点击显示 收起
1 From the School of Biomedical & Molecular Sciences, University of Surrey, Guildford, Surrey, United Kingdom (MDG, IGD, LMM, DJM, and BAG); the Nutritional Sciences Research Division, King's College London, London, United Kingdom (TABS, FL, and SS); the Centre for the Genetics of Cardiovascular Disease, Royal Free and University College London Medical School, London, United Kingdom (JAC); the Medical Research Council Cardiovascular Research Group, Wolfson Institute, Barts and The London Queen Mary's School Medicine and Dentistry, London, United Kingdom (GLM)
2 Supported by the UK Food Standards Agency. Unilever Research and Mills DA, Norway, provided the spreads and the salmon spread, respectively. 3 Address reprint requests to BA Griffin, School of Biomedical and Molecular Sciences, University of Surrey, Guildford, Surrey, United Kingdom. E-mail: b.griffin{at}surrey.ac.uk.
ABSTRACT
Background: Insulin resistance is associated with elevated plasma triacylglycerol, low HDL concentrations, elevated postprandial lipemia, and a predominance of small, dense LDLs (sdLDLs). It has been hypothesized that the dietary ratio of n–6 to n–3 (n–6:n–3) polyunsaturated fatty acids (PUFAs) may have favorable effects on these risk factors by increasing insulin sensitivity.
Objective: The objective was to measure changes in insulin sensitivity, lipoprotein size, and postprandial lipemia after a 6-mo alteration in n–6:n–3.
Design: In a randomized, parallel design in 258 subjects aged 45–70 y, we compared 4 diets providing 6% of energy as PUFAs with an n–6:n–3 between 5:1 and 3:1 with a control diet that had an n–6:n–3 of 10:1. The diets were enriched in -linolenic acid, eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), or both. Insulin sensitivity was assessed with the homeostatic model assessment of insulin resistance and the revised quantitative insulin sensitivity test.
Results: Dietary intervention did not influence insulin sensitivity or postprandial lipase activities. Fasting and postprandial triacylglycerol concentrations were lower, and the proportion of sdLDLs decreased (by 12.7%; 95% CI: –22.9%, 2.4%), with an n–6:n–3 of 3:1, which was achieved by the addition of long-chain n–3 PUFAs (EPA and DHA).
Conclusions: Decreasing the n–6:n–3 does not influence insulin sensitivity or lipase activities in older subjects. The reduction in plasma triacylglycerol after an increased intake of n–3 long-chain PUFAs results in favorable changes in LDL size.
Key Words: Blood lipids • n–3 fatty acids • n–6 fatty acids • insulin sensitivity • lipoprotein lipase • LDL • HDL • long-chain polyunsaturated fatty acids
INTRODUCTION
Insulin resistance and elevated fasting and postprandial blood insulin concentrations are associated with an increased risk of ischemic heart disease (IHD) (1, 2). The prevalence of insulin resistance increases with age, obesity, and low physical activity and has reached epidemic proportions in most developing countries (3). Part of the increased risk of IHD associated with insulin resistance and hyperinsulinemia appears to be associated with dyslipidemia, commonly referred to as the atherogenic lipoprotein phenotype, which is characterized by elevated VLDL concentrations, impaired postprandial clearance of triacylglycerol-rich lipoproteins, low HDL-cholesterol concentrations, and a preponderance of small, dense LDL particles. The mechanisms leading to this dyslipidemia may, in part, be explained by an inappropriate release or overspill of nonesterified fatty acids (NEFAs) from adipose tissue and failure to activate lipoprotein lipase, which, in turn, leads to an overproduction and impaired clearance of triacylglycerols, respectively (4). Data from animal studies suggest that dietary polyunsaturated fatty acids (PUFAs) may influence insulin sensitivity (5, 6). Consequently, modifying the intake of PUFAs may correct some of these lipid abnormalities by increasing the sensitivity of tissues to insulin. Previous studies have shown that n–6 and n–3 fatty acids have different effects on lipid metabolism, particularly with regard to plasma triacylglycerol, LDL, HDL, and lipoprotein subclasses (7, 8). An intake of long-chain n–3 PUFAs (LC-PUFAs) in the range of 3 to 6 g/d, in the form of eicosapentaenoic acid (20:5n–3; EPA) and docosahexaenoic acid (22:6n–3; DHA)supplements, has