Atherosclerosis, 81 (1990) 199-208 Elsevier Scientific Publishers Ireland,
ATHERO
199 Ltd.
04447
Effects of safflower oil and evening primrose oil in men with a low dihomo-y-linolenic level Robert D. Abraham
*, Rudolph
A. Riemersma, Robert A. Elton, Cecilia Macintyre and Michael F. Oliver
Cardiovascular Research Unit and Medical Statistics Unit, University of Edinburgh, Edinburgh, Scotland (U.K.) (Received 12 July, 1989) (Revised, received 29 November, 1989) (Accepted 18 December, 1989)
Low levels of essential polyunsaturated fatty acids of the n - 6 series are associated with coronary heart disease. Linoleic acid, but not y-linolenic acid requires the activity of A6-desaturase for its conversion to dihomo-y-linolenic and arachidonic acid. Evening primrose oil (EPO) and safflower oil (SO) are rich in linoleic acid, but EPO contains also 9% y-linolenic acid. The effect of EPO (10, 20 and 30 ml/day) and SO (20 ml/day) for 4 months on the deposition of linoleic acid metabolites in adipose tissue of 4 groups of 6-9 men with low adipose dihomo-y-linolenic acid was examined. EPO but not SO increased adipose dihomo-y-linolenic acid level from 0.080 * 0.005% to 0.101 + 0.005% (P < 0.01; 20 ml/day for 4 months). Adipose dihomo-y-linolenic/linoleic acid ratio increased with EPO from 0.99 f 0.16 x lo* to 1.13 + 0.14 x lo2 and fell on SO from 1.04 + 0.10 X lo* to 0.90 + 0.07 x lo2 (P < 0.01). Similar qualitative changes in the relative amount of dihomo-y-linolenic acid in serum triglyceride and cholesteryl ester fractions were observed. At the dose of 20 ml/day, SO and EPO did not differ in their effect on serum cholesterol (7.13 + 0.43 vs. 7.33 f 0.42 mmol/l (NS)), LDL-cholesterol (5.10 + 0.32 vs. 4.88 5 0.46 mmol/l (NS)) nor did the 2 oils differ in their effect on HDL-cholesterol. These results suggest that linoleic acid is not readily converted to dihomo-y-linolenic acid due to a low activity of A6-desaturase in these highly selected men. EPO was not an effective hypocholesterolaemic agent in this study.
Key words:
Evening primrose oil; Efamol; lipids; A6-Desaturase; Dietary
* Present address: Department
Dihomo-y-linolenic acid; Linoleic acid; Adipose supplement; Essential fatty acid; Men
Introduction of Cardiology,_~ Blacktown
Hospital, Blacktown, NSW, Australia. Correspondence to: Rudolph A. Riemersma, Cardiovascular Research Unit Department of Medicine (RIE), Hugh Robson Building, George Square, Edinburgh EH8 9XF, Scotland, U.K.
0021-9150/90/$03.50
tissue; Serum
0 1990 Elsevier Scientific
Publishers
Ireland,
There is now substantial evidence to support the view that low levels of the essential fatty acid linoleic acid, reflecting long term dietary intake of Ltd.
200 this fatty acid, predisposes to coronary heart disease (CHD) [l]. Adipose tissue linoleate from apparently healthy men reflects existing gradients in CHD mortality in Europe, with low levels in the high CHD mortality areas and high levels in the low CHD areas [2,3]. These gradients cannot be explained by the confounding influences of regional differences in serum cholesterol or other classical risk factors for CHD. Low plasma phospholipid linoleate has also been linked with the development of CHD in a prospective, Finnish study [4]. In Scotland, men with occult CHD also have lower adipose linoleate [5] and the increased risk of angina or first myocardial infarction is also independent of classical risk factors [6] and is of the same order of magnitude as that of serum cholesterol [7]. The essentiality of linoleic acid depends on its conversion to other highly unsaturated, long-chain fatty acids. An important and rate limiting step in this pathway is the desaturation of linoleic acid to y-linolenic acid by A6-desaturase [8]. Other enzymes, chain elongase and A5-desaturase then convert y-linolenic readily to the biologically active fatty acids dihomo-y-linolenic and arachidonic acid. Interestingly, low levels of adipose tissue dihomo-y-linolenic acid also reflect regional gradients in CHD mortality [3] and were more strongly TABLE
associated with occult CHD than its precursor linoleic acid [5] suggesting that a ‘block’ at the level of A6-desaturase might be more important. This prompted us to select apparently healthy volunteers with a low level of dihomo-y-linolenic acid and to study the fatty acid composition of adipose tissue and of serum lipids after supplementation with evening primrose oil, which because of its y-linolenic acid content does not depend on the activity of A6-desaturase for the conversion to dihomo-y-linolenic and arachidonic acid. Safflower oil which does not contain y-linolenic acid, but is otherwise similar to evening primrose oil, was used as a control. Methods Study population The study population was drawn from a random sample of apparently healthy Edinburgh men aged 35-54 years who had undergone adipose tissue sampling as part of a previous study in 1983-1984 [6]. With permission of the general practitioners, 66 men representing 90.4% of the lowest quintile of the adipose tissue dihomo-ylinolenic acid distribution, were invited. Thirty-five (48.0%) men were enrolled into the study. The men were randomized into 4 groups to receive 10
1
CLINICAL EVENING
CHARACTERISTICS OF THE SUBJECTS, PRIMROSE OIL (EPO)
TAKING
20 ml/day
SAFFLOWER
OIL (SO) OR 10, 20 OR 30 ml/day
Mean (SEM). so 20(n=9)
EPO
Quetelet’ index (kg/m’)
5I(2) 33 6.89 (0.34) 26.1
52(2) 33 6.76 (0.53) 27.0
49(2) 50 6.78 (0.45) 22.9
49(2) 33 6.91 (0.23) 26.1
Adipose
linoleate
(1.3) 10.1
(0.9) 9.9
(4.3) 8.3
(0.8) 9.1
Adipose
DGLA
(0.8) 0.099 (0.006) 78 17
(0.4) 0.089 (0.004) 63 31
(0.3) 0.080 (0.022) 25 59
(0.3) 0.085 (0.005) 56 29
(5)
(5)
,ASe (yrs) Cigarette smokers (%) Serum cholesterol (mmol/l)
(X) (W)
Alcohol drinkers (%) Alcohol consumption of drinkers (g/day) DGLA
= dihomo-y-linolenic
acid.
lO(n=9)
20(n=8)
(18)
30 (n = 9)
(8)
201 evening (n = 9), 20 (n = 8) or 30 (n = 9) ml/day primrose oil (Efamol, Scotia Pharmaceuticals Ltd., Guildford, U.K.) or 20 ml/day (n = 9) safflower oil. Of the 35 men enrolled into the study, 5 dropped out. One could not tolerate EPO, and another could not attend due to changing employment commitments. Three dropped out: 2 because of illness (diabetes, and unstable angina), and one unexplained sudden death. No differences in age, relative weight, smoking habit or alcohol consumption were observed between the 4 groups (Table 1). Informed, written consent was obtained from all subjects. Ethical approval for the study was obtained from the Ethics of Medical Research Sub-Committee (Medicine and Clinical Oncology) Lothian Health Board. Study design All were seen initially in August 1986. Subjects underwent clinical examination and lipoprotein determination. A second baseline examination and venesection were performed 1 month later and the subjects underwent the first of 3 subcutaneous adipose tissue biopsies under local anaesthesia [5]. Blood was also collected for fatty acid composition of serum lipids. The first month’s supply of oil, safflower oil (placebo) or evening primrose oil was dispensed at this visit. The fatty acid composition of the 2 oils is presented in Table 2. The oil was packaged under nitrogen in identical 20-ml glass bottles in a sealed box with written instructions on dose (l/2, 1 or 1 l/2 bottle/day, i.e. 10, 20 or 30 ml/day). Neither the investigators nor the subjects were aware of the nature of the oil. Subjects attended for monthly visits when they received new supplies of oil, their blood pressure
TABLE
2
FATTY ACID COMPOSITION (S) OF SAFFLOWER (SO) AND EVENING PRIMROSE OIL (EPO) Fatty acid
so
EPO
Palmitic Stearic Oleic Linoleic y-Linolenic
7 3 15 14 _
5 2 10 74 9
OIL
and weight were recorded, fasting blood was collected (lipoproteins) and they were interviewed concerning any adverse symptoms. A second subcutaneous adipose tissue and serum samples were obtained after 4 months for fatty acid analysis. Further fasting venous blood samples were obtained 1 and 2 months after completing the oil supplement period and the third subcutaneous adipose tissue and serum lipid samples were obtained for fatty acid composition at the second post-supplement visit. There was a mean 2.2 kg increase in body weight in the group receiving 30 ml evening primrose oil (P < 0.001) and small non-significant increases in the other groups. Laboratory procedures The laboratory was not aware of the subjects’ group allocation. Adipose tissue samples were stored at - 40 o C until the conclusion of the study when all samples were analysed by gas-liquid chromatography (GLC) together [5]. For this study. each sample was analysed in triplicate to improve the precision of quantitation of dihomo-y-linolenic acid present in low concentration. Serum lipids were extracted and purified by thin-layer chromatography for the GLC determination of fatty acid composition. For complex peaks, gas chromatography/ mass spectrography had shown that the methyl esters of acid eicosanoid acid (C20 : 0) and y-linolenic (Cl8 : 3 n - 6) coincide, as do those of eicosaenoic acid (C20 : 1 n - 9) and cu-linolenic acid (Cl8 : 3 n - 3). We did not try to determine the relative amounts of the constituents of these peaks. The mean coefficients of variation for the repeated analyses of each sample was 0.7% for linoleic acid (Cl8 : 2n - 6), 1.5% for dihomo-y-linolenic acid (C20 : 3n - 6), 0.8% for arachidonic acid (C20 : 4n - 6). Lipoprotein fractions were analysed on fresh serum by ultracentrifugation [9]. Cholesterol and triglyceride levels were determined enzymatically [lo]. The laboratory is standardised to the WHO Lipid Reference Laboratory (Prague). Statistical analysis Statistical comparisons of baseline data between multiple groups were made by one-way analysis of variance, and time related effects were
202 examined by two-way analysis of variance. Since the purpose of the study was to examine the effects of evening primrose oil, the main statistical analysis was to compare evening primrose oil (20
TABLE
ml/day) examine primrose compare
with safflower oil (20 ml/day) and to the effect of the dose levels of evening oil. Two sample t-tests were used to the effects of evening primrose oil (20
3
EFFECT OF SAFFLOWER OIL (SO) AND EVENING PRIMROSE OIL (EPO, 10, 20 and 30 ml/day) LONG-CHAIN POLYUNSATURATED FATTY ACIDS OF THE n - 6 FAMILY Mean percentage Oil
(SEM). The results of all other fatty acids are available n
Baseline
4 Months
ON ADIPOSE
on request.
