点击显示 收起
1 From the Hugh Sinclair Unit of Human Nutrition, School of Food Biosciences, University of Reading, Reading, United Kingdom (YEF, AMM, ECL-F, and CMW); the Institute of Human Nutrition, School of Medicine, University of Southampton, Southampton, United Kingdom (SK and PCC); the Unilever Health Institute, Unilever R&D Vlaardingen, Vlaardingen, Netherlands (GWM); and Roche Vitamins Ltd, Basel, Switzerland (RM).
2 Supported by a grant from the Department for Environment, Food & Rural Affairs, Biotechnology and Biological Sciences Research Council, Roche Vitamins Ltd, Basel, and by Unilever Research, Vlaardingen, under the Agri-Food LINK programme (AFQ111). Roche Vitamins Ltd, Basel, provided the fish oil and fish oil capsules and Unilever Research, Vlaardingen, provided the margarines. 3 Reprints not available. Address correspondence to CM Williams, Hugh Sinclair Unit of Human Nutrition, School of Food Biosciences, University of Reading, Reading RG6 6AP, United Kingdom. E-mail: c.m.williams{at}reading.ac.uk.
ABSTRACT
Background: Dietary -linolenic acid (ALA) can be converted to long-chain n-3 polyunsaturated fatty acids (PUFAs) in humans and may reproduce some of the beneficial effects of eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) on cardiovascular disease risk factors.
Objective: This study aimed to compare the effects of increased dietary intakes of ALA and EPA+DHA on a range of atherogenic risk factors.
Design: This was a placebo-controlled, parallel study involving 150 moderately hyperlipidemic subjects randomly assigned to 1 of 5 interventions: 0.8 or 1.7 g EPA+DHA/d, 4.5 or 9.5 g ALA/d, or an n-6 PUFA control for 6 mo. Fatty acids were incorporated into 25 g of fat spread and 3 capsules to be consumed daily.
Results: The change in fasting or postprandial lipid, glucose, or insulin concentrations or in blood pressure was not significantly different after any of the n-3 PUFA interventions compared with the n-6 PUFA control. The mean (± SEM) change in fasting triacylglycerols after the 1.7-g/d EPA+DHA intervention (-7.7 ± 4.99%) was significantly (P < 0.05) different from the change after the 9.5-g/d ALA intervention (10.9 ± 4.5%). The ex vivo susceptibility of LDL to oxidation was higher after the 1.7-g/d EPA+DHA intervention than after the control and ALA interventions (P < 0.05). There was no significant change in plasma -tocopherol concentrations or in whole plasma antioxidant status in any of the groups.
Conclusion: At estimated biologically equivalent intakes, dietary ALA and EPA+DHA have different physiologic effects.
Key Words: -Linolenic acid • eicosapentaenoic acid • docosahexaenoic acid • polyunsaturated fatty acids • n-3 fatty acids • lipids • plasma fatty acid • LDL oxidation • moderately hyperlipidemic subjects • triacylglycerol
INTRODUCTION
Long-chain (LC) n-3 fatty acids, principally eicosapentaenoic acid (EPA; 20:5n-3) and docosahexaenoic acid (DHA; 22:6n-3) have been found to have a variety of beneficial effects on risk factors for coronary heart disease (CHD) (1), and there is general consensus that consumers should increase their intakes of these polyunsaturated fatty acids (PUFAs) (2–4). However, an important issue for the scientific community and regulatory health bodies relates to the amount and type of n-3 PUFAs that should be recommended in the diet. Although the cardioprotective effects of low doses of EPA+DHA (0.85 g/d) have been confirmed in those with existing heart disease (5), this level of intake is 5 times higher than habitual Western intakes (6, 7).
An alternative source of n-3 PUFAs is -linolenic acid (ALA; 18:3n-3), which in stable-isotope studies in humans was shown to be desaturated and elongated to EPA and DHA (8, 9). Mean dietary intakes of ALA are much higher than those of the LC n-3 PUFAs (6, 7, 10), and there is potential for increasing ALA intakes through the use of ALA-rich oils (eg, rapeseed oil). However, the evidence of a protective effect of ALA on CHD is equivocal. The Norwegian vegetable oil experiment (1965–1966) failed to see an effect of ALA on CHD endpoints after 1 y of supplementation with 5.5 g ALA/d compared with a sunflower oil control (11). One secondary prevention trial reported a substantial reduction in coronary events and death in subjects following a Mediterranean-style diet that also included an ALA-rich margarine (12). However, definitive conclusions cannot be drawn regarding specific effects of ALA because the intervention involved changes in nutrients other than ALA. Two prospective epidemiologic studies reported a protective effect of increased dietary ALA intake on the relative risk from fatal CHD (13, 14). This raises the question as to whether ALA has effects similar to those attributed to the LC n-3 PUFAs.
