Changes in LDL particle composition after the consumption of meals containing different amounts and types of fat

¹ From the Oxford Lipid Metabolism Group, the Oxford Centre for Diabetes, Endocrinology and Metabolism, Radcliffe Infirmary, Oxford, United Kingdom.

² Supported by a contract from the Ministry of Agriculture, Fisheries and Food (LKMS and HB) and by a Medical Research Council–Glaxo-Wellcome Collaborative Research Studentship (JC).

³ Reprints not available. Address correspondence to KN Frayn, Oxford Lipid Metabolism Group, Radcliffe Infirmary, Oxford OX2 6HE, United Kingdom. E-mail: keith.frayn{at}oxlip.ox.ac.uk.

ABSTRACT  
Background: Remodeling of lipoprotein particles in the postprandial period is considered to be an important source of atherogenic particles, but acute changes occurring after meals have been little studied.

Objective: We sought to characterize changes in LDL particle composition occurring after a single meal, with particular reference to potential lipid exchange with particles carrying dietary fatty acids.

Design: In a balanced design, 8 healthy subjects ingested isoenergetic meals of different fat content: low-fat, rich in saturated fatty acids (SFAs), and rich in polyunsaturated fatty acids (PUFAs). We investigated changes in LDL composition 4 and 6 h after meal ingestion.

Results: The LDL triacylglycerol-to-protein ratio closely mirrored the plasma triacylglycerol concentrations after each of the meals, and there was a strong association between these variables in both the fasting and postprandial states (P < 0.001). A postprandial increase in LDL triacylglycerol was associated with a decrease in LDL cholesterol. There were no effects of the ingestion of a single meal on the LDL density profiles for protein or for any of the lipid components. The fatty acid composition of total LDL lipids changed in the postprandial period, with an enrichment in PUFA after the PUFA-rich meal and in SFA after the SFA-rich meal.

Conclusions: The changes observed in LDL composition after single meals are in accord with the proposition that there is neutral lipid exchange in the postprandial period, with triacylglycerol enrichment of LDL particles at the expense of cholesteryl esters. The change in the fatty acid composition of LDL particles implies significant lipid exchange with particles containing dietary fat.Am J Clin Nutr 2002;76:–50.

Key Words: Postprandial lipemia • cholesteryl ester transfer protein • triacylglycerol • dietary fatty acids • LDL

INTRODUCTION  
It is now recognized that the atherogenic potential of hypertriglyceridemia is mediated in part through alterations in the density profile of LDL particles. Elevated triacyglycerol concentrations are associated with a predominance of smaller, dense LDL particles that are more susceptible to oxidation and are more closely associated with atherosclerosis than are larger, buoyant LDL particles (1–3). The remodeling of LDL particles may be associated particularly with elevations in triacylglycerol concentrations in the postprandial period (4,5). Neutral lipid exchange, mediated by the cholesteryl ester transfer protein, has been postulated to lead to triacylglycerol enrichment of LDL particles at the expense of cholesteryl ester. Subsequent hydrolysis of the triacylglycerol may then lead to the formation of smaller, lipid-depleted LDL particles (4,6).

Despite the potential importance of the postprandial period in LDL remodeling, there have been few studies of the acute changes in LDL particles that might occur in the period after a single meal, especially a meal of relatively normal fat content. Redgrave and Carlson (7) found no change in LDL particle size 6 h after the ingestion of 100 g soybean oil. Attia et al (8) showed a shift in the distribution of LDL classes toward larger particles in control subjects and toward smaller particles in subjects with type 2 diabetes 8 h after a meal containing 80 g fat. Ashida et al (9) made semiquantitative observations of a decrease in LDL particle diameter in 3 men given 150 g fat, and Pirro et al (10) found a decrease in LDL particle size after a high-fat meal (65 g/m² surface area) only in women with fasting hypertriglyceridemia. In studies with meals of relatively normal fat content, Dubois et al (11) found enrichment of LDL with triacylglycerol and depletion of cholesteryl esters 3 h after meals, with the extent of the change reflecting the fat content of the test meal ( 50 g). They also found a strong positive relation between the elevation in plasma triacylglycerol concentration and the triacylglycerol enrichment of LDL, which mirrors our findings in a preliminary study (12).

