Another study links being ‘overweight’ with lowest risk of death

When it comes to official advice about body weight, the norm is still to base recommendations on the body mass index (weight in kg divided by the square of height in metres). We are traditionally encouraged to conform to a ‘normal’ or ‘healthy’ BMI of 18.5-24.9. It is amazing to me just how rarely (if at all) this ‘desirable’ BMI band is justified in any meaningful way. Any justification is usually rich in rhetoric but low in scientific facts.

And perhaps some of the reason for that is that the science shows the traditional BMI banding is, well, wide of the mark. One could argue that perhaps the best judge of the effect of any lifestyle on health is its impact on overall risk of death. And when BMI values have been judged by this most fundamental of outcomes, time and again it turns out that the BMI category associated with the lowest is the ‘overweight’ one (BMI 25-29.9).
See here and here for relatively recent blog posts which explore this finding.

I don’t think anyone knows for sure why being bigger than is traditionally advised appears to be better for survival. One theory that has been put forward, however, is that some surplus fat can come in handy during a critical illness, in that it offers an energy reservoir the body can draw on in times of need. Critical illnesses are more likely to occur in elderly, which raises the suggestion that being ‘overweight’ would be particularly beneficial to those of advanced years.

This concept was studied recently in a study published in the Journal of the American Geriatrics Society [1]. In this study, more than 9000 Australian women aged 70-75 (at the start of the study) were monitored over a 10-year period. It turns out that individuals in the ‘overweight’ category, compared to those of ‘normal’ weight, were at a 13 per cent reduced risk of death. These findings are reminiscent of another quite-recent study in which higher body fat levels were found to be associated with advantages for health and survival in the elderly.

This research adds to an ever-lengthening series of studies which suggest that, on a population basis, ‘healthy’ BMI values are actually too low. And this might be worth bearing in mind should you find yourself labelled as ‘overweight’ and then urged to lose weight.


1 Flicker L, et al. Body Mass Index and Survival in Men and Women Aged 70 to 75. Journal of the American Geriatrics Society 2010;58(2): 234-241

, , , , , , , ,

18 Responses to Another study links being ‘overweight’ with lowest risk of death

  1. Methuselah - Pay Now Live Later 4 February 2010 at 2:53 pm #

    Based on the energy reservoir theory, could the optimum body composition in fact be a muscular one rather than a fatty one? The notion of striving to build or preserve muscle is alien to most elderly people so there’s unlikely to be enough data to tell us… but given the other benefits of muscle retention (mobility, the health benefits of the kind of exercise required) I have a hunch it would be a worthwhile goal for an older person to build and retain as much muscle as possble in later life.

    Of course ironically, a muscular elderly person would be lumped into the same bucket as an obese one by the BMI brigade anyway.

  2. Markus 4 February 2010 at 3:08 pm #

    1) If I recall correctly, Dr. Robert Baron says that the increased mortality at lower body weight is completely explained by smoking and smoking alone (thinner, diseases that make you thin, higher mortality). Source:

    Who is right?

    2) “One theory that has been put forward, however, is that some surplus fat can come in handy during a critical illness, in that it offers an energy reservoir the body can draw on in times of need.”

    Doesn’t this explanation only apply to illness that leads to a loss of appetite or to digestive problems? Otherwise one could get the same “energy reservoir” from food, right? I also wonder if the fairly large fat reserves we have at “normal” weight are not sufficient for most phases of critical illness.

  3. Chris 4 February 2010 at 3:21 pm #

    I always find these studies interesting. Is there an element of cuasation / correlation at play here? What I mean is that in the elderly, drastic weight loss is often a symptom and consequence of serious illness. Can we say that the underweight are more likely to get ill or that getting ill makes you underweight?

    Good post

  4. Jamie 5 February 2010 at 2:55 am #


    Whilst I fully agree that BMI is a complete waste of time for the most part (I’m 27 kg/m2 & ~10% body fat), I think we need to be careful how we pitch this concept at a population level. I think the evidence is likely correct, that in the elderly, cvarrying a bit more body fat does increase survival somewhat. This could be as a result of having a good supply of fuel at a time when (due to things such as poor dentition) nutritional intake isn’t always the best, perhaps due to improved insulation from the cold which the elderly tend to feel, immune modulating factors of fat, or for having a reservoir of things such as fat soluble hormones, vitamin D, etc. These are all things that perhaps need to be explored.

    This should be treated as a separate area from the high body fat levels seen in people 50-60 years younger. Certainly for this group, a group that often looks for any reason not to change their habits, statements such as ‘healthy BMI levels are actually too low’ may be just the thing these people are looking for when clearly, that doesn’t apply to them.

    Perhaps body fat levels should be more of a U-shaped curve – better to be higher for the first & last stages of life, but should fall in between. Instead most tend to have an increasingly upward slope.


  5. Methuselah - Pay Now Live Later 5 February 2010 at 8:42 am #

    Good point Chris.

  6. Adel 5 February 2010 at 6:09 pm #

    I wonder if in lieu of the BMI categories we have now (i.e., 30), had the authors used different cut-offs what they would have seen. Maybe a BMI of 20-28 is really helpful/protective, but below or above that range is detrimental. My point is that the categories themselves are artificial, if the concept/relevance of a BMI itself is not. These are tough questions in epidemiology, particularly since we often use BMI as a categorical as opposed to continuous variable in analyses, and we might be controlling for it incorrectly.

