Ellen W. Freeman, PhD

Premenstrual dysphoric disorder (PMDD) is a cyclical illness characterized by mood, behavioral, and physical symptoms that occur exclusively in the luteal phase of the menstrual cycle and remit with the onset of menses. PMDD is differentiated from premenstrual syndrome (PMS) by the number and specificity of symptoms that are required to make the diagnosis. Several studies have focused on the lifetime comorbidity of mood and anxiety disorders in women with PMDD. However, most PMDD studies have examined populations of women without current comorbid psychiatric disorders, making it difficult to estimate the true rate of psychiatric comorbidity in this group.

This article focuses on diagnostic and treatment dilemmas encountered in women with PMDD and psychiatric comorbidity.

PMDD occurs in about 3% to 8% of women with regular ovulatory cycles (

 PMS, on the other hand, is experienced by about 20% of menstruating women who have at least 1 physical or emotional symptom during the 5 days before menses.

 In PMDD, symptoms must not merely be an exacerbation of another disorder,

 which makes understanding and diagnosing this condition in women who complain of PMS and who already have another Axis I disorder a challenge.

Although the term “PMDD” implies a focus on dysphoria, the most common affective symptoms are irritability and tension. PMDD is associated with a level of impairment that is similar to major depressive disorder and poorer quality of life compared with community norms and should therefore be considered a serious health condition.

Despite the diagnostic criteria that comorbid illnesses must be excluded, studies report a higher rate of comorbid mood and anxiety disorders in women with PMDD than in the general population. It is important to distinguish the concept of comorbidity from premenstrual exacerbation (PME) of an underlying Axis I disorder.

The prevalence of PMEs of Axis I disorders is unknown, but it is probably common in women with mood and anxiety disorders. Kornstein and associates

 found that 63% of premenopausal women with major depression (N = 1500) self-reported PME of their symptoms. Cyclic exacerbations occur when the primary symptoms of the Axis I disorder worsen during the luteal phase of the menstrual cycle. This can be difficult to diagnose if the patient is seen monthly at similar phases of her menstrual cycle. Patients may not realize their symptom pattern is menstrually influenced, so it is important to ask directly about symptom and menstrual cycle patterns. Affected patients may be symptomatic throughout the month or, if treated, may have breakthrough symptoms only during the luteal phase. The key concept in PME is that the symptom profile is similar throughout the month.

In contrast with PME, PMDD is diagnosed as a comorbid condition when the specific constellation of symptoms occurs during the luteal phase, ceases in the follicular phase, and does not replicate the comorbid condition’s main symptoms. As an example, PMDD may be diagnosed in a woman who experiences symptoms of depression characterized by sadness, anhedonia, insomnia, and decreased appetite who reports significant irritability and hypersomnia that occur only premenstrually.

Our recent review suggests that unipolar depression is more common in women with PMDD than in controls.

 Among studies in which PMDD was retrospectively diagnosed, depressive disorders (dysthymia, minor depression, and major depression) occurred in 18% to 69% of women with PMDD.

 In studies in which PMDD was prospectively diagnosed, depressive disorders occurred in 12% to 25% of women with PMDD.

 This phenomenon appears to be similar across cultures. In a recent study of 410 Nigerian university students, depression was the most frequent comorbid diagnosis in women with PMDD; it was present in 24% versus 3.6% of women without PMDD (odds ratio [OR], 8.37).

 Although PMDD was diagnosed retrospectively in this cross-sectional study, the researchers made attempts to distinguish between PME and comorbid PMDD.

 found that 57.6% of their community sample who had prospectively received a diagnosis of PMDD had a history of major depressive disorder, which is consistent with other reports. In general, severe dysphoria during the luteal phase suggests a comorbid depressive disorder.

Suicidality is routinely evaluated in patients with depression but not in the assessment of PMDD. However, suicidal ideation may be more frequent in women with PMDD than expected. Yonkers and associates

 reported that of 426 women in 6 outpatient obstetrics clinics with a current psychiatric disorder who reported premenstrual complaints, 24% with possible PMDD reported suicidal thoughts at any level (several days, more than half the days, or every day); and 20% reported suicidal thoughts for at least several days.