been shown to lower plasma VLDL and triacylglycerol concentrations and to increase LDL particle size and the concentration of HDL2 cholesterol. These effects are not reproduced by their shorter chain precursor -linolenic acid (18:3n–3) at comparable doses (9). Evidence from prospective cohort studies and secondary prevention trials (10, 11) of IHD suggest that altering the dietary ratio of n–6 to n–3 fatty acids (n–6:n–3) decreases the risk of fatal IHD. However, most studies that have examined this relation, and the effect of the n–6:n–3 on insulin sensitivity and postprandial lipemia, have used dietary supplements and not a food-based intervention. They have frequently focused on young, healthy subjects or patients with cardiovascular disease or diabetes, and the interventions have generally been shorter than 3 mo (12–14). The Quantification of the Optimal n–6/n–3 ratio in the UK Diet (OPTILIP) Study was designed to assess the effects of lowering the dietary n–6:n–3 on cardiovascular disease risk factors in older persons. The objective was achieved by using a food-based intervention that involved increasing the relative intake of -linolenic acid or n–3 LC-PUFAs (notably EPA and DHA), or both, in relation to the intake of linoleic acid. The results for changes in hemostatic factors are reported elsewhere (15). This report presents results for insulin sensitivity, postprandial lipid metabolism, and lipoprotein size.
SUBJECTS AND METHODS
Subjects
Men and postmenopausal women aged 45–70 y were primarily recruited from general practices participating in the UK Medical Research Council General Practice Research Framework in the towns of Camberley, Surrey, and North Mymms (Hertfordshire, United Kingdom). An additional 29 subjects were recruited from among the staff of King's College London and its associated hospitals. Exclusion criteria were as follows: body mass index (BMI; in kg/m2) <20 or >35; fasting serum cholesterol >8 mmol/L or triacylglycerol >6.0 mmol/L; abnormal liver function or hematology; clinical history of cholestatic liver disease, pancreatitis, diabetes mellitus, or myocardial infarction; current use of anticoagulants (excluding aspirin). In the younger subjects, postmenopausal status—defined as a span of 1 y since menstruation—was confirmed by measurement of the serum concentration of follicle-stimulating hormone. Subjects taking blood pressure or lipid-lowering medications were eligible if their medication regimens were stable. Suitable subjects were identified from general practice records and were invited by letter to take part in the study. Subjects from King's College London were invited by a broadcast email to take part in the study. The nature of the study was explained to the subjects, particularly that they must be prepared to eat oily fish. To further assess eligibility for the study, we asked the subjects to complete a health and food questionnaire identical to that used in the European Prospective Investigation in Cancer study and to attend a screening clinic. Blood samples were collected after an overnight fast for liver function tests, follicle-stimulating hormone, lipids, plasma glucose, insulin, and routine hematology. Blood pressure (Omron 705CP; Omron Healthcare Inc, Milton Keynes, United Kingdom) was recorded with an automated sphygmomanometer, height was measured with a stadiometer while the subjects were shoeless, and weight was measured while the subjects were wearing minimum indoor clothing. Eligible subjects were then requested to make a 24-h urine collection for the measurement of urinary cotinine and microalbumin concentrations and to complete a 7-d dietary record before beginning the study. The subjects made another urine collection and completed a second dietary record at the end of the study.
The study participants received a modest financial reimbursement for their participation in the study and were provided, at regular intervals, some foods (yellow fat spreads, oil, and fish) for the dietary intervention period. The study protocol was reviewed and approved by the Human Research Ethics Committees of King's College London, East and North Hertfordshire Hospitals Local Research Ethics Committee, and North West Surrey Local Research Ethics Committee. The participants were informed fully about the nature of the study and gave written consent.
