Monday, June 22, 2015

Apple #713: Nutrition Data Labels

I have developed a fondness for a rather terrible junk food.  I've always liked Fritos, especially when inserted into a grilled cheese sandwich.  But now they've got these newfangled souped-up Fritos -- the Honey BBQ Flavor Twists.  The name itself should tell you this is processed pseudo pfood from the pit of badness, but apparently, I don't care.

The twistiness makes the Fritos extra-crunchy and therefore very satisfying, and the honey barbecue flavor, which is probably an affront to all genuine barbecues everywhere, is such a diabolical combination of sweet and spicy that I keep going back for more.

I've been trying to cut back on the sugar these days, and I assumed these junky temptresses would have tons of sugar in them.  I mean, they've got "Honey" in their name.  But when I checked the nutrition label, it says only 1 gram of sugar per serving, where one serving equals 23 chips.

Nutrition Facts for Fritos Honey BBQ Favor Twists.  Recipe for for my downfall, apparently.
(Image from Frito-Lay)

23 chips is a fairly good amount.  I can usually hold myself to about that many per junk-out session with the Fritos.  1 gram of sugar is not that much.  In fact, it's surprisingly little.  Especially when the ingredients include
  • Corn -- among the sweetest of the vegetables. It is, after all, the source of corn syrup.
  • Sugar
  • Brown sugar
  • Honey solids
  • Molasses solids

Either all these sweetening agents are present in very small amounts, or somebody's not doing their math right.

So then of course I had to wonder, who puts together this nutrition data, anyway?  Is it the food manufacturers themselves who say how much of x, y, & z is in their stuff, or is somebody else looking at the food they make and tallying up the goods & the bads?  If it's the food manufacturers themselves, what's to prevent them from lying through their high fructose corn syrup-coated teeth?

Never fear, the Apple Lady has got the skinny for you.

  • It's the FDA that is in charge of the Nutrition Data label.  But when I say "in charge of," I mean they are the ones who say what the label looks like, and what information must be included on it -- you have to say how much trans fats are in your food, how much sodium, how much protein, etc., etc.
  • By the way, they've proposed changing the way the Nutrition Data label looks.  An example of the current label is shown below on the left, and an example of the proposed new label is on the right. 

(sample Nutrition Facts labels, current and forthcoming, from the FDA)

  • They want to make the information that's most important to people bigger and easier to read -- calories in big print, as is the serving size, and the % daily values are closer to the thing they're measuring.  I think it's an improvement. 
  • Speaking of serving size, the FDA are also going to change the way servings sizes are calculated, to reflect what's more realistic.  Instead of making serving sizes say how much of a thing you should eat or drink, they're going to say how much of a thing you actually will eat or drink.  So for a 20-ounce Coke, for example, instead of saying that's a serving and a half or whatever amount they now use, soon they'll say it is one serving, since most people drink the entire 20 ounces as a serving.  Which will mean the amount of sugar per serving displayed on the label of a 20-ounce Coke will most definitely go up.

The exterior of one of the FDA's buildings at its campus headquarters in Silver Spring, MD
(Photo from the FDA's Flickr page)

One of the labs at the FDA. This particular lab is used for drug evaluation, but this gives you an idea of the level of technology the FDA is using on a regular basis to carry out its multitudinous operations.
(Photo from the FDA's Flickr page)

