Abstract
Background:
Few studies have examined the association of long-chain n-3 polyunsaturated fatty acids (LCn-3 PUFAs) with the measures of atherosclerosis in the general population.
Objective:
To examine the relationship of total LCn-3PUFAs, eicosapentaenoic acid (EPA), and docosahexaenoic acid (DHA) with aortic calcification.
Methods:
In a multiethnic population-based cross-sectional study of 1033 asymptomatic men aged 40-49 years (310 US-White, 107 US-Black, 303 Japanese American, and 313 Japanese in Japan), we examined the relationship of serum LCn-3PUFAs to aortic calcification (measured by electron-beam computed tomography and quantified using the Agatston method) using Tobit regression and ordinal logistic regression after adjusting for conventional cardiovascular risk factors and potential confounders.
Results:
Overall 56.5% participants had an aortic calcification score (AoCaS) >0. The means (SD) of total LCn-3PUFAs, EPA, and DHA were 5.8% (3.3%), 1.4% (1.3%), and 3.7% (2.1%), respectively. In multivariable-adjusted Tobit regression, a 1-SD increase in total LCn-3PUFAs, EPA, and DHA was associated with 29% (95% CI= 0.51, 1.00), 9% (95% CI= 0.68, 1.23), and 35% (95% CI= 0.46, 0.91) lower AoCaS, respectively. Results were similar in ordinal logistic regression analysis. In a race/ethnicity-stratified analysis, total LCn-3PUFAs and DHA were inversely and significantly associated with AoCaS among US-White but not in other races/ethnicities.
Conclusions:
This study shows the significant inverse association of LCn-3PUFAs with aortic calcification independent of conventional cardiovascular risk factors among males in the general population. This association appeared to be driven by DHA but not EPA.
Keywords: aorta, atherosclerosis, calcification, docosahexaenoic acid, eicosapentaenoic acid, long-chain n-3 polyunsaturated fatty acid
1. Introduction
Meta-analyses of prospective observational studies and few randomized controlled trials (RCTs) documented a protective effect of long-chain n-3 polyunsaturated fatty acids (LCn-3PUFAs) [eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA)] on cardiovascular health, particularly lowering the risk of cardiac mortality [1-4]. The beneficial effect of LCn-3PUFAs on cardiac mortality is mainly attributed to their antiarrhythmic activity [2, 5]. Other cardiovascular benefits of LCn-3PUFAs include the lowering of triglycerides, blood pressure, resting heart rate, cytokine formation, platelet aggregation, and inflammatory markers; and improvement in endothelial dysfunction, arterial compliance, and vascular reactivity [2]. It is also speculated that LCn-3PUFAs inhibit the atherosclerosis process (the major underlying pathophysiological process of coronary heart disease (CHD)) [6-8] by lowering inflammation, improving endothelial function, and increasing atherosclerotic plaque stability [3, 9-11]. Animal studies [12-14] and basic research [15] strongly support the anti-atherosclerotic properties of LCn-3PUFAs. However, the limited number of studies conducted in a healthy human population reported mixed findings with some documenting no significant association [16-18] and another reporting a significant inverse association [19].
Aortic calcification is a reliable biomarker of generalized atherosclerosis [20] and has a graded and consistent relationship with CHD beyond traditional cardiovascular risk factors [21, 22]. Aortic calcification may be a better measure of atherosclerosis than coronary artery calcification (CAC) - a well-established biomarker of coronary atherosclerosis; because it adds valuable prognostic information of cardiovascular risk beyond CAC [23, 24]; develops earlier and more extensively than in any other vascular bed [25]; and may have a stronger association with cardiovascular risk factors than CAC [23, 24, 26]. To our knowledge, however, no previous study has examined the association of serum biomarkers of LCn-3PUFAs and aortic calcification in the general population.
We examined the relationship of LCn-3 PUFAs to aortic calcification in 1033 asymptomatic middle-aged men in the ERA-JUMP Study [the Electron Beam Computed Tomography (EBCT), Risk-Factor Assessment among Japanese and the United States (US) Men in the Post-World-War-II birth cohort]. Based on our previous finding of a significant inverse association of LCn-3PUFAs with CAC and carotid intima-media thickness (CIMT) [19] and reported differential significant association of DHA compared to EPA with endothelial dysfunction (a precursor of atherosclerosis) [3], we hypothesized that serum total LCn-3PUFAs especially DHA but not EPA would have a significant inverse association with aortic calcification. It is important to examine the relationship between LCn-3PUFAs and atherosclerosis in the general population to gain further insight into the relationship between LCn-3PUFAs and CHD.
2. Materials and methods
2.1. Participants
The ERA-JUMP Study is a population-based study of 1033 men aged 40–49 years comprising US-White, US-Black, Japanese American, and Japanese in Japan. The details of the study protocol have been described previously [27]. Briefly, during 2002–2006, a population-based sample of 1033 men aged 40–49 years, without clinical cardiovascular diseases (CVD) or other severe illnesses, was obtained from 3 centers: 310 White and 107 Black from Pittsburgh, Pennsylvania, US; 303 Japanese American from Honolulu, Hawaii, US; 313 Japanese from Kusatsu City, Shiga, Japan. The study protocol complied with the Helsinki Declaration as revised in 1983. We obtained study approvals from the Institutional Review Boards of University of Pittsburgh, Pittsburgh, US; Kuakini Medical Center, Honolulu, US; Shiga University of Medical Science, Otsu, Japan. All participants gave written informed consent. We excluded participants with missing data for aortic calcification (n=27) and LCn-3PUFAs (n=8). Our final sample size was 998 with 300 US-White, 101 US-Black, 287 Japanese American, and 310 Japanese in Japan.
2.2. Risk factor assessment
As published elsewhere, participants underwent a physical examination, completed a set of lifestyle questionnaire, and a laboratory assessment [28, 29]. Body weight and height were measured while the participant was wearing minimal light clothing without shoes. The formula used to calculate body mass index (BMI) was ‘weight in kilograms divided by the square of the height in meters.' Participants with systolic blood pressure ≥140 mmHg, diastolic blood pressure ≥90 mmHg or use of antihypertensive medications were considered hypertensive. Participants were classified as smokers if they reported current use of cigarettes or having stopped smoking within the past 30 days. The formula used to calculate pack-years of smoking was ‘years of smoking multiplied by the number of cigarettes per day divided by 20’. Medication use (antihypertensive, antidiabetic, or lipid-lowering) was reported as ‘yes/no’. Meat intake was defined as individuals who ate beef, pork, or sausage ≥2 times per week. Self-reported physical activity related to the current job was categorized into sedentary, light, medium, and heavy physical activity.