In the present study, we investigated in more detail the acute changes in LDL particle composition after meals of differing fat content (60 or 5 g) and fat type. Our hypothesis was that modifications of LDL composition would occur in the postprandial period and that these would reflect an accumulation of meal-derived fatty acids in LDL particles. We studied LDL composition after meals of different fat content to obtain specific information about the involvement of dietary fatty acids in lipid exchange.

Some of the data have been published previously in abstract form (12,13). Measurements of the blood flow in adipose tissue, made during the same experiments, have been reported elsewhere, along with the concentrations of plasma hormones and fatty acids (14).

SUBJECTS AND METHODS  
Subjects
Eight healthy, normolipidemic volunteers (4 men) were studied on 3 occasions in a balanced design. Their characteristics are given in Table 1. Four weeks elapsed between occasions for all subjects, except for one man who for unavoidable reasons completed the entire study within 2 wk and one man who consumed 2 of the meals 1 wk apart. Subjects were instructed to consume a low-fat meal the evening before the study and to refrain from smoking, strenuous exercise, and the consumption of alcohol and caffeine for 24 h before the study. The study was approved by the Central Oxford Research Ethics Committee, and all subjects gave informed consent.

View this table:
TABLE 1 . Characteristics of the subjects¹  
Protocol
Having consumed no food and having drunk only water for the previous 12 h, the subjects arrived at the laboratory at 0830 on each study day. To obtain arterialized blood, we inserted a cannula in a retrograde direction into a vein in the subject’s hand, which was warmed in a box at 65 °C. Two basal samples were taken 30 min apart, and the test meal was consumed immediately after the second of these draws. The composition of the 3 isoenergetic test meals is shown in Table 2. Two were high in fat: one in saturated fatty acids (SFAs) and the other in polyunsaturated fatty acids (PUFAs). The third meal was low in fat. LDL was isolated from the samples taken 30 min before and 240 and 360 min after the consumption of the test meal. Subjects rested on a bed for the duration of the study, engaging only in very light activities such as reading or watching television.

View this table:
TABLE 2 . Composition of test meals¹  
Plasma metabolites
Blood was collected in heparinized syringes (Monovette; Sarstedt, Leicester, United Kingdom) for the measurement of plasma glucose and triacylglycerol and into evacuated tubes containing EDTA (Becton Dickinson UK Ltd, Abingdon, United Kingdom) for LDL isolation. Plasma was immediately separated at 4 °C by centrifugation for 10 min at 1800 x g. Plasma glucose concentrations were measured by enzymatic methods (IL Test Glucose; Instrumentation Laboratory, Warrington, United Kingdom) on a Monarch centrifugal analyzer (Instrumentation Laboratory). Plasma triacylglycerol concentrations were also measured enzymatically, with correction for free glycerol (15). Measurements of plasma fatty acids and insulin have been reported elsewhere (14).

