Abstract
I first met Julie two years ago when she was then 42. She had very severe, treatment-resistant depression that had started just before the birth of her fourth child and had worsened in the puerperium. Six years on, she had never been well and had been admitted repeatedly when she was judged to present a high suicide risk. She was prescribed maximal levels of venlafaxine along with lithium, mirtazepine, sodium valproate and ECT, which seemed just to maintain stability. Her psychiatrist wondered whether HRT might help.
This was a very reasonable question. Julie's depression could be traced back to oophorectomy for a dermoid cyst during her last pregnancy. The psychiatrists noted that norethisterone, used to stem a very heavy and prolonged period, had appeared to be associated with deterioration which even the ECT could not stabilize. They wondered whether her increasingly sporadic periods and hot flushes might reflect ovarian failure and not be a side-effect of the drugs. I was asked to help.
Julie came with a support worker from the mental health team. Her affect was clearly flat, she did not make eye contact and answered factually in short sentences. She told me that she never went out of the house alone and could do very little other than think of ways to end it all. She realized that her children and husband were caught up in this, but could offer no other solution.
The assessment process suggested that while her severe daytime flushes and night sweats could have been medication related, when combined with the pattern of waking in the middle of the night (not early morning waking), it was more likely to be estrogen deficient. Increased urinary frequency, vaginal dryness, pruritis and loss of libido reinforced this picture. Her bleeding pattern had been erratic for 18 months with periods 3–6 weeks apart and then a three-month gap followed by the heavy prolonged bleed for which norethisterone had been prescribed. Gynaecological assessment for intermenstrual bleeding had revealed no pathology the year before.
A presumptive diagnosis of perimenopause was agreed and the issues surrounding hormone therapy discussed. Julie was keen to try. We started with patches delivering 100 µg/day estradiol and suggested a levonorgestrel bearing intrauterine system (IUS) to provide endometrial opposition and bleeding control with least systemic impact.
The patches rapidly produced a lift in mood, but within a week of introduction of the IUS intense suicidal thoughts returned and it had to be removed. We appeared to have demonstrated extreme progesterone intolerance, but to what extent could estrogen help? Julie understood what we were trying to do but also our caution. We agreed that we would use estrogen alone for assessment purposes and if there was significant and sustained improvement would then consider the possibility of hysterectomy in order to avoid the progestagens.
Ten weeks later Julie had been to Tesco's on her own for the first time in six years. This might seem trivial but she was delighted. She was still taking all the drugs and her mood was still low though no longer incapacitating: no further ECT had been required. Sleep had improved, her flushes were only sporadic and there had been no bleeding. Vaginal dryness and libido were still a problem. Serum estradiol level was towards the bottom of the reproductive range at 314 pmol/L. Measurement allowed me to argue the case to increase the dose delivered by the patches to 150 µg/24 hours and vaginal estradiol was added.
Three months later the improvement in mood was even more dramatic. As predicted, with physiological levels of estrogen, Julie had started to bleed. This was managed with tranexamic acid. The gynaecology team were well briefed and agreed that hysterectomy with the removal of the residual ovary could be justified on the grounds that the risk of the procedure was less than the risk to Julie's life without estrogen or with a combined regime.
The postoperative road has not been smooth. It has proven difficult to achieve adequate response with patches. Serum estradiol levels without any ovarian production and in the presence of the metabolically stimulating concomitant medication were still only 84 pmol/L when using 2 × 100 µg/24 hour patches of the same variety as before. We resorted to implants and have achieved success. We can chart an almost linear relationship between mood and serum estradiol levels. At 340 pmol/L symptoms start to return but at 674 pmol/L Julie is back to her original personality. She is cheerful, outgoing and independent. She has been able gradually to reduce and then stop all of the psychiatric medication. She is active, four dress sizes smaller and, having previously never left the house, now has a job and is planning a holiday.
Julie is truly a new woman. Her degree of estrogen dependence and progesterone intolerance may be extreme, but is a salutary lesson. It is sobering to think that there may be other women with mental health problems who might benefit from consideration of their hormonal status.