Study design
The study used a randomized, parallel, controlled design of 4 dietary treatments with a control; the duration of the intervention was 6 mo. The subjects were studied in 3 cohorts over a 3-y period. The diets were designed to maintain the intake of saturated and monounsaturated fatty acids constant and to provide 6% of energy from PUFAs with an n–6:n–3 of 10:1 (control), 5:1, or 3:1 when the n–3 fatty acids were provided predominantly as either -linolenic acid (18:3n–3) or long-chain n–3 LC-PUFAs (mainly EPA and DHA), or both. The basis and detailed description of the dietary intervention were described previously (15). The subjects were asked to make 7-d weighed food intake records at baseline and toward the end of the study, as described in detail elsewhere (15).
Fasting venous blood samples were collected twice at baseline and twice at the end of the study for the measurement of blood lipids. The subjects were asked to fast from 2200, and blood samples were collected between 0800 and 1000 on the next day. For the determination of insulin sensitivity and the response to a test meal, the subjects were advised to avoid foods high in fat on the previous day and to abstain from strenuous exercise; the subjects were given a list of foods to avoid and were provided with a frozen low-fat (<10 g fat) evening meal to facilitate compliance with the dietary advice. The next morning a cannula was inserted into a forearm vein, and 3 fasting blood samples were obtained at 5-min intervals for the measurement of glucose and insulin. The subjects then consumed a test meal providing 50 g fat, and additional blood samples were obtained 2, 3, and 6 h after the meal. After the 6-h blood sample was collected, 1500 IU heparin was administered intravenously, and blood samples were collected after 5 min into EDTA-containing tubes. Plasma was separated immediately by centrifugation (1500 x g, 15 min) and stored at –80 °C for measurement of postheparin plasma lipase activities. The test meal consisted of a muffin (85 g) and a strawberry-flavored milkshake (350 mL) that provided 50 g fat (14.8 g saturated fatty acids, 28.2 g oleic acid, 3.4 g linoleic acid, and 0.2 g linolenic acid) 17 g protein, and 75 g carbohydrate within 15 min, which was followed by a 200-mL glass of water. The subjects were provided with a standardized low-fat meal (1.7 MJ), which consisted of a piece of fruit, a low-fat yogurt (<1 g fat), and a glass of water, after the 3-h blood sample was taken. The subjects were advised to avoid strenuous activity throughout the study period but were allowed to leave the clinic to return to home or work between blood samples.
Laboratory methods
Blood was drawn into Vacutainer tubes (Becton Dickinson, Oxford, United Kingdom). Serum lipids were measured and liver function tests were conducted with the use of blood samples collected into Vacutainer tubes without anticoagulant (Vacutainer 368430; Becton Dickinson). The sample was centrifuged (1500 x g for 15 min at room temperature) immediately, and the serum was frozen and stored at –40 °C for the measurement of serum lipids and for the liver function tests at the Department of Clinical Biochemistry, King's College Hospital). For the determination of HDL and LDL size and NEFA and plasma apoprotein B concentrations, blood was collected into Vacutainer tubes containing EDTA (Vacutainer 17644; Becton Dickinson), and plasma was separated by low-speed centrifugation (1500 x g, 15 min, 4 °C) and stored at –80 °C pending analysis. Glucose and insulin measurements were made in fasting and 2-h postprandial blood samples collected into Vacutainer tubes containing fluoride oxalate (Vacutainer 367692; Becton Dickinson) and lithium heparin (Vacutainer 367681; Becton Dickinson), respectively, and plasma was separated by low-speed centrifugation (1500 x g, 15 min, 4 °C) and stored at –80 °C until analyzed. Total serum cholesterol, HDL, and triacylglycerol concentrations were measured by using fully enzymatic procedures with reagents from Wako (Neuss, Germany) on a Technicon DAX48 automated chemistry analyzer (Bayer Diagnostics, Newbury, United Kingdom). The postprandial measurements of plasma triacylglycerol, NEFA, apoprotein B, and glucose concentrations were made at the University of Surrey by enzymatic assays with the use of an autoanalyzer (ACE; Alfa Wassermann, Woerden, Netherlands) with reagents obtained from Randox Laboratories (Crumlin, United Kingdom). Plasma insulin concentrations were measured with an