Abstract
Objectives. To investigate differences in levels of plasma trans fatty acids (TFAs) and a broad set of other markers for cardiovascular disease (CVD) risk in the US adult population between 1999-2000 and 2009-2010.
Methods. Using a nationally representative sample of US adults aged 20 years and older from the 1999–2000 and 2009–2010 waves of the National Health and Nutrition Examination Survey, regression models were estimated to compare levels of TFAs and other markers for CVD risk over time.
Results. Significant declines in levels of plasma TFAs and improvements in a variety of other markers for CVD risk were observed in the US adult population.
Conclusions. Between the 2 time points, 1999–2000 and 2009–2010, there were substantial decreases in TFA levels and improvements in several other important cardiovascular health indicators in US adults.
A diet high in trans fatty acids (TFAs) is a risk factor for cardiovascular diseases (CVDs), such as heart disease and stroke.1 In fact, on a per-calorie basis, the intake of TFAs increases the risk of heart disease by more than any other nutrient.2 With heart disease still ranked as the leading cause of death and stroke not far behind,3 it is important to know how this dietary CVD risk factor has changed over time. There is evidence that levels of plasma TFAs in non-Hispanic White adults in the United States fell between 2000 and 2009,4 but it is unclear whether these reductions in TFA levels were unique to this subgroup in the adult population.
With 1999–2000 and 2009–2010 data on plasma TFAs—the latter of which were released in April 2016—from the National Health and Nutrition Examination Survey (NHANES), I analyzed differences in levels of plasma TFAs in a nationally representative sample of US adults aged 20 years and older. The mechanisms through which TFA consumption increases CVD risk include its adverse effects on cholesterol levels and its promotion of systemic inflammation.5 A complementary analysis examines differences in various other important markers for CVD risk.
METHODS
The NHANES assesses the health and nutritional status of Americans through a combination of interviews and physical examinations and is designed to be representative of the US civilian noninstitutionalized population. I selected adults aged 20 years and older in the 1999–2000 and 2009–2010 waves of the NHANES with information on the following variables for the analysis: 4 TFAs (elaidic acid, vaccenic acid, linoelaidic acid, and palmitelaidic acid); C-reactive protein (CRP); low-density and high-density lipoprotein cholesterol (LDL-C and HDL-C, respectively); triglycerides (TRI); and systolic and diastolic blood pressure. In addition to considering the sum of all 4 TFAs in the analysis, I considered 2 TFAs separately: one that is commonly found in partially hydrogenated oils ([PHOs], i.e., elaidic acid) and another that is commonly found in dairy products (i.e., vaccenic acid).
To estimate differences in TFA levels and other markers for CVD risk between waves, I estimated 3 regression models by ordinary least squares, including (model 1) only a constant term; (model 2) age, gender, race/ethnicity, education, and income; and (model 3) all covariates in model 2 plus indicators for being a smoker, being a high-risk drinker (defined as drinking an average of 4 or more drinks [5 or more drinks] on drinking days in the past year for women [men]),6 and being obese (defined as having a body mass index [defined as weight in kilograms divided by the square of height in meters] of ≥ 30). A comparison of coefficient estimates associated with the main independent variable—an indicator for the 2009–2010 NHANES wave—across these models allows for an examination of differences in plasma TFAs and other CVD risk factors that are independent of differences in other potentially important confounders.
There was significant skewness in the dependent variables, so I used log-transformed variables in the regression analysis. The main independent variable is a binary variable, so I obtained percentage changes over time by applying the formula 100 × exp(coefficient estimate) – 100. I estimated robust SEs using the appropriate sampling weights, strata, and primary sampling units to account for the complex survey design of the NHANES. Information on TFA, LDL-C, HDL-C, and TRI measures is available only for those in the fasting morning session, whereas information on the other measures is available for the full sample. I used fasting sampling weights in regressions involving TFA, LDL-C, HDL-C, and TRI. I used examination sampling weights in regressions involving CRP, systolic blood pressure, and diastolic blood pressure. I conducted all analyses using Stata version 14.1 (StataCorp LP, College Station, TX).
RESULTS
Table 1 shows the coefficient estimates associated with the indicator for the 2009–2010 NHANES wave from unadjusted and adjusted regression models. The estimates from these models are generally similar in magnitude. For brevity, the discussion focuses on the results from adjusted model 2. As shown in Table 1, the analysis revealed that levels of plasma TFAs significantly declined by an average of 52.5% from 1999–2000 to 2009–2010. Levels of elaidic acid fell by a greater amount (58.6%) than did those of vaccenic acid (51.3%). Results from a model that estimated both equations jointly in a seemingly unrelated regression framework indicate that these changes are significantly different from each other (P < .01). Significant improvements in other markers for CVD risk over time were also observed. LDL-C fell by 8.4%, HDL-C rose by 8.1%, TRI fell by 11.8%, CRP fell by 23.4%, systolic blood pressure fell by 2.7%, and diastolic blood pressure fell by 4.4%.
