Objective: On a big data sets to verify how LDL-cholesterol (LDL-C) is determined indirectly according to Friedewald’s formula dependent on triglycerides (TG) concentration. At second, to verify recommendations about necessary sample size for method comparison studies based on regression analysis. Design: Retrospective observational study. Correlational research. Settings: International ad hoc team of authors from – Dôvera Health Insurance Inc., Bratislava, branch Košice, Slovakia; Departamento de Ciéncias Farmacéuticas, Fundacao Universidade Regional De Blumenau, Blumenau, Santa Catarina, Brazil; Clinic of Neurology, Faculty Hospital L. Pasteur and Medical School UPJŠ, Košice, Slovakia; Department of Endocrinology, University of Ghent, Belgium, and MedCalc Software, Mariakerke, Belgium. Material and Methods: Consecutive 10 664 patients with diferent diseases. Total cholesterol, HDL-cholesterol, TG and LDL-C were measured in years 2000 to 2002. The determination of LDL-C was performed using direct Wako method, with DiaSys LDL-C Select FS sets. Second, LDL-C was calculated according to Friedewald’s formula, if TG were less than 4.5 mmol/l. Intermethod comparison was performed using method of Passing and Bablok. Results: Design of study and principle results are described. Difference between LDL-C determined with direct method and calculated according to Friedewald (F – D) is continuously dependent on TG values, from +0.50 mmol/l to -0.59 mmol/l. Second, after randomisation of patients into 51 subgroups, each with 200 persons, was difference F – D 0.24 mmol/l (95% CI, 0.19 to 0.30). Parameters of regression line according Passing and Bablok were in all the 51 subgroups in 95%CI of parameters of group of 10 200 patients. For method comparison studies sample size of 200 seems to be enough.

        Key words: LDL-cholesterol, direct Wako method, Friedewald’s formula, method comparison study.