by Marion Nestle
Jun 4 2014

Guess who funded the contradictory fructose study?

Today, Michael Goran and his colleagues published an NIH-funded study demonstrating that the proportion of fructose in products made with high fructose corn syrup is often higher than 55%—as much as 60% to 67%.

This matters because of concerns that high intake of fructose might induce insulin resistance and other metabolic problems.

Today also, a different group of investigators published a study saying just the opposite.  Fructose in products, it says, is in close agreement with the amount expected.

Who funded this one?  The International Society of Beverage Technologists, whose executive board represents soda companies.

Really, these kinds of results are so predictable that all I have to do is see the results to guess who must have funded the study.

Coincidence?  I don’t think so.



[…] Source: Guess who funded the contradictory fructose study? […]

[…] Source: Guess who funded the contradictory fructose study? […]

[…] Source: Guess who funded the contradictory fructose study? […]

[…] 6. Food Politics » Guess who funded the contradictory fructose study? […]

  • Henry Chin
  • June 5, 2014
  • 10:36 am

If you have a background in analytical chemistry and read the papers carefully, it is clear that the two papers are measuring two different things. Goran, et al’s analytical methods measure only mono and disaccharides. The method in the paper by Hobbs includes the measurement of polysaccharides that are larger than disaccharides. Since the polysaccharides account for about 5% of the total “sugar” in HFCS, not accounting for this faction will result in an error when the % fructose is calculated. In analytical chemistry, there is a saying that a method must be “fit for purpose”. A method might be good for measuring mono and di saccharides, but if it hasn’t been validated to be fit for purpose (in this case measuring the % fructose in HFCS), then the results really can’t be relied upon.

  • Californiaesque
  • June 5, 2014
  • 11:33 am

Henry Chin, Can you explain in laymen’s terms whether either study accurately measures what it purports to study?

  • Henry Chin
  • June 5, 2014
  • 12:57 pm

First, allow me to try to explain the basis of the analytical methods used. All of the methods used in both papers start with a technique called chromatography. Chromatography is a technique used to separate the molecules of interest from other molecules. You then “count” the separated molecules to get information on the composition of the initial mixture. This is akin to having a mixture of balls of different sizes and colors and putting the mixture through a screen. You can then count the balls that go through the screen and get an idea of the composition of the initial mixture. For example, if you had 20 balls in the initial mixture, and you counted 10 red and 10 blue through the screen, the mixture was obviously 50:50. If on the other hand, you had 20 balls and only 18 went through the screen, because the remaining two where too large, well, your results would be inaccurate unless you accounted for those two balls. In the studies in question, Goran et al measured glucose, fructose, sucrose, lactose, maltose, and galactose. In the ball analogy, glucose, fructose and galactose (monosaccharides) are roughly the same size; sucrose, maltose, and lactose are slightly bigger (being dissacharides). If there are no other balls (saccharides) in the sample, then the results are correct. So in fruit juices, where this is the case, the results are accurate. However, HFCS starts with the starch molecules in the corn, thus HFCS contains some larger polysaccharides. In the ball analogy, because the screen is set up for nothing larger than maltose, lactose, etc., these larger balls will not go through the screen and be counted. Thus, if you want to measure the sugar composition of HFCS or foods containing HFCS, you have to adjust your screen to so that you can measure these larger balls. So, the Goran results are probably accurate for fruit juice and foods not containing HFCS, but the reported ratios will be inaccurate for foods containing HFCS. Sorry for the long-winded answer, but hope that this is helpful.

Also, note that in both studies, samples were analyzed by Covance Laboratories. And, as collaboration for what I said above, they reported different results depending on the method.

  • Marko
  • June 5, 2014
  • 5:14 pm
  • Michael Bulger
  • June 5, 2014
  • 6:15 pm

Thanks for the explanation.

Excellent explanation. Thank you!

[…] 5. Guess who funded the contradictory fructose study? – Hint: they make their living from selling fructose. Color me shocked! […]

  • Kevin Klatt
  • June 12, 2014
  • 8:56 am

The polysaccharides in HFCS are made up of glucose. Wouldn’t the logical conclusion be that using methodology that includes these balls would increase the glucose % and decrease relative total fructose?

  • Henry Chin
  • June 12, 2014
  • 11:29 am

Good question. Yes and, there’s more to my answer.

are several facets (or parts) that make up the whole of an analytical
method. In my previous comments, I
addressed how you separate and count (detect in analytical terms) the things of
interest. But, before you can separate
and count, you have to prepare the sample for analysis. The method that you use to prepare the sample
is also dependent upon the question that is being asked (i.e., the purpose of
the test). Using the colored ball
analogy again-if you want to measure the number of “free” blue and red balls in
solution (in this case, glucose and fructose in solution), you try to minimize
what you do to the sample before analysis.
If you want to include the balls that might be bonded together, then you
do something that dissolves the glue that holds the balls together before you
separate and count. Breaking up the
polysaccharides will increase the amount of measured glucose, but this glucose
is not “free” (in the analytical sense) in the sample. For example, starch is a complex
polysaccharide. One way to estimate
starch content is to first remove all of the “free” glucose, then hydrolyze the
starch to glucose, measure the glucose, and calculate the starch content based
upon the measured amount of glucose.

Leave a comment