Venipuncture was performed early in the clinic visit after a 12-hour fast. Blood samples were stored at −70°C and shipped on dry ice from all the centers to the University of Pittsburgh. Serum/plasma samples were assayed for glucose, lipids [including total cholesterol, triglycerides, low-density lipoprotein cholesterol (LDL-C), and high-density lipoprotein cholesterol (HDL-C)], fibrinogen, and C-reactive protein (CRP) as described previously [30]. Participants with fasting glucose ≥7.0 mmol/L or using medications for diabetes were recorded as diabetics.
2.3. n-3 and n-6 polyunsaturated fatty acids assessment
Serum levels of n-3 PUFAs [EPA (20:5n-3), docosapentaenoic acid (DPA (22:5n-3)), DHA (22:6n-3), and α-linolenic acid (ALA (20:3n-3), and n-6 PUFAs [linoleic (LA (18:2n-6)) and arachidonic (ARA (20:4n-6)) acids] were measured using Capillary Gas-Liquid chromatography [27]. Serum fatty acids are reported as weight percentages of total fatty acids. The coefficients of variation between tests in the ERA-JUMP study for EPA, DPA, DHA, ALA, LA, ARA, and total fatty acids were 4.5%, 4.5%, 7.2%, 1.6%, 7.9%, 2.8%, and 5.7%, respectively. Total LCn-3PUFAs was defined as the sum of EPA, DPA, and DHA.
2.4. Aortic calcification assessment
To evaluate aortic calcification, 6 mm EBCT images using a GE-Imatron C150 scanner (GE Medical Systems, South San Francisco, US) [30, 31], were acquired from the aortic arch to the iliac bifurcation. Readings of the scans from all centers were performed centrally at the Cardiovascular Institute, University of Pittsburgh, using a DICOM (Digital Imaging and Communications in Medicine) workstation and software by AccuImage (AccuImage Diagnostic Cooperation, San Francisco, US). The software program implements the widely accepted Agatston scoring method [32]. Trained radiology technicians- blinded to each participant’s characteristics and to the study centers, evaluated the readings. The reproducibility of non-zero Agatston aortic calcification score (AoCaS) had an intra-class correlation of 0.98.
2.5. Statistical Analysis
Continuous variables with approximately normal distribution: total LCn-3PUFAs, EPA, DHA, ALA, LA, ARA, age, BMI, LDL-C, and HDL-C were standardized. Distribution of AoCaS was highly skewed, therefore we created four categories of AoCaS: 0, 1-99, 100-299, and ≥300. Across four categories of AoCaS (i) age and race/ethnicity adjusted BMI, LDL-C, and HDL-C were expressed as means±standard error of mean (SEM); (ii) age and race/ethnicity adjusted triglycerides, years of education, and the pack-years of smoking were expressed as medians and interquartile range; (iii) age and race/ethnicity adjusted categorical variables were expressed in percentages. P-values for trend across the different categories of AoCaS were determined using: linear regression when a response variable was continuous with a normal distribution; quartile regression when a response variable was continuous with skewed distribution; and logistic regression when a response variable was categorical.
We used Tobit conditional regression to determine the independent association of total LCn-3PUFAs, EPA, or DHA with aortic calcification adjusting for potential confounders. For Tobit regression, outcome variable AoCaS was log transformed after the addition of one unit [natural log of (AoCaS + 1)]. Secondarily, we performed ordinal logistic regression to assess the likelihood of study participants being in a higher category of AoCaS. We used the Brant test to evaluate the proportional odds assumption of ordinal regression. For both Tobit regression and ordinal regression: Model I was adjusted for sociodemographic variables (age, race/ethnicity, and years of education); Model II was further adjusted for potential confounders (pack-years of smoking, alcohol consumption, BMI, diabetes, lipid-lowering medications, LDL-C, physical activity at job, and meat intake); Model III was additionally adjusted for intermediary variables (hypertension, HDL-C, triglycerides, CRP, and fibrinogen) in the relationship between LCn-3PUFAs and atherosclerosis/CHD. In model III, we tested for an interaction between race/ethnicity and total LCn-3PUFAs (or EPA or DHA) on aortic calcification. In regression models, we treated total LCn-3PUFAs, EPA, and DHA as categorical variables (race-specific quartiles) or continuous variables, separately. We also assessed the independent association of ALA, LA, and ARA using similar regression techniques and models as mentioned above. The inclusion of variables in the regression models was mainly based on previously published literature on the relationship of LCn-3PUFAs to atherosclerosis/CHD. In Tobit regression and in ordinal regression, a p-value for linear trend across the quartiles of LCn-3PUFAs was calculated using contrast.
Sensitivity analyses were conducted: (i) Excluding Japanese study participants, as serum median levels of LCn-3PUFAs among Japanese in Japan were ≥2 times higher than in the other study participants [27]; and (ii) Stratifying the analysis by race/ethnicity. All p-values were two-tailed and a p-value <0.05 was considered as significant. SAS version 9.4 (SAS Institute, Cary, NC, US) and STATA version 14.0 (StataCorp LP, College Station, TX, US) were used for all statistical analyses.
3. Results
Total number (%) of participants with AoCaS >0 was 434 (56.5): 94 (31.3) US-White, 31 (30.7) US-Black, 199 (64.2) Japanese in Japan, and 110 (38.3) Japanese American. Except for HDL-C, total LCn-3PUFAs, EPA, and DHA, participants with AoCaS >0 had a higher age and race/ethnicity-adjusted: BMI, pack-years of smoking, LDL-C, triglycerides, CRP, and fibrinogen; had a greater proportion with diabetes and hypertension; and were more likely to be on lipid-lowering medications compared to zero AoCaS category (Table 1). The mean (SD) of LCn-3PUFAs (%) among individual races was: US-White 3.8 (1.7), US-Black 3.8 (1.5), Japanese in Japan 9.3 (3.0), and Japanese American 4.8 (2.2).
Table 1.
