Prevalence, penetrance, and coronary heart disease risk of familial hypercholesterolemia across diverse ancestries

Background Contemporary prevalence and penetrance of familial hypercholesterolemia (FH) and the associated coronary heart disease (CHD) risk across diverse genetic ancestries in the United States are unclear. Purpose To estimate prevalence of individuals harboring pathogenic/likely pathogenic (P/LP) variants in LDLR, APOB, and PCSK9, distribution of low-density-lipoprotein cholesterol (LDL-C) levels, and associated CHD risk for carriers in an ancestrally diverse cohort. Methods In 242,167 All of Us (AoU) participants with available whole-genome sequencing data, we obtained LDL-C measurements, statin prescriptions, and ascertained P/LP variants in FH-related genes and presence of CHD based on electronic-health record data. P/LP variants in LDLR, APOB, and PCSK9 were identified in AoU data by filtering for high-impact (frameshift, canonical splice site, indels, and stop gain) and predicted-deleterious missense variants, then selecting those classified as P/LP in ClinVar. Untreated LDL-C levels were estimated assuming a 30% reduction if there was a record of a statin prescription in the preceding 12 months and the highest value among multiple measurements for each individual were selected. Penetrance was defined as LDL-C>155 mg/dl. CHD was determined using a validated phenotyping algorithm based on diagnostic and procedural billing codes. Logistic regression models were used to obtain prevalence and penetrance estimates and odds ratios for CHD, along with associated 95% confidence intervals (95% CI). Results We identified 221 FH-related P/LP variants (LDLR=169, APOB=50, PCSK9=2) in the study cohort. Of 125,841 European (EUR), 53,367 African (AFR), 40,189 admixed American (AMR), and 22,770 other ancestry individuals, 1:218 (0.46%, 95% CI 0.42%-0.50%), 1:307 (0.33%,0.28%-0.38%), 1:350 (0.29%, 0.24%-0.34%), and 1:201 (0.50%, 0.41%-0.59%) were carriers of a P/LP variant. In 92,616 individuals with at least one LDL-C measurement, carriers of a P/LP variant had a higher distribution of LDL-C across all ancestries with similarly high penetrance (Figure 1). Harboring a P/LP variant in LDLR, APOB, and PCSK9 was associated with similarly increased odds of CHD across all the different ancestry groups (Figure 2). Conclusions Monogenic FH was prevalent, penetrant, and significantly associated with CHD similarly across diverse ancestry groups in the AoU cohort. These findings support a ‘genotype first’ approach to detection of FH to enable early intervention across diverse ancestry groups.