Although LC n-3 PUFAs have been shown to affect fasting (15) and postprandial (16) blood lipids, the effects of ALA on blood lipids have received little attention. ALA appears to have effects on blood cholesterol concentrations that are similar to those of linoleic acid (LA; 18:2 n-6) (17, 18), but the data on triacylglycerol concentrations are equivocal (19–23). One limitation of these studies is the short duration of supplementation, which may be insufficient to allow ALA to increase tissue LC n-3 PUFA concentrations. The fact that LC n-3 PUFAs have been reported to produce triacylglycerol-lowering effects at low levels of dietary enrichment (1 g/d) (24, 25) makes it feasible that similar effects may be found with ALA over an extended supplementation period.
The present study investigated the comparative effects of the precursor n-3 PUFA, ALA, with those of EPA and DHA on a range of CHD risk factors. Because ALA is less effective at increasing LC n-3 PUFA tissue concentrations when compared gram for gram than is dietary EPA+DHA (19–21), we compared estimated biologically equivalent amounts of ALA and EPA+DHA. The existing literature suggests that 7 g ALA is approximately equivalent to 1 g EPA+DHA in raising tissue LC n-3 PUFA concentrations (8, 19, 22). Comparison of ALA with EPA+DHA was undertaken over an extended time period (6 mo). This was to ensure that the previous failure to observe significant accumulation of tissue EPA+DHA or significant lowering of triacylglycerols with increased ALA intakes did not reflect inadequate time for conversion and tissue accumulation of the LC products.
SUBJECTS AND METHODS
Subjects
Moderately hyperlipidemic but otherwise healthy adults aged 25–72 y were recruited through a variety of means, including e-mailing the staff at the university a general description of the study, advertising in the local media, and contacting potential participants through a database held at the Department of Clinical Pathology, Royal Berkshire Hospital, Reading, United Kingdom. Those interested in taking part were asked to contact the Hugh Sinclair Unit of Human Nutrition to complete a health and lifestyle questionnaire and to provide a screening blood sample. Exclusion criteria for participation in the study were evidence of cardiovascular disease, including angina; diagnosed diabetes or a fasting glucose concentration > 6.5 mmol/L; liver or other endocrine dysfunction; pregnancy or lactation; smoking > 15 cigarettes/d; exercising strenuously > 3 times/wk; body mass index (in kg/m2) < 20 or > 32; and a hemoglobin concentration < 130 g/L in men or 120 g/L in women. Individuals who were prescribed hypolipidemic or antiinflammatory medication, took fatty acid or antioxidant supplements regularly, or consumed > 2 portions of oily fish/wk were excluded. Vegetarians and nonconsumers of margarine were also excluded. Moderate hyperlipidemia was defined as a fasting total cholesterol concentration between 4.6 and 8.0 mmol/L and a triacylglycerol concentration between 0.8 and 3.2 mmol/L.
To achieve the target sample size (n = 150), 200 suitable individuals were initially recruited into the study to allow for possible dropout in this long-term trial. Reasons for dropout were as follows: unspecified (n = 18), lack of time (n = 13), an inability to consume the portion of margarine (n = 8), and a desire to lose weight (n = 2). One hundred fifty subjects completed the protocol. The study was approved by the University of Reading Ethics and Research Committee and the West Berkshire Health Authority Ethics Committees, and each volunteer gave written informed consent before participating.
Study design
The study was a double-blind, placebo-controlled parallel study carried out in 3 cohorts of 50 subjects each. Cohorts were studied over different months to minimize the effect of seasonal variation. Cohort 1 was studied from August 1998 to May 1999, cohort 2 from May 1999 to November 1999, and cohort 3 from May 2000 to November 2000. Subjects were assigned to 1 of 5 dietary treatment groups (n = 30 per group) by blocked stratified randomization, with the groups matched for fasting triacylglycerol concentration, age, and sex. Subjects were asked to replace their normal margarine or butter with a 25-g portion of a specially formulated margarine and 3 oil capsules to be consumed daily. The first month served as a run-in period, during which all participants consumed the control treatment; this was followed by a 6-mo period (months 0–6) during which participants were assigned to 1 of the 5 treatment groups. Subjects were advised to incorporate the margarine into their usual diet, but were not allowed to use the spreads for baking or frying. Advice was given to the subjects to maintain their habitual diet, physical activity levels, smoking habits, and other lifestyle factors and to maintain a stable body weight. Compliance was assessed by the return of full and empty margarine pots and capsule packs and biochemically by the measurement of plasma phospholipid fatty acid composition.
Fasting blood samples were collected and weight was measured at 0, 2, 4, and 6 mo (n = 150), and a postprandial assessment was carried out at 0 and 6 mo in the first 2 cohorts (n = 98). Blood pressure measurements were taken at baseline and after 4 mo of intervention with the use of a Hawksley random-zero sphygmomanometer (Hawksley & Sons Ltd, Sussex, United Kingdom) for the first 2 cohorts or an OMRON MX2 automatic blood pressure monitor (Omron Healthcare, Hoofddorp, Netherlands) for the final cohort. A series of 3 blood pressure measurements were taken after the subjects had been sitting for 5 min. The procedure was explained to the subject before measurements were taken, and the same observer was used on each occasion. The first reading was discarded and the blood pressure measurement was taken as the mean of the 2 subsequent readings.