Demographic and clinical characteristics by AoCaS categories for the ERA-JUMP Study, 2002-2006 (n=998)
| AoCaS categories | AoCaS= 0 | AoCaS 0 - 99 | AoCaS 100-299 | AoCaS ≥300 | p-trendd |
|---|---|---|---|---|---|
| Total number (%) | 434 (43.5) | 367 (36.8) | 89 (8.9) | 108 (10.8) | - |
| Agea (years) | 44.8 (0.1) | 45.4 (0.2) | 46.2 (0.3) | 46.3 (0.3) | 0.01 |
| BMIa (kg/m2) | 26.2 (0.3) | 28.7 (0.3) | 29.0 (0.5) | 27.3 (0.5) | 0.01 |
| Pack-years of smokingb | 0.0 (0.0, 0.9) | 0.0 (0.0, 1.0) | 0.0 (0.0, 12.8) | 11.0 (0.8, 19.7) | 0.01 |
| Alcoholb (gm/day) | 5.0 (1.0, 15.9) | 3.6 (1.0, 13.9) | 3.5 (1.0, 11.5) | 24.5 (2.1, 35.0) | 0.01 |
| LDL-Ca (mg/dL) | 130.6 (2.6) | 137.9 (2.3) | 131.8 (4.1) | 138.2 (3.8) | 0.16 |
| HDL-Ca (mg/dL) | 50.4 (1.0) | 46.3 (0.9) | 46.2 (1.5) | 49.8 (1.5) | 0.71 |
| Triglyceridesb (mg/dL) | 114.7 (80.7, 154.4) | 130.7 (95.7, 184.9) | 148.7 (105.7, 207.9) | 138.7 (95.7, 240.5) | 0.01 |
| Hypertensionc | 42 (9.5) | 59 (16.1) | 18 (19.9) | 17 (15.2) | 0.02 |
| Diabetesc | 9 (2.1) | 10 (2.7) | 6 (6.5) | 5 (4.6) | 0.01 |
| Anti-lipid medc | 39 (9.0) | 41 (11.0) | 23 (25.8) | 16 (14.5) | 0.01 |
| Meat intakec | 324 (74.7) | 287 (78.0) | 72 (80.0) | 83 (76.4) | 0.61 |
| Years of educationb | 17.0 (16.0, 18.0) | 17.0 (16.0, 18.0) | 16.0 (16.0, 18.0) | 16.0 (16.0, 18.0) | 0.01 |
| CRPb (mg/dL) | 0.8 (0.4, 1.6) | 1.0 (0.6, 1.9) | 1.0 (0.5, 2.0) | 1.0 (0.5, 1.9) | 0.05 |
| Fibrinogena (mg/dL) | 284.6 (5.3) | 293.0 (4.8) | 300.0 (8.2) | 298.6 (7.7) | 0.05 |
| Total LCn-3PUFAsa (%) | 4.2 (0.2) | 3.7 (0.2) | 3.6 (0.3) | 3.7 (0.3) | 0.08 |
| EPAa (%) | 0.9 (0.1) | 0.8 (0.1) | 0.8 (0.1) | 0.8 (0.1) | 0.74 |
| DHAa (%) | 2.6 (0.1) | 2.3 (0.1) | 2.2 (0.2) | 2.3 (0.2) | 0.01 |
| ALAa (%) | 0.3 (0.0) | 0.3 (0.0) | 0.3 (0.0) | 0.3 (0.0) | 0.63 |
| LAa (%) | 30.3 (0.3) | 30.0 (0.3) | 29.2 (0.5) | 28.9 (0.5) | 0.01 |
| ARAa (%) | 0.3 (0.0) | 0.3 (0.0) | 0.3 (0.0) | 0.3 (0.0) | 0.17 |
AoCaS, aortic calcification score; BMI, body mass index; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; CRP, C-reactive protein; EPA, eicosapentaenoic acid; DPA, docosapentaenoic acid; DHA, docosahexaenoic acid; ALA, α-linolenic acid; LA, linoleic acid; ARA, arachidonic acids; Total LCn-3PUFAs, total long chain omega 3 polyunsaturated fatty acids defined as sum of EPA, DPA, and DHA; n(%), number(%);
Values for all variables except age were adjusted for ‘age’ and ‘race/ethnicity’: value of age was fixed at 45.3 years and race was fixed as ‘US White’.
Continuous normally distributed variables were expressed as mean (standard error);
Continuous variables with skewed distribution were expressed as median (inter-quartile range);
Categorical variables were expressed as numbers (%);
p-trend shows a p-value for linear trend across the AoCaS categories;
SI conversion factors: To convert LDL-C and HDL-C to mmol/L, multiply values by 0.0259. To convert triglycerides to mmol/L, multiply values by 0.01129. To convert fibrinogen to μmol/L, multiply values by 0.0294. To convert CRP to nmol/L, multiply values by 9.524.
In Tobit regression, participants in the fourth quartile compared to the first quartile of total LCn-3PUFAs had a 49% lower AoCaS [Model II: Tobit ratio (TR) (95% CI)= 0.51 (0.26, 0.97)] (Table 2). After further adjustment for intermediary variables, this significant inverse association was attenuated and became nonsignificant [Model III: TR (95% CI)= 0.55 (0.28, 1.08)]. In model II, a 1-SD (3.3%) increase in total LCn-3PUFAs was associated with a 29% lower AoCaS [TR (95% CI)= 0.71 (0.51, 1.00)]. EPA was not significantly associated with aortic calcification. In model II, a 1-SD increase in EPA (1.3%) was associated with a 9% lower expected AoCaS [TR (95% CI)= 0.91 (0.68, 1.23)]. Participants in the fourth quartile of DHA compared to the first quartile had significantly lower AoCaS in all models– from unadjusted to fully adjusted model III. There was a significant dose-response relationship between DHA and aortic calcification (p for linear trend <0.05). In model II, a 1-SD (2.1%) increase in DHA was associated with a 35% lower expected AoCaS [TR (95% CI)= 0.65 (0.46, 0.91)]. This significant inverse association remained after further adjustment for intermediary variables [Model III: TR (95% CI)= 0.69 (0.49, 0.98)] (Table 2) and ALA, LA, and ARA [TR (95% CI)= 0.68 (0.47, 0.98)] (Supplementary Table 1-B). In model II, ARA but not ALA or LA had a significant inverse association with aortic calcification [TR (95% CI)= 0.64 (0.48, 0.86)] (Supplementary Table 1-A). When the analysis was repeated using ordinal regression, the findings remained materially the same. There was no significant interaction between race/ethnicity and total LCn–3PUFAs, EPA, DHA, ARA, ALA, or LA on aortic calcification. In a supplementary analysis excluding Japanese in Japan, results were similar to the analysis for all study participants. In multivariable-adjusted regression models, a 1-SD increase in total LCn-3PUFAs (1.9%) and DHA (1.3%) but not EPA (0.7%) was significantly and inversely associated with aortic calcification (Supplementary table 2).
Table 2.
