Prof. John Studd. Women's Health Clinic
clinical gynaecologist
clinical gynaecologist

print this page

Hormones & Depression
in women


August 2005

John Studd, DSc, MD, FRCOG
Consultant Gynaecologist, Chelsea & Westminster Hospital, London

On Boxing Day 1851, Charles Dickens attended the patients Christmas dance at St Luke's Hospital for the insane. On describing his visit in an article for Household Words,. He commented that the experience of the asylum proved that insanity was more prevalent amongst women than men. Of the 18,759 inmates over the century, 11,162 had been women. He adds "it is well known that female servants are more frequently affected lunacy than any other class of persons."

Charles Dickens was as great an observer as any Nobel prize winner and indeed this passage is one of the very few references in Victorian literature that make the point between gender and depression but there are none to my knowledge relating reproductive function to depression. Jane Eyre's red room and Berthe Mason's monthly madness may be coded examples of this from Charlotte Bronte's pen. Modern epidemiology confirms that depression is more common in women than men whether we look at hospital admissions, population studies, suicide attempts or the prescription of antidepressants (1). The challenge remains to determine whether this increase in depression is environmental reflecting women perceived role in contemporary society or whether it is due to hormonal changes.

It is clear that this excess of depression in women starts at puberty and is no longer present in the 6th and 7th decade. The peaks of depression occur at times of hormonal fluctuation in 1) the premenstrual phase, 2) the postpartum phase and 3) the climacteric perimenopausal phase, particularly in the one or two years before the periods cease. This triad of hormone responsive mood disorders, (HRMD) often occur in the same vulnerable woman.

The 45 year old depressed peri-menopausal woman who is still menstruating will often give a history of previous postnatal depression and depression before periods. They will often be in very good mood during pregnancy and also have systemic manifestations of hormonal fluctuation in the form of menstrual headaches or menstrual migraine. Such a woman will often say that she last felt well during her last pregnancy. She then developed post natal depression for several months. When the periods returned, the depression became cyclical and as she approached the menopause the depression became more constant. Reproductive events also affect the course of bipolar disorder in women. 67% of such women had a history of postpartum depression. Of these, all had episodes of depression after subsequent pregnancies. Subsequently women who were not using hormone replacement therapy were significantly more likely than those who were using HRT to report worsening of the depression symptoms during the perimenopause/menopause. The authors Freeman et al conclude that hormonal fluctuations are associated with increased risk of affective disregulation and mood episodes in women with bipolar disorders.

The depression of these patients can be usually treated effectively by oestrogens, preferably by the transdermal route and in a moderately high dose. Transdermal oestrogen patches of 200 mcgs has been the dose used in published placebo controlled studies but the 100 mcg dose is frequently effective.

The problem with this, (to me), clear clinical history of a woman who will probably respond to oestrogens is that the scientists believe that such patients are ideal for the use of antidepressants. This is because they would recognise that they would have had premorbid history of depression and therefore they would have chronic relapsing depressive illness. The fact that this depression is postnatal or premenstrual in timing escapes this. It is sad that both gynaecologists and psychiatrists are victims and products of their own training with too little overlap in knowledge.

The clue to the use of oestrogens came with the important and somewhat eccentric paper by Klaiber (2) who performed the placebo controlled study of very high dose oestrogens in patients with chronic relapsing depression. They had various diagnoses and were both premenopausal and postmenopausal. They were given Premarin 5 mgs daily with an increase in dose of 5 mg each week until a maximum of 30 mg a day was used. There was a huge improvement in depression on these high doses, (figure 1), but this work has not been repeated because of anxiety over high dose oestrogens.