Subjects completed a previously validated 180-question food-frequency questionnaire (26) before the start of the study and toward the end of the study (at 5 mo). Portion sizes were determined by the use of a combination of household measures (27) and a photographic atlas of food portion sizes (28). Nutrient intakes were determined by using FOODBASE software (version 1.3; Institute of Brain Chemistry, London).
Dietary intervention
The aim of the dietary intervention was to supplement the diet with modest amounts of EPA+DHA or estimated biologically equivalent intakes of the precursor, ALA. We hypothesized that 7 g dietary ALA would be approximately equivalent to 1 g EPA+DHA with respect to raising tissue LC n-3 PUFA concentrations. The target intakes of EPA+DHA in the 2 fish oil intervention groups were 0.7 and 1.5 g/d, respectively, whereas target intakes of ALA in the 2 ALA-supplemented groups were 5.0 and 10.0 g/d, respectively. The target intakes were calculated to include the estimated average contribution from the background UK diet of EPA+DHA (0.2 g/d) and ALA (1.5 g/d) (2, 6).
Specially formulated margarines were produced to deliver the intervention doses in a 25-g portion of margarine. Spreads were manufactured containing 85% fat so that a 25-g portion of spread delivered 21 g fat/d. The control margarine was a typical n-6 PUFA–rich margarine based on sunflower and safflower oils (Table 1). The fish oil margarine was a sunflower oil–based margarine containing 0.73 g EPA and 1.12 g DHA per 100 g spread (0.5 g EPA+DHA per 25-g portion). Because it was not possible to incorporate the higher EPA+DHA intervention of 1.3 g EPA+DHA/d into the margarine without compromising the taste and acceptability of the product, 3 fish oil capsules containing 0.8 g EPA+DHA in total were administered to those in the higher EPA+DHA intervention in addition to the fish oil margarine. The other 4 treatment groups were provided with placebo capsules to maintain the double-blind aspect of the study. The moderate- and high-ALA interventions were based on rapeseed and linseed oils. The final composition of the major fatty acid classes in each margarine was similar across all groups.
View this table:
TABLE 1 . Fatty acid composition of the experimental margarines and capsules consumed by the subjects during the study periods1
The amount of vitamin E in each treatment was standardized according to the intake of unsaturated bonds on the basis of the amount and type of PUFAs present in the spreads (29). Analysis of the margarines and capsules showed that the estimated intake of vitamin E [expressed as tocopherol equivalents (TE)] from 25 g spread and 3 oil capsules for each intervention was as follows: control, 11.4 TE; 0.8 g EPA+DHA/d, 14.0 TE; 1.5 g EPA+DHA/d, 15.2 TE; 4.5 g ALA/d, 13.7 TE; and 9.5 g ALA/d, 14.0 TE. The spreads were prepared by Unilever Health Institute (Vlaardingen, The Netherlands) and the capsules by Roche Vitamins Ltd (Basel, Switzerland).
Plasma lipids, glucose, and insulin
Fasting and postprandial blood samples were collected into chilled, 10-mL EDTA-coated tubes. All blood samples except those collected for LDL isolation were centrifuged at 1600 x g for 10 min at room temperature. For HDL-cholesterol analysis, dextran sulfate and magnesium chloride were added to a subsample of plasma, which precipitated all apolipoprotein (apo) B–containing lipoproteins, allowing for HDL-cholesterol analysis (30). The supernatant fluid and the remaining plasma were stored at -20°C until analyzed.
Plasma samples were analyzed for triacylglycerols, total cholesterol, HDL cholesterol, apo B, glucose, and fatty acid concentrations with the use of an IL Monarch Automatic Analyser (Instrumentation Laboratories Ltd, Warrington, United Kingdom) and enzymatic colorimetric kits (Instrumentation Laboratories Ltd) or by immunoturbimetric methods (Apo B; Wako Chemicals Gmbh, Neuss, Germany). LDL-cholesterol concentrations were calculated by using the Friedewald formula (31), modified for molar concentrations. The postprandial triacylglycerol response was expressed as the area under the postprandial triacylglycerol curve (AUC, 0–480 min) calculated by using the trapezoidal rule or as the incremental area under the curve (IAUC, 0–480 min) calculated as the triacylglycerol AUC minus fasting triacylglycerol concentrations. Similar calculations were made for postprandial glucose responses.
Because plasma fatty acid concentrations drop sharply in the immediate postprandial period and rise postprandially as the result of increased chylomicron and adipose tissue lipolysis, the shape of the postprandial fatty acid response is complex, representing several metabolic consequences. In the data analysis, the percentage of fatty acid suppression from 0 to 60 min and the plasma fatty acid AUC for 180–480 min were used as indexes of fatty acid metabolism. The percentage of fatty acid suppression in the first 60 min is thought to reflect insulin induced-suppression of adipose tissue lipolysis, whereas the AUC for 180–480 min reflects postprandial fatty acid accumulation due to chylomicron and adipose tissue lipolysis. Insulin was assayed with the use of a specific enzyme-linked immunosorbent assay kit (Dako Ltd, High Wycombe, United Kingdom).
