Abstract 435: An Electronic Health Records Query for Severe Hypercholesterolemia Identifies Individuals With Undiagnosed Clinical and Molecular FH at a Tertiary Academic Centre

Abstract

Background: Familial hypercholesterolemia (FH) is a common heritable disorder of elevated low density lipoprotein cholesterol (LDL-C) with an estimated prevalence of 1/200-300 in the US; however, fewer than 10% of cases have been identified. We wanted to examine whether a simple EHR query for severe hypercholesterolemia could be used to identify clinically and genetically defined cases of FH. Objectives/Purpose: We tested the hypothesis that querying the EHR using an LDL-C criterion would be a novel way to screen for and ultimately identify undiagnosed cases of FH. Methods: An EHR screening query was used to identify active adult patients with LDL-C ≥ 220 mg/dL in the University of Pennsylvania outpatient EHR database. Patients with secondary causes of hypercholesterolemia and those who had previous genetic testing for FH were excluded. The query identified 3,475 individuals, 120 were subsequently consented and enrolled for molecular testing. This was performed with next-generation sequencing using Progenika’s SEQPRO LIPO IS platform, targeting LDLR, APOB, PCSK9 and LDLRAP1 . A literature search was performed to gather information on identified LDLR variants of unknown significance (VUS). In addition, in-silico analysis was employed to evaluate the pathogenicity of the LDLR and LDLRAP1 VUS. Results: Among the 120 subjects, 53 (44.2%) met the Dutch Lipid Clinic Network (DLCN) criteria for probable or definite clinical FH. Molecularly, 19 FH-related pathogenic mutations were found in 18 (15%) individuals. Four had a common APOB ( R3500Q) mutation, 14 had a LDLR mutation. One individual had a double heterozygous mutation in PCSK9 and LDLR. In addition, 17 LDLR VUS were identified in 16 (13.9%) individuals. A literature review and in-silico analysis predicted that 8 VUS found in 10 subjects were “disease causing”. Therefore, a total of 28 (24.3%) subjects from our cohort carried either a FH causal mutation or likely pathogenic variant. Overall, 59 subjects (49.2%) in our cohort were ascertained to have either a clinical or molecular diagnosis of FH. Conclusion: The use of an EHR screening query for severe hypercholesterolemia was a novel, low investment but relatively high yield approach for identifying undiagnosed cases of FH at a tertiary academic centre.