This condition is mentioned in the fourth century BC by Hippocratic but became a medical epidemic in the nineteenth century. Victorian physicians were aware of menstrual madness, hysteria, chlorosis, ovarian mania, as well as the commonplace neurasthenia. In the 1870's Maudsley(3), the most distinguished psychiatrist of the time, wrote "…The monthly activity of the ovaries which marks the advent of puberty in women has a notable effect upon the mind and body; wherefore it may become an important cause of mental and physical derangement…" This and other female maladies were recognised, rightly or wrongly, to be due to the ovaries. As a consequence bilateral oophorectomy - (Battey's operation(4)) - became fashionable, being performed in approximately 150,000 women in North America and Northern Europe in the 30 years from 1870. Longo(5), in his brilliant historical essay on the decline of Battey's operation, posed the question whether it worked or not. Of course they had no knowledge of osteoporosis and the devastation of long-term oestrogen deficiency, therefore, on balance the operation was not helpful as a long-term solution but it probably did, as was claimed, cure the "menstrual/ovarian madness" which would be a quaint Victorian way of labelling severe PMS. The essential logic of this operation was to remove cyclical ovarian function but happily this can now effectively be achieved by simpler medical therapy. Only in 1931 was the phrase 'premenstrual tension' introduced by Frank(6), who described 15 women with the typical symptoms of PMS as we know it. Greene and Dalton extended the definition to 'premenstrual syndrome' in 1953 (7), recognising the wider range of symptoms.

Severe premenstrual syndrome (PMS) is a poorly understood collection of cyclical symptoms, which cause considerable psychological and physical distress. The psychological symptoms of depression, loss of energy, irritability, loss of libido and abnormal behaviour as well as the physical symptoms of headaches, breast discomfort and abdominal bloating may occur for up to 14 days each month. There may also be associated menstrual problems, pelvic pain, menstrual headaches and the woman may only enjoy as few as 7 good days per month. It is obvious that the symptoms mentioned can have a significant impact on the day-to-day functioning of women. It is estimated that up to 95% of women have some form of PMS but in about 5% of women of reproductive age they will be affected severely with disruption of their daily activities. Considering these figures it is disturbing that many of the consultations at our specialist PMS clinics start with women saying that for many years they have been told that there are no treatments available and that they should simply "live with it". In addition many commonly used treatments of PMS particularly progesterone or progestogens have been shown by many placebo controlled trials not to be effective. In fact they commonly make the symptoms worse as these women are progesterone or progestogen intolerant.

The exact cause is uncertain but fundamentally it is due to the hormonal or biochemical changes, (whatever they are with ovulation), and the resulting complex interaction between ovarian steroids and neuro-endocrine factors that occur with ovulation. This combination produces these varied symptoms in women who are somehow vulnerable to changes in their normal hormone levels. These cyclical chemical changes, probably due to progesterone or one of its metabolites, produce the cyclical symptoms of PMS.

PMS does not occur if there is no ovarian function (8). Obviously, it does not occur before puberty or after the menopause or after oophorectomy. It also does not occur during pregnancy. However, it is important to realise that hysterectomy with conservation of the ovaries does not often cure PMS (9) as patients are left with the usual cyclical symptoms and cyclical headaches. This condition, best-called "the ovarian cycle syndrome" (10) is usually not recognised to be hormonal in aetiology, as there is no reference point of menstruation. The failure to make this diagnosis is regrettable because these monthly symptoms of depression, irritability, mood change, bloating and headaches which might affect the women for most days in the month with only perhaps a good week each month can easily be treated with transdermal oestrogens which suppress ovarian function and thus remove the symptoms.

A medical Battey's operation can be achieved by the use of GnRH analogues and Leather et al (11) have demonstrated that three months of Zoladex therapy cures all of the symptom groups of PMS. The women do, of course, have hot flushes and sweats but these are usually far preferable to the cyclical depression, irritability and headaches. The long-term risk of Zoladex therapy is bone demineralisation but the same group showed that add-back with a product containing 2 mgs of oestradiol valerate and cyclical levonorgestrel (Nuvelle, Schering Health) maintain the bone density at both the spine and the hip (12). Most of the PMS symptoms remain improved with this "add-back" but bloating, tension and irritability recur - probably due to the cyclical progestogen. Livial may be a better add-back preparation. In a Scandinavian study, Sundstrom and colleagues used low-dose GnRH analogues (100 µg buserelin) with good results on the symptoms of PMS, but the treatment still caused anovulation in as much as 56% of patients (13). Danazol is another method to treat PMS by inhibiting pituitary gonadotrophins, but it has side-effects including androgenic and virilising effects. When used in the luteal phase only (14) it only relieved mastalgia but not the general symptoms of PMS even though side-effects were minimal.

