Differential Association of Docosahexaenoic and Eicosapentaenoic Acids With Carotid Intima-Media Thickness
Background and Purpose—Recent studies reported the differential effect of docosahexaenoic (DHA) and eicosapentaenoic acids (EPA). We examined the differential association of DHA and EPA with carotid intima-media thickness (IMT) in Japanese individuals in Japan and in U.S. white individuals and explored whether DHA or EPA contributes to the difference in IMT between the two groups.
Methods—A population-based cross-sectional study in 608 Japanese and U.S. white men aged 40 to 49 was conducted to assess IMT, serum DHA, EPA, and other cardiovascular risk factors.
Results—Japanese compared to U.S. whites had significantly lower IMT (mean±SD, 618±81 and 672±94 μm for Japanese and whites, respectively; P<0.001) and had >2-fold higher levels of DHA and EPA. DHA, but not EPA, had an inverse association with IMT in both Japanese and U.S. whites. The inverse association remained only in Japanese men after adjusting for risk and other factors. The significant difference in multivariable-adjusted IMT became nonsignificant after further adjusting for DHA (mean difference, 17 μm; 95% CI, −8 to 43; P=0.177) but not EPA. In this multivariable-adjusted model, DHA but not EPA was a significant predictor of IMT (P=−0.032 versus 0.863, respectively).
Conclusions—These data suggest that DHA may have a more potent antiatherogenic effect than EPA, especially in levels observed in the Japanese, independent of risk factors.
Marine-derived n-3 fatty acids (FA) have a cardioprotective effect. Both epidemiological and clinical trial studies have shown that marine-derived n-3 FA reduce the risk of coronary heart disease (CHD) death.1 The reduction in CHD death by marine-derived n-3 FA is generally attributed to their antiarrhythmic effect.2 Marine-derived n-3 FA are also reported to have an antiatherogenic effect.3 Results from epidemiological and clinical trial studies in Japan,4,5 where dietary intake of marine-derived n-3 FA is high, ie, 1000 mg/d compared to 100 mg/d in a typical Western diet,6 support a hypothesis that marine-derived n-3 FA have the antiatherogenic effect. To further support the hypothesis, we recently reported that Japanese men had a significant inverse association of serum levels of marine-derived n-3 FA with intima-media thickness (IMT) of the carotid artery, an independent predictor of cardiovascular events.7
Recent studies reported the differential effects of eicosapentaenoic (EPA; 20:5 n-3) and docosahexaenoic acids (DHA; 22:6 n-3), 2 major marine-derived n-3 FA, on cardiovascular risk factors.8–10 Although EPA theoretically can be metabolized to DHA, dietary intake or supplementation of EPA does not increase serum DHA.11 Two previous studies reported the differential association of DHA and EPA with IMT, giving conflicting results.12,13
In the current study, we examined the differential association of EPA and DHA with IMT both in the Japanese individuals in Japan and U.S. white individuals. Additionally, we examined whether EPA or DHA contributes to the difference in IMT between the Japanese and U.S. whites. We examined these questions by electron beam tomography risk factor assessment among Japanese and U.S. men in the post-World War II birth cohort (ERA JUMP Study), a population-based cross-sectional study of men aged 40 to 49 years in Japan and whites in the United States.
Subjects and Methods
Detailed descriptions of methods were published elsewhere.14 Participants were population-based samples of 623 randomly selected men aged 40 to 49 years examined in 2002 to 2006, without clinical cardiovascular disease: 313 Japanese men from Kusatsu, Shiga, Japan, and 310 U.S. men from Allegheny County, Pennsylvania, United States. We excluded 15 subjects with missing data, resulting in 310 Japanese and 298 U.S. whites. Informed consent was obtained from all participants. The study was approved by the Institutional Review Boards of Shiga University of Medical Science, Otsu, Japan, and University of Pittsburgh, Pittsburgh, Pennsylvania, United States.
Venipuncture was performed early during the clinic visit after a 12-hour fast. The samples were stored at −80°C and shipped on dry ice to the University of Pittsburgh. Serum lipids were determined using standardized methods of the Centers for Disease Control and Prevention, serum glucose using an enzymatic assay, serum insulin using a radioimmuno assay, and C-reactive protein using an immunosorbent assay. Serum DHA, EPA, and other FA were determined by capillary gas liquid chromatography, as previously described.14 The coefficients of variation between tests for EPA and DHA were 4.5% and 7.2%, respectively.
A self-administered questionnaire was used to obtain information on demography, frequency of fish intake, and other factors as previously described.14 Pack-years of smoking were calculated as years of smoking multiplied by the number of cigarettes per day divided by 20. Those who exercised were defined as those who regularly exercised ≥1 hour per week. Hypertension was defined as systolic blood pressure (BP) ≥140 mm Hg, diastolic BP ≥90 mm Hg, or use of antihypertensive medications. Diabetes mellitus was defined as fasting serum glucose level ≥7 mmol/L or use of antidiabetic medications.