No studies have examined rates of PMDD in women with a current diagnosis of depression, but because depression is common in women with PMDD, all women with mood disorders should be assessed for PMS and vice versa. In addition, some women appear to be more sensitive to hormonal fluctuations than other women. This is highlighted by the evidence that suggests that PMDD may be a risk factor for postpartum depression.

 The similarity of the cyclical symptomatology and propensity for hypersomnia and hyperphagia in both seasonal affective disorder (SAD) and PMDD highlights the possibility of an overlapping pathophysiology. In one study, PMDD was found in 46% of women with SAD during a summer remission period and in 2% of women in a control group.

 A possible genetic vulnerability for developing the conditions comorbidly has also been suggested by a recent study that found an increased prevalence of affective disorders in first-degree relatives of patients with both SAD and PMDD compared with patients with SAD alone.

CHECKPOINTS 
? Premenstrual dysphoric disorder (PMDD) occurs  exclusively in the luteal phase of the menstrual cycle and remits with  the onset of menses.

? PMDD is  associated with a level of impairment that is similar to major  depressive disorder and poorer quality of life compared with community  norms and should therefore be considered a serious health condition.

? PMDD is diagnosed as a comorbid  condition when the specific constellation of symptoms occurs during the  luteal phase, ceases in the follicular phase, and does not replicate the  comorbid condition’s main symptoms.

?  In the treatment of patients with PMDD, the SSRIs should be considered  first-line, with various dosing strategies to increase efficacy.


The exact nature of the relationship between PMDD and depression is still uncertain, but both disorders have been linked to serotonin (5-HT) dysregulation. In women with PMDD, whole blood 5-HT levels and platelet uptake of 5-HT are decreased.

 In addition, PMS is exacerbated by the depletion of tryptophan, the precursor to 5-HT.

Animal experiments suggest that 5-HT has inhibitory effects on aggression and irritability-predominant symptoms of PMS. For example, fluoxetine reduced aggressive behavior elicited by estradiol and progesterone in ovariectomized rats.

These animal models suggest links between serotonergic transmission and cycle-dependent behaviors that are associated with steroid hormones. The main evidence that links PMDD and unipolar depression to 5-HT dysregulation is the efficacy of the serotonergic antidepressants.

While PMDD is not relieved by nonserotonergic antidepressants, depression may be. Whether women who respond to nonserotonergic antidepressants are less likely to suffer from PMS is not known but would be of interest.

Like mood disorders, our recent review of the data suggests that anxiety disorders are more common in women with PMDD than in controls.

 The evidence suggests that panic disorder co-occurs in 25% of women with PMDD. Social phobia is comorbid in approximately 20% of women with PMDD, and obsessive-compulsive disorder is comorbid in 12%. Rates of comorbid PMDD and generalized anxiety disorder range from 4% to 38%.

There appears to be a relationship between trauma and PMDD. In a sample of 42 women with prospectively assessed PMS, 95% reported at least 1 sexual abuse event.

 In a longitudinal community study of 1251 women with retrospectively diagnosed PMDD, traumatic events were a stronger predictor for the onset of PMDD than depressive disorders (OR, 5.1).

 Although the biggest predictor of a later diagnosis of PMDD was subthreshold PMDD symptoms, previous trauma and a diagnosis of an anxiety disorder were also factors. Women with a history of trauma should be questioned about PMS carefully.

There may be a physiological overlap between PMDD and anxiety disorders. The rate and intensity of carbon dioxide–induced panic attacks are similar in patients with panic disorder and PMDD, while patients with depression have panic attacks similar to controls.

 Differences in g-aminobutyric acid (GABA) plasma levels in women with PMDD have been found, and serial proton magnetic resonance spectroscopic measurements show decreased cortical GABA levels during the follicular phase in women with PMDD compared with controls.

