Nature Reviews Neuroscience came out with a perspectives article today on overeating and obesity, and the evidence behind the food addiction model that is gaining popularity. It's an issue that I myself have given a lot of thought to: is there REALLY such a thing as food addiction that's just like heroin addiction or cocaine addiction? And if there is...what IS it? What are the criteria?
Ziauddeen et al. "Obesity and the brain: how convincing is the addiction model?" Nature Reviews Neuroscience, 2012.
(I suppose I could insert the mainstream media obligatory demeaning headless photo of an overweight person here. But you know what? It's demeaning. Source
EDITED: I have now replaced the sloth picture. I thought sloths were just adorable, and it didn't even occur to me that some people might interpret a baby sloth as "slothful" or "laziness" in this context. I really didn't mean that. Just wanted adorable. This kitten will do instead. Sorry, I really didn't mean to be offensive!)
Whether to call it "food addiction" is something that I have been intellectually wrestling with for some time. I see a paper showing that D2 receptors are decreased in the striatum of obese individuals, like you would see in cocaine addicted individuals, and I start to think the food addiction model may be the way to go. But then I see another study showing the opposite. Or I think of the severe impairments in function present in drug addiction...and overeating just doesn't seem to go there. Or I consider the fact that obesity does not necessarily mean that someone is "food addicted", and conversely, binge eating does not necessarily mean that someone is obese.
So it's very nice to see an article like this, which does a review of where the convincing and less convincing evidence lies, talks about the limitations of the addiction model, and then proposes directions for research and new clinical practice definitions to better fit the current evidence. And it happens to agree with a lot of my ideas of what we might call "food addiction", and what differentiates it from overeating, and especially what differentiates it from obesity.
There are two views for a model of food addiction:
2. Food addiction is a behavior resembling drug addiction seen in a subgroup of people with obesity, as defined by the DSM-IV.
These two views are not mutually exclusive, but the type of evidence available differs. They both rely on environmental factors like the increased availability of highly palatable foods (during my walk to the meeting where I surreptitiously read this paper under the table, I walked past no less than 5 fast food places and 2 Starbucks, increased availability indeed). The first relies more on neurobiological ideas of how drugs of abuse behave, while the second focuses more on the behavioral definitions of addiction as used in the clinic. The authors of this paper consider both perspectives, the evidence as well as the limitations.
Diagnoses in the Clinic
Here are the DSM IV criteria for substance dependence:
Tolerance: increasing amounts of drug required for intoxication
Withdrawal symptoms when the drug is discontinued, including depression, shaking, sweats, seizures, etc.
Persistant desire for and unsuccessful attempts to cut back or cease use
Larger amounts of drug taken than intended
A disproportionate amount of time is spent obtaining the drug, using it, and recovering.
Important social, occupational, etc activities are given up or reduced because of the drug.
Substance use is continued in spite of knowing and experiencing adverse consequences of use.
First, it becomes readily apparent that we are dealing with a small portion of the obese population. The vast majority of overweight or obese people do not binge eat, or even eat that much more than normal weight counterparts. The increased availability of food, decreases in physical activity, and other environmental influences are more than enough to create excess weight without addiction-style eating. When we talk about people who are exhibiting food addiction-like behavior, we are generally talking about people specifically who binge eat. These people may not necessarily be overweight, and so using a BMI in diagnosis isn't going to help here.
But back to the DSM IV for substance abuse. You can make some of these conform pretty easily to chronic binge eating. Binging on larger and larger amounts of food corresponds to tolerance, distress during dieting could account for withdrawal, unsuccessful attempts to cut use clearly corresponds to dieting, and so on.
But at the same time, all of these comparisons have drawbacks. For example, eating larger and larger amounts of food as a measure of tolerance requires a comparison between satiety (when you're full) and the intoxication shown with a drug. While there is some neurobiological evidence for reward (though not intoxication) associated with initial food exposure, it decreases swiftly as the person eating approaches fullness. Not only that, binge eaters regularly eat past the point of satiety, eat when they are not hungry, and continue to eat to physical discomfort, which doesn't relate well to the highs associated with drug abuse. For many of the criteria (particularly the last 5), there's no real established measure for when, exactly, food intake becomes harmful. There is no evidence of food withdrawal (though there is certainly evidence of craving). So while many binge eaters could meet three or more of the criteria of the DSM-IV necessary for a diagnosis of substance dependence (the most common being increasing amounts of food, unsuccessful dieting attempting, and eating larger amounts than intended), it's not clear whether those criteria which apply to substance dependence really apply to food.
So while some of these criteria may apply fairly well for people with binge eating disorders, they do NOT apply to the majority of people who are overweight or obese.
