Objective: We aimed to investigate the possible differences between the analytical assay and estimation with Friedewald formula of LDL cholesterol in patients divided into subgroups according to direct LDL cholesterol, triglyceride, total cholesterol levels, retrospectively.

Materials and Methods: Patients with diagnosis of HIV infection, chronic renal failure or chronic hepatitis were excluded; data of 568 patients (358 female, 210 male) were statistically analyzed. CV values of LDL-cholesterol were 4.95% and 4.91% for level-1, level-2 control materials. % Bias was estimated using the equation:[(calculated LDL-cholesterol) - (direct LDL-cholesterol) / (direct LDL-cholesterol)] x 100. TEa for LDL-cholesterol was 12% according to National Cholesterol Education Program and ratio of staying at within acceptable range of LDL-cholesterol values calculated with friedewald formula for each patient was determined.

Results: The mean bias (±SD) was - 12 ± 7% in the subgroup with triglyceride 150-199 mg/dL, - 17 ± 8% in the subgroup with triglyceride >199 mg/dL. Estimated LDL-cholesterol was in acceptable TEa range in 62.3% of the subjects with triglyceride 150-199 mg/dL, 32.3% of the subjects with triglyceride >199 mg/dL, respectively.

Conclusion: LDL-cholesterol estimated with Friedewald formula in subjects with triglyceride <400 mg/dL can be preferred because this assay has been used in clinical trials documenting the benefits of cholesterol-lowering therapy in routine practice. But measurement of direct LDL-cholesterol instead of Friedewald formula in subjects with triglyceride levels between 200-400 mg/dL should be evaluated after considering cost-benefit analysis.