Samples taken at 0, 2, 4, and 6 mo were stored and analyzed at the end of each cohort in a single batch to minimize interassay variability. The mean intraassay CVs for the triacylglycerol, cholesterol, fatty acid, and glucose assays were 2.26%, 1.64%, 1.48%, and 3.2%, respectively. The interassay variability for the triacylglycerol, cholesterol, fatty acid, and glucose assays was 4.8%, 3.1%, 4.4%, and 2.17%, respectively. For insulin and apo B, results were accepted if the CV for a control sample of known concentration was < 5%.
Postprandial protocol
Cohorts 1 and 2 (n = 98) completed a postprandial assessment at baseline and after 6 mo. The day before the assessment, no alcohol or strenuous exercise was permitted and a relatively low-fat (< 10 g) evening meal was provided to standardize short-term fat intake. After the subjects fasted for 12 h overnight, an indwelling cannula was inserted into the antecubital vein of the forearm and a fasting blood sample was taken. After a test breakfast (0 min) and lunch (330 min) rich in saturated fatty acids, blood samples were taken at 0, 60, 120, 180, 240, 300, 330, 360, 390, 420, and 480 min for measurement of plasma triacylglycerols, glucose, and fatty acids. The nutritional content of the test breakfast was 3.9 MJ energy, 111 g carbohydrate, 19 g protein and 49 g fat, which comprised 29.6 g saturated fatty acids (SFAs), 12.2 g monounsaturated fatty acids (MUFAs), 1.6 g PUFAs, and 2.5 g trans fatty acids. The nutritional content of the test lunch was 2.3 MJ energy, 63 g carbohydrate, 15 g protein, and 29 g fat, of which 14.3 g was SFAs, 7.1 g was MUFAs, 3.0 g was PUFAs, and 2.9 g was trans fatty acids.
Plasma phospholipid fatty acids
The method for isolation and quantification of plasma phospholipid fatty acids was described previously (32).
LDL oxidation and measurement of plasma antioxidants
LDL was isolated from plasma by using a method adapted from Vieira et al (33). The density of plasma (2 mL) was adjusted to 1.21 kg/L with KBr and then overlaid with a solution with a density of 1.006 kg/L in a 5-mL Beckman Optiseal polyallomer centrifuge tube and the solution centrifuged at 65 000 rpm for 50 min at 4 °C in a Beckman near vertical rotor (NVT65; Beckman Instruments UK Ltd, High Wycombe, United Kingdom). The orange band of LDL was extracted and further purified by adjusting the density of the LDL layer to 1.15 kg/L. The solution was overlaid with a solution with a density of 1.063 kg/L in a 5-mL Beckman Optiseal polyallomer centrifuge tube and the tubes were centrifuged at 65 000 rpm for 3 h at 4 °C in the same rotor. LDL, which formed the top layer, was removed and dialyzed against 1.5–2 L phosphate buffer in the presence of Chelex 100 (Sigma-Aldrich Company Ltd, Poole, United Kingdom) for 16 h to remove the EDTA and KBr. The susceptibility of the LDL to oxidation was assessed by measuring the formation of conjugated dienes at an absorbance of 234 nm at 37 °C. Lipid peroxidation in LDL (50 µg LDL protein/mL) was induced by CuSO4, and the lag phase (intercept of the tangent of the lag and propagation phases) was used as the primary index of the susceptibility of the LDL particle to oxidation.
-Tocopherol was extracted from a 500-mL aliquot of plasma by the addition of 500 µL of 10 mmol aqueous sodium dodecyl sulfate/L and 1 mL ethanol and -tocopherol acetate added as the internal standard. -Tocopherol was extracted in three 2-mL portions of hexane, and the sample was redissolved in 2 mL hexane and analyzed by HPLC (Hewlett Packard UK Ltd, Berkshire, United Kingdom) with a fluorescence detector LC240 (Perkin Elmer Ltd, Beaconsfield, United Kingdom). Results were accepted if the CV for a control sample of known concentration fell below 5%. -Tocopherol is expressed as cholesterol standardized amounts (µmol/mmol total plasma cholesterol). The ferric-reducing ability of plasma as a measure of antioxidant power was measured by using the method of Benzie and Strain (34). The mean intraassay CV for the ferric-reducing antioxidant power assay was 3.71%.
Statistical analysis
Sample size (n = 150) was estimated by using power calculations based on the predictive triacylglycerol-lowering effect of a daily intake of 1.5 g EPA+DHA, the required level of significance (P < 0.05) with 80% power. Values are reported as means ± SEMs for each treatment group. Data were checked for normality, and skewed parameters were log transformed before statistical analysis when possible. For normally distributed fasting data, the statistical significances of time and treatment were analyzed by using two-factor repeated-measures analysis of variance (ANOVA) with the baseline value as the covariate. When the interaction between time and treatment was significant, post hoc tests with Bonferroni’s correction were used. For data that were not normally distributed, two-factor repeated-measures ANOVA was carried out on the ranks for the variables to test for any interaction effect. When this was found to be significant, the Kruskal-Wallis test was used to test between-group comparisons and the Friedman test for within-group comparisons. Significance levels were adjusted for multiple comparisons by using Bonferroni corrections, eg, the P value was multiplied by 10, the number of comparisons made between the 5 treatment groups. All statistical analyses were performed by using the SPSS statistical package (version 10; SPSS Inc, Chicago).