 Allopregnanolone-a metabolite of progesterone-has been implicated in the regulation of the GABAA receptor and may provide a link between neurosteroid function and anxiety symptoms.

While PME has been reported in women who have schizophrenia, there are few data to support a link between these 2 disorders. Moreover, studies do not support a strong association between PMDD and bipolar disorder.

As we have highlighted, it is important to elucidate the cyclicity of mood and anxiety symptoms in women. Inquire about the regularity of menstrual cycles, use of hormonal contraception, and use of herbal remedies during the diagnostic interview. If cyclicity is reported, prospective symptom charting should be prescribed, in which the patient must record her daily symptoms for 2 consecutive months.

There are numerous daily symptom rating charts available to help patients record their PMDD symptoms (eg, Daily Symptom Report).

 If there is a 50% change between the premenstrual and postmenstrual symptoms, a diagnosis of comorbid PMDD should be considered. Patients who have PME of their underlying Axis I disorder should be evaluated in a similar manner, even though there are no validated daily rating scales to support the diagnosis. However, clinical studies have used the Beck Depression Inventory and Hamilton Depression Rating Scale; elevated luteal phase scores indicate PME.

The serotonergic antidepressants, particularly the SSRIs, are the first-line treatment for severe PMS and PMDD. In their expert guidelines, Steiner and colleagues

 recommend the following treatment for PME of depression or anxiety:

• Treat the underlying Axis I disorder with full antidepressant dosages.

• If luteal symptoms persist, increase the dosage of the antidepressant for the whole cycle.

• Alternatively, switch to semi-intermittent dosing by maintaining the follicular dosage and then increasing it during the luteal phase only.

A recent study of 9 women with PME of menstrual dysphoric disorder (MDD) showed that variable antidepressant dosing could be effective.

 In this study, the sertraline dosage was increased during the luteal phase by 50% from the follicular phase dosage.

In 5 patients with both PMDD and MDD, cyclical irritability and physical symptoms persisted during treatment with a tricyclic antidepressant.

 In a study of whole-cycle (continuous) versus luteal phase (intermittent) treatment with sertraline, there was a significant decrease in Hamilton Rating Scale for Depression scores in the subgroup that had both PMDD and depression.

Although it is unlikely that PMDD is just a variant of depression, one might suppose that serotonergic antidepressants could address both the PMDD and major depressive symptoms. However, more controlled studies are needed. While patients with PMDD generally respond to dosages of SSRIs at the lower end of the effective range, it is our clinical experience that when there are comorbid Axis I disorders, higher dosages should be prescribed. If a patient is currently taking a nonserotonergic antidepressant, augmentation with an SSRI or switching to an SSRI should be considered.

There are no studies that examine maintenance treatment in women with PMDD and a comorbid psychiatric illness. PMDD is a chronic, recurrent disorder, so it can be assumed that long-term treatment is indicated. Further studies are warranted.

Alprazolam, a benzodiazepine, and buspirone, a 5-HT

 agonist, have shown modest efficacy in some, but not all, PMS studies. Berger and Presser

 found a poorer response to alprazolam in women with follicular phase anxiety and depression than in those with only luteal phase symptoms. Women took the medication regularly during the luteal phase and on an as needed basis during the follicular phase.

If a woman continues to experience significant premenstrual anxiety, despite treatment with an SSRI, luteal phase treatment with an anxiolytic can be considered.

Estradiol affects 5-HT synthesis and reuptake. Pharmacological suppression of ovulation has been shown to decrease symptoms in women with PMS. Young and colleagues

 reported in the analysis of the STAR*D that 1238 women with depression who took a combined oral hormonal contraception (OCP) report less severe depression. Yet, many women reported increased mood and anxiety symptoms while taking synthetic progestins; this needs to be carefully considered in women with comorbid Axis I disorders. OCPs in general do not improve PMS symptoms, although there may be benefit to the FDA-approved use of the OCP that contains drosperinone 3 mg and ethinyl estradiol 20 

g (DRSP/EE) or extended-cycle dosing (fewer hormone-free intervals). In an open-label trial, Joffe and associates

 found that an OCP that contains DRSP/EE given to women with prospectively diagnosed PME of depression was effective. Again, the possible effects of OCPs on women with an underlying mood disorder need to be considered.

