Menstruation and its absence are constant spectres in the experience of anorexia and recovery. And they’re shape-shifting spectres, too. The presence or absence of a regular menstrual cycle can be part of the experience of illness and recovery in multiple ways and can be used to justify pretty much every position on the pro-anorexia to pro-recovery continuum. People I’ve corresponded with via this blog, for example, have expressed variants on all of the following:
- Having no period makes me terrified because I want to have children one day.
- Having no period, as well as plenty of other physical and psychological symptoms, helps focus my mind towards full recovery.
- I’m making excellent progress in recovery, physically and psychologically, but I’m worried because I still don’t have my period: does this, plus the fact that some days in early recovery I kept exercising and only ate 2,000 calories or so, mean I’m in ‘quasi-recovery’ instead of the real thing?
- I feel ill and lost and confused, but not having had my period in four months and having no thyroid gland function makes me fear that if I carry on this way I’ll die, even though I don’t even see myself as emaciated.
- I lost my period when I first lost weight and got it back when I was in the hospital, still at a very low weight. But having my period means my body fat can’t be all that low, can it?
- I have gained quite a bit of weight in recovery from relapse, and my period is regular again, which makes it feel like things should be better by now. But my life is being ruined by my constant desire to eat. I don’t want to go back, but this direction isn’t working either.
- Having gained a lot of weight in recovery is even harder to accept because my period is still absent: if only I had my period I’d feel better about my weight.
- Once I got my period back I started restricting to hold my weight rigidly on the lower boundary of healthy.
- I’ve just been diagnosed with anorexia but I don’t know why: my BMI is only 17 and I know yours was much lower. And I’ve had irregular periods for two years but I don’t feel light enough for this to be because of malnourishment; now I’ve had two periods recently which means that I’m functioning again. I just don’t understand why I’ve got a diagnosis when I’m not thin enough!
- Having a regular period encouraged my medical team not to take my illness seriously, and not to discourage my compulsive exercising.
- My daughter has reached what her therapist called a healthy BMI, but the therapist wants her to continue with weight gain because her period hasn’t yet returned and they want to create a buffer zone in case of future weight loss.
- My daughter has gained 25 pounds, got her period back, and claims her anorexic symptoms are no longer overwhelming: we are confused about when is a good place to stop weight gain and start weight maintenance.
- I stopped having my period as soon as I started dieting, so I knew my body probably wasn’t getting enough food, but I didn’t care — actually I loved it: the confirmation I was doing something wrong meant I had finally managed to defeat my hated body and had control over it.
As for me, fear of damaging my health definitely crystallised around the loss of my period, since it was such a simple on–off undebatable thing that was wrong. Mostly the fears were about osteoporosis; also about my skin, which was awful throughout my illness. Infertility I was less straightforwardly worried about. The child psychiatrist who helped in my first recovery effort, before going to university, asked me to send him a card when I had my first period, and said he’d do a celebratory dance when I told him. I didn’t, because it didn’t come. A couple of years later, aged 20, I was advised to start taking the Pill as hormone replacement, and wrote him a Christmas card, telling him about the artificially kick-started one I’d had. The event left me ambivalent:
Just went to the loo unsuspectingly as usual and found blood on the loo paper — dug out my old battered emergency sanitary towel, felt I was rejoining the ranks of womanhood. Ambivalence — I’ve not wanted to associate myself with breasts and babies and blood, have felt cleaner, purer than people who consider themselves ‘women’ — I feel like a more androgynous creature, not marked by any ‘curse’ of feminine physicality, and I’m rather sad to lose such a feeling — however delusive and destructive. But it is rather a relief to think that I’m not storing up the possibilities of osteoporosis and other delights for later in life — or even infertility, little of a problem as that seems at the moment. Like having a driving licence, I suppose — good to have, even if without any intention of using it. (03.04.02)
Oddly, I also read with surprise now in my diary that I was worried my then partner would be disappointed — even though there was no evidence (apart from him being with me, I suppose) that he idealised androgyny or what I described as the ‘ascetic freedom from physicalities’ in the illogically destructive way I did. I suppose the arrival of the Pill-induced period had in this respect the same implications as weight gain, sparked all the same anxieties: was he attracted to me because of anorexia?
The bitterly elegant ironies of the hormone replacement were also not lost on me. Taking pills that make you bleed and give you pain. Adding more pain, by ingesting something, to a regime that was already full of pain, because of not ingesting enough. Treating the symptoms of an illness of refusal to eat, of minimised consumption, with this minuscule consumption of slow-release refusal to procreate. Feeling something of that obsolete sense of oppression, of being ‘cursed’ by womanhood, arising in me as a result of the privilege of this achievement of modern medicine. Taking the Pill designed to counter the effects of sexual activity in order to counter the effects of something that made sexuality alien. Taking the Pill that frees women from one of the evolved imperatives of femaleness in order to place myself again under its rule — even while wishing I could do so without the blood and pain and proof of it.