Wash-out
Significance SO vs. EPO #
Fatty acid Linoleic acid 7 EPO 10 20 6 30 8 20 8 so y-Linolenic a 7 EPO 10 20 6 30 8 20 8 so Dihomo-y-linolenic 7 EPO 10 20 6 30 8 20 8 SO Arachidonic 7 EPO 10 20 6 30 8 20 8 SO
9.2 8.4 9.1 10.3
(0.4) (0.3) (0.3) (0.8)
9.7 9.2 10.1 11.5
(0.5) (0.3) (0.3) (1.0)
9.9 9.2 10.1 11.5
(0.5) (0.3) (0.3) (1.0)
to
EPO dose response *
Baseline to treatment
Treatment wash-out
NS
NS
NS
0.145 0.129 0.116 0.176
(0.021) (0.025) (0.010) (0.027)
0.138 0.130 0.133 0.151
(0.017) (0.019) (0.010) (0.026)
0.145 0.128 0.139 0.161
(0.022) (0.021) (0.009) (0.024)
NS
NS
< 0.05
0.090 0.080 0.085 0.101
(0.004) (0.005) (0.005) (0.006)
0.097 0.101 0.112 0.100
(0.005) (0.005) (0.007) (0.007)
0.098 0.098 0.112 0.100
(0.004) (0.006) (0.006) (0.007)
< 0.01
NS
NS
0.399 0.465 0.366 0.470
(0.012) (0.036) (0.011) (0.011)
0.404 0.464 0.386 0.457
(0.007) (0.022) (0.011) (0.015)
0.402 0.481 0.388 0.462
(0.008) (0.020) (0.013) (0.011)
NS
NS
NS
Ratios: X 10’ DGLA/Linoleic 7 1.00 (0.10) EPO 10 20 6 0.99 (0.16) 30 8 0.97 (0.13) 20 8 1.04 (0.10) SO Arachidonic/DGLA 7 4.62 (0.29) EPO 10 20 6 6.11 (0.71) 30 8 4.69 (0.62) 8 5.01 (0.49) SO 20 a Complex peak includes C20 : 0. * Two sample r-test comparing the change 20 ml SO vs. 20 ml EPO. * Regression analysis of dose response to treatment. Analysis of variance of all fatty acids within the safflower oil group and y-Iinolenic acid
1.03 1.13 1.14 0.90
(0.08) (0.14) (0.16) (0.07)
1.02 1.10 1.14 0.90
(0.08) (0.15) (0.14) (0.07)
< 0.01
NS
NS
4.19 4.74 3.80 4.99
(0.19) (0.49) (0.54) (0.54)
4.22 5.09 3.60 4.97
(0.28) (0.45) (0.28) (0.45)
< 0.01
NS
NS
in fatty acid pattern
from baseline
10, 20 and 30 ml/day
TISSUE
to treatment
EPO for change
and from treatment
in fatty
to wash-out
acid level from baseline
each group was highly significant (P < 0.01). except for DGLA values in the EPO 10 and 20 ml/day groups.
between
to 4 months
and AA/DGLA
ratio in
203 ml/day) and safflower oil (20 ml/day) on the change in adipose tissue and plasma lipid fatty acid composition from baseline to treatment (4 months) and during the wash-out period (from 4 to 6 months). The relationship between dose of evening primrose oil and the change in fatty acid composition was analysed by linear regression. A P value c 0.05 was considered significant. Results Adipose tissue fatty acid composition The effect of safflower and evening primrose oil on the relative amount of n - 6 fatty acids in adipose tissue is presented in Table 3. Dietary supplementation with safflower and evening primrose oil, both containing large amounts of linoleic acid, increased adipose tissue linoleate levels significantly. At the dose of 20 ml/day, there was no difference between the 2 oils in this respect. The small amounts of dihomo-y-linolenic acid in adipose tissue increased significantly in 2 of the 3 groups receiving evening primrose oil, but not in those on placebo (Fig. 1). The mean increase in dihomo-y-linolenic acid tended to rise with increasing doses of evening primrose oil. The dihomo-y-linolenic level did not fall significantly during the 2 month wash-out period. The adipose tissue concentrations of arachidonic acid tended to increase after 4 months supple-
0.M
A 203
n-6
0.W
(W 0.02
-0.01
_L
1
-0.02 !
10
20
30
ml/day
Fig. 1. Effect of evening primrose oil (EPO) on adipose tissue dihomo-y-linolenic acid levels. The increase in dihomo-y-linolenate after 20 and 30 ml/day EPO supplement is significant, but not at the lower dose of 10 ml/day. Safflower oil (SO), which also contains a high amount of linoleic acid but no y-linolenic acid, did not affect the dihomo-y-linolenic acid levels, The relation between dose of EPO and the rise in dihomo-y-linolenate is not significant.
mentation with evening primrose oil, particularly in the group receiving the highest dose (30 ml/ day). In this group the increased arachidonic acid levels remained after 2 months wash-out period. Arachidonic acid level tended to fall during safflower oil supplementation. There was no significant linear relationship between the dose of evening primrose oil and the percentage increase in adipose tissue arachidonic acid. The mean ratio of dihomo-y-linolenic/ linoleic acid increased significantly in those groups receiving evening primrose oil, but actually fell in those receiving safflower oil, the difference being highly significant (Table 3, P < 0.01). The ratio of arachidonic/ dihomo-y-linolenic acid fell significantly during evening primrose oil supplementation (P < O.Ol), but safflower oil did not affect this ratio. Serum lipid fatty acid composition The linoleic acid content of serum triglycerides tended to rise after both oils (Table 4). The fall during the wash-out period differed and was greater in the safflower oil group. There was no clear relation between the dose of evening primrose oil and the rise in triglyceride linoleate (Table 4). There was also an increase in y-linolenic acid levels during the supplementation period, which was related to the dose of evening primrose oil (Table 4). The levels returned to pre-supplementation values 2 months after stopping the supplement. The increase in dihomo-y-linolenic acid during supplementation with evening primrose oil (20 ml/day) was significantly higher than with safflower oil (20 ml/day), but the decline during the wash-out period was not significantly different between the two oils. The relatively high content of linoleic acid in cholesteryl ester fraction increased with both evening primrose and safflower oil. In contrast the rise in y-linolenyl, dihomo-y-linolenyl and arachidonyl cholesterol were confined to the evening primrose oil supplement but not in a clear, dose-dependent manner (Table 4). A similar pattern emerged for the changes in the relative amount of the n - 6 fatty acids (%) of serum phospholipids (Table 4). The rise in y-linolenic acid in this lipid fraction after evening primrose oil depended on the daily dose of the oil (P < 0.05). Evening primrose oil also reduced the total phospholipid concentration