Greenblatt et al showed the effects of an anovulatory dose of oestrogen implants for the use of contraception (15) and the first study for its use in PMS was by Magos et al (16) using 100 mg oestradiol implants, the dose that had been shown to inhibit ovulation by using ultrasound and day 21 progesterone measurements in earlier studies by the same group. This showed a huge 84% improvement with placebo implants but the improvements of every symptom cluster was greater in the active oestradiol group (Figure 1). In addition the placebo effect usually waned after a few months compared with a continued response to oestradiol. These patients, of course, were also given 12 days of oral progestogen per month to prevent endometrial hyperplasia and irregular bleeding (17). It was clear that the addition of progestogen attenuated the beneficial effect of oestrogen. Subsequently a placebo-controlled trial of cyclical norethisterone in well oestrogenised hysterectomised women reproduced the typical symptoms of PMS (18). This study of cyclical oral progestogen in the oestrogen primed woman is the model for PMS. It is also significant that progestogen intolerance is one of the principal reasons why older, post-menopausal women stop taking HRT (19), particularly if they have a past history of PMS or progesterone intolerance. It is common for progestogens to cause PMS-like symptoms in these women in the same way endogenous cyclical progesterone secretion is the probable fundamental cause of premenstrual syndrome.

Our group still uses oestradiol implants, often with the addition of testosterone for loss of energy and loss of libido, in our PMS clinics but we have reduced the oestradiol dose, never starting with 100 mgs. We will now insert pellets of oestradiol 50 mgs or 75 mgs with 100 mgs of testosterone. These women must have endometrial protection by oral progestogen or a Mirena (Schering Healthcare) levonorgestrel-releasing intra-uterine system (LNG IUS).(20) As women with PMS respond well to oestrogens but are often intolerant to progestogens and it is therefore common-place for us to reduce the orthodox 13 day course of progestogen to 10 or 7 days starting, for convenience, on the first day of every calendar month. Thus, the menstrual cycle is reset.

The Mirena IUS also plays a vital role in preventing PMS-like symptoms as it performs its role of protecting the endometrium without systemic absorption. A recent study has shown a 50% decrease in hysterectomies in our practice since the introduction of the Mirena IUS in 1995,(17). With its profound effect on menorrhagia and the possibility of less progestogenic side-effects, Mirena looks a very promising component of PMS treatment in the future.

Hormone implants are not licensed in all countries and are unsuitable for women who may wish to easily discontinue treatment in order to become pregnant. Oestradiol patches are an alternative and our original double-blind cross-over study used 200 mcgs of oestradiol patch twice weekly (21). This produced plasma oestradiol levels of 800 pmol/l and suppressed luteal phase progesterone and ovulation. Once again this treatment was better than placebo in every symptom cluster of PMS. Figure 4 shows the response to oestradiol treatment and placebo in a six month cross-over study. This is now our treatment of choice in severe PMS.

Subsequently a randomised but uncontrolled observational study from our PMS clinic indicated that PMS sufferers could have the same beneficial response to 100 mcg patches as they do with the 200 mg dose. They also have fewer symptoms of breast discomfort, bloating and there is less anxiety from the patient or general practitioner about high dose oestrogen therapy.(22) 21 day progesterone assays in the patients receiving 100 mcgs showed low anovulatory levels prompting the intriguing question that even this moderate dose might reliably suppress ovulation and be contraceptive. Clearly, a great deal of work must be done before we can suggest that this treatment is effective birth control but it is of great importance because many young women on this therapy for PMS will be pleased if it also was an effective contraceptive. This is a study which needs to be conducted.

