Can a Genetic Risk Score Predict Multiple Sclerosis?
K. Claire Simon, Ph.D.
Harvard School of Public Health
T32ES016645
A large, multinational team of epidemiologists has developed a promising mathematical algorithm for predicting the likelihood of developing multiple sclerosis (MS). With additional refinement, it could become a useful tool in identifying people for early intervention or prevention efforts.
MS is a complex neurological disease with an unknown origin characterized by demyelination of the central nervous system. Disease onset usually occurs in young adults, and it is more common in females. It has a prevalence that ranges between 2 and 150 per 100,000 people. Genetic epidemiology studies indicate that first degree relatives of people with MS are 15-35 times more likely to develop MS themselves. Recent genome-wide association studies have identified a number of genetic loci; however, despite the number of genes confirmed as being involved in MS there is still no clear understanding of the genetic contribution to disease susceptibility. Environmental factors with convincing evidence of involvement with MS include sunshine, vitamin-D, Epstein-Barr viral exposure, and smoking.
The current study attempts to answer the question, can we predict who will develop MS? The team employed a factor called the C statistic, which defines how well a model can differentiate between patients and controls. A model with C statistic equal to 0.5 predicts no better than simply tossing a coin while a perfect model has a C statistic of one. For clinical prediction, a C statistic of 0.8 or higher is considered useful.
In the current study which examines 16 genetic loci associated with MS, the researchers used three different cohorts. The C statistics obtained ranged from 0.64 to 0.72 depending on whether gender, smoking history, and Epstein-Barr virus titers were incorporated. Although below the standard of 0.8, by incorporating other data and environmental factors, this study could lead to the development of a model with strong predictive power, which could identify individuals that would benefit from early intervention efforts.
One member of the research team, K. Claire Simon of the Harvard School of Public Health is supported by an NIEHS training grant.
Citation: De Jager PL, Chibnik LB, Cui J, Reischl J, Lehr S, Simon KC, Aubin C, Bauer D, Heubach JF, Sandbrink R, Tyblova M, Lelkova P; Steering committee of the BENEFIT study; Steering committee of the BEYOND study; Steering committee of the LTF study; Steering committee of the CCR1 study, Havrdova E, Pohl C, Horakova D, Ascherio A, Hafler DA, Karlson EW. Integration of genetic risk factors into a clinical algorithm for multiple sclerosis susceptibility: a weighted genetic risk score. Lancet Neurol. 2009 Dec;8(12):1111-9.
▲ Up: Gene Packaging is Important in Cancer (http://www.niehs.nih.gov/research/supported/sep/2008/gene-packaging/index.cfm)
▼ Down: Natural Food Products Can Treat Cancer (http://www.niehs.nih.gov/research/supported/sep/2009/natural/index.cfm)