Association between LCn-3PUFAs and aortic calcification: Tobit ratio/odds ratio (95% confidence interval) for AoCaS by the quartiles and per 1-SD of total LCn-3PUFAs/EPA/DHA (n=998)
| LCn-3PUFAs quartiles |
Q1 | Q2 | Q3 | Q4 | LCn-3PUFAs (per SD change) |
|
|---|---|---|---|---|---|---|
| Tobit Regression | ||||||
| TR | TR (95% CI) | TR (95% CI) | TR (95% CI) | p-trenda | TR (95% CI) | |
| Total LCn-3PUFAs | ||||||
| Model I | 1 | 0.74 (0.38, 1.46) | 0.73 (0.37, 1.44) | 0.38 (0.19, 0.75) | 0.01 | 0.62 (0.43, 0.88) |
| Model II | 1 | 0.85 (0.45, 1.60) | 0.91 (0.48, 1.72) | 0.51 (0.26, 0.97) | 0.06 | 0.71 (0.51, 1.00) |
| Model III | 1 | 0.87 (0.46, 1.65) | 1.03 (0.54, 1.98) | 0.55 (0.28, 1.08) | 0.14 | 0.76 (0.54, 1.08)b |
| EPA | ||||||
| Model I | 1 | 0.59 (0.30, 1.17) | 0.64 (0.32, 1.28) | 0.57 (0.29, 1.12) | 0.14 | 0.92 (0.67, 1.26) |
| Model II | 1 | 0.60 (0.32, 1.13) | 0.59 (0.31, 1.11) | 0.58 (0.30, 1.11) | 0.10 | 0.91 (0.68, 1.23) |
| Model III | 1 | 0.61 (0.32, 1.15) | 0.65 (0.34, 1.24) | 0.64 (0.33, 1.24) | 0.22 | 0.96 (0.71, 1.29) |
| DHA | ||||||
| Model I | 1 | 0.80 (0.41, 1.58) | 0.59 (0.30, 1.17) | 0.30 (0.15, 0.60) | 0.01 | 0.52 (0.36,0.74) |
| Model II | 1 | 0.85 (0.46, 1.60) | 0.74 (0.39, 1.40) | 0.43 (0.22, 0.83) | 0.01 | 0.65 (0.46, 0.91) |
| Model III | 1 | 0.86 (0.45, 1.62) | 0.83 (0.43, 1.57) | 0.47 (0.24, 0.93) | 0.03 | 0.69 (0.49, 0.98) |
| Ordinal Logistic Regressionc | ||||||
| OR | OR (95% CI) | OR (95% CI) | OR (95% CI) | p-trenda | OR (95% CI) | |
| Total LCn-3PUFAs | ||||||
| Model I | 1 | 0.86 (0.62, 1.20) | 0.85 (0.61, 1.19) | 0.61 (0.43, 0.85) | 0.83 | 0.77 (0.65, 0.92) |
| Model II | 1 | 0.91 (0.65, 1.29) | 0.95 (0.67, 1.35) | 0.69 (0.48, 0.99) | 0.93 | 0.84 (0.70, 1.01) |
| Model III | 1 | 0.94 (0.66, 1.33) | 1.04 (0.73, 1.48) | 0.74 (0.52, 1.07) | 0.81 | 0.88 (0.73, 1.06)d |
| EPA | ||||||
| Model I | 1 | 0.78 (0.56, 1.09) | 0.82 (0.59, 1.14) | 0.72 (0.52, 1.01) | 0.25 | 0.94 (0.81, 1.09) |
| Model II | 1 | 0.78 (0.55, 1.10) | 0.76 (0.54, 1.07) | 0.73 (0.51, 1.04) | 0.19 | 0.95 (0.81, 1.12) |
| Model III | 1 | 0.78 (0.55, 1.11) | 0.80 (0.56, 1.14) | 0.78 (0.55, 1.12) | 0.21 | 0.98 (0.84, 1.16) |
| DHA | ||||||
| Model I | 1 | 0.91 (0.65, 1.26) | 0.78 (0.56, 1.08) | 0.55 (0.40, 0.78) | 0.93 | 0.72 (0.60, 0.86) |
| Model II | 1 | 0.94 (0.67, 1.33) | 0.87 (0.62, 1.23) | 0.65 (0.45, 0.92) | 0.83 | 0.80 (0.67, 0.97) |
| Model III | 1 | 0.95 (0.67, 1.35) | 0.93 (0.65, 1.32) | 0.70 (0.49, 1.01) | 0.82 | 0.84 (0.70, 1.02) |
AoCaS, aortic calcification score; Q1, Q2, Q3, Q4: quartiles of LCn-3PUFAs/EPA/DHA; IQR, interquartile range; TR, Tobit ratio; OR, odds ratio; CI: confidence interval; SD: standard deviation; EPA, eicosapentaenoic acid; DPA, docosapentaenoic acid; DHA, docosahexaenoic acid; Total LCn-3PUFAs, total long chain omega 3 polyunsaturated fatty acids defined as the sum of EPA, DPA, and DHA;
For total LCn-3PUFAS, median (IQR) for Q1, Q2, Q3, and Q4 equals to 2.7 (2.3, 5.0), 3.8 (3.1, 7.7), 4.9 (4.0, 9.0), and 7.3 (5.9, 11.8) respectively; For EPA, median (IQR) for Q1, Q2, Q3, and Q4 equals to 0.4 (0.4, 0.9), 0.6 (0.6, 1.8), 0.9 (0.7, 2.4), and 1.9 (1.2, 3.6) respectively; For DHA, median (IQR) for Q1, Q2, Q3, and Q4 equals to 1.6 (1.3, 3.4), 2.5 (1.8, 4.9), 3.2 (2.6, 5.9), and 4.9 (4.0, 7.3) respectively; One SD of total LCn-3PUFAs, EPA, and DHA equals to 3.3%, 1.3%, and 2.1% respectively.
Model I: Fatty acids (Total LCn-3PUFAs or EPA or DHA), age, race, years of education;
Model II: Model I + pack-years of smoking, alcohol consumption, BMI, diabetes, lipid-lowering medications, LDL-C, physical activity at the job, and meat intake;
Model III: Model II + hypertension, HDL-C, triglycerides, CRP, and fibrinogen;
p-trend shows a p-value for linear trend across the quartiles of total LCn-3PUFAs/EPA/DHA calculated using contrast;
Tobit ratio of 0.76 means that with every SD (3.3%) increase in total LCn-3PUFAs, there is a 26% decrease in the AoCaS;
For ordinal logistic regression, AoCaS categories used were 0, 1-99, 100-299, and ≥300;
OR of 0.88 means that a 1-SD (3.3%) increase in total LCn-3PUFAs made a study participant in this study 12% less likely to be in a higher AoCaS category;
In a race/ethnicity-stratified analysis, in US-White, in multivariable-adjusted Tobit regression, there was a significant inverse association of total LCn-3PUFAs [Model III: TR (95% CI)= 0.43 (0.22, 0.86)] and DHA [Model III: TR (95% CI)= 0.47 (0.26, 0.85)] but not of EPA [Model III: TR (95% CI)= 0.60 (0.27, 1.34)] with aortic calcification. In US-Black, Japanese in Japan, and Japanese American, total LCn-3PUFAs and DHA were inversely and nonsignificantly associated with aortic calcification. Results were consistent in ordinal logistic regression analysis.
