September 29, 2009 update: repaired broken link to "Final Rule on Labeling of Trans Fatty Acids"
On a recent Southwest Airlines flight, the snack wasn't the classic tiny bag of peanuts, but a tiny bag of "Plane Crackers" (sometimes it seems that the packaging weighs as much as the food inside). When I saw a Nabisco label in the corner, I immediately took a close look at the label. The label (shown below, with some crucial things underlined) indicates 3.5 grams of total fat per serving. The fats are also broken into categories, with individual listings for each one. This is where it gets messy. A serving contains
The sum of the categorized fats is only 3.0 grams, one-half gram less than the total fat. Is it an error? Probably not. The one-half gram deficiency is most likely intentional — and perfectly legal. Let me explain...
Take a look at the ingredients list above (click image to see an larger version). One of the items is partially hydrogenated cottonseed oil, a trans fat (trans fats are notorious health hazards because they raise "bad" cholesterol levels and lower "good" cholesterol levels, thus increasing the chance of heart disease or stroke. More details at the American Heart Association or the Harvard School of Public Health). With that ingredient in the cracker, why is the trans fat content zero grams? The reason is rounding.
Fat quantities on the nutrition labels are expressed in 0.5 gram increments below 5 grams, and in 1.0 gram increments above 5 grams (according to 21 CFR 101.9(c)(2)). If a serving contains less than 0.5 grams of a fat, it can be expressed as zero grams on the label. And thus a serving can have almost 0.5 grams of trans fat and still have a zero on the trans fat line.
The Federal Register notice on the Final Rule on Labeling of Trans Fatty Acids explains why the FDA allowed manufacturers to "zero out" the trans fats on the label:
FDA notes that while these recommended levels might be quantifiable by laboratories using GC methodology such as that described in AOAC method 996.06 (Official Methods of Analysis of AOAC International, 17th edition, Revision 1, 2002) (Ref. 105), they will pose a problem for laboratories that are set up to quantify trans fatty acids by infrared spectroscopy (IR) methodology because the detection limits of the currently available IR methods are higher than those of the GC methods. More importantly, however, there are no unambiguous methods for confirming the very low levels suggested by the comment [0.1 g]. (from Volume 68, Number 133, page 41463)
In other words: it's too hard. This from an agency that has already approved technologies that splice bits of bacteria DNA into the DNA of plants, and wants to approve cloned animals for food?
With 100-calorie packs being all the rage in some circles, and with serving sizes being tuned so that the trans fat content is just under 0.5 grams (and thus 0.0 grams on the label), it would be easy to eat quite a few servings of "zero trans fat" foods in a day. The 2005 Dietary Guidelines for Americans says that you should "keep trans fatty acid consumption as low as possible," so it would be a good idea for FDA to find a way to show trans fats with accuracy of 0.1 grams or better.
The next time you see a Nutrition Facts label with "0 g trans fat," remember that in the FDA's world, zero is not zero. Check the ingredients list to be sure. Although meat and dairy products naturally contain small amounts of trans fat, most trans fats are artificially produced through the process of hydrogenation, so the key phrases to look for are "hydrogenated" and "partially hydrogenated."