Depression is more common in women than in men during the reproductive years, particularly in association with reproductive transitions and several common gynecological conditions. In the first part of this chapter, we review specific evidence supporting an association of mood disturbance with puberty in girls, across the menstrual cycle, with polycystic ovary syndrome, with infertility, and during the transition to menopause. Because of the prevalence of depression in women during the reproductive years, there are some important implications for pharmacologic treatment. Some psychotropic medications can have adverse effects on the hypothalamic–pituitary–gonadal (HPG) axis, which manifest in menstrual dysfunction, as well as potentially disruptive interactions with both endogenous and exogenous reproductive hormones. In the second part of this chapter, we review what is known about the effect of psychotropic medications on the HPG axis, and discuss important interactions with reproductive hormones. Given that women of reproductive age comprise a large proportion of psychiatric patients, knowledge about these special considerations will improve treatment outcomes for a large number of women.

Case illustration: Marta is a 13-year-old girl brought to see you by her mother, who is concerned because Marta has appeared listless, sad, and withdrawn. Marta had been a straight. A student until a couple of months ago, when she lost interest in school and had difficulty focusing on homework. Developmental history reveals that Marta’s mother has had recurrent episodes of major depression, with a particularly severe, prolonged episode while pregnant with Marta. Marta was born with low birth weight. Within her first 2 years of life, she caught up and exceeded normal weight for height. Early developmental milestones were normal. Family life is characterized by frequent verbal altercations between Marta’s parents. Marta tends to brood a lot after hearing these; by contrast, her older brother tends to distract himself. Marta began showing breast development by age 9, with menarche at age 11. Two months ago, Marta’s best friend rejected her in favor of a more socially popular set of friends. Marta confides in you that she feels “fat and ugly,” which makes her feel hopeless about ever being socially desirable. She eats for comfort even when not hungry, and then feels guilty. She has had thoughts about killing herself by overdosing on pills from the medicine cabinet, but says she would never actually do that because it would be a sin.

EPIDEMIOLOGY

One of the most striking aspects of the epidemiology of depression is the marked increase in prevalence in girls, but not boys, which arises in mid-puberty. In prepubertal children, studies consistently show that rates of depressive disorder are either similar in boys and girls, or slightly higher in boys.¹ Beginning at about age 13, the prevalence of depressive disorders in girls rises to about twice that in boys. In the National Comorbidity Survey – Adolescent Supplement, a nationally representative face-to-face survey of over 10,000 adolescents aged 13 to 18 years in the United States, the lifetime prevalence of major depressive disorder or dysthymia was 15.9% in girls and 7.7% in boys.²

PATHOPHYSIOLOGY (BOX 8-1)

The marked gender difference in vulnerability to depression that arises during puberty may be best understood from a biocultural framework, with gender-linked biological factors interacting with cognitive, environmental, and sociocultural contexts. Research to date has identified the following factors as especially influential:

• 
  

  Low birth weight. Low birth weight is a biomarker for in utero stress, which may confer health risks later in life. In the prospective, longitudinal Great Smoky Mountains Study, the cumulative prevalence of depression over the 3-year period from 13 to 16 years old was 8.4% among girls with normal birth weight, but 38.1% among girls with low birth weight.³ Low birth weight appeared to potentiate other pregnancy-linked risk factors, such as premature birth and labor complications. The effect of low birth weight on adolescent depression was considerably less marked for boys than for girls.

  
  
• 
  

  Changes in gonadal hormones. A study comparing the influence of age, pubertal stage, and circulating hormone levels found that the gender difference in depression prevalence arose at a specific stage of puberty—Tanner Stage III, when breast tissue grows beyond areola and pubic hair coarsens— regardless of age. Statistical modeling showed that the link between pubertal stage and depression was largely due to estradiol and testosterone levels.⁴

  
  
• 
  

  Pubertal timing. Most, but not all, studies have found that the earlier girls begin puberty, the more likely they are to develop depressive symptoms in adolescence.^5,6 The reasons for this have not yet been empirically established. One posited explanation is that girls who appear physically more developed may be treated by others as more mature than they are and they may not be developmentally ready for this. Another is that early puberty is a biomarker for early life stressors, such as family discord and/or absent father, that might confer increased risk for subsequent depression. A retrospective study suggests that only girls with certain estrogen receptor gene polymorphisms are vulnerable to experiencing early menarche in the context of family discord,⁷ supporting a posited link between genetic vulnerability, early life stress, and hormonal activation of depression in adolescent girls. Other studies are exploring the role of weight in mediating links between early puberty and depression. Low-birth-weight babies who have substantial “catch up” growth early in life are vulnerable to childhood obesity. This leads to increases in leptin, a hormone manufactured in adipose tissue that plays a role in initiating the cascade of physiologic changes leading to menarche.⁸ Such effects can be transgenerational; maternal antenatal depression can increase the risk of low birth weight in infants,⁹ contributing to a sequence of biopsychosocial events, including childhood excessive weight gain, early puberty, and adolescent onset depression.

