Abstract. The blood eicosapentaenoic and docosahexaenoic
acid (EPA+DHA) concentration is an important inverse risk
factor for sudden cardiac death. However, it is not known what
kinds of factors influence the EPA+DHA levels in the total
phospholipid fraction in red blood cells (RBC EPA+DHA) in
Japan, who regularly eat more fish with increasing age. Four
hundred and fifty-six healthy individuals (320 men and 136
women, 18 to 70 years old) were recruited between 2002 and
2005. RBC EPA+DHA were measured by gas chromatography
and questionnaires were administered. Multivariate analysis
indicated that there were significant correlations between RBC
EPA+DHA and (i) dietary EPA+DHA (‚=0.31), (ii) age
(‚=0.33), (iii) gender (‚=–0.15) and (iv) physical activity
(‚=–0.11) but not with body mass index or smoking.
The percentage of eicosapentaenoic and docosahexaenoic
acids (EPA+DHA) in the blood is a useful biomarker of
long-chain n-3 fatty acid intake. It was termed the Omega-
3 Index, and might be an important negative risk factor for
coronary heart disease (CHD) deaths (1). This index was
actually shown to be significantly correlated with human
myocardial n-3 fatty acid content (2), and strongly
associated with reduced risk of sudden cardiac deaths (3).
Recently Sands et al. (4) determined the Omega-3 Index
in red blood cells (RBCs) and other basic characteristics
from 163 adults, and found that four factors significantly
and independently influenced the Omega-3 Index in
RBCs: fish servings, age, body mass index (BMI) and
diabetes. In the present study the relationship between the
EPA+DHA levels in the total phospholipid fraction in
RBCs (RBC EPA+DHA) and several factors including
physical activity was investigated in 456 Japanese subjects.
Considering that Japanese people regularly eat fish and
that the rate of CHD death is only one fourteenth of all-
cause mortality in Japan (5), it might also be interesting to
compare RBC EPA+DHA in Japan and the Omega-3
Index of the United States.
Materials and Methods
We performed four intervention studies in the Tokyo, Shizuoka
and Toyama areas between 2002 and 2005. Intervention included
changes in lipid nutrition. The fatty acid composition of the total
phospholipid fraction in RBCs was measured at both start and end
of each intervention study. In the present study, the data obtained
at the start of studies (while study participants were still under no
effects of intervention) were compiled for statistical calculation.
Each intervention trial was approved by the ethics committee of
Toyama Medical and Pharmaceutical University (the former name
before its union with University of Toyama) and written informed
consent was obtained from each participant. Those who were under
treatment of diabetes mellitus, hypertension and hyperlipidemia
were not eligible for any of the four studies described above. Those
who were taking supplements were also excluded.
Food intake including alcoholic beverages was calculated with a
semi-quantitative food-frequency questionnaire and the food
calculation software, Eiyokun 3.0 (Kenpakusha Co. Ltd., Tokyo).
To assess nutrients, study participants were asked how often on
average they had eaten a portion size of each food during the
previous 4 weeks. The daily physical activity and smoking status
(smoker, non-smoker, ex-smoker) were also determined with a
simple questionnaire. With regard to the daily physical activity,
participants were asked to choose one of four activity levels,
namely low, relatively low, moderate or high. For each activity
level, there was a simple explanation with approximate hours spent
for standing, walking, rapid walking and strong muscle activity.
Fasting EDTA-anticoagulated blood samples were collected
early in the morning and RBCs were separated from the samples.
