OBJECTIVE We evaluated factors that caused differences between calculated low-density lipoprotein cholesterol (C-LDL-C) and direct-measured LDL-C (D-LDL-C) and compared them in subjects with diabetes mellitus (DM) or taking lipid-lowering medications.

METHODS 21,452 subjects (9,177 women, 12,275 men; 8.1% with DM and 8.5% on lipid-lowering medications) were included in the analysis. Participants were classified into 3 groups, i.e., group 1: the subjects without DM and not on lipid-modifying drugs (n = 18,287), group 2: without DM and on lipid-modifying drugs (n = 1,423), and group 3: with DM (n = 1,742). LDL-C concentrations were either directly measured by a homogenous method or calculated by Friedewald formula.

RESULTS There was a significant correlation between C-LDL-C and D-LDL-C (r = 0.966, P < .001). The absolute values of the differences between two LDL-C values were 7.0 ± 6.2 mg/dl and 6.6 ± 7.3% (6.6 ± 5.9 mg/dl and 6.0 ± 6.5%, 8.8 ± 6.7 mg/dl and 9.1 ± 9.7%, and 10.1 ± 7.3 mg/dl and 10.7 ± 10.1% in group 1, 2, and 3 respectively, P < .001). The subjects with the absolute value of the differences of LDL-C ≥10% was 20.2% (17.3%, 31.3%, and 41.1% in group 1, 2, and 3 respectively, P < .001). In the multiple logistic regression analysis, high triglyceride (≥150 mg/dl), low high-density lipoprotein cholesterol (HDL-C) (<40 mg/dl), male gender, obesity (body mass index ≥25 kg/m(2)), DM and taking lipid-lowering drugs were significant associated with high LDL-differences (the absolute value of the differences ≥10% or ≥10 mg/dl).

CONCLUSION D-LDL-C was generally higher by 5 mg/dl or 5% than C-LDL-C. The differences C-LDL-C and D-LDL-C were higher in subjects with DM and on lipid-lowering medications. Male gender, high triglyceride, low HDL-C, and obesity were also associated with the greater differences between C-LDL-C and D-LDL-C.