4. Discussion
In this community-based sample of healthy middle-aged men, blood levels of total LCn-3PUFAs were significantly and inversely associated with aortic calcification independent of cardiovascular risk factors. This significant inverse association appeared to be driven by DHA. DHA was significantly and inversely associated with aortic calcification after adjustment for conventional cardiovascular risk factors and other fatty acids including ALA, LA, and ARA. The race/ethnicity stratified analysis suggest that US-White men may differ from other race/ethnicity in their response LCn-3PUFAs. To our knowledge, this is the first community-based study examining the relationship between blood biomarkers of LCn-3PUFAs and aortic calcification in asymptomatic middle-aged men across different races/ethnicities from two countries in a standardized manner.
In contrast to our findings, He et al. in the Multi-Ethnic Study of Atherosclerosis [16] of 5488 healthy US adults from four different races/ethnicities, aged 45–84 years without clinical CVD and Heine-Broring et al. in the Rotterdam Study [17] of 1570 asymptomatic participants aged >55 years reported no significant association of dietary intake of LCn-3PUFAs with CAC as measured by EBCT. Similarly, Shang et al. in the Melbourne Collaborative Cohort Study of 312 asymptomatic participants aged 45–64 years reported no significant association between dietary intake of LCn-3PUFAs and aortic calcification as measured by lateral thoracolumbar radiography and dual-energy X-ray absorptiometry [18]. Several plausible explanations for the contrasting findings between these studies [16-18] and ours may include differences in the age distribution, subclinical atherosclerosis assessment techniques, vascular bed examined, and the use of blood biomarkers of LCn-3PUFAs in our study as opposed to self-reported dietary assessment of fatty acids intake- which may lead to LCn-3PUFAs misclassification [2].
In our study, the association of total LCn-3PUFAs, EPA, and DHA with aortic calcification was partly attenuated after adjusting for intermediary variables suggesting that the relationship was partly mediated through the effects of LCn-3PUFAs on blood pressure, lipids, and inflammation. Atherosclerosis is a systemic chronic inflammatory disease of the vessel walls. Inflammation resulting from the interaction of modified atherogenic lipoproteins, inflammatory cells, and smooth muscle cells of vessel wall plays a major role in the initiation and progression of atherosclerotic plaque. Available evidence from epidemiological and experimental studies suggest that LCn-3PUFAs exerts their antiatherosclerotic effect through several anti-inflammatory pathways including lowering expression of nuclear factor к-B, regulators of inflammation, and oxidative stress, improving endothelial function, and increasing atherosclerotic plaque stability [3, 9-11].
Our study shows a significant inverse association of DHA but not of EPA with aortic calcification. This finding concords with our previous observation of a significant inverse association of DHA but not of EPA with CIMT [19] as well as the results of a prospective cohort study among postmenopausal women with CHD where DHA but not EPA was significantly associated with less progression of coronary atherosclerosis [33]. Additionally, indirect evidence from short-term clinical trials in humans reported that DHA compared to EPA is more potent in lowering blood pressure [34, 35], resting heart rate [36], the expression of pro-inflammatory cytokines, cell adhesion molecules, and monocyte adhesion to endothelial cells [37].
To date, no RCT has assessed the effect of pure DHA or compared the effect of pure DHA with EPA on cardiovascular outcomes or atherosclerosis. However, RCTs of pure EPA conducted among diabetics [8], patients with stable angina [7], and CHD [6] reported an inverse association of EPA with atherosclerosis. The RCTs in diabetics [8] and CHD patients [6] reported an increase in serum EPA but not in DHA (although EPA theoretically can be metabolized to DHA) among intervention groups supporting an antiatherogenic effect for EPA. The significant inverse association of DHA but not of EPA with aortic calcification in our study may imply that DHA may be more antiatherogenic than EPA. RCTs are warranted to disentangle the differential association of EPA and DHA with cardiovascular outcomes and atherosclerosis.
In our study, in US-White total LCn-3PUFAs and DHA were significantly and inversely associated with aortic calcification but not in US-Black, Japanese in Japan or Japanese American. This finding may suggest that US-White may respond differently to LCn-3PUFAs compared to other races/ethnicities. However, the differential association of LCn-3PUFAs by races/ethnicities with aortic calcification in this study may be a random finding due to cross-sectional study design and small sample size.
Our study also shows a significant inverse association of ARA with aortic calcification. This finding concords with findings reported in a meta-analysis of prospective observational studies and RCTs [38]. Although ARA-derived eicosanoids such as leukotriene-B4 and thromboxane-A2, are thought to be pro-inflammatory, other ARA-derived eicosanoids such as epoxyeicosatrienoic acid and lipoxins are considered to lower inflammation [39]. The anti-inflammatory mechanism of ARA is also supported by animal studies and observational studies in humans showing an inverse association of ARA metabolites with cardiovascular risk [2].
Our study has several limitations. First, blood levels of LCn-3PUFAs reflect short-term intake and may not reflect long-term dietary intake. However, blood levels of LCn-3PUFAs vary randomly, therefore the actual association between blood levels of LCn-3PUFAs and aortic calcification may be stronger than reported in the current study. Second, we examined healthy men aged 40-49 years in Japan and the US; therefore, the results of the study cannot be generalized to females, other populations, or age groups. Third, although we have controlled for a variety of sociodemographic and clinical characteristics, the possibility of residual confounding cannot be excluded. Fourth, we cannot establish a causal association between blood levels of LCn-3PUFAs and aortic calcification based on cross-sectional analysis. Strengths of the current study include: (i) the community-based nature of the study design with randomly selected study participants increasing the external validity of the study findings; (ii) Standardized measurement techniques across all study-centers; (iii) The use of EBCT to detect aortic calcification, allowing the accurate visualization of small calcific deposits without image blurring; (iv) and the use of blood biomarkers of LCn-3PUFAs-reducing the possibility of recall bias, as opposed to self-reported dietary assessment of fatty acids.
Our study findings have a public health significance. All over the world the commercialization of the use of fish oil or LCn-3PUFAs is rapidly expanding. This growing enthusiasm needs to be supported by the robust scientific data on intake of LCn-3PUFAs. Evidence concerning LCn-3PUFAs and atherosclerosis is limited in the general population. Evidence generated from this study adds to the evidence on the anti-atherosclerotic property of LCn-3PUFAs especially DHA in healthy middle-aged men. The findings of this study, if replicated in larger and longer follow-up studies, would help support intake of LCn-3PUFAs policy in the general population.