204 TABLE 4 EFFECT OF SAFFLOWER OIL (SO; 20 ml/day) AND EVENING PRIMROSE OIL (EPO; 10, 20 and 30 ml/day) RELATIVE AMOUNT OF SERUM LIPID POLYUNSATURATED FATTY ACIDS OF THE n - 6 FAMILY
ON THE
Mean percentage (SEM). The results of all other fatty acids are available on request. Oil
n
Baseline
4 Months
Wash-out
Significance SO vs. EPO #
Triglycerides: fatty acid Linoleic 12.1 (1.7) EPO 10 9 10.6 (2.6) 20 6 13.0 (1.9) 30 8 12.8 (2.5) SO 20 8 y-Iinolenic a 0.155 (0.033) EPO 10 9 0.155 (0.022) 20 6 0.224 (0.059) 30 8 0.145 (0.022) so 20 8 Dihomo-y-linolenic 0.130 (0.012) EPO 10 9 0.131 (0.030) 20 6 30 8 0.186 (0.044) 0.160 (0.032) SO 20 8 Arachidonic EPO 10 9 0.706 (0.060) 0.770 (0.141) 20 6 0.924 (0.121) 30 8 0.728 (0.087) SO 20 8 Cholesterol esters Linoleic EPO 10 9 20 6 30 8 SO 20 8 y-Linolenic a EPO 10 9 20 6 30 8 so 20 8 Dihomo-y-Enolenic EPO 10 9 20 6 30 8 SO 20 8 Arachidonic EPO 10 9 20 6 30 8 SO 20 8
54.1 47.4 54.0 53.2
(3.4) (4.1) (1.2) (2.3)
16.8 (2.6) 14.4 (2.4) 17.7 (2.4) 18.7 (2.2)
11.6 (1.1) 11.6 (1.5) 13.5 (2.0) 13.4 (2.1)
EPO dose response *
Baseline to treatment
Treatment to wash-out
NS
< 0.05
NS
0.205 0.309 0.521 0.154
(0.030) (0.088) (0.122) (0.017)
0.152 0.157 0.222 0.132
(0.030) (0.034) (0.075) (0.022)
NS
NS
< 0.05
0.261 0.252 0.357 0.204
(0.053) (0.050) (0.048) (0.027)
0.133 0.126 0.186 0.151
(0.016) (0.018) (0.043) (0.023)
< 0.05
NS
NS
0.788 0.980 1.262 0.804
(0.078) (0.203) (0.163) (0.075)
0.744 0.846 0.944 0.820
(0.081) (0.075) (0.171) (0.069)
NS
NS
NS
NS
< 0.01
NS
58.9 51.1 55.6 60.4
(3.3) (2.1) (1.2) (1.8)
58.4 52.3 50.9 54.0
(3.5) (2.9) (2.1) (1.9)
0.51 0.68 0.54 0.64
(0.06) (0.07) (0.05) (0.08)
0.90 1.13 1.42 0.67
(0.13) (0.12) (0.18) (0.06)
0.54 0.47 0.84 0.55
(0.07) (0.05) (0.19) (0.09)
NS
-=z0.01
NS
0.48 0.41 0.62 0.56
(0.02) (0.05) (0.04) (0.04)
0.76 0.71 1.00 0.58
(0.11) (0.10) (0.14) (0.04)
0.52 0.49 0.60 0.57
(0.04) (0.03) (0.08) (0.05)
< 0.01
< 0.05
NS
4.24 4.84 5.11 4.77
(0.63) (0.74) (0.74) (0.25)
5.37 6.28 6.61 4.90
(0.91) (0.97) (0.69) (0.29)
5.04 5.97 6.10 5.48
(0.75) (0.77) (0.56) (0.29)
i 0.05
NS
NS
205 TABLE Oil
4 (continued) n
Baseline
4 Months
Wash-out
Significance SO vs. EPO *
Phospholipidy Linoleic EPO 10 9 20 6 30 8 so 20 8 y-Linolenic ’ EPO 10 9 20 6 30 8 so 20 8 Dihomo-y-linolenic EPO 10 9 20 6 30 8 so 20 8 Arachidonic EPO 10 9 20 6 30 8 SO 20 8
22.4 20.0 22.0 22.3
(1.6) (1.2) (1.0) (0.9)
0.075 0.056 0.034 0.038
-
(0.022) (0.009) (0.006) (0.010)
22.7 18.8 21.7 24.2
(1.9) (1.2) (1.2) (0.8)
0.050 0.053 0.082 0.034
21.6 20.4 19.8 21.5
(0.009) (0.012) (0.020) (0.010)
(1.6) (1.0) (1.1) (0.7)
0.048 0.029 0.049 0.029
(0.017) (0.005) (0.007) (0.010)
EPO dose response *
Baseline to treatment
Treatment wash-out
< 0.05
< 0.05
NS
NS
NS
< 0.05
to
2.73 2.73 2.74 2.82
(0.18) (0.33) (0.19) (0.24)
4.36 3.82 4.65 3.31
(0.29) (0.59) (0.70) (0.29)
2.79 2.75 3.12 3.14
(0.18) (0.23) (0.16) (0.33)
NS
NS
NS
7.54 8.69 7.95 8.14
(0.53) (0.95) (0.69) (0.26)
9.03 10.48 9.75 8.83
(0.78) (1.39) (0.56) (0.37)
8.45 10.00 9.70 9.05 -
(0.78) (0.82) (0.60) (0.40)
< 0.05
NS
NS
“ Complex peak including C20 : 0. * and #: see Table 3 for explanation.
expressed in mmol/l, but such an effect was not seen in cholesteryl ester or triglyceride concentration (not shown). The relative amount of a-linolenic and eicosapentaenoic acid (n - 3) in serum phospholipids decreased with evening primrose oil supplement. Similar results were also obtained for the cx-linolenic acid content in serum cholesteryl esters. The decrease was always related to the dose of evening primrose oil (not shown) and remained throughout the wash-out period. Serum lipids There were few effects of the 2 oils. Neither safflower oil, nor evening primrose oil had marked effects on serum lipids (Table 5). Serum triglycerides were stable throughout the study. The apparent higher fasting triglyceride concentration in those taking 20 ml/day evening primrose oil reflects the baseline values in this group. There
were no consistent and HDL-cholesterol
differences in total, VLDLbetween the 2 supplements.
Discussion Fatty acid composition Dietary supplementation with evening primrose oil increased the levels of dihomo-y-linolenic acid in adipose tissue. The men in this study were selected to have a low concentration of adipose tissue dihomo-y-linolenate, because it was expected that they were most likely to show an increase in this fatty acid in adipose tissue in response to dietary evening primrose oil supplement. In contrast, there was no change in adipose dihomo-y-linolenate concentration in the control group given safflower oil. This oil is equally rich in linoleic acid, but contains no y-linolenic acid. The conversion of y-linolenic acid to the longchain polyunsaturated n - 6 fatty acids (dihomoy-linolenic, arachidonic and arachidic acid)
206 TABLE
5
EFFECT OF SAFFLOWER OIL (SO; 20 ml/day) AND OF EVENING PRIMROSE OIL (EPO; IO, 20 and 30 ml/day) ON FASTING SERUM LIPIDS Mean percentage
(SEM).
so
EPO
20 (n=8) Cholesterol Total VLDL LDL HDL
Triglycerides Total
(mmol/l) 7.13 (0.43) 0.52 (0.12) 5.10 (0.32) 1.34 (0.09) (mmol/l) 1.61 (0.35)
;;=9)
;;,6)
;;=8)
6.89 (0.53) 0.51 (0.10) 4.95 (0.45) 1.42
6.38 (0.47) 0.28 (0.07) 4.58
(0.12)
7.33 (0.42) 0.83 (0.12) 4.88 (0.46) 1.27 (0.13)
1.52 (0.26)
2.27 * (0.26)
1.07 (0.18)
(0.44) 1.49 (0.13)
Note: No significant difference between the 2 oils by analysis of variance. Regression analysis of dose response to EPO also not significant for all lipid fractions. * This value is not significantly higher than those in the other groups and reflects the tendency to higher baseline values in this group.
