Are personality disorders overdiagnosed in EDs?

Lots of other mental health conditions commonly co-occur with eating disorders. I’ve blogged before about the links between OCD and EDs, and I’ve tangentially talked about links to depression and other anxiety disorders. But a paper published this week in the journal Psychopathology (von Lojewski, Fisher, & Abraham, 2013) reminded me that I haven’t yet blogged about the relatively strong and frequent co-occurrence of personality disorders and EDs.

First, some definitions. Quoth Wikipedia on the subject of personality disorders:

These behavioral patterns in personality disorders are typically associated with substantial disturbances in some behavioral tendencies of an individual, usually involving several areas of the personality, and are nearly always associated with considerable personal and social disruption. Additionally, personality disorders are inflexible and pervasive across many situations, due in large part to the fact that such behavior may be ego-syntonic (i.e. the patterns are consistent with the ego integrity of the individual) and are, therefore, perceived to be appropriate by that individual. This behavior can result in maladaptive coping skills, which may lead to personal problems that induce extreme anxiety, distress or depression. The onset of these patterns of behavior can typically be traced back to early adolescence and the beginning of adulthood and, in some instances, childhood.

Psychologists have broken personality disorders down into three main clusters. Cluster A is described as “odd or eccentric” disorders, and include paranoid personality disorder (PD), schizoid PD, and schizotypical PD. Cluster B contains the “dramatic, emotional, or erratic” disorders, which are antisocial PD, borderline PD, histrionic PD, and narcissistic PD. Lastly, Cluster C contains the “anxious and fearful” disorders, which are avoidant PD, dependent PD, and obsessive-compulsive PD (not to be confused with OCD). {The links lead to the Wikipedia pages on each disorder- not peer-reviewed research, but adequate for a brief intro to the subject.}

 Personality disorders in EDs

The two PDs most frequently seen in EDs are borderline PD and obsessive-compulsive PD. Other Cluster B and C PDs are also seen in EDs, although somewhat less frequently. Broadly speaking, the Cluster B PDs are most commonly associated with EDs that have binge/purge features, whereas the Cluster C PDs are seen more frequently in restrictive-type EDs, although exceptions are common. Generally, a psychologist or researcher will diagnose a personality disorder either after doing a structured clinical interview or by self-report and questionnaire.

Several studies have found that personality disorders are much more common in EDs than in the non-eating disordered population. According to the authors of the recent study, other studies have found that anywhere between 27-95% of ED patients also have a PD. The idea that 95% of ED patients also have a clinical personality disorder seems rather absurd to me. Even if it weren’t, the range is rather large, which suggested to the researchers that ways of assessing for PDs in people with EDs needed further study.

Instead of administering a paper questionnaire, the researchers in the Psychopathology study used a structured interview by a trained psychologist. Self-reporting on questionnaires, they wrote, tended to “greatly overestimate the prevalence of personality pathology.”  They interviewed 132 female inpatients at the Northside Clinic outside of Sydney, Australia between 2005 and 2010. Of the 132 inpatients, 20.5% had the restricting subtype of AN, 19.7% had the binge-purge subtype of AN, 32.6% had BN, and 27.3% had EDNOS. Ages ranged from 18-55.

Notably, the researchers also attempted to account for the neuropsychological effects of the ED when they interviewed the patients. “To avoid the impact of negative energy balance and associated behaviours on personality evaluation, interviews were conducted after patients had been on a supervised re-feeding or regular eating program, as appropriate, for at least 3 weeks,” the authors wrote.

The researchers found that avoidant PD was the most common diagnosis, with fully one-quarter of the patients qualifying for a definite or probable diagnosis. Borderline PD and obsessive-compulsive PD were the two other, relatively common PD diagnoses in the group.Self-induced vomiting was significantly associated with borderline PD, and was the only ED behavior associated with a PD diagnosis. The full table of their findings is below.

PDs in EDs

So what do these findings mean?

Although the researchers did try to control for the effects of malnutrition on the results, I doubt that three weeks of regular eating would fully eliminate any signs of ED-induced personality psychopathology. Personality disorders are, by definition, fairly stable over time (although new research is finding that they might not be as stable and life-long as originally thought), so ED recovery should not make a valid personality disorder go away.

When researchers looked at 54 patients who had recovered from their eating disorder (defined as no longer meeting the DSM-III-R criteria for anorexia and/or bulimia, which isn’t a very good definition of recovery, in my opinion), they found that 26% met the criteria for a personality disorder (Matsunaga et al., 2000). In a separate study of women who had been in recovery for more than one year, researchers found that 42% of patients with a history of AN qualified for a diagnosis of obsessive-compulsive PD, and so did 20% of people with a history of BN (Wagner et al., 2006). In addition, they found that certain personality traits, like novelty seeking and harm avoidance, persisted even after recovery and are thought to be present before the onset of the ED. A study of twins found that the dysregulated cluster of personality profiles (which corresponds roughly to symptoms seen in Cluster B PDs) were most strongly associated with bulimic symptoms, although it’s important to note that the researchers did not diagnose PDs or BN, just the symptoms of these disorders (Slane et al., 2013).

