Over-Exercise is Associated With Suicidality in Individuals with Disordered Eating

Last week, I blogged about a study that examined personality traits and clinical variables associated with excessive exercise in eating disorder patients. In that study, 2 out of 5 participants engaged in excessive exercise. Today, I’m going to discuss a study that suggests over-exercise in disordered eating patients is associated with suicide behaviour.

Suicide rates in eating disorder patients are high. One meta-analysis suggested that out of all eating disorder related deaths, 1 in 5 are suicides. (Keep in mind, these numbers are really hard to pin down as they depend a lot on the sample population, sample size, and how the authors did their statistics, among other things.)

Another analysis found that the standardized mortality ratio (ratio of observed deaths in the study sample/expected deaths in the population of the same age but without the disease/disorder you are studying) for suicide in eating disorders was 31 for patients with anorexia nervosa and 7.5 for patients with bulimia nervosa. Moreover, around 25-35% of bulimia nervosa and 3-20% of anorexia nervosa patients attempt suicide at least once in their lifetime (Bulik et al., 1999; Corcos et al., 2002; Franko and Keel, 2006).

Clearly, understanding what sub-population of ED patients is most vulnerable to suicide ideation and suicide attempts is important to develop proper screening and treatment approaches.

In 2005, Thomas Joiner developed a theory to attempt to understand and explain suicide. The interpersonal-psychological theory of suicide (IPTS) posits that there are three essential components that must be in place before one commits suicide:

  1. thwarted belongingness (feelings of loneliness, perception that one has no meaningful relationships)
  2. perceived burdensomeness (feeling that one’s death is worth more than one’s life)
  3. acquired capability of suicide (ability to inflict potentially lethal self-harm which is achieved by subjecting the body to painful experiences)

The authors of the present study wanted to use the IPTS framework to examine there’s a relationship between acquired capability as a result of eating disorder behaviours and suicidal behaviour.

We were specifically interested at looking at the disordered eating behavior of over-exercise due to its noted association with pain tolerance (e.g., Ryan and Kovacic, 1966) and suicidality (e.g., Brown and Blanton, 2002), and because over-exercise often results in pain and injury (e.g., Veale, 1987; McKenzie, 1999). Behaviors such as vomiting and laxative abuse involve pain and bodily damage […] Further, prolonged periods of restriction require one to over come hunger pains and intense discomfort. However, we hypothesized that over-exercise would be a stronger predictor of acquired capability for suicide than other compensatory behaviors.

The authors conducted four different studies to test the following ideas:

  1. over-exercise is associated with suicidal behaviour
  2. over-exercise is associated with acquired capability for suicide
  3. pain insensitivity mediates the relationship between over-exercise and acquired capability for suicide
  4. acquired capability for suicide accounts for relationship between over-exercise and suicide.

Here’s a graphic to explain their hypothesis (with the studies testing each hypothesis in brackets):

Smith - 2012 - Figure 1


Study 1: Over-exercise was the only significant predictor of suicidal behaviours and suicide attempts in a clinical sample of bulimia nervosa patients. It was more predictive of suicidal behaviours than vomiting, dietary restraint, laxative abuse, and demographic factors such as age.

Study 2: In a sample of 171 undergraduates, over-exercise predicted the acquired capability for suicide above other eating disorder symptoms (vomiting, dietary restraint, etc.), suggesting that “over-exercise is a unique contributor to the acquired capability for suicide.”

Study 3: This study of 427 undergraduate revealed that over-exercise was related to pain insensitivity, and that pain insensitivity accounted for the relationship between over-exercise and acquired capability for suicide. Two important limitations should be noted: the authors measured pain insensitivity and not pain tolerance (which are related, but different), and they examined the relationship only at one time point, so longitudinal studies  are needed to clarify whether over-exercises leads to pain insensitivity.

Study 4: In the final study, the authors found that in a sample of 512 undergraduates, acquired capability for suicide accounted for the relationship between over-exercise and suicide attempts.

The authors conclude that,

Overall, the results of four studies converge to suggest that over-exercise is related to suicide attempts through its association with pain insensitivity and acquired capability for suicide […]These results suggest that over-exercise can be hazardous to individuals not only directly—via overuse injuries—but also indirectly, via the acquired capability for suicide. 

The authors caution that these results are preliminary, and prospective and longitudinal studies are needed to replicate these findings.

There are important limitations to these studies: (1) the definition of over-exercise varied between the four studies (however, the definition has also varied widely in previous studies); (2) only one study used a clinical sample (exhibiting bingeing/purging behaviour), whereas studies 2-4 used undergraduates, thus it is unclear whether the findings are generalizable to a clinical population.

