Statins: Still Overhyped After All These Years

In a rather self-serving review article entitled “A historical perspective on the discovery of statins,” Japanese biochemist Akira Endo hits all the conventional and PC notes in his 10-page (including references) trip down memory lane. From the get-go, in the abstract itself he tells us that…

“Cholesterol is essential for the functioning of all human organs, but it is nevertheless the cause of coronary heart disease. Building on that knowledge, scientists and the pharmaceutical industry have successfully developed a remarkably effective class of drugs–the statins–that lower cholesterol levels in blood and reduce the frequency of heart attacks.”

We would expect nothing less from Endo, considered the father of statin drugs, had this article been published in the 1990s. But, appearing as it did in May, 2010, in the Proceedings of the Japan Academy, his astonishingly uninformed portrayal of coronary disease etiology; and the modest benefits of statin therapy–let alone their devastating side effects–give a whole new meaning to the term “Ivory Tower.”

Endo was inspired by the work of Nobel prize winning bacteriologist Alexander Fleming, who discovered the antibiotic substance benzylpenicillin (Penicillin G) from the fungus Penicillium notatum. As such, on his mission to save the world from cholesterol, Endo began searching in 1968 for fungi that might produce an inhibitor for Hydroxymethylglutaryl-coenzyme A reductase, a key enzyme in the biosynthesis of cholesterol. Inasmuch as he acknowledges that cholesterol is “essential for the functioning of all human organs,” one wonders why he thought grossly interfering with its synthesis would be a good idea.

Indeed, Endo’s describing his search for a HMG-CoA Reductase inhibitor in the same context with one of the first–and still widely used–antibiotics is perilously close to medical blasphemy. Nonetheless, Endo’s work eventually led to lovastatin, the first FDA-approved statin drug (August, 1987). This drug would be followed by several others, making statins the best-selling class of pharmaceuticals in history.

Endo should have been aware that the majority of studies touting the efficacy of statins use the statistical chicanery of highlighting the relative, rather than the absolute risk reduction–even though all the relevant raw numbers are detailed in the publications. Consider the oft-heard claim that “Statins reduce the risk of a heart attack by 30%.”

It may have been that the risk of having a heart attack in a particular cohort was reduced from 0.5% to 0.35%. In relative terms, this is a 30% reduction, but in absolute terms is far less impressive. In fact, based on meta-analysis, statins show their best numbers among those with pre-existing heart disease, with an average absolute risk reduction of 3%. Among those without pre-existing heart disease, absolute risk reduction is 1.6%. In both cases, the risk reductions were well below the reported incidence of side effects.

As to those side effects–which include myopathy, memory loss, and hyperglycemia–more often than not, patients’ reports of these are played down by their physicians. And, in the case of memory loss, junk science is used in a truly pathetic, nay idiotic attempt to obscure them…

In a recent JAMA article entitled “Statin therapy and risk of acute memory impairment,” the authors admit that all cholesterol-lowering drugs induce memory loss, but that must be an artifact, since they found similar results with dissimilar (statin and non-statin) drugs. Lead author Brian Strom calls this artifact “detection bias.” That is, patients taking a new drug are highly attuned to their health, and to reported side effects.

Or, to put it another way, results contrary to the mainstream gospel are always the result of detection bias, because people are just so stupid. Never mind that 25% of the cholesterol in your body is found in your brain, where it plays important roles in such things as membrane function, or that in the elderly, the best memory function is observed in those with the highest levels of cholesterol.

But wait, there’s more. A couple of weeks ago, two studies presented at the American Society for Clinical Oncology conference in Chicago, are being touted as showing that statins may help prevent cancer. However, there are at least three flaws, as pointed out by Malcolm Kendrick MD, and others:

1.     The studies are observational, not randomized and controlled. As such, associations between two variables can be elucidated, but causality cannot be proved.

2.     Risk reduction is reported in relative, rather than absolute terms, which, as discussed above, is highly misleading.