immunochemiluminometric assay (Molecular Light Technologies, Cardiff, United Kingdom). The precision (% SD/mean) of 4 controls over 3 assays were as follows: 7.7% at 47 pmol/L (6.5 mU/L), 4.2% at 151 pmol/L (21.0 mU/L), 3.6% at 603 pmol/L (84 mU/L), and 4.3% at 1213 pmol/L (169 mU/L). Total lipase and hepatic lipase activities were measured in the postheparin plasma (collected as above), and lipase activities were determined with commercially available kits that use a fluorescent triacylglycerol substrate (kit nos. PR2003 and PR2004, respectively; Technoclone, Hampshire, United Kingdom). Total lipoproteins (density < 1.22 kg/L) were isolated by ultracentrifugation in a fixed-angle rotor (70.1 Ti; Beckman Coulter, High Wycombe, United Kingdom), and the HDL subclasses were separated with the use of commercially available precast nondenaturing, polyacrylamide gradient (4–30%) gels with the use of a PGGE Pore Gradient Lipoprotein Electrophoresis System (C.B.S. Scientific Company, Del Mar, CA) (16). LDL density was determined as a close surrogate of LDL particle size on prestained plasma sample on an iodixanol gradient with a Beckman NVT 65 near-vertical rotor as previously described (17). Urinary cotinine was measured with an enzyme-linked immunosorbent assay with horseradish peroxidase–labeled cotinine (Cozart Diagnostics, Abingdon, United Kingdom). Urinary microalbumin was measured by using an immunoturbidometric assay, and creatinine concentrations were measured by using the Jaffé reaction on an Advia 1650 analyzer (Bayer Diagnostics, Newbury, United Kingdom). The homeostatic model assessment of insulin resistance (HOMA-IR) was calculated as the product of fasting glucose (mmol/L) x fasting insulin (U/L) divided by 22.5 (18). The revised quantitative insulin sensitivity test (RQUICKI) was calculated as the sum of the reciprocals of the log10 concentrations of insulin, glucose, and NEFAs.
Statistical analysis
Data were log normalized before the statistical analysis, and outliers (3SD outside the mean) were excluded (<1% sample) from the analysis to stabilize the variance. Changes in treatment were analyzed as the difference in log-transformed data and are expressed as the percentage increase or decrease with 95% CIs; probabilities were adjusted for multiple comparisons by using the Bonferroni correction factor. When appropriate, the probabilities were adjusted by analysis of covariance for smoking, BMI, age, sex, and hormone replacement therapy. Data for HDL size was analyzed by using the nonparametric Kruskal-Wallis and Mann Whitney U tests because of a significant difference in variances between groups. To examine the relation between n–6:n–3 and markers of coronary disease risk, the latter were compared against quartiles of n–6:n–3 across all 5 groups. All statistical analyses were performed by using SPSS version 12.01 (SPSS Inc, Chicago, IL).
RESULTS
A total of 354 subjects were screened, and 46 subjects were excluded because they did not meet the inclusion criteria. The remaining subjects were randomly assigned to 1 of 5 diets; subjects living together were allocated to the same treatment. Forty-four subjects withdrew from the study because of personal reasons (eg, family commitments, change in work circumstances) or the inability or unwillingness to comply with the study requirements, and 4 subjects withdrew for reasons of ill health (2 because of cancer, 1 because of mental illness, and 1 because of orthopedic surgery). A total 258 subjects (88 from Camberley, 141 from North Mymms, and 29 from King's College London) satisfactorily completed the 6-mo dietary intervention; subject characteristics are shown in Table 1. Most (76%) of the subjects had a serum cholesterol concentration >5.0 mmol/L, 16% had a low HDL concentration (<1.0 mmol/L for men and <1.2 mmol/L for women), 41% had a blood pressure >140 (systolic)/90 (diastolic) mm Hg, and 16% were smokers (smoking habits remained stable throughout the study on the basis of urinary cotinine excretion). Three subjects were taking lipid-lowering medication (2 simvastatin, 1 pravastatin), 7 were taking thyroxine medication, and 39 were receiving estrogen hormone replacement therapy (HRT)—34 of whom were taking oral estrogens and 5 of whom were using transdermal patches or implants. Thirteen of the subjects were taking medication for the treatment of hypertension. No statistically significant changes in the urinary microalbumin:creatinine ratios or the results of liver function tests were observed (data not shown).
