New AAP Guidelines on Childhood Obesity

Several weeks ago, the American Academy of Pediatrics released new guidelines for pediatricians on childhood obesity. These replace guidelines from 2007 that emphasized nutrition counseling and possible referrals outside of pediatricians’ offices. The new guidelines differ, in short, in arguing for a more aggressive approach to obesity in children and adolescents, up to and including medication and surgery. The guidelines still state that the most effective treatment is lifestyle and behavior changes, and that these new suggested measures are another tool in the toolbox.

Issuing guidelines like this is a huge lift, and I do not envy the amount of work that went into them by the committee. Especially since, to be frank, the reception has been decidedly cool, bordering on frigid. Criticism of the guidelines has come from all sides. Within the past week, I read fairly scathing criticism from both Vinay Prasad and Virginia Sole-Smith. To be clear, those two disagree vehemently with each other; what they can agree on is that the guidelines are problematic.

I have been considering why the reaction has been so negative. In doing so, I thought about the comparison to other guidelines. To be sure, the AAP has released other guidelines that commentators found problematic (here I am on its breastfeeding guidelines last year). But the organization also releases a lot of documents that are useful but not noteworthy.

For example: there are AAP guidelines on vision screening, which encourage doctors to be vigilant about screening for vision issues and to refer patients to specialists if necessary. Vision problems in children are treated in a variety of ways, ranging from eye drops to glasses to contact lenses. In a sense, these guidelines share a structure with the obesity guidelines: they recommend screening and treatment. The reaction is, of course, totally different.

The clearest reason for this difference is the way we, as a society, think about weight. Implicitly or explicitly, society has associated overweight and obesity with value. I am no expert on this; I’d point you to the work of Sole-Smith, Aubrey Gordon, and this excellent piece. These issues are often especially fraught in interactions with doctors. (Even the word obesity is uncomfortable for many people; I will use it here for precision, since it is the one used in the pediatric guidelines.)

So, stigma is one issue. However, in reflecting on the pediatric guidelines, I believe they would run into criticism even if we completely turned off these value judgements. Because from a data standpoint, there are at least two central conclusions on which these guidelines are based, and on which people disagree. Today, I want to unpack those data pieces.

Issue 1: Is childhood obesity a health concern?

Consider the vision guidelines and ask a parallel question — Is childhood myopia a health concern? — and the answer would clearly be yes. Not being able to see impacts the ability to do activities (see the blackboard at school, compete in many sports, etc.). It’s a concern for quality of life.

When we turn to weight, the AAP guidelines make clear that they are starting from the point that childhood obesity is a health concern. But not everyone agrees with this.

The disagreement is closely related to disagreements about the relationship between weight and health in adults. On one hand, it is clearly visible in the data that increased weight (holding height constant) is, on average, associated with worse health outcomes. People who have obesity are more likely to have other health issues — Type 2 diabetes, heart disease, stroke, cancer, sleep apnea, and others. In at least some of these cases, there is a lot of reason to think that these are mechanistically linked; that is, that obesity has an underlying relationship to the disease.

On the other hand, the link between weight, or body mass index, and health is by no means simple. BMI itself is just a ratio, and these “overweight” and “obese” cutoffs in medical definitions are arbitrary. And while it may be the case that, on average, people have more health issues at a BMI of 30 than a BMI of 24, this is definitely not true in all cases. Other health behaviors really, really matter. Someone with a BMI of 29 who jogs three miles five days a week is likely to have much better cardiovascular fitness than someone with a BMI of 22 who never exercises.

Attempting to summarize a full picture of health with a single number is misleading and counterproductive, on top of being judgmental.

The issues are even more complicated with children. There is a link between childhood obesity and adult obesity, but the predictive power is not 100%. About half of children with obesity go on to have obesity in adolescence, and about 70% of adolescents with obesity go on to have obesity at age 30. This could argue for lowering the focus on childhood and waiting to address health issues in adulthood. However, habit formation in childhood around food and exercise does carry over to adulthood, so it may be easier to intervene on these concerns with children. In addition, there is an argument that treating obesity directly in kids would make them less likely to develop conditions like Type 2 diabetes later. Better to treat obesity now than wait for other health problems later.

If you are in the camp of thinking childhood obesity is not a problem to be solved, then the entire premise of the AAP guidelines is flawed. As it turns out, though, even if we agree that there is a health concern, there is a second fundamental disagreement.

Issue 2: Is there anything effective to do about this problem?

Let’s return to childhood vision. There are a lot of effective interventions to treat myopia — glasses, contacts. Recommending that people do these things is uncontroversial. That is not true of obesity.

The AAP makes a number of recommendations. The first is that all children over the age of 2 with obesity should be referred to what is called intensive health behavior and lifestyle treatment (IHBLT). This involves family engagement with a high-intensity treatment program designed to change nutrition and physical activity. The gold-standard programs here deliver 26 or more hours of face-to-face contact over a 3- to 12-month period.

