High density lipoproteins (HDLs) are an important link in the pathway which transports cholesterol from extrahepatic tissues to the liver. A low concentration of HDL cholesterol is a powerful predictor of premature coronary heart disease (CHD). The need to treat a low HDL concentration depends on the concentrations of the other lipids and the patient's risk factors. Non-pharmacological measures, such as weight loss and exercise, should usually precede drug treatment. The observation that interventions which raise HDL concentrations reduce CHD require confirmation.

A high plasma cholesterol is associated with an increased risk of developing coronary heart disease (CHD). We also know that a reduced concentration of high density lipoprotein (HDL) cholesterol is an equally powerful predictor of CHD.1 However, few doctors understand what HDLs are and most are uncertain about how (or whether) to treat patients with low HDL concentrations.

What are HDLs and what is their function?
HDLs are cholesterol-rich lipoproteins which account for about 25% of the cholesterol in normal human plasma. The concentration of HDL cholesterol is normally in the range 1.1-1.8 mmol/L in men and 1.2-2.0 mmol/L in women. Very rarely, levels may be even higher than 2 mmol/L and contribute to an elevation of plasma cholesterol. Usually, however, the HDL cholesterol has very little influence on the concentration of plasma total cholesterol. While the concentration of LDL cholesterol (the major cholesterol fraction in human plasma) may be deduced from the plasma total cholesterol, the HDL cholesterol concentration can be determined only by direct assay.

HDLs act as the initial acceptors of tissue cholesterol in the pathway which transports cholesterol from peripheral tissues to the liver either for recycling or for excretion in the bile. This function may underlie the protective effect of HDLs on CHD.

What is the relationship between HDL cholesterol levels and CHD?
A low concentration of HDL cholesterol predicts the development of CHD with a power as great as or greater than an increased level of plasma total cholesterol.1 The PROCAM study2 reported the incidence of cardiac events over a 6 year follow-up period in 4407 men aged 40-65 years who had no known cardiac disease at the start of the study. The relationships between plasma total cholesterol, plasma triglyceride and HDL cholesterol concentrations and the incidence of cardiac events in the PROCAM study are shown in Tables 1 and 2.

Several points about the risk of myocardial infarction emerged from this study:

  • at all levels of plasma total cholesterol, the risk increased as the HDL cholesterol level decreased
  • at all levels of HDL cholesterol, the risk increased as the plasma total cholesterol increased
  • in men with plasma total cholesterol levels above 6.5 mmol/L, the finding of an HDL cholesterol level <0.9 mmol/L indicated that the risk over the next 6 years was as high as 20-30%
  • at all levels of plasma triglyceride, an HDL cholesterol level of 0.9 mmol/L or less was associated with an increased risk. The highest risk was in subjects who had mild to moderate hypertriglyceridaemia (>=2.3 mmol/L) in combination with a low level of HDL cholesterol. This observation of increased risk associated with the combination of high triglyceride and low HDL cholesterol has also been observed in several other large studies
Table 1
Incidence of CHD as a function of plasma total cholesterol and HDL cholesterol in the PROCAM study
Plasma total cholesterol* Percentage of subjects having a mycocardial infarction over a 6 year period
HDL-C > 1.4 HDL-C 0.91 - 1.4 HDL-C <= 9
< 5.2 0.6% 1.6% 3.3%
5.2 - 6.4 1.4% 2.5% 11.0%
6.5 - 7.7 3.3% 4.3% 21.1%
> 7.7 13.3% 11.7% 29.4%

The PROCAM study reported the incidence of myocardial infarction over a 6 year follow-up period in 4407 asymptomatic men aged 40-65.

* Plasma total cholesterol and HDL concentrations are given in mmol/L.

