Indications for cholesterol-lowering medication: comparison of risk-assessment methods

BACKGROUND Recommendations for the prescription of lipid-lowering drugs emphasise the importance of an assessment of absolute coronary heart disease (CHD) risk based on all risk factors, rather than simply the serum cholesterol concentration. If, however, the methods recommended for risk assessment are inaccurate, recommended prescribing will not occur. We compared several sets of guidelines for such treatment in a series of patients referred to a lipid clinic, to assess the difference in degree of risk of CHD at which lipid-lowering medication is recommended by each set of guidelines. METHODS For a series of 570 patients (50% men) without pre-existing clinical evidence of atherosclerosis referred to a lipid clinic, we compared the algorithms, charts, and tables used by the US National Cholesterol Education Program (NCEP), the joint guidelines of the European Society of Cardiology, the European Atherosclerosis Society, and the European Society of Hypertension, and the report of the UK Standing Medical Advisory Committee with the Framingham risk equation programmed into a computer. FINDINGS In 386 patients for whom the NCEP and UK guidelines could be compared, 62% of the men and 72% of the women met NCEP criteria for lipid-lowering medication, whereas only 9% of the men and less than 1% of the women met the UK criteria. The Framingham equation estimated a CHD risk of more than 3% per year in 22% of the men and 7% of the women, which shows that the UK tables underestimated CHD risk. European guidelines could be applied to only 261 patients, and were reasonably accurate in assessment of a CHD risk of 2% per year. INTERPRETATION Guidelines for the use of statin treatment in patients with CHD differ in their assessment of CHD risk. The method of risk assessment recommended in future guidelines for CHD prevention should be critically tested in relevant groups of patients.