It’s clear, then, that you can make bleeding or not bleeding once a month justify anything. Generalising from these examples, we can see that the underlying structure is the same in all of them: regular menstruation is being used as a proxy for health. A brief look at the online literature on menstruation and eating disorders confirms the idea of menstruation as a ‘barometer for health’ (Hello Clue, here), or, more strongly, that ‘Regular periods are a sign of hormonal balance and a regular menstrual cycle shows that everything is working tickety-boo just as it should do’ (Tabitha Farrar, here). The steps are easy to take from resumption of menses to ‘hormonal balance’ to ‘optimal [bodily] functioning’. But that’s an awful lot of distance to cover in two leaps. Even if the specific hormones involved in ovulation and menstruation are in balance, all kinds of other hormonal interactions might not be. And then, of course, making health nothing more than hormone balance is a pretty blinkered way of looking at things.
The hormonal basics
To understand better what’s at stake here, let’s first do a brief recap of what happens when menstruation stops because of malnutrition. (See a list of further reading on the topic here.)
- Functional hypothalamic amenorrhea (FHA) sets in when levels of ghrelin (the ‘hunger hormone’) increases and its counterpart leptin decreases (when you eat less), or when cortisol levels increase (e.g. when you’re stressed, or over-exercising).
- Either of these changes (and various other candidates; see a helpful diagram of the possible hormonal pathways to amenorrhea, in Fourman and Fazeli, 2015, Figure 2) reduces the frequency of regular pulses of gonadotropin-releasing hormone (GnRH) released by the hypothalamus.
- This makes the pituitary gland release less luteinising hormone and follicle-stimulating hormone.
- And this in turn means that less oestrogen is produced, the ovarian follicles (the sacs that contain the eggs) don’t grow, and ovulation may not occur.
Overall, FHA involves low oestrogen and progesterone, increased testosterone, and loss of the usual peaking of luteinising hormone and oestrogen around ovulation. These changes entail others, most importantly for the eating disorder context the loss of oestrogen for calcium absorption: low oestrogen production from the ovaries creates a risk of loss of bone density (osteopenia) which can lead to significant loss of density, porosity, and susceptibility to breakage, in full-blown osteoporosis. Other direct effects of low oestrogen include night sweats, disturbed sleep, poor concentration, poor skin, and irritability or depression — symptoms sometimes described as premenopausal.
What changes in anorexia and recovery, and when and why?
The obvious next question for our purposes is what kinds of physical change are most likely to trigger onset of FHA, and what changes in the opposite direction are most likely to bring this aspect hormonal functioning back to normal. The simple answer is: not eating enough and/or exercising too much makes your period likely to stop, and stopping doing these things makes it likely to come back. But beyond this rough rule of thumb there’s a great deal of complexity. The (sadly now dormant) Science of EDs blog has a good post exploring this question through analysis of a 1997 paper (Golden et al.) which deals with amenorrhea in the eating disorder context. The study asks what the best predictors are for when menses will resume after treatment for anorexia.
The 1997 study (and as far as I’m aware, there’s been no direct update) involved 100 adolescent girls with anorexia who had been amenorrhoeic for an average of 22 months, 69 of whom got to 1-year follow-up, and 59 to 2-year follow-up. (It’s not quite clear how being available for follow-up or not relates to the time actually spent in treatment.) The experimenters measured bodyweight and body fat percentage, as well as levels of luteinising hormone (LH), follicle-stimulating hormone, and estradiol (an oestrogen steroid hormone) at baseline and every 3 months during treatment until menstruation resumed (defined as two or more consecutive spontaneous menstrual cycles). The treatment consisted of a mixture of medical, nutritional, and psychiatric intervention aimed at nutritional rehabilitation, weight restoration, and resolution of underlying psychological conflicts, with ‘limitations on exercise’ until resumption of menses.
Following up one after one year of treatment, 47 (68%) of 69 patients had resumed menses, with no significant differences in bodyweight, BMI, or body fat percentage at that point compared with those who hadn’t. The only significant differences at the one-year mark were hormonal: those who hadn’t resumed menses had lower levels of LH and estradiol at follow-up (estradiol being the strongest correlate). There were significant differences in the pre-treatment measures, however: the group at high risk for remaining amenorrheic had lower levels of FSH and GnHR, and undetectable levels of LH, before the start of treatment, and had lower bodyweight and body fat at that point too. Undetectable LH levels correlated with a tenfold increase in likelihood of still being amenorrheic after a year.