Among the most effective hormonal treatments for PMS suppression is the gonadotropin-releasing hormone (GnRH) agonist leuprolide, which is administered by monthly injection in depot form, and intranasal buserelin. GnRH agonists decrease pituitary release of follicle-stimulating hormone and luteinizing hormone, thereby suppressing ovulation. One key note about the treatment of comorbid depression and PMDD is that the GnRH agonists are less effective when there is an underlying depressive disorder.

 In addition, GnRH agonists can be difficult for patients because of the resulting hypoestrogenism. Adverse effects of GnRH treatment include hot flashes, headaches, and osteoporosis.

reported a significant increase in Beck Depression Inventory scores in 26 previously asymptomatic women after 4 weeks of treatment with leuprolide. The mood-related adverse effects of a GnRH agonist should be monitored closely in all women; this is especially important in women who are predisposed to mood disorders.

Calcium carbonate reduces premenstrual depression, fatigue, edema, and pain significantly more than placebo in women with PMS.

Although there are no data in women with Axis I comorbidity and PMDD, calcium is a reasonable first-line choice for women with mild PMS symptoms or as an adjunct for women with moderate to severe PMS. Calcium should be used carefully in patients with constipation or renal stones.

 describes the steps to diagnose PMDD or PME in the context of another psychiatric disorder.

In patients with PMDD there is a 30% to 70% and a 14% to 16% lifetime risk of major depression and anxiety, respectively. Patients with PMDD need to be screened for depression and anxiety and vice versa. There are few data to guide treatment in these patients, but the use of SSRIs should be considered first-line with various dosing strategies to increase effectiveness.

Articles

Diagnostic Dilemmas-Effective Treatment Approaches

Premenstrual Dysphoric Disorder and Psychiatric Comorbidity

Although the exact etiology of premenstrual syndrome (PMS) remains unknown, it appears to be a chronic mood disorder that continues for many years in reproductive-age women. Approximately 2% to 10% of cycling women report severe symptoms resulting in a disruption of work or relationships (Logue and Moos, 1986). Identifying appropriate treatments and the optimal dosing of medications is important for the large number of women who experience marked disruptions of their work relationships, productivity and quality of life because of distressing premenstrual symptomatology.  

Research has shown that serotonergic antidepressants taken daily throughout the menstrual cycle are effective for severe PMS. This efficacy is based on continuous daily dosing of the antidepressant medications, thus treating the asymptomatic phase as well as the symptomatic premenstrual phase of the cycle.

Several preliminary studies discussed in this article have reported that half-cycle dosing of a serotonergic antidepressant in the symptomatic premenstrual phase is effective for severe PMS and its 

 diagnosis of premenstrual dysphoric disorder (PMDD). These results suggest that PMS/PMDD patients do not require continuous daily dosing. The findings are important because intermittent dosing may spare PMS patients unnecessary medication and reduce the risks and costs of treatment.

The most recent report of intermittent dosing for premenstrual symptoms comes from the University of Pennsylvania PMS Program (Freeman et al., 1999). This study used sertraline (Zoloft), a selective serotonin reuptake inhibitor (SSRI), in a flexible dosing schedule and compared half-cycle with full-cycle dosing regimens. More subjects improved in half-cycle treatment than in full-cycle treatment. These preliminary data add support to other pilot studies reporting the efficacy of sertraline, clomipramine (Anafranil) and fluoxetine (Prozac) administered only premenstrually to PMS patients who had no other comorbid disorders. 

The women in the University of Pennsylvania Study fulfilled criteria for severe PMS when they entered a preceding three-arm treatment trial, in which daily administration of two antidepressant medications and a placebo were compared. At the end of this short-term treatment trial, the double-blind protocol was not compromised. Thirty-one volunteers, half of whom were improved and half of whom were unimproved, were randomized in double-blind fashion to receive either full-cycle or half-cycle sertraline for three additional months. All study participants took capsules daily, with the half-cycle treatment group receiving placebo capsules in the first two weeks of the cycle, both to control for placebo effects and to maintain the double-blind protocol.   

