Abstract

Background Most people with autosomal dominant familial hypercholesterolaemia (FH) remain undetected, which represents a missed opportunity for coronary heart disease prevention. Objective To evaluate the performance of two-stage adult population screening for FH. Design Using data from UK Biobank, we estimated the screening performance of different low-density lipoprotein cholesterol (LDL-C) cut-offs (stage 1) to select adults for DNA sequencing (stage 2) to identify individuals with FH-causing variants in LDLR, APOB, PCSK9 and APOE. We estimated the number of additional FH cases detected by cascade testing of first-degree relatives of index cases and compared the overall approach with screening in childhood. Setting UK Biobank. Participants 140 439 unrelated participants of European ancestry from UK Biobank with information on circulating LDL-C concentration and exome sequence. Main outcome measures For different LDL-C cut-offs, we estimated the detection and false-positive rate, the proportion of individuals who would be referred for DNA sequencing (stage 1 screen positive rate), and the number of FH cases identified by population screening followed by cascade testing. Results We identified 488 individuals with an FH-causing variant and 139 951 without (prevalence 1 in 288). An LDL-C cut-off of >4.8 mmol/L had a stage 1 detection rate (sensitivity) of 40% (95% CI 36 to 44%) for a false-positive rate of 10% (95% CI 10 to 11%). Detection rate increased at lower LDL-C cut-offs but at the expense of higher false-positive and screen positive rates, and vice versa. Two-stage screening of 100 000 adults using an LDL-C cut-off of 4.8 mmol/L would generate 10 398 stage 1 screen positives for sequencing, detect 138 FH cases and miss 209. Up to 207 additional cases could be detected through two-generation cascade testing of first-degree relatives. By comparison, based on previously published data, childhood screening followed by cascade testing was estimated to detect nearly three times as many affected individuals for around half the sequencing burden. Conclusions Two-stage adult population screening for FH could help achieve the 25% FH case detection target set in the National Health Service Long Term Plan, but less efficiently than childhood screening and with a greater sequencing requirement.