View this table:
TABLE 1. Baseline characteristics of the subjects according to allocation to dietary treatment1
The dietary intakes of all subjects after the 6-mo dietary intervention are shown in Table 2. Energy intakes after the dietary intervention were not significantly different from intakes at baseline, but body weight increased on follow-up for the whole group (
View this table:
TABLE 2. Dietary intakes at baseline and after a 6-mo dietary intervention containing additional n–3 fatty acids as linolenic acid or n–3 long-chain polyunsaturated fatty acids (LC-PUFAs) or both1
View this table:
TABLE 3. Plasma glucose and insulin concentrations and homeostatic model assessment for insulin resistance (HOMA-IR) and revised quantitative insulin sensitivity check index (RQUICKI) at baseline and after a 6-mo dietary intervention containing additional n–3 fatty acids either as linolenic acid or n–3 long-chain polyunsaturated fatty acids (LC-PUFAs) or both1
The results of the lipoprotein analyses are shown in Table 4. No significant diet x sex interaction was observed for serum cholesterol (P = 0.015), and the diet x sex interaction remained significant after the exclusion of women receiving HRT (P = 0.01) with adjustments for age, BMI, smoking, and alcohol use. A significant interaction on follow-up was observed between HRT status and serum cholesterol in the groups that received additional n–3 LC-PUFAs (P = 0.002). A subgroup analysis showed a 10.6% (P = 0.004) increase in plasma LDL cholesterol in the women who received additional n–3 LC-PUFAs who were not receiving HRT. For the women receiving HRT, a nonsignificant trend was observed in the opposite direction. A significant follow-up effect, but no significant treatment effect, was observed for HDL cholesterol. However, a significant effect of diet on the change in the proportion of HDL2 was observed on follow-up (P = 0.02; Kruskal Wallis test); the values were significantly greater with the n–3 LC-PUFA + linolenic acid treatment than with the control treatment (P = 0.03). Fasting and postprandial serum triacylglycerols and the proportion of HDL2 after the diets with and without n–3 LC-PUFAs, adjusted for the intake of linolenic acid, are shown in Table 5. Both fasting serum triacylglycerol and postprandial lipid concentrations were significantly lower in the subjects who received additional n–3 LC-PUFAs than in those who did not. The proportion of sdLDLs was also significantly lower in this group of subjects, and there was a nonsignificant trend for the proportion of HDL2 to be greater in this group.
View this table:
TABLE 4. Serum lipoprotein concentrations and postheparin lipase activities at baseline and after a 6-mo dietary intervention containing additional n–3 fatty acids as linolenic acid or n–3 long-chain polyunsaturated fatty acids (n–3 LC-PUFAs) or both1
View this table:
TABLE 5. Fasting and postprandial serum triacylglycerol (TG) concentrations and proportion of HDL2 after a 6-mo dietary intervention with or without additional long-chain n–3 polyunsaturated fatty acids (n–3 LC-PUFAs)1
The proportion of sdLDLs was significantly greater in the men than in the women (32.1 ± 20.5% compared with 18.9 ± 13.3%; P < 0.0001) at baseline. Although the proportion of sdLDLs was significantly associated with serum triacylglycerol (r = 0.516, P < 0.0001), the ratio of triacylglycerol to HDL cholesterol (r = 0.712, P < 0.001), and hepatic lipase activity (r = 0.277, P < 0.001) after adjustment for age, sex, and BMI, the proportion of HDL2 was inversely associated with serum triacylglycerol (r = –0.328, P < 0.0001) and hepatic lipase activity (r = –0.306, P < 0.0001) after the same adjustments. Hepatic lipase activity was also significantly greater in the men than in the women (35.5 ± 10.8% compared with 30.0 ± 10.6%; P < 0.001) at baseline and was positively related to fasting insulin (r = 0.173, P = 0.03), HOMA-IR (r = 0.183, P = 0.0), and waist-hip ratio (r = 0.172, P = 0.03). Examination of the relation between quartiles of n–6:n–3 and markers of coronary disease risk (total, LDL, and HDL cholesterol; fasting and postprandial triacylglycerol; percentage of sdLDLs), irrespective of dietary group, showed no significant associations (data not shown).
