When should surgical treatment be considered for premenstrual dysphoric disorder?

Introduction

Behavioural and psychiatric manifestations in the premenstrual week(s) cover a spectrum from absent or mild (premenstrual molimina) to severe, distressing and potentially disabling (premenstrual syndrome [PMS] or premenstrual dysphoric disorder or PMDD). In the last two decades, progress has been made in establishing definitions and diagnostic criteria and a range of therapeutic approaches have been successfully employed. ¹

Despite these advances, the precise aetiology of premenstrual mood disorders has eluded researchers in the field. An aberrant central nervous system (CNS) response to normal fluctuations in ovarian hormones appears to be the common link, ^2,3 but beyond that the reason as to why some women are affected and others are not remains speculative. Women with PMDD are reported to have an allelic variation on the estrogen receptor [alpha] gene that could account for abnormal estrogen signalling during the luteal phase, in turn leading to premenstrual affective, cognitive and somatic symptoms. ⁴

Successful treatments for PMDD have fallen into four broad categories: (1) lifestyle and counselling approaches, (2) treatments targeted towards specific symptoms (such as anxiolytics), (3) treatments that alter CNS function (such as selective serotonin and norepinephrine re-uptake inhibitors [SSRI, SNRI]) and (4) therapies designed to suppress cyclical ovarian hormonal secretion.

For mild-to-moderate symptoms, especially when menstrual symptomatology is a significant contributor to the overall perception of distress, certain combined hormonal contraceptives (those containing drospirenone) ^5,6 or specific regimens (extended cycle or continuous) ⁷ have proven to be modestly effective first-line therapies. For more severely affected individuals, SSRIs and SNRIs given continuously ⁸ or as luteal phase therapy ⁹ have emerged as the treatments of choice.

There remains a group of individuals who cannot or will not accept these approaches or for whom side-effects preclude ongoing therapy. Side-effects of SSRIs and SNRIs are not uncommon and at times are drug specific. Treatment of PMDD could involve prolonged exposure to SSRIs or SNRIs although luteal phase therapy alone may afford short breaks in treatment. ³ Common complaints among women using SSRIs and SNRIs are nausea, headache, sleep disturbance, diminished libido and anorgasmia. ^10,11 Relationship issues are frequently exacerbated in couples where the woman suffers from PMDD and a satisfactory sexual relationship is often of paramount importance to such women. The incidence of sexual dysfunction with SSRIs is reported to range from 30% to 70%. ¹² The sexual side-effects most often associated with SSRIs in women are inhibited or decreased sexual desire, inhibited sexual excitement, and difficulty in achieving orgasm. ¹³ In women with depression SSRI treatment-emergent sexual dysfunction is a principal reason for poor treatment adherence and treatment non-compliance. ¹⁴

When SSRIs and SNRIs are not acceptable, suppression of cyclical ovarian activity is often the next line of therapy for women with severe PMDD. Gonadotropin-releasing hormone analogues (GnRHa) are highly effective and well tolerated when accompanied by appropriate low-dose hormone therapy. Long-term use of GnRHa and add-back hormone therapy until menopause when cyclical ovarian activity abates naturally is acceptable to many women; however, the costs can be staggering if prolonged use is required. For some women who have experienced the relief from PMDD afforded by medical ovarian suppression, the option of a one-time surgical solution to both PMDD and ancillary menstrual concerns is appealing. This paper will review the evidence for a surgical approach in these cases and will suggest an algorithm for therapy and counselling of potential candidates.

Does surgical therapy work for PMDD?

Unfortunately, the attempt to treat all manner of psychiatric illnesses by removal of uterus or ovaries in the late 1900s and early 20th century is a sad chapter in gynaecological history. ¹⁵ Some have criticized the use of surgery for PMDD as a modern-day version of this overzealous effort to cure a complicated illness with a blunt instrument. ¹⁶ Nevertheless, there are a variety of lines of evidence that support a surgical approach in selected cases.

PMDD is clearly linked to ovarian activity. It does not appear prior to puberty but may onset at any subsequent age when cyclical ovarian activity remains intact. PMDD is absent when ovarian cyclicity disappears during pregnancy or at times of hyper- and hypo-gonadotropic amenorrhea. PMDD disappears at the time of menopause.

There are contradictory reports about the effect of hysterectomy alone on premenstrual symptoms. Several case series report relief for some women but not others when ovarian function was preserved at the time of hysterectomy. ^17–21 Such reports must be interpreted with caution since there was often no clear documentation of the presence or severity of premenstrual symptomatology before surgery. Women could therefore be describing the effect of hysterectomy on the summation of premenstrual and menstrual complaints. It is hardly surprising that elimination of bleeding and dysmenorrhea would result in an overall improvement in the perimenstrual experience.

In the only study with careful pre- and post-hysterectomy monitoring of symptoms and hormonal profiles in women with PMS hysterectomy-alone failed to eliminate premenstrual symptoms. ²² Indeed, typical premenstrual symptoms recurred during the luteal phase as confirmed by the appearance of urinary pregnanediol, a progesterone metabolite.