  
  
• 
  

  Environmental stress. Girls are substantially more likely than boys to experience sexual abuse and/or rape. Girls’ exposure to these risks increases substantially at adolescence, explaining a portion of the gender difference in vulnerability to depression.¹⁰ In addition to differential exposure, research suggests the possibility of gender differences in types of reactivity to stress. Ebstein Barr virus (EBV) titers are a biomarker for stress-linked immunocompromise. Traumatic life events have been shown to increase EBV titers in girls, but not boys, in early adolescence.¹¹

  
  
• 
  

  Cognitive style. Certain characteristic ways of perceiving the world and reacting to stress are correlated with increased likelihood of becoming depressed. These include rumination, expectations of abandonment, and beliefs that one is defective or unworthy. By contrast, a problem-solving approach to stressors can be protective against depression. As compared with adolescent boys, adolescent girls are more likely to endorse cognitive styles conducive to depression. There is a reciprocal relationship among cognitive styles, stress, and depressive symptoms; cognitive vulnerabilities contribute to experiencing more stress, and depressive symptoms and stress intensify cognitive vulnerabilities.¹²

  
  
• 
  

  Sociocultural gender role expectations. Perceived failure to live up to gender role expectations can influence the development of depression. In social environments that value slenderness in girls, body image at puberty is a particularly salient influence on depression. At puberty, self-perception of being overweight and efforts to lose weight, regardless of actual body mass index, correlate significantly with depressive symptoms in girls but not in boys.¹³

CLINICAL PRESENTATION

As compared to adolescent boys, adolescent girls with depressive disorders endorse more symptoms of guilt, body image dissatisfaction, self-blame, disappointment in themselves, fatigue, hypersomnia, and impairment in concentration. As anxiety disorders are more prevalent in adolescent girls as compared to boys, depression may present with anxiety disorder spectrum comorbidity. Girls are also substantially more likely than boys to attempt suicide; this gender difference declines in young adulthood.¹⁴ Girls are less likely than boys to experience anhedonia and morning symptom intensification.¹⁵

COURSE AND NATURAL HISTORY

Among adolescents who develop major depressive episodes, female sex is associated with longer episode duration and a greater likelihood of recurrent episodes in young adulthood. Among adolescent girls, but not boys, conflict with parents predicts young adult recurrence of depressive episodes.^16,17

ASSESSMENT AND DIFFERENTIAL DIAGNOSIS

Recognition that girls enter a high-risk time for developing depression at mid-puberty can alert clinicians to assess for depression at that time. Detection of major depression can be improved by routine screening with a validated tool. The 9-item Patient Health Questionnaire is a public domain self-report depression screening tool that is linked with Diagnostic and Statistical Manual criteria for major depression. It has been validated for use in adolescents,¹⁸ and can be used to measure symptom severity and track outcome after treatment. In addition to evaluating pubertal girls for the presence and severity of depressive symptoms, assessing the contextual variables in Table 8-1 can help guide treatment.

Directly asking about thoughts of suicide is important, given the especially high risk of suicide attempts in adolescent girls. If suicidal thoughts are present, ascertain whether the patient has a specific plan, the level of her intent to act on the plan, and her access to means (e.g., firearms or medications in the home). Also assess for protective factors, such as willingness to share her thoughts with supportive others, hopefulness about the future, impulse control, and religious beliefs.

The differential diagnosis of pubertal depression (Box 8-2) in girls is similar to that of depression in general. However, certain conditions deserve special emphasis in this population:

• 
  

  Bipolar disorder. The first episode of bipolar disorder in females is more often depressive, while the first episode in males is more often manic.¹⁹

  
  
• 
  

  Use of addictive substances causing or contributing to depressive symptoms.

  
  
• 
  

  Iron deficiency anemia. Menstruation increases a girl’s risk of developing iron deficiency. Low ferritin levels are associated with increased depressive symptoms.²⁰

TREATMENT (BOX 8-3)

The treatments with the most robust evidence for efficacy in adolescents with major depression are fluoxetine and cognitive-behavioral therapy (CBT), particularly in combination.²¹ Other antidepressants that can be useful for adolescent depression are sertraline, citalopram, or escitalopram. When prescribing (explaining why just these specific agents are noted) antidepressants to adolescent girls, it is helpful to discuss contraception for those who are sexually active and at risk for unintended pregnancy. It is important to monitor closely for emerging suicidal thoughts or impulses during the course of treatment of adolescents with depression.

CBT can be especially helpful for girls with trauma histories, ruminative coping styles, and self-deprecatory or perfectionist beliefs. Other types of psychotherapy can be considered as well. Interpersonal therapy (IPT) may be especially helpful for girls whose depressive symptoms seem linked to interpersonal stressors or difficulties with navigating social roles. Family-focused therapy can be effective for girls whose depressive episodes are linked with family discord and communication difficulties.

