Genetic analyses to test the causal relevance of lipoprotein(a) levels to coronary disease

Heart attacks and strokes account for almost 1 in every 2 adult deaths worldwide, and coronary heart disease is the single leading killer in the UK. Major established causes of heart disease include diabetes, smoking, and higher levels of blood pressure and blood cholesterol. If further blood components can be shown to be additional important causes of heart disease, then this should lead to improvements in the prediction and prevention of the disease, exemplified by measurement and modification of blood cholesterol levels.

Lipoprotein(a) (?Lp[a]?) is a very large protein wrapped around a ?bad cholesterol? molecule that may promote narrowing of arteries that precedes heart attacks. Over the past few decades, there have been many reports of correlations between blood levels of Lp(a) and subsequent risk of heart disease. It is not clear from such studies, however, whether Lp(a) concentration is mainly a cause or a consequence of heart disease.

We have proposed a series of large genetic studies in order to help judge whether Lp(a) is likely to have a cause-and-effect relation relationship with heart disease. Because genetic factors that influence blood Lp(a) level are fixed at conception, such genetic information may be used in sufficiently large studies to help distinguish causality from mere correlation.

In particular, we will characterise the genetic, biochemical and other correlates of blood Lp(a) levels in 20,000 European and South Asian people, in whom comprehensive genetic profiling has already been completed. We will use this information to pinpoint the key genetic determinants of blood Lp(a) levels in the same 20,000 people. These key genetic factors will then be assessed in the stored genetic samples of about 30,000 patients with heart disease and of about 70,000 healthy people. Statistical tests for consistency will be done across these databases to help determine whether genetic information supports a causal role for Lp(a) in heart disease.

The findings of this work should provide important scientific insights, helping to prioritise efforts to develop disease prevention strategies against Lp(a) and to interpret findings from ongoing large clinical trials of niacin (a drug that moderately lowers blood levels of Lp[a] and several other blood fats).

Lipoprotein(a) (Lp[a]), a very large glycoprotein wrapped around an LDL-like particle, has been implicated in coronary heart disease (CHD), but causality remains uncertain. The current proposal involves a series of genetic studies to help test causality by:

(1) assay of Lp(a) levels in 20,000 healthy European and South Asian people in whom genome wide association scans, biochemical profiling and other detailed phenotyping have already been done, enabling powerful and cost-effective characterisation of Lp(a) correlates and distributions;

(2) combination of genome wide and gene?centric approaches in the same 20,000 participants to identify key genetic variants at the LPA gene (the major determinant of Lp[a] levels) that influence circulating levels of Lp(a);

(3) genotyping a relatively small number of key SNPs that show reproducible statistical associations with Lp(a) levels in a total of about 29,000 CHD cases and 69,000 controls in an existing consortium; and

(4) integration of the gene-Lp(a) and gene-CHD data generated in the studies above with data on Lp(a)-CHD associations, such as from the separately funded Emerging Risk Factors Collaboration of 120,000 participants (8500 incident CHD cases) with Lp(a) levels.

Assessment for numerical consistency of these varying sources of data (ie, a ?Mendelian randomisation analysis?) will enable judgement of the likelihood and magnitude of any causal association of Lp(a) concentration with CHD. Findings from the proposed studies should have considerable translational implications for existing drugs (eg, niacin non-specifically lowers Lp[a] levels by about 20%), for the development of new preventative strategies, and for novel mechanistic insights.