The body mass index (BMI) is the standard and most commonly-used method for assessing body weight by health professionals. It is calculated by dividing an individual’s weight in kg by the square of their height in metres. The fact that some arithmetic is required to ascertain the BMI makes it look quite scientific and robust. The fact of the matter is that it tells us nothing about body composition nor the distribution of fat in the body. Oh, and there’s another problem with the BMI: individuals in the ‘overweight’ category (BMI 25-29.9) seem to enjoy at least a good overall health if not better than those in the ‘healthy’ category. See here for more on this.
As a measurement of health status as it is currently applied the BMI is a misleading waste of time. The good news is that there are other more easily determined and more useful measurements out there. One of these is the circumference of the waist. Basically, the bigger this is, the more fat is likely to be congregated around the abdomen, and it is this sort of fat that seems to be most strongly associated with chronic diseases such as heart disease and type 2 diabetes. In short, the thought is that the bigger one’s belly, the more likely one is to get sick and die.
In a recent study published in the American Journal of Epidemiology, the relationship between waist circumference and overall risk of death was assessed in almost 250,000 men and women over a 9-year period. As part of the study, the researchers isolated individuals who from a BMI perspective were ‘healthy’ in terms of weight (BMI 18.5 ” 24.9). In this ‘healthy’ group, the overall risk of death in those with considered to have large waist circumferences (for men and women this was a waist size of 102 and 88 cm or more respectively) was compared to those with individuals of normal waist circumferences (measurements less that 102 and 88 cm in men and women respectively). The results of this study were that:
Men with the higher waist circumferences were found to be a 23 per cent increased risk of death
Women with the higher waist circumferences were found to be a 22 per cent increased risk of death
These findings, I think, serves to highlight the importance of waist size as a measure of health status. They also serve to highlight the inadequacy of the BMI as a tool or assessing risk of chronic disease and death.
For those wishing to trim their waists, my suggestion is to consider a diet lower in carbohydrates that tend to destabilise blood sugar and insulin levels. The rationale here is that ‘abdominal obesity’ is often a feature of a condition known as ‘metabolic syndrome’, in which elevated levels of insulin and/or ‘insulin resistance’ are usually features. It can be that a glut of carbohydrate in the diet can cause insulin surges which predispose to fat deposition around the abdomen. So, one obvious approach to this problem is to cut back on carbs.
I recently reported on a study in which what was essentially a low-carb ‘primal’ diet was found to be effective for weight loss and waist circumference. The study has some deficiencies, but it did mirror what I almost invariably find to work in practice for the purposes of waist disposal.
I’ve read your blog for quite a few months but not commented. I’d like to say I find your posts interesting and informative.
I’ve been following a moderately low carb diet since February this year and lost 10cm from my waist (as well as similar amounts from hips and bust) so can add some more anecdotal evidence that a low carb approach works. Just got to get rid of another 5cm and I will be in the health range according to that study!
I completely agree with this concept. We have known for quite awhile that waist circumference is positively correleated with heart disease risk.
Further, muscle weighs more than fat, so two people could be the same height and weight, with two completely difference fat/lean ratios but the same BMI. Clearly, the person with the greater lean percentage would have a smaller waist size.
I don’t get this. Why is waist size an indicator of abdomen size? My waist isn’t big, (size 12), but my abdomen IS.
Tiggy.
It is good to know that science comes round to the truth. Please find the best way to measure the waists of large people, for consistency, because the rolls of fat for large people, can be wider than at the tummy button and changes as they lose the fat.
Thank you for Dr Briffa
my only issue with this is that it does not take into account a persons height or overall body build – it is still a load of poppy-cock sorry but what is so hard about saying to be healthy we need to eat less if not zero sugar which is actually according to the studies & information I have read the number one poor health indicator. in that the higher your carb/sugar consumption the poorer your health in all respects.
I believe this is another “last ditch” attempt to shore up the “fat kills!” dogma, this time shifting the blame to visceral or “internal” body fat.
Published in the April edition of the American Journal of Clinical Nutrition, the CDC and National cancer institute (of the U.S.A.) assessed various body measurements; BMI [weight (kg)/height (m)2]; percentage of body fat measured by bioelectrical impedance; skin-fold thickness; circumference of waist, hip, and arm and waist-hip ratio; waist-height ratio for all-cause mortality. Their finding were;
● The fractions of all-cause mortality associated at low levels of all the variables were similar and small, ranging from 0.3% to 2.5%.
● For intermediate levels of each body measure compared with the reference measurements, the associations with all causes of mortality were negative (-1.8% to -9%) ” meaning, having higher than ideal numbers were associated with lower mortality ” and the association with percentage body fat was only 0.1%.
● Finally, for the highest measurements ” regardless of the anthropometric or body composition variable ” there were no tenable correlations to excess mortality ” they were all statistically weak, ranging from -1.7% to 1.5% compared to “ideal” numbers.
● Even comparing the different body measurements to each other “showed no significant differences between estimates of excess deaths for any pair of variables at any level.