They were washed twice with saline and frozen at –80ÆC until
analysis. The fatty acid composition of the total phospholipid
fraction of washed RBCs was determined as described elsewhere
Correspondence to: Tomohito Hamazaki, Division of Clinical
Application, Department of Clinical Sciences, Institute of Natural
Medicine, University of Toyama, 2630 Sugitani, Toyama-city,
Toyama 9300194, Japan. e-mail: email@example.com
Key Words: Age, fish intake, gender, Omega-3 Index, physical
in vivo 22: 131-136 (2008)
Factors Influencing EPA+DHA Levels
in Red Blood Cells in Japan
MIHO ITOMURA1, SHUNTARO FUJIOKA1, KEI HAMAZAKI1, KOUJI KOBAYASHI1,
TETSURO NAGASAWA1, SHIGEKI SAWAZAKI2, YUKO KIRIHARA1and TOMOHITO HAMAZAKI1
1Section of Clinical Application, Department of Clinical Sciences,
Institute of Natural Medicine, University of Toyama, Toyama 930-0194;
2Hida City Hospital, 725 Higashi-machi, Kamioka, Hida-City, Gifu 506-1111, Japan
and expressed as area % (6). Briefly, total lipids were extracted
with chloroform/methanol; the total phospholipid fraction was
separated by thin-layer chromatography; after transmethylation
with HCl-methanol, the fatty acid composition was analyzed by gas
chromatography (GC14A Shimadzu Corporation, Kyoto) with a
capillary column DB-225 (0.25 mm i.d., 30 m length, 0.25 Ìm; J&W
Scientific, Folsom, CA, USA).
Data are expressed as means±SD. Univariate regression and
multivariate analyses were performed to assess the associations
between RBC EPA+DHA and characteristics of the participants.
In the case of multivariate analysis, the following items were
included into the calculation as independent predictors: age,
gender, BMI, smoking status, physical activity and dietary
EPA+DHA. The mean dietary EPA+DHA levels in age groups in
both sexes were compared with two-way ANOVA. SPSS,
version13.0 (SPSS 13.0J Base System, 2005, SPSS Japan Inc.,
Tokyo) was used for calculation. P<0.05 was considered as
From the four intervention studies above, 456 complete
data sets were collected. The study participants constituted
320 Japanese men and 136 women. The age distribution was
as follows: age 18-29, 47; age 30-39, 142; age 40-49, 126; age
50-59, 126 and age 60-70, 15 with an average age of
42.5±10.6 years. Other baseline characteristics of the study
participants were as follows: BMI, 22.5±3.3 (range: 16-38),
smokers (men: 51%, women: 24%), and physical activity
(low: 27%, relatively low: 41%, moderate: 29%, and high:
3%). The average values of RBC EPA, DHA and
EPA+DHA were 1.6±0.7% (range: 0.4-4.4%), 6.8±1.3%
(range: 3.5-10.4%) and 8.5±1.8% (range 3.9-12.9%),
respectively. There was only one participant whose RBC
EPA+DHA was below 4%.
Univariate and multivariate analyses of RBC EPA+DHA
and independent variables are shown in Table I. Dietary
EPA+DHA expressed as a weight % of the total fatty acid
intake (dietary EPA+DHA %), age, gender, BMI and
smoking status were significantly correlated with RBC
EPA+DHA on univariate analysis, but physical activity was
not correlated with RBC EPA+DHA. However, on
multivariate analysis, BMI and smoking ceased to be
significant, while physical activity became significant.
Figure 1 shows the relationship between RBC
EPA+DHA and dietary EPA+DHA %. Figure 2 shows the
relationship between RBC EPA+DHA and age in both
sexes. Figure 3 shows the relationship between dietary
EPA+DHA % and age in both sexes.
The Omega-3 Index is strongly associated with reduced risk
of sudden cardiac death (3). Sands et al. (4) proposed that
an Omega-3 Index in RBCs of 8% or above was the target
in vivo 22: 131-136 (2008)
Figure 1. The relationship between RBC EPA+DHA and the dietary ratios
of EPA+DHA to the total fatty acid intake. r=0.40, p<0.0001, n=456.
Table I. Relation between RBC EPA+DHA and participants' characteristics.
Univariate analysisMultivariate analysis
95% CI95% CI
Independent variabler Lower Upper
p-valuestandardized ‚ Lower Upper
Dietary EPA+DHA %
Results from 456 healthy participants. RBC EPA+DHA=the area percentage of EPA and DHA in the total phospholipid fraction in RBCs.
Dietary EPA+DHA =the weight percentage of the total fatty acid intake.
cardioprotective level. The average RBC EPA+DHA in the
present study participants were above that value, actually
8.5%. In contrast, the average value in Sands et al.'s study
was only 4.9% (4). They also proposed that the high risk
horizon of the Omega-3 Index in RBCs was 4%.