There are an enormous number of reasons why this recommendation is practically problematic. These programs are expensive and not widely available. They are typically not covered by insurance. They take huge amounts of time that families do not have. Attrition rates are high, perhaps 60% in many populations.

In addition, the impacts — even in experimental trials, which are likely to way overstate the population impacts — are small. This JAMA meta-analysis combines the randomized trials of this treatment. In commenting on the impact of trials with 26 to 51 contact hours, the authors say: “However, across all levels of contact, children in both groups showed a wide range of effects, as demonstrated by large SDs [standard deviations] relative to the mean change: some children in both groups showed fairly large reductions in excess weight, some showed no or modest changes, and some continued to gain excess weight.”

Basically, the impacts of these trials are really imprecise. Some show reductions, some do not. The effects are also small in magnitude, with the treatment groups mostly maintaining weight and the control group gaining 5 to 17 pounds on average. In the four trials that followed kids beyond 12 months, two of them showed continued impacts and two did not.

This isn’t to say that there isn’t some possible impact of these interventions, especially as you get into a range over 52 contact hours. But they aren’t some kind of magic panacea. And, again, they are pretty impractical as a public policy.

These results reflect a broad reality: weight loss through behavior change is extremely difficult to implement, especially long-term. There have been lots of kinds of interventions done with both children and adults, and behaviors are intractable. Diets usually do not have long-term impacts for the vast majority of people. Policies like “more gym time at school” also do not matter for weight.

Perhaps reflecting this reality, the AAP recommendations also suggest that pediatricians consider treatment with medication for children 12 and over, and bariatric surgery for children 13 and over with severe obesity.

The discussion of surgery in the document, while somewhat jarring (permanent, major surgery in kids who are 14 or 15 is hard to contemplate), is in a way fairly measured. They are clear that this is only one possible discussion in very extreme cases, with significant health-related complications. I do not think that this recommendation is likely to change practice in any meaningful way.

On the other hand, the medication discussion feels more meaningful — more like it might change how obesity is treated.

The guidelines discuss a large number of medications. For the most part, those don’t really work, or the side effects are prohibitive. For example, metformin (for which there is the most evidence) seems to have only marginal impacts on weight, along with substantial gastrointestinal side effects. Orlistat causes so much gas and “fecal urgency” that it isn’t widely used. Something called topiramate in principle suppresses appetite, but it causes cognitive slowing to the point of interfering with daily living, and the only trial of its use in children showed no impacts on weight.

I cannot imagine anyone prescribing any of these. There would be no reason!

However, there are new developments in weight-loss medication in the form of glucagon-like peptide-1 receptor agonists, like semaglutide, which suppress appetite. These are delivered through weekly injections and, in adults, seem to have significant impacts on weight (perhaps a 15% reduction). The idea is that these drugs would be taken continuously long-term to keep appetite suppressed.

Studies of these in kids are limited. A couple of tiny (like 12 kids) trials have shown BMI reductions. But these also have adverse effects (nausea, vomiting, possible increased risk of thyroid cancer).

Here is where we find ourselves. There is a broadly impractical recommendation for high-dose counseling interventions that, at best, are likely to have small impacts. There is at least one more extensively studied medication (metformin) that doesn’t work. And there’s an exciting new treatment with no real data in children, with no long-term follow-up in adults as an obesity treatment, that requires weekly injections, possibly forever, and has serious side effects.

Where to go from here

As I said, I’m glad I do not have to write these guidelines. I do not see some obvious solution. There are, however, two things that occur to me.

First: we need more evidence on what might work. There have been many efforts made — to improve school food, to increase access to fruits and vegetables, to change how food is marketed to kids, to create places for play, to educate about nutrition, and so on. All of these need more research dollars, along with a deeper dive on how we can improve nutrition in the household. Focusing on weight as a primary outcome misses the point. We’ve spent a lot of time focused on weight and relatively less time on trying to understand how households eat and what might nudge them to healthier eating.

The second is that there is not even close to enough discussion of the mental health impacts of these interventions on kids. Trials that look at weight loss tend not to consider either short- or long-term impacts on eating disorders in kids. This discussion is completely missing the possibility that by trying to address one issue, you may actually create another, potentially worse, one.

I listened to an interview about these guidelines in which someone suggested that, basically, we need to intervene extensively with kids because kids are bullied for their weight. That strikes me as an absolutely terrible argument. Rather than trying to address the problem of bullying, or help kids be resilient, we’re going to tell them that if they were just thinner, that wouldn’t happen? Or, effectively, bully them through the medical system? This is hardly a solution, and, plausibly, it would make things worse.

This seems to me the fundamental conflict with the entire AAP document. It argues that there is a problem, but the solutions presented are ultimately not compelling. It is a very hard problem; to go back to the vision comparison, we are about as far from “slap a pair of glasses on your face and you will be able to see” as you can get. In the end, hard problems will require work to solve — work that we should do but haven’t done enough of yet.