bypasses the rate limiting A6-desaturase step. Therefore the rise in dihomo-y-linolenate is almost certainly due to the presence of y-linolenic and not of linoleic acid in evening primrose oil. Linoleic acid in large quantities inhibits its conversion to y-linolenic and dihomo-y-linolenic acid and the decrease in adipose dihomo-y-linolenic/ linoleic acid ratio during safflower oil supports this view. Thus we cannot exclude the possibility that the rise in adipose dihomo-y-linolenic acid during evening primrose oil was partly balanced by an opposing effect on account of its linoleic acid content. It would be of interest to examine the effect of concentrated y-linolenic acid. Adipose dihomo-y-linolenic acid levels remained elevated for at least 2 months after stopping the evening primrose oil supplement. Our data do not allow us to estimate the turnover of this essential fatty acid, but it is clear that it is low. Evening primrose oil also increased the relative amount of dihomo-y-linolenic acid in serum
cholesteryl esters and triglycerides, and of arachidonic acid in serum cholesteryl ester and phospholipid fractions. But, in contrast to adipose tissue fatty acid composition, that of serum lipids returned quickly to baseline after the oil supplement. It is worthy of mention that evening primrose oil specifically decreased the relative amount of a-linolenic acid (n - 3) and of eicosapentaenoic acid (n - 3) in phospholipids in a dose-dependent manner. A similar decrease was also observed for the cy-linolenic acid content of cholesteryl esters, but not for triglycerides. The mechanism of these changes is not clear, but is almost certainly not a simple competition between the n - 3 and n - 6 polyunsaturated fatty acids for incorporation into these serum lipids. This conclusion is deduced from the fact that the relative amount of these n - 3 but not n - 6 fatty acids persisted throughout the 2-month wash-out period. The activity of A6-desaturase activity was probably low in these men, since little or no conversion of linoleic acid from safflower oil to adipose dihomo-y-linolenic acid was observed. For obvious reasons, we could not measure the activity of this enzyme in liver microsomes in our study. Nevertheless, the possibility of low A6-desaturase activity should not be surprising, in view of our selection procedure. The question arises therefore whether low A6-desaturase activity is typical for all Edinburgh men. In animals, A6-desaturase activity is influenced by risk factors for coronary heart disease, such as ‘stress’, alcohol and cholesterol [S]. This possibility clearly requires further study in men. Another, perhaps even more pertinent question is whether low A6-desaturase activity leads to CHD [ll]. Support for this hypothesis comes from the observation that gradients in CHD mortality rates in Europe are inversely related to adipose dihomo-y-linolenic acid levels [3]. Patients with CHD have lower dihomo-y-linolenic acid levels than their controls [5], although this was not confirmed in another study [6]. The effect of evening primrose oil on adipose tissue arachidonate was not as consistent as that on dihomo-y-linolenate levels. However, there was a consistent and significant fall in the ratio of adipose arachidonic/ dihomo-y-linolenic acid. Thus, A5-desaturase may become the rate limiting step in the conversion of y-linolenic to arachidonic
207 acid under these conditions. The results of the effects of the 2 oils on serum cholesteryl ester and phospholipid arachidonic acid levels support this view. The effect of the diet on A’-desaturase activity is not certain. Data from rat studies suggest that a high saturated fat diet leads to higher As-desaturase activity than a diet rich in linoleate [12]. Taken together, this could imply that y-linolenic acid may specifically affect A5-desaturase activity. Lipids and lipoprotein fractions We were unable to demonstrate any consistent effect of either evening primrose oil or safflower oil on the fasting concentrations of total cholesterol and triglycerides, or on VLDL-, LDLand HDL-cholesterol subfractions. This was somewhat surprising, since both evening primrose and safflower oil contain rather large amounts of linoleic acid. According to the Hegsted [13] or Keys [14] formulae, we might have expected a reduction in total cholesterol of approximately 0.14 mmol/l due to the linoleate of evening primrose oil alone. The hypocholesterolaemic effect of y-linolenic acid is not well documented. Early data suggest an increased potency relative to linoleic acid of 2 [15], whilst Horrobin and Manku [16] believe it might be 170 times as potent, depending on initial serum cholesterol levels. Taking this higher value, we should have expected a fall in serum cholesterol of 2.46 mmol/l with 20 ml/day evening primrose oil. Thus our results are at variance with the Horrobin and Manku study [16] of patients mostly with atopic eczema, a condition in which A6-desaturase is known to be impaired [ll]. It is not clear whether this is responsible for their anomalous results. Our results suggest that the estimate of the hypocholesterolaemic effect of ylinolenic acid may well be substantially less. This view is supported by the results of other, smaller studies all showing slight reductions in serum cholesterol, ranging from - 0.18 to - 0.36 mmol/l, some significant others not [17-201. It would be very difficult to show such a small cholesterol reduction convincingly, as it is in the order of magnitude of the analytical error alone. The high-dose evening primrose oil (30 ml/day) increased the HDL/total cholesterol ratio (not shown). Other studies in man using smaller doses
of the oil do not confirm increased HDL levels, and this effect could be due to an increase in total fat intake [21]. In animal studies, the cholesterol lowering effects of evening primrose oil are clearer and are in full agreement with each other [22-241. Jnterestingly, these show that evening primrose oil and also safflower oil are not hypocholesterolaemic per se, but inhibit the rise in serum cholesterol induced by large quantities of dietary cholesterol. If these results can be extrapolated to men at all, they suggest that evening primrose oil may only be a hypocholesterolaemic agent, when raised serum cholesterol is caused by high levels of cholesterol in the diet. Acknowledgements This study was supported by a grant from Efamol Ltd. Robert D. Abraham was supported by a grant from the Blacktown Staff Specialist Trust Fund. We are grateful to Mary Walker, Karin Lyall and Margaret Millar for excellent technical support and to Christina Anderson for typing the manuscript. References Kingsbury, K.J., Brett, C., Stovold, R., Chapman, A., Ander, J. and Morgan, D.M., Abnormal fatty acid composition and human atherosclerosis. Postgrad. Med. J.. 50 (1974) 425. Logan, R.L.. Riemersma, R.A.. Thomson, M.. Oliver, M.F.. Olsson, A.G., Walldius, G., Rossner, S., Kaijser, L., Callmer, E., Carlson, L.A., Lockerbie, L. and Lutz, W., Risk factors for ischaemic heart disease in normal men aged 40. Edinburgh-Stockholm Study. Lancet, 1 (1978) 949. Riemersma, R.A., Wood, D.A., Butler, S., Elton, R.A., Oliver, M., Sale, M., Nikkari, T., Vartiainen. E., Puska, P., Gey, F., Rubba, P., Mancini, M. and Fidanza, F., Linolenic acid content in adipose tissue and coronary heart disease. Br. Med. J., 292 (1986) 1423. Miettinen, T.A., Naukkarinen, V., Huttunen, J.K., Mattill, S. and Kumlin, T., Fatty acid composition of serum lipids predicts myocardial infarction. Br. Med. J.. 282 (1982) 993. Wood, D.A., Butler, S., Riemersma, R.A.. Thomson, M., Oliver, M.F., Fulton, M., Birtwhistle, A. and Elton, R., Adipose tissue and platelet fatty acids and coronary heart disease in Scottish men. Lancet, 2 (1984) 117. Wood, D.A., Riemersma, R.A.. Butler, S., Thomson. M., Macintyre, C., Elton, R.A. and Oliver, M.F., Linoleic and eicosapentaenoic acids in adipose tissue and platelets and risk of coronary heart disease. Lancet, 1 (1987) 177.
208 7 The Pooling Research Group. Relationship of blood pressure, serum cholesterol, smoking habit, relative weight, and ECG abnormalities to incidence of major coronary events. Final report of the pooling project. J. Chron. Dis., 31 (1978) 201. 8 Brenner, R.R., Nutritional and hormonal factors influencing desaturation of essential fatty acids. Prog. Lipid Res., 20 (1982) 41. 9 Carlson, K.J., Lipoprotein fractionation. J. Clin. Pathol., 26 Suppl. 5 (1973) 32. 10 Notghi, A., Riemersma, R.A., Anderton, J.L. and Oliver, M.F., Effect of pindolol versus atenolol on lipid profile in hypertensive patients. Atherosclerosis, 77 (1989) 215. 11 Horrobin, D.F. and Huang, Y-S., The role of linoleic acid and its metabolites in the lowering of plasma cholesterol and the prevention of cardiovascular disease. Int. J. Cardiol., 17 (1987) 241. 12 Hagve, T-A. and Christophersen, B.O., Effect of dietary fats on arachidonic acid and eicosapentaenoic acid biosynthesis and conversion to C22 fatty acids in isolated rat liver cells. Biochim. Biophys. Acta, 796 (1984) 205. 13 Hegsted, D.M., McGandy, R.B., Myers, M.L. and Stare, F.J., Quantitative effects of dietary fat on serum cholesterol in man, Am. J. Clin. Nutr., 17 (1965) 281. 14 Keys, A., Anderson, J.T. and Grande, F., Prediction of the serum cholesterol responses of man to changes in fats in the diet. Lancet, 2 (1957) 959. 15 Worne, H.E. and Smith, L.W., Effects of certain pure long chain polyunsaturated fatty acid esters on blood lipids of man. Am. J. Med. Sci., 237 (1959) 710. 16 Horrobin, D.F. and Manku, M.S., How do polyunsaturated fatty acids lower plasma cholesterol levels? Lipids, 18 (1983) 558.
17 Blaton, V., Muls, E., Herman, A., Criel, A., Declerck, B. and Hollez, D., Effects of dietary alpha- and gamma-linolenate supplement on serum lipids, prostaglandins and platelet function in men. Clin. Chem., 32 (1986) 242 (Abstr.). 18 Ishikawa, T., Fujiyama, Y., Igarashi, 0. et al. Effects of gamma-linolenic acid on plasma lipoproteins and apolipoproteins. Atherosclerosis, 75 (1989) 95. 19 Boberg, M., Vessby, B. and Selinus, I., Effects of dietary supplementation with n - 6 and n - 3 long-chain polyunsaturated fatty acids on serum lipoproteins and platelet function in hypertriglyceridemic patients. Acta Med. Stand., 220 (1986) 153. 20 Szczeklik, A., Gryglewski, R.J., Kosta-Trabka, E. and Zmuda, A., Dihomo-gamma-linolenic acid in patients with atherosclerosis: effects on platelet aggregation, plasma lipids and low-density lipoprotein induced inhibition of prostacycline generation. Thromb. Haemostas., 51 (1984) 186. 21 Knuiman, J. and West, C.E., Differences in HDL cholesterol between populations: no paradox? Lancet. 1 (1983) 296. 22 Huang, Y.S., Manku, MS. and Horrobin, D.F., The effects of dietary cholesterol on blood and liver polyunsaturated fatty acids and on plasma cholesterol in rats fed various types of fatty acid diet. Lipids, 19 (1984) 664. 23 Sugano, M., Ishida, T., Yoshida, K., et al., Effect of mold oil containing gamma-linolenic acid on the blood cholesterol and eicosanoid levels in rats. Agric. Biol. Chem., 50 (1986) 2483. 24 Sugano. M., Ide, T., Ishida, T. and Yoshida, K., Hypocholesterolemic effect of gamma-linolenic acid as evening primrose oil in rats. Ann. Nutr. Metab., 30 (1986) 289.
