Trans fatty acids (TFA) are monounsaturated or polyunsaturated fats with at least 1 trans double bond, rather than the cis double bond normally synthesized by mammals and most plants (1). Small amounts of TFA occur naturally in meats and dairy products, but most TFA in modern diets are the result of industrially produced partially hydrogenated vegetable oils (PHVO). Partial hydrogenation converts vegetable oils to semisolid fats that have attractive commercial properties for cooking, baking, and frying. The use of PHVO — and corresponding TFA content of foods — increased substantially during the 20th century. In recent years, evidence for cardiovascular and metabolic harms of TFA (1–4) in conjunction with federal nutrition labeling and regional limitations on TFA (5–7) have spurred some US companies to reformulate their products to reduce TFA. Estimates from National Health and Nutrition Examination Survey (NHANES) data suggest that mean US consumption of industrially produced TFA (IP-TFA) decreased between 2003 and 2006 (8) and, on the basis of blood TFA levels, that average US exposure declined by approximately 50% between 2000 and 2009 (9). We previously reported on 83 brand-name foods from US supermarkets and restaurants that were all reformulated to reduce TFA from 1993–2006 through 2008–2009 (10). Among these reformulated products, average TFA content was reduced 92%; we found a mean reduction of 1.8 g per serving (standard deviation [SD], 2.3 g/serving) in supermarket foods and 3.2 g per serving (SD, 1.5 g/serving) in restaurant foods, generally without increases in saturated fatty acids (SFA). However, we did not evaluate the denominator (ie, the products that were not reformulated).

Overall, this evidence suggests that PHVO use in the United States is decreasing. However, the pace at which such reformulations are occurring remains unclear, and it may also vary over time, by type of product, and between companies. We investigated the trends in reformulation of major brand-name TFA-containing products in the United States from 2007 through 2011, including whether the pace of reformulations increased, decreased, or remained similar over time and how reformulation efforts varied by type of food product and company.

The 270 US brand-name TFA-containing foods in our primary analysis comprised an extensive range of products, representing 48 major brands or manufacturers including General Mills, Kellogg Company, Heinz, ConAgra, Safeway, Giant, Campbell Soup, Sara Lee, and Walmart (Appendix A). By 2011, 178 products (66%) had been reformulated to reduce TFA (Table 1), with a mean absolute decline in TFA among these 178 products of 1.5 (SD, 1.3) g per serving and a mean percentage decline in TFA of 78% (SD, 26) (Table 2). Among these 178 products, 146 (82%) listed 0 g TFA on their Nutrition Facts panels in 2011, but only half of these (n = 73) had completely eliminated IP-TFA (no PHVO in the ingredients list). Half (n = 73) still contained PHVO and were assumed to contain 0.25 g TFA in all analyses. Thirteen of the original 270 products increased their TFA content from 2007 through 2011, with an average increase of 0.7 (SD, 0.2) g per serving (see supplementary Table 1, Appendix B).

Among all 270 products, the mean TFA content decreased from 1.9 (SD, 1.3) g per serving in 2007 to 0.9 (SD, 1.2) g per serving in 2011 (P for change < .001), an overall decline of 49%. This decline was predominantly due to reformulations that reduced TFA to less than 0.5 g per serving. Among products that continued to contain 0.5 g or more per serving, average TFA content did not significantly change over time (1.9 [SD, 1.3] g/serving in 2007 vs 1.9 [SD, 1.3] g/serving in 2011).

The pace of TFA reduction significantly decreased over time (Figure 1). The mean reduction among all products was 0.6 (SD, 1.2) g per serving (30% decline) from 2007 through 2008, 0.3 (SD, 0.9) g per serving (12% decline) from 2008 through 2010, and 0.1 (SD, 0.4) g per serving (3% decline) from 2010 through 2011 (P value for trend, < .001) (Table 2). This slowing in TFA reduction was due to fewer reformulations over time; restricting calculations to only those products that contained TFA at the start of each period, 40% were reformulated from 2007 through 2008, 31% from 2008 through 2010, and 15% from 2010 through 2011 (P value for trend, < .001) (Table 2). Even restricting analyses to those products that were reformulated, the magnitude of TFA reduction also declined over time (P for trend < .001). Among products that reduced TFA during each period, the average reduction was 75% (1.5 [SD, 1.3] g/serving decline) during 2007 and 2008, compared with 61% (1.0 [SD, 0.7] g/serving decline) during 2010 and 2011 (P = .009 comparing 2007–2008 with 2010–2011). The average TFA reduction from 2008 through 2010 was intermediate. Similarly, restricting to products containing 0.5 or more g per serving TFA at the start of each period, the pace of percentage decline in TFA slowed across years (P value for trend = .02) (Table 2).

