Once more, for the cheap seats: body weight is NOT a lifestyle

CW: discussion and critique of an article claiming complex associations between body weight (and other factors) and chronic illness. I use the phrase “body weight” or BMI instead of this term. I’ve written here about why I don’t use that term.

When you’re a teacher, you know that some lessons are harder to learn than others. When I was a student, the subjunctive tense in Italian never quite sunk in (mi dispiace!) I teach a lot of logic, and I know through long experience where the pitfalls lie (e.g. necessary vs. sufficient conditions; a not-very-clear explanation is here, and a really super-long explanation is here). We teachers do what we can, and usually the confusion clears by exam time.

I really wish the same were true with medical researchers and body weight.

In case you missed it, I’m referring to an article published this spring in JAMA Internal Medicine called “Association of Healthy Lifestyle with Years Lived Without Major Chronic Diseases”. The researchers were looking for a correlation between some combo of what they considered to be lifestyle factors and onset of major chronic illness in a database of 116K people, followed for 15+ years. Here are the factors they used:

Four baseline lifestyle factors (smoking, body mass index, physical activity, and alcohol consumption) were each allocated a score based on risk status: optimal (2 points), intermediate (1 point), or poor (0 points) resulting in an aggregated lifestyle score ranging from 0 (worst) to 8 (best). Sixteen lifestyle profiles were constructed from combinations of these risk factors.

Okay, so smoking and alcohol intake are standardly considered to be health-related behaviors– things we do that affect our health outcomes (e.g. what diseases we get and when). Physical activity also has been well-documented to affect health outcomes; however, calling it a lifestyle oversimplifies it, for instance ignoring the many physical and economic and other barriers to activity that are beyond people’s control.

But then we get to body mass index (BMI). BMI is listed as a lifestyle factor? Huh?

Why don’t I think that BMI is a lifestyle factor? Let me turn this over to one of the several responses to the article, published just this week. Authors Kyle, Nadglowski and Stanford write the following:

Treating BMI as a lifestyle behavior obscures the complex etiologies that contribute to BMI… Perhaps more importantly, it promotes a mistaken notion that is the foundation for weight bias and stigma—that [one’s BMI] is [something] that patients choose for themselves through behaviors they elect. The resulting weight bias is well-documented to harm both health and quality of life of patients [with BMI >30].

Body mass index itself is neither a behavior nor a lifestyle, even though health behaviors and lifestyle factors can influence BMI. Many other factors are contributors. 

Yeah, what they said. There are, oh, about three zillion studies showing that body weight is largely genetic and/or heritable (55–70% in many research papers). That means it’s not a health behavior in the way that smoking, alcohol consumption, or physical activity are. Health behaviors affect body weight (just like they affect our cholesterol levels), but that doesn’t make them lifestyle factors, rather they are biological measures used for many purposes.

The original authors (Nyberg, Singh-Manoux and Kivimaki) respond, saying (I’m excerpting but it’s not out of context, I promise):

Maintenance of healthy weight [BMI <25] is indeed part of a healthy lifestyle…

No. Clearly we need to back up and start again.

In this commentary on the article, we get another good try at explaining the situation (this is edited to insert BMI as a term):

Such is the nature of implicit bias about [BMIs >30]. … in their hearts, even some very smart people remain certain that body size must be a matter of choice.

Yes, yes, yes. Very many very smart people (including both the study authors and the editors at JAMA Internal Medicine) still believe that body size is a matter of choice. But it’s not. The replies to this article all cite loads of articles in showing that body size is largely heritable, and if you want some refs, ask me in the comments, and I’ll reply with some standard ones.

So, one last time: body weight is not a lifestyle. But I found this website with 50 lifestyle choices to browse among, if you’re feeling like a change. I claim no responsibility for anything having to do with decisions made on the basis of looking, by the way.

I hope this clears things up.