ATHERO
199 Ltd.
04447
Effects of safflower oil and evening primrose oil in men with a low dihomo-y-linolenic level Robert D. Abraham
*, Rudolph
A. Riemersma, Robert A. Elton, Cecilia Macintyre and Michael F. Oliver
Cardiovascular Research Unit and Medical Statistics Unit, University of Edinburgh, Edinburgh, Scotland (U.K.) (Received 12 July, 1989) (Revised, received 29 November, 1989) (Accepted 18 December, 1989)
Low levels of essential polyunsaturated fatty acids of the n - 6 series are associated with coronary heart disease. Linoleic acid, but not y-linolenic acid requires the activity of A6-desaturase for its conversion to dihomo-y-linolenic and arachidonic acid. Evening primrose oil (EPO) and safflower oil (SO) are rich in linoleic acid, but EPO contains also 9% y-linolenic acid. The effect of EPO (10, 20 and 30 ml/day) and SO (20 ml/day) for 4 months on the deposition of linoleic acid metabolites in adipose tissue of 4 groups of 6-9 men with low adipose dihomo-y-linolenic acid was examined. EPO but not SO increased adipose dihomo-y-linolenic acid level from 0.080 * 0.005% to 0.101 + 0.005% (P < 0.01; 20 ml/day for 4 months). Adipose dihomo-y-linolenic/linoleic acid ratio increased with EPO from 0.99 f 0.16 x lo* to 1.13 + 0.14 x lo2 and fell on SO from 1.04 + 0.10 X lo* to 0.90 + 0.07 x lo2 (P < 0.01). Similar qualitative changes in the relative amount of dihomo-y-linolenic acid in serum triglyceride and cholesteryl ester fractions were observed. At the dose of 20 ml/day, SO and EPO did not differ in their effect on serum cholesterol (7.13 + 0.43 vs. 7.33 f 0.42 mmol/l (NS)), LDL-cholesterol (5.10 + 0.32 vs. 4.88 5 0.46 mmol/l (NS)) nor did the 2 oils differ in their effect on HDL-cholesterol. These results suggest that linoleic acid is not readily converted to dihomo-y-linolenic acid due to a low activity of A6-desaturase in these highly selected men. EPO was not an effective hypocholesterolaemic agent in this study.
Key words:
Evening primrose oil; Efamol; lipids; A6-Desaturase; Dietary
* Present address: Department
Dihomo-y-linolenic acid; Linoleic acid; Adipose supplement; Essential fatty acid; Men
Introduction of Cardiology,_~ Blacktown
Hospital, Blacktown, NSW, Australia. Correspondence to: Rudolph A. Riemersma, Cardiovascular Research Unit Department of Medicine (RIE), Hugh Robson Building, George Square, Edinburgh EH8 9XF, Scotland, U.K.
0021-9150/90/$03.50
tissue; Serum
0 1990 Elsevier Scientific
Publishers
Ireland,
There is now substantial evidence to support the view that low levels of the essential fatty acid linoleic acid, reflecting long term dietary intake of Ltd.
200 this fatty acid, predisposes to coronary heart disease (CHD) [l]. Adipose tissue linoleate from apparently healthy men reflects existing gradients in CHD mortality in Europe, with low levels in the high CHD mortality areas and high levels in the low CHD areas [2,3]. These gradients cannot be explained by the confounding influences of regional differences in serum cholesterol or other classical risk factors for CHD. Low plasma phospholipid linoleate has also been linked with the development of CHD in a prospective, Finnish study [4]. In Scotland, men with occult CHD also have lower adipose linoleate [5] and the increased risk of angina or first myocardial infarction is also independent of classical risk factors [6] and is of the same order of magnitude as that of serum cholesterol [7]. The essentiality of linoleic acid depends on its conversion to other highly unsaturated, long-chain fatty acids. An important and rate limiting step in this pathway is the desaturation of linoleic acid to y-linolenic acid by A6-desaturase [8]. Other enzymes, chain elongase and A5-desaturase then convert y-linolenic readily to the biologically active fatty acids dihomo-y-linolenic and arachidonic acid. Interestingly, low levels of adipose tissue dihomo-y-linolenic acid also reflect regional gradients in CHD mortality [3] and were more strongly TABLE
associated with occult CHD than its precursor linoleic acid [5] suggesting that a ‘block’ at the level of A6-desaturase might be more important. This prompted us to select apparently healthy volunteers with a low level of dihomo-y-linolenic acid and to study the fatty acid composition of adipose tissue and of serum lipids after supplementation with evening primrose oil, which because of its y-linolenic acid content does not depend on the activity of A6-desaturase for the conversion to dihomo-y-linolenic and arachidonic acid. Safflower oil which does not contain y-linolenic acid, but is otherwise similar to evening primrose oil, was used as a control. Methods Study population The study population was drawn from a random sample of apparently healthy Edinburgh men aged 35-54 years who had undergone adipose tissue sampling as part of a previous study in 1983-1984 [6]. With permission of the general practitioners, 66 men representing 90.4% of the lowest quintile of the adipose tissue dihomo-ylinolenic acid distribution, were invited. Thirty-five (48.0%) men were enrolled into the study. The men were randomized into 4 groups to receive 10
1
CLINICAL EVENING
CHARACTERISTICS OF THE SUBJECTS, PRIMROSE OIL (EPO)
TAKING
20 ml/day
SAFFLOWER
OIL (SO) OR 10, 20 OR 30 ml/day
Mean (SEM). so 20(n=9)
EPO
Quetelet’ index (kg/m’)
5I(2) 33 6.89 (0.34) 26.1
52(2) 33 6.76 (0.53) 27.0
49(2) 50 6.78 (0.45) 22.9
49(2) 33 6.91 (0.23) 26.1
Adipose
linoleate
(1.3) 10.1
(0.9) 9.9
(4.3) 8.3
(0.8) 9.1
Adipose
DGLA
(0.8) 0.099 (0.006) 78 17
(0.4) 0.089 (0.004) 63 31
(0.3) 0.080 (0.022) 25 59
(0.3) 0.085 (0.005) 56 29
(5)
(5)
,ASe (yrs) Cigarette smokers (%) Serum cholesterol (mmol/l)
(X) (W)
Alcohol drinkers (%) Alcohol consumption of drinkers (g/day) DGLA
= dihomo-y-linolenic
acid.
lO(n=9)
20(n=8)
(18)
30 (n = 9)
(8)
201 evening (n = 9), 20 (n = 8) or 30 (n = 9) ml/day primrose oil (Efamol, Scotia Pharmaceuticals Ltd., Guildford, U.K.) or 20 ml/day (n = 9) safflower oil. Of the 35 men enrolled into the study, 5 dropped out. One could not tolerate EPO, and another could not attend due to changing employment commitments. Three dropped out: 2 because of illness (diabetes, and unstable angina), and one unexplained sudden death. No differences in age, relative weight, smoking habit or alcohol consumption were observed between the 4 groups (Table 1). Informed, written consent was obtained from all subjects. Ethical approval for the study was obtained from the Ethics of Medical Research Sub-Committee (Medicine and Clinical Oncology) Lothian Health Board. Study design All were seen initially in August 1986. Subjects underwent clinical examination and lipoprotein determination. A second baseline examination and venesection were performed 1 month later and the subjects underwent the first of 3 subcutaneous adipose tissue biopsies under local anaesthesia [5]. Blood was also collected for fatty acid composition of serum lipids. The first month’s supply of oil, safflower oil (placebo) or evening primrose oil was dispensed at this visit. The fatty acid composition of the 2 oils is presented in Table 2. The oil was packaged under nitrogen in identical 20-ml glass bottles in a sealed box with written instructions on dose (l/2, 1 or 1 l/2 bottle/day, i.e. 10, 20 or 30 ml/day). Neither the investigators nor the subjects were aware of the nature of the oil. Subjects attended for monthly visits when they received new supplies of oil, their blood pressure
TABLE
2
FATTY ACID COMPOSITION (S) OF SAFFLOWER (SO) AND EVENING PRIMROSE OIL (EPO) Fatty acid
so
EPO
Palmitic Stearic Oleic Linoleic y-Linolenic
7 3 15 14 _
5 2 10 74 9
OIL
and weight were recorded, fasting blood was collected (lipoproteins) and they were interviewed concerning any adverse symptoms. A second subcutaneous adipose tissue and serum samples were obtained after 4 months for fatty acid analysis. Further fasting venous blood samples were obtained 1 and 2 months after completing the oil supplement period and the third subcutaneous adipose tissue and serum lipid samples were obtained for fatty acid composition at the second post-supplement visit. There was a mean 2.2 kg increase in body weight in the group receiving 30 ml evening primrose oil (P < 0.001) and small non-significant increases in the other groups. Laboratory procedures The laboratory was not aware of the subjects’ group allocation. Adipose tissue samples were stored at - 40 o C until the conclusion of the study when all samples were analysed by gas-liquid chromatography (GLC) together [5]. For this study. each sample was analysed in triplicate to improve the precision of quantitation of dihomo-y-linolenic acid present in low concentration. Serum lipids were extracted and purified by thin-layer chromatography for the GLC determination of fatty acid composition. For complex peaks, gas chromatography/ mass spectrography had shown that the methyl esters of acid eicosanoid acid (C20 : 0) and y-linolenic (Cl8 : 3 n - 6) coincide, as do those of eicosaenoic acid (C20 : 1 n - 9) and cu-linolenic acid (Cl8 : 3 n - 3). We did not try to determine the relative amounts of the constituents of these peaks. The mean coefficients of variation for the repeated analyses of each sample was 0.7% for linoleic acid (Cl8 : 2n - 6), 1.5% for dihomo-y-linolenic acid (C20 : 3n - 6), 0.8% for arachidonic acid (C20 : 4n - 6). Lipoprotein fractions were analysed on fresh serum by ultracentrifugation [9]. Cholesterol and triglyceride levels were determined enzymatically [lo]. The laboratory is standardised to the WHO Lipid Reference Laboratory (Prague). Statistical analysis Statistical comparisons of baseline data between multiple groups were made by one-way analysis of variance, and time related effects were
202 examined by two-way analysis of variance. Since the purpose of the study was to examine the effects of evening primrose oil, the main statistical analysis was to compare evening primrose oil (20
TABLE
ml/day) examine primrose compare
with safflower oil (20 ml/day) and to the effect of the dose levels of evening oil. Two sample t-tests were used to the effects of evening primrose oil (20
3
EFFECT OF SAFFLOWER OIL (SO) AND EVENING PRIMROSE OIL (EPO, 10, 20 and 30 ml/day) LONG-CHAIN POLYUNSATURATED FATTY ACIDS OF THE n - 6 FAMILY Mean percentage Oil
(SEM). The results of all other fatty acids are available n
Baseline
4 Months
ON ADIPOSE
on request.