Among the 31 products that reduced TFA from 2010 through 2011 (average reduction 1.0 [SD, 0.7] g/serving), the average SFA content was also reduced by 0.4 (SD, 1.5) g per serving, for an overall decrease in combined TFA and SFA of 1.4 (SD, 1.6) g per serving. Among the 15 products that increased TFA during 2010 and 2011 (average increase, 0.7 [SD, 0.5] g/serving), the average SFA content also increased by 0.2 (SD, 2.2) g per serving, and the total combined TFA and SFA content by 0.9 (SD, 2.1) g per serving.

Among food categories, the largest gram per serving declines in TFA were evident for doughnuts (−2.8 g/serving), crackers (−1.9 g/serving), and pies (−1.4 g/serving), and the largest percentage declines were seen in French fries/other potatoes (88% reduction), doughnuts (81%), and ice creams (73%), from 2007 through 2011 (Figure 2) (see supplementary Table 2, Appendix B). The smallest percentage declines were seen for rolls (15%), margarines (19%), and popcorns (19%). In all years, popcorn products contained the most TFA; popcorn products contained an average of 4.5 g per serving in 2007 and 3.8 g per serving in 2011, representing a modest decline (0.7 g/serving). In comparison, all 18 French fries/other potato products containing TFA in 2007 were reformulated to 0.5 g or less TFA per serving by 2008 and remained at this level through 2011. By 2011, only 5 of these 18 French fries/other potatoes contained PHVO and IP-TFA, while 13 had completely eliminated those substances. All 7 doughnut products and all 5 muffin products were also reformulated by 2011 to contain less than 0.5 g per serving TFA, with large gram per serving declines for doughnuts (−2.8 g/serving) and smaller declines for muffins (−0.3 g/serving). All 12 of these doughnut and muffin products continued to contain PHVO as an ingredient in 2011, indicating they still contained small amounts of TFA.

Among all 270 products, changes in TFA content were different by company or manufacturer (Figure 3) (see supplementary Table 3, Appendix B). These products represented 48 parent companies, with the most from General Mills (48 products), Kellogg (22 products), and H. J. Heinz (18 products) (Appendix A). Among the 20 companies having at least 4 TFA-containing products in 2007, the largest percentage declines in TFA occurred in products from Cole’s Quality Foods (100%), Schwan Food Company (90%), and Tasty Baking Company (87%), and the smallest declines were in foods made by American Pies (3%), Giant Foods (12%), and ConAgra Foods (13%). The largest gram per serving declines occurred in products from Schwan Food Company (2.6 g/serving) and Tasty Baking Company (2.5 g/serving), down to average TFA contents of 0.3 g and 0.2 g per serving, respectively. In all years, products from American Pies contained the most average TFA (>4 g/serving), with little decline between 2007 and 2011. In all years, products from Pinnacle Foods and Giant contained on average more than 2.0 g per serving TFA, with little change by 2011.

This variation in reformulations across companies was not simply due to differences in the types of products they sold across food categories. The observed differences persisted in analyses stratified by food category and parent company. For example, among identified popcorn products, the TFA content of Safeway’s product fell from 4.5 to 0 g per serving from 2007 through 2011, while the General Mills product increased from 5 g to 6 g per serving during the same period. (see supplementary Figures 1–16, Appendix B for complete stratified analysis by food categories and by parent companies). By 2011, only 3 of the 20 major companies (Cole’s Quality Foods, Interstate Bakeries, Tasty Baking Company) had reformulated all of their products identified in this analysis to contain less than 0.5 g per serving TFA. Of these, only Cole’s Quality Foods had completely eliminated PHVO from its 4 bread-based products in 2011. Among the 28 other companies having fewer than 4 TFA-containing products in our 2007 survey, 16 companies had reformulated all their products to contain less than 0.5 g per serving TFA, and 6 of these companies had fully eliminated PHVO in all their products by 2011.

In sensitivity analyses, we further evaluated the other 90 TFA-containing products that were discontinued by 2011. Across major food categories, the proportion of discontinued products varied from 0% to 56% (see supplementary Table 4, Appendix B). To test the theoretically largest trends in reformulations over time, we assumed that these discontinued products were “reformulated” to contain no PHVO and 0 g TFA (see supplementary Tables 5–6, Appendix B). As expected, estimated magnitudes of TFA declines were modestly larger. For instance, the theoretical average TFA content of all 360 foods was calculated to be 0.7 (SD, 1.1) g per serving in 2011, compared with the actual 0.9 (SD, 1.2) g per serving of the 270 foods still on the market. However, our findings for the numbers and amounts of TFA reformulations, and the slowing of reformulations over time, were similar.

In this analysis of major US store brand and brand-name food products containing TFA in 2007, we found that two-thirds (66%) were reformulated to reduce TFA by 2011. Among reformulated products, 82% reduced TFA to less than 0.5 g per serving, although only half of these had completely eliminated IP-TFA. Of all products, reformulation eliminated an average of approximately 1 g of TFA per serving, representing an average percentage reduction of 49%. These findings suggest that some food manufacturers, supported by oil processors, seed developers, and farmers, have made substantial progress in recent years in replacing PHVO with alternative oils.

