A preponderance of small, dense (sd) LDL is strongly associated with the development of coronary heart disease, but the method for the measurement of sd LDL is too laborious for clinical use. We report a simple method for the quantification of sd LDL that is applicable to an autoanalyzer. This method consists of two steps: first, to precipitate the lipoprotein of density (d) Ͻ 1.044 g/ml using heparin-magnesium; and second, to measure LDL-cholesterol in the supernatant by the homogenous method or apolipoprotein B (apoB) by an immunoturbidometric assay. LDL particles are heterogenous with respect to size and density (d) of lipid composition. Two distinct phenotypes based on LDL particles have been recognized: pattern A, with a higher proportion of large, buoyant LDL particles, and pattern B, with a predominance of small, dense (sd) LDL particles (1, 2, 3). It has been suggested that compared with buoyant LDL, sd LDLs are highly atherogenic as a result of their higher penetration into the arterial wall, their lower binding affinity for the LDL receptor, prolonged plasma half-life, and lower resistance to oxidative stress (4, 5). Several studies have reported a 2-to 3-fold increase in coronary heart disease (CHD) risk among patients with pattern B (1, 2). We have also reported that sd LDL is highly associated with CHD events in Japanese, an ethnic group with lower LDL-cholesterol levels, compared with Western populations (6, 7). Therefore, sd LDL has been highlighted as a new potent risk marker for CHD.LDL particle size is usually measured by gradient gel electrophoresis (GGE) using nondenaturing polyacrylamide according to the method of Krauss and Burke (8). However, this procedure requires a long assay time, i.e., overnight electrophoresis, staining, and destaining. Of course, this assay does not allow quantitative determination of sd LDL. Ultracentrifugation is the standard technique for the isolation of the sd LDL fraction (9, 10) and allows quantification of sd LDL. Griffin et al. (10) reported that LDL-III (equivalent to sd LDL, with d ϭ 1.044-1.060 g/ml) concentration was significantly increased in CHD patients, and the relative CHD risk was increased 4.5-fold in individuals having LDL-III concentrations (protein plus lipid) Ͼ 100 mg/dl, compared with those with lower concentrations. Their study suggests that in addition to measurement of LDL size, quantification of sd LDL is also useful for the assessment of CHD risk. However, the ultracentrifugation technique is too laborious for general clinical use, because it requires special equipment and a long running time.It is well known that a combination of divalent cations and polyanions precipitates apolipoprotein B (apoB)-containing lipoproteins, which allows for the measurement of HDL-cholesterol. However, we found that the combination of heparin and Mg did not precipitate all of the apoBcontaining lipoproteins; the denser part of LDL remained in the supernatant. Here we report a simple precipitation method for the direct measurement of sd LDL in serum