Eating disorders… (I)

“Dermatologists have an important role in the early diagnosis of eating disorders since skin signs are, at times, the only easily detectable symptoms of hidden anorexia and bulimia nervosa. Forty cutaneous signs have been recognized”

The full title of the book is Eating Disorders and the Skin, but there’s a lot of stuff about eating disorders in general in this book as well, and I figured I’d mostly focus on the ‘general stuff’ in this post. Here’s my goodreads review of the book, which I gave 3 stars.

Here are the DSM-IV-TR diagnostic criteria for anorexia nervosa:

“1. Refusal to maintain body weight at or above a minimally normal weight for age and height (e.g., weight loss leading to maintenance of body weight less than 85% of that expected, or failure to make expected weight gain during period of growth, leading to body weight less than 85% of that expected).

2. Intense fear of gaining weight or becoming fat even though underweight.

3. Disturbance in the way in which one’s body weight or shape is experienced, undue influence of body weight or shape on self-evaluation, or denial of the seriousness of the current low body weight.

4.4. In postmenarcheal females, amenorrhea, i.e., the absence of at least three consecutive menstrual cycles.”

Interestingly, aside from anorexia [-AN] and bulimia [-BN] (diagnostic criteria here), there’s also a big category called ED-NOS – Eating Disorder Not Otherwise Specified. That’s for cases that don’t really fit into the standard criteria for specific eating disorders; they note than an example of this type could be a female fitting all diagnostic criteria for AN except that she has regular menses. It is perhaps worth mentioning here that surprisingly enough (…to me), menstrual irregularities are not limited to cases of AN, thus: “In almost 50% of bulimic patients, menstrual irregularities, such as oligomenorrhea or amenorrhea, take place”. They note in the book that there’s been some concern about the validity of the ED-NOS category, which makes up almost 60% of patients with an eating disorder. Eating disorders are much more common in females than in males (“Males are generally reported to account for 5–10% of anorectics and 10–15% of bulimics identified in the general population”), and particular subgroups mentioned to be at high risk are athletes, models and dancers. It’s noted in the book that most epidemiological studies are conducted in high-risk settings, whereas epidemiological studies assessing risk in the general population are somewhat rarer. One problem complicating matters a little in terms of estimating risk is that an eating disorder cannot be diagnosed through a self-report questionnaire; you need a structured or semi-structured interview to make a diagnosis, which makes things more expensive. As in other contexts one way to get around this issue, at least to some extent, is to employ multi-step screening protocols – in this case a two-step procedure in which individuals at high risk are identified at the first step through inexpensive means, and these individuals are then later assessed more carefully in the second step, employing more accurate (and expensive) methods.

They note that in Western countries, point prevalence of AN in female adolescent (the highest risk sub-group) is estimated at 0.2-1% of the population, whereas the prevalence studies on bulimia nervosa indicates that this eating disorder is somewhat more common, with the majority of studies finding prevalences of 1.5-5%; do recall again that most studies as already mentioned look at high-risk subgroups, so total population prevalence is likely to be lower than this. They observe in the book that general-practice studies find that the incidence of anorexia nervosa is less than one in ten-thousand per year (8 per 100,000 per year); so full-blown AN certainly is likely quite rare in low-risk populations.

On lifetime risk, the book notes that:

“Most of the epidemiological studies on ED [eating disorders] have evaluated the prevalence of full syndromes of both AN [anorexia nervosa] and BN [bulimia nervosa]. The few studies that have evaluated partial or subclinical manifestations of EDs in young females, however, found lifetime prevalence rates of 5–12% for atypical AN and 1–4.8% for atypical BN and up to 14.6% in adolescent samples”.

A review of epidemiological studies concluded that there’s no evidence of either a secular increase in AN or BN over time; to the extent that the number of people with diagnosed BN has increased over time, changes in diagnostic and referral practices likely account for this. On a related topic it is noted in the book that “It is a common idea among clinicians that early-onset cases of anorexia nervosa (AN) are increasing, but few data in the literature are available to demonstrate this trend.”

AN most commonly present among females at the age of 15-19, whereas BN presents a little later, most commonly at the age of 20-24. But eating disorders are not limited to teenagers and young adults: “Even if anorexia nervosa and bulimia nervosa occur characteristically in females during adolescence and young adulthood, there have been case reports of illness beginning after the age of 25 and even after the menopause, and some authors suggest that the rates of eating disorders in older patients may be increasing [2]. Clinical impression suggests that the late-onset cases present with more depressive features than the adolescent counterpart. […] dieting is considered one of the most salient precipitating factors.”