5. Conclusions
Our study demonstrated for the first time that in a general male population, LCn-3PUFAs are significantly inversely associated with subclinical atherosclerosis, defined by aortic calcification, independent of conventional cardiovascular risk factors. This significant inverse association was mainly attributed to DHA. Follow-up population-based studies are needed to further clarify the effect of LCn-3PUFAs on the incidence and progression of atherosclerosis as well as to disentangle the differential effect of EPA and DHA, and on the underlying biological mechanisms. Inverse association of LCn-3PUFAs with aortic calcification for US-White emphasizes the importance of studying the effects of LCn-3PUFAs in a variety of ethnic and cultural groups.
Supplementary Material
Table 3.
Association between LCn-3PUFAs and aortic calcification by races/ethnicities: Tobit ratio/odds ratio (95% confidence interval) for AoCaS per 1-SD of total LCn-3PUFAs/EPA/DHA
| Races/ Ethnicities |
Regression Technique |
Regression Models |
Polyunsaturated Fatty Acids | ||
|---|---|---|---|---|---|
| Total LCn- 3PUFAs |
EPA | DHA | |||
| US White (n= 300) | Tobit Regression | Model I | 0.34 (0.18, 0.65) | 0.42 (0.19, 0.93) | 0.40 (0.23, 0.71) |
| Model II | 0.48 (0.25, 0.92) | 0.57 (0.26, 1.26) | 0.54 (0.31, 0.94) | ||
| Model III | 0.43 (0.22, 0.86)b | 0.60 (0.27, 1.34) | 0.47 (0.26, 0.85) | ||
| Ordinal Logistic | Model I | 0.48 (0.31, 0.74) | 0.56 (0.33, 0.94) | 0.54 (0.37, 0.78) | |
| Model II | 0.58 (0.36, 0.92) | 0.66 (0.37, 1.16) | 0.63 (0.43, 0.93) | ||
| Model III | 0.60 (0.35, 0.94)c | 0.70 (0.40, 1.28) | 0.61 (0.40, 0.94) | ||
| US Black (n= 101) | Tobit Regression | Model I | 0.27 (0.07, 1.06) | 0.34 (0.06, 1.85) | 0.31 (0.10, 0.95) |
| Model II | 0.42 (0.11, 1.56) | 0.56 (0.10, 3.09) | 0.45 (0.16, 1.28) | ||
| Model III | 0.35 (0.09, 1.29) | 0.44 (0.08, 2.50) | 0.41 (0.14, 1.20) | ||
| Ordinal Logistic | Model I | 0.52 (0.22, 1.19) | 0.56 (0.21, 1.52) | 0.54 (0.26, 1.11) | |
| Model II | 0.62 (0.24, 1.60) | 0.75 (0.22, 2.52) | 0.62 (0.29, 1.33) | ||
| Model III | 0.54 (0.20, 1.45) | 0.63 (0.17, 2.29) | 0.59 (0.26, 1.32) | ||
| Japanese in Japan (n= 310) | Tobit Regression | Model I | 0.88 (0.41, 1.90) | 1.10 (0.60, 2.01) | 0.73 (0.31, 1.71) |
| Model II | 0.82 (0.40, 1.68) | 0.94 (0.53, 1.65) | 0.75 (0.34, 1.64) | ||
| Model III | 0.95 (0.46, 1.96) | 1.08 (0.61, 1.92) | 0.86 (0.38, 1.91) | ||
| Ordinal Logistic | Model I | 0.96 (0.75, 1.23) | 1.02 (0.83, 1.24) | 0.93 (0.70, 1.22) | |
| Model II | 0.96 (0.72, 1.27) | 0.98 (0.79, 1.23) | 0.94 (0.69, 1.28) | ||
| Model III | 1.02 (0.75, 1.34) | 1.04 (0.83, 1.31) | 0.98 (0.71, 1.34) | ||
| Japanese American (n= 287) | Tobit Regression | Model I | 0.54 (0.27, 1.03) | 1.25 (0.67, 2.32) | 0.47 (0.25, 0.90) |
| Model II | 0.72 (0.38, 1.35) | 0.92 (0.51, 1.64) | 0.66 (0.35, 1.24) | ||
| Model III | 0.71 (0.38, 1.33) | 0.85 (0.48, 1.51) | 0.68 (0.36, 1.27) | ||
| Ordinal Logistic Regression | Model I | 0.73 (0.51, 1.03) | 0.90 (0.65, 1.23) | 0.69 (0.49, 0.96) | |
| Model II | 0.83 (0.58, 1.19) | 0.94 (0.67, 1.30) | 0.81 (0.57, 1.15) | ||
| Model III | 0.83 (0.58, 1.19) | 0.90 (0.65, 1.26) | 0.82 (0.57, 1.18) | ||
AoCaS, aortic calcification score; IQR, interquartile range; TR, Tobit ratio; OR, Odds ratio; CI: confidence interval; SD: standard deviation; EPA, eicosapentaenoic acid; DPA, docosapentaenoic acid; DHA, docosahexaenoic acid; Total LCn-3PUFAs, total long chain omega 3 polyunsaturated fatty acids defined as the sum of EPA, DPA, and DHA;
For US-White, a 1-SD of total LCn-3PUFAs, EPA, and DHA equals to 1.7%, 0.5%, and 1.2%, respectively;
For US-Black, a 1-SD of total LCn-3PUFAs, EPA, and DHA equals to 1.5%, 0.5%, 1.1%, respectively;
For Japanese in Japan, a 1-SD of total LCn-3PUFAs, EPA, and DHA equals to 3%, 1.4%, and 1.7%, respectively;
For Japanese American, a 1-SD of total LCn-3PUFAs, EPA, and DHA equals to 2.2%, 0.9%, and 1.4%, respectively;
Model I: Fatty acids (Total LCn-3PUFAs or EPA or DHA), age, and years of education;
Model II: Model I + pack-years of smoking, alcohol consumption, BMI, diabetes, lipid-lowering medications, LDL-C, physical activity at the job, and meat intake;
Model III: Model II + hypertension, HDL-C, triglycerides, CRP, and fibrinogen;
For ordinal logistic regression, four AoCaS categories used were 0, 1-99, 100-299, and ≥300;
Tobit ratio of 0.43 means that in US-White for every SD (1.7%) increase in total LCn-3PUFAs, there is a 57% decrease in the AoCaS;
OR of 0.88 means that a 1-SD (1.7%) increase in total LCn-3PUFAs made a US-White participant in this study 40% less likely to be in a higher AoCaS category;
Financial support:
The work was supported by grants HL068200 and HL071561 from the National Institutes of Health, USA (Bethesda, Maryland, USA), B 16790335 and A 13307016, 17209023, and 21249043 from the Japanese Ministry of Education, Culture, Sports, Science and Technology (Tokyo, Japan)
A list of abbreviations
- ALA
α-linolenic acid
- AoCaS
Agatston aortic calcification score
- ARA
arachidonic acid
- BMI
body mass index
- CAC
coronary artery calcification
- CI
confidence interval
- CIMT
carotid intima-media thickness
- CHD
coronary heart disease
- CRP
C-reactive protein
- CVD
cardiovascular disease
- DHA
docosahexaenoic acid
- DPA
docosapentaenoic acid
- EBCT
Electron beam computed tomography
- EPA
eicosapentaenoic acid
- HDL-C
high-density lipoprotein cholesterol
- LA
linoleic acid
- LCn-3PUFAs
