It is sad to acknowledge, but eating disorders such as anorexia nervosa have been reported in individuals as young as 5 years old. However, as one may imagine, eating disorders in young children are quite rare and thus very difficult to study. Fortunately, there are very sophisticated surveillance systems that have been established in Canada, the United Kingdom, and Australia. These systems asked physicians working in the community to report on individuals that they encounter who are diagnosed with a rare condition. The physicians report on this case to the surveillance system and then are sent a more detailed questionnaire to complete about the child’s diagnostic history. One of the priorities of the surveillance system was to better understand the emergence and course of eating disorders when they appear in children. This method of data collection is quite important because it enables us to understand what eating disorders look like in the general population. Often, research is published in children who present to specialty treatment clinics. While the presentation of these two groups may be similar the only way to know this is to examine both.
It is of additional importance to understand a bit about the history of how psychiatric disorders were first identified and named to appreciate the sophistication of the work done in this paper. Historically, we learned about clinical conditions through very detailed case studies. These descriptive narratives would paint, in great detail, the clinical presentation of rare cases. We still employ this methodology today. It is of great utility in discovering clinical conditions or features that had not been previously observed. However, the field of psychiatry and psychology wanted to advance beyond these descriptive narratives and to employ more data-driven approaches in understanding mental illness. The shift to data-driven approaches has been steadily evolving over decades with our statistical methods growing increasingly sophisticated.
One method for examining profiles of symptoms that may occur together is called the latent class analysis. Basically, what this method does is to examine whether there are clusters of features that seem to co-occur - or more precisely, that there are certain patterns of responses that occur together and that are distinct from a different class of responses that tend to occur together. Once you have identified these patterns, a researcher can examine if these co-occurring patterns of features map on to current clinical syndromes or whether they reveal new syndromes of which we’ve been ignorant.
This methodology had been employed previously in adolescent and adult samples and more or less confirmed the importance of diagnoses such as anorexia nervosa and bulimia nervosa. The authors of the current study were interested in employing this statistical methodology to better understand eating disorder diagnoses that emerged in children. Their reasoning was that eating disorders in children may appear differently than eating disorders in adolescents and adults. Thus analyzing children, adolescents, and adults as one group may muddy the waters and lead to less understanding about what eating disorders look like in kids.
The basic design of the study was that the researchers gather data from the surveillance system in Canada, the United Kingdom and Australia for a period of time ranging from 14 months to 36 months. Every month, physicians and psychiatrists in these locations were asked whether they had seen any new case of eating disorders that emerged in childhood. Cases were defined as children who were ≤12 years old (though this was extended to up to 13 years in cases collected in Australia). To be considered an “case”, a child in this age range had to exhibit what the authors referred to as “determined food avoidance” combined with weight loss or failure to make expected weight gains. What is meant by determined food loss is not defined specifically, but usually refers to intentional food avoidance (as opposed to limited food intake due to cruel circumstances such as poverty). Physicians who identified a case were then sent a questionnaire detailing the child’s basic demographic information (for example, sex, race) as well as medical complications and medical history. One impressive aspect of the surveillance system is how responsive the physicians were with it: up to 96% of physicians completed the follow-up surveys they were sent! Very impressive!
So what did they find?
Well, before I tell you, it is always important to know a bit more about these children so that you can better compare the findings to your own children. The age ranges they examined were children from five years old to 12 years old. However, one slightly confusing aspect from my perspective is that they eliminated children with overlapping eating disorder symptoms. I’ll explain why there may further unexplored important results in a bit.
