More evidence that the ‘overweight’ are at the lowest risk of death

The standard way of assessing weight as it relates to health is the body mass index (BMI) – calculated by dividing an individual’s weight in kilograms by the square of their height in metres. According to conventional wisdom, a BMI of 18.5 ” 24.9 is to be regarded as ‘healthy’, while BMIs of 25.0-29.9 are considered ‘overweight’. Those of 30 or above are considered ‘obese’. While the BMI usually forms the basis of the advice health professionals give to individuals about their weight, there are reasons to be mistrustful of it, I think.

My main gripe with this measurement is that while the BMI tells us something about the relationship between height and weight, it tells us nothing about body composition. It is entirely possible, therefore, to have a heavy, muscularly built, really very healthy individual whose BMI marks him or her out as ‘overweight’ or even ‘obese’. The other side of this, of course, is that someone can have a body composition that is low in muscle and high in fat, but at the same time be classified as ‘healthy’ from a BMI perspective.

The fact that the BMI tells us nothing about body composition means that, in all likelihood, it’s not going to tell us much about health status either. While it might be enshrined in medical lore that a ‘healthy’ BMI is one that ranges between 18.5 and 24.9, there is actually some evidence which calls this wisdom into question.

One way to assess the appropriateness of the BMI categories is to measure the overall risk of death in each category. If BMIs of 18.5-24.9 are deemed as being most healthy, then individuals with BMIs in this range should be at the lowest risk of dying too, right? Well, the biggest and most comprehensive study ever to look into the relationship between BMI and risk of death was published in Journal of the American Medical Association in 2007 [1].

Some notable findings from this study included:

  • Being overweight was not associated with an increased risk of death from cardiovascular disease
  • Being overweight was not associated with an increased risk of death from cancer
  • Being overweight was associated with a reduced risk of deaths not related to cardiovascular disease or cancer
  • Being overweight was associated with the lowest risk of death overall

Recently, some doctors acknowledged this evidence, but then reinforced the usual message that individuals should keep out of the overweight category. You can read about this here.

This week sees the publication of yet more evidence for the notion that ‘ideal’ and ‘healthy’ weights as determined by the current classification of BMI are wide of the mark [2].

In this study, researchers followed more than 11,000 Canadians over a 12-year period. They then calculated overall risk of death in each category (underweight, ‘healthy’ or ‘normal’, overweight, obese and extremely obese.

Compared to individuals in the ‘healthy’ category (BMI 18.8-24.9), overall risk of death for the other categories was as follows:

Underweight (BMI <18.5): 73 per cent increased risk of death

Overweight (BMI 25.0-29.9): 17 per cent reduced risk of death

Obese (BMI 30.0-34.9): No statistically significant difference in risk of death

Extremely obese (BMI 35 or more): 36 per cent increased risk of death

Here, again, we find that the lowest risk of death was found individuals classified as ‘overweight’. And this result was statistically significant. Perhaps even more surprising than this, though, is the finding that being ‘obese’ did not appear to put individuals at a significantly increased risk of death.

Risk of death is not the only way to measure health. Risk of illness and quality of life should be borne in mind too. However, some would argue that risk of death is probably the best single measure of the impact of a body measurement or lifestyle factor on health. I mean, there is an argument for believing that the most fundamental measure of health is whether someone is dead or alive.

We now have large studies which have found that being ‘overweight’ is associated with reduced risk of death. There is a case for changing the name of the ‘overweight’ category to ‘healthy’ and the ‘healthy’ category to ‘underweight’. And bearing in mind the failing of the BMI in terms of assessing body composition on a personal level, there’s an argument for ignoring it altogether. For information regarding what measurement is a better guide to health status, see here.


