Saturday, November 16, 2013

The Cardiologist Giveth, then the Cardiologist Taketh Away: Revision of the Cholesterol Guidelines

There has been quite a stir this week with the publication of the newest revision of the ACC/AHA guidelines for the treatment of cholesterol.  The New York Times is awash with articles summarizing or opining on the changes and many of the authors are perspicacious observers:
As the old Spanish proverb states, "rio revuelto, ganancia de pescadores" - when the river is stirred up, the fishermen benefit.  I will admit that I'm gloating a bit since I consider the new guidelines to be a tacit affirmative nod to several posts on the topic of the cholesterol hypothesis (CH).  (More posts here and here and here, among several others - search for "cholesterol" or "causal pathways" on the Medical Evidence Blog search bar.)

One main message of the new guidelines is that lowering cholesterol by any other means than the use of statins is now considered by the authors suspect and not recommended.  By inference, the CH is under a brutal siege (for an exposition of how drug trials help us sort out causal pathways see this post).  In aggregating and analyzing the evidence, the guideline authors conceeded that lowering cholesterol by any means does not prevent atherosclerotic cardiovascular disease (ASCVD).  Only lowering it with statins does.  (Which is why I quit taking niacin after the results of the AIM-HIGH study were revealed.)

A corollary of this message and also a main proposition of the new guidelines is that targets for cholesterol lowering are unsupported by the evidence.  One way of viewing this is that measuring a surrogate marker of the effect of statins, namely their effect on cholesterol levels, is not an adequate or reliable marker of their effect on cardiovascular outcomes of interest to patients - because they likely have other effects that we have yet to elucidate.

Another message in the guidelines is that the general utility of statins is high because of the high prevalence of cardiovascular disease, but the specific utility of the drugs is related to the specific risk in an individual patient (see yesterdays blog post on Status Iatrogenicus for an admittedly crude formalization of this concept).  I commend the authors for recognizing in Table 7 on page 34 of the guidelines that the absolute rather than the relative risk reduction is what counts for individual patients.  Too often do we succumb to the temptation to compartmentalize risk, because relative risk reductions are more impressive than absolute ones.  In some patients, such as those receiving hemodialysis or with advanced heart failure, statins may not be of benefit in spite of high risk - either these patients will die before statins' benefits can accrue, or there is time dependency of causal pathways - the proverbial horse is out of the barn.  Finally, as one of the above authors has argued in BMJ, the absolute risk of disease that justifies chronic treatment with statins is a judgment call.  That may be a future blog post.

I applaud the authors of these guidelines with a standing ovation.  I think they have done a stand up job at putting aside their cherished beliefs, thumbing their noses at the status quo, and focusing on the actual evidence.  This is real progress.

The cardiologist giveth, then the cardiologist taketh away.  But alas, the cardiologist hath seen the light.


  1. Here is an absolutely wonderful editorial in JAMA about the new guidelines: Psaty Editorial

  2. I had a high cholesterol test (~190) in 2004 and was prescribed lovastatin, which progressively made my joints ache beyond my years. I complained to my doctor repeatedly, then refused to take any more. We tried a few other drugs, but all had side effects, until success with generic Lipitor about 5 years ago (10mg keeps my cholesterol below 150). I exercise a lot and love salads now, and still enjoy mexican food or pizza once in a while, thanks to Generic Lipitor.

  3. I would check the new guidelines. You may find that the course you have taken is totally consistent with them. In any case, your result as you describe it would be hard to improve upon.