Hypercholesterolemia has been recognised as a primary risk factor for coronary heart disease. Reduction of plasma levels of total and LDL cholesterol has been shown to decrease coronary atherosclerosis. Plasmapheresis represents an useful non-pharmacological tool to treat severe hypercholesterolemias. We have evaluated the effectiveness of a system of plasmapheresis using a cascade filtration method in two young male subjects (aged 16 and 26 years) with homozygous familial hypercholesterolemia. Both showed severe coronary atherosclerosis as determined by angiography. Procedures were performed at intervals of 7 days in each case. We observed a mean reduction of plasma levels of total cholesterol of 59.5% (range 31.0-75.5%); LDL-cholesterol, 61.6% (range 32.6-77.1%); triglycerides, 48.1%; HDL-cholesterol, 31.1%; apo A-I, 30.8%; and apo B, 57.6%. We also noted a reduction of other parameters, such as fibrinogen (49.9%) and Lp(a) (59.9%). At the end of each procedure about 8 g of cholesterol was removed from the total body pool. A decrease of total proteins (26.9%) and albumin (19.6%) was also observed, but this was completely restored before the next apheresis (1 week). These data show the effectiveness of the removal of LDL in a cascade filtration system, which obtains results not very different from other more selective methods. The lack of selectivity is not much of a problem, since it also reduces other risk factors such as Lp(a) and fibrinogen.