RESULTS
Study population
The characteristics of the different intervention groups at baseline are shown in Tables 2, 3, and 4. There were no significant differences among the treatment groups at baseline with respect to age, sex profile, BMI, body weight, systolic or diastolic blood pressure or any of the lipid indexes to be investigated. There was a small increase in body weight over the course of the study (P < 0.05), but there was no significant difference between any of the groups in the absolute or percentage change in weight from baseline to 6 mo. Thus, because any potential effect of this small weight gain on plasma lipids should affect both the control and intervention groups equally, the hypothesis under investigation was not affected. Systolic and diastolic blood pressure tended to decrease over the course of the study but were not significantly different from baseline. There was no significant difference between the control group and the n-3 intervention groups in the absolute or percentage change in either systolic or diastolic blood pressure from baseline to 4 mo.
View this table:
TABLE 2 . Characteristics at baseline of the subjects in the different treatment groups1
View this table:
TABLE 3 . Physiologic and fasting biochemical indexes in all 5 dietary treatment groups over the course of the study period1
View this table:
TABLE 4 . Postprandial lipid and glucose responses to high-fat meals at baseline and after 6 mo of dietary intervention1
Compliance and dietary intakes
Mean compliance with the interventions as measured by the return of empty margarine pots and capsule packs was > 92% and > 88%, respectively, across all treatment groups and was not significantly different among the groups, indicating comparable acceptability of the products across the groups. Calculation of total dietary intakes was adjusted for individual compliance with the intervention. Analysis of the food-frequency questionnaires administered before the intervention showed no significant differences among groups in habitual energy intake; the percentage of energy derived from protein, carbohydrate, alcohol, fat, total SFAs, total MUFAs, or total PUFAs; or the intakes of individual n-6 (LA and arachidonic acid) and n-3 PUFAs (ALA, EPA, and DHA) (data not shown).
Reported daily intakes during the intervention are shown in Table 5. Mean total ALA intakes with the moderate- and high-ALA diets were 4.5 and 9.5 g/d, respectively, and were significantly different from ALA intakes in all other groups including the control. Mean EPA+DHA intakes with the moderate- and high-EPA+DHA diets were 0.8 and 1.7 g/d, respectively, and were significantly different from EPA and DHA intakes in all other groups including the control. There was a significant difference in total n-6 PUFA and LA intakes between the ALA-supplemented and non-ALA-supplemented groups because of the substitution of LA with ALA in the margarine. There were no significant differences among treatment groups with respect to the percentage of energy derived from protein, carbohydrate, alcohol, or fat or in total SFAs, MUFAs, or PUFAs. There was a significant difference in the total energy intake between the control group and the group receiving 4.5 g ALA/d.
View this table:
TABLE 5 . Reported daily intakes of energy, macronutrients, and fatty acids in the different groups during the dietary intervention1
The mean percentage of energy from fat increased significantly (P < 0.05) within all groups during the intervention compared with baseline; the mean increase was 3.6 ± 0.4% in all groups (data not shown). This mostly represented an increase in the percentage of energy from PUFAs (2.9 ± 0.18%; P < 0.001) and was accompanied by a significant reduction in the percentage of energy derived from carbohydrate (-2.7 ± 0.4%; P < 0.05). However, there were no significant differences between the control and the n-3 intervention groups in the changes that occurred in macronutrient intake, and total energy intakes did not change significantly over the course of the study. Correlation analysis showed no significant association between the change in the percentage of energy from fat and the change in blood lipids.
Plasma phospholipid fatty acids
No significant differences were detected among groups in plasma fatty acid composition at baseline, with the exception of DHA in plasma phospholipid (Table 6). The proportion of DHA in plasma phospholipid at baseline was significantly lower in both the 0.8-g/d EPA+DHA and the 9.5-g/d ALA groups than in the control group (P < 0.05). For this reason, the absolute changes in plasma phospholipid fatty acids were compared between groups (Figure 1). There were no significant changes in the composition of plasma fatty acids of interest in the control group over the course of the study. The changes that occurred in the treatment groups indicated compliance with the dietary intervention.
View this table:
TABLE 6 . Plasma phospholipid composition of the fatty acids of interest for subjects at baseline and after 3 and 6 mo of dietary intervention1
FIGURE 1. . Mean (± SEM) change in plasma phospholipid fatty acid composition from baseline to 6 mo. ALA, -linolenic acid; AA, arachidonic acid; EPA, eicosapentaenoic acid; DPA, docosapentaenoic acid; DHA, docosahexaenoic acid. *Significantly different from change in the control group, P < 0.05.