Relatively few systematic studies have been completed about the efficacy of lifestyle changes and self-management for girls with pubertal-onset depression. Preliminary data are promising for positive effects of aerobic exercise²² and technology-enabled biofeedback.²³

SUMMARY

The risk of developing depressive disorders increases substantially at mid-puberty for girls. This heightened risk may be influenced by genetic vulnerability, epigenetic influences in utero, gonadal hormones, environmental stressors, cognitive and coping styles, and pubertal timing. As compared to adolescent boys, adolescent girls with depressive episodes may be at greater risk of suicide attempts, longer episode duration, and subsequent recurrence. There is robust evidence for the efficacy of fluoxetine and CBT for treating major depression in adolescents.

Case outcome. Marta was diagnosed with a major depressive episode. Factors contributing to her vulnerability to depression may have included genetic predisposition, epigenetic effects of her in utero environment, early puberty, family discord, cognitive style, coping style, and current social stressors. She began fluoxetine 20 mg daily and CBT. The latter began with behavioral activation, aided by a mobile phone app, which helped Marta keep track of the activity schedule she’d set up for herself. Marta then learned problem-solving and assertiveness skills, which she used to explain to her parents how their arguing affected her and to work with them on changing family communication patterns. She also worked on identifying and reframing self-deprecatory core beliefs, by logging them on her mobile phone when they occurred and testing them against evidence. She achieved full remission, but remained on fluoxetine for 6 months to prevent recurrence.

Case illustration: Denise is a 23-year-old woman whose boyfriend has asked her to seek help. He has noticed that she seems unusually grouchy at times, sometimes even throwing things toward him or storming out of the room. He has also noticed that she tends to call in sick at work on those days. He thinks there might be a premenstrual pattern to these mood and behavior changes. Denise agrees, and has noticed that a few days before each menstrual period, she becomes more irritable, anxious, and overwhelmed. She eats more, especially fatty foods, feels tired and bloated, has headaches, and sleeps more. The oral contraceptive pills she takes have not had any effect on these symptoms. When her boyfriend mentioned his concerns, Denise found an online daily symptom log and charted her symptoms across two menstrual cycles. This showed that her symptoms began 5 to 6 days before menses, peaked in severity 2 days before menses, and resolved the day after menstrual bleeding began. She is worried about getting a suboptimal performance review at work, and about ruining her relationship with her boyfriend.

EPIDEMIOLOGY

There are three types of depression-related disorders linked with the menstrual cycle: premenstrual dysphoric disorder (PMDD), psychological symptoms manifesting as part of a premenstrual syndrome (PMS), and premenstrual exacerbation of depressive disorders (PME-DD) (Fig. 8-1). PMDD is defined as having five or more characteristic symptoms during the last week of the luteal (postovulatory) phase of all or most menstrual cycles. By definition, symptoms begin to remit within a few days after the onset of menstrual bleeding and are gone by a week after the onset of menstrual bleeding. Symptoms must be severe enough to interfere with functioning. The symptoms may include irritability, depressed mood, hopelessness, anxiety, tension, mood lability, anhedonia, change in appetite, poor concentration, fatigue, lack of energy, sleep disturbance, feeling overwhelmed, and/or physical symptoms, such as bloating or breast tenderness.²⁴ Using this definition, 1% to 8% of menstruating women meet criteria for PMDD in studies using prospective daily charting to confirm a premenstrual pattern.²⁵ PMS has no fully consensual definition; the term is often used broadly to refer to premenstrual mood and/or physical changes, regardless of distress or functional impairment. When defined as subsyndromal PMDD, for example, fewer than five symptoms, but causing distress and/or impairment, PMS has been found to affect 12% to 18% of menstruating women.²⁶ It is important to note that while psychological symptoms are common in PMS, diagnostic criteria for this condition can be met in the absence of any mood disturbance when physical symptoms are prominent (e.g., cramping and abdominal pain) and interfere with daily activities and/or are distressing.

The prevalence of PME-DD has not yet been rigorously studied. In a sample of 433 premenopausal depressed women from the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study, none of whom were using oral contraceptives, 64% reported that they perceived premenstrual worsening of their depressive symptoms.²⁷ Among 58 menstruating 13- to 53-year-old females with depressive disorders in a community sample, 58% had one or more depressive symptoms worsen before menses, based on prospective daily charting.²⁸ Similarly, in a cohort of 293 premenopausal women with bipolar disorder enrolled in the Systematic Treatment Enhancement Program for Bipolar Disorder (STEP-BD) study,²⁹ 65% reported premenstrual worsening of their underlying bipolar disorder.

