Substantial strides have been made in the area of genetic diagnoses, particularly for rare diseases such as Huntington’s chorea (affecting 5 per 100,000 persons) in which individuals with a pathogenic mutation are virtually assured of developing the disease during their lifetimes.
Dr. Mathew R.G. Taylor
Progress in identifying genetic risks for more common diseases has been slower, although some headway has been made with the introduction of genomewide association studies. Of the common diseases, perhaps none has greater impact in Western society than do the cardiovascular diseases, including coronary artery disease, which plagues more than 16,000,000 Americans.
In the case of premature myocardial infarctions in younger individuals, genetic factors may indeed account for the majority of the disease risk. The expectation is that eventually, some common genetic variants that confer modestly elevated coronary disease risk will be identified to explain these observations.
In the past 2 months, two studies have been reported that reflect both the promise and pitfalls of research efforts to identify genetic-based risk factors for coronary disease.
In the past 5 years, several groups have published studies involving large cohorts of patients with CAD that link variants in the kinesinlike protein 6 (KIF6) gene to coronary risk. One variant in particular, Trp719Arg, was shown in several studies to increase the risk of CAD in four randomized, controlled trials. In each case, the genetic work was done as a substudy, leveraging DNA that is increasingly collected from subjects in large studies. These early studies led to others that found data supporting a risk reduction that was mediated through gene-drug interaction with statins.
A meta-analysis of seven trials estimated that carriers of the 719Arg allele had a 20% increased risk of CAD, compared with noncarriers. As the evidence accumulated for KIF6’s role, companies began to look at the development of clinical assays that culminated in a KIF6 genetic-testing assay that was marketed to cardiologists and internists. To date, more than 150,000 tests have been ordered in the United States, with plans to extend this to the European market.
However, a large and well-designed meta-analysis of 19 studies, conducted by Dr. Themistocles L. Assimes of Stanford (Calif.) University and colleagues and published last month, refutes much of the previous work, seriously calling into question the validity of the KIF6–coronary artery disease link as well as the clinical testing activity that was built around KIF6 (J. Am. Coll. Cardiol. 2010 Oct 7 [doi:10.1016/j.jacc.2010.06.022]).
The meta-analysis pooled data from 17,000 cases and 39,369 controls of largely European descent, and could find no contribution of KIF6 to MI risk either in the analysis of all subjects or in a subgroup analysis that focused on early-onset disease.
In an accompanying editorial, Dr. Eric Topol and Dr. Samir Damani of Scripps Translational Science Institute and the Scripps Research Institute, La Jolla, Calif., pointed out that a reliance on biased methods (such as candidate gene analysis) and a lack of a known biological mechanism for KIF6 to increase coronary disease risk likely contributed to earlier findings. Furthermore, Dr. Assimes’s study measured actual MI events, a firm end point, instead of ‘softer’ end points such as imaging evidence of CAD.
Although the recent study will likely stimulate further discussions of the KIF6 data, it appears that the outcome of this story is a \"step backward\" and the lesson a cautionary one in terms of translating genetic findings into the clinical practice.
A more optimistic story is building behind data presented by Steven Rosenberg, Ph.D., of CardioDx Inc. in Palo Alto, Calif., and colleagues showing that a specific pattern of gene expression has the potential to predict existing coronary artery disease (Ann. Intern. Med. 2010;153:425-34).
In this study, the team built on the earlier identification of 23 human genes that were differentially expressed in CAD, compared with those in healthy individuals. They performed a validation study of their predictive algorithm in 526 nondiabetic patients who underwent clinical angiography. The authors used the Diamond-Forrester risk score, which relies on age, sex, and chest pain type to assess CAD risk. They asked whether the addition of the gene-expression algorithm improved the sensitivity of predicting actual coronary artery disease.
Overall, the combined model that integrated gene-expression findings with the Diamond-Forrester score performed better than did the Diamond-Forrester alone.
The accompanying editorial cautioned that this work is in its early stages and needs to be studied further before it is released into clinical practice. However, as the gene-expression profile avoids the need for radiation and contrast agents, it remains an attractive "step forward" in the field. Whether and how soon this type of gene expression profiling will infiltrate cardiology and internal medicine practices remain sources of speculation.