Genes don't add much value for predictions of type 2 diabetes risk

At Gene Expression, p-ter points to two studies in this week's New England Journal of Medicine examining the predictive value of known genetic markers for type 2 diabetes.

Both studies find the additional predictive power of the genetic markers beyond traditional predictors (like age, sex, family history, body-mass index, fasting glucose levels, systolic blood pressure, high-density lipoprotein cholesterol levels, and serum triglyceride levels) to be extremely small, to the point of complete clinical insignificance. Thus while massive genome-wide association studies for T2D have been informative in terms of uncovering the molecular pathways underlying the disease, they haven't yet yielded much of value in terms of individual risk prediction.

As p-ter notes, this result is quite unsurprising given the small fraction of the overall risk variance explained by these markers, but it's still important to quantitate the extent of the disappointment...

One small ray of hope: one of the articles notes that "[t]he discriminative power of genetic risk factors improved with an increasing duration of follow-up, whereas that of clinical risk factors decreased". In other words, genetic information increases in value relative to clinical predictors when there is a long time lag between testing and disease onset - again unsurprising given that clinical variables change over time while genetic risk factors are fixed, but still important to bear in mind. This suggests that genetic information will be most valuable to clinicians for predicting the risk of adult-onset diseases in children, or predicting the risk of late-onset diseases in adults.

Subscribe to Genetic Future.

More like this