A clear example of how values matter in judging statins for primary prevention


Category : Medical Rants

The new JAMA has an article and an editorial that favor using statins for primary prevention. The clinical evidence and the editorial focus on the relative risk improvement with statins. However, the clinical evidence reports on a number needed to treat for 5 years of 138 to prevent 1 death. So 5 years of statin therapy for 138 patients to prevent 1 death. Is this a reasonable trade off?

The lipidologists obviously see this as a victory for their hypothesis. They quote data that patients tolerate statin therapy as well as they tolerate placebo. They now have the evidence. We should give many more patients statins!

Many critics will view these data differently. 690 years of statins to prevent 1 death seems like a weak result (I multiplied 138 by 5 to get the number of statin years). How do we value a year of statins? Since the studies mostly used pravastatin, the cost to the patient should be around $50 year. What about side effects? Primary care physicians have a different view of statin side effects than do lipid specialists. Patients have a different viewpoint.

Is there a correct answer? Is our own goal decreasing deaths? Some would argue it is the primary goal. Others would argue that every additional drug has its own hazards. They would dispute the side effect data in the randomized controlled trials.

And this is why we have controversies in medicine. Our decisions do depend on how we define problems. The definition of our prevention goals remain fuzzy to many physicians.

Comments (4)

I think using number needed to treat to guide clinical decisions is the best way to go. Risk reduction (either relative or absolute) can be make it seem that interventions are doing more that they are. I would be interested in knowing the numbers to treat for “lifestyle” interventions- e.g. if you bring your BMI from X to Y, or exercise 30min a day, if you stop smoking you do this, etc. I agree with your assessment that 690 statin years seems weak for primary prevention. On a population scale, it may seem good, but that’s a lot of people on medications who will NOT have benefit.

One issue in any field is separating the mathematical from the practical. Yes there is a possible mathematical improvement, but at these values is it due to chance, or is it a real value. Additionally, how do we translate this information into everyday use?

An engineer will tell you the strength of a bolt or piece of steel, but do I need the very strongest, most expensive item to do my job?

We also have to look at association bias: if I spend my whole life working on an area of research, no matter what that area, this then becomes the most important area of research in my field.

As a businessman I am looking for meaningful numbers, not something that is three digits to the right of a decimal point. I also understand the desire of those working in the field for self justification.

The most important thing is I understand the desire of the drug companies to sell product and have seen enough evidence of manipulation, and downright lies, to make me as a patient question any seemingly small mathematical relationship. Sadly, we also see doctors using some small item to justify decisions made in their practice. Statins and the associated blood test and office visits has become a practice driver.

A recent post on another blog highlighted how a doctor should never order a test or prescribe a medication that is not in the best interest of the patient. The reality is today many doctors do not follow this time tested rule. The result is this rather abstract discussion of the benefits of this drug.

Remember the AMA has a vested interest in this issue given its control of the RUC, and the specialist who make up the majority of that committee want to see more medical spending, not less. The AMA then generates a large income over the sale of the resulting computer codes, which translates into practice standards.

Steve Lucas

What is a good NNT? Lots of things we do have worse NNTs than statins and cost alot more (mammogram screening 781-2000 depending on the study, screening colonoscopy infinity as there are no published randomized trials of the effectiveness of screening colonoscopy in average risk patients ) So are you ready to quit doing these things? Lss deadly than MI and stroke and worse side effect profile.

I have a video on YouTube (http://youtu.be/2BlUhW6Zu2E) reviewing the guidelines and the issues around primary prevention. There are 2 meta-analyses of primary prevention done recently: the Cochrane meta-analysis (reviewed in the JAMA article) and the CTT metaanalysis in lancet. The CTT metaanalysis is reviewed in my video and makes a compelling argument for more statins as does a paper in health affairs (also reviewed in my video) on the benefits of more statin usage.

For some reason there is a bias out there at taking more statins is a bad thing but yet there is no data put forth to support that view. Make a compelling argument for not expanding statin usage in primary prevention. Leading causes of death in this country are MI and stroke….both absolutely documented to be reduced in primary prevention trials in patients at all risk levels (look at the CTT metaanalysis). They are cheap. The societal benefits outweigh costs 4:1 if you review the Health Affairs paper I reference. Yes some patients cant tolerate them but most do.

I will be blogging about these issues on my blog also: ebmteacher.com

This may be of interest to this and other discussions regarding patient care:


Steve Lucas

Post a comment