The original studies outlined in this paper are all scientifically valid placebo-controlled trials showing a considerable improvement in PMS symptoms with oestrogens. Although this treatment is used by most gynaecologists in the United Kingdom, its value has not been exploited by psychiatrists anywhere in the world. We believe that the benefit of this therapy in severe PMS is due to the inhibition of ovulation but there is probably also a central mental tonic effect. Klaiber et al (1) in his study of high dose Premarin showed this and our other psycho-endocrine studies of climacteric depression(23)and post-natal depression(24) have shown the benefit of high dose transdermal oestrogens for these conditions.

Ultimately there are some women who, after treatment with oestrogens and Mirena coils will prefer to have a hysterectomy in order to remove all cycles with a virtual guarantee of improvement of symptoms. This should not be seen as a failure or even treatment of last resort as it does carry many other advantages (25).

It is important that these women who have had a hysterectomy and bilateral salpingo-oophorectomy have effective replacement therapy, ideally with replacement of the ovarian androgens. Implants of oestradiol 50 mgs and testosterone 100 mgs are an ideal route and combination of hormones for this long-term therapy post-hysterectomy with a continuation rate of 90% at 10 years (17). We have a study of 50 such patients who have had a hysterectomy, bilateral salpingo-oophorectomy and implants of oestradiol and testosterone for severe PMS who have gone through many years of treatment with transdermal oestrogens and cycle progestogens or Mirena coil. The symptoms are removed in all patients and all but one was very satisfied with the outcome.

Postnatal depression is another example of depression being caused by fluctuations of sex hormones and having the potential to be effectively treated by hormones. It is a common condition which affects 10-15% of women following childbirth and may persist for over one year in 40% of those affected. There does seem to be a lack of any overall influence of psychosocial background factors in determining vulnerability to this postpartum disorder although it can be recurrent.

Although common, the disease is often not reported to the health care professional, particularly the general practitioner or the visiting midwife as the exhaustion and depression is regarded as normal. Indeed the symptoms of postnatal depression may be confused with the normal sequelae of childbirth. The symptoms can consist of depressed mood with lack of pleasure with the baby or any interest in her surroundings. There may be sleep disturbance, either insomnia or hypersomnia. There may be loss of weight, loss of energy and certainly loss of libido together with agitation, retardation and feelings of worthlessness or guilt. Frequent thoughts of death and suicide are common.

Postnatal depression is not more common after a long labour, difficult labour, caesarean section, separation from the baby after birth, nor is it determined by education or socio-economic group. The only environmental factor which seems to be important is the perceived amount of support given by the partner. There is no doubt that the first 6 or more months after delivery can be an exhausting time, full of anxiety and insecurity in mothers with the new responsibility of the baby. Even allowing for that, there does seem to be a clear hormonal aspect to this condition.

Postnatal depression is severe and more prolonged in women who are lactating and lower oestradiol levels are found in depressed women following delivery than with controls. It is probable that the low oestradiol levels with breast feeding and the higher incidence of depression are related in a causative way.

We studied the effect of high dose transdermal oestrogens in this condition in an attempt to close the circle of studies treating this triad of hormone responsive depressions - premenstrual depression climacteric depression and postnatal depression. This was a double blind placebo controlled trial of 60 women with major depression which began within 3 months of childbirth and persisted for up to 18 months postnatally.(24) They had all been resistant to antidepressants and the diagnosis of postnatal depression was made by two psychiatrists who are expert in the field. We excluded breast feeding women from the study. They were given either placebo patches or transdermal oestradiol patches 200 mcgs daily for 3 months without any added progestogen. After 3 months, cyclical Duphaston 10 mgs daily was added for 12 days each month. The women were assessed monthly be a self-rating of depressive symptoms on the Edinburgh postnatal depression score, "EPDS" and by clinical psychiatric interview. Both groups were severely depressed with a mean EPDS score of 21.8 before treatment. During the first month of therapy the women who received oestrogen improved rapidly and to a greater extent than controls. None of the other factors, age, psychiatric, obstetric or gynaecological history, severity and duration of current episode of depression and concurrent antidepressant medication influenced the response to treatment.