  • The FDA also dictates the way percentages of daily values are calculated (you would expect this to be obvious and simple, but it's not), and they stipulate how the samples of a particular food are to be determined, how many must be sampled, what math to use to get rid of outliers in the sample, and so on.
  • I want to pause on this concept of sampling for a moment.  Let's say your product is dried plums.  You grow your plums in several different orchards in different locations around the country.  The plums grown in California during a drought will be slightly different -- maybe smaller with a darker & thicker skin -- than the plums grown in Michigan during an especially rainy season -- maybe larger and with more moisture -- and those will again differ slightly from the plums grown in Idaho during an especially sunny time -- maybe these will have a lighter-colored skin but darker flesh.  
  • Because of these variations, the nutritional content of the plums in each of these regions will also vary slightly.  Maybe the droughty California plums will have a higher fiber content, while the rainy Michigan plums have more sugar, while the sunny Idaho plums have more Vitamin C.
  • The FDA therefore has rules about how to account for such variations in crops.  Either you use a database that's already been complied for your product (in this case, plums) that has been let's say "normalized" to take into account all the variations -- growing season, have the crops been transported and how far and for how long, in what soil were they grown, how were they processed, and so on.  Or if the FDA doesn't have a database already, you can develop your own, but the FDA has very strict rules about how you do that.
  • So, ultimately, the FDA is in charge of what the label looks like, how the samples are compiled, and what math is used to arrive at the numbers that go on the label.
  • But the real meat & potatoes of the business -- how the food itself is tested and analyzed -- the FDA is not in charge of that.
  • The testing & analysis of the food is the responsibility of the food manufacturer.  In the case of our Honey BBQ Twisty Fritos, that would be Frito-Lay.
  • However, most food manufacturers do not test & analyze their own foods.  They don't have the expertise or capabilities to do that.  Most food manufacturers hire a laboratory to do the testing & analysis for them.
  • Who are the food testing laboratories?  They are legion, across the country and around the world.
  • Think of all the things that foods can be tested for.  Here is a short list:
    • Nutritional data -- fats, vitamins, minerals, calories, etc.
    • Presence of toxins -- bacteria, salmonella, fungal toxins, etc.
    • Organically grown or not, and the effect on nutrition
    • Influence of pesticides or herbicides
    • Presence of GMOs or influence of GMOs on nutrition
  • That gives you one clue as to how many food testing labs there are out there.  

An FDA inspector and a store employee conducting a pretty simple on-site test to determine the pH of rice.  With a pH of < 4.6, the rice can be left at room temperature while the sushi, right, is prepared.
(Photo from the FDA's Food Safety Flickr page)

  • Maybe a better indicator of the number of food testing labs is the range of methods that can be used to analyze food. Those, too, are now legion.
    • One of my favorite methods is by incineration.  That's right, they burn the food.  Then they analyze the ash. The ash can tell them lots of things, such as levels of carbon or potassium, or the alkalinity of the food.  Two notable observations: if they're analyzing a sugar or a gelatin and there's an unusually high amount of ash after burning, it means the sugar is of low quality.  Or when analyzing other types of food, a high amount of ash can indicate "the presence of an adulterant," like dirt or sand.
    • Most methods these days are extremely advanced, and the names for those methods use very big words.  I've listed some of these techniques along with a super-simplified description of how the technique is used to identify a food's chemical components.
      • mass spectrometry -- identifies chemicals by their mass
      • nuclear magnetic resonance spectroscopy -- by the magnetic fields surrounding the nuclei
      • polymerase chain reaction -- by their DNA replicated on a large scale
      • gas chromatography -- by the gases present when the material is vaporized
      • high-performance liquid chromatography -- by the way each component interacts with another liquid
      • capillary electrophoresis -- by the way the ions react after voltage is applied
      • supercritical fluid extraction -- by the way the compounds behave in the presence of a non-organic solvent super-heated and super-pressurized so it is between a liquid and a gas

This is what a mass spectrometry machine looks like, circa 2005.
(Photo by Nayu Kim on Flickr)

  • You can see how a food manufacturer would not possess the expertise or the equipment to run such tests.
  • The methods the labs use to run their tests are stipulated, for the most part, by the industry associations to which they belong.  There are several such associations -- the American Association of Cereal Chemists, the American Oil Chemists Society, the Institute of Food Technologists, to name a few.