LDL separation and analysis
Plasma for LDL isolation was centrifuged immediately after it was collected to remove a chylomicron-rich fraction. Duplicate 2-mL portions of plasma were layered under a solution of sodium chloride with a density of 1.006 kg/L and centrifuged at 45 900 x g (average) for 30 min at 4 °C in an SW55Ti rotor (Beckman Coulter UK Ltd, High Wycombe, United Kingdom). After aspiration of the chylomicron-rich fraction, LDL was isolated by ultracentrifugation with the use of a method based on that of Griffin et al (16). The chylomicron infranatant fluid (1.5 mL) was removed, and its density was adjusted to 1.090 kg/L by the addition of solid potassium bromide. The density-adjusted infranatant fluid and a 7-step sodium chloride–potassium bromide gradient were introduced sequentially into polyvinyl alcohol-coated UltraClear tubes (Beckman Instruments Inc, Palo Alto, CA) (17) over a cushion of density 1.186 kg/L by a peristaltic pump (Watson-Marlow, Falmouth, United Kingdom). Density solutions contained 0.04% Na₂EDTA and 0.1% NaN₃. All densities were checked with a digital densitometer (Paar Scientific Ltd, London). The gradient was centrifuged in an SW55Ti rotor at 287 000 x g (average) for 19 h at 23 °C with slow acceleration and no braking. The resulting density gradient containing LDL was displaced upward from the tube by an inert, dense, hydrophobic material (Fluorinert FC-40; Sigma-Aldrich Chemical Co, Poole, United Kingdom) that was introduced under the plasma layer by a peristaltic pump. Sixteen 200-µL fractions were collected over the density range 1.020–1.063 kg/L. Agarose gel electrophoresis showed that fractions 4–13 collected after ultracentrifugation contained LDL that was uncontaminated by other lipoprotein particles. These fractions cover the density range 1.021–1.052 kg/L.

The concentrations of cholesterol, triacylglycerol, and phospholipid in each of the LDL fractions were measured by enzymatic colorimetric methods (Sigma Diagnostics cholesterol reagent from Sigma-Aldrich Chemical Co and a phospholipids kit from the Diagnostics and Biochemicals Division of Boehringer Mannheim UK Ltd, Lewes, United Kingdom) on a Monarch centrifugal analyzer. The protein concentrations were measured by a bicinchoninic acid-based colorimetric method (BCA protein assay reagent kit; Pierce & Warriner, Chester, United Kingdom).

The fatty acid composition of all LDL lipids was determined with gas chromatography by using the combined fractions. The lipids were first extracted into chloroform:methanol (2:1, vol:vol), and the fatty acids were then transesterified to methyl esters by incubation with methanolic sulfuric acid at 70 °C. Methyl esters of fatty acids were analyzed by capillary gas chromatography (Chrompack UK Ltd, Millharbour, United Kingdom). The fatty acid composition of the test meals was measured in the same way after a portion of each meal had been homogenized in a Waring blender and dissolved in chloroform.

Calculations and statistics
Plasma concentrations of LDL components were calculated as the area under the curve (AUC) from fractions 4–13 inclusive, corrected for dilution factors to give concentrations per milliliter of plasma. The midpoint of the AUC was used as an estimate of any shift in density of the LDL profile and was calculated as the density that split the AUC into halves.

Changes in plasma and LDL constituents at different time points were analyzed by repeated-measures analysis of variance. When significant effects were found, post hoc comparisons were made with correction for multiple comparisons by using the Bonferroni method. When there were multiple samples per subject, correlations between variables were analyzed by using analysis of covariance with the subject as a fixed factor (18). Calculations were carried out with SPSS for WINDOWS, release 8.0 (SPSS Inc, Chicago). For clarity, only significant results are listed in the figure legends.

RESULTS  
Plasma metabolites
The mean plasma triacylglycerol concentration increased after ingestion of both high-fat meals and peaked at 3 h (Figure 1). Approximately one-half of the increase in plasma triacylglycerol was attributable to an increase in chylomicron triacylglycerol (Figure 1). There was only a gradual increase in the plasma triacylglycerol concentration after the low-fat meal, and the effect of this meal on the plasma triacylglycerol concentration was significantly different from that of the high-fat meals (Figure 1). Despite differing baseline values, the time courses of the increase in the mean plasma triacylglycerol concentration from fasting values were not significantly different after the SFA- and PUFA-rich meals.

