The obesity epidemic re-visited

When we measure the prevalence of obesity, it is usually by way of a survey over a defined period. Thus the National Adult Nutrition Survey (NANS) here in Ireland was conducted over a 12 month period in 2009-2010.  This gives us a single measure in time but it tells us nothing about the dynamics of obesity. In other words, any individual selected at random from within the NANS database, might have acquired any excess weight at any time prior to being measured and even a subject within a normal weight range might have been fat at some previous time. Thus, there is a growing literature in the use of birth cohorts to gain a more accurate picture of the dynamics of the present obesity epidemic. Such studies seek to examine the separate effects of age, period and cohort (APC studies). We know that as we get older, our body fat rises and our lean body mass falls. The question is, does this happen at an equal rate independent of year of birth or period of life. Birth cohorts are groups of subjects born the same year or over a small number of years. They grow old together and they experience major period effects (war, depression, economic boom, technological innovation etc) at the same age. the period effect is the consequences of some event which effects all ages when it occurs. Thus the advent of the internet is a period, in this sense, which all ages encounter at a given time.

All studies show that as we age, there is a gradual rise in our level of body-fat. It rises gradually from late teens up to mid-fifties and then it begins to decline. So age effects obesity. Looking at data from 1976 to 2000^[1], there is a rise in the prevalence in obesity. Thus the prevalence was about 12% in 1976. Some 8 years later in 1984, it had increased to just 15%. A further 8 years saw the obesity prevalence rise to 22% and by 2000 it was 28%. So period has an effect.  The key question is whether every birth cohort experiences the same effect of age in the development of obesity. The Reither paper shows that for birth cohorts starting in 1895, there was a gradual but constant rise in the % of the population that were obese, reaching a peak in the late 1920s.  Then the rate of obesity declined until the late 1950s when it took off again. Clearly, this paper shows that each birth cohort or in lay terms, each generation, experiences the age related gain in weight in a different manner. A second paper ^[2] also looked at US birth cohorts and also found a similar effect. This second paper presented data in a novel manner such that we can look at what the authors call iso-BMI lines. This is like looking at a map and examining the contours of mountainous regions where specific heights appear on a continuous line. The closer the lines get, the steeper is the climb. Thus we can look at birth year on one axis and age on another axis and see the manner in which iso-BMI contours change. Those born in 1900, could expect to have a BMI of 21 by age 40.  However, for those born ten years later, that BMI of 21was achieved at 30 years of age. A decade later, it was achieved at 20 years of age. If we look at a BMI of 25, this was achieved by 20 in 1980, by 30 in 1950, by 40 in 1940, by 50 in 1920 and by about 65 in 1900. The paper of Komlos & Brabec also shows that the rise in obesity came in waves. There was a surge between 1900 and 1920 and then a decline with a second surge in the late 1940s. This pattern of surges differed according to both race and sex with African American women showing the highest level of obesity at all times and the highest post WW2 surge.

These data tell us quite a lot about the dynamics of the obesity epidemic. Firstly, more recent generations experience greater levels of obesity and at an earlier age the older generations. One could argue that this is a period effect. There is greater sedentary leisure time and a more abundant obesogenic environment. However, one could also look at a biological explanation and argue that uterine programming is involved and that as each generation experiences obesity, it somehow enhance the likelihood of even greater obesity in the next generation. In my recent book, I called my chapter on obesity “A tsunami of lard” citing other data from the US and Europe on the cyclical growth of obesity which began well over a century ago. In fact, data from pooled US and Canadian actuarial data involving 34 insurance companies examined health related risks of obesity for 163,000 policyholders were deemed overweight. Of the overweight men, those whose body fat was distributed around their abdomen had a higher risk of death than other overweight men. Among the severely overweight men, those with abdominal obesity had a 52 percent increased risk of death, whereas those without abdominal obesity had only a 35 percent greater risk compared to the general population^[3].

So when the likes of David Kessler writes in his book “The end of overeating” that the incidence of obesity soared from the late 1980s, he is ignoring an indisputable fact that obesity more or less tracked the industrial revolution. This is of huge importance. If Kessler chooses to ignore the early origins of obesity, then he can be comfortable blaming the advent of foods high in salt, sugar and fat^[4]. Others can comfortably blame the advent of high fructose corn syrup, fast food, sugar sweetened beverages. Well consider Lucius Columella in his great work De Re Rustica (On Agriculture) in the year of 65 AD. when he wrote: "The consequence is that ill health attends so slothful a manner of living; for the bodies of our young men are so flabby and enervated that death seems likely to make no change to them." It is a simple fact of  life that obesity is one of the drawbacks of affluence where food is abundant and where labour saving devices (and slave labour) are accessible. This is not for one iota to play down the health consequences of obesity. It is to simply of enormous importance in understanding the causes of obesity.

---

^[1] Reither EN et al (2009) Soc Sci Med November;69 (10) 1439-1448

^[2] Komlos J & Brabec M (2010) Am J Hum Biol;22, 631-638

^[3] Kahn HS & Williamson DE (1995) Int J Obes Relat Meat Disord; 18 (10) 686-691

^[4] That will be the subject of a future and highly critical blog on Kessler’s book.