient to support a rec­ommendation for universal screening for depression among prenatal patients, although such screening should be considered.¹ There is considerable variability among obstetrical pro­viders regarding the practice of depression screening; tools to be used if such screening is done; and screening frequency through the pregnancy.

Discernment of depression is difficult. Many somatic symptoms of depression overlap with common prenatal complaints and, consequentially, can be overlooked. Among a sample of 700 pregnant women, for example, 56% complained of lack of energy; 19%, of insomnia; and 19%, of appe­tite changes.² Weight change, of course, is universal.

The 10-question self-rating Edinburgh Postnatal Depression Scale has been vali­dated for use during pregnancy and post­natally. This screening instrument can be helpful for differentiating purely physical complaints from mental distress due to depressive symptoms.^2,3

When an obstetrical provider suspects a depressive disorder, or one has been diagnosed, she (he) faces the problem of what to do with that information. Women of low socioeconomic status and victims of domestic violence are at increased risk of depression during pregnancy, but barri­ers to appropriate referral can seem nearly insurmountable because they lack insur­ance and social support.^4-9

In addition, within the setting of numer­ous tasks that need attending during the relatively short prenatal period, it is com­mon for women newly given a diagnosis of depression to fail to follow up on a refer­ral to a mental health provider.

Although most providers will “check in” with a depressed or at-risk patient at each prenatal visit about her mood, any effort at follow-up can be overshadowed by tangible physical concerns, such as pre­term contractions, fetal growth restriction, and coordination of routine testing that has been delayed because of scant prena­tal care. All these physical concerns and circumstances of care are associated with maternal depression, as we will discuss.

Preterm labor and birth
Preterm labor is defined as uterine contrac­tions that lead to cervical change before 37 weeks gestational age. Preterm labor increases the risk of preterm birth; pre­term labor precedes 50% of preterm births. Preterm birth is the leading cause of neona­tal mortality in the United States, and rates of morbidity and mortality increase as ges­tational age decreases.¹⁰ Common neonatal complications related to prematurity are shown in the Figure.¹¹

Women who suffer from depression have an increased risk of preterm labor and preterm birth, as many studies of treated and untreated depressed pregnant women have shown.^12-20 The causative mechanism is unknown; it has been proposed that the increase in maternal cortisol production asso­ciated with depression and distress triggers overproduction of placental cortisol releasing hormone, which is thought to be involved in initiation of parturition.^21,22 Depression also is associated with other risk factors for preterm birth, such as low socioeconomic status, sub­stance use, and smoking.

Intrauterine growth restriction
Women who have depression during preg­nancy have an increased risk of intrauterine growth restriction (IUGR), which leads to delivery of an infant who is small for ges­tational age (SGA) or of low birth weight (LBW) (weighing <2,500 g at birth), or both.²³ Again, the basis of the association between depression and IUGR and SGA is unknown; it is theorized that increased levels of cortisol and catecholamines associated with maternal distress might, by increasing blood pressure and inducing vasoconstriction, cause placen­tal hypoperfusion.^24,25

It also is possible that the association of depression with other risk factors for IUGR, such as smoking, substance use, obesity, and poor prenatal care, puts the infants of depressed women at risk of growth restric­tion.²⁶ Several large-scale studies showed that the association between LBW and depres­sion is lost when smoking and substance use are accounted for; other studies, however, found a persistent association in untreated depressed women when smokers, substance users, and drinkers were excluded.^17,26,27

IUGR infants are at increased risk of iat­rogenic prematurity and stillbirth. Fetuses that weigh <10th percentile for their ges­tational age are delivered no later than 40 weeks; delivery can be indicated as early as 32 weeks, depending on the results of other antenatal tests. Women who have a growth-restricted infant have a higher risk of cesarean delivery because growth-restricted infants often have less reserve and poorer tolerance of labor.

Preeclampsia and eclampsia
Preeclampsia is defined as blood pressure >140/90 mm HG on at least 2 occasions, with proteinuria, that occurs later than the twentieth week of pregnancy in women who did not have hypertension or renal dysfunction at baseline. Preeclampsia is a progressive disease that can cause severe maternal morbidity, including renal failure, stroke, hepatic rupture, pulmonary edema, and heart failure.

