Treating anorexia nervosa is hard. Treating chronic and severe anorexia nervosa is a lot harder. Although the situation seems to be improving, there are really no evidence-based treatments for anorexia nervosa – particularly for those who have been sick for a long time.
Patients with severe and enduring anorexia nervosa have one of the most challenging disorders in mental health care (Strober, 2010).They have the highest mortality rate of any mental illness with markedly reduced life expectancy (Harbottle et al., 2008). At 20 years the mortality rate is 20%, and given the young age of onset this results in many young adults dying in their 30s, and a further 5–10% every decade thereafter (Steinhausen, 2002)… Patients are often under- or unemployed, on sickness benefits, suffer multiple medical complications… have repeated admissions to general and specialist medical facilities, and are frequent users of primary care services (Birmingham and Treasure, 2010; Robinson, 2009).. pose a significant burden to parents and carers (Treasure et al., 2001).
The treatments that many often claim are evidence-based are often only applicable to a select subgroup of ED patients, and even then, the evidence is usually weak. (I’m referring to Maudsley/FBT (family-based therapy) for adolescents with <3 yrs duration of AN and CBT for bulimia nervosa.) But what about those with long-standing anorexia nervosa? In a recent review, Phillipa Hay and colleagues set out to conduct a systematic review of randomized controlled trials of treatment for chronic AN.
Randomized controlled trials or RCTs are at the heart of evidence-based medicine:
The randomized controlled trial is one of the simplest but most powerful tools of research…[it is] a study in which people are allocated at random to receive one of several clinical interventions… the methodology of the randomized controlled trial has been increasingly accepted and the number of randomized controlled trials reported has grown exponentially… and they have become the underlying basis for what is currently called “evidence-based medicine”.
Hay et al searched the literature to identify RCTs where, among other criteria, the mean duration of illness was at least three years. They found eleven studies, but they could only confirm that a majority had a mean duration of over 3 years in just four of those studies.
I’ve summarized the studies in Table 1 (click to enlarge), highlighting the differences that were found between the treatment and control groups. (Side note, those that I labelled barely significant had p values of around 0.3):
In addition to the aforementioned RCTs, Hay et al identified several case series and studies (which were not RCTs) that have specifically tried to address the perpetuating factors that lead to long-standing anorexia nervosa. The table below summarizes these novel treatment approaches.
One novel approach (Fairburn, 2010) uses a variant of cognitive behavioural therapy (CBT), developed specifically for chronic anorexia, called CBT-E (E for extended). In addition to the regular CBT modules, CBT-E includes additional modules that focus on factors that maintain and perpetuate AN, as well as dealing with mood intolerance, interpersonal relationships, perfectionism and self-esteem. In a study utilizing this approach conducted by Byrne et al. (2011), out of the 34 patients with AN (2 males, 32 females) with a mean duration of illness slightly over 10 years, 12 completed the study, 3 attained partial remission and 3 attained full remission. Full remission was defined as having a BMI of 18.5 and above, Eating Disorder Examination (EDE) scores less than 2.46 (which is within 1 standard deviation from the community norm), as well as the cessation of restricting and/or binge/purging behaviours.
That sounds pretty okay so far (remember the average duration of the illness is over 10 years). But here’s the kicker, I went to check the original paper: the “full remission” is based on the absence of symptoms for… drum-roll… 28 days. Err what?
The other trials showed similar trends; there’s some improvement but often the weight gain is not sufficient to be in the normal weight range and the symptoms of rigidity and perfectionism are still present. That’s not to say that these approaches are useless. They are not: any improvement is better than no improvement – especially for someone with a long history of AN.
If you take a look at Table 2, you’ll see that a lot of the approaches are centered around CBT in combination with motivational interviewing. The idea, at least in part, is to address the thinking styles common in AN that tend to perpetuate the disorder. Also note that while weight gain is a goal in many treatment approaches, it is not mandatory and/or emphasized as the primary goal.
Now, I am predicting that a lot of Maudlsey/FBT proponents will comment about the importance of weight restoration and how psychological and cognitive changes cannot happen until normal weight is restored (or at least wont be as effective). That might be true for some BUT it is definitely not true for all. To say that it is, is to deny the very real experiences of many ED sufferers.
But, there are a few even more salient points I want to make:
- we are talking about long-lasting, chronic anorexia nervosa: at least 3+ years, but in many cases it is 6, 8 and 10, even 15+ years.
