Current Medical Research and Opinion (1996), 13, No. 7, 379-390

Gold Horizontal Line

LDL subfractions and atherogenicity:
an hypothesis from the
University of Glasgow

C. J. Packard, MRCPath, DSc

Professor of Pathological Biochemistry, Department of Pathological Biochemistry, Glasgow University,
Glasgow G4 0SF,Scotland

Accepted: 8th April 1996

Gold Horizontal Line


The variation in the size and atherogenicity of the low density lipoprotein (LDL) particles has attracted a great deal of recent attention. In particular, attention has focused on the role of plasma triglyceride concentrations in driving the lipoprotein exchange that determines the concentration of the smaller, denser, more atherogenic LDL fraction. In a study at Glasgow University, researchers analysed the distribution of LDL subfractions among normocholesterolaemic men, with or without coronary artery disease, survivors of myocardial infarction, and normal controls. The results showed that the risk of coronary artery disease or myocardial infarction is considerably greater in those groups with higher plasma concentrations of small, dense LDL. In a second study, eight patients with hypercholesterolaemia were treated with fenofibrate. Radioisotope tracers showed that fenofibrate shifts the distribution of LDL subfractions from small, dense, atherogenic particles towards larger, lighter, less atherogenic ones. The efficacy of fenofibrate derives from its hypotriglyceridaemic activity. Triglycerides may have further atherogenic and thrombogenic effects: they may cause endothelial cell dysfunction in the artery wall, stimulating the recruitment of macrophages into the endothelium. They may also promote the synthesis of thrombogenic mediators, suppressing local plasmin synthesis and accelerating intra-arterial fibrin deposition. This evidence has led to an increasing recognition of the central role of triglycerides in the process of atherogenesis.

Key Words

Fibrates - Triglycerides - LDL-subfractions - Atherogenesis - Cholesterol ester transfer protein


The standard classification of lipoproteins assigns molecules to either the HDL, LDL, IDL or VLDL families, according to their density. However, the variation within each of the traditional divisions is considerable, and that of the LDL particles has recently attracted much attention, for two reasons. Firstly, the distribution of the LDL subfractions depends on a dynamic lipoprotein exchange driven by plasma triglycerides, and studies have demonstrated an epidemiological link between plasma triglyceride levels and coronary disease.1 Secondly, LDL density appears to have a direct correlation with atherogenicity.2

Atherogenic triglyceride

The relationship between plasma triglyceride and coronary heart disease is probably mediated through three separate mechanisms:
  1. High triglyceride levels transform the nature of LDL, making it smaller and denser, and more atherogenic. The mechanism underlying this relationship is discussed later.
  2. High triglyceride levels also transform HDL lipoproteins into smaller and denser HDL particles, which are cleared from the bloodstream more rapidly. Consequently, less HDL is available to enter the cardioprotective pathway by which HDL returns cholesterol from the periphery to the liver.
  3. Cholesterol ester transfer protein (CETP) initiates a dynamic lipid exchange in which triglyceride-rich VLDL and chylomicrons lose much of their triglyceride to cholesterol-rich LDL and HDL, and cholesterol ester moves in the reverse direction.3, 4 As triglyceride levels rise, this exchange accelerates. The resulting cholesterol-enriched VLDL or chylomicron remnants are probably the only known naturally occurring lipoproteins that mediate cholesterol uptake by macrophages. As one of the earliest stages of atherogenesis, cholesterol is deposited by the macrophages in the developing atheromatous plaque in the artery wall.

LDL size and coronary risk

Dividing LDL into categories by density is, necessarily, arbitrary. However, the classification has some value, and at Glasgow University we use the size ranges shown in Figure 1.

The small dense LDL III particles are more atherogenic than the lighter fractions. Poor recognition by the LDL receptor-mediated clearance mechanism allows them to stay in the plasma compartment for longer, thereby penetrating the arterial intima more readily. In addition, they are more readily oxidised and may contain less antioxidant, and are more easily taken up by macrophages to create foam cells. The overall effect is highly atherogenic.