I think future studies should examine if these findings hold up in clinical sample of anorexia nervosa patients as well (given than excessive exercise seems to be most prevalent among purging-type anorexia nervosa patients, see previous post), and if other aspects of the model (thwarted belongingness, perceived burdensomeness) are also related to over-exercise.

Nonetheless, the strength of this paper is that the findings from individuals studies (which had different participants) converge on the same model. If these findings are replicated, the implications are obvious. For one, screening for over-exercise might identify patients who are at a higher risk of suicide, and then treatment approaches can be tailored to work on reducing over-exercise and dealing with other factors that might feed into suicidal ideation (such as the other components of the IPTS theory).


Smith, A., Fink, E., Anestis, M., Ribeiro, J., Gordon, K., Davis, H., Keel, P., Bardone-Cone, A., Peterson, C., Klein, M., Crow, S., Mitchell, J., Crosby, R., Wonderlich, S., Grange, D., & Joiner, T. (2012). Exercise caution: Over-exercise is associated with suicidality among individuals with disordered eating Psychiatry Research DOI: 10.1016/j.psychres.2012.11.004

Liked it? Take a second to support tet on Patreon!


Tetyana is the creator and manager of the blog.


  1. I wrote a lengthy comment on your previous post regarding over-exercise and EDs describing how my exercise behaviours were strongly linked to control of anxiety and mood. Compulsive over-exercise was the core feature of my ED. Exercise was what I thought about constantly; not weight, shape or eating.

    On a personal level I cannot identify with the model above and have never attempted suicide or self harmed. However, I have often had strong thoughts about suicide. These were most intense when my weight dropped so low that exercise became impossible. My suicidal thoughts were directly associated with an inability to exercise.

    My over-exercising led to serious bodily damage in the form of over-training syndrome, injuries and accidents. The latter included getting knocked off my bicycle a number of times (in part because I rode my bike recklessly) and fracturing my pelvis. I ended up in hospital on a few occasions due to cardiovascular collapse associated with over-exercising.

    I didn’t ‘use’ exercise for weight control but for mood control. I am not an impulsive person, which is something of a relief, because my immediate impulse, if unable to exercise, was to kill myself. I just didn’t feel I could cope with life unless I was near permanently ‘drugged’ by the effects of exercise.

    • I will add that I am actually very pain-sensitive – which might make one wonder how I could flog myself during exercise… The answer is that I was pretty much trance-like when I was exercising hard. I didn’t feel much at all. Perhaps the analgesic effects of endorphins released during exercise played a role?

      • I’m not a psychologist, so I don’t really know much about the psychology of suicide and how validated this model is (anyone?).

        Keep in mind, the acquired ability for suicide is just one of the 3 components of the IPTS, too. So, it would have to be, according to this model, over-exercise plus probably other components that lead to this, and the other two I mentioned.

        I think my biggest questions are about the IPTS theory, of which I don’t know much about. It strikes me as a bit odd that pain tolerance would be a component–many overdose on pills, for example or use other relatively painless methods. I don’t know if I am sold on the fact that the link between over-exercise and suicdality is through pain insensitivity, my guess would be that over-exercise is associated with increased severity of the psychological and physical symptoms of eating disorders, and that, not the pain insensitivity or pain tolerance, would increase the risks of suicide.

        I also can’t personally identify with it, but that doesn’t mean much, I’m just an n=1.

        • Have a look at this paper: I found it quite interesting when I read it a few years ago. I like that it is qualitative, although I would suggest that primary exercise dependence certainly exists.


          As far as I’m aware, the first author had an ED herself.

  2. Excess exercise causes hemo-dynamic changes. Hemo-dynamic overload influences the renin-angiotensin system.
    As for suicide relevance:
    “An association between the II genotype of the angiotensin-converting enzyme (ACE) insertion (I)/deletion (D) polymorphism and suicide was found ….”
    (source = http://www.ncbi.nlm.nih.gov/pubmed/19439995/
    (also = http://www.ncbi.nlm.nih.gov/pubmed/18521860/
    (more recently = http://www.ncbi.nlm.nih.gov/pubmed/20797799 “Decreased aldosterone in the plasma of suicide attempters with major depressive disorder”)

  3. My history of and motivation for my decade of overexercising is eerily similar to Cathy’s above. For me I will add that although I’ve suffered with depression/OCD/overexericse/ED crap for more than a decade, I have never “seriously” contemplated suicide. However, I know that with the overexercising and the damage I know that it causes, I am actually very passively killing myself.

    It’s slow, it can be veiled as “healthy” by outsiders, etc. and so it can remain a secretive form of self-harm, depending on the level of isolation. I would never take a gun to my head, but I exercise to make any anxiety wane. When I can’t exercise, it’s pretty much the only time I seriously freak out or feel much of anything. Just my two cents…

Comments are closed.