Abstract
BACKGROUND: Risk at high-density lipoprotein cholesterol (HDL-C) extremes may reflect genetic predisposition, ancestry, or region, but their relative contributions remain unknown. We characterized extreme HDL-C associations with atherosclerotic cardiovascular disease and mortality, and examined whether genetic-phenotypic discordance identifies risk beyond measured HDL-C. METHODS: Atherosclerotic cardiovascular disease–free participants from UKB (UK Biobank) and All of Us (AoU) with extremely low (2.3 mmol/L; based on standard clinical guidelines) HDL-C were analyzed. A 247-variant polygenic score stratified European ancestry participants at each HDL-C extreme into tertiles; discordant groups had measured HDL-C contradicting genetic predisposition. RESULTS: Among 408 353 UKB and 115 423 AoU participants, extremely low HDL-C was associated with higher atherosclerotic cardiovascular disease (UKB: hazard ratio [HR], 1.33 [95% CI, 1.27–1.39]; AoU: HR, 1.32 [95% CI, 1.08–1.61]) and mortality (UKB: HR, 1.25 [95% CI, 1.20–1.30]; AoU: HR, 1.64 [95% CI, 1.37–1.96]) across ancestries. Extremely high HDL-C was associated with excess mortality in UKB European ancestry (HR, 1.28 [95% CI, 1.19–1.38]), driven by COPD (HR, 2.64 [1.94–3.59]) and lung infections (HR, 2.24 [95% CI, 1.48–3.39]), but not in AoU European ancestry. In UKB, this was confined to genetically discordant high HDL (high-density lipoprotein; polygenic score HDL T1; HR, 1.53 [95% CI, 1.22–1.92]); genetically concordant high HDL showed progressively lower mortality. Genetically discordant low HDL (polygenic score HDL T3) carried higher atherosclerotic cardiovascular disease (UKB: HR, 1.20 [95% CI, 1.08–1.34]; AoU: HR, 1.44 [95% CI, 1.15–1.81]) and mortality (UKB: HR, 1.22 [95% CI, 1.11–1.34]; AoU: HR, 1.75 [95% CI, 1.21–2.54]) than concordant low HDL (polygenic score HDL T1), comparable in magnitude to a 10 mm Hg increment in systolic blood pressure. CONCLUSIONS: Extreme HDL-C levels carry different cardiovascular and mortality risk depending on whether they reflect or contradict genetic predisposition, with additional variation by ancestry and region.