Human Limits

Exploring performance and health with Michael J. Joyner, M.D.

Photo of Michael J. Joyner, M.D.

Statins, 23andMe, and Personalized Medicine

What opinion leaders in biomedical research were thinking 20 or 30 years ago seems pretty remote from the news last month that the American College of Cardiology and American Heart Association released new guidelines on statin therapy to reduce the risk of cardiovascular disease.   It also seems pretty remote from the more recent news that the FDA told the genetic screening company 23andME to suspend their direct to consumer health related genetics testing and advice.   At first glance these two stories are unrelated, but bear with me while I make the case that both stories are fundamentally about the limitations of something called personalized medicine and the routine use of genetic testing to customize therapy for individual patients.


Statin Guidelines

The net effect of the guidelines is to recommend that far more apparently healthy people be treated with statins to prevent cardiovascular disease.   The new guidelines have been controversial from the start and criticized for reasons including: 1) what data and studies were used to develop the guidelines, 2) the accuracy of the calculator used to estimate who is at what risk and in need of treatment, and 3) the deemphasizing of cholesterol monitoring to adjust therapy.    There was also yapping about links between some of the guideline writers with the pharmaceutical industry and just the confusion associated with changing recommendations on how commonly used drugs should be used.   What most people fail to realize when they read about the next best medical thing is that medical evidence and recommendations shift and change over time.  In a recent review of how medical evidence changes more than a third of what is considered state of the art underwent significant challenge or revision over 10 years.  So the fact that the new statin guidelines are controversial and why they are controversial is not surprising.


Direct to Consumer Genetics Testing

23andMe was essentially banned from offering to direct consumer information about disease risk based on genetic testing.  The main concerns outlined by the FDA relate to the basic validity and reliability of the testing, how assessments of risk were being communicated to the consumer, and the unpredictable nature of how a consumer might respond to a “positive” test.  For common non-communicable diseases (the stuff the kills most of us like diabetes, cardiovascular disease, obesity, and many forms of cancer) there are hundreds of gene variants that put people at increased risk.  However the increased risk for each variant is typically tiny and pales in comparison to behavioral risks much of the population engages in.  In reality no one really has a good handle on the real risk associated with most gene variants and there is much less insight into how they should influence medical decision making.  The other dirty little secret here is that it has been difficult to replicate risk variants across studies, and for a number of diseases the gene variant profiles are the same in people with and without the clinical disease.  There are also the predictable issues of what findings in one ethnic group might mean in a different group.


Not so Fast: The Statin 23andME Link

Sometime in the 1980s or early 90s the idea took hold that if we simply knew enough about the individual genetic profile of each person that we would be able to customize drug therapy in specific and medical therapy in general to better treat, predict, and prevent a wide variety of disease.  Most prominent academic medical centers now have huge programs based on this basic proposition.  The FDA’s message to 23andME can be viewed as a “not so fast” message.  Whoever is doing the testing, there is a lot left to learn before personalized medicine might really make a difference for most people seeking advice from most Drs. for most of what ails us.

In this context, the statin guidelines say the same thing the FDA did by what they don’t recommend. Instead of recommending that statin therapy be tailored based on each patients genetic profile and cholesterol level, they advocate use of a generic risk calculator derived from decades old epidemiology studies and then largely abandon the idea that it is important to follow how the drugs affect cholesterol levels.   Maybe the next step is simply putting statins in the water.

As far as personalized medicine in general goes, research might ultimately get us to a genetically based diagnostic and therapeutic promised land.  However, don’t be surprised if personalized medicine based on your individual genetic profiles ends up becoming the biomedical version of atomic fusion, an attractive idea that just can’t quite get there for most diseases.



Leave a Reply