The evidence in the medical literature for an association between LDL density and increased cardiovascular risk is as follows:

Figure 1

The lipid shuttle

In vitro experiments have demonstrated that CETP must be present for these lipid exchanges to occur.3, 4 If purified triglyceride-rich lipoproteins are mixed with LDL or HDL in vitro, without the presence of CETP, the molecules maintain their original structure. However, if CETP is added to the lipoprotein mixture, the lipid exchange process is initiated and will continue until equilibrium is reached.

The lipid shuttle is the essential process by which the triglyceride-rich lipoproteins exert their atherogenic effect. The speed of the lipid shuttle is determined not by the activity of the enzyme but by the concentration of the triglyceride-rich lipoproteins. High plasma triglyceride levels promote a more rapid cholesterol ester transfer to the LDL subfractions, thereby increasing the atherogenicity of the lipoprotein profile. Conversely, lower plasma triglyceride levels slow the lipid exchange, reducing the atherogenicity of the LDL subfractions.

The shuttle in practice

The role of plasma triglyceride levels in CETP-mediated lipid exchange has direct clinical relevance. Many practitioners will have encountered the patient who has undergone several months on a lipid-lowering diet, but still has total and LDL cholesterol concentrations at the top end of the normal range. The key information is supplied by the patient's plasma triglyceride concentration: if, for example, it is 2.0 mmol/l (the top 5% of the normal range), and the HDL cholesterol is 0.8 mmol/l, one could assume that the patient's lipoproteins were driven by the CETP reaction. The LDL subfractions would probably be small and dense. The triglyceride-rich particles, loaded with cholesterol, would be highly atherogenic. This is the classic presentation of the atherogenic lipoprotein phenotype (ALP).

The atherogenic lipoprotein phenotype carries a very high risk of premature cardiovascular disease. Roughly speaking, normal total cholesterol, together with a triglyceride level above 2.0 mmol/l and an HDL cholesterol below 1.0 mmol/l (the hallmarks of ALP) carries the same risk as a total cholesterol of 6.5 mmol/l in a patient whose triglycerides and HDL are within the normal range. If that cholesterol level is accompanied by a triglyceride level of 2.5 mmol/l and an HDL cholesterol of less than 1 mmol/l the risk is more than doubled (Figure 2).

Figure 2

Everyone now accepts that a cholesterol of 6.5 mmol/l demands intervention and, because of the 4S study, 7 cardiologists are starting to treat patients whose total cholesterol exceeds 5.5 mmol/l. The atherogenic lipoprotein phenotype deserves the same attention.

The treatment strategy must aim at lowering the triglyceride level in order to slow the CETP lipid shift and drive the dynamic exchange in the reverse direction. The result would be a shift towards larger LDL and HDL subfractions, and a loss of cholesterol ester from the triglyceride-rich lipoproteins. Consequently, the overall atherogenic potential would fall (Figure 3).

Figure 3

One of the most useful effects of the fibrates is their effect on the size distribution and metabolic fate of the LDL subfractions, a direct consequence of their powerful hypotriglyceridaemic effect. The size of the lipoprotein particles seems to be reflected in the length of time that they stay in the circulation. The large light fractions that bind tightly to the LDL receptors are removed quite quickly. The small dense ones, that bind much more weakly to the receptors, are much more persistent, and this persistence is one of the factors that makes them so atherogenic.

The Glasgow experience

The LDL-subfraction study

We recently looked at the distribution of the LDL subfractions among four groups of normocholesterolaemic men. Groups 1 and 2 were defined by the presence or absence of coronary artery disease. Group 3 were normocholesterolaemic men who had survived myocardial infarction, and group 4 were normocholesterolaemic controls. We analysed the distribution of LDL subfractions, with particular reference to the concentration of the small dense fraction, LDL III subfraction (Table I).