  7. Adel 5 February 2010 at 6:11 pm #

    Sorry – the BMI categories being less than 18, 18-25, 25-30, 30-35, greater than 35 – don’t know why that didn’t work. (HTML glitch, I suspect.)

  8. Hilda Glickman 5 February 2010 at 6:11 pm #

    Weight is not the same for all people. It depends on how you get there. If you eat lots of junk and do no exercise that is different from someone who eats all the right food and is still overweight. Genes have a lot to do with it. If you study adadption in animals you will see that animals living in cold climateshave lots of fat wheras those in warm climates do not. In Israel there are two interesting types-semitic from the Middle East and Eastern european from the cold regions there. Those are almost spherical in cases while the semitic are usually lean. This goes along with little body hair and dark skin.We know that pale skinned people make vit D more easily from the sun (Less sun in these regions) so why not more fat?

  9. Dr Willip 5 February 2010 at 8:09 pm #

    I agree that the upper limit for the BMI should be increased to somewhere near 27. This would place athletes within the normal, healthy range. And, if you look carefully at the chart comparing BMI to disease, the inflection point seems to be close to 27, in my estimation. Having a little extra fat in your muscles and liver is not crippling, and may come in handy in times of illness. Health at a BMI of 27 also depends on a number of independent factors.

  10. Peter Deadman 6 February 2010 at 1:07 am #

    It may be that normal BMI measures are on the low side – particularly because they don’t seem to allow for developed muscle mass, nor for unique individual variation, but evidence of higher mortality at lower weights is meaningless unless the study finds some way of measuring only the healthy underweight. At ages 70-75 there will be a relatively high number of subjects suffering from the early or late stages of debilitating diseases such as cancer. Epidemiological studies, e.g. carried out on the Okinawan population, have found lifetime low caloric intake and low BMI correspond to increased mortality.

  11. Peter Silverman 6 February 2010 at 4:33 pm #

    As more apple-shaped men die off, leaving lots of pear shaped women, the average person has a higher BMI. But it’s not good news for the apple shaped men.

  12. Hilda Glickman 6 February 2010 at 11:56 pm #

    Maybe it is just the case that weight per se is not a factor in health at all but level of antioxidants, other nutrients, etc. The obsession with weight might just be spurious.

  13. Chris 7 February 2010 at 9:42 pm #

    I’d agree with Hilda’s nutritionist perspective that direction to include more of the ‘good’ things such as antioxidant rich food sources is generally sound advice for people on nutritionally depleted 21st century diets. (Link to ‘Origins’ paper, lead author Loren Cordain) However, an article from the printed edition of the Daily Mail (p7, Jan 28) caught my attention. A study on lab animals has explored the possibility that an excess of antioxidants can impair muscle function. It’s something to do with an axis of vasodilation moderated by hydrogen peroxide, apparently; though it is a bit early to rush to conclusions on the ramifications for the human species.
    The fastest man on two legs, Usain Bolt, does not, allegedly and according to this ‘sound bite’, take his nutrition too seriously. (100m hero powered by chicken nuggets and yams)
    John, you recently drew attention to a claim by anthropologist and author Peter McAllister that our paleolithic ancestors could have outpaced todays’ top athletes.
    It is all a bit confusing but it does make for interesting contemplation. Would dietary direction and intervention such as eating an extra ‘Shredded Wheat’ at breakfast, or including more of a given macro nutrient, or more of certain micro-nutrients and anti-oxidants, or perhaps even excluding certain food groups, improve Usain Bolts track performance?

  14. Chris 7 February 2010 at 9:46 pm #

    I omitted to add that I had tried to source ‘Manthroplogy’ but the book is not published in this territory, hence my library service could not oblige.

  15. Sue T 10 February 2010 at 11:47 am #

    My less ‘worthy’ observation – how nice to think that I may be able to choose to have a little more of what I fancy when I am elderly – and it will be good for me! Very encouraging, thank you.

  16. Nadia Mason - Discover Nutrition 11 February 2010 at 2:55 pm #

    Perhaps there is a case for raising the optimal BMI range for the older population. After all, adipose tissue is not just ‘padding’ – it has endocrine activity that can benefit post-menopausal women by increasing circulating levels of estogen.

    It’s therefore conceivable that menopausal women with greater amounts of fat would be less vulnerable to conditions such as osteoporosis and cardiovascular disease.

  17. peterlepaysan 27 February 2011 at 12:10 pm #

    Who established the BMI?

    I am led to understand that it was inspired in the USA by life insurance companies, in the 1920’s.

    It was done for actuarial purposes.

    It was based on statistics, based on a population comprised of white male European descended males.

    There is no science in it.

    A bunch of insurance sales people established the BMI.

    Whoop de whoo! this is real science.


  1. Another study links being 'overweight' with lowest risk of death … | Health News - 16 February 2010

    […] rest is here: Another study links being 'overweight' with lowest risk of death … Share and […]

Leave a Reply