Trans-fats have been in the news sporadically in recent years, thanks largely to the bans passed by the New York City Health Department and the Indiana State Fair. Even consumers who don’t read the news have undoubtedly become familiar with the term because of food manufacturers labeling their products “0 Trans Fats!” (often with a small-print “per serving” which usually means there are trans fats in the product, just less than .5 g per serving) or “Trans-fat free!” whether or not they ever contained trans-fats in the first place.
But there still seems to be considerable confusion about what trans-fats are and why they might be bad for your health, which has likely been complicated by the long, stupid demonization of fats qua fats and saturated fats as a supposed cause of high cholesterol and heart disease.
Trans-fats are trans isomers of fatty acids, and although they occur naturally in small amounts in meat and dairy products from ruminants like cows and goats (in the form of vaccenic acid), the primary source of trans-fats in most Americans’ diets is hydrogenated vegetable oils. Most vegetable oils are composed primarily of unsaturated fats, which tend to be liquid at room temperature. In the early 20th Century, when the U.S. started importing soybeans as a source of cheap protein, soybean oil became readily available as a byproduct and was far cheaper than butter or lard. However, liquid and unsaturated fats get rancid much more quickly than solid fats, have a lower smoke point, and were unsuited to many American culinary traditions—biscuits and pastry crusts or all “short breads”* absolutely depend on solid fats to create their flaky texture, as explained in the note at the end.
Hydrogenation, a process first developed by French and German chemists around the turn of the century, provided the solution: heating the liquid, unsaturated fats in the presence of hydrogen turned them into solids at room-temperature. Apparently, cottonseed oil was also far cheaper than the beef tallow used in candles, so that one of the first uses of hydrogenated oils. It took a little marketing work to convince people it was also good eating–the major campaign for years was “Use Crisco, it’s digestible!” (okay, actually that probably just reflected the central nutritional concerns of the early 20th C: indigestion and dispepsia, see Hillel Schwartz’s Never Statisfied.
However, the process of hydrogenation also creates trans fats, and a different kind from the ones present in beef and dairy products. Unlike saturated fats, trans-fats produced through hydrogenation have been repeatedly correlated with coronary heart disease, including fatal heart attacks, in large, long-term epidemiological studies, including the Framingham Study. A review article on the available research on the relationship between dietary fat and coronary heart disease (CHD) published this past September concluded:
According to the classic ‘diet-heart’ hypothesis, high intake of SFAs [saturated fatty acids] and cholesterol and low intake of PUFAs [poly-unsaturated fatty acids] increase serum cholesterol levels and risk of CHD. However, few within-population studies have been able to demonstrate consistent associations with any specific dietary lipids, with the exception of trans fats and n–3 fatty acids.
In other words, everything you’ve heard in the last decade about trans fats (bad) and omega-3s (good) actually seems to be supported by the available research, unlike everything you’ve been told for the last five decades about saturated fats.
However, many of the claims about the threat posed by trans-fats allege that trans-fats raise LDL levels. And it’s not at all clear to me that anyone should be concerned about the fact that trans fats might be associated with increased cholesterol, even "bad" cholesterol, for reasons I discuss in the second entry on saturated fat. The only thing makes me think trans fats might actually be bad for people’s health is the consistent, strong association between trans fat consumption and increased risk of CVD and myocardial infarction. I know correlation =/= causation, and I haven’t found any good evidence about a proposed mechanism. On the basis of the current evidence, it seems like there’s a difference between the naturally-occurring trans-fats and the ones produced by hydrogenation:
The association was only seen for for trans fatty isomers from hydrogenated vegetable oils. The mainly different trans isomers from ruminant fats did not show such an association. A case-control study in 239 people suffering an acute myocardial infarction found that after adjustment for age, sex and energy intake, intake of trans fatty acids was directly related to risk of myocardial infarction . Those with the highest intake of trans fatty acids had twice the risk of myocardial infarction as those with the lowest intakes after adjusting for other cardiovascular risk factors. As with the Nurses Health Study, the association was only seen for trans isomers from partially hydrogenated vegetable oils. (from a 1995 lit review in the British Food Journal)
Is that because the hydrogenated oils contain linoleic acid, and meat and milk with naturally-occurring trans-fats don’t? I don’t know. I think the most interesting finding from Harvard’s Nurse’s Health Study, which is one of the studies that did show a weak but significant correlation between saturated fat consumption and CVD, is that the consumption of trans fats was associated with a much higher incidence of CVD than saturated fats. That finding, which is nearly two decades old, should probably be pretty infuriating to anyone who’s eaten hydrogenated-vegetable-oil margarine or shortening in the last two decades because it was supposedly a “healthier” alternative to butter or lard. Similar findings in several other studies prompted the more recent review article to conclude:
The observational evidence that TFA are independently associated with increased risk of CHD events is convincing, though based on a more limited body of evidence.
There is probably no direct relation between total fat intake and risk of CHD.
So the persistent recommendation from public health and nutrition authorities to reduce total fat consumption? Not supported by the available evidence. But the reason this isn’t a "things that won’t kill you" entry is because, well, trans-fats might.
What it means for how I eat
The short version: lard before margarine or shortening, except when the lard is shelf-stable or the margarine/shortening is produced using fractionation and palm oil.
