The association between elevated levels of low-density lipoprotein (LDL) cholesterol and an increased risk of premature coronary heart disease (CHD) is well documented. Most clinical laboratories estimate LDL cholesterol concentrations according to the Friedewald formula. It provides a relatively reliable estimate of LDL cholesterol concentration, provided the triglyceride concentration is or =400 mg/dL. The interactions between lipoproteins and surfactants, divalent cations, sugars, and lectins were investigated, and we developed a new assay protocol to chemically measure the LDL cholesterol level in serum that does not require immunoseparation or centrifugation. The assay protocol was evaluated by measuring serum samples obtained from 88 patients and 20 healthy volunteers. The triglyceride levels of the patient samples ranged from 66 to 2199 mg/dL, and the samples were classified as or =400 mg/dL (n=52; 23, 3, and 26 patients had type IIb, type III, and type IV hyperlipoproteinemia, respectively) for comparative studies. The accuracy and precision of our assay protocol fulfilled the criteria of the NCEP Lipid Standardization Panel, and no matrix effect influenced the measurements. The assay protocol is less sensitive to LDL-I than to LDL-II and LDL-III. LDL cholesterol measurements correlated well with those obtained by the ultracentrifugal assay of normotriglyceridemic and hypertriglyceridemic samples. This evidence shows that the results obtained with our assay protocol are superior to those obtained with the Friedewald formula.