1. Flegal KM, et al. Cause-Specific Excess Deaths Associated With Underweight, Overweight, and Obesity. JAMA. 2007;298(17):2028-2037

2. Orpana HM, et al. BMI and Mortality: Results From a National Longitudinal Study of Canadian Adults. Obesity 2009 Jun 18. [Epub ahead of print]

12 Responses to More evidence that the ‘overweight’ are at the lowest risk of death

  1. Peter Deadman 10 July 2009 at 9:11 am #

    It’s obvious that there has to be a distinction between healthy muscular/strong/fat people and unhealthy skinny people.
    These weight/risk of death studies go back to the 19th century where they were originally led by life insurance companies. One issue that has been identified as a possible confounder of the more controversial higher weight=lower risk of death studies, is whether or not they exclude smokers and the chronically ill. Since both smoking and chronic and degenerative illness are associated with lower weight, clearly – unless these have been discounted – the findings will be flawed.
    There are also epidemiological studies, for example conducted on Okinawans, that low calorie consumption and low body weight are associated with gains in longevity.

  2. Jane Dodds 10 July 2009 at 3:59 pm #

    I am confused. We are often told on this site that waist circumference is the most important indicator in terms of health risk, including heart disease, diabetes and some cancers. Surely not everyone in the BMI overweight category in these studies also happened to have the waist measurement deemed healthy for their gender?

    My BMI is 25, yet my waist measurement is 3 inches more than is recommended for a woman. So, should I lose weight, thus reducing my waist size and my BMI and, according to these studies, increase my risk of dying…..?!

  3. Dr John Briffa 11 July 2009 at 7:55 am #


    The blog here suggests that the BMI categories are misleading when applied to a population (they give a false impression, I think of the health ‘hazards’ of being overweight. When applied to an individual, the BMI is even more useless, as it tells us nothing about body composition nor fat distribution. So when looking at an individuals (e.g. you) I think the BMI value has practically no relevance at all.

    I suggest you ignore it, and focus (if you want) on more useful body measurements such as, as you the suggest, waist circumference.

  4. Karen Harrison 14 July 2009 at 3:08 pm #

    Does anyone know where the definitions for BMI i.e. healthy, overweight etc. came from? It seems that the calculation and definition came about arbitrarily and there is no easily accessible evidence what science this was founded on.

  5. Trinkwasser 16 July 2009 at 4:00 pm #

    Changes in BMI may be more useful than the number per se.

    I’ve always been on the lower end of normal and was actually underweight while I had undiagnosed gallstones.

    However this never stopped me from having severe Impaired Glucose Tolerance, only a gnat’s whisker away from diagnosible diabetes.

    After being attacked by a Dietician my BMI rapidly increased to borderline overweight, and equally rapidly decreased back to *my* normal when I started eating a rational diet instead.

    Meanwhile one of my aunts, with a different subset of the same genes, was comfortably overweight with “the blood pressure of a 30 year old” in her eighties, while my tiny mother has had hypertension for 40 or 50 years. This seems to be a distinct familial pattern where the “overweight” measure significantly better on cardiovascular risk factors than us skinnes. Maybe this is not just a pattern in our family but is more widespread.

    Michael Eades calls it “metabolic obesity” where you measure like an obese person without the overweight. Presumably there’s also a corollary of “metabolic slimness” where your BMI is higher than your risk factors.

  6. Hilda Glickman 17 July 2009 at 8:28 am #

    Weight itself might not matter as much as the reason why someone is overweight. Some people have low thyroid and this can affect the heart and also cause weight gain. Some people are genetically prone to weight gain even if they eat a good diet. Others ar overweight because they eat junk. These are probably the ones who get ill because of what they not because of the weight. So obesity may be a symptom and not an illness. Hilda Glickman nutitionist

  7. Jamie Scott 18 July 2009 at 11:47 am #

    I think the key measure here is quality of life & health while alive. Measuring death rates as related to BMI, weight, etc, is a bit spirious in my mind. For example, having to spend large periods of time sitting on renal dialysis is quite distinct, in terms of quality of life, from someone who gets to spend the same period of time riding their bike in the sun. But both individuals could however die at the same age/stage of life.

