Editorials

Cholesterol: how low is low enough?

Reaching target levels may be better than relative reductions 

Not so very long ago many of us did not realise the importance of cholesterol lowering in patients with coronary disease. After the World Health Organisation's clofibrate trial¹ many patients with hyperlipidaemia, with or without manifest coronary disease, were left without cholesterol lowering treatment. Now, after several large placebo controlled trials, the message is clear: patients with coronary disease and high or normal serum cholesterol concentrations benefit from cholesterol lowering treatment, by a 20-40% reduction in coronary events. ^{2 3} What remains less clear is by how much to lower those concentrations and whether it is the absolute concentration or the percentage reduction that matters most.

Current guidelines recommend a treatment goal for low density lipoprotein cholesterol of 2.6 mmol/l in patients with coronary disease. ^{4 5} However, statins, the most widely used drugs for cholesterol lowering, reduce cholesterol values not to a specific level but in proportion to pretreatment values. Thus the absolute reduction in concentration will be greater in patients with high initial values, but in these patients the target value will also be harder to achieve. Furthermore, angiographic trials have shown that percentage changes in stenoses are significantly correlated not with the achieved concentration of low density lipoprotein cholesterol but with its percentage reduction.⁶ Some have suggested that it may be more practicable to recommend the percentage by which low density lipoprotein cholesterol concentrations should be lowered rather than setting target levels.⁵

Lower serum cholesterol concentrations are associated with lower risk of coronary disease throughout the range considered normal in Western populations. This is a dynamic process: as population levels of serum cholesterol decrease, as a result of dietary changes, so does mortality from ischaemic heart disease.⁷ Data from a Chinese population show that the positive relation between coronary risk and serum cholesterol concentration continues down to values well below the range of Western populations, with no evidence of a threshold effect.⁸ In observational studies a prolonged difference in usual serum cholesterol value of 0.6 mmol/l is associated with an almost 30% reduction in risk of coronary disease.⁹ The effects of a similar difference in the randomised controlled trials of statins are smaller, possibly because the trials did not extend beyond five years. More prolonged cholesterol lowering may result in greater reductions in risk.

Data derived from several sources show a log linear relation between low density lipoprotein cholesterol concentration and risk of coronary disease.⁵ As the absolute reduction in low density lipoprotein cholesterol induced by cholesterol lowering drugs will be larger with higher initial levels and, as the relation between low density lipoprotein cholesterol and risk of coronary disease is curvilinear, this larger reduction will result in a proportionately greater net reduction in coronary events. In terms of reduction in absolute risk in a particular individual, the reduction will be determined as much by his or her overall coronary risk, as by the initial cholesterol concentration. The presence of coronary disease means a higher absolute risk and that the same net benefit in terms of reduced coronary events can be expected at lower initial lipid levels in patients with established coronary disease.

So far, clinically important reductions have been achieved only in patients at high risk. Both the Scandinavian Simvastatin Survival Study (4S) and the Cholesterol and Recurrent Events (CARE) study show, furthermore, that the proportional reduction in risk is similar, irrespective of the initial cholesterol concentration, except for those in the CARE study with baseline low density lipoprotein values below 3.2 mmol/l. It is unknown whether lowering of low density lipoprotein cholesterol to values below those achieved by the 4S study or the CARE study will reduce coronary events further. Extrapolation from observational and randomised controlled studies, however, suggests that achievement of lower low density lipoprotein concentration in patients with coronary disease with high or normal cholesterol values may reduce coronary events by perhaps as much as 15-20%.

Is this effort worthwhile? In the absence of evidence from trials of more intensive cholesterol lowering this is, so far, conjectural. However, considerable scope exists for attempting to improve the prognosis of patients with coronary disease. In the study population of the 4S trial the risk of coronary death in the treated group at 5.4 years was still 5%. This should be compared with the risk of 1.7% during 4.9 years' follow up in the placebo group in a primary prevention trial of pravastatin in men with hypercholesterolaemia but no prior infarction.¹⁰ Consequently, a further reduction in risk in a population at high risk could lead to a not unimportant number of lives saved. Whether the additional numbers needed to be treated are justified by the number of coronary events saved is not clear, nor is whether prolonged intensive cholesterol lowering by medication will lead to excess non-coronary mortality. Also there are probably many patients at high risk who are not being treated at all. Nevertheless, to the best of our current knowledge, settling for proportional lowering of serum cholesterol values, instead of attempting to achieve target levels, will mean less than optimal treatment in a fair share of patients with coronary disease.