Table 2
Incidence of CHD as a function of plasma triglyceride and HDL cholesterol in the PROCAM study
Plasma triglyceride* Percentage of subjects having a mycocardial infarction over a 6 year period
HDL-C > 1.4 HDL-C 0.91 - 1.4 HDL-C <= 0.9
< 1.7 1.0% 2.7% 8.3%
1.7 - 2.2 + 2.7% 11.2%
>= 2.3 + 4.0% 12.8%

The PROCAM study reported the incidence of myocardial infarction over a 6 year follow-up period in 4407 asymptomatic men aged 40-65.

* Plasma triglyceride and HDL cholesterol concentrations are given in mmol/L.

+ The number of subjects in these groups is insufficient for analysis to be meaningful.

What causes low HDL cholesterol concentrations
The concentration of HDL cholesterol becomes a matter of concern if it is less than 1.Ommol/L. Such levels are encountered for a variety of reasons (Table 3).

Genetic disorders
Several genetic disorders result in very low levels or even an absence of plasma HDL. These conditions are extremely rare and are of little importance in normal clinical practice.

Lifestyle factors
The low HDL cholesterol concentrations associated with a sedentary lifestyle, obesity and smoking may increase coronary risk. However, the low HDL cholesterol which is common in people who eat very low-fat, high-carbohydrate diets is not necessarily a cause for concern, as it is generally offset by a low level of LDL cholesterol.

Low HDL in other disorders
HDL cholesterol is commonly lower in patients with obesity and non-insulin dependent diabetes mellitus (NIDDM). In both conditions, the low concentration of HDL cholesterol tends to coincide with hypertriglyceridaemia with sufficient frequency for the term 'high triglyceride-low HDL cholesterol syndrome' to have been coined.

The antioxidant drug, probucol, lowers the concentration of cholesterol in both HDL and LDL. However, there is no evidence that the reduction in HDL cholesterol by probucol is a cause for concern. The HDL cholesterol concentration may be reduced and the plasma triglyceride increased by some beta blockers and diuretics. Such drugs should therefore be used with caution in patients with dyslipidaemia.

Table 3
Conditions associated with low HDL cholesterol

Genetic disorder (rare)

Lifestyle factors

  • sedentary existence
  • smoking
  • low-fat, high-carbohydrate diet


Non-insulin dependent diabetes mellitus



  • probucol
  • beta blockers
  • diuretics

Now can HDL cholesterol levels be increased?
Non-pharmacological measures are often effective in raising the concentration of HDL cholesterol. The most effective are weight reduction in those who are overweight and aerobic exercise in those with a sedentary lifestyle. Stopping smoking often results in an increase in the HDL cholesterol concentration.

The concentration of HDL cholesterol may also be increased by several drugs (Table 4), especially those such as gemfibrozil and nicotinic acid. These drugs also reduce plasma triglyceride. Plasma cholesterol lowering drugs such as the HMG CoA reductase inhibitors, simvastatin and pravastatin, and the bile acid sequestering resins, cholestyramine and colestipol, lead to mild increases in HDL cholesterol levels. The mechanism by which these drugs raise HDL levels is not known.

HDL concentrations may also be increased by alcohol consumption, but the benefit of drinking more alcohol to raise HDL is uncertain.

What is the evidence that raising HDL cholesterol levels reduces the risk of CHD?
A low concentration of HDL cholesterol, especially if accompanied by elevations of plasma total cholesterol or triglyceride, is a powerful predictor of premature CHD. Studies of people who are free of CHD (primary prevention), such as in the Helsinki Heart Study, suggest that the incidence of future cardiac events may be reduced by raising HDL levels in people who have low HDL concentrations. The people with the highest coronary risk in the Helsinki Heart Study were those in whom mild to moderate elevations of plasma triglyceride and plasma total cholesterol were accompanied by reduced levels of HDL cholesterol. Treatment with gemfibrozil resulted in an increase in HDL cholesterol, a decrease in the levels of plasma triglyceride and plasma total cholesterol and a 65% reduction in the incidence of cardiac events.3

There is also evidence that interventions which increase the level of HDL cholesterol are of benefit to patients who already have CHD (secondary prevention). In one such secondary prevention study, an increase in HDL cholesterol was found to correlate independently with regression of coronary atherosclerosis.4 Observations such as this support a case for aggressive therapy in patients in whom existing CHD coincides with a low level of HDL cholesterol.