EPA+DHA-supplemented groups
The proportion of EPA in plasma phospholipid was significantly increased compared with baseline after the 0.8-g/d and 1.7-g/d EPA+DHA interventions. However, the change in the proportion of EPA was not significantly different from the change in the control group (Figure 1). There was little effect of the EPA+DHA interventions on the proportion of docosapentaenoic acid in plasma phospholipids. The proportion of DHA was significantly increased after both EPA+DHA interventions compared with baseline and the change was significantly greater than the change after the control treatment. The proportion of arachidonic acid tended to decrease after the 1.7-g/d EPA+DHA intervention but was not significantly lower after 6 mo.
ALA-supplemented groups
The proportion of ALA in plasma phospholipids was significantly higher after supplementation with either 4.5 or 9.5 g ALA/d and the change in ALA was significantly different from the change in the control group. The proportion of EPA in plasma phospholipids was increased significantly after both ALA interventions. However, only the change in EPA after the 9.5-g/d ALA intervention was significantly different from the change with the control. The proportion of docosapentaenoic acid was significantly higher after the 9.5-g/d ALA intervention, but the change was not significantly different from that with the control. The proportion of DHA did not change after either ALA intervention. The proportion of arachidonic acid increased slightly, but not significantly, within the groups and was not significantly different from the change in the control group.
Fasting plasma measures
Fasting plasma lipid, apo B, glucose, insulin, and fatty acid concentrations during the study are shown in Table 3. The two-factor ANOVA detected an effect of time (P < 0.01), treatment (P < 0.05), and a time-dependent effect of treatment (time x treatment interaction, P < 0.05) on mean fasting triacylglycerol concentrations. Treatment with both EPA+DHA interventions was associated with a significant decrease (15%) in mean fasting triacylglycerol concentrations after 2 but not after 6 mo. Fasting triacylglycerol concentrations remained lower than baseline concentrations after 6 mo with the 1.7-g/d EPA+DHA intervention (1.40 mmol/L), although the effect was not significant. The percentage changes from 0 to 6 mo in fasting triacylglycerols in these groups were not significantly different from the change in the control group. However, the change in fasting triacylglycerols (0–6 mo) after the 1.7-g/d EPA+DHA intervention was significantly different from the change after the 9.5-g/d ALA intervention (P < 0.05).
Significant time-related increases in mean fasting total cholesterol, LDL-cholesterol, HDL-cholesterol, and apo B concentrations and in the ratio of LDL to apo B were observed, but no significant main effect of treatment or time x treatment interactions were found. The increase in LDL cholesterol from baseline to 6 mo was particularly marked after the 1.7-g/d EPA+DHA intervention, but there was no significant different in the percentage change in total cholesterol, LDL cholesterol, HDL cholesterol, apo B, or LDL:apo B from baseline to 6 mo between the n-3 intervention groups and the control group.
A significant effect of time on mean fasting glucose and insulin concentrations was observed, but no significant main effect of treatment or treatment x time interactions were found. There was a tendency for a time-dependent reduction in mean fasting fatty acid concentrations (P < 0.05) in all groups, particularly after the 1.7-g/d EPA+DHA intervention. There was no significant difference in the percentage change in plasma fatty acid, glucose, or insulin concentrations from 0 to 6 mo between the n-3 intervention groups and the control group.
Postprandial plasma measures
At baseline, there was no significant difference between groups in the postprandial triacylglycerol response to a high-fat meal as measured by the AUC or IAUC (Table 4). The mean plasma triacylglycerol response to the high-fat test breakfast and lunch tended to be greater in all groups after 6 mo of intervention, with the exception of the 1.7-g/d EPA+DHA intervention (Figure 2). Two-factor ANOVA detected an effect of time (P < 0.01), treatment (P < 0.05), and a time-dependent effect of treatment (time x treatment interaction, P < 0.01) on the mean postprandial triacylglycerol concentration. The mean postprandial triacylglycerol AUC and the percentage change from 0 to 6 mo was significantly (P < 0.05) different between the 1.7-g/d EPA+DHA intervention (0.8%) and the 4.5-g/d ALA intervention (23%).
FIGURE 2. . Mean (± SEM) plasma triacylglycerol (TG) concentrations over 8 h after consumption of a standard test breakfast (49 g fat at 0 min) and a test lunch (29 g fat at 330 min) at baseline (PP1) and after 6 mo of dietary intervention (PP2). EPA, eicosapentaenoic acid; DHA, docosahexaenoic acid; ALA, -linolenic acid. Two-factor ANOVA detected significant effects of time (P < 0.01) and treatment (P < 0.05) and a significant time x treatment interaction (P < 0.01).
The mean postprandial triacylglycerol IAUC tended to increase over the 6 mo (time effect, P < 0.05), but no significant treatment or time x treatment interaction was observed. The percentage change in postprandial triacylglycerol AUC and IAUC from baseline to 6 mo between the n-3 intervention groups and the control group was not significantly different.