Interestingly, there was only one participant with a value
below 4% in the present study. This low incidence is
probably one of the major reasons why the Japanese do not
frequently die from acute myocardial infarction.
In the present study, we analyzed the total phospholipid
fraction of RBCs. This fraction was different from that most
of the other workers used for analysis. Taking into account
that about 95% of all fatty acids in RBCs are located in the
total phospholipid fraction (not published data), it might be
reasonably safe to compare the differences in EPA+DHA
levels between the total lipid (not total phospholipid)
fraction in RBCs measured by others and our RBC
EPA+DHA, especially where there was a sizable difference
in the fatty acid composition between two study populations.
Nevertheless, we should realize that there are a number of
different methods to measure fatty acids and to describe the
fatty acid composition, and that all these differences make a
direct comparison difficult, especially if the difference in the
fatty acid composition was not great.
It has generally been thought that fish consumption
increases with age in Japan and some data support this
belief (7). As shown in Figure 3, our data concur with such
a belief. Interestingly, as a factor, age survived as a
significant predictor of RBC EPA+DHA even after
adjustment for n-3 fatty acid intake. This finding was similar
to what Sands et al. found in their study (4) and also to an
animal experiment of Gudbyarnason (8) who showed that
DHA contents in the myocardium of old rats were higher
than those of young rats, although both groups of rats ate
the same diet.
BMI did not remain a significant predictor of RBC
EPA+DHA on multivariate analysis (Table I). This was
different from the results of Sands et al. (4). There was a big
difference in average BMIs between the two study
populations (22.5±3.3 in the present study vs. 26.2±4.8 (4)).
The relationship between BMI and EPA+DHA might be
different in a country where the BMI is rather high such as
in the United States. Another reason may be the difference
in the unit of EPA+DHA intake. In their study (4), n-3
fatty acid intake was registered as the frequency of fish
servings, which was not adjusted to participants' body
weights or their total fat (or energy) intake, whereas the
EPA+DHA intake in the present study was adjusted for the
Itomura et al: Factors Influencing RBC EPA+DHA in Japan
Figure 2. The relationship between RBC EPA+DHA and age in both genders on a univariate analysis. Men, closed columns; women, open columns. Bars
indicate one SD.
total fatty acid intake. Consequently, our data of
EPA+DHA intake were less influenced by obesity.
In the present study, physical activity was inversely related
to RBC EPA+DHA (Table I). The mechanism of this
relationship is not immediately clear, but there is a possibility
that EPA and DHA might be preferentially catabolized for
energy production when physical activity is increased. EPA
and DHA are two of the least viscous fatty acids (9) and are
probably not suitable for storage in the adipose tissue. In fact,
the percentages of EPA and DHA in the human adipose
tissue did not exceed 1% even after more than one year of
daily administration of 10 g fish oil (10).
In previous studies on a smaller scale (11-13), n-3 fatty
acid levels tended to be rather higher in women than in
men. The same trend was seen in the report of Sands et al.
before adjustment for some confounding factors, but this
was reversed after adjustment (4). Our study showed that
RBC EPA+DHA was higher in men than in women both
before and after adjustment for confounding factors (see
Table I). For women who eat enough EPA and DHA, such
as those women in our study, a greater capacity to produce
EPA and DHA from ·-linolenic acid in women (14) does
not seem to influence RBC EPA+DHA.
Hibbeln et al. reported that smoking status predicted
lower levels of RBC n-3 fatty acids in schizophrenic patients
(13). However, the effect of smoking was not confirmed by
Sands et al. (4). In our study with essentially normal
participants, a positive association of smoking with RBC
EPA+DHA was found before adjustment, but this
association disappeared after adjustment (Table I).
In conclusion, multivariate analysis indicated that there
were significant correlations between RBC EPA+DHA and
(i) dietary EPA+DHA, (ii) age, (iii) gender and (iv)
physical activity, but not with BMI or smoking in Japan.
We thank Dr. Yoshihiro Terashima, Ms Shizuko Takebe, Mr.