3.     And in the most massive flaw, it is well known that high cholesterol is protective against cancer. Thus, those patients taking statins would likely be the ones with higher cholesterol, and therefore be more likely to show lower risk.

Yes friends, this flapdoodle is now what passes for mainstream science. As Shakespeare’s Queen Gertrude famously said: “The lady doth protest too much, methinks.”

Advertisements

9 responses to “Statins: Still Overhyped After All These Years

  1. Good article, thanks.

  2. Michael, great job as usual. I would add a couple of other points.
    1) At the time I published my book, there had been five major statin studies that were placebo controlled. At least three of these had active participation in the planning and execution of the trials by employees of the pharmaceutical company whose drug was being tested.

    2) None showed improvement in outcomes that exceeded 2-3% in absolute risk which is of no practical significance at the low levels of incidence reported no matter how much jiggering of the “p” values is done. This is about the same level of “benefit” as that conferred by low dose aspirin or other platelet inhibitors.

    3) In all of these studies there was a total disconnection between the improved outcomes and both the initial level of blood cholesterol and the degree of cholesterol lowering attained. In other words, “protection” against coronary heart disease was the same whether the initial cholesterol level was high or low and the degree of protection did not correlate with the degree of cholesterol lowering. Subjects with normal cholesterol levels at the outset showed the same amount of benefit as those with high cholesterol. Those subjects whose cholesterol went down by a relatively small amount benefited to the same degree as those whose cholesterol levels declined a lot.

    In experimental science this disconnection is called a “lack of normal exposure-response” and means that the factor under investigation is not the true cause of the disorder. In other words, the small “benefit” conferred by the statin drugs was not due to their effect on cholesterol levels but rather to some other cause.

  3. @Ernest–
    Thanks for your additional remarks. I find it most interesting that Akira Endo seems positively clueless as to the real world results of statins. Is it even possible that the father of the best-selling class of drugs in history is totally ignorant of the actual outcomes?

    Astonishing!

    • ernestncurtis

      The human mind is capable of extraordinary feats of rationalization when one’s financial well being or reputation are at stake. Endo shows his ignorance in his opening statement: “Cholesterol is essential for the functioning of all human organs, but it is nevertheless the cause of coronary heart disease.” The first phrase is certainly true but the second is patently false. There is zero scientifically credible evidence that cholesterol is even a factor in the causation of CHD much less “the” cause.

      • @Ernest–
        Indeed! While I was writing the piece, even though the references to Endo were to set a historical perspective, and not be the main theme, I just could not wrap my mind around the fact that the guy is in total denial.

  4. This is a request for information on the statin trials. My understanding is that the end point in (some? many?) the original efficacy trials was not coronary events in themselves, but the supposed markers for coronary events or just a reduction in cholesterol itself. I am sure someone here knows how many trials actually used coronary events as the end point (prior to the drugs being approved) and that the impacts on coronary events are only from post-approval studies.

    My point is that, given the acknowledgement that cholesterol levels themselves are not a risk factor, how many of the efficacy trials are now invalid?

    • @Rob–
      I’m sure that Dr. Curtis will reply to your question. Not to spoil anything, though, the trials were all basically crapola.

    • Rob, many of the statin trials did use surrogate end points. One that was popular for awhile was “regression of coronary atherosclerosis” as measured by angiographic and/or ultrasonic criteria. Others include episodes of chest pain, hospitalizations for chest pain, etc. All of these are unreliable for a variety of reasons. But even the coronary events that are most commonly used are fairly inaccurate. The three most commonly used are number of heart attacks, death from coronary heart disease, and total (all cause) mortality.The latter is the only truly “hard” or reliable statistic. The criteria used to determine death due to coronary disease or incidence of heart attack bring a high degree of error into these statistics.
      One problem with accepting the bad science that shows epidemiologic correlations with coronary disease and certain markers is that the markers then become the end points and thus moves the analysis another step from reality. As a famous economist once remarked in regard to such a process it is “nonsense on stilts”.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s