TABLE 1—
Differences in Plasma TFA Levels and Other Markers of CVD Risk Among Adults Aged 20 Years and Older: NHANES, United States, 1999–2000 and 2009–2010
| Dependent Variable (Logged) | Unadjusted Model,a Regression Coefficient (SE) | Adjusted Model 1,b Regression Coefficient (SE) | Adjusted Model 2,c Regression Coefficient (SE) | Sample Size, NO. | Change in Dependent Variablec % (95% CI) |
| Total TFAsd | –0.751 (0.021) | –0.734 (0.020) | –0.745 (0.020) | 3344 | –52.5 (–54.3, –50.6) |
| IP-TFA | –0.885 (0.022) | –0.868 (0.022) | –0.882 (0.022) | 3463 | –58.6 (–60.3, –56.8) |
| R-TFA | –0.730 (0.021) | –0.712 (0.021) | –0.720 (0.021) | 3490 | –51.3 (–53.3, –49.3) |
| LDL-C | –0.079 (0.013) | –0.086 (0.013) | –0.088 (0.013) | 3547 | –8.4 (–10.7, –6.2) |
| HDL-C | 0.079 (0.013) | 0.067 (0.011) | 0.078 (0.011) | 3626 | 8.1 (5.8, 10.5) |
| TRI | –0.107 (0.023) | –0.108 (0.022) | –0.126 (0.022) | 3623 | –11.8 (–15.6, –8.0) |
| CRP | –0.209 (0.037) | –0.207 (0.036) | –0.266 (0.034) | 7962 | –23.4 (–28.3, –18.1) |
| SYS-BP | –0.018 (0.004) | –0.024 (0.004) | –0.027 (0.004) | 7699 | –2.7 (–3.4, –2.0) |
| DIA-BP | –0.036 (0.005) | –0.042 (0.005) | –0.045 (0.005) | 7657 | –4.4 (–5.3, –3.6) |
Note. CI = confidence interval; CRP = C-reactive protein; CVD = cardiovascular disease; DIA-BP = diastolic blood pressure; HDL-C = high-density lipoprotein cholesterol; IP-TFA = elaidic acid; LDL-C = low-density lipoprotein cholesterol; NHANES = National Health and Nutrition Examination Survey; R-TFA = vaccenic acid; SYS-BP = systolic blood pressure; TFA = trans fatty acid; TRI = triglyceride. Each cell contains the coefficient estimate (SE) associated with an indicator for the 2009–2010 NHANES wave from a separate regression model estimated by ordinary least squares. I used NHANES sampling weights in all regressions. Robust SEs are in parentheses. Each coefficient estimate shown is statistically significant at the 1% or higher level.
Includes only a constant term.
Adjusted for age, gender, race/ethnicity, education, and income.
Adjusted for age, gender, race/ethnicity, education, income, and indicators for being a smoker, being a high-risk drinker, and being obese.
Total TFAs = sum of elaidic acid, vaccenic acid, linoelaidic acid, and palmitelaidic acid.
DISCUSSION
From 1999–2000 to 2009–2010, significant declines in levels of plasma TFAs were observed in the US adult population. Interestingly, a parallel analysis uncovered significant improvements in a host of other important cardiovascular health indicators, some of which have been causally linked to TFA consumption (e.g., LDL-C, HDL-C, and CRP).5 Taken together, the evidence shows that the cardiovascular health of US adults has improved along several dimensions over the course of a decade in the recent past.