I often see (and blog about) new neurobiological and genetic correlations with binge eating and obesity that are similar to changes seen in drug abuse. With regard to genetics, there are some studies showing that family histories of alcoholism are associated with increased obesity risk. A polymorphism in the mu-opioid receptor gene associated with increased reward sensitivity is also associated with drug addiction. The personality trait of impulsivity is associated with both increase obesity risk and increased substance abuse risk. But some similarities in genetic predispositions don't necessarily mean that the clinical presentations of drug abuse and binge eating will present the same way in the clinic. Not only that, it doesn't mean that drug abuse and binge eating occur via the same mechanisms.
Of course there are the much vaunted PET and fMRI studies. In many of these studies, the researchers focus on the D2 receptor, a receptor for the neurotransmitter dopamine, a neurotransmitter associated heavily with reward and responses drugs of abuse.
D2 receptors, in response to drugs like cocaine, are very predictable. D2 receptor levels in the striatum of cocaine addicts are decreased. Always. It's been so repeated it's practically gospel. You see decreased D2 receptors in response to drugs of abuse in humans, monkeys, rats, mice. If drugs of abuse behave this way, and food was acting as a drug of abuse in obese patients, you might expect to see similar decreases in D2 receptors.
And sometimes, yes you do. But sometimes you don't. The results have been highly variable, depending on the groups used. Results are similar for things like activity in various brain regions as assessed by fMRI: that is, highly inconsistent. Many people see changes, but there's no real pattern to them. There may be a distinct neurobiological underpinning to binge eating or obesity, but it doesn't appear to really match that of drug abuse.
And finally, there's the issue of the highly complex biological and environmental interactions that contribute to the development of obesity. The authors argue that the complexity of factors contributing to obesity means that it may be hard to isolate the causes behind food addiction. I'm not sure about this point, myself, as the factors that contribute to obesity (genetics, environment, social factors, access, etc) all apply pretty equally to drug addiction. But I think the point about a lack of consistency in imaging studies is a strong one to make. There is great consistence in the drug abuse literature as to what changes you see in addicts, but the consistency has been much harder to find in obese individuals. I wonder what would happen if you limited the studies only to binge eaters, regardless of body weight.
Animal models currently show the strongest evidence for food addiction. You can you specific schedules and highly palatable foods to induce an addiction-like phenotype in rats, where they will escalate food intake and exhibit binge-like eating behaviors. High sugar diets in this paradigm can even produce a kind of withdrawal state, with increases in headshakes and forepaw tremor (though this hasn't been seen in humans). The D2 receptor levels in these animals fall in a manner similar to that seen in cocaine.
But there are limitations here, too. While the rats given highly palatable food may seek it out, press levers for it, and binge...they don't become obese. They tend to decrease intake of their normal food instead (which may be an important difference between rodents and humans). Not only that, the paradigms required to get rats to self-administer food like this are highly specific, much much more so than you would see in any human. So while I think that animal models could make great strides in understanding human obesity, we have to be careful in making claims.
Where do we go from here?
So, given all these caveats...is there a place for food addiction? The authors think so, and I am inclined to agree. However, it needs to be much more stringent than the current model of food addiction that many people want to embrace (the idea that sugar makes you addicted or that being overweight means you have a problem). Changes need to be made.
First off, it's important to separate food addiction from obesity. Binge eating does not necessarily mean you are overweight, and being overweight does not necessarily mean that you binge eat. Ranking by BMI is not going to work.
Secondly, it's important to differentiate food addiction from substance abuse criteria, as so many of the clinical criteria for substance abuse don't really translate to food addiction. The authors call for a set of measurable behaviors which can be clearly defined, and which demonstrate a change from normal food intake to compulsive eating. I would make the criteria here similar to those of binge eating disorder (or even merge the two).
Finally, things will have to change in how we study "food addiction" in humans. Getting a bunch of people with BMIs over 40 and popping them in a scanner isn't going to cut it. In order to get a picture of what food addiction really is, we need to use the measures that we could come up with that apply specifically to food addiction, and study only those who meet those measures. Again, just categorizing by BMI does not mean you get the correct subset of people.
All this doesn't mean we should throw the food addiction model out with the bathwater. The animal literature and binge eating in humans show indications that the food addiction model may be a good one. What this perspective suggests instead is that we need to take a more careful approach in adopting the model, to make sure that we are, in fact, modeling the right things. By modeling the right things and coming up with a useful definition of food addiction, we may be able to create a consistent picture of what changes are taking place.
Ziauddeen, H., Farooqi, I., & Fletcher, P. (2012). Obesity and the brain: how convincing is the addiction model? Nature Reviews Neuroscience DOI: 10.1038/nrn3212