Lipoprotein(a): A Genetically Determined, Causal, and Prevalent Risk Factor for ASCVD — AHA Scientific Statement 2021
Authors, Journal, Affiliations, Type, DOI
- Gissette Reyes-Soffer, Henry N. Ginsberg, Lars Berglund, P. Barton Duell, Sean P. Heffron, Pia R. Kamstrup, Donald M. Lloyd-Jones, Santica M. Marcovina, Calvin Yeang, Marlys L. Koschinsky; on behalf of the AHA Council on Arteriosclerosis, Thrombosis and Vascular Biology; Council on Cardiovascular Radiology and Intervention; and Council on Peripheral Vascular Disease
- Arteriosclerosis, Thrombosis, and Vascular Biology (AHA Scientific Statement; endorsed by International Atherosclerosis Society)
- Approved by AHA Science Advisory and Coordinating Committee, August 2021
- DOI: 10.1161/ATV.0000000000000147
Overview
This 2021 AHA Scientific Statement provides comprehensive foundational coverage of Lp(a) — its structure, genetics, measurement challenges, pathophysiology, and emerging therapies. Lp(a) levels are 70–90% genetically determined via KIV2 copy number variation in the LPA gene and remain largely stable across the lifespan. Lp(a) is established as an independent, causal risk factor for ASCVD and calcific aortic valve disease (CAVD) based on mechanistic, observational, and Mendelian randomization evidence. The statement identifies three major clinical gaps: lack of assay standardisation (mg/dL vs nmol/L), absence of universal screening thresholds, and no completed outcomes trial for targeted Lp(a)-lowering therapy.
Keywords
AHA Scientific Statements · apolipoprotein B100 · atherosclerotic cardiovascular disease · cholesterol, low-density lipoprotein · lipoprotein(a)
Key Takeaways
Historical Perspective
- Lp(a) first identified in 1963 by geneticist Kåre Berg as a unique antigen in the LDL fraction; linked to CHD by 1974
- Strong genetic evidence for causal ASCVD role first reported in 2009 via large genetic epidemiological studies
- LPA gene evolved by duplication of the plasminogen (PLG) gene — accounts for structural homology and proposed prothrombotic mechanisms
Structure and Biology
- Lp(a) consists of a lipid-rich core (primarily cholesteryl esters) plus apoB100 and apolipoprotein(a) [apo(a)] covalently bound via a single disulfide bond
- Apo(a) contains 10 kringle IV subtypes (KIV1–KIV10); KIV type 2 is present in multiple copies, creating a highly variable molecular mass (300–800 kDa)
- LPA gene is unique to humans, Old World monkeys, apes, and hedgehogs — limits relevance of animal model data
- Assembly is a 2-step process: (1) noncovalent lysine-dependent interaction between apo(a) KIV and apoB N-terminal domain; (2) covalent disulfide bond formation extracellularly
Determinants of Plasma Lp(a) Levels
- Lp(a) levels are 70% to ≥90% genetically determined
- KIV2 copy number variant (CNV): inversely related to Lp(a) concentration (more repeats → larger apo(a) → lower Lp(a)); accounts for 19–69% of interindividual heterogeneity
- Numerous LPA SNPs also independently associate with Lp(a) levels, including some not in linkage disequilibrium with KIV2 CNV
- APOE ε2 allele associates with lower Lp(a) levels (~0.5% of variation); APOH (β2-glycoprotein 1) also associated
- Black individuals of African descent and South Asian populations have higher median Lp(a) than White or East Asian individuals
- Lp(a) levels are stable across the lifespan; intraindividual biological variability up to 20% — consider averaging 2 measurements for borderline cases
- Large apo(a) isoforms are retained longer in the endoplasmic reticulum and subjected to increased proteasomal degradation, explaining the inverse isoform size–plasma level relationship
- Liver is the primary site of Lp(a) catabolism; key clearance receptors not definitively identified (LDL-R may play a role with statins + PCSK9 inhibitors; SR-B1, LRP1, LRP8 implicated)
Quantification and Assay Standardisation
- Lp(a) measured by immunoassays using antibodies specific to apo(a); two major problems:
- Apo(a) size variability: different isoform sizes cause under- or overestimation with some assays; isoform-insensitive assays (5-calibrator ELISA) partially mitigate this
- Dual units: historical calibration in mg/dL (total Lp(a) mass — assumes constant component proportions, which is invalid); gold-standard ELISA calibrated in nmol/L of apo(a) (number of Lp(a) particles)
- Reference standard: WHO/IFCC SRM-2B — expressed in nmol/L; multiple guidelines recommend traceable assays in nmol/L
- No unbiased conversion factor exists from mg/dL to nmol/L — clinicians should not convert between units
- A mass spectrometry candidate reference method now validated with high correlation to gold-standard ELISA
- Lp(a) cholesterol is included in all clinical LDL-C assays (β-quantification); attempts to subtract 30% of Lp(a) mass are unreliable due to variable cholesterol content
Genetic Approaches to Ascertaining Risk
- Lp(a) is an excellent Mendelian randomization candidate due to its high genetic determination
- Key Mendelian randomization findings:
- Each 2-fold higher genetically determined Lp(a) associated with 22% greater risk of MI (n>40,000; Kamstrup 2009)