Other measures that we might expect to be significant – age, age at which menstruation had started, duration of amenorrhea, duration of illness, amount of weight lost, psychological measures, and exercise — were not predictors for resumption of menstruation by one year. (Caveats here include the strange fact that none of the psychological measures was ever taken again after admission, which makes one wonder about the solidity of their criteria for discharge, and the fact that once treatment ended people were free to exercise again, and exercise was self-reported so likely under-reported.) So basically, what this study found is that how ill you were in psychological terms was unrelated to the likelihood of your period return by one year, and the only relevant factors you might plausibly be aware of when monitoring your own physical state were bodyweight and bodyfat before you started treatment, not bodyweight when you finished. (This counters claims like Gwyneth Olwyn's that menstruation will return once you reach your body's 'optimal weight set point'.)
Using US population norms to establish a generic ‘mean standard bodyweight’, Golden and colleagues found that resumption of menses happened on average at 91.6% of that standard bodyweight level (though this varied between 75% and 115%) and that 48 (86%) of the 56 patients who achieved the mean standard bodyweight resumed menses within 6 months of reaching that weight. (It’s pleasantly surprising that so many did, given how many studies of this kind don’t allow for weight restoration to anything near a population-level mean; see my discussion of this problem here, for example.) On average, people regained their period at a weight 2.05 kg higher than the weight at which they’d first lost it (though how accurate the reporting about the original loss can have been is another question, which the authors don’t comment on). But bodyweight and body fat at the first visit were predictors — the lower these were, the less likely resumption of menses was.
So what’s the upshot from this? Basically, it’s complicated! This is why amenorrhea was removed from the diagnostic criteria for anorexia in the latest edition of the DSM (as well as the fact that it’s useless for men, and prepubescent and postmenopausal women). Don’t expect anyone to be able to tell you in advance when your period will come back. Don’t expect the return of your period to have anything systematic to do with your improvement on any other measure except estradiol and luteinising hormone. Indeed, an interesting early study (Nillius et al., 1975) did things the other way round, using an experimental intervention to change hormone status and see what happens. It found that treating people with a synthetic LH-releasing hormone induces ovulation and menstruation, but doesn’t result in any change to mood, eating, bodyweight, or libido. And in a more recent descriptive study (Miller et al., 2004), higher body fat and hence leptin predicted menstruation, but there were only small differences in psychological status (measured by the Eating Disorders Inventory) between the menstruating and the not menstruating groups — and, interestingly, in the opposite direction from what you’d expect: those menstruating had slightly more severe symptoms on this scale. (For more discussion of these and related studies, see Södersten et al., 2008, pp. 450-451.) It’s also important to remember that quite a lot of people never lose their period despite severe illness (Miller et al., 2004). So essentially, you shouldn’t use regular menstruation as a sign that everything is fine, and you shouldn’t be frightened if you’re still not menstruating despite doing everything else right. Menstruation may be a helpful rule of thumb as a necessary criterion for recovery if you were menstruating regularly before you became ill, but if it’s necessary, it’s absolutely not sufficient.
And by the way, according to the latest American Psychiatric Society guidelines for treatment of eating disorders (2010), the evidence on use of hormone replacement therapy to improve bone mineral density is poor (essentially limited to one unreproduced study with very-low-weight adolescents, Klibanski et al., 1995; see also Drabkin et al., 2017). There’s no replacement for recovering properly via full weight restoration, and having a monthly bleed may make it easier for you to pretend to yourself that everything is OK. Calcium and vitamin D supplements may help a bit, but basically what you need to do is eat.
What evolution can teach us
To help us grasp the sheer breadth of the gap between recovery and resumption of menses, it may be interesting to zoom out and ask on the evolutionary scale why menstruation happens and why it stops. Not all female animals menstruate in the way we do. The human version of menstruation (shared by our close evolutionary relatives like chimps) is a form of overt menstruation, where blood and the lining of the uterus are lost to the body in every cycle. The other option (which happens in most other mammals) is covert menstruation, where the blood is reabsorbed rather than being lost to the organism. Actually, in humans, about two-thirds is in fact reabsorbed, but the length of pregnancy and hence the size of the uterus mean it isn’t practical to absorb all of it. The same considerations apply to the question of why we don’t just continuously keep the uterus lining in a state where it could support a fertilised egg: the energy costs of doing that are greater than the costs of having to rebuild it every fertility cycle.
