Wednesday, August 24, 2011

EPIC Study Says: Vegetables, Fish, Dairy, Pasta & Rice Reduce, Softdrinks, Processed Meat, Margarine, Spirits and Potatoes Increase Waist Circumference - Really!?

Image 1: The EPIC project unites scientists
from all over Europe (EPIC, IARC. 2011)
I don't know whether it is a good idea, for an anti-epidemiologist like me (I mean, come on, epidemiology vs. controlled experiments is like astrology vs. astronomy) to post this, but a 13-year (1992-2005) meta-analysis (Romaguera. 2011) of data from 48,631 men and women from 5 countries participating in the European Prospective Investigation into Cancer and Nutrition (EPIC) study is probably as good as as epidemiology can ever get and with scientists from 12 different institutes all over Europe, there is at least some hope that the results of their investigation into the association of food groups/items consumption "on prospective annual changes in 'waist circumference for a given BMI' (WCBMI), prox for abdominal adiposity".

Unfortunately the first of the typical shortcomings of EPIC studies (pun intended) like this becomes obvious in the previous citation, already, as the quotation marks before "waist circumference" indicate that the latter "was measured either at the midway between the lowest rib and the iliac crest (the Netherlands, and Potsdam-Germany) or at the narrowest torso circumference (the other centres)." And even worse,
at follow-up examinations, participants in Norfolk (United Kingdom) and Doetinchem (the Netherlands) were measured by trained technicians using the same protocols as at baseline, whereas other centres provided self-reported data.
I guess you can imagine how "exact" these waist circumference are - if not, try and "measure" your personal waist circumference, you will be astonished how easy it is to "diet down" from 35" to 32"... it's a pitty that this kind of 'weight reduction' is not sustainable ;o)

Be that as it may, it argues in favor of the scientists that they did a validity test with 408 men and women from the Danish cohort beforehand and it turned out that "a high correlation between the self-reported and technician measured WC was found". Furthermore, the scientists were able to extrapolate association with baseline BMI and used these later in the regression analysis to make up for potential measuring "mistakes" on part of their subjects. If we thusly assume that at least this part of the data is reliable - I don't think I have to go into any detail as far as notoriously inaccurate food frequency questionnaires are concerned - the study provides some tangible insights into the influence the eating habits of the general population has on the accrual / loss of body fat.
Figure 1: Association (expressed as beta² coefficients, left axis) of certain foodstuffs and beverages (daily consumption in grams, right axis) with increases (positive beta²) and decreases (negative beta²) in waist circumference in 48,631 subjects from the EPIC-DiOGenes project (data adapted from Romaguera. 2011)
If you have a closer look at the data in figure 1, it's quite easy to recognize the major offenders, if you compare the association strengths (beta²) of a given food with the total amount of this foodstuff/beverage in g/day the men and women in the EPIC-DiOGenes project consumed. High beta² values indicate a positive association, which means that study participants who report (!) that they eat much of this food / drink much of this type of beverage, were more likely to increase their waist circumference.

Image 2: If you have any interest in milk, colostrum or dairy in general, make sure you read my recent write-up on milk and milk consumption!
Now, while spirits may have an even higher association with abdominal adiposity than soft drinks, the overall contribution of Jack Daniels and Co to the obesity epidemic, which has a similar handle on Europe as on the US, is nevertheless negligible, because even men "booze" way to seldom to get fat just from booze. On the other hand, the negative association of fish (beta²=-0.05) and waist circumference is much stronger than that of milk, but since fish consumption is equally low for both men and women, chances are that, overall, milk (beta²=-0.01) may contribute just as much or even more to keep (at least some of) Europeans, who love their milk (you can read more on milk if you follow the link beyond image 2) and despise fish, from bursting out of their pants and skirts than fish does, despite having the higher and thus less beneficial beta² coefficient.

I bet, some of you are just about to freak out that both "vegetable oils", as well as pasta and rice and (worst of all) breakfast cereals are associated with reductions in waist circumference. At least in the eyes of the paleo-cultists out there, this may disqualify the study as devil's work ;-) People like Robb Wolf or Matt Lalonde, whose pointed analyses go beyond the common black-and-white cave-paintings you will find on the bazillion of paleo-blogs out there, would yet notice that the men and women in the study actually consumed relatively low amounts of these foods. Roughly 63g of rice and pasta for men and 52g for women is certainly a tolerable amount. And to be honest, I have not seen anybody die from a well filled tablespoon of vegetable oil per day, either ;-)

Almost "experimental" seems the last finding of the study I want to mention, which is the (calculated) effect the replacement of (-100kcal energy from) soft drinks, margarine, processed meat and white breads (the major offenders) with either (+100kcal from) dairy or fruits would have on the changes in waist circumference in the course of one year.
Figure 2: Model calculation of the beneficial (lower beta² coefficients) effects of iso-caloric (100kcal/day) replacements of soft drinks, margarine, processed meat and white bread by fruits or dairy (data adapted from Romaguera. 2011)
As the data in figure 2 goes to show the effect the statistical model predicts that the effects of fruit would be more profound... this, however, is something I dare question, simply because the different macronutrient composition of dairy (moderate protein, moderate fat, relatively low carb) vs. fruit (no fat, no protein, high carb) simply forbids such statistical thimblering that neglects the macronutrient composition and focuses solely on the caloric value... ahh, one last word: Whenever, this crucial distinction between eating calories and eating food (I am starting to think that maybe some of the scientists actually eat calories?) will reach mainstream science, I will immediately let you know ;-)