View larger version (21K):
FIGURE 1. . Mean (± SEM) plasma (closed symbols) and chylomicron (open symbols) triacylglycerol concentrations in 8 subjects after each of the 3 test meals: squares, meal rich in polyunsaturated fatty acids (PUFAs); circles, meal rich in saturated fatty acids (SFAs); triangles, low-fat meal. Repeated-measures ANOVA showed the following: plasma triacylglycerol, a significant effect of time (P < 0.001) and a significant meal-by-time interaction (P < 0.001); chylomicron triacylglycerol, a significant effect of time (P < 0.001), a significant effect of meal (P = 0.002), and a significant meal-by-time interaction (P < 0.001). Post hoc tests showed the following significant effects: plasma triacylglycerol, a significant meal-by-time interaction for the PUFA-rich meal compared with the low-fat meal (P < 0.003) and for the SFA-rich meal compared with the low-fat meal (P < 0.003); chylomicron triacylglycerol, a significant effect of meal for the PUFA-rich meal compared with the low-fat meal (P < 0.003) and for the SFA-rich meal compared with the low-fat meal (P = 0.024) and a significant meal-by-time interaction for the PUFA-rich meal compared with the low-fat meal (P < 0.003) and for the SFA-rich meal compared with the low-fat meal (P = 0.003).

 
LDL particle composition
Concentrations of LDL lipid components were analyzed with respect to LDL protein (Figure 2). There were significant changes with time in both LDL triacylglycerol and LDL cholesterol when expressed in this way, but not in LDL phospholipid.

View larger version (34K):
FIGURE 2. . Mean (± SEM) plasma triacylglycerol (TG) concentrations and molar ratios of LDL lipid components to LDL protein by diet type; n = 8. , baseline; , 240 min after the meal; , 360 min after the meal. Repeated-measures ANOVA showed the following: plasma TG, see the legend to Figure 1; LDL TG:protein, a significant effect of time (P = 0.015) and a significant meal-by-time interaction (P = 0.011); LDL cholesterol:protein, a significant effect of time (P = 0.038); LDL phospholipid:protein, no significant effects. PUFA, polyunsaturated fatty acid; SFA, saturated fatty acid.

 
The shapes of the density profiles of all LDL components were similar. There were no effects of meal type or time on the midpoints of the AUCs of the LDL density profiles for protein or for any of the lipid components. Mean LDL protein density profiles before and after the PUFA-rich meal are shown as an example (Figure 3).

View larger version (14K):
FIGURE 3. . Mean LDL protein density profiles before and after the meal rich in polyunsaturated fatty acids; n = 8. , 30 min before the meal; •, 240 min after the meal; , 360 min after the meal.

 
Relations between LDL composition and plasma analytes
The LDL triacylglycerol-to-protein ratio closely mirrored the plasma triacylglycerol concentrations after each of the meals (Figure 2, left panels). These 2 variables were positively correlated when all data points were included (r² = 0.68, P < 0.001 by analysis of covariance). This correlation was also evident even when only the data representing the fasting state were examined (r² = 0.73, P < 0.001 by analysis of covariance), which suggested that the day-to-day and interindividual variations in plasma triacylglycerol are accompanied by corresponding variations in LDL triacylglycerol:protein. The correlation was maintained during postprandial lipemia.

As was found in a pilot study (12) and by Dubois et al (11), there was a positive correlation between the postprandial increase in the plasma triacylglycerol concentration and that in LDL triacylglycerol:protein (r² = 0.35, P = 0.001 by analysis of covariance).

Interrelations among LDL components
Triacylglycerol as a molar percentage in LDL correlated negatively with the molar percentage of LDL cholesterol. The molar percentage of LDL phospholipid did not show this correlation with LDL cholesterol. The data for the PUFA-rich meal are shown in Figure 4. In addition, the molar percentage of LDL triacylglycerol was found to correlate very strongly and linearly with LDL triacylglycerol:protein (r² = 0.99, P < 0.001), whereas the molar percentage of LDL cholesterol correlated less strongly with LDL cholesterol:protein (r² = 0.76, P < 0.001 by analysis of covariance). Similarly strong inverse relations between LDL triacylglycerol and LDL cholesterol content were seen after the other meals (Figure 4).