PATHOPHYSIOLOGY

Numerous studies have found alterations of neurotransmission of serotonin and gamma amino butyric acid (GABA) in women with PMDD, often specific to the luteal phase.^30,31 By contrast, most studies have found no difference in HPG axis functioning between women with and without PMDD. However, eliminating ovulation with gonadotropin releasing hormone (GnRH) analogs usually alleviates PMDD symptoms, while adding back estrogen, progesterone, or both causes symptom recurrence.³² To explain these findings, it is posited that women with PMDD have abnormal neurotransmitter responses to normal gonadal hormonal flux. A specific hypothesis being investigated is that women with PMDD have reduced central nervous system sensitivity to, and/or altered levels of, the progesterone metabolite allopregnanolone, which functions as a neuroactive GABA receptor modulator.³³

Functional neuroimaging studies are being conducted to elucidate the functional neuroanatomy of PMDD. The most consistent findings to date involve increased amygdala reactivity to negative stimuli in the luteal phase, suggestive of enhanced negative emotional processing.^34,35 This is consistent with findings from studies using functional tasks, such as facial discrimination, demonstrating that in the luteal phase women with PMDD display a negative bias in the processing of nonverbal emotional stimuli.³⁶

CLINICAL PRESENTATION (BOX 8-4)

PMDD appears to be qualitatively distinct from major depressive disorder, apart from its premenstrual temporal pattern. The most commonly endorsed mood symptoms are not depressed mood or anhedonia, but rather irritability and anger, followed by anxiety, lethargy, and mood lability.³⁷ Another contrast is that, although changes in appetite can be a symptom of either major depressive disorder or PMDD, the appetite change in PMDD is distinct. It is nearly always increased rather than decreased appetite, often with desires or cravings for food items that are sweet and high in fat content, happening specifically in the luteal phase.^38,39

Although thoughts of death or suicide are not listed as a definitional symptom in PMDD, they are relatively common during symptomatic phases. In a large epidemiologic survey in the United States, women with PMDD were significantly more likely to report suicidal ideation, plans, and attempts than were women with no premenstrual symptoms.⁴⁰

There is wide variability from woman to woman in the duration of symptoms across the menstrual cycle. Most often the most severe symptoms are reported to occur during the two days prior to menses, with milder symptoms preceding. Within individual women, symptoms tend to remain qualitatively similar from cycle to cycle, although severity may vary.³⁷

In keeping with some women’s experiences that they “become a different person” while premenstrual, personality scales change significantly from the follicular to the luteal phase in women with PMDD.⁴¹ Locus of control, generally a relatively fixed characteristic, is more external in the luteal phase than in the follicular phase in women with PMDD.⁴² In addition, subjective, behavioral, and neurophysiologic measures of stress reactivity are increased in the luteal phase in women with PMDD.⁴³ Small but significant changes in performance on attention and memory tasks are also seen in the luteal phase of women with PMDD, confirming subjective reports of reduced ability to concentrate and remember things during the symptomatic phase.³⁹

COURSE AND NATURAL HISTORY

Although premenstrual dysphoria is relatively common in adolescent girls,⁴⁴ PMDD does not typically begin at menarche. Once established, the symptom pattern tends to remain stable over time but can worsen with age and childbearing.⁴⁵ To date, there are no rigorous longitudinal studies of the course of PMDD as women enter perimenopause. Case reports suggest that some women may experience symptom worsening during perimenopause, with PMDD symptoms not necessarily limited to the premenstrual phase given that cycles are irregular and ovarian steroid fluctuations are unpredictable. PMDD remits after menopause.

ASSESSMENT AND DIFFERENTIAL DIAGNOSIS (BOX 8-5)

The key to accurate diagnosis of PMS, PMDD, and PME is prospective daily symptom charting across at least two menstrual cycles. This distinguishes premenstrual mood disorders from other conditions that can present with mood shifts, such as rapid-cycling bipolar disorder and personality disorders. There are several types of daily symptom charts available in the public domain, often online and as applications for mobile devices. An instrument that is especially helpful for clinical diagnosis is the Premenstrual Tension Syndrome (PMTS) scale.⁴⁶ This has a self-rated version in the form of a visual analog scale (PMTS-VAS), which averages 68 seconds to complete, and an observer version in the form of a Likert scale (PMTS-OR). Each of these has 11 items, which correspond to the Diagnostic and Statistical Manual definitional symptoms of PMDD.

To guide interview-based diagnosis of PMDD, the Structured Clinical Interview for DSM-IV-TR-defined PMDD (SCID-PMDD) is an instrument with high sensitivity and reliability.⁴⁷ It is helpful to schedule two evaluation appointments, one in the follicular phase of the cycle and one in the late luteal phase, to observe mental status changes directly.

