When I was first introduced to the Minnesota Semi-Starvation Study (MSSS), published in 1950 by Ancel Keys and his team, I was overwhelmed by the sheer volume of data they were able to accumulate about the physical and psychological ramifications of semi-starvation in humans.

It was mind-blowing. Then I came across the graph on page 106 of the hardbound edition called “Over-all Changes in Body Weight in the Minnesota Experiment” with the subtitle: “Expressed as a percentage of the control value for the particular group of subjects”. I invite you to Google it.

The Y axis (the vertical one) of this remarkable graph showed “% of control weight achieved” (100% at the start of the study, before the subjects were dieted). On the X axis (the horizontal one) was the “number of weeks of the experiment,” which stretched out to 58, or one year and 1.5 months.

Over the course of the study, the subjects were fed a calorie-reduced diet designed to mimic war-time semi-starvation rations; the subjects’ body weights could be seen to drop to about 77% of “normal” by about week 24. So they were dieted for 24 weeks.

Thereafter came 12 weeks of “structured re-feeding,” which is what we do at Kartini Clinic -- a process designed to give a healthy mixture of nutrients, but in a controlled fashion. Around the 12th week of re-feeding the subjects were allowed to eat “ad-lib” (or “intuitively” in the parlance of our day), with no content restrictions. This style of eating caused a very steep rise in the subjects’ rate of weight gain and resulted in them gaining 110% of their control weight. In other words, their weight went up higher than it had ever been before the diet.

The University of Minnesota published a video about this experiment 15 or so years ago. In it, they interviewed one of the “survivors” of this study, now an older man. He said that once they were allowed to eat what they wanted, they ate everything, often indulging in the sweets and desserts they had denied themselves during the period of semi-starvation. “We got fat” he said, and seemed to find it amusing. Following the above-referenced curve out to the 58th week of the study, though, you can see that the overeating appears to have subsided and the subjects’ weight to curve back down towards their normal weight -- although they never really reached as low a weight as they had before they started dieting.

Taking a structured approach to refeeding at Kartini Clinic

OK, it may have amused some of the young men to “get fat,” but I reasoned that our patients (and their families) would not find this amusing at all. At Kartini Clinic, we are searching for the state of health that allows children to grow and thrive. And, in an effort to keep the growing body from “over-shooting the mark” with re-feeding and unnecessarily promoting binge eating (which appears to be common following semi-starvation and weight loss even in people who have never binged before), we decided to remove “hyper-palatable” foods for about a year encompassing weight restoration. Hyper-palatable foods are those with high sugar and fat content -- neuroscientists’ definition, not ours.

If you look at Key’s graph, waiting a year to re-introduce these foods should take us out of the area of binge eating into more normally driven food intake. Simple, right?

I think that nothing that we have innovated at Kartini Clinic has caused us to be so roundly criticized as this suggestion that our patients avoid sweets during the initial phase of re-feeding. We immediately ran afoul of the “intuitive eating” proponents although, to date, no one has explained the medicine/science behind their objections. Ancel Keys’ research subjects did not have eating disorders and 12 weeks of ordered refeeding did not cause them to “fear foods” or to develop an eating disorder. Eventually, they went back to normal eating...a year plus later. That is our plan as well.

Once binge eating behavior starts, it’s hard to stop

In addition to studies about the bingeing and food seeking behavior that can follow dieting or restricting, there is also a plethora of animal studies specifically looking at bingeing and hyper-palatable foods. Bingeing in mammals is a product of gene-environment interactions. The “gene” part is how vulnerable to binge eating your brain is because of your genotype, the “environment” part is the food you are exposed to.

You cannot tell whether someone is binge-prone just by looking at them. From rat studies, we know that this tendency sometimes does not surface until puberty or later. In binge-prone rats “intake of just a morsel of PF [palatable foods] or simply the smell of PF is known to trigger relapses back to binge eating."[1]

For those of you who like reading the science, Chapter 8 of the same text, by Michopoulos, Moore and Wilson, discusses the biology of diet choice and emotional feeding in times of stress, which will vary from one being to another. While brain chemistry causes these behaviors, they can appear to be lifestyle “choices” for which a patient is blamed.

Other authors (Ventura, Latagliata et al) note in their chapter on “Food Seeking In Spite of Harmful Consequences”: "…repeated stimulation of reward pathways through highly palatable food consumption may lead to neural adaptations that make consumption more compulsive.”

I ask you, why go there? In anorexia nervosa, simply refraining from adding hyper-palatable foods to our patients’ meal plans would seem to offer the possibility of a smooth return to normal eating. As we say: First do no harm, then innovate.

And as I say: Disagree? Show me the science.


[1] Animal Models of eating Disorders; Avena, N (Ed.); Humana Press 2013 ISBN 978-1-62703-103-5) page 8.