I mentioned earlier that latent class analysis is about looking at profiles of symptoms. The researchers chose four variables to look at based on expert opinion and prior studies of key features of eating disturbance in children. The four features were 1) a fear of being fat or gaining weight, 2) misperception of body size or weight (for example, believing that you are normal weight when actually you’re quite thin), 3) over exercising, and 4) somatic complaints (for example, physical complaints like stomachaches). The results indicated that the researchers discovered two key groups. One group corresponded with the child version of anorexia nervosa. Much like the presentation of anorexia nervosa in adolescents and adults, this group had greater fears of weight gain, were preoccupied with their weight, were more likely to misperceive their body size, over-exercised and were preoccupied with food. This group in general was of lower weight, was more likely to be hospitalized, and was more likely to have unstable vital signs – a sign of medical compromise due to low weight. In contrast, the second group appeared to correspond to the diagnosis of Avoidant Restrictive Food Intake Disorder.
In addition to an absence of the symptoms of misperception of weight, over exercise, and fear gain, the trajectory of weight loss was also different between groups. The anorexia nervosa group appeared to lose a great deal of weight over a short period of time, in contrast to the low weight of the ARFID group that appeared to have begun to lose weight at an earlier age and the weight that was lost (or not gained with growth) occurred more gradually. The lower amount of weight lost and the more gradual weight loss may have resulted in less need for medical hospitalizations and less signs of physical instability.
There were several interesting features about the ARFID group. First, this group was more likely to have a comorbid psychiatric disorder – most notably anxiety. This is notable because individuals with anorexia nervosa are also hallmarked by their anxious temperament and premorbid anxiety disorders. Thus, to have a group that’s even more anxious than those with anorexia nervosa is worth pondering over! Second, the second group was more likely to have somatic complaints- symptoms such as stomach aches, or headaches, for example.
So what is this mean for parents or children trying to figure out their child’s or their own diagnosis?
Well to answer that, it would first be interesting to know how many children were excluded from the latent class analysis because of overlapping eating disorder symptoms. ***Stay tuned for interview with Dr. Sloan Madden to learn more about the details of the study.**
The seemingly simple answer to this question is that anorexia nervosa is marked by a fear of weight gain and preoccupation with food and body weight and that those with ARFID do not have these fears and preoccupations. This seems to be a straightforward distinction. In clinical practice, however, these distinctions may not be so straightforward, particularly in the arena of the symptoms of fear of weight gain. In anorexia nervosa, the motivation for fear of weight gain is thought to be a fear of fatness. This is often thought of as motivated by societal pressures to maintain a thin physique, a path communicated by societal forces that is alleged to lead to greater popularity and acceptance by others. However, the motivation for weight loss in anorexia nervosa can’t be that simple. The majority of individuals face these societal pressures, yet only a very small fraction develops anorexia nervosa. Thus, one theory, one investigated by our own group, is that one of the functions of weight loss in anorexia nervosa is to mute aversive sensory somatic experiences and emotional experiences (e.g., feeling bloated, sad and heavy; feeling the flutterings of butterflies in your gut or the pit of guilt or dread). Losing a great deal of weight may “quiet everything down” and contribute to those with anorexia nervosa “feeling better”.
A similar, though not completely overlapping, phenomena may be occurring in children with ARFID as well. Given their higher levels of somatic complaints, their weight loss may be negatively reinforced as these children also do not like to feel strong sensations from their bodies. Thus, while they are not afraid of weight gain due to societal pressures of weight stigma, gaining weight may conjure up different fears of discomfort, bloating and fullness. Thus it will be interesting to see how many children were eliminated from the study due to overlapping eating disorder symptoms and examine what the nature of this overlap may be (see the image below for some ideas).
For a parent, exploring with your child what it feels like after they’ve eaten a meal, whether they have any fears related to getting stronger and healthier by gaining weight, and how they react to sensations such as fullness or the physical sensations of emotional experience may help you to develop a better understanding of what they’re going through.
In summary, this was an exceedingly elegant study of a sophisticated data collection mechanism. By having physicians and psychiatrists be on the lookout and monitor the status of rare diseases, we can learn more about these dangerous conditions in a way that would not be possible without this data collection surveillance system.
By Nancy Zucker, PhD