Discriminative ability of LDL-cholesterol to identify patients with familial hypercholesterolemia: a cross-sectional study in 26,406 individuals tested for genetic FH

Abstract

Background: Screening for familial hypercholesterolemia (FH) within affected families is often based on cutoff values for low-density lipoprotein cholesterol (LDL-C). However, the diagnostic accuracy of LDL-C levels is influenced by the magnitude of the LDL-C overlap between FH patients and unaffected relatives. The purpose of the current study was to assess to what extent this overlap is influenced by the severity of specific FH mutations.

Methods and results: Individuals were eligible if they underwent family screening for FH between 2003 and 2010. The entire cohort was then compared with those who were investigated for the presence of the most severe mutations (class 1). The area under the receiver operating characteristics curve and the sensitivity of the 90th percentile of LDL-C were calculated for both cohorts. We included 26 406 individuals, of whom 9169 (35%) carried an FH-causing mutation. In the entire cohort at baseline, mean LDL-C was 4.63 ± 1.44 mmol/L for FH carriers (n=5372) and 2.96 ± 0.96 mmol/L for unaffected relatives (n=15 148); P<0.001. The corresponding operating characteristics curve (95% CI) was 86.6% (85.9%-87.2%), and the cutoff level of LDL-C above the 90th percentile showed a sensitivity of 68.5%. The operating characteristics curve and sensitivity significantly improved when the 5933 individuals tested for class 1 mutations were assessed separately; 96.2% (95.3%-97.1%) and 91.3%, respectively.

Conclusions: In summary, the overlap in terms of LDL-C levels between those with molecularly proven FH and unaffected relatives is to a large extent because of the high prevalence of modestly severe LDL-receptor mutations in the Netherlands.