Reduced All-Cause Mortality in Overweight Adults: Should we be Surprised?

Scale-A-Week Project

Today’s post comes courtesy of my friend and frequent collaborator Dr Jean-Philippe Chaput.  It is a Letter to the Editor that was written by Dr Chaput along with Angelo Tremblay and Eric Doucet in response to a recent paper in the Journal of the American Medical Association.

In the new paper, Katherine Flegal and colleagues perform a systematic review and meta analysis of the available evidence linking body mass index (BMI) with mortality.  They included 97 studies, with a combined sample of 2.88 million people.  Here is what they found (my emphasis):

Relative to normal weight, both obesity (all grades) and grades 2 and 3 obesity were associated with significantly higher all-cause mortality. Grade 1 obesity [BMI of 30-34.9] overall was not associated with higher mortality, and overweight was associated with significantly lower all-cause mortality. The use of predefined standard BMI groupings can facilitate between-study comparisons.

Not surprisingly, the paper has generated a lot of discussion, most notably from Harvard epidemiologist Walter Willett who claimed that “Stated politely, the paper is a pile of rubbish“. Drs Chaput and colleagues wrote a letter in response to this paper, which unfortunately was not accepted for publication by JAMA.  But he has graciously allowed me to post it below, and given the amount of discussion on Peter’s previous post explaining why BMI is a poor measure of your health, I think it will be of interest to our readers.   I have bolded some sections to emphasize the key arguments. Enjoy!

The results recently reported by Flegal and colleagues1 about reduced all-cause mortality in overweight individuals compared with normal weight individuals seem counterintuitive but are concordant with what would be predicted by normal human physiology. The research on adipose tissue over the last decades has indeed emphasized that the role of fat cells in biological regulation goes much beyond the storage of excess energy/lipid. A gain of 5 to 10 kg of body fat has numerous biological effects which seem to a priori promote what would appear to be a series of relatively positive outcomes rather than detrimental physiological and metabolic effects.

These effects become more apparent when overweight and class 1 obese individuals are subjected to weight-reducing programs to lead to a reduction of what could be initially perceived as excessive fat storage. As recently reviewed,2,3 fat loss promotes the following adverse effects: 1) an increase in the drive to eat, 2) a greater than predicted decrease in energy expenditure being at least partly related to a release from adipose tissue of persistent organic pollutants and to a decrease in plasma leptin levels, 3) an increase in the Beck Depression Inventory score associated to hypoglycemia and reduced plasma triiodothyronin levels, and 4) an increase in a posteriori weight gain (regain) related with reduced adipose tissue lipolysis.

In our opinion, we can draw from these observations that early body fat gains that characterize an overweight status likely exert a protective homeostatic role. In fact, increased energy storage in the form of body fat may help to manage an environment that is seemingly becoming more toxic, at least as far as energy balance is concerned. An increasing number of obesogenic factors related to our modern way of life, such as demanding mental work4 and an increase in atmospheric CO2 levels5, to name a few, are being identified and characterized.

Thus, the increase in body fat characterizing overweight people is more likely to represent a normal allostatic response than a pathological process. In this context, it is not surprising to note a reduced mortality in overweight individuals. However, as for many regulatory physiological processes, fat gain cannot entirely compensate for a chronic detrimental stimulation, which is again consistent with the Flegal et al. data indicating that mortality rates are increased in grades 2 and 3 obesity. A preventive approach that attempts to reduce the obesogenic nature of our society is probably the only long-term viable solution to improve our health, even if it does not easily fit within the priorities of an economically globalized world.

Angelo Tremblay, PhD

Éric Doucet, PhD

Jean-Philippe Chaput, PhD


  1. Flegal KM, Kit BK, Orpana H, Graubard BI. Association of all-cause mortality with overweight and obesity using standard body mass index categories: a systematic review and meta-analysis. JAMA. 2013;309(1):71-82.
  2. Chaput JP, Doucet E, Tremblay A. Obesity: a disease or a biological adaptation? An update. Obes Rev. 2012;13(8):681-691.
  3. Tremblay A, Chaput JP. Obesity: the allostatic load of weight-loss dieting. Physiol Behav. 2012;106(1):16-21.
  4. Chaput JP, Drapeau V, Poirier P, Teasdale N, Tremblay A. Glycemic instability and spontaneous energy intake: association with knowledge-based work. Psychosom Med. 2008;70(7):797-804.
  5. Hersoug LG, Sjödin A, Astrup A. A proposed potential role for increasing atmospheric CO2 as a promoter of weight gain and obesity. Nutr Diab. 2012;2:e31.

ResearchBlogging.orgFlegal, K. (2013). Association of All-Cause Mortality With Overweight and Obesity Using Standard Body Mass Index CategoriesA Systematic Review and Meta-analysisAll-Cause Mortality Using BMI Categories JAMA, 309 (1) DOI: 10.1001/jama.2012.113905

Related Posts Plugin for WordPress, Blogger...
This entry was posted in Guest Post, Obesity Research, Peer Reviewed Research and tagged . Bookmark the permalink.

14 Responses to Reduced All-Cause Mortality in Overweight Adults: Should we be Surprised?

  1. Natalie says:

    An interesting take on what overweight might mean in terms of overall health. The only sad part is the number depicted on the scale: in order to be overweight at 120 lb. (this is clearly a lb. scale, not a metric one), the person in question would have to be 4’10”, and basically a dwarf. Pictures like this give the wrong message to women and girls, the vast majority of whom are nowhere near overweight at 120 lb. Please don’t feed into eating disorders, even if inadvertently, by posting such triggering and inaccurate photos.

    • WRG says:


      I am 4’8″ and not a dwarf, just really, really short. Dwarfism is a condition that not all people of short stature have.

