Using changes in attributable risk to predict long-term efficacy of simvastatin treatment

It is a common perception that coronary heart disease (CHD) deaths will decline by 2% for every 1% decrease in serum total cholesterol (TC). We investigated the changes in serum TC concentrations obtained with simvastatin, an inhibitor of hydroxymethylglutaryl-CoA reductase, in 70 hypercholesterolemic subjects. Although simvastatin reduced serum TC concentrations by 30% after 6 months, calculations from age- and risk-related TC quintiles show that the absolute difference in attributable CHD deaths would be just > 1%. These results are reminiscent of the experience from several large prospective primary drug trials that demonstrated impressive relative benefit ratios but much smaller absolute differences (< 2%) in CHD endpoints between the treated and untreated groups. This epidemiologically based approach of assessing efficacy and potential long-term benefit of a lipid-lowering drug may be more appropriate and potentially less misleading than just the percentage change of TC (or other lipoproteins) so frequently reported in short-term drug trials.