Low-density-lipoprotein (LDL) apheresis has the theoretical advantage over anion-exchange resins and hydroxymethylglutaryl coenzyme A inhibitors of decreasing lipoprotein(a) as well as LDL. To confirm this advantage, patients with heterozygous familial hypercholesterolaemia and coronary artery disease were randomised to receive LDL apheresis fortnightly (with disposable dextran sulphate/cellulose columns) plus simvastatin 40 mg daily, or colestipol 20 g plus simvastatin 40 mg daily. Quantitative coronary angiography was repeated after a mean of 2.1 years in 20 patients undergoing apheresis and in 19 on combination drug therapy. Changes in serum lipoproteins were similar in both groups apart from greater lowering by apheresis of LDL cholesterol (3.2 vs 3.4 mmol/L in drug group, p = 0.03) and lipoprotein(a) (geometric means 14 vs 21 mg/dL, p = 0.03). There were no significant differences in primary angiographic endpoints per patient but lesion-based and segment-based secondary endpoints were biased in favour of the drug group (change in minimum lumen diameter of lesions 0.07 vs -0.004 mm, p = 0.046; change in mean lumen diameter of segments 0.02 vs -0.06 mm, p = 0.01). None of the angiographic changes correlated with lipoprotein(a) concentrations. Per patient changes in % diameter stenosis and minimum lumen diameter in the two groups were as or more favourable than those observed in five published trials that assessed lipid-lowering drug therapy by quantitative coronary angiography. Although LDL apheresis combined with simvastatin was more effective than colestipol plus simvastatin in reducing LDL cholesterol and lipoprotein(a), it was less beneficial in influencing coronary atherosclerosis and should be reserved for patients unresponsive to drugs. Decreasing lipoprotein(a) seems to be unnecessary if LDL cholesterol is reduced to 3.4 mmol/L or less.