The postprandial plasma fatty acid response followed a normal pattern, falling sharply after consumption of the test breakfast and lunch and gradually rising to fasting levels 5 h after breakfast. There was no significant change in percentage of fatty acid suppression at 60 min after 6 mo of intervention compared with baseline or control, although the greatest suppression occurred after the 1.7-g/d EPA+DHA intervention (10.1%). The postprandial plasma fatty acid AUC (180–480 min) was lower in all groups after the intervention (time effect, P < 0.05), particularly after the 1.7-g/d EPA+DHA and the 9.5-g/d ALA interventions. There was no significant difference in the percentage of change in plasma fatty acid AUC (180–480 min) between the control and n-3 intervention groups.
The mean postprandial glucose responses as measured by the AUC and IAUC did not change significantly over the 6 mo (no significant time effect), and there was no significant difference in the percentage change in the glucose AUC or IAUC between the control and n-3 intervention groups.
LDL oxidation and plasma antioxidants
LDL lag phase was significantly reduced after both 2 and 6 mo compared with baseline after the 1.7-g/d EPA+DHA diet only (Table 7). The mean reduction in the lag phase was 9 ± 1.9 min and was significantly different from the change in the control group (2 ± 1.8 min), the 4.5-g/d ALA group (-1.9 ± 2.3 min), and the 9.5-g/d ALA group (0.3 ± 3.3 min). The lag phase tended to be reduced after the 0.8-g/d EPA+DHA intervention but was not significantly different from baseline or from the change in the control group. There was no significant effect of time, treatment, or any time x treatment interaction on plasma -tocopherol concentrations or in the ferric-reducing antioxidant power of plasma.
View this table:
TABLE 7 . LDL lag phase, plasma -tocopherol content, and ferric-reducing ability of plasma at baseline and after 2 and 6 mo of dietary intervention in the 5 treatment groups1
DISCUSSION
The major aim of this study was to determine whether long-term supplementation with the precursor n-3 fatty acid, ALA, would produce effects comparable with those of the longer-chain n-3 PUFAs (EPA and DHA) on a range of established and emerging risk factors for CHD. In this study, we introduced the concept of comparing biologically equivalent amounts of ALA and EPA+DHA. The plasma phospholipid fatty acid composition data show that when compared in this way, dietary ALA is as effective as low-dose preformed EPA+DHA at increasing the proportions of EPA. However, in agreement with the results of many shorter-term studies (17, 32, 35, 36), long-term increased ALA intake failed to increase plasma phospholipid DHA concentrations, suggesting that ALA cannot be used as a replacement for preformed dietary DHA. The lack of effect of a substantial increase in dietary ALA on the proportion of AA, a finding also observed in other studies (17, 32, 35), questions conventional wisdom regarding the competition between LA and ALA.
Overall, there does not appear to be an obvious beneficial effect of increased long-term dietary EPA+DHA or ALA intake on the lipoprotein risk factor profile at the amounts investigated in this study in comparison with an n-6 PUFA intervention (control). Both intakes of EPA+DHA used in this study did show an initial trend toward triacylglycerol lowering, confirming observations from other studies of modest, nonsignificant reductions in fasting triacylglycerol concentrations with EPA+DHA intakes of between 1 and 2 g/d (37, 38). However, the initial hypotriglyceridemic effect of the fish oils in this study was not sustained, particularly with the lower dose of EPA+DHA (0.8 g/d). The mean EPA and DHA content of both plasma and platelet phospholipids remained constant between 2 and 6 mo, confirming continued compliance. Few long-term studies of low-dose EPA+DHA have been carried out, but both a gradual attenuation of the triacylglycerol-lowering effect over time (39) and sustained effects over 2 y (40) have been reported. It is possible that an EPA+DHA intake of 1 g/d represents the limit below which the detection of significant differences in triacylglycerol concentrations becomes difficult, unless sample sizes larger than those in the present study are used.
It is generally considered that the beneficial effects of ALA are due to conversion to LC n-3 PUFAs rather than to any effect of ALA per se. Dietary ALA at intakes up to 9.5 g/d failed to reduce plasma triacylglycerol concentrations despite changes in plasma phospholipid EPA comparable with those observed with increased dietary EPA+DHA, confirming previous short-term studies (19–21). Conversely, we observed a tendency for fasting triacylglycerol concentrations to increase after ALA supplementation compared with the EPA+DHA intervention, suggesting possible divergent effects of the plant and marine n-3 PUFAs on plasma triacylglycerols. In a study by Bemelmens et al (41), a significant increase in fasting triacylglycerols was observed after a 2-y dietary intervention with 6.3 g ALA/d compared with an LA control. Although the lack of a hypotriglyceridemic effect may be ascribed to the rather modest increases in plasma phospholipid EPA formed from ALA, compared with the greater overall increases in total phospholipid EPA and DHA from preformed EPA+DHA, the tendency toward a hypertriglyceridemic effect of ALA is harder to explain. It has been suggested that in addition to undergoing ß-oxidation, ALA can be recycled into other fatty acids (42), which hypothetically would increase the amount of fatty acid substrates available for triacylglycerol synthesis.
