Combined elevation of lipoprotein(a) and polygenic risk in index patients identifies family members at high risk of coronary artery disease
M E Annink, E S G Stroes, L F ReeskampAbstract
Background
Lipoprotein(a) [Lp(a)] and genome-wide polygenic risk scores (PRS) are heritable, lifelong markers of coronary artery disease (CAD) risk. Guidelines recommend measuring Lp(a) at least once in adulthood, and clinical implementation of PRS is advancing. When elevated levels are found in an index patient, their first-degree relatives (FDRs) are expected to share a substantial proportion of this genetic CAD risk, yet evidence to support cascade screening for these markers is lacking. Whether index Lp(a) and PRS predict clinically meaningful CAD risk elevation in FDRs, and whether their combined assessment identifies family members at particularly high risk, has not been established.
Purpose
To determine whether Lp(a) and a validated CAD PRS measured in index participants independently and jointly predict incident CAD in FDRs.
Methods
We performed a family-based cohort study in the UK Biobank. FDR pairs were identified using genetically inferred kinship, with the older eligible individual assigned index. Index participants required availability of Lp(a) and PRS, and no prevalent atherosclerotic cardiovascular disease (ASCVD); FDRs were required to be free of prevalent ASCVD. Exposures were index Lp(a) and a validated genome-wide CAD PRS (GPSMult), excluding variants in the LPA region. Cox models estimated associations across percentile strata and per standard deviation (SD) increase, adjusted for FDR age, sex, smoking, systolic blood pressure, diabetes, body mass index, LDL-cholesterol, HDL-cholesterol, triglycerides, glomerular filtration rate, and statin use. Joint effects were assessed using four categories based on high (≥80th percentile) versus low (<80th percentile) Lp(a) and GPSMult. Continuous net reclassification improvement (NRI) was evaluated.
Results
Among 27,333 FDR pairs (25,480 unique index participants; 25,062 unique FDRs), 1,046 incident CAD events occurred in FDRs over median 13.9 years of follow-up. Index and FDR Lp(a) and GPSMult correlated (r=0.38 and r=0.51, resp.). Adjusted hazard ratios (aHR) for CAD in FDRs increased across index Lp(a) strata (Ptrend=0.002), reaching 1.40 (95% CI 1.14–1.70) for the top decile; per-SD aHR 1.10 (1.04–1.17). Index GPSMult showed a steeper gradient (P-trend<0.001), with the top decile showing an aHR of 1.74 (1.43–2.13); per-SD aHR 1.24 (1.17–1.33). Combined elevation of these index markers identified FDRs at markedly increased CAD risk: aHR 2.14 (1.43–3.20; P<0.001), with non-significant interaction (P=0.50) indicating additive effects. Adding index Lp(a) and GPSMult to a model with FDR clinical risk factors correctly reclassified 21% of FDRs (NRI 0.21, 95% CI 0.13–0.28).
Conclusion
Lp(a) and polygenic CAD risk in index patients independently predict incident CAD in FDRs, with combined elevation more than doubling risk. These findings support cascade screening of FDRs when elevated Lp(a) or polygenic risk is identified, extending this strategy beyond monogenic disorders.Figure 1Figure 2