Self-report metrics can only help you so much when you’re trying to assess risk; a major problem in this context is that denial of illness is a very common feature in these patient populations (so you certainly can’t just ask people if their relationship with food/exercise etc. might be unhealthy…): “typically, [the] course [of an eating disorder] is characterized by a high fluidity between the diagnostic classes; furthermore, the patient often denies even to himself the psychiatric nature of the disease” (recall also that “denial of the seriousness of the current low body weight” is included in the diagnostic criteria). The book covers a lot of symptoms which relate to low body weight – like cold intolerance, bradycardia (slow heart rate), acrocyanosis (bluish discoloration of the hands and feet, caused by slow circulation), systemic hypotension (low blood pressure), lots of skin signs (I haven’t decided yet how much detail I’ll go into, so let’s leave it at that now) – or e.g. to purging behaviours (throat and tooth pain due to vomiting and enamel erosion), but it would go much too far to discuss all these in detail here. One to me interesting aspect of the coverage was that whereas BMI is a useful sign, it’s not itself a diagnostic criterion; the authors note that a BMI below 18.5 is considered pathological, but when listing main signs of anorexia nervosa the most important diagnostic sign (or at least the first one listed) is a BMI below 17.5; I assume part of the discussion surrounding the validity of the ED-NOS category probably relate to individuals who’re in this ‘border area’; they likely have some symptoms due to low body mass (like e.g. cold intolerance), but they don’t have full-blown AN (there are a lot of things that can go wrong when you have low body mass – there are a lot of symptoms described in this book!). It’s also important to note that very different symptom patterns can be present at similar levels of BMI, as the severity of symptoms also relate to how fast body mass decreases – the body is actually capable of adjusting quite well to lower energy intake states (in the short run at least), and so “if weight loss is gradual, it is possible to maintain, even for a long time, an apparent metabolic equilibrium.”

Anorexics have high mortality rates: “From a meta-analysis of 119 studies involving 5,590 patients, Steinhausen reported a crude mortality rate of 5% which exceeded 9% in a followup of 10 years.” Remember when thinking about those estimates that most of the people in these studies were likely young women – these numbers are high, and the authors note that anorexia nervosa “represents the major cause of death of young women in the age between 12 and 25 years.”

Most deaths are due to ventricular arrhythmia; the book goes into some detail about how anorexia affects the cardiovascular system, but I won’t discuss this in detail. An important observation is that: “Cardiac findings tend to disappear with weight recovery.” I assume this comment relates mostly to findings like QTc prolongation, QTc dispersion, and mitral valve prolapse, all of which are found in anorexics, whereas I’d be surprised if cardiac abnormalities related to direct damage to the heart muscle resolve themselves after weight gain, but the book does not go into details on this topic, except in the sense that it is noted that heart failure is uncommon in anorexics. Among those who survive their illness, osteoporosis is a major irreversible long-term problem. People with higher body mass tend to have a higher bone mineral density and thus a lower risk of osteoporosis (unless they get type 2 diabetes, in which case the situation is, well, complicated), so perhaps it’s not really surprising that women with AN and very low body mass index tend to develop osteoporosis. They certainly do:

“Osteopenia and osteoporosis represent one of the most relevant and potentially not reversible complications of eating disorders. This complication is particularly severe when eating disorders have an early onset […] Bone loss is an early effect of the disease, already present after 6–12 months […] In untreated patients, bone loss ranges from 4% up to 10% per year […]. In case of recovery, the progressive loss of BMD [bone mineral density] stops, but in most cases, a normal bone mass is not restored [64].”

It’s noted that bone loss is due to both hormonal and metabolic factors; estrogen plays a role, and “BMD loss in AN is more rapid and severe than in other hypoestrogenic conditions”. Despite this observation weight gain is considered the primary treatment modality of osteoporosis in this context (i.e., not estrogen therapy), and research using estrogen therapy to try to boost bone mineral density in anorexics who did not also gain weight has not been successful.

A to me interesting aspect of the coverage which I could not help but discuss here is how eating disorders relate to diabetes; the book has a few remarks on this topic:

“The concurrence of an eating disorder with insulin-dependent diabetes has been outlined by several researchers: especially bulimia nervosa and disorder not otherwise specified (EDNOS) are reported to be significantly higher in females with type 1 diabetes […] In case of comorbidity, ED onset followed the diagnosis of IDDM in 70% of the patients [10]. Specific aspects of diabetes and its management could, in fact, potentially increase a particular susceptibility to the development of an eating disorder: weight gain, associated with initiation of insulin treatment and dietary restraint, might, in fact, trigger body dissatisfaction and the drive for thinness with consequent weight control behaviors ranging from healthy to very unhealthy behaviors […] insulin omission [is] a common weight loss behavior in girls with IDDM and eating disorder […] APA Guidelines 2006 suggest that insulin omission should be considered a specific type of purging behavior in the next DSM revision”.