It remains uncertain whether the reduction in coronary artery disease in either primary or secondary prevention studies is the consequence of the increase in HDL cholesterol or the decrease in other plasma lipids. However, this may be irrelevant in terms of practical management since:

  • low levels of HDL cholesterol frequently co-exist with elevations of other plasma lipids and
  • therapies which decrease plasma triglyceride and cholesterol tend also to increase the level of HDL cholesterol.

Nevertheless, there is a need for further research to elucidate the mechanisms by which low HDL levels may predispose to CHD. There is also an urgent need for additional prospective studies to show that an increase in HDL levels results in a reduced incidence of CHD.

Table 4
Effects of lipid lowering drugs on HDL cholesterol
Drugs inducing major increases in HDL cholesterol

nicotinic acid

Drugs inducing mild increases in HDL cholesterol


Should patients with low levels of HDL cholesterol receive treatment?
The finding of a low HDL cholesterol concentration should be considered in the context of the levels of other plasma lipids and the presence of other CHD risk factors. For example, in patients whose low HDL is secondary to a low-fat, high-carbohydrate diet, the plasma total cholesterol is likely also to be low and there is little indication for intervention. If, however, the low HDL cholesterol is associated with even a mild elevation of plasma total cholesterol, especially if the plasma triglyceride is also elevated, coronary risk is markedly increased and intervention is indicated. If the patient also has NIDDM, the case for treating the dyslipidaemia is especially strong.

Proposed guidelines for treating low HDL cholesterol
Low HDL cholesterol with plasma total cholesterol normal

e.g. HDL cholesterol <1.0 mmol/L
Plasma total cholesterol <5.0 mmol/L

The low HDL level in this case is either secondary to the consumption of a low-fat, high-carbohydrate diet, in which case therapy is not indicated, or it is of genetic origin and will probably be unresponsive to therapy. There is thus little case for intervention.

Low HDL cholesterol with mild elevations of other lipids

e.g. HDL cholesterol <1.0 mmol/L
Plasma total cholesterol 5.0-6.5 mmol/L

The coronary risk associated with this profile is considerably increased and may be especially high if the patient also has NIDDM. Therapy should start with non-pharmacological measures (stop smoking, weight reduction, exercise program). If the lipids do not respond, especially if the patient has either a personal or a strong family history of premature CHD, there is a powerful case for drug treatment. Both gemfibrozil and nicotinic acid will raise the HDL, although the capacity of nicotinic acid to decrease glucose tolerance precludes its use in diabetics.

Low HDL cholesterol with moderate elevations of other lipids

e.g. HDL cholesterol <1.0 mmol/L
Plasma total cholesterol >6.5 mmol/L and/or plasma triglyceride >4.0 mmol/L

This lipid profile indicates a very high CHD risk. If non-pharmacological measures (low saturated fat diet, weight reduction, appropriate exercise program, stop smoking) are not successful, there is a strong case for drug intervention particularly if there are other risk factors. This case is especially compelling if the patient has NIDDM or has either existing CHD or a family history of premature CHD in first degree relatives. Other factors which increase risk are smoking and hypertension. The choice of drug therapy depends on whether the predominant elevation is plasma total cholesterol (cholestyramine, colestipol, simvastatin, pravastatin) or plasma triglyceride (gemfibrozil, nicotinic acid) or both (gemfibrozil, simvastatin, pravastatin, nicotinic acid). In most patients, successful reductions of both plasma total cholesterol and triglycerides will generally be accompanied by some degree of increase in the concentration of HDL cholesterol.

Editorial note:

Professor Barter has given readers an excellent summary of the role of HDL in the pathogenesis of atherosclerotic vascular disease, together with his personal recommendations for instituting lipid lowering therapy when HDL is low (<l mmol/L). These differ slightly from the PBAC's recently revised official guidelines. (see extract below).