TREATMENT (BOX 8-6)

In keeping with the posited role of serotonergic dysfunction in the etiology of PMDD, serotonergic antidepressants are effective in its treatment. The serotonin-selective reuptake inhibitors (SSRIs) fluoxetine, sertraline, paroxetine, and escitalopram are approved by the Food and Drug Administration (FDA) for this indication. By contrast, antidepressants that have limited to no effect on serotonergic function, such as bupropion, desipramine, and maprotiline, have not performed significantly better than placebo for PMDD.^48–50

In addition to continuous dosing (taking the same antidepressant dose every day), numerous studies and a meta-analysis have shown that luteal phase dosing (taking antidepressants only during the 14 days prior to menses) is effective for PMDD.⁵¹ Some studies have also found that symptom-onset dosing is effective, with a mean duration of use of 6 to 9 days during the late luteal phase of the menstrual cycle.^52,53 For PME-DD, preliminary data suggest that boosting the dose in the luteal phase is effective.⁵⁴ The rapid response and effectiveness of these dosing patterns imply that the mechanism of antidepressant action for premenstrual mood syndromes differs from that in major depressive disorder. There is some evidence to support the hypothesis that serotonergic medications alleviate premenstrual dysphoria via a rapid effect on central nervous system sensitivity to neuroactive hormonal GABA agonists.⁵⁵

Patients who are candidates for luteal phase dosing regimens are:

• 
  

  Those who prefer to minimize medication use, due to side effects and/or values.

  
  
• 
  

  Those who have regular, predictable menstrual cycles that can be tracked (for luteal phase dosing and dose escalation) or who can recognize early symptoms (for symptom-onset dosing).

  
  
• 
  

  Those who do not experience problematic transient side effects upon starting or discontinuing antidepressants.

Given the link between hormonal cyclicity and symptoms in PMDD, many have wondered whether hormonal interventions would be effective. Progesterone-only and combined oral contraceptive (COC) pills, at one time used relatively widely for treating PMS, have been found to be no more effective than placebo. The one exception is the COC drospirenone/ethinyl estradiol when used in a formulation that has 24 days of active agent and 4 days of placebo. Drospirenone, the progestin, is an antiandrogenic analog of spironolactone, a diuretic shown to alleviate PMS. It is not known why this particular COC is effective for treating PMDD, while others have not been shown to be effective. Factors contributing to its efficacy may include properties of its unique progestin and the shorter hormone-free interval, which maintains better suppression of the endogenous hypothalamic-pituitary-ovarian axis. This agent is FDA approved for treatment of PMDD, although the potential benefits must be weighed against a heightened risk of blood clots and other adverse effects of COCs.

Studies have begun to evaluate the efficacy of using COCs continuously, with active hormones daily and no placebo pills, therefore, no menstrual bleeding. The results of three randomized, double-blind, placebo-controlled trials of the efficacy of continuous COCs for PMDD are inconsistent. All showed a high placebo response rate; only 1 of the 3 found significantly better response to active treatment than to placebo.⁵⁶

While COCs have not been shown to worsen PMS/PMDD, dysphoria and adverse mood symptoms have been reported as side effects. It has been difficult to definitely establish whether or not there is a relationship between COCs and depressive symptoms, in part because of the numerous varieties of COCs, but if depression is a side effect it appears to be an uncommon one.⁵⁷ One large epidemiologic study found that in 71% of women taking a COC, mood remained stable or improved, even among those with a history of depression.⁵⁸

Other hormonal therapies are effective for PMDD, but are rarely used due to side effects. GnRH analogs, which create a temporary “chemical menopause” by suppressing estradiol and progesterone, are effective because they eliminate the hormone fluctuations, but long-term use is limited by risks associated with a hypoestrogenic menopausal state.⁵⁹ This option may be best reserved for women who are expected to reach menopause within one year, though it’s use may be limited by expense and limited insurance coverage for off-label medication use. Danazol, an androgenic and anti-gonadotrophic agent, is also effective with either continuous or luteal phase dosing,⁶⁰ but is rarely used due to androgenic side effects.

Elimination of fluctuating estradiol and progesterone through bilateral oophorectomy is expected to eliminate premenstrual dysphoric symptoms, although few empiric data exist.⁶¹ This surgical menopause should be considered as a tertiary intervention because of adverse health sequelae (hypoestrogenism, osteoporosis, sexual dysfunction, possible negative long-term health effects). Postoperative hormone replacement can cause symptom recurrence. If hysterectomy is performed with preservation of at least one ovary, symptoms are expected to persist because the hormones will continue to fluctuate despite the absence of a menstrual marker.