Wash-out
Significance SO vs. EPO #
Fatty acid Linoleic acid 7 EPO 10 20 6 30 8 20 8 so y-Linolenic a 7 EPO 10 20 6 30 8 20 8 so Dihomo-y-linolenic 7 EPO 10 20 6 30 8 20 8 SO Arachidonic 7 EPO 10 20 6 30 8 20 8 SO
9.2 8.4 9.1 10.3
(0.4) (0.3) (0.3) (0.8)
9.7 9.2 10.1 11.5
(0.5) (0.3) (0.3) (1.0)
9.9 9.2 10.1 11.5
(0.5) (0.3) (0.3) (1.0)
to
EPO dose response *
Baseline to treatment
Treatment wash-out
NS
NS
NS
0.145 0.129 0.116 0.176
(0.021) (0.025) (0.010) (0.027)
0.138 0.130 0.133 0.151
(0.017) (0.019) (0.010) (0.026)
0.145 0.128 0.139 0.161
(0.022) (0.021) (0.009) (0.024)
NS
NS
< 0.05
0.090 0.080 0.085 0.101
(0.004) (0.005) (0.005) (0.006)
0.097 0.101 0.112 0.100
(0.005) (0.005) (0.007) (0.007)
0.098 0.098 0.112 0.100
(0.004) (0.006) (0.006) (0.007)
< 0.01
NS
NS
0.399 0.465 0.366 0.470
(0.012) (0.036) (0.011) (0.011)
0.404 0.464 0.386 0.457
(0.007) (0.022) (0.011) (0.015)
0.402 0.481 0.388 0.462
(0.008) (0.020) (0.013) (0.011)
NS
NS
NS
Ratios: X 10’ DGLA/Linoleic 7 1.00 (0.10) EPO 10 20 6 0.99 (0.16) 30 8 0.97 (0.13) 20 8 1.04 (0.10) SO Arachidonic/DGLA 7 4.62 (0.29) EPO 10 20 6 6.11 (0.71) 30 8 4.69 (0.62) 8 5.01 (0.49) SO 20 a Complex peak includes C20 : 0. * Two sample r-test comparing the change 20 ml SO vs. 20 ml EPO. * Regression analysis of dose response to treatment. Analysis of variance of all fatty acids within the safflower oil group and y-Iinolenic acid
1.03 1.13 1.14 0.90
(0.08) (0.14) (0.16) (0.07)
1.02 1.10 1.14 0.90
(0.08) (0.15) (0.14) (0.07)
< 0.01
NS
NS
4.19 4.74 3.80 4.99
(0.19) (0.49) (0.54) (0.54)
4.22 5.09 3.60 4.97
(0.28) (0.45) (0.28) (0.45)
< 0.01
NS
NS
in fatty acid pattern
from baseline
10, 20 and 30 ml/day
TISSUE
to treatment
EPO for change
and from treatment
in fatty
to wash-out
acid level from baseline
each group was highly significant (P < 0.01). except for DGLA values in the EPO 10 and 20 ml/day groups.
between
to 4 months
and AA/DGLA
ratio in
203 ml/day) and safflower oil (20 ml/day) on the change in adipose tissue and plasma lipid fatty acid composition from baseline to treatment (4 months) and during the wash-out period (from 4 to 6 months). The relationship between dose of evening primrose oil and the change in fatty acid composition was analysed by linear regression. A P value c 0.05 was considered significant. Results Adipose tissue fatty acid composition The effect of safflower and evening primrose oil on the relative amount of n - 6 fatty acids in adipose tissue is presented in Table 3. Dietary supplementation with safflower and evening primrose oil, both containing large amounts of linoleic acid, increased adipose tissue linoleate levels significantly. At the dose of 20 ml/day, there was no difference between the 2 oils in this respect. The small amounts of dihomo-y-linolenic acid in adipose tissue increased significantly in 2 of the 3 groups receiving evening primrose oil, but not in those on placebo (Fig. 1). The mean increase in dihomo-y-linolenic acid tended to rise with increasing doses of evening primrose oil. The dihomo-y-linolenic level did not fall significantly during the 2 month wash-out period. The adipose tissue concentrations of arachidonic acid tended to increase after 4 months supple-
0.M
A 203
n-6
0.W
(W 0.02
-0.01
_L
1
-0.02 !
10
20
30
ml/day
Fig. 1. Effect of evening primrose oil (EPO) on adipose tissue dihomo-y-linolenic acid levels. The increase in dihomo-y-linolenate after 20 and 30 ml/day EPO supplement is significant, but not at the lower dose of 10 ml/day. Safflower oil (SO), which also contains a high amount of linoleic acid but no y-linolenic acid, did not affect the dihomo-y-linolenic acid levels, The relation between dose of EPO and the rise in dihomo-y-linolenate is not significant.
mentation with evening primrose oil, particularly in the group receiving the highest dose (30 ml/ day). In this group the increased arachidonic acid levels remained after 2 months wash-out period. Arachidonic acid level tended to fall during safflower oil supplementation. There was no significant linear relationship between the dose of evening primrose oil and the percentage increase in adipose tissue arachidonic acid. The mean ratio of dihomo-y-linolenic/ linoleic acid increased significantly in those groups receiving evening primrose oil, but actually fell in those receiving safflower oil, the difference being highly significant (Table 3, P < 0.01). The ratio of arachidonic/ dihomo-y-linolenic acid fell significantly during evening primrose oil supplementation (P < O.Ol), but safflower oil did not affect this ratio. Serum lipid fatty acid composition The linoleic acid content of serum triglycerides tended to rise after both oils (Table 4). The fall during the wash-out period differed and was greater in the safflower oil group. There was no clear relation between the dose of evening primrose oil and the rise in triglyceride linoleate (Table 4). There was also an increase in y-linolenic acid levels during the supplementation period, which was related to the dose of evening primrose oil (Table 4). The levels returned to pre-supplementation values 2 months after stopping the supplement. The increase in dihomo-y-linolenic acid during supplementation with evening primrose oil (20 ml/day) was significantly higher than with safflower oil (20 ml/day), but the decline during the wash-out period was not significantly different between the two oils. The relatively high content of linoleic acid in cholesteryl ester fraction increased with both evening primrose and safflower oil. In contrast the rise in y-linolenyl, dihomo-y-linolenyl and arachidonyl cholesterol were confined to the evening primrose oil supplement but not in a clear, dose-dependent manner (Table 4). A similar pattern emerged for the changes in the relative amount of the n - 6 fatty acids (%) of serum phospholipids (Table 4). The rise in y-linolenic acid in this lipid fraction after evening primrose oil depended on the daily dose of the oil (P < 0.05). Evening primrose oil also reduced the total phospholipid concentration