Table I. Small dense LDL and coronary risk
Groups12 34
LDL III (mg/dl)
With CAD (n = 46)
Without CAD (n = 24)
MI survivors (n = 40)
Controls (n = 58)
25 (54%)
5 (21%)
29 (73%)
16 (28%)
21 (46%)
19 (79%)
11 (27%)
42 (72%)
Odds ratio*
(95% CI)
4.5 (p < 0.01)

6.9 (p < 0.001)


Adapted from Reference 8.
*The ratio of the risk of CAD or MI between the high and low LDL III groups. Equal risk gives a ratio of 1.0, a higher risk in the low LDL III group gives a ratio less than 1.0, a higher risk in the high LDL III group gives a ratio greater than 1.0

Our results suggest that the risk of coronary artery disease or myocardial infarction is considerably greater in those groups with higher concentrations of plasma LDL III (> 100 mg/dl).8 The relative risk for coronary artery disease associated with LDL III > 100 mg/dl was 4.5 (p < 0.01), and that for myocardial infarction 6.9 (p < 0.001).

The fenofibrate study

Ten years ago we showed that fenofibrate could shift the balance of LDL metabolism from the slow moving, atherogenic fraction, to the less atherogenic, fast moving, fraction.9 There were two possible explanations for that effect; fenofibrate could alter the nature of the LDL receptor, or it could modify the structure of the lipoprotein. We felt that the drug was more likely to change the lipoprotein than the receptor and recently we carried out a rather more detailed study to see if that was correct. 10

We took 8 patients with hypercholesterolaemia (cholesterol > 7.0 mmol/l, triglycerides < 2.3 mmol/l) and treated them with fenofibrate (100 mg tid) for eight weeks. Radioisotope tracers were used to follow the paths of the fast- and slow-moving LDL fractions. The method involved treating LDL with cyclohexanedione to modify its structure so that it no longer binds to the LDL receptor (although in all other respects it behaves like normal LDL). Each patient's LDL was divided it into two portions, one labelled with 125I and the other with131I. The 131I fraction was treated with cyclohexanedione and the 125I fraction was left unchanged. In this way, we obtained two markers that followed the fast- and slow-moving LDL fractions, respectively (Figure 4).

Figure 4

The total mass of cholesterol in each of the LDL pools was calculated, as was the amount of cholesterol that passed into and out of each pool per day (Figure 5).

Figure 5

The results

Fenofibrate lowered the plasma level of cholesterol and triglyceride by 29 and 35%, respectively ( p < 0.001) It reduced LDL cholesterol from 6.0 to 3.86 mmol/l but had no effect on HDL cholesterol (which was normal to begin with). However, none of these changes was particularly surprising. We were much more interested in the behaviour of the fast- and slow-moving lipid pools. The results showed that, although the LDL content of the fast-moving, less atherogenic, pool A changed relatively little, the rate at which cholesterol moved into and out of the pool increased by 76%. On the other hand, the slow-moving, more atherogenic, pool B almost halved in size because of the reduction in the rate of synthesis of the smaller particles.

Unpublished results also showed that fenofibrate shifts distribution of LDL subfractions from the small dense particles derived from pool B to the larger, lighter ones that bind to the receptor. The results suggest that the effect is due to some fundamental influence on the synthesis of LDL particles that increases both their average size and their affinity for the LDL receptor. In these patients, fenofibrate, because of its effect on plasma triglyceride levels, was able to correct the underlying abnormality in LDL metabolism.


Darwin and CETP

Why has evolution provided us with CETP? Rats and several of the higher mammalian species manage quite well without it. Primates and several other mammals, on the other hand, seem to need it in order to function normally. What unites the CETP-producing species is their susceptibility to atherosclerosis and premature cardiovascular disease. Experiments, in which the gene that governs CETP synthesis was introduced to a rat, showed that the procedure increases the rat's susceptibility to atherosclerosis.11 In humans, studies have shown that smoking enhances the enzyme's activity,12, 13 suggesting that CETP, and the lipid exchange that it promotes, are central to atherogenesis.

LDL concentration and atherogenicity

Evidence for the role of triglycerides, and their effect on LDL concentration, in the atherogenic process can be found in studies of familial hypercholesterolaemia. A report of father and son from just such a family, where each family member has the same LDL receptor mutation, illustrates the point.14 The son recently underwent his third bypass graft. His total cholesterol was 9.1 mmol/l. His 86-year-old father, with exactly the same LDL receptor defect and the same total cholesterol, had no sign of coronary disease. The son had a triglyceride level of 2.6 mmol/l and a low HDL cholesterol (0.9 mmol/l); his father had low triglycerides (1.4 mmol/l) and high HDL (1.4 mmol/l), a protective combination that, apparently, outweighed the risk associated with the genetically determined hypercholesterolaemia.