long-chain n-3 polyunsaturated fatty acids
- LDL-C
low-density lipoprotein cholesterol
- RCT
randomized controlled trial
- SD
standard deviation
- SEM
standard error of mean
- TR
tobit ratio
Footnotes
Conflict of interest: None
References
- [1].Alexander DD, Miller PE, Van Elswyk ME, Kuratko CN, Bylsma LC. A Meta-Analysis of Randomized Controlled Trials and Prospective Cohort Studies of Eicosapentaenoic and Docosahexaenoic Long-Chain Omega-3 Fatty Acids and Coronary Heart Disease Risk. Mayo Clinic Proceedings: Elsevier; 2017. p. 15-29. [DOI] [PubMed] [Google Scholar]
- [2].Mozaffarian D, Wu JH. Omega-3 fatty acids and cardiovascular disease: effects on risk factors, molecular pathways, and clinical events. Journal of the American College of Cardiology. 2011;58:2047–67. [DOI] [PubMed] [Google Scholar]
- [3].Mori TA. Dietary n-3 PUFA and CVD: a review of the evidence. Proceedings of the Nutrition Society. 2014;73:57–64. [DOI] [PubMed] [Google Scholar]
- [4].Mozaffarian D, Rimm E. Fish intake, contaminants, and human health: evaluating the risks and the benefits part 2-health risks and optimal intakes. Cardiol Rounds. 2006;10:1–6. [DOI] [PubMed] [Google Scholar]
- [5].Siscovick DS, Barringer TA, Fretts AM, Wu JH, Lichtenstein AH, Costello RB, et al. Omega-3 Polyunsaturated Fatty Acid (Fish Oil) Supplementation and the Prevention of Clinical Cardiovascular Disease: A Science Advisory From the American Heart Association. Circulation. 2017:CIR. 0000000000000482. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [6].Watanabe T, Ando K, Daidoji H, Otaki Y, Sugawara S, Matsui M, et al. A randomized controlled trial of eicosapentaenoic acid in patients with coronary heart diseae on statins. Journal of cardiology. Accepted in Aug 2017. [DOI] [PubMed] [Google Scholar]
- [7].Niki T, Wakatsuki T, Yamaguchi K, Taketani Y, Oeduka H, Kusunose K, et al. Effects of the Addition of Eicosapentaenoic Acid to Strong Statin Therapy on Inflammatory Cytokines and Coronary Plaque Components Assessed by Integrated Backscatter Intravascular Ultrasound. Circulation Journal. 2016;80:450–60. [DOI] [PubMed] [Google Scholar]
- [8].Mita T, Watada H, Ogihara T, Nomiyama T, Ogawa O, Kinoshita J, et al. Eicosapentaenoic acid reduces the progression of carotid intima-media thickness in patients with type 2 diabetes. Atherosclerosis. 2007;191:162–7. [DOI] [PubMed] [Google Scholar]
- [9].Calder PC. The role of marine omega-3 (n-3) fatty acids in inflammatory processes, atherosclerosis and plaque stability. Molecular Nutrition & Food Research. 2012;56:1073–80. [DOI] [PubMed] [Google Scholar]
- [10].Cawood AL, Ding R, Napper FL, Young RH, Williams JA, Ward MJ, et al. Eicosapentaenoic acid (EPA) from highly concentrated n– 3 fatty acid ethyl esters is incorporated into advanced atherosclerotic plaques and higher plaque EPA is associated with decreased plaque inflammation and increased stability. Atherosclerosis. 2010;212:252–9. [DOI] [PubMed] [Google Scholar]
- [11].Thies F, Garry JM, Yaqoob P, Rerkasem K, Williams J, Shearman CP, et al. Association of n-3 polyunsaturated fatty acids with stability of atherosclerotic plaques: a randomised controlled trial. The Lancet. 2003;361:477–85. [DOI] [PubMed] [Google Scholar]
- [12].Burgess N, Reynolds T, Williams N, Pathy A, Smith S. Evaluation of four animal models of intrarenal calcium deposition and assessment of the influence of dietary supplementation with essential fatty acids on calcification. Urological Research. 1995;23:239–42. [DOI] [PubMed] [Google Scholar]
- [13].Schlemmer C, Coetzer H, Claassen N, Kruger M, Rademeyer C, Van Jaarsveld L, et al. Ectopic calcification of rat aortas and kidneys is reduced with n-3 fatty acid supplementation. Prostaglandins, Leukotrienes and Essential Fatty Acids. 1998;59:221–7. [DOI] [PubMed] [Google Scholar]
- [14].Abedin M, Lim J, Tang T, Park D, Demer L, Tintut Y. N-3 fatty acids inhibit vascular calcification via the p38-mitogen-activated protein kinase and peroxisome proliferator-activated receptor-γ pathways. Circulation Research. 2006;98:727–9. [DOI] [PubMed] [Google Scholar]
- [15].Massaro M, Scoditti E, Carluccio M, Campana M, De Caterina R. Omega-3 fatty acids, inflammation and angiogenesis: basic mechanisms behind the cardioprotective effects of fish and fish oils. Cellular and Molecular Biology (Noisy-le-Grand, France). 2009;56:59–82. [PubMed] [Google Scholar]
- [16].He K, Liu K, Daviglus ML, Mayer-Davis E, Jenny NS, Jiang R, et al. Intakes of long-chain n–3 polyunsaturated fatty acids and fish in relation to measurements of subclinical atherosclerosis. Am J Clin Nutr. 2008;88:1111–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [17].Heine-Bröring RC, Brouwer IA, Proença RV, van Rooij FJ, Hofman A, Oudkerk M, et al. Intake of fish and marine n– 3 fatty acids in relation to coronary calcification: the Rotterdam Study. Am J Clin Nutr. 2010;91:1317–23. [DOI] [PubMed] [Google Scholar]
- [18].Shang X, Sanders KM, Scott D, Khan B, Hodge A, Khan N, et al. Dietary α-linolenic acid and total ω-3 fatty acids are inversely associated with abdominal aortic calcification in older women, but not in older men. The Journal of Nutrition. 2015:jn211789. [DOI] [PubMed] [Google Scholar]