Among the many herbal products used in an attempt to alleviate premenstrual dysphoric symptoms, only one, vitex agnus castus extract (VACE), has demonstrated efficacy in several prospective, randomized, double-blind, placebo-controlled trials. VACE, a plant with weak estrogenic effects, is popularly known as chaste tree berry. VACE has efficacy superior to placebo and equal to fluoxetine overall, although fluoxetine is somewhat more effective for mood symptoms.⁶² Optimal dosing for PMDD is 20 mg daily.⁶³ Side effects are similar to those of the placebo arm in controlled trials.⁶⁴

The efficacy of psychotherapy for premenstrual mood disorders is unclear. There have been several studies of CBT with inconsistent results. This is in part due to methodologic limitations, but also due to a lack of clarity or consensus about optimal conceptual approaches to psychotherapeutic interventions for premenstrual dysphoria. For example, some therapies focus on helping women to reduce premenstrual stress by scheduling fewer activities or deadlines at that time, while other focus on maintaining normal functioning throughout the cycle. Mindfulness and acceptance-based therapies are being explored as promising interventions for premenstrual dysphoric mood states.⁶⁵

Lifestyle interventions for PMS and PMDD are also being explored. In several studies, most relatively small and nonrandomized, aerobic exercise has led to symptom improvement.⁶⁶ The most promising nutritional intervention is calcium supplementation (1200 mg daily), which has been found to be superior to placebo in alleviating premenstrual dysphoric symptoms.⁶⁷

Treatment studies for PME-DD are limited, with an open-label trial showing that augmentation of an antidepressant with the COC drospirenone/ethinyl estradiol alleviated premenstrual mood symptoms⁶⁸ and case reports suggesting that premenstrual dose escalation of antidepressants may be helpful.⁶⁹

SUMMARY

PMDD, PMS, and PME-DD are prevalent disorders that can cause substantial distress and impairment. It is posited that they arise in women who have abnormal central nervous system responses to gonadal hormonal flux, particularly in neurotransmission involving serotonin and GABA. Diagnosis is aided by prospective daily symptom charting across 2 menstrual cycles. PMDD can be treated with serotonergic antidepressants, using continuous dosing, luteal phase dosing, or symptom-onset dosing. The hormonal contraceptive drospirenone/ethinyl estradiol 24/4 is also effective for PMDD. Other promising interventions are VACE, CBT, aerobic exercise, and calcium supplementation. PME-DD may response to luteal phase escalation of antidepressant dose or augmentation with drospirenone/ethinyl estradiol 24/4, but further studies are needed.

Case outcome. Denise was diagnosed with PMDD. She was reluctant to take medication, but felt comfortable trying low-dose sertraline (50 mg daily) during the luteal phase only. This substantially reduced her irritable and anxious mood, but only somewhat alleviated her fatigue, headaches, and bloating sensation. She began taking VACE 20 mg daily, and started using an exercise digital video disc DVD three to four times per week. Her symptoms improved. She now has only occasional mild fatigue, headaches, and irritability for 1 or 2 days before some of her menstrual cycles. She charts her cycles to predict vulnerable days, and maintains heightened awareness of her feelings and her reactions to her boyfriend and coworkers on those days.

Case illustration: Anne is a 32-year-old graphic designer who requests evaluation for depression. Her supervisor had become concerned after noticing that Anne, whose work had been consistently exemplary in the past, had procrastinated on several projects and had missed a couple of important deadlines over the past 3 months. Anne acknowledges that she has felt down and sluggish over that period of time, with reduced energy and difficulty getting out of bed. She has lost interest in her work, which she used to love, and has struggled to focus on it. When asked if she feels down about anything in particular, she begins to sob. She explains that she has been trying unsuccessfully to become pregnant for the past year. She is considering infertility treatment, but fears learning that something is wrong with her reproductive system, because her menses have always been irregular. She feels shaky in her marriage. She has gained weight over the past few years, especially lately, and thinks her husband finds her unattractive. He has always wanted to be a father. She fears he will leave her if evaluation reveals she is infertile. Examination is notable for hirsutism, mild acne, and central obesity.

EPIDEMIOLOGY

PCOS is an endocrine disorder with mutually interacting reproductive and metabolic abnormalities. Based on international consensus criteria, women are diagnosed with PCOS when at least two of three cardinal signs—hyperandrogenism (biochemical or symptomatic hirsutism, acne, male-pattern balding), anovulation or oligo-ovulation, and polycystic ovarian morphology on ultrasound—are present. These are often accompanied by insulin resistance and features of metabolic syndrome. Obesity is present in about half of women with PCOS.