The differential effects of ALA and EPA+DHA on the postprandial triacylglycerol AUC largely reflected the changes occurring in the fasting triacylglycerols because there was no differential effect of the treatments on the postprandial triacylglycerol IAUC, which essentially eliminates the fasting effect. The EPA+DHA interventions did not produce the expected reduction in the postprandial triacylglycerol response, although the postprandial triacylglycerol AUC after 6 mo on the 1.7-g/d EPA+DHA diet tended to be lower than in the other groups. Lovegrove et al (43) also observed no significant effect on the postprandial triacylglycerol response to an SFA-rich test meal after an intake of 1.6 g EPA+DHA/d in the form of enriched foods for 3 wk.
The delivery of fatty acids to the liver is a major determinant of hepatic VLDL-triacylglycerol secretion and there is evidence that fish oil supplementation reduces fasting plasma fatty acid concentrations (43, 44), which may represent a mechanism for the triacylglycerol-lowering effects of EPA+DHA. In this study, we observed a trend toward a greater reduction in fasting fatty acids (-14%; P = 0.06) and greater fatty acid suppression at 60 min postprandially after the 1.7-g/d EPA+DHA intervention. This occurred without any significant change in fasting insulin or glucose concentrations and may reflect an increase in insulin sensitivity as far as fatty acid suppression is concerned.
In the present study, the ALA interventions had no significant effect on fasting plasma total, LDL-, or HDL-cholesterol concentrations compared with an n-6 control, which confirms the observations of others (18). The trend toward a greater increase in total and LDL cholesterol after the 1.7-g/d EPA+DHA interventions is surprising given the low intakes of EPA+DHA used in this study. Harris (15) concluded in his analysis of 32 parallel trials with an average EPA+DHA intake of 3–4 g/d that LC n-3 PUFAs do not cause an increase in total cholesterol concentrations but that LDL-cholesterol concentrations increase on average by 5%. The fish oil blend used in the current study was rich in DHA, with a DHA-to-EPA ratio of 1.5:1. A study of the differential effects of EPA and DHA reported an 8% increase in LDL cholesterol after supplementation with DHA but not EPA (45). However, the increase in LDL cholesterol observed with fish oil has been associated with the LDL particles becoming larger and less dense, resulting in a reduced atherogenic risk (38, 46). The tendency toward an increased LDL:apo B in the 1.7-g/d EPA+DHA group observed in this study may reflect an increased LDL particle size counteracting possible negative effects of an increase in LDL-cholesterol concentrations.
At the levels of supplementation used in this study, effects of n-3 PUFAs on blood pressure were not observed. Meta analyses suggest that a minimum daily amount of 3 g EPA+DHA/d may be needed for a significant reduction (47), whereas an intake of 9.2 g ALA/d over 7 wk failed to affect blood pressure (48). Our results are consistent with these findings.
Dietary EPA+DHA at an intake of 1.7 g/d increased the potential susceptibility of LDL to oxidation in vitro, whereas increasing dietary ALA had no effect on LDL susceptibility to oxidation. The effect of EPA+DHA occurred despite the addition of -tocopherol to the spreads and capsules. In a similar study using margarine enriched with 1 g EPA+DHA/d, a significant reduction in lag phase (5 min) was reported after 21 d of intervention (38). Conversely, supplementation of the diet with capsules containing 0.3, 0.6, or 0.9 g EPA+DHA/d for 16 wk had no effect on the lag phase (25). Although we have confirmed a possible advantage to increasing intakes of ALA rather than LC n-3 PUFAs on subsequent in vitro susceptibility of LDL to oxidation, the extent to which this measurement of peroxidation potential reflects actual in vivo oxidation is not known. Measurements of -tocopherol and ferric-reducing antioxidant power did not suggest detrimental effects on antioxidant status as a result of any of the interventions. The measurement of LDL oxidation products such as hydroxy fatty acids, oxysterols, or F2-isoprostanes is now considered preferable to copper-induced LDL oxidation as an index of oxidative stress (49).
In conclusion, the present data support the view that dietary ALA and EPA+DHA have different physiologic effects. We conclude that long-term consumption of modest amounts of dietary ALA does not reproduce the effects of preformed EPA+DHA on triacylglycerol concentrations, on tissue DHA concentrations, or on in vitro susceptibility to oxidation of LDL. The results of the present study do not support the argument that the previous failure to observe significant accumulation of tissue DHA or triacylglycerol lowering in ALA intervention trials was due to inadequate time for conversion and tissue accumulation of the LC n-3 PUFAs.
ACKNOWLEDGMENTS
We acknowledge GWM, Jan Luff (University of Reading), and Tapati Banerjee (University of Southampton) for technical assistance; Patrick Kelly (University of Reading) for contributions to statistical analysis; DL Williams (Royal Berkshire Hospital, Reading) for identification of suitable volunteers; and Paul Robinson (Royal Berkshire Hospital, Reading) for biochemical analysis. We thank the study subjects for their participation in our studies.
The authors contributed as follows: study design, all authors; data collection, YEF, AMM, ECL-F, and SK; data analysis, YEF, AMM, and SK; and writing the manuscript, all authors.
REFERENCES