Eclampsia refers to onset of seizures in the setting of preeclampsia. These 2 hyperten­sive disorders are the third leading world wide cause of maternal mortality.²⁸

Depressed women have an elevated risk of preeclampsia. The association between preeclampsia and depression might be caused by the presence of increased levels of inflammatory mediators^29,30; other comorbidities, such as increased body mass index, also might be involved, but the risk for preeclampsia in depressed women still is increased after controlling for obesity.³¹

The presence of preeclampsia is respon­sible for a high percentage of iatrogenic pre­term births, because the cure for the disorder is delivery—even at early or previable gesta­tional age. Complication rates for mother and infant are high.

The presence of preeclampsia is a sig­nificant risk factor for intrauterine fetal demise. Treating the mother after delivery involves administration of IV magnesium for 24 hours; often, the mother is separated from her infant for a day after birth.

Impact on prenatal care
Depression increases odds that women will have fewer prenatal visits.³² During pregnancy, women typically initiate pre­natal care during the first trimester, when pregnancy-dating ultrasonography and early screening tests for chromosomal abnormalities are performed. Prenatal vis­its occur monthly until the third trimes­ter, then every 2 weeks between 32 and 36 weeks’ gestation, increasing to weekly after 36 weeks’ gestation.

The increased number of visits in late pregnancy allows for early detection and treatment of hypertensive disorders; assesses fetal well-being; and decreases the risks of morbidity and mortality for mother and fetus.³³ Because women who suffer from depression are at increased risk of an array of adverse pregnancy outcomes, the impor­tance of regular and timely prenatal care cannot be understated.

In addition, the prenatal visit gives the obstetrician the opportunity to connect women with other specialists for manage­ment of any unmet medical needs. One study showed that, when women have adequate prenatal care (measured by the number of visits), the association between preterm birth and self-reported maternal depression was eliminated.³⁴

Substance use
Substance use and depression often co-exist.^35,36 Unlike screening for depression, screening for substance use is universal dur­ing prenatal care. Studies have shown that women who screen positive for depression are at higher risk of a number of comorbidi­ties, including substance use.^37,38 Conversely, women who use substances are more likely to screen positive for depression.

Evidence suggests that best practice might be to screen for depression in any woman who has a positive drug screen, if a provider is not routinely screening their general patient population.³⁹ Substance use in pregnancy is associated with a number of poor outcomes, including placental abrup­tion (cocaine use); dysmorphic facies and congenital anomalies (alcohol); and neonatal abstinence syndrome (heroin).

Antidepressants in pregnancy
A full discussion of the risks and ben­efits associated with pharmacotherapy for depression in pregnancy is beyond the scope of this article. Generally, antidepressant use is fraught with concerns over teratogenic­ity and adverse fetal outcomes. Although ACOG states that (1) pharmacotherapy for depression should be individualized and (2) most selective serotonin reuptake inhibi­tors (SSRIs) are not considered major terato­genic agents, many obstetricians and patients feel uncomfortable using these medications in pregnancy.⁴⁰ Often, pre-pregnancy antide­pressants are discontinued in the first trimes­ter; one large population-based study found that only 0.9% of women who had depres­sion filled their antidepressant prescription consistently throughout their pregnancy.⁴¹

It is unclear whether antidepressant use in pregnancy contributes to the risk of preterm birth seen in women who have depression. In a large population-based study, use of antidepressants in the second trimester was associated with preterm delivery but severe depression was not.¹⁸ A recent meta-analysis revealed an increased risk of preterm birth in women who used an antidepressant, com­pared with healthy women and untreated depressed women.⁴²

Research limits, unanswered questions. Regrettably, it is difficult to untangle risk factors for preterm birth among depressed women without randomized controlled studies that are not ethically feasible. It can­not be said with certainty whether antide­pressant pharmacotherapy is associated with a higher risk of preterm birth than depression alone.

Likewise, it is difficult to clarify the extent to which antidepressants contribute to infant growth restriction, if at all. Two recent meta-analyses concluded that exposure to antidepressants is associated with a statisti­cally significant risk of LBW.^42,43 However, increased severity of depressive symptoms generally is associated with exposure to anti­depressants during pregnancy, and a ran­domized controlled trial is, again, impossible to conduct for ethical reasons.