Depending on diagnostic criteria used, the general population prevalence for PCOS ranges between 6% and 18% of women.⁷⁰ Women with PCOS are especially vulnerable to depressive symptoms. A meta-analysis of 26 studies found a significant increase in depressive as well as anxiety symptom scores in women with PCOS as compared to women without PCOS, with a moderate effect size.⁷¹ Studies using Diagnostic and Statistical Manual diagnoses have found that the current prevalence of major depressive disorder is 26% to 35% of study participants with PCOS as compared to about 10% of control women.^72,73

PATHOPHYSIOLOGY

The relationship between PCOS and depression is likely bidirectional and multidetermined, as shown in Figure 8-2. Stigma and distress from some of the symptoms of PCOS, such as hirsutism, acne, male-pattern balding, and obesity, may contribute to the risk of depression. However, studies show that these factors do not fully account for the increase in depressive symptoms in women with PCOS as compared to controls.⁷¹ Similarly, consequences of PCOS, such as subfertility, do not fully explain vulnerability to depression. When compared with women with non-PCOS infertility, women with PCOS were found to have more depressive symptoms,⁷⁴ and among women with PCOS, unfulfilled wishes to have a child have not been found to increase depressive symptoms.⁷⁵

Studies investigating a posited link between gonadal hormone levels and depressive symptoms in women with PCOS have been inconclusive. Lower circulating androgen levels have been found in some studies to correlate with more depressive symptoms in women with PCOS, but other studies have found no such link.^76,77 A more consistent finding is that women with PCOS have heightened physiologic and emotional responses to stress as compared to control women. Abnormalities include elevations in hypothalamic-pituitary-adrenal (HPA) axis response, elevated cardiovascular response, and increased psychological distress.⁷⁸ These abnormalities could contribute to depressive symptoms, although a direct link has not yet been demonstrated. Sleep apnea is more common among women with PCOS than among BMI-matched controls. Levels of pro-inflammatory cytokines have been found to be elevated in women with PCOS,⁷⁹ suggesting that PCOS includes a chronic, low-grade inflammatory state. These are other possible links with increased risk of depressive symptoms.

CLINICAL PRESENTATION

PCOS comprises a cluster of reproductive and metabolic abnormalities. Table 8-2 shows typical clinical features. The most frequent initial presenting symptoms are hirsutism and irregular menstrual cycles. Some cases are discovered due to infertility or unexplained weight gain. Although depression and anxiety are not core clinical features of PCOS,⁸⁰ some women seeking treatment for depression and anxiety may have undiagnosed comorbid PCOS.

COURSE AND NATURAL HISTORY

Symptoms and signs of PCOS usually begin to be manifest in late puberty, after menarche. Since irregular menstrual cycles are relatively common in normal adolescence, often the diagnosis is not made until later. As women with PCOS become older, androgen levels and associated symptoms (e.g., hirsutism and acne) tend to decline. However, metabolic risk factors tend to worsen, with increases in blood pressure, total cholesterol, low density lipoprotein (LDL)-cholesterol, triglycerides, fasting blood sugar, insulin, and (BMI) over time.^81,82 There is no clear age link with comorbid depressive symptoms; the risk of depression appears to be elevated at all stages of the illness.

ASSESSMENT AND DIFFERENTIAL DIAGNOSIS

For women who initially present with depression, it is important to consider a diagnosis of PCOS if they also have hirsutism, irregular menses, infertility, or other common PCOS symptoms. The key steps in assessing women for PCOS include the following:

• 
  

  Assessing for the cardinal symptoms of hyperandrogenism, polycystic ovaries, and anovulation/oligo-ovulation.

  
  
• 
  

  Ruling out disorders with overlapping clinical features, many of which also confer a higher risk of depressive symptoms (Box 8-7).

  
  
• 
  

  Assessing for specific cardiovascular and other risk factors once a diagnosis of PCOS is confirmed. Consensus guidelines⁸³ recommend that all women with PCOS be assessed regularly for BMI, waist circumference, blood pressure, and lipid profile. Oral glucose challenge tests are recommended for those with BMI >30 kg/m², age >40 years, history of gestational diabetes mellitus, or family history of Type 2 diabetes mellitus.

Given the prevalence and risks of depressive symptoms, an expert panel recommends routine assessment for depression in all women with PCOS, even when depressive symptoms are not part of the presenting complaints.⁸³

TREATMENT

Treatment for women with PCOS is focused on alleviating current symptoms and on reducing long-term health risks. A patient’s goals may include reducing signs of hyperandrogenism (e.g., hirsutism, acne), regulating menses, conceiving and maintaining a healthy pregnancy, losing weight, reducing risks of cardiovascular disease and diabetes, and treating comorbid depressive symptoms. Relevant interactions between depression and PCOS treatment are summarized below.

• 
  

  Lifestyle changes. A key component of successful treatment of PCOS is to maintain a healthy diet and adequate aerobic exercise. These interventions reduce metabolic and menstrual abnormalities, improve inflammatory profiles, increase responsiveness to insulin, help maintain normal weight, and reduce cardiovascular risk. Treating comorbid depression may increase the likelihood that women with PCOS can maintain healthy eating habits and adequate physical activity. In turn, exercise and healthy eating can reduce depressive symptoms. This bidirectional improvement has not yet been studied in detail in PCOS, although preliminary results are promising. In a small open trial of eight sessions of CBT for adolescents with PCOS, obesity, and depression, participants had significant reductions in both weight and depressive symptoms.⁸⁴ Similarly, dietary and exercise interventions significantly improved depression scores in women with PCOS who had no other specific antidepressant treatment.⁸⁵