204 TABLE 4 EFFECT OF SAFFLOWER OIL (SO; 20 ml/day) AND EVENING PRIMROSE OIL (EPO; 10, 20 and 30 ml/day) RELATIVE AMOUNT OF SERUM LIPID POLYUNSATURATED FATTY ACIDS OF THE n - 6 FAMILY
ON THE
Mean percentage (SEM). The results of all other fatty acids are available on request. Oil
n
Baseline
4 Months
Wash-out
Significance SO vs. EPO #
Triglycerides: fatty acid Linoleic 12.1 (1.7) EPO 10 9 10.6 (2.6) 20 6 13.0 (1.9) 30 8 12.8 (2.5) SO 20 8 y-Iinolenic a 0.155 (0.033) EPO 10 9 0.155 (0.022) 20 6 0.224 (0.059) 30 8 0.145 (0.022) so 20 8 Dihomo-y-linolenic 0.130 (0.012) EPO 10 9 0.131 (0.030) 20 6 30 8 0.186 (0.044) 0.160 (0.032) SO 20 8 Arachidonic EPO 10 9 0.706 (0.060) 0.770 (0.141) 20 6 0.924 (0.121) 30 8 0.728 (0.087) SO 20 8 Cholesterol esters Linoleic EPO 10 9 20 6 30 8 SO 20 8 y-Linolenic a EPO 10 9 20 6 30 8 so 20 8 Dihomo-y-Enolenic EPO 10 9 20 6 30 8 SO 20 8 Arachidonic EPO 10 9 20 6 30 8 SO 20 8
54.1 47.4 54.0 53.2
(3.4) (4.1) (1.2) (2.3)
16.8 (2.6) 14.4 (2.4) 17.7 (2.4) 18.7 (2.2)
11.6 (1.1) 11.6 (1.5) 13.5 (2.0) 13.4 (2.1)
EPO dose response *
Baseline to treatment
Treatment to wash-out
NS
< 0.05
NS
0.205 0.309 0.521 0.154
(0.030) (0.088) (0.122) (0.017)
0.152 0.157 0.222 0.132
(0.030) (0.034) (0.075) (0.022)
NS
NS
< 0.05
0.261 0.252 0.357 0.204
(0.053) (0.050) (0.048) (0.027)
0.133 0.126 0.186 0.151
(0.016) (0.018) (0.043) (0.023)
< 0.05
NS
NS
0.788 0.980 1.262 0.804
(0.078) (0.203) (0.163) (0.075)
0.744 0.846 0.944 0.820
(0.081) (0.075) (0.171) (0.069)
NS
NS
NS
NS
< 0.01
NS
58.9 51.1 55.6 60.4
(3.3) (2.1) (1.2) (1.8)
58.4 52.3 50.9 54.0
(3.5) (2.9) (2.1) (1.9)
0.51 0.68 0.54 0.64
(0.06) (0.07) (0.05) (0.08)
0.90 1.13 1.42 0.67
(0.13) (0.12) (0.18) (0.06)
0.54 0.47 0.84 0.55
(0.07) (0.05) (0.19) (0.09)
NS
-=z0.01
NS
0.48 0.41 0.62 0.56
(0.02) (0.05) (0.04) (0.04)
0.76 0.71 1.00 0.58
(0.11) (0.10) (0.14) (0.04)
0.52 0.49 0.60 0.57
(0.04) (0.03) (0.08) (0.05)
< 0.01
< 0.05
NS
4.24 4.84 5.11 4.77
(0.63) (0.74) (0.74) (0.25)
5.37 6.28 6.61 4.90
(0.91) (0.97) (0.69) (0.29)
5.04 5.97 6.10 5.48
(0.75) (0.77) (0.56) (0.29)
i 0.05
NS
NS
205 TABLE Oil
4 (continued) n
Baseline
4 Months
Wash-out
Significance SO vs. EPO *
Phospholipidy Linoleic EPO 10 9 20 6 30 8 so 20 8 y-Linolenic ’ EPO 10 9 20 6 30 8 so 20 8 Dihomo-y-linolenic EPO 10 9 20 6 30 8 so 20 8 Arachidonic EPO 10 9 20 6 30 8 SO 20 8
22.4 20.0 22.0 22.3
(1.6) (1.2) (1.0) (0.9)
0.075 0.056 0.034 0.038
-
(0.022) (0.009) (0.006) (0.010)
22.7 18.8 21.7 24.2
(1.9) (1.2) (1.2) (0.8)
0.050 0.053 0.082 0.034
21.6 20.4 19.8 21.5
(0.009) (0.012) (0.020) (0.010)
(1.6) (1.0) (1.1) (0.7)
0.048 0.029 0.049 0.029
(0.017) (0.005) (0.007) (0.010)
EPO dose response *
Baseline to treatment
Treatment wash-out
< 0.05
< 0.05
NS
NS
NS
< 0.05
to
2.73 2.73 2.74 2.82
(0.18) (0.33) (0.19) (0.24)
4.36 3.82 4.65 3.31
(0.29) (0.59) (0.70) (0.29)
2.79 2.75 3.12 3.14
(0.18) (0.23) (0.16) (0.33)
NS
NS
NS
7.54 8.69 7.95 8.14
(0.53) (0.95) (0.69) (0.26)
9.03 10.48 9.75 8.83
(0.78) (1.39) (0.56) (0.37)
8.45 10.00 9.70 9.05 -
(0.78) (0.82) (0.60) (0.40)
< 0.05
NS
NS
“ Complex peak including C20 : 0. * and #: see Table 3 for explanation.
expressed in mmol/l, but such an effect was not seen in cholesteryl ester or triglyceride concentration (not shown). The relative amount of a-linolenic and eicosapentaenoic acid (n - 3) in serum phospholipids decreased with evening primrose oil supplement. Similar results were also obtained for the cx-linolenic acid content in serum cholesteryl esters. The decrease was always related to the dose of evening primrose oil (not shown) and remained throughout the wash-out period. Serum lipids There were few effects of the 2 oils. Neither safflower oil, nor evening primrose oil had marked effects on serum lipids (Table 5). Serum triglycerides were stable throughout the study. The apparent higher fasting triglyceride concentration in those taking 20 ml/day evening primrose oil reflects the baseline values in this group. There
were no consistent and HDL-cholesterol
differences in total, VLDLbetween the 2 supplements.
Discussion Fatty acid composition Dietary supplementation with evening primrose oil increased the levels of dihomo-y-linolenic acid in adipose tissue. The men in this study were selected to have a low concentration of adipose tissue dihomo-y-linolenate, because it was expected that they were most likely to show an increase in this fatty acid in adipose tissue in response to dietary evening primrose oil supplement. In contrast, there was no change in adipose dihomo-y-linolenate concentration in the control group given safflower oil. This oil is equally rich in linoleic acid, but contains no y-linolenic acid. The conversion of y-linolenic acid to the longchain polyunsaturated n - 6 fatty acids (dihomoy-linolenic, arachidonic and arachidic acid)
206 TABLE
5
EFFECT OF SAFFLOWER OIL (SO; 20 ml/day) AND OF EVENING PRIMROSE OIL (EPO; IO, 20 and 30 ml/day) ON FASTING SERUM LIPIDS Mean percentage
(SEM).
so
EPO
20 (n=8) Cholesterol Total VLDL LDL HDL
Triglycerides Total
(mmol/l) 7.13 (0.43) 0.52 (0.12) 5.10 (0.32) 1.34 (0.09) (mmol/l) 1.61 (0.35)
;;=9)
;;,6)
;;=8)
6.89 (0.53) 0.51 (0.10) 4.95 (0.45) 1.42
6.38 (0.47) 0.28 (0.07) 4.58
(0.12)
7.33 (0.42) 0.83 (0.12) 4.88 (0.46) 1.27 (0.13)
1.52 (0.26)
2.27 * (0.26)
1.07 (0.18)
(0.44) 1.49 (0.13)
Note: No significant difference between the 2 oils by analysis of variance. Regression analysis of dose response to EPO also not significant for all lipid fractions. * This value is not significantly higher than those in the other groups and reflects the tendency to higher baseline values in this group.