The practice setting

For the clinician planning a therapeutic strategy, it is essential to know the patient's LDL cholesterol level, and how that cholesterol is distributed among the LDL subfractions. Measuring the distribution of the LDL subfractions is a tricky process that is unlikely to become a part of routine patient management. However, the strong correlation between plasma triglyceride concentration and size distribution in the LDL subfractions (70-80%) provides all the information the clinician needs. The higher the plasma triglyceride level, the faster the CETP-induced lipid shift, and the smaller and more atherogenic the LDL.

The modes of action of the two main classes of agents in this area differ in significant ways. We have discussed the fibrates and their selective activity in reducing triglyceride levels and LDL III concentration. By contrast, the statins have a less selective effect. They work by depleting the liver cell of cholesterol, forcing the cell to take more cholesterol from the plasma. As a result, plasma LDL levels fall. The effect is non-selective: the statins reduce the concentration of all the LDL subfractions equally, although they may have a mildly preferential effect on the larger lighter fractions that bind more tightly to the LDL receptor.15 In patients with raised LDL who have normal triglyceride levels, this is relatively unimportant because most of the LDL will be in the larger, less dense fractions. But, if the plasma triglyceride concentration is high, the proportion of LDL in the small dense atherogenic fraction is greater, and maximum lipid reduction (and clinical benefit) may not be achieved.

Atherogenesis and thrombogenesis: a unified field

The old view of atherogenesis and thrombogenesis as two separate entities has been replaced by a 'unified field theory' that links them closely together. The theory envisages a process by which triglyceride-rich lipoproteins come into contact with the artery wall, causing endothelial cell dysfunction and stimulating the recruitment of macrophages into the endothelium. The macrophages will have accumulated cholesterol from the triglyceride-rich lipoproteins, which is then deposited within the artery wall. In this way, the macrophages contribute to plaque extension along the artery wall.

A second effect of the triglycerides on the vascular endothelium is their ability to stimulate the synthesis of thrombogenic mediators such as plasminogen activator inhibitor or PAI-1. There is a great deal of evidence that PAI-1 activity is enhanced (and fibrinolysis impaired) in patients with hypertriglyceridaemia.16, 17 It appears that arterial endothelial cells secrete PAI-1, suggesting that enhanced production could suppress local plasmin synthesis and accelerate intra-arterial fibrin deposition. Evidence for this comes from studies in which human arterial endothelial cells that have been incubated with VLDL from hypertriglyceridaemic patients speed up their production of PAI-1.18 VLDL from normolipaemic individuals, on the other hand, has no such effect.19 This strongly suggests a direct link between hypertriglyceridaemia and impaired fibrinolysis.


The initial search for a simple direct relationship between plasma triglyceride and coronary risk failed because it overlooked the complexity of lipid exchange dynamics. These interchanges, driven by the dynamic metabolism of the triglyceride-rich lipoproteins, centre around the CETP-mediated lipid shuttle. They are responsible for distributing cholesterol among the lipoprotein fractions and, indirectly, for moving cholesterol from the blood into the arterial wall.

The prevention of coronary heart disease has focused exclusively on hypercholesterolaemia. These insights into the role of triglycerides suggest that clinicians should extend their vigilance to another group of patients, those with raised triglyceride, low HDL cholesterol and moderately raised cholesterol that characterise the atherogenic phenotype. The Glasgow experience suggests that treatment with a third generation fibrate may bring a reduction in risk of the same order as that seen in the simvastatin trial.7


1. Hokanson, J., and Austin M., (1993). Plasma triglycerides and coronary risk. A meta-analysis. Circulation, 88, 1-510.

2. Austin, M.A., Breslow, J.L., Hennekens, C.H., et al., (1988). Low-density lipoprotein subclass patterns and risk of myocardial infarction. JAMA, 260, 1917-21.

3. Tall, A.R., (1995). Plasma cholesterol ester transfer protein and high density lipoproteins: new insights from molecular genetic studies. J Intern Med., 237, 5-12.