Over the study period, 2 major policy initiatives were implemented with the goal of reducing TFA consumption. The Food and Drug Administration has required food manufacturers to declare TFA content on the nutrition facts label of packaged foods since 2006, and, starting in 2007, several localities and states have restricted the use of PHOs in food service establishments. The percentage of the population living in a jurisdiction with a PHO restriction grew from an estimated 3% to 20% between 2007 and 2010.7 It is possible that these regulations contributed to improvements in the cardiovascular health indicators observed in this study.8,9 The larger decline in a TFA commonly found in PHOs may be viewed as supporting this interpretation, because PHOs are the primary source of industrially produced TFAs, which are easier to eliminate from the food supply than are ruminant TFAs.10
It is important to note, however, that this study was limited to only 2 waves of the NHANES and that these are separated by a decade and the analysis cannot determine the causes of the differences in the outcomes analyzed. There were policy changes over the study period, for example, that did not target TFA consumption (e.g., antismoking regulations), which may have contributed to improvements in some of the cardiovascular health indicators I analyzed. An important direction for future research is to isolate and compare the causal impacts of a variety of policy actions on TFA levels and other CVD risk factors. Results from such a research endeavor may be useful as inputs into policymaking decisions regarding effective actions to reduce TFA consumption. This is especially true for European countries that are currently considering regulation of TFA levels in the food supply in an effort to improve cardiovascular health at the population level.11
PUBLIC HEALTH IMPLICATIONS
Analysis of a nationally representative sample of US adults aged 20 years and older revealed that levels of plasma TFAs fell substantially between 1999–2000 and 2009–2010. In addition, there were significant improvements in several other important markers for CVD risk, including blood pressure and levels of LDL and HDL cholesterol and triglycerides in the blood. About 1 in every 6 national health care dollars is spent on CVD. If the improvements in the cardiovascular health indicators I observed persist, they may help to mitigate the projected rise in the economic burden of CVD.12
HUMAN PARTICIPANT PROTECTION
No protocol approval was necessary because this study used publically available, de-identified data.
REFERENCES
- 1.Mozaffarian D, Katan MB, Ascherio A, Stampfer MJ, Willett WC. Trans fatty acids and cardiovascular disease. N Engl J Med. 2006;354(15):1601–1613. doi: 10.1056/NEJMra054035. [DOI] [PubMed] [Google Scholar]
- 2.European Commission. Report from the commission to the European Parliament and the Council regarding trans fats in foods and in the overall diet of the Union population. 2015. Available at: http://ec.europa.eu/food/safety/docs/fs_labelling-nutrition_trans-fats-report_en.pdf. Accessed October 12, 2016.
- 3.National Center for Health Statistics. Health, United States, 2015: with special feature on racial and ethnic health disparities. 2016. Available at: http://www.cdc.gov/nchs/data/hus/hus15.pdf. Accessed October 12, 2016. [PubMed]
- 4.Vesper HW, Kuiper HC, Mirel LB, Johnson CL, Pirkle JL. Levels of plasma trans-fatty acids in non-Hispanic White adults in the United States in 2000 and 2009. JAMA. 2012;307(6):562–563. doi: 10.1001/jama.2012.112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Mozaffarian D, Aro A, Willett WC. Health effects of trans-fatty acids: experimental and observational evidence. Eur J Clin Nutr. 2009;63(suppl 2):S5–S21. doi: 10.1038/sj.ejcn.1602973. [DOI] [PubMed] [Google Scholar]
- 6.US Department of Health and Human Services. 2015–2020 dietary guidelines for Americans. 2015. Available at: http://health.gov/dietaryguidelines/2015/guidelines. Accessed October 12, 2016.
- 7.National Restaurant Association. State and local enacted trans fat bans. 2013. Available at: https://www.restaurant.org/downloads/pdfs/advocacy/maps/map_transfat.pdf. Accessed October 12, 2016.
- 8.Downs SM, Thow AM, Leeder SR. The effectiveness of policies for reducing dietary trans fat: a systematic review of the evidence. Bull World Health Organ. 2013;91(4):262–269H. doi: 10.2471/BLT.12.111468. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Restrepo BJ, Rieger M. Trans fat and cardiovascular disease mortality: evidence from bans in restaurants in New York. J Health Econ. 2016;45:176–196. doi: 10.1016/j.jhealeco.2015.09.005. [DOI] [PubMed] [Google Scholar]
- 10.Stender S. In equal amounts, the major ruminant trans fatty acid is as bad for LDL cholesterol as industrially produced trans fatty acids, but the latter are easier to remove from foods. Am J Clin Nutr. 2015;102(6):1301–1302. doi: 10.3945/ajcn.115.123646. [DOI] [PubMed] [Google Scholar]
- 11.World Health Organization Regional Committee for Europe. European food and nutrition action plan 2015–2020. 2014. Available at: http://www.euro.who.int/__data/assets/pdf_file/0008/253727/64wd14e_FoodNutAP_140426.pdf. Accessed October 12, 2016.
- 12.Heidenreich PA, Trogdon JG, Khavjou OA et al. Forecasting the future of cardiovascular disease in the United States: a policy statement from the American Heart Association. Circulation. 2011;123(8):933–944. doi: 10.1161/CIR.0b013e31820a55f5. [DOI] [PubMed] [Google Scholar]