Whereas a plausible biological mecha­nism associating IUGR, SGA, and LBW with depression exists, the same cannot be said for antidepressants. In one study, exposure to maternal depression altered the expres­sion of certain placental genes but exposure to SSRIs did not cause further changes. This suggests that, on a cellular level, placental function might differ in depressed women.⁴⁴ Although antidepressants do cross the pla­centa, it remains to be seen whether fetal growth is impacted as a result. One study found decreased fetal head circumference in infants who had been exposed to antidepres­sants during pregnancy, but no increased risk for having a SGA or LWB infant.⁴⁵

Obstetrical management and mental health implications
Treated or not, women who suffer depres­sion are a high-risk group when it comes to preterm birth and a host of other pregnancy comorbidities. Women with serious compli­cations of pregnancy often are hospitalized for observation, and can undergo a pro­longed stay when close proximity to medical services or a surgical suite is required.

For example, hospitalization until deliv­ery is the standard of care for women who have preterm premature rupture of mem­branes or preeclampsia before 34 weeks’ gestation. Prolonged inpatient admis­sions and associated restriction of activity is profoundly deleterious on mood, with depression and anxiety significantly cor­related with length of stay.^46,47 Given the associations between depression and pre­term birth, it might be reasonable to con­sider screening antenatal inpatients at risk of preterm birth for depression on a regu­lar basis, so that treatment can be initiated if needed.

Depression during pregnancy is rela­tively common; an estimated 12.7% of pregnant women are affected at some time between conception and birth.⁴⁸ Not only does depression appear to have deleterious effects on pregnancy outcomes, it also plays a pivotal role in the qualitative experience of pregnancy for the mother.

Bottom Line
Awareness of obstetrical complications associated with depression in pregnancy is important for the entire care team, including the psychiatrist and obstetrician. Depression not only appears to have deleterious effects on pregnancy outcomes, it also plays a pivotal role in the qualitative experience of pregnancy for the mother. Antidepressant use generally is fraught with concerns over teratogenicity and adverse fetal outcomes.

Related Resources
• Freeman MP. Some SSRIs are better than others for pregnant women (audio interview). Current Psychiatry. 2014;13(7). http://www.currentpsychiatry.com/specialty-focus/practice-trends/article/some-ssris-are-better-thanothers-for-pregnant-women/e3adb4704e25492f3e15331fc1cc058d.html.
• Freeman MP, Joffe H, Cohen LS. Postpartum depression: Help patients find the right treatment. Current Psychiatry. 2012;11(11):14-16,19-21.

Disclosures
Dr. Habecker reports no financial relationships with any company whose products are mentioned in this article or with manufacturers of competing products.

Dr. Freeman is a member of the advisory board of JDS Therapeutics, Sunovion Pharmaceuticals, Inc., and Takeda Pharmaceutical Co. She receives research grant support from Takeda Pharmaceutical Co.

Challenges in recruiting women to postpartum depression (PPD) antidepressant treatment trials, which we encountered when conducting a trial of escitalopram, contribute to the limited body of knowledge about PPD treatment. Here we discuss results from a preliminary trial of escitalopram for PPD, and challenges of research in this area.

Escitalopram, the S-enantiomer of citalopram, is a selective serotonin reuptake inhibitor with high selectivity and potency that is FDA-approved for treating major depressive disorder (MDD) and generalized anxiety disorder. An agent with antidepressant and anxiolytic effects is particularly desirable for PPD because anxiety is more common in postpartum major depressive episodes than non-postpartum MDD.¹ Anxiety and depressive disorders commonly are comorbid in postpartum women.²

We conducted an open-label trial of escitalopram for women with PPD and anxiety. We initially attempted to recruit 20 women.

Methods

Patients received 8 weeks of treatment with escitalopram, 10 to 20 mg/d (flexible dose). After completing the initial phone screen, patients had 5 follow-up visits, once every 2 weeks for 8 weeks. The institutional review board at Massachusetts General Hospital approved this study and we obtained written informed consent from all patients at the first visit. Twelve patients completed the phone screen and 7 eligible patients were enrolled in the study over 32 months. Reasons for ineligibility included having a history of psychosis, onset of symptoms >3 months postpartum, or presenting >6 months after onset. Others declined to participate because of concern about the time commitment or because they pursued nonpharmacologic treatments after the evaluation visit. One patient was lost to follow-up. Three patients completed the study. The study was halted because of the slow pace of recruitment.