4. Tato, F., Vega, G.L., Tall, A.R., et al., (1995). Relation between cholesterol transfer protein activities and lipoprotein cholesterol in patients with hypercholesterolaemia and combined hyperlipidaemia. Arterioscler Thromb Vasc Biol., 15, 112-20.

5. Watts, G.F., Lewis, B., Brunt, J.N.H., et al., (1992). Effects on coronary artery disease of lipid-lowering diet, or diet plus cholestyramine in the St Thomas' Atherosclerosis Regression Study (STARS). Lancet, 339, 563-9.<

6. Blankenhorn, D.H., Azen, S.P., Crawford, D.W., et al., (1991). Effects of colestipol-niacin therapy on human femoral atherosclerosis. Circulation, 83, 438-47.

7. Scandinavian Simvastatin Survival Study Group, (1994). Randomised trial of cholesterol lowering in 4444 patients with coronary heart disease: the Scandinavian Simvastatin Survival Study. Lancet, 344, 1383-9.

8. Griffin, B.A., Freeman, D.J., Tait, G.W., et al., (1994). Role of plasma triglyceride in the regulation of plasma low density lipids to coronary heart disease risk. Atherosclerosis, 106, 241-53.

9. Shepherd, J., Caslake, M.J., Lorimer, A.R., et al., (1985). Fenofibrate reduces low density lipoprotein catabolism in hypertriglyceridemic subjects. Arteriosclerosis, 5, 162-8.

10. Caslake, M.J., Packard, C.J., Gaw, A., et al., (1993). Fenofibrate and metabolic heterogeneity in hypercholesterolaemia. Arterioscler Thromb., 13, 702-11.

11. Dinchuk, J., Hart, J., Gonzalez, G., et al., (1995). Remodelling of lipoproteins in transgenic mice expressing human cholesterol ester transfer protein. Biochim Biophys Acta., 1255, 301-10.

12. Dullaart, R.P., Hoogenberg, K., Dikkeschei, B.D., et al., (1994). Higher plasma lipid transfer protein activities and unfavorable lipoprotein changes in cigarette-smoking men. Arterioscler Throm., 14, 1581-5.

13. Freeman, D.J., Griffin, B.A., Holmes, A.P., et al., (1994). Regulation of plasma HDL cholesterol and subfraction distribution by genetic and environmental factors. Associations between the TaqI B RFLP in the CETP gene and smoking and obesity. Arterioscler Thromb., 14, 336-44.

14. Kotze, M.J., Davis, H.J., Bissbot, S., et al., (1993). Intrafamilial variability in the clinical expression of familial hypercholesterolaemia: importance of risk factor determination for genetic counselling. Clinical Genetics, 43, 295-99.

15. Cheung, M.C., Austin, M.A., Moulin, P., et al., (1993). Effects of pravastatin on apolipoprotein-specific high density lipoprotein subpopulations and low density lipoprotein subclass phenotypes in patients with primary hypercholesterolaemia. Atherosclerosis, 102, 107-19.

16. Cigolini, M., Targher, G., and Seidell, J.C., (1994). Relationships of plasminogen activator inhibitor-1 to anthropometry, serum insulin, triglycerides and adipose tissue fatty acids in healthy men. Atherosclerosis, 106, 139-47. 17. Asplund-Carlson, A., Hamsten, A., Wiman, B., et al., (1993). Relationship between plasma plasminogen activator inhibitor-1 activity and VLDL triglyceride concentration, insulin levels and insulin sensitivity: studies in randomly selected normo- and hypertriglyceridaemic men. Diabetologica, 13, 1865-73.

18. Stiko-Rahm, A., Wiman, B., Hamsten, A., et al., (1990). Secretion of plasminogen activator inhibitor-1 from cultured human umbilical vein endothelial cells is induced by very low density lipoproteins. Arteriosclerosis, 9, 134-9.

19. Mussoni, L., Mannuci, L., Sirtori, M., et al., (1992). Hypertriglyceridaemia and regulation of fibrinolytic activity. Arterioscler Thromb., 12, 19-23.