DISCUSSION
The OPTILIP Study set out to determine the optimal ratio of dietary n–6 to n–3 PUFAs for cardiovascular health in the UK population with the use of surrogate risk markers for cardiovascular disease. The study focuses on indexes of insulin sensitivity, postprandial lipemia, and lipoprotein particle size. The rationale for selecting an older population of men and women was because of their substantial risk of cardiovascular events. Most of the subjects (89%) had one or more of the established risk factors for IHD [serum cholesterol >5.0 mmol/L, blood pressure >140 (systolic)/90 (diastolic), cigarette smoking], and 60% of the subjects had impaired fasting glucose according to the revised definition based on a cutoff 5.5 mmol/L (19). On the basis of the definition for the metabolic syndrome proposed by the World Health Organization (3), 28% of the subjects had the metabolic syndrome, which has been associated with a 37% attributable risk of IHD in nondiabetic persons aged >50 y (20). BMIs were slightly lower than those reported in the National Dietary and Nutritional Survey (21) for this age group, and the prevalence of cigarette smoking was also lower. Although a substantial number of the subjects were taking medication, this overall pattern of characteristics is typical of this age group, consequently, the results reported herein can be broadly generalized to this segment of the population. A unique feature of the OPTLIP Study was that it used a food-based intervention rather that dietary supplements or a single item of food (eg, margarine). It also had a relatively long intervention period (ie, 6 mo) and a large sample size compared with previous studies (13, 14, 22, 23). This study reports total intakes of -linolenic acid, EPA, and DHA rather than supplemental intakes, as reported in other studies that did not take background intakes into account. The ratios of n–6 to n–3 fatty acids that were achieved with each diet were close to the predicted target values, whereas further evidence for compliance with the dietary advice was provided by increases in the proportions of EPA and DHA in erythrocyte membrane phospholipids. At baseline, the respective intakes in men and women were 1.3 and 1.0 g for -linolenic acid and 0.5 and 0.4 g/d for n–3 LC-PUFAs; the proportion of the dietary energy derived from -linolenic acid (0.5%) and n–3 LC-PUFAs (0.2%) was not different between sexes. After the dietary intervention, intakes of -linolenic acid ranged from 0.5% to 1.1% of energy and those for n–3 LC-PUFAs ranged from 0.2% to 0.7% of energy. The intake of n–3 LC-PUFAs achieved, 1g/d, was not different from that achieved through dietary supplementation in the GISSI (Gruppo Italiano per lo Studio della Sopravvivenza nell'Infarto Miocardico) trial (24) and has subsequently been adopted as an intake that may confer protection against IHD. Although these dietary intakes were substantially different from the subjects' background diet, -linolenic acid intakes were much lower than those in other studies that used flaxseed oil or margarine made from flaxseed oil and achieved intakes between 1.5% and 15% of energy (9, 22, 25). However, these intakes represent the maximum changes likely to be achieved through dietary advice to the general population in the United Kingdom and are of practical relevance because they reflect recent dietary guidelines in Europe (26) and the Unites States (27).
There was no evidence of a dietary effect of n–6:n–3 or of n–3 fatty acids on measures of insulin resistance. This finding is consistent with other published studies that used dietary supplements (13, 14) and is of particular relevance because 60% of the study population had impaired fasting glucose concentrations. Moreover, there is no consistent evidence in the literature to suggest that n–3 fatty acids exert any specific effects on insulin sensitivity in humans, over and above that of other poly- or monounsaturated fatty acids at the intakes consumed in typical human diets.