The study showed that the mean EPDS score was less with the active group at one month and then maintained for eight months and that the percentage with EPDS scores over 14, (diagnostic of postnatal depression) was reduced by 50% at one month and 90% at 5 months. This was much better than the placebo response. Fig

Not only did this study show that transdermal oestrogens were effective for the treatment of postnatal depression but a subsequent study by Lawrie et al(26) showed that depot-progestogen was worse than placebo causing deterioration in the severity of postnatal depression. Thus we have again the picture of the mood elevating effect of oestrogens and the depressing effect of progestogen.

An uncontrolled study showed similar improvements using sublingual oestradiol in 23 women with major postnatal depression. (26) These women had plasma levels of 79.0 pmol/l before the treatment with sublingual oestradiol. The oestradiol levels were 342 pmol/l at one week and 480 pmol/l at 8 weeks.. There was improvement in 12 out of the 23 patients at one week and after two weeks there was recovery in 19 of the 23 patients.

The mean Montgomery Asberg depression rating scale was 40.7 before treatment, 11 at one week and 2 at eight weeks. At the end of the second week of treatment the MADRS scores were compatible with clinical recovery in 19 out of the 23 patients. This study stressed the rapidity of response to the oestradiol therapy and this was our observation also. However, it must be stressed that this is an uncontrolled study in women with a very low almost postmenopausal level of oestradiol. Another placebo controlled study is required together with information about bleeding patterns to support or refute our original paper.(24)

It would support the hormonal pathogenesis of this condition if we could mimic postnatal depression by hormonal manipulation. This was done by a study by Bloch,(27) who studied 16 women, 8 with a history of postnatal depression. They induced hypogonadism with leuprolide acetate and stimulated pregnancy by "add back" supraphysiological doses of oestradiol and progesterone for 8 weeks and then withdrawing both steroids. Five of the eight women, 62.5% with a history of postnatal depression and none of the women without a prior history developed significant mood symptoms during the withdrawal period.

This study supported the view that there was an involvement of the reproductive hormones, oestrodiol and progesterone in the development of postpartum depression in a set group of women. Furthermore, the study showed that women with a history of postpartum depression are differentially sensitive to the mood destabilising effects of gonadal steroids.

Like many aspects of depression in women, the diagnosis of climacteric depression and its treatment remains controversial. Whereas gynaecologists who deal with the menopause have no difficulty in accepting the role of oestrogens in the causation and the treatment of this common disorder, psychiatrists seem to be implacably opposed to it. This may be because there is no real evidence of an excess of depression occurring after the menopause, nor any evidence that oestrogens help postmenopausal depression or what used to be called "involutional melancholia". This is quite true and indeed many women with longstanding depression improve considerably when the periods stop. This is because the depression created by premenstrual syndrome, heavy painful periods, menstrual headaches and the exhaustion that attend excess blood loss disappears. Therefore, the longitudinal studies of depression carried out by many psychologists, particularly those as notable as Hunter(28), have shown no peak of depression in a large population of menopausal women. The depression that occurs in women around the time of the menopause is at its worst in the two or three years before the periods stop. This, of course, is perimenopausal depression and is no doubt, related to premenstrual depression as it becomes worse with age and with falling oestrogen levels.

The earliest placebo controlled study which defined the precise menopausal syndrome showed that oestrogens helped hot flushes, night sweats and vaginal dryness. They also had a mood elevating effect.(29) This work was further supported by the work of Campbell and Whitehead(30) who used Premarin and by the study of Montgomery et al(23) using higher dose oestradiol implants. This study of 90 peri and postmenopausal women with depression showed considerable improvement in the treatment group compared with placebo but only in the perimenopausal women. There is no improvement in the depression in the postmenopausal women with this treatment when compared with placebo.

At last, after 15 years, psychiatrists, particularly in the USA are coming round to the view that transdermal oestrogens are effective in the treatment of depressed perimenopausal women. Soares et al in 2001 studied 50 such women 26 with major depressive disorder, 11 with dysthymic depression and 30 with minor depressive illness. They treated them with 100 mcg oestradiol patches in a 12 week placebo controlled study. There was a remission of depression in 17 out of 25 of the treatment patients, (68%) and only 5 out of the 25 placebo patients (20%). This improvement occurred regardless of the DSM-IV diagnosis.

