Boston, MA - A genotype score derived from a combination of validated single-nucleotide polymorphisms (SNPs) previously associated with cholesterol levels has been shown to be an independent risk factor for incident cardiovascular disease, a new study has shown [1]. While the score did not improve risk discrimination among the general cohort, it did improve clinical reclassification for subjects considered intermediate risk, with the addition of the score reclassifying 26% of patients into a higher- or lower-risk category.

"This is really a proof-of-concept study based on SNPs that have been previously reported," lead investigator Dr Sekar Kathiresan (Massachusetts General Hospital, Boston) told heartwire. "The field is moving very rapidly, and there are many more variants for lipids, as well as for heart attack, and these will all need to be integrated into the score. This paper shows that the combination of multiple polymorphisms is likely going to be a useful way to identify individuals at high risk and then possibly focus on more aggressive treatment for those considered high risk."

Publishing their findings in the March 20, 2008 issue of the New England Journal of Medicine, the investigators focused on nine gene variants shown to have a role in the modulation of cholesterol levels, including six SNPs associated with LDL cholesterol from APOB, PCSK9, LDLR, CETP, LIPC, and LDL. The investigators analyzed data from 5414 Swedish adults participating in the cardiovascular cohort of the Malmö Diet and Cancer Study. They first validated the association between SNPs and LDL or HDL cholesterol and then created a genotype score, ranging from 0 to 18, based on the number of unfavorable alleles patients carried.

Investigators report that with increasing genotype scores, LDL-cholesterol levels rose, ranging from 152 mg/dL to 171 mg/dL, while HDL cholesterol levels decreased. During a median follow-up of 10.6 years, 238 subjects had a first cardiovascular event, and investigators report that the genotype score was associated with incident disease in a model adjusted for other covariates, including baseline lipid levels.

Testing for the panel of nine SNPs was not better than standard risk factors for predicting cardiac events in the overall population, with the use of the genotype score not improving clinical risk prediction as assessed by the C statistic, a measure of the area under the receiver operating characteristic (ROC) curve.

"At the population level, above and beyond the standard risk factors, the score based on the nine variants did not improve the C statistic," said Kathiresan. "But if you look at specific subgroups within the population—let's say, groups that are intermediate risk for heart disease based on standard risk factors—using the genotype score based on the nine SNPs did improve the classification of individuals who were truly high risk vs those who were low risk."

In addition, one advantage of genetics for risk prediction is that identifying individuals based on genotype at a young age would allow clinicians to focus on either lifestyle or pharmacologic interventions, something that would be useful regardless of whether or not it improves risk prediction at the population level, said Kathiresan.

Not included in the genotype score were six new gene loci related to LDL cholesterol, HDL cholesterol, and triglycerides discovered recently, a discovery made through the joint efforts of Kathiresan's group and other researchers at the University of Michigan. Those findings, reported by heartwire, were recently published in Nature Genetics [2,3].

"Two major things are needed to go forward," he said. "We need to discover all of the variants related to lipid levels and heart-attack risk. This is just the start. There are only nine variants here. We have already discovered additional alleles, those in the Nature paper, that are not part of the genotype score."

Kathiresan anticipates that in one year or so there will be approximately 50 to 100 validated genetic variants related to risk factors such as cholesterol or directly related to MI risk without working through LDL- and HDL-cholesterol modulation, such as variants at chromosome 9, and these will be incorporated into a genotype score for risk assessment. Then, future studies will need to show that the better score relates to heart-attack risk, he said, as well as determine whether acting on this new genetic knowledge improves clinical outcomes.