Hiroto Nishizawa and Ms. Hiroko Hamatani for their technical
assistance. This work was partly supported by Polyene Project Ltd.,
Nissui Co. Ltd., Yashima Co. Ltd., and Mr. Kanji Nakajima.
1 Harris WS and Von Schacky C: The Omega-3 Index: a new risk
factor for death from coronary heart disease? Prev Med 39:
Harris WS, Sands SA, Windsor SL, Ali HA, Stevens TL,
Magalski A, Porter CB and Borkon AM: Omega-3 fatty acids
in cardiac biopsies from heart transplantation patients:
correlation with erythrocytes and response to supplementation.
Circulation 110: 1645-1649, 2004.
Albert CM, Campos H, Stampfer MJ, Ridker PM, Manson JE,
Willett WC and Ma J: Blood levels of long-chain n-3 fatty
acids and the risk of sudden death. N Engl J Med 346: 1113-
Sands SA, Reid KJ, Windsor SL and Harris WS: The impact of
age, body mass index, and fish intake on the EPA and DHA
content of human erythrocytes. Lipids 40: 343-347, 2005.
in vivo 22: 131-136 (2008)
Figure 3. The relationship between the mean dietary EPA+DHA and age in both genders. Men, closed columns; women, open columns. There was a
significant difference in the mean dietary EPA+DHA % among age groups (p<0.0001, two-way ANOVA).
5 Health and Welfare Statistics Association: Koseino shihyo. J Download full-text
Health Welfare Stat 51(Suppl.): S45-S57, 2004 (in Japanese).
Hamazaki K, Itomura M, Huan M, Nishizawa H, Sawazaki S,
Tanouchi M, Watanabe S, Hamazaki T, Terasawa K and
Yazawa K: Effect of ˆ-3 fatty acid-containing phospholipid on
blood catecholamine concentrations in healthy volunteers: a
randomized, placebo-controlled, double-blind trial. Nutrition
21: 705-710, 2005.
Iso H, Date C, Noda H, Yoshimura T, Tamakoshi A; JACC
Study Group: Frequency of food intake and estimated nutrient
intake among men and women: the JACC Study. J Epidemiol
15(Suppl. 1): S 24-S42, 2005.
Gudbyarnason S: Dynamics of n-3 and n-6 fatty acids in
phospholipids of heart muscle. J Intern Med 225 (Suppl. 1):
Hamazaki T, Kobayashi S, Urakaze M, Yano S and Fujita T:
Viscosities of some triglycerides and ethylester of fatty acids
frequently found in cell membranes – a possible effect of
viscosity of fatty acids in phospholipids on hemorheology.
Biorheology 22: 221-226, 1985.
10 Leaf DA, Connor WE, Barstad L and Sexton G: Incorporation
of dietary n-3 fatty acids into the fatty acids of human adipose
tissue and plasma lipid classes. Am J Clin Nutr 62: 68-73, 1995.
11 Kuriki K, Nagaya T, Tokudome Y, Imaeda N, Fujiwara N, Sato
J, Goto C, Ikeda M, Maki S, Tajima K and Tokudome S:
Plasma concentrations of (n-3) highly unsaturated fatty acids
are good biomarkers of relative dietary fatty acid intakes: a
cross-sectional study. J Nutr 133: 3643-3650, 2003.
12 Giltay EJ, Gooren LJ, Toorians AW, Katan MB and Zock PL:
Docosahexaenoic acid concentrations are higher in women than
in men because of estrogenic effects. Am J Clin Nutr 80: 1167-
13 Hibbeln JR, Makino KK, Martin CE, Dickerson F, Boronow J
and Fenton WS: Smoking, gender, and dietary influences on
erythrocyte essential fatty acid composition among patients with
schizophrenia or schizoaffective disorder. Biol Psychiatry 53:
14 Burdge GC and Wootton SA: Conversion of alpha-linolenic acid
to eicosapentaenoic, docosapentaenoic and docosahexaenoic
acids in young women. Br J Nutr 88: 411-420, 2002.
Received July 16, 2007
Revised November 6, 2007
Accepted December 3, 2007
Itomura et al: Factors Influencing RBC EPA+DHA in Japan