The results showed that more subjects improved with half-cycle treatment as compared with full-cycle treatment (89% versus 46%, respectively; p-02). Response to the previous treatment was not associated with response in the half-cycle dosing study. Of the subjects who were improved when they entered the second study, 88% (14 of 16) remained improved for the study duration, regardless of whether the sertraline was taken throughout the cycle or only premenstrually. Of the subjects who were unimproved with previous treatment, 53% (8 of 15) improved. The subjects who were unimproved at the outset either improved swiftly in the first month of treatment or remained unimproved throughout the study.  A greater percentage of initially unimproved subjects improved with half-cycle dosing than with full-cycle dosing (80% versus 40% respectively), though the small numbers in this comparison were not statistically significant.

The sertraline dose was 50 mg/day (one capsule) in the first month. In the absence of improvement or dose-limiting side effects, the dose was raised to 100 mg/day (two capsules) for the second month and to 150 mg/day (three capsules) for the third month of treatment. The mean sertraline dose at the end of three months was 80 mg/day in the half-cycle group and 100 mg/day in the full-cycle group-a nonsignificant difference. 

Side effects were reported by 81% (25 of 31) of the subjects; there were no statistically significant differences between the full- and half-cycle dosing groups. The side effects (nausea, dry mouth, headache, dizziness, fatigue and decreased libido) were typically transient and diminished in the second and third months of treatment. No subject reported symptoms after discontinuation of sertraline, though the half-cycle dosing group discontinued active drug on day 3 and resumed active drug on day 14 in each treatment cycle. 

These results of half-cycle dosing with sertraline for premenstrual symptoms are consistent with recent findings reported for the SSRI citalopram (Celexa) (Wikander et al., 1998). Intermittent treatment of premenstrual dysphoria with citalopram was more effective than placebo, and intermittent treatment also appeared more effective than continuous or semi-intermittent administration of the drug.

Two other studies reported that intermittent treatment with sertraline was superior to placebo for dysphoric premenstrual symptoms. Sertraline administered for two weeks premenstrually produced improvement significantly greater than placebo, and appeared equivalent in efficacy to sertraline given during the entire menstrual cycle (Halbreich and Smoller, 1997). In this fixed-dose study using 100 mg/day of sertraline, only subjects who improved after a month of daily sertraline were randomized to intermittent dosing.

Another fixed-dose study examined the response to half-cycle sertraline without first establishing improvement with daily treatment (Young et al., 1998). Again, sertraline was significantly better than the placebo, with 50 mg sertraline administered in a randomized, double-blind, placebo-controlled crossover design for 14 days before menses.

Successful intermittent administration of an antidepressant for premenstrual symptoms was first reported for clomipramine, a potent but nonselective serotonin reuptake inhibitor, at daily doses of 25 mg to 75 mg (Sundblad et al., 1993). Twenty-nine subjects were randomized for three months of treatment with either clomipramine or placebo for 14 days before menses. Treatment response was significantly better with clomipramine than the placebo. This intermittent treatment trial demonstrated a short onset of action when using clomipramine for premenstrual symptoms.

An open study of fluoxetine at 20 mg/day compared continuous and intermittent treatment for three menstrual cycles (Steiner et al., 1997). Seventy-one percent of the subjects improved. Again, more women responded in the intermittent dosing group than in the continuous dosing group, though the difference was small and not statistically significant.

More research and investigation is necessary, as all intermittent dose studies thus far have been preliminary. Some have investigated intermittent dosing only after establishing a treatment response with daily dosing. Some have compared intermittent dosing to continuous dosing, while other studies have compared only half-cycle dosing to placebo. Nonetheless, the consistent results suggest that half-cycle dosing with an SSRI can control PMS symptoms.