A review of the effects of n–3 fatty acids on plasma lipids by the British Nutrition Foundation Task Force on Unsaturated Fatty Acids (28) concluded that there was consistent evidence of a lowering of fasting and postprandial triacylglycerol concentrations with intakes of n–3 LC-PUFAs in excess of 2–3 g /d and noted that, in some subgroups, LDL-cholesterol concentrations increased. It has also been shown that the duration and magnitude of postprandial lipemia after a fat-containing meal are decreased in subjects whose diet was modified to contain 5 g n–3 LC-PUFAs (8). However, there has been a lack of reliable information on the effects of n–3 LC-PUFA intakes in the order of 1 g/d, which is the intake associated with a decreased risk of IHD and that promoted in dietary guidelines. Cross-sectional studies have reported an inverse association between the dietary intake of -linolenic acid and plasma triacylglycerol (29), although dietary interventions at intakes of between 4 and 20 g have, in general, failed to reproduce the triacylglycerol-lowering effects of n–3 LC-PUFAs (9, 23, 25, 30). Modification of the n–6:n–3 can be achieved in several ways; by either raising n–3 and lowering n–6 or by making disproportionate adjustments in both of these components. In the present study, only small differences in n–6 PUFAs were observed between diets (1–2% of energy), with the total mass of PUFAs being held constant at 6% of energy. Changes in the ratio were achieved, in the main, by the addition of -linolenic acid and long-chain n–3 LC-PUFAs. The results indicate that dietary advice to modify the n–6:n–3 by increasing the intake of -linolenic acid will have no effect on fasting or postprandial triacylglycerol concentrations. However, the addition to the diet of 0.5% of energy from n–3 LC-PUFAs resulted in a modest decrease in fasting triacylglycerol and a decrease in the incremental area under the curve for plasma triacylglycerol. This change appeared to be most marked in men allocated to the n–3 LC-PUFA + linolenate diet. A previous study, which tested a similarly low intake of n–3 LC-PUFAs, was unable to show such an effect, probably as a consequence of its small sample size (22).
It is well established that a plasma triacylglycerol concentration >1.5 mmol/L is associated with an increased proportion of sdLDLs and that dietary supplementation with 3 g n–3 LC-PUFAs/d decreases the proportion of sdLDLs, particularly in men with an atherogenic lipoprotein phenotype (23). In the present study, only 8.1% of the subjects had an atherogenic lipoprotein phenotype (triacylglycerol >1.7 mmol/L, sdLDL > 50%, and HDL cholesterol <1.05 mmol/L in men and < 1.29 mmol/L in women). Exclusion of these subjects from the statistical analysis did not alter the significance of the finding that the proportion of sdLDLs decreased after consumption of the diets containing the additional n–3 LC-PUFAs. A tendency for LDL to increase, particularly in women not receiving HRT, with an increased intake of n–3 LC-PUFAs was observed. Two recent studies have shown that low intakes of DHA in the range of 0.7 –1.5 g/d increase LDL cholesterol (31, 32). Possible explanations for this effect are that n–3 LC-PUFAs promote the formation of a smaller VLDL that is preferentially converted to LDL and through an increase in LDL size (33). The reduction in plasma triacylglycerol was also accompanied by an increase in the proportion of HDL2 with no change in total HDL-cholesterol concentration, which agrees with previous reports. In the present study, plasma triacylglycerol concentration and hepatic lipase activity were correlated with the proportions of sdLDL and HDL2. This finding is consistent with the role of hepatic lipase in remodeling LDL and HDL into smaller and denser fractions found in an atherogenic lipoprotein phenotype (34).
In conclusion, dietary advice to decrease the n–6:n–3, chiefly by altering the mass of n–3 PUFAs, does not influence insulin sensitivity or postheparin plasma lipase activities in older men and women. However, increasing the dietary intake of n–3 LC-PUFAs from 0.2% to 0.7% of energy (1 g/d) from foods promotes favorable alterations in lipoprotein particle size that can be attributed to a decrease in basal and postprandial plasma triacylglycerol.
ACKNOWLEDGMENTS
We thank Roy Sherwood for assistance with the clinical chemistry, Robert Gray for technical assistance, and all of the study subjects for their time and dedication to the trial.
TABS, DJM, and GJM were responsible for the original concept and study design. BAG and LMM were coinvestigators responsible for the study design and day-to-day management of the study. MDG, FL, and SS were in charge of the subject recruitment, dietary intervention, postprandial tests, and data collection. IGD conducted the lipoprotein fraction analyses. JAC was responsible for the statistics and randomization. TABS and BAG were the lead writers. All authors helped to refine the manuscript. None of the authors had any personal or financial conflict of interest.
REFERENCES