  • The AOAC is a global association of academics, government agency members, laboratories, instrument manufacturers, and other providers of chemical technology services and equipment.  Since 1884, these members have contributed their knowledge and experience toward the development of standards of chemical analysis -- not just for testing food for nutritional content, but they also test supplements & vitamins, infant formula, livestock feed, fertilizer, soil & water, and human & animal pharmaceuticals.
  • They publish what seems to be the industry bible that describes the best way to run all sorts of tests, whichever method you're using, or for whatever component(s) you want to identify.  Their guide, Official Methods of Analysis, is the must-have reference tool for the industry (though the current edition, the 19th, is sold out at the moment).
  • It's all voluntary, people's participation in this association and their contribution to the development of better & better standards.  But if they screwed up their own standards, they'd only be making things harder for themselves.  Having standards is also a way they can tell which companies are doing a good job and which companies are doing shoddy work and so might be giving their fellow chemists a bad name.
  • So it's in an industry association's best interests to come up with the best practices that they can and to keep everybody as informed about those best methods.  This is why the FDA seems to feel pretty confident in relying on the chemical & food research community to develop and maintain their own methods for analyzing foods.
  • So, to recap, a food manufacturer hires one of these food analysis labs that are most likely a member of the AOAC or similar organization.  The lab runs whatever super-duper tests they have on the food -- in our case, the FritoLay Honey BBQ Flavor Twists--and then they report the results of their multiple & complex tests to the food manufacturer.
  • It is up to the food manufacturer to make sure that the information on the Nutrition Data label is accurate and is presented in line with the FDA's regulations (the math is done correctly, etc.).
  • The manufacturer sends the information about the label to their packager, the packager prints the labels on the packaging & ships out the product, the grocery store puts it on the shelf, we pick it up, read the label, say, "Only one gram of sugar? Heck, that's hardly anything," and buy the stupid bag of Fritos.

The Frito-Lay Honey BBQ Flavor Twist packaging being produced.  The material used is a pre-printed plastic film which extends product shelf life. Frito-Lay's Fayetteville, TN plant is working to find more ways to recycle this material.  Their scraps from making potato chips are sold to a dog food manufacturer.
(Photo from Maintenance Technology)

  • But the FDA isn't entirely and totally trusting of this process.  They will randomly spot-check a manufacturer's products.  They'll run their own tests and compare the results against the Nutrition Data label. (The FDA follows the AOAC's Methods in conducting their tests, by the way.)
  • The FDA is in charge of most of the foods we eat.  The USDA polices all things meat & poultry -- slaughtered, processed, packaged meats, chicken, and eggs -- and the FDA covers everything else.  It's the USDA's policy that if the stuff that's actually in the package is within 20% of what the label says, the food is in compliance.  I can't find anything that says so exactly, but I'm assuming that the FDA has a similar tolerance allowance in the neighborhood of 20%.
  • If what's in the package is more than 20% different than what the label says, then you're out of compliance and . . . I don't know what happens next.  All the out-of-compliance stuff I found was for really serious things, like foodborne illnesses (salmonella, listeria, etc.) which can be really dangerous and which were found to be present because some element of the company's operation was out of compliance with the FDA's requirements for how you're supposed to set up your shop.  
  • I couldn't find any discussion of occasions when what was in a package of Fritos differed in some way greater than 20% of what the label said, and what the FDA did about it.
  • I suppose there's probably a hierarchy of penalties -- fines, requirements to change your labels and maybe your procedures, and then if the problem persists, maybe you lose your approval to sell your foodstuffs.  But I'm just guessing at this part of it.

FDA food safety inspectors examining cans of infant formula to see if they're within their expiration date.
(Photo from the FDA's Food Safety Flickr page)

  • So, what does this all mean about my Fritos nutrition data label?  I think it means that, unless I'm willing to go out and get my own mass spectrometer and run my own spectrometry tests, I should probably accept that the industry professionals know what they're doing -- within a 20% range of accuracy, and for a normalized sample of Fritos.  
  • It is possible I could have gotten a rogue bag of Fritos that has more than 20%  of what the label says of sugar in it, which would amount to > 1.2 grams of sugar per serving as opposed to 1 gram per serving.  But the odds of that happening are not high.  
  • And I'd probably eat the crunchy twisty Fritos anyway.

Mmm, look how crunchy.
(Screenshot from this video of someone eating a bag of the Honey BBQ Flavor Twists)

USDA, Frequently Asked Questions, What are the regulations for creating food product and nutrition labels?
US FDA, Guidance for Industry: Nutritional Labeling Manual - A Guide for Developing and Using Data Bases
US FDA, Proposed Changes to the Nutrition Facts Label
US FDA, Compliance Manuals
Leo M. L. Nollet, Handbook of Food Analysis: Physical characterization and nutrient analysis
AOAC International 
Alejandro Cifuentes, Food Analysis: Present, Future, and Foodomics, ISRN Analytical Chemistry, Volume 2012, Doc ID 801607
The National Restaurant Association, The rise of nutrition analysis
Chem Guide UK, The Mass Spectrometer