Carotid Artery Scanning
The scanning procedures were previously described.14 A Toshiba 140A scanner equipped with a 7.5-MHz linear-array imaging probe was used for carotid scanning at both centers. The sonographers scanned the right and left common carotid arteries, the carotid bulbs, and the internal carotid arteries. The scans were recorded and sent for scoring to the Ultrasound Research Laboratory, University of Pittsburgh. Trained readers digitized the best image for scoring. Our current study used common carotid arteries.14 The readers were blinded to participant characteristics and the study centers. Correlation coefficients of IMT between sonographers and between readers were 0.96 and 0.99, respectively.
To compare risk factors between the populations, a t test or the Mann-Whitney U test for continuous variables or χ2 test for categorical variables was used. To examine associations of EPA, DHA, or each of other FA with IMT, we made tertile groups of each FA for each of the Japanese and U.S. whites. Then, we compared age-adjusted and multivariable-adjusted tertile-specific levels of IMT. For the multivariable adjustment, we first adjusted for BP, high-density lipoprotein cholesterol, and triglycerides, which have significant associations with marine n-3 FA.3 Then, we further adjusted for low-density lipoprotein cholesterol, pack-years of smoking, glucose, insulin, and body mass index, which are risk factors for IMT. Finally, we further adjusted for C-reactive protein and other factors. To examine the linear trend of tertile-specific levels of IMT and whether EPA or DHA was associated with lower IMT in the Japanese than U.S. whites, general linear model analyses were used. All P values were 2-tailed. P<0.05 was considered significant. PSAW Statistics (release 18.0; IBM) software was used for all statistical analyses.
The Japanese as compared to U.S. whites had a less favorable or similar profile of many major cardiovascular risk factors (Table 1), including BP, low-density lipoprotein cholesterol, smoking, and diabetes. Meanwhile the Japanese had a more favorable profile of some risk factors, including body mass index, high-density lipoprotein cholesterol, and C-reactive protein. More than 40% of the Japanese ate fish 4 times per week or more as opposed to 3% of U.S. whites.
The Japanese had markedly higher serum levels of both EPA and DHA (Figure 1). The lower 5th percentiles of EPA and DHA in the Japanese were still higher than the higher 75th percentiles in U.S. whites (0.90% versus 0.89%, respectively, for EPA and 3.34% versus 3.03%, respectively for DHA). The Spearman correlation of EPA and DHA was 0.349 (P<0.001) for the Japanese and 0.139 (P=0.017) for U.S. whites. The Supplemental Table (http://stroke.ahajournals.org) shows means and standard deviations of other FA.
In the age-adjusted analysis, DHA had an inverse association with IMT both in the Japanese and U.S. whites (Figure 2). The inverse association was marginally significant in the Japanese (P=0.060) and significant (P=0.007) in U.S. whites. In contrast, EPA had no significant association with IMT in either group. Among other FA, α-linolenic and transfatty acids had significant and marginally significant positive associations with IMT, respectively, only in U.S. whites (Figure 3). Linoleic acid had a nonsignificant inverse association with IMT in U.S. whites. Other serum FA shown in the Supplemental Table did not have significant inverse associations with IMT in either the Japanese or the U.S. whites (data not shown).
After adjusting for BP, high-density lipoprotein cholesterol, and triglycerides, DHA in the Japanese but not in U.S. whites had a significant inverse association with IMT (model I in Table 2). The significant inverse association remained after further adjusting for other factors (models II and III in Table 2). In contrast, EPA had no significant association with IMT in either group (Table 2).
The significant difference in IMT between the Japanese and U.S. whites remained after further adjusting for traditional and other risk factors (models I–III in Table 3). However, further adjusting for DHA but not EPA made the significant difference in multivariable-adjusted IMT nonsignificant. It is noted that DHA itself was a significant predictor of IMT in this model (P=0.032). The effect of DHA was independent of EPA (Table 3) or other FA (data not shown).
This study shows that DHA but not EPA in the Japanese had an inverse association with IMT and that the inverse association was significant after adjusting for various risk and other factors. This study also shows that DHA contributed to the lower IMT in the Japanese than in U.S. whites. Our results suggest that DHA may have an antiatherogenic effect.
Our finding that DHA but not EPA has a significant inverse association with IMT is consistent with the result from a previous study among subjects with primary hyperlipidemia in Spain,12 where fish intake is high compared to other western countries. This study reported a significant inverse association of DHA but not EPA with IMT after adjusting for risk and other factors. Although a population-based study of 487 Swedish men reported a significant inverse correlation of EPA but not DHA with IMT,13 the association does not remain significant after adjusting for covariates.