Patient selection. Patients were screened for a major depressive episode with postpartum onset within 3 months of childbirth; depressive symptoms may have developed during pregnancy and worsened postpartum to meet criteria for MDD. Women were eligible for the study if they:

• were age 18 to 45
• experienced a major depressive episode with symptoms developing within 3 months of childbirth
• presented within 6 months of childbirth
• had a Montgomery-Åsberg Depression Rating Scale (MADRS) score >15
• had a Beck Anxiety Inventory (BAI) score >10.

Patients who were pregnant or breast-feeding were excluded from the study per an agreement with the sponsor. In addition, women were excluded if they had taken any psychotropic medication within 2 weeks of enrollment; had active suicidal ideation, homicidal ideation, or presence of psychotic symptoms; had chronic depression or dysthymia; had chronic or treatment-resistant anxiety disorders; had a history of mania or hypomania; or had active alcohol or substance abuse within the past year.

Treatment. Patients received escitalopram, 10 mg/d, after the baseline visit. At the investigator’s discretion, the dose could be increased to 20 mg/d or lowered to 5 mg/d if side effects occurred.

Measures. At the first visit, patients were assessed with the Mini-International Neuropsychiatric Interview to verify MDD and exclude diagnoses that would determine ineligibility. MADRS and Edinburgh Postnatal Depression Scale (EPDS) were used at each visit to measure depressive symptoms.^3,4 The BAI was completed at each visit to measure anxiety symptoms. Obsessions and compulsions were measured with the Yale-Brown Obsessive Compulsive Scale (Y-BOCS)⁵ at baseline, and at all following visits if the patient scored >8 at baseline. The Clinical Global Impression Scales for severity and improvement were completed at each visit.⁶

Results

Of 7 patients enrolled, 3 completed the study, 2 were ineligible after the baseline visit, and 2 did not participate after the baseline visit (1 selected to pursue psychotherapy, and 1 was lost to follow-up).

Two of 3 patients responded to escitalopram (≥50% decrease on MADRS), and both were remitters (MADRS score <7). All 3 patients were responders on EPDS and BAI. One patient had Y-BOCS >8 at baseline (Total Y-BOCS score of 9, and final Y-BOCS score of 8) (Table).

Table

Symptom rating scale scores at baseline and study end

Discussion

Patients who stayed in treatment improved during the course of this study. Recruitment was difficult; we were able to recruit only 7 patients out of a projected 20 for the screening visit. We solicited feedback from local obstetrics health care providers and social workers on recruitment and attractiveness of the study as part of our routine collaboration with obstetrical services that screen for PPD. Primary reasons patients were not referred were that they were breast-feeding or they stated they would prefer to receive treatment from their primary care doctor. Recruitment difficulty in this study was in stark contrast to other recent studies completed at our center. For example, we have successfully recruited for menopausal depression and premenstrual dysphoric disorder treatment studies, and have completed large naturalistic studies of women with unipolar depression and bipolar disorder across pregnancy and postpartum. We suspect that many patients who were eligible for the study preferred to seek care from an obstetrician or primary care doctor with whom they already had a therapeutic alliance, and we also suspect that many women with PPD do not seek treatment at all, which is consistent with findings from other research groups.

Lessons learned from PPD research include:

• Including women who are breast-feeding is important because many women choose to breast-feed and suffer from PPD. Because antidepressant use during breast-feeding has been closely studied, it is appropriate to include potential research participants who are breast-feeding as long as they receive adequate information and are able to provide informed consent.
• Participants in PPD studies may require accommodations that take into account their role as a new mother, such as on-site childcare, home visits, or other strategies.
• Because of recruitment challenges in postpartum patients, multisite trials may be required to include adequate numbers of participants.

Related Resource

• Freeman MP, Joffe H, Cohen LS. Postpartum depression: Help patients find the right treatment. Current Psychiatry. 2012;11(11):14-21.

Drug Brand Names

• Citalopram • Celexa
• Escitalopram • Lexapro

Disclosures

Dr. Freeman has received grant or research support from Eli Lilly and Company, Forest Laboratories, and GlaxoSmithKline, is on the advisory boards of Otsuka and Takeda/Lundbeck, and is a consultant for PamLab LLC.

Dr. Joffe has received grant or research support from Cephalon/Teva, and is a consultant to Noven and Sunovion.