Yet, both the pace of reformulation and magnitude of TFA reduction significantly decreased over time. This slowing reformulation was related to both little reformulation in certain food categories and failure of certain parent companies or manufacturers to embrace reformulation to replace PHVO with more healthful oils. For example, we identified substantial heterogeneity in TFA reformulations by both type of food product and parent company.

Differences in reformulation could partly reflect technical challenges of reformulating certain foods. Most French fries or other potato products, ice creams, and doughnuts were reformulated to contain less than 0.5 g per serving TFA by 2011, while foods in other food categories, such as popcorns, pies, margarines, and rolls, still averaged more than 1.5 g per serving TFA. However, in each of the latter food categories, we also identified individual products that reduced TFA to less than 0.5 g per serving. These findings suggest it is possible to produce marketable forms of these products with reduced or eliminated TFA; therefore, reasons other than technical difficulties with reformulation may be contributing to remaining high TFA contents in these types of products.

Differences in TFA reformulations among companies were also identified. Some manufacturers markedly reduced TFA contents of their products while others showed little evidence of progress. Our investigation indicated instances in which TFA reformulation efforts have been successful, both in food categories and among parent companies, and identified other food categories and companies where additional focus on reformulation strategies would be most relevant. For example, a recent study using data from the 1999–2002 NHANES survey listed French fries, popcorns, pies, and margarines among the top 10 food group contributors to TFA intake in the US population during that period (12). Our more contemporary analysis suggests that most French fry manufacturers have subsequently made strides toward eliminating IP-TFA, while popcorns and pies still require intensive reformulation efforts. Our results also indicate that, although TFA consumption in the United States decreased from 2006 through 2009 (8,9), many major national products in the food supply continue to contain high amounts of TFA, and reformulations have slowed after 2008 and especially after 2010.

By 2011, 73 (50%) of the 146 products listing 0 g TFA on the Nutrition Facts panel still listed PHVO in the ingredients panel. This indicates continued presence of IP-TFA in these 73 products, theoretically up to 0.49 g per serving. However, the average reduction in TFA in these products was still large: 1.4 g per serving. This finding suggests that consumers can view a “0 g TFA” listing on a reformulated product with cautious optimism, as a substantial proportion of TFA content has been eliminated. Nonetheless, when consumed across multiple servings and different products, these remaining amounts of PHVO pose a health risk and should be eliminated. Given the many products that were not reformulated at all, our findings suggest that public health attention should focus on reformulation of current high-TFA foods; additional methods to fully eliminate TFA in foods with less than 0.5 g TFA can follow.

Our study has limitations. The observed trends in TFA reformulations are specific to the 360 identified products, and these foods did not represent an exhaustive accounting of every possible TFA-containing food in the United States or a random sampling of all products made by all companies. Certain food categories, such as frosting, may have been underrepresented in our initial survey. However, our large sample from major US grocers likely provides a reasonable national representation of major packaged food products in the United States and of progress in TFA reformulation in these foods from 2007 through 2011. In addition, our investigation represents the most comprehensive contemporary assessment of TFA reformulations of packaged foods sold in the United States. We evaluated total TFA above 0.5 g per serving, as current labeling methods do not distinguish between industrial and natural sources. Some TFA on the Nutrition Facts panel could represent TFA from ruminant-derived products or nonhydrogenated vegetable oils (eg, due to deodorization). However, ruminant TFA would represent a small proportion of total TFA at or above 0.5 g per serving; below 0.5 g per serving, we based our assessment on the presence of PHVO. We did not assess new products introduced after 2007, some of which could contain TFA. However, a recent US Department of Agriculture evaluation of new products introduced between 2006 and 2010 found a minority of new products (5.3%) containing TFA, with decreasing introductions over time (13). Inclusion of these new products could increase the absolute exposure of the US population to TFA but would not alter our conclusions regarding reformulations of existing TFA-containing products. Nonetheless, even these new product introductions will further slow the pace of elimination of IP-TFA from the food supply, and introduction of new products that contain IP-TFA should be eliminated. We also did not collect any information on the prices of the products, which may have changed over time.

Our findings suggest that some food manufacturers have made progress in reducing TFA in their US products, but substantial variation exists by food type and by parent company, and overall progress has significantly slowed over time. Our findings also indicate that, even among this subset of national products, many products remain in the food supply that contain substantial amounts of IP-TFA. Because TFA consumption is harmful even at low levels (1,14), our results emphasize the need for continued major efforts and commitment toward reformulating (or discontinuing) foods to eliminate PHVO, with particular focus on certain food categories and companies. The heterogeneity of success also suggests that nationally coordinated policy efforts, including both voluntary and regulatory strategies and surveillance of key foods and companies, would be most effective to eliminate exposure to IP-TFA in the general population. For instance, the Food and Drug Administration could solve the problem expeditiously by declaring PHVO as no longer “generally recognized as safe,” thereby placing limits on the presence of IP-TFA in foods (small amounts occur when vegetable oil is purified), as several other nations including Austria, Denmark, Iceland, Sweden, and Switzerland have done (15).