View larger version (19K):
FIGURE 4. . Interrelations between LDL components expressed as molar percentages. Each subject is distinguished by a different symbol; there are 3 points for each subject, representing values at baseline and at 4 and 6 h after the meal. Solid symbols show the data for triacylglycerol (TG). Overall relations for the meal rich in polyunsaturated fatty acids (PUFAs) were significant for TG (r² = 0.97, P < 0.001) and protein (r² = 0.63, P = 0.002), but not for phospholipid (P = 0.29), by analysis of covariance. Overall relations for the meal rich in saturated fatty acids (SFAs) and for the low-fat meal were significant for TG (r² = 0.77, P = 0.001 and r² = 0.97, P < 0.001, respectively).

 
LDL fatty acid composition
The fatty acid composition of total LDL lipids changed after the consumption of the high-fat test meals. The data for the major fatty acids present in LDL and in the test meals are shown in Table 3. There was an enrichment of LDL with PUFA at the expense of SFA after the PUFA-rich meal, with the opposite occurring after the SFA-rich meal. These changes are illustrated in Figure 5, which shows the PUFA-to-SFA ratios of total LDL fatty acids 30 min before and 240 and 360 min after the meal and a significant meal-by-time interaction.

View this table:
TABLE 3 . Fatty acid composition of total lipid from LDL and from test meals as percentage of total fatty acids by weight¹  

View larger version (18K):
FIGURE 5. . Mean (± SEM) ratios of polyunsaturated fatty acids (PUFAs) to saturated fatty acids (SFAs) of total LDL lipids by weight. n = 7 for the PUFA- and SFA-rich meals, and n = 5 for the low-fat meal. Two-way repeated-measures ANOVA showed a significant main effect of time (P < 0.01) and a significant meal-by-time interaction (P = 0.015). , 30 min before the meal; , 240 min after the meal; , 360 min after the meal.

 

DISCUSSION  
We have shown that there are significant changes in LDL particle composition in the period after fat-rich meals and that these changes differ from those seen after a low-fat meal. The lack of change in LDL protein with time suggests that the number of LDL particles present in the blood is not affected at this stage after the consumption of a meal, and the lack of change in density profiles suggests that the effects we observed are due to lipid exchange rather than to net accumulation. In keeping with the postulated effects of cholesteryl ester transfer protein in the postprandial period, LDL particles accumulated triacylglycerol but lost cholesterol. The changes were not quantitatively different after meals rich in SFAs or PUFAs, although the nature of LDL fatty acids changed to reflect the fatty acids in the meal.

The close inverse relation observed between LDL cholesterol and triacylglycerol, when each was expressed as a molar percentage in the particle, was interesting and is a new observation as far as we are aware. It might seem obvious that, as the percentage of one constituent increases, the percentage of another must decrease, but the fact that no such relation existed for LDL protein and phospholipid suggests a specific relation between the particle contents of cholesterol and triacylglycerol. This seems to be further powerful evidence for the specific exchange of neutral lipids with other particles in vivo. In this context, it would have been valuable to estimate separately the free and the esterified cholesterol contents of LDL, but, with multiple analyses to perform, we did not have sufficient sample for that step. Nevertheless, our results for total LDL cholesterol show very clearly the relation with LDL triacylglycerol.

Because we did not see a transformation to smaller, denser LDL particles, we have presumably observed only the first stages in their creation. The triacylglycerol-enriched particles that accumulated during the postprandial period could later be substrates for the triacylglycerol-lipolytic activity of hepatic lipase (4,6). It is interesting that in the studies of Attia et al (8) there were shifts in the distribution of LDL subclasses 8 h after a high-fat meal. This might suggest that the action of hepatic lipase becomes more prevalent in the period 6–8 h after meals. Certainly, triacylglycerol enrichment in LDL reached a peak in our study coincident with the peak in plasma triacylglycerol concentrations, and the subjects in the study of Attia et al consumed a somewhat larger amount of fat (70 g, as compared with 60 g in our study). In that respect, it is also interesting that the accumulation of triacylglycerol in LDL was not consistent over the postprandial period; rather, it reached a peak and then declined. Possibly, therefore, we were observing a reversible exchange of triacylglycerol and cholesterol between lipoprotein species. It must be remembered in this context that our subjects were healthy and normolipidemic, and thus the accumulation of small, dense LDL particles was unlikely. It could also be that, as smaller particles were formed, they were efficiently removed by receptors.