  
  
• 
  

  Insulin sensitizing agents are used in the treatment of women with PCOS in order to improve glucose tolerance and regulate menstrual cycles. The most frequently used agent is metformin. Newer agents are being studied that may have greater neuroprotective properties. A medication of particular interest in this regard is pioglitazone, a thiazolidinedione. In a 6-week double-blind trial comparing metformin to pioglitazone for women with mild to moderate major depressive episodes and PCOS, pioglitazone was superior to metformin in reducing depressive symptoms. In the absence of other treatment, 20% of study participants taking pioglitazone experienced remission of depression, as compared to none in the metformin group.⁸⁶ Further study is needed before pioglitazone could be considered for first-line use, especially studies weighing long-term benefits and risks. Pioglitazone and related agents have been implicated in increased risk of fractures and bladder cancer.

  
  
• 
  

  Hormonal therapies used to treat PCOS include clomiphene to induce ovulation and improve fertility outcomes, intermittent doses of progestins to induce menses and protect the endometrium, and COCs to reduce hyperandrogenism and its associated symptoms.

  
  
• 
  

  Lipid-lowering agents, particularly statins, are used to target elevated (LDL-C). Although some studies have suggested that depression may be a side effect of cholesterol-lowering medications, most data, including a recent meta-analysis, have not supported this link.⁸⁷

  
  
• 
  

  Antidepressants, when used for women (Box 8-8) with PCOS, can be chosen based on a side-effect profile that does not heighten preexisting risk factors. In particular, it is helpful to avoid medications with a higher likelihood of causing weight gain, such as mirtazapine. Among antidepressant augmenting agents, olanzapine is the most likely to contribute to weight gain, cholesterol elevation, and impaired glucose tolerance.⁸⁸

SUMMARY

Women with PCOS are at increased risk for depressive symptoms and major depressive episodes. Untreated depressive symptoms can add to the long-term health risks associated with PCOS, and can impair ability to adhere to treatment. Exercise, healthy eating habits, and maintenance of a normal weight are first-line treatments for PCOS. Treating depression can improve adherence to these lifestyle interventions, and in turn these interventions can alleviate depressive symptoms. Choosing antidepressants that do not increase weight or worsen metabolic abnormalities is important for women with PCOS.

Case outcome. Anne was diagnosed with a major depressive episode of moderate severity. Findings of an elevated free androgen index, anovulatory bleeding patterns, and multiple ovarian cysts on ultrasound confirmed a concomitant diagnosis of polycystic ovary syndrome. Her BMI was 31, consistent with obesity. It became clear that her depressive symptoms had contributed to weight gain and psychomotor slowing, and would interfere with her ability to increase exercise and achieve healthier eating habits. Therefore, her initial treatment plan consisted of bupropion and cognitive behavioral therapy (CBT). As depressive symptoms remitted, Anne was able to establish and maintain a routine of daily aerobic exercise and strength training. She felt less of a drive toward mood-linked overeating, and was able to adhere to healthy eating guidelines recommended by a nutritionist. These focused on mindfulness about eating, and food choices based on glycemic index, rather than on calorie counting. She found this was more effective than the weight-loss diets she had tried in the past. After 9 months of treatment, she had sustained remission of depression, her BMI was 27, and her free androgen index was reduced. She had a higher percentage of ovulatory menstrual cycles. She felt more secure in her marriage, having overcome distorted negative thinking. She felt ready to resume trying to conceive. She decided to taper and discontinue bupropion, but to maintain booster CBT sessions as needed.

EPIDEMIOLOGY

Infertility, the inability to conceive after having regular, unprotected sexual intercourse for a year, affects 6.0% of married women ages 15 to 44 years in the United States according to a National Survey of Family Growth conducted on 3,088 women.⁸⁹ Once pregnant, 10.9% of women in this age group are unable to carry a pregnancy to a live birth (impaired fecundity).⁸⁹ In a direct comparison among women with various medical conditions, psychological symptoms associated with infertility were similar to those related to cancer, hypertension, and cardiac rehabilitation.⁹⁰

Evidence to date suggests that infertility does not increase the risk of major depression, though most studies have found an increased prevalence of subsyndromal depressive symptoms. Rates of moderate to severe depressive symptoms in women with infertility vary considerably from study to study. Nelson et al., for example, found that 19% of infertile women had moderate and 13% had severe depression.⁹¹ Much higher rates of severe depression were found in a study by Drosdzol and Skrzypulec, with 35.4% of infertile women (in comparison to 19.47% of fertile women) scoring above the cut-off for severe symptoms of depression.⁹⁷ Methodological differences, such as varying definitions of infertility and timing of evaluation, make these findings difficult to compare. Rates of depression seem to be especially high in countries where family status and childbearing are highly valued.⁹³