Dr. Cohen has received research support from AstraZeneca, Bayer HealthCare Pharmaceuticals, Bristol-Myers Squibb, Forest Laboratories, GlaxoSmithKline, National Institute of Mental Health, National Institute on Aging, National Institutes of Health, Ortho-McNeil Janssen, and Pfizer and has served on an advisory board for PamLab LLC.

This study was funded as an investigator-initiated trial by Forest Pharmaceuticals.

Challenges in recruiting women to postpartum depression (PPD) antidepressant treatment trials, which we encountered when conducting a trial of escitalopram, contribute to the limited body of knowledge about PPD treatment. Here we discuss results from a preliminary trial of escitalopram for PPD, and challenges of research in this area.

Escitalopram, the S-enantiomer of citalopram, is a selective serotonin reuptake inhibitor with high selectivity and potency that is FDA-approved for treating major depressive disorder (MDD) and generalized anxiety disorder. An agent with antidepressant and anxiolytic effects is particularly desirable for PPD because anxiety is more common in postpartum major depressive episodes than non-postpartum MDD.¹ Anxiety and depressive disorders commonly are comorbid in postpartum women.²

We conducted an open-label trial of escitalopram for women with PPD and anxiety. We initially attempted to recruit 20 women.

Methods

Patients received 8 weeks of treatment with escitalopram, 10 to 20 mg/d (flexible dose). After completing the initial phone screen, patients had 5 follow-up visits, once every 2 weeks for 8 weeks. The institutional review board at Massachusetts General Hospital approved this study and we obtained written informed consent from all patients at the first visit. Twelve patients completed the phone screen and 7 eligible patients were enrolled in the study over 32 months. Reasons for ineligibility included having a history of psychosis, onset of symptoms >3 months postpartum, or presenting >6 months after onset. Others declined to participate because of concern about the time commitment or because they pursued nonpharmacologic treatments after the evaluation visit. One patient was lost to follow-up. Three patients completed the study. The study was halted because of the slow pace of recruitment.

Patient selection. Patients were screened for a major depressive episode with postpartum onset within 3 months of childbirth; depressive symptoms may have developed during pregnancy and worsened postpartum to meet criteria for MDD. Women were eligible for the study if they:

• were age 18 to 45
• experienced a major depressive episode with symptoms developing within 3 months of childbirth
• presented within 6 months of childbirth
• had a Montgomery-Åsberg Depression Rating Scale (MADRS) score >15
• had a Beck Anxiety Inventory (BAI) score >10.

Patients who were pregnant or breast-feeding were excluded from the study per an agreement with the sponsor. In addition, women were excluded if they had taken any psychotropic medication within 2 weeks of enrollment; had active suicidal ideation, homicidal ideation, or presence of psychotic symptoms; had chronic depression or dysthymia; had chronic or treatment-resistant anxiety disorders; had a history of mania or hypomania; or had active alcohol or substance abuse within the past year.

Treatment. Patients received escitalopram, 10 mg/d, after the baseline visit. At the investigator’s discretion, the dose could be increased to 20 mg/d or lowered to 5 mg/d if side effects occurred.

Measures. At the first visit, patients were assessed with the Mini-International Neuropsychiatric Interview to verify MDD and exclude diagnoses that would determine ineligibility. MADRS and Edinburgh Postnatal Depression Scale (EPDS) were used at each visit to measure depressive symptoms.^3,4 The BAI was completed at each visit to measure anxiety symptoms. Obsessions and compulsions were measured with the Yale-Brown Obsessive Compulsive Scale (Y-BOCS)⁵ at baseline, and at all following visits if the patient scored >8 at baseline. The Clinical Global Impression Scales for severity and improvement were completed at each visit.⁶

Results

Of 7 patients enrolled, 3 completed the study, 2 were ineligible after the baseline visit, and 2 did not participate after the baseline visit (1 selected to pursue psychotherapy, and 1 was lost to follow-up).

Two of 3 patients responded to escitalopram (≥50% decrease on MADRS), and both were remitters (MADRS score <7). All 3 patients were responders on EPDS and BAI. One patient had Y-BOCS >8 at baseline (Total Y-BOCS score of 9, and final Y-BOCS score of 8) (Table).