The changes we observed were relatively small: the PUFA-to-SFA ratio in LDL on average increased by 8% after the PUFA-rich meal and decreased by 9% after the SFA-rich meal. Nevertheless, our hypothesis would be that small changes occurring after every meal might eventually lead, in those at risk, to large cumulative effects.

There has been considerable discussion of the major particles involved in neutral lipid exchange in the postprandial period. Lassel et al (19) suggested that large VLDL particles were more important donors of cholesteryl ester to LDL than were chylomicrons in the postprandial period. Other work suggests that, among the triacylglycerol-rich lipoproteins, participation in neutral lipid exchange is governed more by the amount of triacylglycerol per particle than by the number of particles (20) and thus that chylomicrons might be more important than VLDL. In the present studies, the change in LDL fatty acids to reflect the fatty acids in the meal suggests the movement of triacylglycerol mainly from the chylomicron fraction. Cholesteryl esters are themselves enriched in PUFA, and thus the loss of cholesteryl ester through the cholesteryl ester transfer protein mechanism will result in a loss of PUFA. Therefore, it is probably not surprising that the increase in the PUFA-to-SFA ratio after the PUFA-rich meal was not significant, although the difference from the decrease in the ratio seen after the SFA-rich meal was highly significant. This observation might also reflect a mechanism for restraining PUFA accumulation in LDL lipids. That would be beneficial, because LDL particles that are highly enriched in PUFAs are usually found to be better substrates for lipid peroxidation (21,22). It would, of course, have been ideal to specifically separate LDL triacylglycerol fatty acids for analysis, but the amount of sample available made it necessary to use total LDL fatty acids.

In conclusion, we have shown acute changes in LDL composition in the period after meals. Although the changes we observed in healthy, normolipidemic subjects are not likely to lead to the production of more atherogenic particles, it could easily be envisaged that more prolonged or exaggerated postprandial lipemia occurring after successive meals, as is seen in many subjects at increased risk of coronary artery disease (4,23), could lead to the changes in LDL particle density that are associated with atherogenesis.

ACKNOWLEDGMENTS  
We thank Sandy Humphreys for statistical advice.