Conclude Wagner et al:

A wide range of classic ED symptoms persist after recovery and do not differ between subtypes of ED, suggest- ing that they are traits rather than state-related disturbances. Individuals who recover from an ED look remarkably similar to each other, and appear to have lower rates of some psychopathology, such as Cluster B disorders, than are typically reported in ill individuals. These findings may be useful in understanding factors associated with good outcome. As clusters of ED are more defined by personality features than by classical eating pathology, these data are relevant for treatment and prevention in childhood.

It could be that an ED exacerbates underlying personality issues and amplifies the PDs that are already there. I couldn’t find any studies that looked at the severity of PDs over the course of ED treatment. Still, although PDs might be overdiagnosed (borderline PD is a label that I’ve found has been readily applied to any “difficult” patient or one who regularly self-injures), especially when self-administered or assessed by untrained professionals, they are still a significant co-occurring condition that can affect what type of treatment is most appropriate and a person’s response to that  treatment.

Twitter Digg Delicious Stumbleupon Technorati Facebook Email

8 Responses to “Are personality disorders overdiagnosed in EDs?”

  1. Oh, very interesting! Did the researchers note any differences in PDs between ED groups? I’ve heard before that people with Bulimia nervosa are more likely to be Borderline.

    PS – I thought I was on ScienceofEDs for a minute due to the new format of your posts!

    • Not in the most recent paper, but there were in some of the others. I didn’t look at that aspect quite as closely, but borderline PD is commonly associated with bulimic symptoms, and obsessive-compulsive PD in restrictive EDs. But there are lots of exceptions; this is more of an over-arching tendency rather than a hard and fast rule.

  2. This is a very interesting article. I think treating each patient as an individual as opposed to making assumptions regarding personality disorders that may be present is the most effective approach to treatment.
    Eating disorders are highly complex and there are typically multiple factors that play into the development of the eating disorder.

    • This is very true. I think the question the researchers were trying to ask was the nest way to diagnose PDs and the best time at which to do so.

  3. I think an important thing about PDs in GENERAL, is they only get diagnosed when they are a problem. In case that sounds flip, what I mean is EACH of us have typical personality styles and traits that “fit” some one or the other (or more) of the DSM clusters. But the difference between a Trait and a Disorder is a matter of *degree* and not *type*. So even with the reported methodology, the interviewing clinician still knew that they were looking at a patient with at least one mental health diagnosis. And because MH dxs cluster *anyway*, this still means there is potential bias towards over-documenting.

    I will leave aside my BPD rant altogether, except to second what you wrote, and opine that of the PDs, this one is the most problematic.

    That said, to my mind, it’s super hard to accurately diagnose ANY PD when you’ve got frank symptomatology from some other organic process going on. Personality is a function of the brain, and if your brain is wonky, your personality is going to be affected; if your brain is wonky enough that you are currently an INPATIENT, your personality cannot be assumed to accurately reflect your baseline-functioning self.

    OTOH, as you say, PDs are definitionally pervasive across domains. Whereever you go, there you are, and all that. Our personality styles – disordered or not – have enormous determination over how we respond to which kind of treatment(s), what kinds of supports and ancillaries we need, what kinds of strengths and weaknesses we bring to our problem-solving. Understanding whether and WHICH PD may be co-occurring is a legitimate and important diagnostic task.

    In particular, OCPD has been identified as a predictor of poor therapeutic outcomes. OCDP is egosyntonic, and the person with it tends to frame the world in a both a B/W & very zero-sum kind of way. IOW, they believe the problem is that not enough of the universe is doing things their way*. It’s associated with poor treatment outcomes because A) ppl rarely present of their own accord – their is almost always coercion of one form or another B) they have low motivation for change / don’t see the problem C) they are critical of methods that don’t match what they already do and D) they rarely stay with treatment. On the plus side, if you can convince them you know what you’re doing and it will help them get something they want, they are EXTREMELY compliant.

    I think this paper is a nice step towards refining how people with ED get evaluated, and identifying important information that is relevant both to treatment and outcomes.

    *I read this thing about the differential btwn OCD and OCDP from a therapist. It was along the lines of “If you’re unsure which your client has, look at who is miserable. If it’s your patient, then it’s OCD. If it’s their family and co-workers, then it’s OCPD.”

  4. What about when having an eating disorder overlaps with having Asperger’s syndrome? Women with undiagnosed Asperger’s also have been misdiagnosed with personality disorders when actually they had Asperger’s all along but they were just flying under the radar because it is harder to detect it in women than in men as the traits are significantly different in the different sexes. The ‘female’ traits of Asperger’s are also not as familiar to doctors as the ‘male’, classical traits of the condition.

    • Very true. I’ve been wanting to blog on autism spectrum conditions and EDs, although I didn’t cover it here since it was a little off-topic.


  1. EDs Behind Bars: Eating Disorders in Prisoners | ED Bites - April 2, 2013

    […] stimulant drugs, and show signs of EDs (Rosler et al., 2009). Given the known associations between borderline personality disorder, substance abuse, ADHD, and EDs, it’s not unreasonable to hypothesize that EDs are also […]