Abstract
BACKGROUND: The diagnosis of familial hypercholesterolemia (FH) in children is primarily based on main criteria including low-density lipoprotein cholesterol (LDL-C) levels, increased in the proband and relatives, and its inheritance. Two other relevant parameters are symptoms, rarely occurring in children, as rare are the FH homozygous patients, and the mutation detection of related genes. The latter allows the final diagnosis, although it is not commonly available. Moreover, the application of diagnostic scores, useful in adults, is poorly applied in children. The aim of this study was to compare the reliability of criteria here applied with different scores, apart from genetic analysis, for FH diagnosis. The latter was then confirmed by genetic analysis. METHODS: n. 180 hypercholesterolemic children (age 10.2 ± 4.6 years) showing LDL-C levels ≥95th percentile (age- and sex-related), the dominant inheritance pattern of hypercholesterolemia (including LDL-C ≥95th percentile in one parent), were considered potentially affected by FH and included in the study. The molecular analysis of the LDLR, APOB and PCSK9 genes was applied to verify the diagnostic accuracy. Biochemical and family history data were also retrospectively categorized according to European Atherosclerosis Society (EAS), Simon Broome Register (SBR), Pediatric group of the Italian LIPIGEN (LIPIGEN-FH-PED) and Dutch Lipid Clinic Network (DLCN) criteria. Detailed kindred biochemical and clinical assessments were extended to three generations. The lipid profile was detected by standard laboratory kits, and gene analysis was performed by traditional sequencing or Next-Generation Sequencing (NGS). RESULTS: Among 180 hypercholesterolemic subjects, FH suspected based on the above criteria, 164/180 had the diagnosis confirmed, showing causative mutations. The mutation detection rate (MDR) was 91.1%. The scoring criteria proposed by the EAS, SBR and LIPIGEN-FH-PED (resulting in high probable, possible-defined and probable-defined, respectively) showed high sensitivity (~90%), low specificity (~6%) and high MDR (~91%). It is noteworthy that their application, as a discriminant for the execution of the molecular investigation, would lead to a loss of 9.1%, 9.8% and 9.1%, respectively, of FH-affected patients, as confirmed by the genetic analysis. DLCN criteria, for which LDL-C cut-offs are not specific for childhood, would lead to a loss of 53% of patients with mutations. CONCLUSIONS: In the pediatric population, the combination of LDL-C ≥95th percentile in the proband and the dominant inheritance pattern of hypercholesterolemia, with LDL-C ≥95th percentile in one parent, is a simple, useful and effective diagnostic criterion, showing high MDR. This pattern is crucial for early FH diagnosis. EAS, SBR and LIPIGEN-FH-PED criteria can underestimate the real number of patients with gene mutations and cannot be considered strictly discriminant for the execution of molecular analysis.