REFERENCES  

1. Austin MA, Breslow JL, Hennekens CH, Buring JE, Willett WC, Krauss RM. Low-density lipoprotein subclass patterns and risk of myocardial infarction. JAMA 1988;260:1917–21.
2. Chait A, Brazg RL, Tribble DL, Krauss RM. Susceptibility of small, dense low-denstiy lipoproteins to oxidative modification in subjects with the atherogenic lipoprotein phenotype, pattern B. Am J Med 1993;94:350–6.
3. Griffin BA, Freeman DJ, Tait GW, et al. Role of plasma triacylglycerol in the regulation of plasma low density lipoprotein (LDL) subfractions: relative contribution of small, dense LDL to coronary heart disease risk. Atherosclerosis 1994;106:241–53.
4. Miesenböck G, Patsch JR. Postprandial hyperlipidemia: the search for the atherogenic lipoprotein. Curr Opin Lipidol 1992;3:196–201.
5. Karpe F, Tornvall P, Olivecrona T, Steiner G, Carlson LA, Hamsten A. Composition of human low density lipoprotein: effects of postprandial triacylglycerol-rich lipoproteins, lipoprotein lipase, hepatic lipase and cholesteryl ester transfer protein. Atherosclerosis 1993;98:33–49.
6. Lagrost L, Gambert P, Lallemant C. Combined effects of lipid transfers and lipolysis on gradient gel patterns of human plasma LDL. Arterioscler Thromb 1994;14:1327–36.
7. Redgrave TG, Carlson LA. Changes in plasma very low and low density lipoprotein content, composition, and size after a fatty meal in normo- and hypertriglyceridemic man. J Lipid Res 1979;20:217–29.
8. Attia N, Durlach V, Paul JL, Soni T, Betoulle D, Girard Globa A. Modulation of low density lipoprotein subclasses by alimentary lipemia in control and normotriglyceridemic non-insulin-dependent diabetic subjects. Atherosclerosis 1995;113:197–209.
9. Ashida Y, Takagi A, Ikeda Y. Improved method for non-denaturing polyacrylamide gradient gel electrophoresis for detection of small-sized LDL produced during postprandial hypertriglyceridaemia. Scand J Clin Lab Invest 1999;59:663–70.
10. Pirro M, Lupattelli G, Siepi D, et al. Postprandial lipemia and associated metabolic disturbances in healthy and hyperlipemic postmenopausal women. Metabolism 2001;50:330–4.
11. Dubois C, Beaumier G, Juhel C, et al. Effects of graded amounts (0–50 g) of dietary fat on postprandial lipemia and lipoproteins in normolipidemic adults. Am J Clin Nutr 1998;67:31–8.
12. Callow J, Frayn KN. Enrichment of low-density lipoprotein with triacylglycerol after a high fat meal. Int J Obes Relat Metab Disord 1996;20(suppl):151 (abstr).
13. Callow J, Summers LKM, Frayn KN. Acute changes in low-density lipoprotein composition after consumption of meals of varying fat composition. Atherosclerosis 1999;145:434 (abstr).
14. Summers LKM, Callow J, Samra JS, Macdonald IA, Matthews DR, Frayn KN. The effect on adipose tissue blood flow of isoenergetic meals containing different amounts and types of fat. Int J Obes Relat Metab Disord 2001;25:1294–9.
15. Humphreys SM, Fisher RM, Frayn KN. Micro-method for measurement of sub-nanomole amounts of triacylglycerol. Ann Clin Biochem 1990;27:597–8.
16. Griffin BA, Caslake MJ, Yip B, Tait GW, Packard CJ, Shepherd J. Rapid isolation of low density lipoprotein (LDL) subfractions from plasma by density gradient ultracentrifugation. Atherosclerosis 1990;83:59–67.
17. Holmquist L. Surface modification of Beckman Ultra-Clear centrifuge tubes for density gradient centrifugation of lipoproteins. J Lipid Res 1982;23:1249–50.
18. Bland JM, Altman DG. Correlation, regression, and repeated data. BMJ 1994;308:896.
19. Lassel TS, Guerin M, Auboiron S, Chapman MJ, Guy Grand B. Preferential cholesteryl ester acceptors among triacylglycerol-rich lipoproteins during alimentary lipemia in normolipidemic subjects. Arterioscler Thromb Vasc Biol 1998;18:65–74.
20. Fisher RM, Coppack SW, Gibbons GF, Frayn KN. Post-prandial VLDL subfraction metabolism in normal and obese subjects. Int J Obes Relat Metab Disord 1993;17:263–9.
21. Mata P, Alonso R, Lopez-Farre A, et al. Effect of dietary fat saturation on LDL oxidation and monocyte adhesion to human endothelial cells in vitro. Arterioscler Thromb Vasc Biol 1996;16:1347–55.
22. Hau M-F, Smelt AHM, Bindels AJGH, et al. Effects of fish oil on oxidation resistance of VLDL in hypertriglyceridemic patients. Arterioscler Thromb Vasc Biol 1996;16:1197–202.
23. Karpe F. Postprandial lipid metabolism in relation to coronary heart disease. Proc Nutr Soc 1997;56:671–8.