  2. Travis Saunders, MSc, CEP says:

    That hadn’t even crossed my mind. I just wanted a picture of a scale since we’re talking about weight. I’ve changed it for another image I found on Flickr of a metric scale with a weight that would place a person of average height in the overweight category.


  3. Natalie says:

    Thanks for your willingness to understand and immediately change the picture. The problems of anorexia and bulimia and diabulimia are certainly much rarer than obesity, but in a way, they’re the other side of the coin. And as the article so correctly points out, the major concern for all of us needs to be health, not weight per se! Kudos to you! :-)

  4. Pingback: Home altitude tied to obesity risk – Reuters | HNN

  5. Di B says:

    I have often wondered whether it is accounted for that many illnesses make people lose weight, to the point of being underweight, and these people may be un-diagnosed and living in the sampled general population. And people with more than usual muscle mass are likely to measure in the overweight/level 1 obese category. These would both tend to confound statistics, unless specifically adjusted for.
    I am doing research that looks at mechanisms, and fully appreciate that generating plausible theory through mechanisms is useful and often seductive, but thorough consideration of confounders is all part of the move towards understanding the phenomenon (or is is collection of phenomena?). I so want to believe this one, too, as a level 1 body-type myself, albeit physically active.

    • Travis says:

      Great questions. One of the criticisms of this paper is that some (e.g. Walter Willet, whose comments I linked to above) feel that the study didn’t adequately adjust for people who may have had an underlying illness reducing their body weight. It could be having a small effect on the results, but I’d point out that this study isn’t an outlier – lots of studies have shown that health risk really increases most at extremely high or extremely low BMIs, with little or no health risk seen in the overweight and first obese category, even when they exclude people with underlying illnesses.

      It’s true that muscle weighs more than fat, but that affects very few people at the population level. Aside from varsity athletes, it’s uncommon to find a person who is overweight due to their muscle, and extremely uncommon to find a person who is obese due to their muscle mass (I think that Shaq is the only example that I know of).

      There are other more pressing concerns with BMI when it comes to health, the main one being that it tells us nothing about body fat distribution, which matters much more than total weight or even total body fat content. This is especially a concern when we talk about different populations, as Asians tend to have much more abdominal fat than some other ethnic groups for a given BMI, which means that a BMI that is quite healthy for a White or Black American may not be healthy for an Asian American. We’ve outlined those in more detail here:

      • Di B says:

        Good answers, thanks for that – I was not looking at the paper, evidently, just on the commentary. In my study I have measured waist and hips (women) and wrist size as well as height and weight (it is a pilot, aimed to find out acceptability and feasibility as well as a notion of effect sizes and variability). Not sure whether these will sufficiently account for the inconsistencies in BMI, at least I am trying something… Thanks for the background reading. I have sevaral other quite detailed questionnaire measures also, so this is just one area.

        • Travis says:

          Waist circumference is a very good one to have. Between that and BMI you’re able to get a much better handle on health risk than with BMI alone. Good luck with your study!

  6. Di B says:

    Thanks – I will need it – keeping up with the data analysis etc. is an organisational and intellectual challenge – as of course it is menat to be….

  7. S says:

    Well now, that’s hardly fair. I got laughed out of grad school for saying what Drs. Tremblay, Doucet, and Chaput say here. They are basically saying that endocrine disturbances promote fat accumulation, and fat accumulation increases hunger. Adipose tissue is the largest endocrine organ in your body. It doesn’t just expand because you ate another cookie. You’ll just burn off the cookie.

    Glad to see someone in academia is taking this approach.

    • Travis says:

      I don’t think they’re saying that. When they talk about the adverse effects of weight loss, it’s important to remember that they are not necessarily the same things that caused the weight gain in the first place.

      Here’s what I take to be the key message.

      An increasing number of obesogenic factors related to our modern way of life, such as demanding mental work and an increase in atmospheric CO2 levels, to name a few, are being identified and characterized.

      They are saying that there are lots of factors that contribute to positive energy balance, and that storing excess energy in adipose tissue can be viewed as a natural response to such an “obesogenic environment,” and helps to bring the body back into neutral energy balance. JP and I go into this in more depth in another paper, available here.

      I’m sure there’s a role played by the endocrine system (it’s a pretty hot topic at the moment thanks to Gary Taubes, Bob Lustig, etc), but I tend to think that it is one of many factors that link the “obesogenic environment” with weight gain, rather than thinking it is the one and only cause.

      • S says:

        Hmm….how would homeostatic changes in response to obesogenic environmental factos *not* involve the endocrine system? Adipose tissue itself is an endocrine organ.

        I’m with Gary Taubes that energy balance is the “how” , not the “why”, adipose tissue expands. Sure, eat *too many* cookies and you mess with dozens of hormones, including insulin, and the body needs to do something to get back to balance. What it has to do to get back in balance results in weight gain.

        Likewise, weight loss only occurs *after* favorable hormonal shifts. Weight loss surgery has been show to dramatically alter several hormones *within weeks*, that is, before substantial weight has been lost.

        • Travis says:

          I don’t really disagree with you… and most of this is just a matter of what paradigm you want to use to talk about obesity. The area where I really disagree Taubes is the idea that the endocrine system (and insulin specifically) explains pretty much everything with respect to obesity. There is lots of good evidence that the endocrine system has a big role to play there, but there are too many exceptions insulin alone to be the One True Cause that Taubes claims it to be. There is a terrific discussion of this issue at Whole Health Source.

          In the end it doesn’t really matter whether I think Intervention A prevents weight gain because it reduces positive energy balance, and you think Intervention A prevents weight gain because it gets the endocrine system back on track. They’re probably both involved, and so long as Intervention A prevents weight gain, the actual mechanisms are ultimately of lesser importance.