- LPA SNPs with strongest CHD association in large case-control study (PROCARDIS, n>3,100; Clarke 2009)
- LPA SNP associated with CAVD: OR 2.05 per allele (GWAS 2013) — stronger than CHD association
- LPA is the only monogenic risk factor identified for aortic stenosis
- Stroke association: weaker signal; 13% lower stroke risk per 1 SD lower genetically determined Lp(a) vs ~30–40% lower CHD/PAD/CAVD risk
Cardiovascular Risks and Risk Modification
- UK Biobank (Patel et al): n=460,000; median Lp(a) 19.6 nmol/L overall (19, 31, 75, 16 nmol/L in White, South Asian, Black, Chinese participants; women 22 vs men 17 nmol/L)
- Risk is log-linear above the median, with increasing risk at higher concentrations
- HR 1.11 per 50 nmol/L increment (95% CI 1.10–1.12) for ASCVD, independent of traditional risk factors; similar across all ancestry groups
- Practical formula for Lp(a)-adjusted 10-year ASCVD risk (PCE framework):
Updated 10-y risk = Predicted PCE risk × 1.11^(Lp(a) in nmol/L ÷ 50)
Example: 10% × 1.11^(250÷50) = 10% × 1.69 = 16.9% - Apo(a) size independent of Lp(a) level: conflicting evidence; not currently used for clinical risk stratification — molar Lp(a) concentration sufficient
Pathophysiology
- Atherogenesis: apoB moiety of Lp(a) similar to LDL; Lp(a) selectively retained in arterial wall via apo(a) binding to extracellular matrix proteins; carries oxidised phospholipids (OxPLs) covalently bound to apo(a) — recognised by innate immune system as danger signals → sterile inflammation
- CAVD: OxPLs → upregulate procalcific and osteogenic genes in valve interstitial cells (VICs); Lp(a) + apo(a) components colocalise with CAVD calcification; autotaxin (ATX) and its product lysophosphatidic acid implicated; 18F-NaF PET shows enhanced aortic valve calcification activity with elevated Lp(a)
- Inflammation: FDG-PET demonstrates enhanced arterial wall inflammation with elevated Lp(a); Lp(a)-lowering with ASO reduces monocyte inflammatory gene expression and transendothelial migration
- Thrombosis: apo(a) has plasminogen-like domains but inhibits fibrinolysis in vitro; however, ex vivo clot lysis times unchanged after Lp(a) lowering with ASO — antifibrinolytic properties may be masked when apo(a) is covalently bound to apoB; VTE association not supported by genetic data (except very high Lp(a) levels)
Lp(a)-Lowering Therapies
| Therapy | Lp(a) Reduction | Status (2021) | Notes |
|---|---|---|---|
| Lipoprotein apheresis | 50–85% acutely (every 2 weeks) | FDA-approved: FH+CAD, LDL-C >100 mg/dL; Lp(a) ≥60 mg/dL indication | Also lowers LDL 60–85%; definitive CV outcomes data lacking |
| PCSK9 inhibitors (alirocumab, evolocumab) | 25–30% | Approved for LDL lowering | Post hoc: alirocumab Lp(a) lowering independently contributed to MACE reduction (only in Lp(a) >13.7 mg/dL and LDL-C <70 mg/dL at baseline) |
| Niacin | 25–40% | Available | CV benefit unknown; adverse effects with statins concern |
| Statins | Minimal (may modestly increase) | Approved | Do not lower Lp(a); still indicated for overall LDL/ASCVD management |
| Pelacarsen (TQJ230, IONIS-APO(a)-LRx) | ~80% at 20 mg SC weekly | Phase 3 CVOT (NCT04023552, n=7,682) | First ASO targeting LPA expression; also reduces monocyte inflammation |
| ARO-LPA (AMG 890) | Not reported | Phase 2 (NCT04270760) | siRNA targeting apo(a) mRNA |
| SLN360 | Not reported | Phase 1 (NCT04606602) | siRNA targeting apo(a) mRNA |
- No completed RCT has demonstrated that targeted Lp(a) lowering reduces adverse cardiovascular outcomes at time of publication (2021)
- Dietary intervention: very modest effect on Lp(a) levels
Limitations of the Document
- No cardiovascular outcomes data for targeted Lp(a)-lowering therapy — all causal evidence is mechanistic, observational, or Mendelian randomization
- Assay non-standardisation limits cross-study and cross-population comparison
- Lack of a proper animal model (species restriction of LPA) limits mechanistic studies
- Stroke data are inconsistent across studies; causal role less firmly established than for CHD or CAVD
- Racial/ethnic differences in Lp(a) biology and risk not fully characterised; most data from European-descent cohorts
Key Concepts Mentioned
- concepts/Lipoprotein-a — core subject of the statement
- concepts/ASCVD-Risk-Assessment — Lp(a) as risk-enhancing factor; UK Biobank risk formula
- concepts/Aortic-Stenosis — LPA is the only monogenic risk factor for aortic stenosis; OxPL-mediated CAVD
- concepts/Gene-Silencing-Therapy — pelacarsen (ASO), ARO-LPA and SLN360 (siRNA) targeting LPA expression
Key Entities Mentioned
- entities/ATTR-Amyloidosis — not referenced
- UK Biobank (Patel et al) — largest study of Lp(a) and ASCVD risk by ancestry (n=460,000)
Wiki Pages Updated
wiki/sources/lpa-aha-2021.md— createdwiki/concepts/Lipoprotein-a.md— updated: structural biology, assay standardisation, risk formula, source_count incrementedwiki/concepts/ASCVD-Risk-Assessment.md— updated: PCE-era Lp(a) risk formula addedwiki/sourceindex.md— updatedwiki/wikiindex.md— updated