When most people think of bulimia nervosa, they think of binge eating and self-induced vomiting. While that is not incorrect, it is not the full picture either. In the current edition of the Diagnostic and Statistical Manual (DSM-IV), there are two subtypes of bulimia nervosa: purging (BN-P) and nonpurging (BN-NP). The difference lies in the types of compensation methods: patients with BN-P engage in self-induced vomiting, or the misuse of laxatives, diuretics, or enemas whereas patients with BN-NP use fasting or excessive exercise to compensate for binge eating.
How common in BN-NP? It is very hard to say. A small population-based study in Finland (less than 3,000 participants) found that 1.7% of the sample that bulimia nervosa, 24% had BN-NP (or 0.4% of the entire sample) (Keski-Rahkonen et al., 2009). (I couldn’t find much else on prevalence of BN-NP.)
Unfortunately, however, there’s been very little research on BN-NP.
So little, in fact, that many have wondered if it make sense to subtype bulimia nervosa patients into purging and nonpurging groups? And are there differences between patients with BN-NP and BN-P or patients with BN-NP and binge eating disorder (BED)?
These questions matter because diagnoses are meant to be useful:
The main purpose of the DSM is to be clinically useful, namely to improve the assessment and care of individuals with mental disorders. Therefore, a crucial element of the diagnostic criteria is predictive validity: a diagnostic label should make it possible for clinicians to obtain and apply information on course, complications, and effective treatment options.
In 2009, Daphne van Hoeken and colleagues reviewed and summarized studies comparing BN-P with BN-NP and BN-NP with BED in an attempt to answer these questions.
Between 1992 and 2009, they found 23 studies that included BN-NP and BN-P and/or BED subjects. The studies were generally small: 7 had <25 subjects for BN-NP, 2 had <25 subjects for BN-P, and 1 had <25 for BED. Only 2 studies addressed the course of the disorder and none compared treatment response.
Out of those 23, only 14 addressed course of illness, complications/comorbidities (more comorbidities usually suggests a more severe course of illness and less favourable outcome), and treatment response. I’ve summarized the main findings below.
SUMMARY OF MAIN FINDINGS
Comparing Bulimia Nervosa Subtypes: BN-NP and BN-P
- Course of Illness:
- 2 studies, <1 year follow-up (Bulik et al., 2007; Hay and Fairburn, 1998)
- Rates of recovery were lower for BN-P than BN-NP (and highest for those with BED) (Hay and Fairburn, 1998)
- Bulik et al. (2007) studies pregnant women and found that rates of remission were lower for women with BN-P than for those with BN-NP. (But it is difficult to know how generalizable these are to nonpregnant women.)
- 11 studies
- No differences:
- Suicide attempts (Corcos et al., 2002; Milos et al., 2004; Youssef et al., 2004)
- General psychiatric symptoms (Hay and Fairburn, 1998)
- Comorbid axis I disorders (Striegel-Moore et al., 2001; Tobin et al., 1997; Walters et al., 1993)
- Lower rates of impulsivity for BN-NP than BN-P (Favaro et al., 2005)
- Lower rates of major depression, alcohol abuse, anxiety disorders, and social phobia in women with BN-NP than BN-P in a study by Garfinkel et al. (1996)
Comparing Nonpurging Bulimia Nervosa and Binge Eating Disorder
- Course of Illness:
- 2 studies (same as above)
- Hay and Fairburn (1998): higher remission rate and less severe ED symptoms in BED compared with BN-P and BN-NP
- Bulik et al. (2007): lower remission rate and more new onsets of BED than BN-P or BN-NP. Again, unclear how generalizable these findings are to nonpregnant women and to what extent “appetite changes associated with pregnancy influenced the evaluation of binge eating.”