If HDL cholesterol is <1 mmol/L in association with a variety of high risk situations, Professor Barter recommends lipid lowering therapy (after non-pharmacological measures have been tried) if total cholesterol is in the range 5.0-6.5 mmol/L. The PBAC has explicitly specified these high risk categories and approved subsidy for these patients with a slightly higher threshold, range 5.5-6.5 mmol/L.

If HDL cholesterol is <1 mmol/L but there are no other cardiovascular risk factors, the PBAC threshold for lipid lowering drug therapy is a total cholesterol of 6.5 mmol/L. Professor Barter suggests that there may be benefits from treatment at lower levels of total cholesterol.

If HDL is <1 mmol/L and fasting triglycerides are >4 mmol/L, Professor Barter believes there is a strong case for lipid lowering therapy. The PBAC guidelines make no mention of this combination, but allow subsidy for patients with triglycerides >4 mmol/L in men aged 35-75 and postmenopausal women.

These are all relatively minor differences. They should not detract from the major advances in the new PBAC guidelines. These are:

  • Patients with the highest risk profile are the ones who gain the most benefit from lipid lowering therapy, hence the lower threshold for instituting drug therapy in these patients.
  • A low HDL level is explicitly recognised as an important risk factor in addition to total cholesterol in high risk patients. If HDL cholesterol is low, lipid lowering therapy can be instituted (and receive subsidy) at a lower level of total cholesterol than if HDL is normal.
  • The PBAC has therefore endorsed the need to measure HDL cholesterol in high risk patients with total cholesterol concentrations greater than 5.5 mmol/L.

PBAC qualifying criteria for dyslipidaemia
Measurement of plasma lipids

The decision to commence drug therapy should be based on at least two measurements at an accredited laboratory.

Cholesterol triglyceride levels
In addition to dietary therapy, drug therapy is indicated in patients in accordance with the table below:

Patient category Lipid levels for subsidy
Patients with one or more of the following:
prior cardiovascular disease cholesterol > 6.5 mmol/L or cholesterol > 5.5 mmol/L and HDL < 1 mmol/L
peripheral vascular disease
family history of cardiovascular disease (first degree relative less than 60 years of age)
familial hypercholesterolaemia
diabetes mellitus
Patients with:
HDL <1 mmol/L cholesterol >6.5 mmol/L
Patients without above risk factors:
men 35-75 years cholesterol > 7.5 mmol/L or triglyceride > 4 mmol/L
postmenopausal women up to 75 years
Patients not included in the above:
cholesterol >9 mmol/L or triglyceride > 8 mmol/L

Note: In all cases where drug therapy is instituted, dietary therapy should continue.


  1. Gordon DJ, Probstfield JL, Garrison RJ, Neaton JD, Castelli WP, Knoke JD, et al. High-density lipoprotein cholesterol and cardiovascular disease: four prospective American studies. Circulation 1989;79:8-15.
  2. Assmann G, Schulte H. Results and conclusions of the Prospective Cardiovascular Munster (PROCAM) Study. In: Assmann G, editor. Lipid metabolism disorders and coronary heart disease. 2nd enl ed. Munchen: MMV Medizin Verlag, 1993:19-67.
  3. Manninen V, Tenkanen L, Koskinen P, Huttunen JK, Manttari M, Heinonen OP, et al. Joint effects of serum triglyceride and LDL cholesterol and HDL cholesterol concentrations on coronary heart disease risk in the Helsinki Heart Study: implications for treatment [see comments]. Circulation 1992;85:37-45. Comment in: Circulation 1992;85:365-7. Comment in: Circulation 1993;87:300-1.
  4. Brown G, Albers JJ, Fisher LD, Schaefer SM, Lin JT, Kaplan C, et al. Regression of coronary artery disease as a result of intensive lipid-lowering therapy in men with high levels of apolipoprotein B [see comments]. N Engl J Med 1990;323:1289-98. Comment in: N Engl J Med 1990;323:1337-9. Comment in: N Engl J Med 1991;324:1133-5.