bypasses the rate limiting A6-desaturase step. Therefore the rise in dihomo-y-linolenate is almost certainly due to the presence of y-linolenic and not of linoleic acid in evening primrose oil. Linoleic acid in large quantities inhibits its conversion to y-linolenic and dihomo-y-linolenic acid and the decrease in adipose dihomo-y-linolenic/ linoleic acid ratio during safflower oil supports this view. Thus we cannot exclude the possibility that the rise in adipose dihomo-y-linolenic acid during evening primrose oil was partly balanced by an opposing effect on account of its linoleic acid content. It would be of interest to examine the effect of concentrated y-linolenic acid. Adipose dihomo-y-linolenic acid levels remained elevated for at least 2 months after stopping the evening primrose oil supplement. Our data do not allow us to estimate the turnover of this essential fatty acid, but it is clear that it is low. Evening primrose oil also increased the relative amount of dihomo-y-linolenic acid in serum
cholesteryl esters and triglycerides, and of arachidonic acid in serum cholesteryl ester and phospholipid fractions. But, in contrast to adipose tissue fatty acid composition, that of serum lipids returned quickly to baseline after the oil supplement. It is worthy of mention that evening primrose oil specifically decreased the relative amount of a-linolenic acid (n - 3) and of eicosapentaenoic acid (n - 3) in phospholipids in a dose-dependent manner. A similar decrease was also observed for the cy-linolenic acid content of cholesteryl esters, but not for triglycerides. The mechanism of these changes is not clear, but is almost certainly not a simple competition between the n - 3 and n - 6 polyunsaturated fatty acids for incorporation into these serum lipids. This conclusion is deduced from the fact that the relative amount of these n - 3 but not n - 6 fatty acids persisted throughout the 2-month wash-out period. The activity of A6-desaturase activity was probably low in these men, since little or no conversion of linoleic acid from safflower oil to adipose dihomo-y-linolenic acid was observed. For obvious reasons, we could not measure the activity of this enzyme in liver microsomes in our study. Nevertheless, the possibility of low A6-desaturase activity should not be surprising, in view of our selection procedure. The question arises therefore whether low A6-desaturase activity is typical for all Edinburgh men. In animals, A6-desaturase activity is influenced by risk factors for coronary heart disease, such as ‘stress’, alcohol and cholesterol [S]. This possibility clearly requires further study in men. Another, perhaps even more pertinent question is whether low A6-desaturase activity leads to CHD [ll]. Support for this hypothesis comes from the observation that gradients in CHD mortality rates in Europe are inversely related to adipose dihomo-y-linolenic acid levels [3]. Patients with CHD have lower dihomo-y-linolenic acid levels than their controls [5], although this was not confirmed in another study [6]. The effect of evening primrose oil on adipose tissue arachidonate was not as consistent as that on dihomo-y-linolenate levels. However, there was a consistent and significant fall in the ratio of adipose arachidonic/ dihomo-y-linolenic acid. Thus, A5-desaturase may become the rate limiting step in the conversion of y-linolenic to arachidonic
207 acid under these conditions. The results of the effects of the 2 oils on serum cholesteryl ester and phospholipid arachidonic acid levels support this view. The effect of the diet on A’-desaturase activity is not certain. Data from rat studies suggest that a high saturated fat diet leads to higher As-desaturase activity than a diet rich in linoleate [12]. Taken together, this could imply that y-linolenic acid may specifically affect A5-desaturase activity. Lipids and lipoprotein fractions We were unable to demonstrate any consistent effect of either evening primrose oil or safflower oil on the fasting concentrations of total cholesterol and triglycerides, or on VLDL-, LDLand HDL-cholesterol subfractions. This was somewhat surprising, since both evening primrose and safflower oil contain rather large amounts of linoleic acid. According to the Hegsted [13] or Keys [14] formulae, we might have expected a reduction in total cholesterol of approximately 0.14 mmol/l due to the linoleate of evening primrose oil alone. The hypocholesterolaemic effect of y-linolenic acid is not well documented. Early data suggest an increased potency relative to linoleic acid of 2 [15], whilst Horrobin and Manku [16] believe it might be 170 times as potent, depending on initial serum cholesterol levels. Taking this higher value, we should have expected a fall in serum cholesterol of 2.46 mmol/l with 20 ml/day evening primrose oil. Thus our results are at variance with the Horrobin and Manku study [16] of patients mostly with atopic eczema, a condition in which A6-desaturase is known to be impaired [ll]. It is not clear whether this is responsible for their anomalous results. Our results suggest that the estimate of the hypocholesterolaemic effect of ylinolenic acid may well be substantially less. This view is supported by the results of other, smaller studies all showing slight reductions in serum cholesterol, ranging from - 0.18 to - 0.36 mmol/l, some significant others not [17-201. It would be very difficult to show such a small cholesterol reduction convincingly, as it is in the order of magnitude of the analytical error alone. The high-dose evening primrose oil (30 ml/day) increased the HDL/total cholesterol ratio (not shown). Other studies in man using smaller doses
of the oil do not confirm increased HDL levels, and this effect could be due to an increase in total fat intake [21]. In animal studies, the cholesterol lowering effects of evening primrose oil are clearer and are in full agreement with each other [22-241. Jnterestingly, these show that evening primrose oil and also safflower oil are not hypocholesterolaemic per se, but inhibit the rise in serum cholesterol induced by large quantities of dietary cholesterol. If these results can be extrapolated to men at all, they suggest that evening primrose oil may only be a hypocholesterolaemic agent, when raised serum cholesterol is caused by high levels of cholesterol in the diet. Acknowledgements This study was supported by a grant from Efamol Ltd. Robert D. Abraham was supported by a grant from the Blacktown Staff Specialist Trust Fund. We are grateful to Mary Walker, Karin Lyall and Margaret Millar for excellent technical support and to Christina Anderson for typing the manuscript. References Kingsbury, K.J., Brett, C., Stovold, R., Chapman, A., Ander, J. and Morgan, D.M., Abnormal fatty acid composition and human atherosclerosis. Postgrad. Med. J.. 50 (1974) 425. Logan, R.L.. Riemersma, R.A.. Thomson, M.. Oliver, M.F.. Olsson, A.G., Walldius, G., Rossner, S., Kaijser, L., Callmer, E., Carlson, L.A., Lockerbie, L. and Lutz, W., Risk factors for ischaemic heart disease in normal men aged 40. Edinburgh-Stockholm Study. Lancet, 1 (1978) 949. Riemersma, R.A., Wood, D.A., Butler, S., Elton, R.A., Oliver, M., Sale, M., Nikkari, T., Vartiainen. E., Puska, P., Gey, F., Rubba, P., Mancini, M. and Fidanza, F., Linolenic acid content in adipose tissue and coronary heart disease. Br. Med. J., 292 (1986) 1423. Miettinen, T.A., Naukkarinen, V., Huttunen, J.K., Mattill, S. and Kumlin, T., Fatty acid composition of serum lipids predicts myocardial infarction. Br. Med. J.. 282 (1982) 993. Wood, D.A., Butler, S., Riemersma, R.A.. Thomson, M., Oliver, M.F., Fulton, M., Birtwhistle, A. and Elton, R., Adipose tissue and platelet fatty acids and coronary heart disease in Scottish men. Lancet, 2 (1984) 117. Wood, D.A., Riemersma, R.A.. Butler, S., Thomson. M., Macintyre, C., Elton, R.A. and Oliver, M.F., Linoleic and eicosapentaenoic acids in adipose tissue and platelets and risk of coronary heart disease. Lancet, 1 (1987) 177.
208 7 The Pooling Research Group. Relationship of blood pressure, serum cholesterol, smoking habit, relative weight, and ECG abnormalities to incidence of major coronary events. Final report of the pooling project. J. Chron. Dis., 31 (1978) 201. 8 Brenner, R.R., Nutritional and hormonal factors influencing desaturation of essential fatty acids. Prog. Lipid Res., 20 (1982) 41. 9 Carlson, K.J., Lipoprotein fractionation. J. Clin. Pathol., 26 Suppl. 5 (1973) 32. 10 Notghi, A., Riemersma, R.A., Anderton, J.L. and Oliver, M.F., Effect of pindolol versus atenolol on lipid profile in hypertensive patients. Atherosclerosis, 77 (1989) 215. 11 Horrobin, D.F. and Huang, Y-S., The role of linoleic acid and its metabolites in the lowering of plasma cholesterol and the prevention of cardiovascular disease. Int. J. Cardiol., 17 (1987) 241. 12 Hagve, T-A. and Christophersen, B.O., Effect of dietary fats on arachidonic acid and eicosapentaenoic acid biosynthesis and conversion to C22 fatty acids in isolated rat liver cells. Biochim. Biophys. Acta, 796 (1984) 205. 13 Hegsted, D.M., McGandy, R.B., Myers, M.L. and Stare, F.J., Quantitative effects of dietary fat on serum cholesterol in man, Am. J. Clin. Nutr., 17 (1965) 281. 14 Keys, A., Anderson, J.T. and Grande, F., Prediction of the serum cholesterol responses of man to changes in fats in the diet. Lancet, 2 (1957) 959. 15 Worne, H.E. and Smith, L.W., Effects of certain pure long chain polyunsaturated fatty acid esters on blood lipids of man. Am. J. Med. Sci., 237 (1959) 710. 16 Horrobin, D.F. and Manku, M.S., How do polyunsaturated fatty acids lower plasma cholesterol levels? Lipids, 18 (1983) 558.
17 Blaton, V., Muls, E., Herman, A., Criel, A., Declerck, B. and Hollez, D., Effects of dietary alpha- and gamma-linolenate supplement on serum lipids, prostaglandins and platelet function in men. Clin. Chem., 32 (1986) 242 (Abstr.). 18 Ishikawa, T., Fujiyama, Y., Igarashi, 0. et al. Effects of gamma-linolenic acid on plasma lipoproteins and apolipoproteins. Atherosclerosis, 75 (1989) 95. 19 Boberg, M., Vessby, B. and Selinus, I., Effects of dietary supplementation with n - 6 and n - 3 long-chain polyunsaturated fatty acids on serum lipoproteins and platelet function in hypertriglyceridemic patients. Acta Med. Stand., 220 (1986) 153. 20 Szczeklik, A., Gryglewski, R.J., Kosta-Trabka, E. and Zmuda, A., Dihomo-gamma-linolenic acid in patients with atherosclerosis: effects on platelet aggregation, plasma lipids and low-density lipoprotein induced inhibition of prostacycline generation. Thromb. Haemostas., 51 (1984) 186. 21 Knuiman, J. and West, C.E., Differences in HDL cholesterol between populations: no paradox? Lancet. 1 (1983) 296. 22 Huang, Y.S., Manku, MS. and Horrobin, D.F., The effects of dietary cholesterol on blood and liver polyunsaturated fatty acids and on plasma cholesterol in rats fed various types of fatty acid diet. Lipids, 19 (1984) 664. 23 Sugano, M., Ishida, T., Yoshida, K., et al., Effect of mold oil containing gamma-linolenic acid on the blood cholesterol and eicosanoid levels in rats. Agric. Biol. Chem., 50 (1986) 2483. 24 Sugano. M., Ide, T., Ishida, T. and Yoshida, K., Hypocholesterolemic effect of gamma-linolenic acid as evening primrose oil in rats. Ann. Nutr. Metab., 30 (1986) 289.