Rasgon et al, (32) studied 16 perimenopausal women with unipolar major depressive disorder for an 8 week open protocol trial comparing low dose 0.3 mg Premarin plus Fluoxetine daily. There was a greater response with oestrogen alone. All but two of the total patients responded but the response was greater in the ERT patients and it was significant that the reduction of depression scores began rapidly after the first week of treatment.

More recently, Harlow(34) studied a large number, (976) of perimenopausal women with a history of major depression and those without. The patients with the history of depression had higher FSH levels and lower oestradiol levels at enrolment to the study and those women with a history of antidepressant medication had three times the rate of early menopause. A similar excess rate was found in perimenopausal women who had had a history of severe depression.

It is reassuring for those "menopausologists" who have been trying to persuade the world of psychiatry that oestrogens have a place in the treatment of depressed women and pleasing to read at last the comments from Soares who states that periods of intense hormonal fluctuations have been associated with the heightened prevalence and exacerbation of underlying psychiatric illness, particularly the occurrence of premenstrual dysphoria, puerperal depression and depressive treatment during the perimenopause. It is speculated that sex steroids such as oestrogens, progestogens, (SIC), testosterone and DHEA exert a significant modulation of brain functioning…..

There are preliminary, although promising data on the use of oestradiol, (particularly transdermal oestradiol to alleviate depression during the menopause "at last we are getting through".

Progestogen Intolerance
These women having moderately high dose oestrogen therapy must of course have cyclical progestogen if they still have a uterus in order to prevent irregular bleeding and endometrial hyperplasia. The problem is that women with hormone responsive depression enjoy a mood elevating effect with oestrogens but this is attenuated by the necessary progestogen. This hormone can produce depression, tiredness, loss of libido, irritability, breast discomfort and in fact, all of the symptoms of premenstrual syndrome, particularly in women with a history or previous history of PMS. A randomised trial of Norethisterone against placebo in oestrogenised hysterectomised women, already referred to clearly showed this and in fact the paper was subtitled a "A model for the causation of PMS".(16)

If women become depressed with 10 to 12 days of progestogen, it may be necessary to halve the dose, decrease the duration or change the progestogen used.(35) It is our policy to routinely shorten the duration of progestogen in women with hormone responsive depression because adverse side-effects with any gestogen are almost invariable. We would therefore use transdermal oestrogens either 100 mcgs or 200 mcgs of an oestradiol patch or a 50 mg oestradiol implant and then we would reset the menstrual bleeding by prescribing Norethisterone 5 mgs for the first 7 days of each calendar month. This will produce a regular bleed on about day 10 or 11 of each calendar month.

If heavy periods occur, (and they usually do not), to extend the duration of progestogen to the more orthodox 12 days. At this stage many women would prefer to have a Mirena coil inserted so there will be no bleeding, no cycles nor any need to take oral progestogen with its side effects. It is It is not unusual for women at this stage who understand the benefits of oestrogens and the problems of their menstrual cycles, to request hysterectomy and bilateral salpingo-oophorectomy with hormone replacement therapy with oestradiol and testosterone.(33) This is a fact of medical life and patient choice but it will be at least another 15 years before psychiatrists attempt to leap over that hurdle.


  1. Oestrogen therapy is effective for the treatment of postnatal depression, premenstrual depression and perimenopausal depression the triad of hormone responsive mood disorders.
  2. Transdermal oestradiol 100 mcg or 200 mcg patches producing plasma levels
    approximately of 500 pmol/l and 800 pmol/l respectively should be used.
  3. These patients often require plasma levels of more than 600 pmol/l for efficacy.
  4. Consider adding testosterone for depression libido and energy.
  5. They require a cyclical progestogen or Mirena IUS if the patient still has a uterus.
  6. The most effective longterm medical therapy is oestradiol patches or an implant of oestradiol and testosterone with a Mirena IUS in situ.

Ultimately, a hysterectomy plus bilateral salpingo-oophorectomy and implants with testosterone may be requested.