Table

Symptom rating scale scores at baseline and study end

Discussion

Patients who stayed in treatment improved during the course of this study. Recruitment was difficult; we were able to recruit only 7 patients out of a projected 20 for the screening visit. We solicited feedback from local obstetrics health care providers and social workers on recruitment and attractiveness of the study as part of our routine collaboration with obstetrical services that screen for PPD. Primary reasons patients were not referred were that they were breast-feeding or they stated they would prefer to receive treatment from their primary care doctor. Recruitment difficulty in this study was in stark contrast to other recent studies completed at our center. For example, we have successfully recruited for menopausal depression and premenstrual dysphoric disorder treatment studies, and have completed large naturalistic studies of women with unipolar depression and bipolar disorder across pregnancy and postpartum. We suspect that many patients who were eligible for the study preferred to seek care from an obstetrician or primary care doctor with whom they already had a therapeutic alliance, and we also suspect that many women with PPD do not seek treatment at all, which is consistent with findings from other research groups.

Lessons learned from PPD research include:

• Including women who are breast-feeding is important because many women choose to breast-feed and suffer from PPD. Because antidepressant use during breast-feeding has been closely studied, it is appropriate to include potential research participants who are breast-feeding as long as they receive adequate information and are able to provide informed consent.
• Participants in PPD studies may require accommodations that take into account their role as a new mother, such as on-site childcare, home visits, or other strategies.
• Because of recruitment challenges in postpartum patients, multisite trials may be required to include adequate numbers of participants.

Related Resource

• Freeman MP, Joffe H, Cohen LS. Postpartum depression: Help patients find the right treatment. Current Psychiatry. 2012;11(11):14-21.

Drug Brand Names

• Citalopram • Celexa
• Escitalopram • Lexapro

Disclosures

Dr. Freeman has received grant or research support from Eli Lilly and Company, Forest Laboratories, and GlaxoSmithKline, is on the advisory boards of Otsuka and Takeda/Lundbeck, and is a consultant for PamLab LLC.

Dr. Joffe has received grant or research support from Cephalon/Teva, and is a consultant to Noven and Sunovion.

Dr. Cohen has received research support from AstraZeneca, Bayer HealthCare Pharmaceuticals, Bristol-Myers Squibb, Forest Laboratories, GlaxoSmithKline, National Institute of Mental Health, National Institute on Aging, National Institutes of Health, Ortho-McNeil Janssen, and Pfizer and has served on an advisory board for PamLab LLC.

This study was funded as an investigator-initiated trial by Forest Pharmaceuticals.

Challenges in recruiting women to postpartum depression (PPD) antidepressant treatment trials, which we encountered when conducting a trial of escitalopram, contribute to the limited body of knowledge about PPD treatment. Here we discuss results from a preliminary trial of escitalopram for PPD, and challenges of research in this area.

Escitalopram, the S-enantiomer of citalopram, is a selective serotonin reuptake inhibitor with high selectivity and potency that is FDA-approved for treating major depressive disorder (MDD) and generalized anxiety disorder. An agent with antidepressant and anxiolytic effects is particularly desirable for PPD because anxiety is more common in postpartum major depressive episodes than non-postpartum MDD.¹ Anxiety and depressive disorders commonly are comorbid in postpartum women.²

We conducted an open-label trial of escitalopram for women with PPD and anxiety. We initially attempted to recruit 20 women.

Methods

Patients received 8 weeks of treatment with escitalopram, 10 to 20 mg/d (flexible dose). After completing the initial phone screen, patients had 5 follow-up visits, once every 2 weeks for 8 weeks. The institutional review board at Massachusetts General Hospital approved this study and we obtained written informed consent from all patients at the first visit. Twelve patients completed the phone screen and 7 eligible patients were enrolled in the study over 32 months. Reasons for ineligibility included having a history of psychosis, onset of symptoms >3 months postpartum, or presenting >6 months after onset. Others declined to participate because of concern about the time commitment or because they pursued nonpharmacologic treatments after the evaluation visit. One patient was lost to follow-up. Three patients completed the study. The study was halted because of the slow pace of recruitment.

Patient selection. Patients were screened for a major depressive episode with postpartum onset within 3 months of childbirth; depressive symptoms may have developed during pregnancy and worsened postpartum to meet criteria for MDD. Women were eligible for the study if they:

• were age 18 to 45
• experienced a major depressive episode with symptoms developing within 3 months of childbirth
• presented within 6 months of childbirth
• had a Montgomery-Åsberg Depression Rating Scale (MADRS) score >15
• had a Beck Anxiety Inventory (BAI) score >10.