    There are any number of factors that relate to premature death & high body fat levels are one of those. What we want however, is not necessarily a population that lives any longer, but one that is healthy for the time it is alive.

    As a clinical nutritionist & exercise physiologist, I get a bit uncomfortable when I see headlines such as “More evidence that the ‘overweight’ are at the lowest risk of death” as I know how eagerly some people will seize on them & misrepresent them from their original intention in order to justify their current poor eating & exercise habits & health status. That is fine for them & very much their individual choice. However, here in New Zealand, I don’t get the choice as to whether I want my taxes spent on propping up their health when it all goes wrong!

  8. M. Cawdery 21 July 2009 at 12:03 pm #

    Another reference by Flegal:
    JAMA. 2005 Apr 20;293(15):1861-7.
    Excess deaths associated with underweight, overweight, and obesity.
    Flegal KM, Graubard BI, Williamson DF, Gail MH.
    National Center for Health Statistics, Centers for Disease Control and Prevention, Hyattsville, Md 20782, USA.
    Overweight was not associated with excess mortality (-86,094 deaths; 95% CI, -161,223 to -10,966). That is, Overweight individuals LIVED LONGER than normal weight persons!
    Check for yourself on Pubmed. This study has been widely ignored by the medical establishment.


    Incidentally the first paper published was shown to have had arithmetic mistakes. The above paper had to be written to correct the errors. BEWARE of medical statistics!

  9. SteveMD 30 August 2009 at 6:56 pm #

    There are clearly some protective advantages in carrying excess fat, particularly in conditions that cause inflammation and cell damage. Fat and obese ICU patients tend to have better outcomes than thin or underweight patients, – a 30% better survival rate. These figures are, like those above, adjusted for age, smoking and wasting diseases. There are also increased risks associated with carrying, excess fat. On the whole I would tend to agree that excess fat, below the extremes noted above, is largely benign and the case against it vastly exaggerated.

    To be specific, the anti-fat stance is mostly, in my opinion, a matter of fashion, not medical concern.

    Exercise is far more important, with regard to health, than the percentage of body fat. Even with this we manage to scare people away from it more than get them to take part. Firstly; we must abandon the “minimum requirement” that health professionals insist is needed for improvement. We overwhelm people with images of pain and hard work before they even begin to exercise. My own view is that ANY increase in physical activity, no matter how little is a good thing. I know several, previously inactive, people who have improved their standard of life simply by buying a Wii games console. On the whole I just encourage people to do a little more each day and emphasise that I mean a LITTLE. This way they don’t even notice that they are increasing the amount they do and tend to stick at it much longer, until low and behold, they are doing a useful amount of exercise regularly, as part of their daily routine. Instil the habit first, this is the most important aspect of regular exercise, then build the amount – slowly.

  10. Aristotle 24 August 2011 at 6:35 pm #

    I think it’s very possible that these conclusions actually reverse the casuality here, and that it may not really be the fat tissue itself that provides the protective advantage, but that the fat tissue reflects a more fundamental parameter.

    Pretty much every single chronic disease known to man is caused by chronically-elevated blood sugar, or at least clusters around the condition that has come to be known as the Metabolic Syndrome. It is very likely that people that are a little overweight are in better overall health because they are more likely to have good bloodsugar regulation (i.e., putting on some fat mass is the physiologically-normal and healthy reaction to elevated blood sugar). People who are extremely obese on the other hand, have likely developed some degree of insulin resistance, and therefore have chronically-developed the same poor blood sugar regulation as someone underweight.


  1. Low Carb Age » Does Being Overweight Harm Your Health? - 10 July 2009

    […] people would live longer, but as it turns out, several studies are showing that they die sooner. Dr. Biffa’s blog recaps the results of two recent studies that show surprising results: Compared to individuals in […]

  2. The Financial Cost of Obesity | San Jose Health Conscious Consumers Community News - 28 July 2011

    […] link:… […]

Leave a Reply