Which women will respond to half-cycle dosing and which will require continuous dosing is not known. However, it should be noted that the existing information on half-cycle dosing is based entirely on women who were free of premenstrual symptoms in the follicular phase of the cycle and did not have other comorbid diagnoses.

The psychopharmacologic reasons for the possible efficacy of intermittent dosing are not known. Possibly SSRIs have a more rapid onset of action when used to treat PMS, i.e., several days rather than the one to four weeks when used to treat depression. The mechanisms underlying PMS may differ from those of depression, the model for studies of serotonergic actions. The predominant PMS symptoms such as irritability and mood swings are not exclusively depressive symptoms, and the syndrome frequently includes anxiety and physical symptoms. The onset (and remission) of PMS symptoms is typically rapid, rather than gradual as in depression.

Nonetheless, evidence is accumulating that alterations in serotonin transmission may be important in PMS. For example, administration of tryptophan, an essential amino acid that serves as a dietary precursor for serotonin, and d-fenfluramine, which stimulates serotonin neurotransmission, reduced symptoms in PMS subjects. Alternatively, depletion of tryptophan aggravated PMS symptoms. In addition, short-term improvement of PMS subjects in response to administration of the serotonin agonist meta-chlorophenylpiperazine suggested that dysregulation of either the hypothalamic-pituitary-adrenal (HPA) axis or serotonin control of the HPA axis may be associated with the symptoms of PMS (Su et al., 1997). 

These preliminary studies of half-cycle dosing for PMS patients suggest that administration of a serotonergic antidepressant in the symptomatic half of the menstrual cycle may effectively control premenstrual symptoms. This is an important clinical issue for patients who might control symptoms with less medication, for medical costs, and for an increased understanding of PMS mechanisms and its relationship with depression and other mood disorders. Additional well-designed studies with sample sizes sufficient for definitive and statistically significant results are needed to support or refute these intriguing findings.

Freeman EW, Rickels K, Arredondo R et al. (1999), Full- or half-cycle treatment of severe premenstrual syndrome with a serotonergic antidepressant. J Clin Psychopharmacol 19(1):3-8.
 

Halbreich U, Smoller JW (1997), Intermittent luteal phase sertraline treatment of dysphoric premenstrual syndrome. J Clin Psychiatry 58(9):399-402.
 

Logue CM, Moos RH (1986), Perimenstrual symptoms and risk factors. Psychosom Med 48(6):388-414.Steiner M, Korzekwa M, Lamont J et al. (1997), Intermittent fluoxetine dosing in the treatment of women with premenstrual dysphoria. Psychopharm Bull 33(4):771-774.
 

Su TP, Schmidt PJ, Danaceau MA et al. (1997), Fluoxetine in the treatment of premenstrual dysphoria. Neuropsychopharmacol 16(5):346-356.
 

Sundblad C, Hedberg MA, Eriksson E (1993), Clomipramine administered during the luteal phase reduces the symptoms of premenstrual syndrome: a placebo-controlled trial. Neuropsychopharmacol 9(2):133-145.
 

Wikander I, Sundblad C, Andersch B et al. (1998), Citalopram in premenstrual dysphoria: Is intermittent treatment during luteal phases more effective than continuous medication throughout the menstrual cycle? J Clin Psychopharmacol 18(5):390-398.
 

Young SA, Hurt PH, Benedek DM, Howard RS (1998), Treatment of premenstrual dysphoric disorder with sertraline during the luteal phase: a randomized, double-blind, placebo-controlled crossover trial. J Clin Psychiatry 59(2):76-80.

Although the exact etiology of premenstrual syndrome (PMS) remains unknown, it appears to be a chronic mood disorder that continues for many years in reproductive-age women. Approximately 2% to 10% of cycling women report severe symptoms resulting in a disruption of work or relationships.

Ellen W. Freeman, PhD

Articles

Premenstrual Dysphoric Disorder and Psychiatric Comorbidity

Comparing Full- and Half-Cycle Treatment of Serotonergic Antidepressants for Severe PMS