And the idea of energy costs brings us to the evolutionary reasons why periods stop. Remember the evolutionary algorithm: copying with variation and selection creates design (e.g. Blackmore, 2017). This bafflingly simple rule explains the vast complexity of every living organism on this planet. And what is it that’s being copied, varied (through random mutation), and selected (by environmental pressures)? Not organisms, but genes (and arguably other replicators, like memes [e.g. Blackmore, 1999]). In this sense, a human being is a mere carrier of his or her genes (and memes): the temporary housing in an aeons-old evolutionary process. And this, remember, is why we don’t ovulate and menstruate in times of famine (low energy availability, high energy expenditure). Certainly not because of any care that any divine plan takes of our wellbeing.
At the evolutionary scale, childbirth is significant because it allows for the survival of the genes into the next generation. (Before I go any further I should note that the relationship between amenorrhea and infertility isn’t a completely neat one-to-one mapping: it is possible to be fertile when not menstruating, and it is possible to be infertile despite menstruating. One blog reader described getting accidentally pregnant while not menstruating, giving birth, and going on to have a second child, all while seriously ill and in and out of the hospital for her children’s first ten years. But for simplicity’s sake let’s assume that menstruation and fertility are roughly coterminous.)
So imagine a distant ancestor in evolutionary history, maybe around the Middle Palaeolithic (or Middle Stone Age) between 300,000 and 50,000 years ago, where homo sapiens emerged and developed into behavioural modernity. Her period carries on when a famine hits. She has sex, she conceives, an increasing proportion of her energy intake is used to grow the foetus inside her, other physical functions become increasingly compromised, and she dies without giving birth to a viable infant, or dies when the infant is technically viable but lacks other support (e.g. from the male out doing hunter-gatherer stuff) to allow it to grow and survive postnatally without her. She and the child die, and her genes are eliminated from the gene pool.
Now imagine another distant ancestor, maybe in the same tribe, whose period stops when that same famine hits. She too has sex, but she is not ovulating, so she doesn’t conceive. She survives the end of the famine, begins to eat more and forage less, and begins to ovulate again. Then she has sex, she conceives, she adjusts her energy intake to accommodate the extra demands of the growing foetus, and she gives birth to a healthy infant. Her genes remain in the gene pool for future selection. Thus a small selective advantage is accrued by the genes which selected for ‘stop period during famine’, and in her children that response to famine is again more likely to result in successful genetic transmission than the ‘keep periods happening’ response.
It’s interesting here also to bring to mind other variations on these possibilities. There’s the woman who conceives but miscarries due (directly or indirectly) to lack of available energy. She survives to attempt conception, gestation, childbirth, and genetic transmission again in future; this is evolution’s last-minute correction mechanism. There’s the woman who doesn’t die after giving birth successfully, but who is too malnourished to breastfeed and otherwise feed and care for her child. If the infant is adopted by another lactating female it may survive and pass its genes to the next generation; if it isn’t, it dies too, and the genetic line ends. Or there’s the woman whose fertility is never restored after long famine, and who has no children despite favourable environmental conditions returning before she dies; and so the genetic line ends with her, irrespective of the quality of her life post-famine.
So, a woman fails to pass on her (and her sexual partner’s) genes either because she doesn’t give birth to children (who carry the genes into the next generation), perhaps thanks to long-term infertility for environmental or other reasons, or because she has children but the children don’t survive, for example, because she can’t feed and protect them. So, our genetic makeup now has been selected for by all the countless examples of women who had children who survived to pass on their own genes, versus all those women who didn’t have children when they could have or did have children when they shouldn’t have — in circumstances radically far removed from our own now.
What are the conspicuous absences in all of this? The woman’s health in any respect other than her ability to create viable descendants who in turn transmit their own genes through sex and childbirth. The woman’s survival at any point after her children become capable of survival without her. The woman’s happiness. The woman’s desire to be alive, or not. The woman’s desire to have children, or not. Evolution is blind to all of this.
To believe something like ‘I’ve got my period, I must be fine’ is to let your concept of what it means to be fine — healthy, happy — be dictated by the selection pressures of early hominids. And why would you? We now (me writing this blog post, and most of you who are reading it — though the world’s inequalities in this respect are profound) have the unprecedented privilege of getting to decide that health means more to us than the survival of our genes. As far as evolution is concerned, you’re a carrier for another organism that carries your genes. As far as you’re concerned — are you more than that?