In 1931, Frank ²³ reported the use of ovarian irradiation to suppress the ovarian cycle in severe PMS. Before the widespread use of GnRHa for medical ovarian suppression other treatments that eliminated ovarian cyclicity (such as Danazol ^24,25 and estradiol implants ²⁶ ) were shown to eliminate premenstrual symptoms.

GnRHa are a means to reversibly suppress ovarian activity resulting in estrogen levels that fall to the menopausal range. The first attempts to use these agents for treatment of PMDD often resulted in less than satisfactory symptom improvement. ²⁷ Cyclical symptoms of PMS were supplanted by continuous symptoms of hypoestrogenism (joint aches, vasomotor symptoms, mood lability, sleep disruption, etc.). With the use of GnRHa and low-dose estrogen add-back therapy to counteract menopausal symptoms, this approach became much more acceptable to women with PMDD. ^28–30

Long-term unopposed estrogen carries the risk of endometrial hyperplasia or malignancy so that any plan for long-term medical ovarian suppression in a woman with an intact uterus requires periodic or continuous progestogen therapy. Progestogens themselves are often associated with adverse effects on mood and the addition of progestogen to standard estrogen replacement has been shown to lead to a recrudescence of premenstrual symptoms in some ³¹ but not all ^32,33 former PMS sufferers.

Three reports with careful documentation of premenstrual symptomatology have evaluated PMS before and after hysterectomy and bilateral salpingo-oophorectomy. ^34–36

Fourteen women with debilitating PMS unresponsive to conventional interventions were treated with Danazol (400 mg daily or higher as needed to suppress ovulation) to determine their response to medical ovarian suppression. ³⁴ All showed a dramatic reduction in premenstrual symptomatology that returned when Danazol was subsequently stopped (Figure 1). Rather than continuing Danazol long term (with its attendant androgenic side-effects) all 14 women elected hysterectomy and bilateral salpingo-oophorectomy with low-dose estrogen replacement. Subsequent prospective symptom monitoring demonstrated dramatic relief from PMS symptoms. OPEN IN VIEWER

This beneficial effect of hysterectomy and removal of ovaries was confirmed in a second study with careful pre- and postoperative symptom monitoring of 14 women with severe PMS – all of whom demonstrated complete ongoing symptom relief while receiving low-dose estrogen replacement therapy. ³⁵

In 2004, a further retrospective evaluation of 47 women treated by hysterectomy and bilateral salpingo-oophorectomy demonstrated complete resolution of symptoms in 94% and women declared that they were satisfied or very satisfied with treatment in 96% of cases. ³⁶

The second report ³⁶ summarizes interviews with 47 PMS sufferers treated by hysterectomy and bilateral salpingo-oophorectomy over a six-year period. Although the precise time since surgery is not stated estrogen continuation remained high (94%) and 96% reported that they were either ‘satisfied’ or ‘very satisfied’ with their treatment and outcome. Eighty-nine percent would recommend the treatment to an affected friend.

Satisfaction rates with hysterectomy in general are high and this is more likely to result with appropriate patient selection and counselling. ^38,39 Surgery is not without potential short- and long-term complications and appropriate counselling is paramount. ³⁹

In addition, counselling about the importance of adherence to estrogen therapy after surgery may be critical to long-term health. Oophorectomy without estrogen replacement is associated with accelerated bone loss and the risk of premature osteoporosis. ⁴⁰ Estrogen replacement effectively counters this bone loss. ⁴¹ Epidemiological studies suggest that premature loss of ovarian function also accelerates coronary artery disease. ⁴² A recent cohort study comparing coronary artery calcium deposition in oophorectomized women with and without estrogen replacement clearly demonstrated that estrogen therapy would counter this premature plaque accumulation. ⁴³

Although many women and their physicians worry about increasing risk of breast cancer with longer-term estrogen therapy, this overlooks the fact that breast cancer risk diminishes significantly with early removal of endogenous ovarian hormone production. ^44,45 One model has predicted that cessation of ovarian hormone production 10 years early could reduce breast cancer incidence by as much as 67% and if 15 years early by as much as 80%. ⁴⁵ Accordingly, the benefit of low-dose estrogen replacement for quality of life as well as bone and heart protection until the natural age of menopause will greatly exceed the risks for the majority of women.

Conclusion

For the severely affected PMS/PMDD sufferer who has completed her family and who cannot tolerate SSRIs or their side–effects, a trial of medical ovarian suppression with low-dose estrogen add-back is both diagnostic and therapeutic. If dramatic relief from symptoms is evident and the treatment is well tolerated then cyclical or continuous progestogen therapy should be incorporated into the hormonal add-back regimen. If the additional progestogen is well tolerated and the woman is in her mid-to-late 40s then medical ovarian suppression may be continued until natural cessation of ovarian cyclicity at menopause. However, for some women the cost or inconvenience of monthly injections to suppress ovarian activity may be prohibitive and a surgical solution should be considered. Optimally this treatment should be reserved for those women who cannot anticipate becoming naturally menopausal within the next five years. Potential candidates for surgical therapy should be counselled about the pros and cons of surgery, the implications of premature menopause and the need for longer-term estrogen replacement.

Competing interests

None declared.