Genetic variants associated with Lp(a) lipoprotein level and coronary disease
Robert Clarke, John F Peden, Jemma C Hopewell, Theodosios Kyriakou, Anuj Goel, Simon C Heath, Sarah Parish, Simona Barlera, Maria Grazia Franzosi, Stephan Rust, Derrick Bennett, Angela Silveira, Anders Malarstig, Fiona R Green, Mark Lathrop, Bruna Gigante, Karin Leander, Ulf de Faire, Udo Seedorf, Anders Hamsten, Rory Collins, Hugh Watkins, Martin Farrall, PROCARDIS Consortium - The New England Journal of Medicine, 2009
Two common LPA gene variants were strongly associated with both higher Lp(a) lipoprotein levels and higher coronary-disease risk. Crucially, after adjustment for the measured Lp(a) level the genotype-score association with coronary disease was abolished, which the authors interpret as supporting a causal role for Lp(a). This genotype-based (Mendelian-randomization-style) result favours the causal lipid hypothesis specifically for Lp(a).
Key findings
- Two LPA variants were each associated with both higher Lp(a) and higher coronary-disease risk: rs10455872 (odds ratio 1.70, 95% CI 1.49-1.95) and rs3798220 (odds ratio 1.92, 95% CI 1.48-2.49).
- A genotype score combining the two SNPs gave odds ratios of 1.51 (95% CI 1.38-1.66) for one variant and 2.57 (95% CI 1.80-3.67) for two or more variants.
- After adjustment for the Lp(a) lipoprotein level, the association between the LPA genotype score and coronary disease was abolished, which the authors read as supporting a causal role for Lp(a).
- Discovery sample of 3145 cases and 3352 controls (PROCARDIS), with replication adding 4846 cases and 4594 controls (about 15,937 total).
Effect measures
- Odds Ratio: 1.7095% CI 1.49-1.95
- Odds Ratio: 1.9295% CI 1.48-2.49
- Odds Ratio: 1.5195% CI 1.38-1.66
- Odds Ratio: 2.5795% CI 1.80-3.67
Why this evidence tier (High)
- Risk of bias:
- Large multi-stage genetic-association study with independent replication; using LPA genotype as an instrument and testing mediation by the Lp(a) level gives Mendelian-randomization-style protection against confounding and reverse causation.
- Precision:
- Tight, significant odds ratios across discovery and replication (e.g. rs10455872 OR 1.70, 95% CI 1.49-1.95).
- Directness:
- Coronary-disease endpoint and an Lp(a)-specific genetic exposure; the mediation analysis (effect abolished after adjusting for Lp(a)) speaks directly to causality.
- Consistency:
- Concordant with Kamstrup 2009 and the broader Lp(a) literature.
- Funding / COI:
- Funded by the Wellcome Trust, the UK Medical Research Council (grant MC_U137686857), and the British Heart Foundation. No structured COI statement is shown on the PubMed record, and full author disclosures held by NEJM were not retrievable (NEJM full text returned HTTP 403 this session).
High certainty supporting a causal role for Lp(a) in coronary disease, established through genotype-based (Mendelian-randomization-style) inference with replication.
- Population:
- European adults: a discovery sample of 3145 case subjects with coronary disease and 3352 controls (PROCARDIS), with replication in three independent populations adding 4846 cases and 4594 controls (about 15,937 total).
- Conflicts of interest:
- No structured conflict-of-interest statement is displayed on the PubMed record (PMID 20032323); full author disclosure forms are held by NEJM and were not retrievable this session because the NEJM full text returned HTTP 403. Funding came from the Wellcome Trust, the UK Medical Research Council, and the British Heart Foundation. No retraction or erratum was found.
- Funding:
- Wellcome Trust, UK Medical Research Council (grant MC_U137686857), and the British Heart Foundation.
Limitations
- European-ancestry case-control populations, limiting generalizability to other ancestries.
- Causal inference rests on a genotype-score mediation argument (the association is abolished after adjusting for the Lp(a) level) rather than a randomized intervention.
- Full per-author conflict-of-interest disclosures were not retrievable this session because the NEJM full text was paywalled (HTTP 403).