The complex upshot
Combined, the clinical and the evolutionary perspectives on the breadth of the gap between health and wellbeing on the one hand and menstrual bleeding on the other, make it incredible to me that the conclusion of Golden and colleagues’ 1997 paper was:
A weight approximately 90% of standard bodyweight was the average weight at which ROM [resumption of menses] occurred and is a reasonable treatment goal weight, because 86% of patients who achieved this goal resumed menses within 6 months. (p. 16, my italics)
To be fair, later in the paper they add a critical qualifier:
we recommend a treatment goal weight at or above 95% of median weight for height and age. (p. 20, my italics)
But the implication is loud and clear: once the current state of the ongoing genetic selection process results, for only 86% of an already minuscule number of individual organisms, in a physiological status which is in probabilistic terms more likely to result in successful genetic transmission than not, you’re done.
The decision to call a halt to weight restoration (and almost inevitably, thereby, to encourage restriction to maintain the new 'healthy' weight) is to be based on nothing but this?
The trouble is, people, and the systems we create, love clarity. Medicine, in particular, loves things that give measurable, yes-no, healthy-unhealthy answers. This is one reason why mental illness is still so badly understood and treated: the world of the mind is not a world of measurable, yes-no, healthy-unhealthy answers. So although every mental illness always has physical components, and the physical ones may be easier to measure than the mental ones (bodyweight/BMI in anorexia being a classic example), reducing mental illness to objective criteria doesn’t tend to work very well, because the suffering is about the intangible mental stuff as much as the more tangible physical stuff. One person can be eating the exact same as another for an entire week and one be entirely healthy and the other be profoundly unwell. That is why so many people get to a pseudo-recovered state in which everything physical seems fine (including the menstrual cycle) and yet everything is not fine: it’s partial recovery and the person knows it, or is eventually prompted to realise it.
Only you can tell whether you’re better from your eating disorder. Other people can take measurements, at every point on the spectrum from measuring your luteinising hormone to asking you in-depth questions, but this existence is your reality. Be worried about the specific damage that fertility-related hormonal imbalances entail for you: about making yourself more likely to have easily fractured bones, or about sleeping badly now. Be pleased about the reduction and reversal of that damage when your period returns. Be thrilled at the likelihood of being able to have children being vastly increased when it returns if that’s important to you. But don’t kid yourself the monthly bleeding means any more than what it does in fact mean. Don’t assume it means you’re doing something wrong if you’re progressing well in recovery and it’s still absent (the timescales of all these interaction pathways are complex), and don’t kid yourself you’re fine if you’re menstruating but other stuff is still bad.
And any time you find yourself wondering, for any reason at all, whether you’re really ill enough to warrant recovery, ask yourself: Well, what do I actually want for my life?
Blackmore, S.J. (1999). The meme machine. Oxford: Oxford University Press. Author's webpage for the book here.
Blackmore, S.J. (2017). Replicator power. Edge annual question 2017: What scientific term or concept ought to be more widely known? Full text here.
Drabkin, A., Rothman, M. S., Wassenaar, E., Mascolo, M., & Mehler, P. S. (2017). Assessment and clinical management of bone disease in adults with eating disorders: a review. Journal of Eating Disorders, 5(1), 42. Open-access full text here.
Fourman, L. T., and Fazeli, P. K. (2015). Neuroendocrine causes of amenorrhea—an update. The Journal of Clinical Endocrinology & Metabolism, 100(3), 812-824. Open-access full text here.
Golden, N. H., Jacobson, M. S., Schebendach, J., Solanto, M. V., Hertz, S. M., and Shenker, I. R. (1997). Resumption of menses in anorexia nervosa. Archives of Pediatrics & Adolescent Medicine, 151(1), 16-21. Direct PDF download here.
Klibanski, A. N. N. E., Biller, B. M., Schoenfeld, D. A., Herzog, D. B., & Saxe, V. C. (1995). The effects of estrogen administration on trabecular bone loss in young women with anorexia nervosa. The Journal of Clinical Endocrinology & Metabolism, 80(3), 898-904. Paywall-protected journal record here.
Miller, K. K., Grinspoon, S., Gleysteen, S., Grieco, K. A., Ciampa, J., Breu, J., ... & Klibanski, A. (2004). Preservation of neuroendocrine control of reproductive function despite severe undernutrition. The Journal of Clinical Endocrinology & Metabolism, 89(9), 4434-4438. Open-access full text here.
Nillius, S. J., Fries, H., and Wide, L. (1975). Successful induction of follicular maturation and ovulation by prolonged treatment with LH-releasing hormone in women with anorexia nervosa. American Journal of Obstetrics & Gynecology, 122(8), 921-928. Paywall-protected journal record here.
Södersten, P., Nergårdh, R., Bergh, C., Zandian, M., & Scheurink, A. (2008). Behavioral neuroendocrinology and treatment of anorexia nervosa. Frontiers in Neuroendocrinology, 29(4), 445-462. Full-text PDF here.