Patients who were pregnant or breast-feeding were excluded from the study per an agreement with the sponsor. In addition, women were excluded if they had taken any psychotropic medication within 2 weeks of enrollment; had active suicidal ideation, homicidal ideation, or presence of psychotic symptoms; had chronic depression or dysthymia; had chronic or treatment-resistant anxiety disorders; had a history of mania or hypomania; or had active alcohol or substance abuse within the past year.

Treatment. Patients received escitalopram, 10 mg/d, after the baseline visit. At the investigator’s discretion, the dose could be increased to 20 mg/d or lowered to 5 mg/d if side effects occurred.

Measures. At the first visit, patients were assessed with the Mini-International Neuropsychiatric Interview to verify MDD and exclude diagnoses that would determine ineligibility. MADRS and Edinburgh Postnatal Depression Scale (EPDS) were used at each visit to measure depressive symptoms.^3,4 The BAI was completed at each visit to measure anxiety symptoms. Obsessions and compulsions were measured with the Yale-Brown Obsessive Compulsive Scale (Y-BOCS)⁵ at baseline, and at all following visits if the patient scored >8 at baseline. The Clinical Global Impression Scales for severity and improvement were completed at each visit.⁶

Results

Of 7 patients enrolled, 3 completed the study, 2 were ineligible after the baseline visit, and 2 did not participate after the baseline visit (1 selected to pursue psychotherapy, and 1 was lost to follow-up).

Two of 3 patients responded to escitalopram (≥50% decrease on MADRS), and both were remitters (MADRS score <7). All 3 patients were responders on EPDS and BAI. One patient had Y-BOCS >8 at baseline (Total Y-BOCS score of 9, and final Y-BOCS score of 8) (Table).

Table

Symptom rating scale scores at baseline and study end

Discussion

Patients who stayed in treatment improved during the course of this study. Recruitment was difficult; we were able to recruit only 7 patients out of a projected 20 for the screening visit. We solicited feedback from local obstetrics health care providers and social workers on recruitment and attractiveness of the study as part of our routine collaboration with obstetrical services that screen for PPD. Primary reasons patients were not referred were that they were breast-feeding or they stated they would prefer to receive treatment from their primary care doctor. Recruitment difficulty in this study was in stark contrast to other recent studies completed at our center. For example, we have successfully recruited for menopausal depression and premenstrual dysphoric disorder treatment studies, and have completed large naturalistic studies of women with unipolar depression and bipolar disorder across pregnancy and postpartum. We suspect that many patients who were eligible for the study preferred to seek care from an obstetrician or primary care doctor with whom they already had a therapeutic alliance, and we also suspect that many women with PPD do not seek treatment at all, which is consistent with findings from other research groups.

Lessons learned from PPD research include:

• Including women who are breast-feeding is important because many women choose to breast-feed and suffer from PPD. Because antidepressant use during breast-feeding has been closely studied, it is appropriate to include potential research participants who are breast-feeding as long as they receive adequate information and are able to provide informed consent.
• Participants in PPD studies may require accommodations that take into account their role as a new mother, such as on-site childcare, home visits, or other strategies.
• Because of recruitment challenges in postpartum patients, multisite trials may be required to include adequate numbers of participants.

Related Resource

• Freeman MP, Joffe H, Cohen LS. Postpartum depression: Help patients find the right treatment. Current Psychiatry. 2012;11(11):14-21.

Drug Brand Names

• Citalopram • Celexa
• Escitalopram • Lexapro

Disclosures

Dr. Freeman has received grant or research support from Eli Lilly and Company, Forest Laboratories, and GlaxoSmithKline, is on the advisory boards of Otsuka and Takeda/Lundbeck, and is a consultant for PamLab LLC.

Dr. Joffe has received grant or research support from Cephalon/Teva, and is a consultant to Noven and Sunovion.

Dr. Cohen has received research support from AstraZeneca, Bayer HealthCare Pharmaceuticals, Bristol-Myers Squibb, Forest Laboratories, GlaxoSmithKline, National Institute of Mental Health, National Institute on Aging, National Institutes of Health, Ortho-McNeil Janssen, and Pfizer and has served on an advisory board for PamLab LLC.

This study was funded as an investigator-initiated trial by Forest Pharmaceuticals.