- most of these patients are adults, in their mid to late 20’s, and some are older (average age in one study in Table 1 was ~45)
- these patients likely failed (were kicked-out or dropped-out) from previous treatments, probably in inpatient care or residential treatment centers, or they relapsed soon after discharge, in other words: other treatments didn’t work for them, the treatments failed them as much as they failed the treatment.
I’m going to harp on this a little more because it is an important but often ignored issue. You wouldn’t treat stage I cancer the same way you’d treat advanced stage IV cancer, right? Nor do you treat mild cases of depression the same you’d treat long-lasting chronic depression, right? To keep trying to same approach, or slight variants of the same approach for all patients, regardless of age, gender, sexual orientation, ethnic background, family relations, socioeconomic status, duration of illness, comorbid psychiatric disorders, and so on, is.. well, stubborn to say the least. Almost dogmatic.
And don’t get me wrong: I do not think long-lasting anorexia (or bulimia) is untreatable, nor do I that high heritability means genetic determinism (it doesn’t). Eating disorders, not unlike cancer, are heterogeneous. To quote myself from my very first blog post:
… when comes to anorexia, genetics and the environment both play important roles. Furthermore, the precise influence of genes to environment will vary from person to person. For some patients, genetics may be the main factor, while for others, the environment might play a more important role in leading to anorexia. Naturally, this complicates the picture, in terms of research but also in terms of treatment. Perhaps – and this is just my hypothesis – this may explain why patients respond differently to various treatments. Perhaps a large part of it is due to differences in the degree to which genetic and environmental factors played a role in the development and maintenance of the patients eating disorder. Perhaps if we were able to determine the extent of, say, genetic influence in any given patient, we would be able to tailor treatment for that person…
But anyway, the other reason I’m bringing all of this up is because I want to point out a study that essentially used a harm reduction approach. Williams et al. (2010) (see Table 2) utilized a treatment approach meant “to improve quality of life and minimize harm.” The goals and the pace of therapy are “set by the client and the focus is on symptom management, skill development and understanding benefits and risks of symptoms.” Moreover, this approach doesn’t require inpatient or residential stays (which are problematic in their own ways): it is set in the patient’s own community. Medical safety, however, is a “non-negotiable” goal.
In 18 patients with a mean duration of AN for 15.2 years, this approach, called COPP (Community Outreach Partnership Program), led to significant improvements in some areas (but not all) outcomes that were evaluated after 25 month of treatment.
Note the link above (or here) leads to the full Williams et al paper, I shared it through my Dropbox folder because I think it is an important read. I strongly suggest everyone checks it out (although the excerpts about patients with EDs can be somewhat triggering or upsetting). We cannot forget about the patients that drop-out of treatment, relapse or stay chronically sick in and out of hospitals. And we cannot keep using the same treatment approaches and hoping that they will magically work, even though they’ve failed before.
In this ‘new kind of paradigm’ for chronic illness there is an emphasis on early treatment alliance and not hastening into a specific intervention, weight gain is not the principal objective, re-feeding is negotiated carefully and collaboratively. Nutritional improvement, however, is encouraged, routines of social activity are established, hobbies and pursuits encouraged, regular medical examinations done, and regular meetings are undertaken with family members and supportive others. [Strober (2010)] emphasises the professional challenge of working with these patients, the ability to ‘not require successes measured by patient progress…to face frailty and profound sickness with ease’ and therapists avoiding the ‘risks of treading too heavily on their (the patient’s) efforts at salvaging “marginal comforts.”’…
The authors stress the need for more randomized controlled trials in evaluating the effectiveness and efficacy of treatments, but at the same time noting the importance of open trials in exploring new treatment approaches.
I think when looking at treatment approaches and determining what works and what doesn’t, we cannot put all anorexic patients into one group, or simply divide them based on the presence or absence of bingeing and purging behaviour. I think the effectiveness of therapies that might work incredibly well for a subpopulation of patients gets diluted in a large trial by subpopulations for whom the treatment is ineffective (or harmful). I think clinicians and health professionals who work with ED patients need to listen to them as well as their families/caregivers/partners. Alliance with the patients and, when applicable, the patients family, is crucial for treatment to be successful.
It is hard. There’s no question about that. But treatment that may work for a self-aware, highly educated single female in her late 20’s or early 30’s, who is estranged from her family, will be different than treatment that may work for an adolescent female who is in denial of her illness and lives with her stable middle-class family. Replace “female” with “male” or “trans female” or “trans male”, or add some psychiatric comorbidities, and without question the best approaches to treatment change, too.