- 8 studies
- 5 studies found no significant differences in comorbidity between BED and BN-NP (Ramacciotti et al., 2005; Santonastaso et al., 1999; Striegel-Moore et al., 2001; Vervaet et al., 2004 and Tobin et al., 1997)
- Hay and Fairburn (1998): lower levels of psychopathology and better social adjustment in BED than BN-P
At least one study did not make it into this review because it came out 2 years after this paper was published. In 2011, Nuzen-Navarro et al. published a study comparing 34 BED, 34 BN-NP, and 34 BN-P patients on clinical and personality variables. While thirty-four is still a small sample size, it is much better than many of the studies reviewed by van Hoeken and colleagues.
They found that BED patients were significantly older and were more likely to be married than either of the BN groups. Interestingly, they found no differences in personality traits across the three disorders. But there were some important differences:
The dimensions of clinical severity suggest dimensional differences across the three diagnosis with BN-P representing the most severe and BED the least severe. The sole but important difference that emerged was on obesity and related family history of obesity. Obesity is much more strongly associated with BED than with either form of BN.
That seems to hint at the idea that BN-NP is more closely linked with BN-P than BED. But again, that’s hard to see because we are not entirely sure on the prevalence of BN-NP. That might be at least in part due to the fact that “there may be a problem in diagnosing BN-NP] subjects. Individuals who would qualify for the diagnosis BN-NP may go unnoticed or be wrongly diagnosed as BED or ED-NOS as a result of incomplete assessment of nonpurging compensatory behaviors.”
van Hoeken et al. point out that both dieting and exercise are common in the general population and are not necessarily pathological. Moreover,
There is no clear criterion to decide at what point the amount of exercising and dieting exceeds a cut-off point and becomes abnormal. This does not mean that nonpurging compensatory behaviors are clinically irrelevant [...] their frequency is associated with severe maladaptive core beliefs and they are associated with impaired social functioning.
All in all, some studies found that BN-NP is closer to BN-P while others found that it is closer to BED. But most of these studies had small sample sizes, making it difficult to form any strong conclusions. Nonetheless, there does seem to be support for a hierarchy in severity with BN-P being more severe than BN-NP, and BN-NP being more severe than BED.
Of course, subtyping based on the presence or absence of compensatory behaviours or based on the types of compensatory behaviours is just one option. There are others.
van Hoeken et al. conclude by stating that there is little evidence for the utility of the BN-NP subtype and it is not clear whether it is closer to BN-P or BED.
It seems that the DSM-5 Eating Disorder Work Group agreed. As far as I understand, the subtypes will be removed from the upcoming DSM-V (scheduled to be released next month!) It seems that this decision is, at least in part, based on (1) unclear definitions of what exactly counts as “nonpurging compensatory behaviour” and (2) the conclusion that research suggests patients with BN-NP resemble those with BED more than BN-P.
Personally, I agree with the first point, but I’m not sure I agree with the second. That’s not to say I disagree, it is more that I am unconvinced either way. I don’t think there’s enough research out there for us to make any concrete conclusions. I do feel that classifying these disorders based on compensation methods is not necessarily misguided and may be the most optimal way to go about things (we need more research!) but I also don’t quite understand why misuse of laxatives and diuretics counts as “purging” but excessive exercise doesn’t.
I do, however, think that self-induced vomiting suggests a more severe psychopathology as does multiple compensatory behaviours (and this seems to be supported by the research). But I am having a hard time with the idea that patients with BN-NP who engage in excessive exercise (say, so excessive that most of us will agree it is excessive) or extreme fasting may suddenly find themselves in a weird limbo state when it comes to diagnosis–not quite BN-P (according to the criteria) and not really BED. Surely there are at least some differences in illness course, complications, effective treatment modalities and/or treatment outcomes?
Before I wrap-up, I want to be clear on the following: stating that patients with BED are less clinically severe as a group DOES NOT mean that all patients with BED are “less severe” than all patients with BN-P or BN-NP.
All in all, I remain pretty agnostic about this issue, but I am interested to see what will happen after the DSM-V comes out. Readers, what are your thoughts? Do you think there’s a better way to classify BN and BED (remember, classifications need to be useful)?