I don’t know if this suggested change has been implemented at this point, but it would make a lot of sense. To people who don’t know what this ‘insulin omission’ they talk about is all about, the short version is that if you’re a type 1 diabetic in need of regular insulin injections, if you don’t take enough insulin you lose weight and you can eat pretty much whatever you like without gaining weight; which is of course an unfortunate though likely very attractive option for young women to have. The downside of engaging in systematic insulin omission behaviour of that kind is that you’ll likely go blind from your diabetes and/or die of kidney failure or DKA if you do that for an extended period of time.

January 2, 2016 - Posted by | books, diabetes, medicine, Psychology


  1. Yet another thing the DSM and its typological bent have screwed up. People with mental disorders don’t get help because they don’t think they have a problem, or don’t think their problem is “serious enough”, and we confirm it by telling them their BMIs aren’t low enough to qualify for help. “Lose 15 more pounds and maybe we’ll pay attention.” It’s not surprising prevalence rates for AN are so low when we define it practically out of existence.

    Comment by human | January 2, 2016 | Reply

    • I don’t really have strong opinions on this topic, but my take on this is perhaps a bit different. An important point here is that there’s really a lot of variation in terms of how even what most people would consider ‘normal people’ approach lifestyle variables such as food and exercise, so I imagine it’s somewhat difficult to make hard and consistent boundaries which are able to separate people who behave in a ‘normal manner’ from people who do not, just as it is difficult to make strong boundaries between behaviours which are ‘healthy’ and behaviours which are not. You also see this phenomenon/problem at the other end of the weight spectrum; it’s really unhealthy to have a BMI of 35 and a lot of psychological mechanisms definitely contribute to an unhealthy weight like that, but many people, including certainly a not insignificant proportion of the people who are actually in that weight range, do not agree that it’s useful to categorize people in that weight range as ‘having an illness’, despite many of them dying as a result of their life style. They’re fat, yeah, but they don’t die from ‘fat disease’ – they die of ‘heart disease’ or ‘diabetes’, despite the fact that both of these things might easily in a different setting with slightly different medical categorization schemes be conceptualized as ‘complications to their fat disease’. The wide variation in lifestyle behaviours across individuals and over time I think contribute to a certain hesitancy among medical people to categorize people as ‘sufficiently outside the norm for intervention to be mandated’.

      In that context both anorexia nervosa and bulimia nervosa are definitely conditions where it is not hard to argue that behaviours and mental models of the people affected are far away from the norm, and both conditions are definitely harmful. If you include people without these conditions in your model of ‘disease behaviour related to eating/exercise/etc.’, which is what you implicitly do when you also look at people with EDNOS, the lines become more blurry. The discussion about the usefulness and validity of the EDNOS category I think likely reflect in part the fact that some of these people may either not really need treatment, or treating them may not be cost-effective (I imagine treatment compliance is a big problem to be worked around here, and dealing with that is not easy; I would also expect treatment effects to be positively correlated with disease severity). The cost-effectiveness point relates to the fact that it’s expensive and complicated to treat eating disorders, so it makes a lot of sense to not medically intervene unless the problem is somewhat severe. I wouldn’t rule out that GPs might be able to do more to address these kinds of issues before they ‘blow up’ and the individual develop full-blown anorexia nervosa, but opportunistic screening tends to be very expensive. However the authors of the book would certainly be in favour of a model where health care providers able to potentially intervene knew more about how to figure out who are and who are not in trouble – but this relates more to the part of the book’s coverage which I have yet to cover; there are a lot of skin signs and signs which a dentist might pick up on which are useful for health care providers to know about, and which may facilitate earlier intervention, thus improving outcomes. One of the primary aims of this book is presumably to tell people who otherwise would not have known this that there are a lot of subtle signs, especially skin signs, which might lead to a suspicion that this individual has unhealthy eating patterns, and they share this knowledge presumably in order to make intervention in these cases easier.

      Comment by US | January 2, 2016 | Reply

      • Honestly, I don’t believe a strong behavioral focus with regard to eating disorders is very productive, precisely because of inter-individual variability. The impact of lifestyle, hormonal and metabolic variation on the markers used by the DSM – BMI, amenorrhea – will result in corresponding behavioral adaptations even holding psychological factors constant, meaning that the AN and BN categories as currently defined measure non-psychiatric properties to a significant extent… which they shouldn’t, since this is the the bible of *psychiatry*, after all. The current definition surely includes individuals who are not significantly impaired or troubled by their “anorexia”, who do not restrict or purge to any great degree, do not obsess about their weight to any great degree, and do not experience a great degree of mental suffering as a result. Worse, it excludes individuals who do heavily restrict, purge, over-exercise, obsess about weight for literally hours per day, and are plagued with crippling insecurity and self-loathing, but who fail to meet the necessary criteria (underweight BMI and amenorrhea) because of different, shall we say, somatic calibrations. The DSM construct is measuring the wrong things, because it forgot that it’s supposed to be about mental properties.