- [19].Sekikawa A, Kadowaki T, El-Saed A, Okamura T, Sutton-Tyrrell K, Nakamura Y, et al. Differential association of docosahexaenoic and eicosapentaenoic acids with carotid intima-media thickness. Stroke. 2011;42:2538–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [20].Lewis JR, Schousboe JT, Lim WH, Wong G, Zhu K, Lim EM, et al. Abdominal aortic calcification identified on lateral spine images from bone densitometers are a marker of generalized atherosclerosis in elderly women. Arteriosclerosis, Thrombosis, and Vascular Biology. 2016;36:166–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [21].Gonçalves FB, Voûte MT, Hoeks SE, Chonchol MB, Boersma EE, Stolker RJ, et al. Calcification of the abdominal aorta as an independent predictor of cardiovascular events: a meta-analysis. Heart. 2012;98:988–94. [DOI] [PubMed] [Google Scholar]
- [22].Iribarren C, Sidney S, Sternfeld B, Browner WS. Calcification of the aortic arch: risk factors and association with coronary heart disease, stroke, and peripheral vascular disease. JAMA. 2000;283:2810–5. [DOI] [PubMed] [Google Scholar]
- [23].Criqui MH, Denenberg JO, McClelland RL, Allison MA, Ix JH, Guerci A, et al. Abdominal aortic calcium, coronary artery calcium, and cardiovascular morbidity and mortality in the Multi-Ethnic Study of Atherosclerosis. Arteriosclerosis, Thrombosis, and Vascular Biology. 2014;34:1574–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [24].Criqui MH, Kamineni A, Allison MA, Ix JH, Carr JJ, Cushman M, et al. Risk Factor Differences for Aortic Versus Coronary Calcified Atherosclerosis The Multiethnic Study of Atherosclerosis. Arteriosclerosis, Thrombosis, and Vascular Biology. 2010;30:2289–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [25].Kuller LH, Matthews KA, Sutton-Tyrrell K, Edmundowicz D, Bunker CH. Coronary and Aortic Calcification Among Women 8 Years After Menopause and Their Premenopausal Risk Factors The Healthy Women Study. Arteriosclerosis, Thrombosis, and Vascular Biology. 1999;19:2189–98. [DOI] [PubMed] [Google Scholar]
- [26].Hoffmann U, Massaro JM, D'Agostino RB, Kathiresan S, Fox CS, O'Donnell CJ. Cardiovascular Event Prediction and Risk Reclassification by Coronary, Aortic, and Valvular Calcification in the Framingham Heart Study. Journal of the American Heart Association. 2016;5:e003144. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [27].Sekikawa A, Curb JD, Ueshima H, El-Saed A, Kadowaki T, Abbott RD, et al. Marine-derived n-3 fatty acids and atherosclerosis in Japanese, Japanese-American, and white men: a cross-sectional study. Journal of the American College of Cardiology. 2008;52:417–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [28].Fujiyoshi A, Sekikawa A, Shin C, Masaki K, Curb JD, Ohkubo T, et al. A cross-sectional association of obesity with coronary calcium among Japanese, Koreans, Japanese Americans, and US whites. European Heart Journal-Cardiovascular Imaging. 2013:jet080. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [29].Choo J, Ueshima H, Curb JD, Shin C, Evans RW, El-Saed A, et al. Serum n– 6 fatty acids and lipoprotein subclasses in middle-aged men: the population-based cross-sectional ERA-JUMP Study. Am J Clin Nutr. 2010;91:1195–203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [30].Sekikawa A, Ueshima H, Kadowaki T, El-Saed A, Okamura T, Takamiya T, et al. Less subclinical atherosclerosis in Japanese men in Japan than in white men in the United States in the post–World War II birth cohort. American Journal of Epidemiology. 2007;165:617–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [31].Sekikawa A, Ueshima H, Zaky WR, Kadowaki T, Edmundowicz D, Okamura T, et al. Much lower prevalence of coronary calcium detected by electron-beam computed tomography among men aged 40–49 in Japan than in the US, despite a less favorable profile of major risk factors. International Journal of Epidemiology. 2005;34:173–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [32].Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M, Detrano R. Quantification of coronary artery calcium using ultrafast computed tomography. Journal of the American College of Cardiology. 1990;15:827–32. [DOI] [PubMed] [Google Scholar]
- [33].Erkkila AT, Matthan NR, Herrington DM, Lichtenstein AH. Higher plasma docosahexaenoic acid is associated with reduced progression of coronary atherosclerosis in women with CAD. J Lipid Res. 2006;47:2814–9. [DOI] [PubMed] [Google Scholar]
- [34].Mori TA, Woodman RJ. The independent effects of eicosapentaenoic acid and docosahexaenoic acid on cardiovascular risk factors in humans. Current Opinion in Clinical Nutrition & Metabolic Care. 2006;9:95–104. [DOI] [PubMed] [Google Scholar]
- [35].Morris MC, Sacks F, Rosner B. Does fish oil lower blood pressure? A meta-analysis of controlled trials. Circulation. 1993;88:523–33. [DOI] [PubMed] [Google Scholar]
- [36].Mori TA, Bao DQ, Burke V, Puddey IB, Beilin LJ. Docosahexaenoic Acid but Not Eicosapentaenoic Acid Lowers Ambulatory Blood Pressure and Heart Rate in Humans. Hypertension. 1999;34:253. [DOI] [PubMed] [Google Scholar]
- [37].De Caterina R, Liao JK, Libby P. Fatty acid modulation of endothelial activation. Am J Clin Nutr. 2000;71:213S–23S. [DOI] [PubMed] [Google Scholar]
- [38].Chowdhury R, Warnakula S, Kunutsor S, Crowe F, Ward HA, Johnson L, et al. Association of dietary, circulating, and supplement fatty acids with coronary riska systematic review and meta-analysis. Annals of Internal Medicine. 2014;160:398–406. [DOI] [PubMed] [Google Scholar]
- [39].Serhan CN. Novel lipid mediators and resolution mechanisms in acute inflammation: to resolve or not? The American Journal of Pathology. 2010;177:1576–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.