This whole series of entries was inspired by the casual research I was doing to figure out what kind of fat I should use in the Christmas cookies I wanted to send to distant relatives. Before that, I generally used butter when baking, primarily for its flavor but also because of a vague belief I had that butter is more "natural" and thus potentially "healthier" than margarine or shortening—at least partially-inspired by the explosion of negative publicity about trans-fats in the last decade.
However, the goal of the Christmas cookies was to replicate recipes used by my grandmother, and those generally call for shortening. "Shortening" can refer to either animal or vegetable-based semi-solid fat, named for the "short breads" or "shortcake" they produce,* but in my family, it’s only ever meant one thing: big blue cans of Crisco, which even with the new formula, still includes trans-fat.
The new Crisco formula does contain substantially less trans-fat than previous formulas, and it’s worth noting that the studies showing that margarine-eating is associated with increased heart disease and death generally involved stick or hydrogenated margarine—which is where some people have gotten the idea that stick margarines are worse for you than tub margarines. To the extent that the former usually require greater hydrogenation in order to be firmer and replicate butter for baking, and thus contain more trans fats, that’s probably true. Margarines produced through fractionation, which contain no trans-fats, could be just as healthy as butter, perhaps healthier if they contain omega-3s, perhaps less if the saturated fat in butter is a protective against cholesterol oxidation. At that point it’s definitely in the realm of “it’s complicated and there’s no clear evidence either way.”
If I had been making the cookies for myself, I might have just substituted butter and never bothered to look into it further. But aside from the flavor difference, butter also produces a thinner, crisper, and more delicate cookie because it’s only about 80% fat compared to shortening’s 100% (the other 20% is mostly milk solids and water). That’s fine if you like that sort of thing, but in this case was a problem both for verisimilitude and stability—I need them to get to their destinations at least mostly intact.
Lard seemed like the next best choice, being, after all, the other traditional solid fat that hydrogenated vegetable shortening was developed to mimic. But I felt hesitant about that because even though the family members I’m sending them to aren’t vegetarians, lots of people react with horror to the idea of eating lard. I would basically never dare to make something with lard and take it to a potluck lest people ask for the ingredients and then react the way people did when this guy took lard-containing liver pate to a Christmas dinner last year: making horrid faces and saying lard "will kill you." It’s not like they could have expected it to be vegetarian or low-fat—it was liver pate. People are just scared and disgusted by the idea of eating lard. Even I had this lingering sense that lard was supposedly "artery-clogging" or somehow terrible for you, and assumed that was because it was high in saturated fat.
As any readers who’ve been swept up in the minor lard revival probably already know, lard is primarily composed of unsaturated fatty acids (56-62% depending on where on the pig it comes from and how it’s processed). So even if the reason you think lard is scary and bad is that you think saturated fat is scary and bad, lard is considerably better than butter (~36% unsaturated and 64% saturated), which people seem generally way less freaked out by.
BUT…the lard I bought, a green and white Armour container just like the one below, stocked in many supermarket’s meat coolers even though it doesn’t need to be refrigerated, contains not just rendered pork fat, but also hydrogenated lard. Or, as succinctly visually annotated by District Plates:
I’m still working through the lard, and deeply impressed by the pie crust and biscuits it produces, but I’m not convinced it’s a good idea to eat it. So ultimately, for the cookies, I decided to use a 50-50 combination of butter and the vegan shortening produced by Spectrum Organics, which is composed primarily of palm oil—naturally a solid at room temperature.
I can’t be entirely confident that butter made from corn-fed, antibiotic and hormone-treated cows is necessarily "healthier" than almost entirely non-hydrogenated margarines, like Smart Balance. After a lot of mostly-fruitless efforts to find more information about what trans-fats might actually do to human health, I finally concluded that, as with most issues of diet and health, it’s complicated. Which is to say, there’s not a lot of information out there, likely because it would be extremely difficult and expensive to study with enough control and for a long enough duration to make very good conclusions. But I also can’t find any evidence that butter, even from industrial-agriculture cows, is unhealthy.
The cookies tasted buttery, but weren’t too fragile, and apparently got to their destinations mostly intact.
In sum, the list of fats I’m happy eating and feeding to my loved ones:
- Non-hydrogenated lard
- Palm and Coconut Oils
- Margarine produced through fractionation rather than hydrogenation
- Liquid vegetable oils, though generally not in combination with saturated fats
And the fats I try to avoid:
- Hydrogenated margarine, lard, or shortening
Beyond that, I think you might as well choose on the basis of taste and texture preferences or moral considerations (animal welfare, environmental sustainability, labor issues, etc., not that those are any clearer or less complicated).
*Short breads are distinguished from yeasted breads by their lack of long gluten strands, which can’t form when the protein in the flour is coated in fat, although that’s not, apparently, the origin of the name, which instead refers to the crumbly, flaky texture. The use of the verb "shorten" to mean "to make something friable or crumbly" was first used in reference to the effect of sand in soil and applied to breads long before the existence and function of gluten was understood. So that seemed like a neat coincidence, at first, but then I realized if it was really just about coating the flour in fat, there’s no reason only solid fats would be referred to a "shortening" because oil could do that just as well, if not better. But you can’t just substitute oil for shortening and still get flaky, crumbly biscuits or pie crust. What actually produces the flaky texture is the combined lack of gluten strands and the chunks of solid fat that melt as the bread bakes, creating thin layers between the layers of flour. So what initially seemed like a really neat faux-etymology turns out to not be as cool after all. Le sigh.