        As for cost effectiveness, I’d say that treating those with disordered eating should, priority-wise, be roughly on par with treating depression, OCD, and other major mood disorders, especially considering the comorbidity rates and the probably causal link between the two. Which is to say, much higher than either eating disorders or mood disorders currently are, given their societal (not to mention personal) costs.

        Comment by human | January 2, 2016

      • “The impact of lifestyle, hormonal and metabolic variation on the markers used by the DSM – BMI, amenorrhea – will result in corresponding behavioral adaptations even holding psychological factors constant”

        I’m not sure I understand what you’re saying here; can you restate this point?

        “The current definition surely includes individuals who are not significantly impaired or troubled by their “anorexia”, who do not restrict or purge to any great degree, do not obsess about their weight to any great degree, and do not experience a great degree of mental suffering as a result.”

        First, if you have AN you have an: “Intense fear of gaining weight or becoming fat even though underweight.” That’s a diagnostic criterion, so there are no individuals with ‘properly diagnosed AN’ who do not “obsess about their weight” – at least I figure if you have an intense fear of gaining weight, you do obsess about your weight, and if you don’t do that you don’t have AN.

        Second, in terms of people who are underweight ‘but there’s really nothing wrong with them’, there are very likely some individuals like that which might still receive a diagnosis of AN or BN, but I would question how many of those people actually get a diagnosis; the relatively strict criteria for AN for example would indicate to me that there’s a greater emphasis on avoiding over-diagnosis than on avoiding under-diagnosis, presumably in part because interventions can be a really big deal; in severe purging subtypes of AN not only do you need to watch everything the individual eats and drinks, you also need to watch them all hours of the day outside of the meals because if you don’t they’ll just vomit up the food or take laxatives to avoid gaining weight. The different approaches to treatment of overweight and underweight I incidentally find interesting; if you’re starving yourself to death, you’ll be locked up in a small room and force-fed until you’ve gained weight, whereas if you’re eating yourself to death, well, …

        The current diagnostic criteria will include some women who are not sick, and miss some who are, but I think the number of women *with diagnosed AN or BN* (as opposed to women who formally meet the diagnostic criteria, but do not have a diagnosis) in the first category is likely to be relatively small (as many of them presumably will never get a diagnosis) and the EDNOS category is included to take care of the others:

        “Worse, it excludes individuals who do heavily restrict, purge, over-exercise, obsess about weight for literally hours per day, and are plagued with crippling insecurity and self-loathing, but who fail to meet the necessary criteria (underweight BMI and amenorrhea) because of different, shall we say, somatic calibrations.”

        As far as I can see, those individuals are (should be!) picked up by the EDNOS category, and picking such individuals up is presumably a major part of the reason why this category still exists.

        As for behavioural or physiological metrics, I’m not so sure in particular physiological metrics are not useful; there’s pretty good data on for example at which point on the BMI scale that bone health starts to be affected, and using physiologically derived cut-offs to inform treatment decisions seems to me like a no-brainer; you also have for example a criterion where underweight women with very low heart rates should be admitted immediately, even if they’re completely asymptomatic (“Main guidelines advise to hospitalize [1, 2] and strictly monitor a subject presenting a heart rate below 40/min, even if asymptomatic”). Cut-offs like these make a lot of sense to me – and it’s important to remember that one of the main uses of diagnostics is to inform clinical treatment decisions. A major implicit point of this book is also that a lot of the women who have AN or BN likely have symptoms they might not associate with their weight; symptoms which would go away if they gained weight – I haven’t discussed the symptoms in detail, but there are a lot of them. An implicit point made in the context of the admission criterion mentioned above (and the mortality rates reported in the post) is also that ‘mental suffering’ is not the only relevant outcome at stake; ‘psychological parameters’ are important, but they’re not the only relevant ones here – many people die from AN or BN, and then you’re back to the age-old discussion of to which extent ‘society’ should allow people to kill themselves/engage in (/informed?) risky behaviours.

        I would probably agree with you that an argument could be made that from a cost-effectiveness point of view more resources should be spent on mental illnesses (including eating disorders) and I definitely agree with you that in this area comorbidities are really important, and I would add that this particular literature (eating disorders, that is, not mental illness in general – I have read some stuff on the latter in the past) is not a literature with which I’m familiar, so I don’t really have a good idea of the extent to which current treatment regimes are informed by CE considerations. My own impression would be that search costs are potentially quite large in this area, and that this aspect, perhaps much more than direct costs of treatment, is what limits health care provision to this (potential) patient population, but this is based on work I’ve done on screening models in general settings and as already mentioned my knowledge of what the CE-literature has to say on this particular topic is pretty much non-existent.

        Thanks for the comments, incidentally!

        Comment by US | January 3, 2016

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: