Analysis of total cholesterol results measured in the initial period of the Croatian screening program for familial hypercholesterolemia: a pilot study

Abstract Objectives In 2023, Croatia implemented a screening program for familial hypercholesterolemia (FH) targeting children entering the first grade of primary school. The program is based on total cholesterol (TC) concentration measurements. This pilot study aimed to assess and compare TC results obtained using different analytical platforms during the program’s initial phase (2023–2024). Methods Retrospective data from laboratories across Croatia were analyzed. Results covered the following analyzers: Beckman Coulter AU (BC), Abbott Alinity/Architect c (AA), Roche Cobas c (RC) and Siemens Atellica CH (SA). Statistical analyses included Kruskal–Wallis and chi-square tests using the program’s cut-off values. Bland–Altman and Passing-Bablok analyses were used to assess differences in a direct comparison study of 40 samples. To reduce variability, TC values were also expressed as multiples of the median (MoM), calculated separately for each platform. Results The study included 17,694 children. Median TC values (IQR) differed significantly across platforms: BC (4.3 [3.8–4.7] mmol/L), AA (4.1 [3.7–4.5]), RC (4.1 [3.7–4.5]), and SA (4.0 [3.6–4.4]); p<0.001. BC and SA results differed significantly from all other platforms, while no significant difference was found between RC and AA. The distribution of TC results also varied significantly (p<0.001). In the comparison study, all platforms showed statistically significant differences (p<0.001). In contrast, MoM values showed no statistically significant differences between analyzers (p=0.106) or in distribution when applying various MoM-based cut-offs. Conclusions Although variations in TC measurements across platforms are small, they are statistically significant and affect FH classification. MoM normalization may support harmonization in future screening protocols.