Results from a clinical trial and a recent review on the association of tissue n-3 FA with nonfatal CHD support the hypothesis that DHA has the antiatherogenic effect. The Estrogen Replacement and Atherosclerosis Trial reported that DHA but not EPA is associated with reduced progression of coronary atherosclerosis.15 The recent review found that tissue DHA had a significant inverse association with nonfatal CHD.
Results from the Japan Eicosapentaenoic Acid Lipid Interventions Study (JELIS) show EPA reduces nonfatal CHD. JELIS shows 900 mg of highly purified EPA per day is effective in reducing the risk of nonfatal CHD.5 Dietary intake or supplementation of EPA increases serum EPA but not DHA,11 although JELIS did not report the changes in serum DHA during the trial. Thus, the reduction of nonfatal CHD in JELIS is likely to be mediated through EPA, not DHA.
Therefore, our results may suggest that DHA has a more potent antiatherogenic effect than EPA. Despite extensive research on the effects of marine-derived n-3 FA on cardiovascular risk factors, only a small number of studies examined the differential effects of DHA and EPA.16 DHA but not EPA reduces platelet aggregatory responses ex vivo and platelet-derived thromboxane B2.9 DHA but not EPA improves dilator and constrictor responses in the forearm microcirculation.8 In vitro studies show that DHA but not EPA decreases the expression of proinflammatory cytokines10 and cell adhesion molecules,17 which contribute to the development of atheroslerosis.18 DHA and EPA are equally effective in reducing urinary F2-isoprostane.19
We observed higher prevalence of diabetes in the Japanese than in U.S. whites, although it was not statistically significant. This is in accordance with the fact that the Japanese are more susceptible to the development of diabetes than whites.20 We observed significantly lower rates of lipid-lowering medications in the Japanese than in the U.S. whites. This is partly because of the difference in clinical practice between the 2 countries.21
Although the significant difference in multivariable-adjusted IMT between the Japanese and U.S. whites became nonsignificant after further adjusting for DHA, our intent was not to explain the difference in IMT, but rather to demonstrate that DHA substantially attenuated the difference. Many factors besides traditional risk factors account for the difference in IMT, including genetics, age-ethnicity interaction, and cohort effect. However, among these factors, cohort effect is unlikely to account for the difference. This is because in this birth cohort, levels of serum total cholesterol and BP have been similar throughout their lifetime, and prevalence of diabetes is similarly high between the 2 groups.14
Several limitations of the study warrant discussion. Serum EPA and DHA reflect short-term dietary intake and may not reflect long-term dietary intake. However, in populations in which fish intake is high and stable like in the Japanese in Japan, a single measurement of serum EPA and DHA reflects the ranking of habitual intake of EPA and DHA, respectively.22 Additionally, because the variation in serum DHA occurs randomly, the actual association of DHA with carotid IMT would be stronger than was observed in our study. Our cross-sectional study does not allow us to infer whether a lifelong exposure to high DHA is necessary to observe its effect. The sample size is relatively small. Our study included men and only those aged 40 to 49 years, and the results may not be generalizable to women or other age groups. The study is observational and we cannot exclude the possibility of residual or unmeasured confounding.
Our results suggest that DHA may have a more potent antiatherogenic effect than EPA, especially in levels observed in the Japanese. Further research of the differential effects of DHA and EPA is needed.
Sources of Funding
This research was supported by HL68200 from the National Institutes of Health and B 16790335 and A 13307016, 17209023, and 21249043 from the Japanese Ministry of Education, Culture, Sports, Science, and Technology.
The online-only Data Supplement is available at http://stroke.ahajournals.org/cgi/content/full/STROKEAHA.110.613042/DC1.
- Received January 4, 2011.
- Accepted February 10, 2011.
- © 2011 American Heart Association, Inc.
- Wang C,
- Harris WS,
- Chung M,
- Lichtenstein AH,
- Balk EM,
- Kupelnick B,
- et al.
- Balk E,
- Chung M,
- Lichtenstein A,
- Chew P,
- Kupelnick B,
- Lawrence A,
- et al
- Iso H,
- Kobayashi M,
- Ishihara J,
- Sasaki S,
- Okada K,
- Kita Y,
- et al.
- Lorenz MW,
- von Kegler S,
- Steinmetz H,
- Markus HS,
- Sitzer M
- Mori TA,
- Watts GF,
- Burke V,
- Hilme E,
- Puddey IB,
- Beilin LJ
- Arterburn LM,
- Hall EB,
- Oken H
- Sala-Vila A,
- Cofan M,
- Perez-Heras A,
- Nunez I,
- Gilabert R,
- Junyent M,
- et al.
- Erkkila AT,
- Matthan NR,
- Herrington DM,
- Lichtenstein AH
- De Caterina R,
- Liao JK,
- Libby P