Monday, November 29, 2010

I Warned Them This Drug Might Kill People...

by Chris Masterjohn

I was happy to see last night that the two letters to the editor I've written about vitamin E have been cited recently. 

I'm sad that another recent paper did not cite my letter about why their drug might be killing people, which we'll get to below.

The first letter was published in the Journal of the American College of Cardiology.  It was a response to the famous "one meal of saturated fat will kill you" study where the authors found that making a carrot cake and a milk shake with coconut oil rather than safflower oil led to lower anti-inflammatory properties of HDL particles isolated from the subjects' blood.  They attributed the effect to the saturated fat in the coconut oil, and I pointed out that it was probably a result of the vitamin E in the safflower oil.  I developed this argument in much greater detail in my article on this site, "Myth: One Meal of Saturated Fat Can Be Bad."

Little did I know until last night that the widely buzzed-about Krauss paper debunking the association between saturated fat and heart disease in the American Journal of Clinical Nutrition. 

They cited my letter very accurately, apart from the little typo at the end:
In the acute postprandial phase following a meal enriched in saturated or polyunsaturated fat, HDL collected from individuals after a coconut meal compared with a safflower or unsaturated fat meal was associated with a 50-70% increase in intercellular adhesion molecule and vascular cell adhesion molecule (131). Attribution of this effect specifically to the saturated fat of the coconut meal may, however, be confounded by the high [low] concentrations of tocopherol found in coconut oil (132).
Thanks guys!  If they liked my vitamin E letter, I hope I can convince them in the future to change their minds that "substitution of dietary polyunsaturated fat for saturated fat has been shown to lower CVD risk."  But I'm certainly happy that they concluded that "there are few epidemiologic or clinical trial data to support a benefit of replacing saturated fat with carbohydrate" and that "dietary efforts to improve the increasing burden of CVD risk associated with atherogenic dyslipidemia should primarily emphasize the limitation of refined carbohydrate intakes and a reduction in excess adiposity."

On the other hand, a paper about aromatherapy massage oils prepared from virgin coconut oil published in the Journal of the American Oil Chemists' Society used my letter in the most preposterous way imaginable:
Masterjohn [25] demonstrated that coconut oil had anti-inflammatory properties, which could help to reduce the inflammation occurring in muscles.
I did?  I wasn't even a graduate student yet when I wrote that letter.  I'm baffled that I "demonstrated" anything in my bedroom where I typed that letter, except the logical flaws in the paper I was responding to.  The study I commented on was about coconut oil diminishing the anti-inflammatory properties of isolated HDL.  I don't think that study design provided a realistic evaluation of the physiological effects of the coconut oil, but it certainly didn't demonstrate an anti-inflammatory effect of coconut oil!  And what it had to do with muscular function is beyond me.

I imagine they did a pubmed search for "coconut oil anti-inflammatory."  My letter is fourth out of twelve results.  Its title is "The anti-inflammatory properties of safflower oil and coconut oil may be mediated by their respective concentrations of vitamin E."  The authors must have taken "no abstract available" to mean you only have to read the title.  But geez, the whole letter is about one paragraph long.

The second letter was published in The American Heart Journal and argued that safety trials of anacetrapib, a cholesterol ester transfer protein (CETP) blocker, should test the effect of the drug on vitamin E transport.  This was in response to a short, 8-week study of anacetrapib showing no adverse effects.

CETP blockers are the first drugs ever used to test the "reverse cholesterol transport" hypothesis in humans.  This hypothesis says that HDL is beneficial because it takes cholesterol back to the liver and that a high total-to-HDL cholesterol ratio indicates that this process is occuring less efficiently and plays a causal role in atherosclerosis.  The first drug used to test this hypothesis, torcetrapib, killed people left and right.  But they want to believe this is because it increased blood pressure and had other "off-target toxicity," instead of believing that their hypothesis was falsified.  So they invented anacetrapib, another drug that does the same thing without the effects on blood pressure.

My letter cited test tube and cell studies showing that CETP plays an important role in transferring vitamin E to HDL particles, and that HDL particles have a very specific role in delivering vitamin E to the endothelial cells that line the blood vessel wall, where the vitamin E helps suppress the oxidation of LDL and thereby suppresses the development of atherosclerosis.  I concluded as follows:
An 8-week preliminary trial of the CETP inhibitor torcetrapib found no increase in adverse effects [9], but this drug ultimately increased the risk of cardiovascular events and death from all causes [10]. Because part of the protective effect of HDL may be due to its role in vitamin E transport, safety and efficacy trials of CETP inhibitors should take into account their effect on this process.

In other words, the fact that anacetrapib showed no adverse effects after eight weeks should give us little comfort that longer trials will not show it to kill people left and right, because the early safety trials with torcetrapib showed the same thing!

The authors responded that the role of CETP in transporting vitamin E to HDL had not been demonstrated conclusively in live people, and that since large trials using alpha-tocopherol did not prevent heart disease, it is not clear that alterations in vitamin E metabolism would have any effect either.  They came to the following conclusion:
A study to assess long-term efficacy and safety of anacetrapib in patients with coronary heart disease is currently in progress (clinical NCT00685776).  The safety information provided by this trial will determine whether or not to initiate a phase III outcome trial with anacetrapib.  Ultimately, it is only by conducting such an outcome trial that the true efficacy and safety of anacetrapib will emerge.

Indeed.  However, it still seems worthwhile to do a very simple measurement of the vitamin E contents of these HDL particles just in case, perhaps, a long-term trial shows that it kills more people.

I was quite happy to see last night that a recent review on vitamin E transport in the journal Molecular Nutrition and Food Research cited this letter.  I'm a bit sad, however, that the people who matter -- the people studying anacetrapib -- continue to ignore it.
As Stan (Heretic) and Peter (Hyperlipid) recently pointed out, the initial results from the anacetrapib trial just came in.  They are published in the New England Journal of Medicine here.
Searching the pdf for "tocopherol" or even "vitamin" yields no results.
The results are a little disturbing, but almost impossible to interpret. 
Among the just over 1600 people who were randomized to receive either anacetrapib or the placebo, forty percent more deaths occurred in the anacetrapib group:
The difference, however, was not statistically significant.  This means that we can't be confident it wasn't simply the result of chance.  This is unsurprising, given the small number of people that died.

There were four times as many deaths from cardiovascular causes in the anacetrapib group:

Again, the difference was not statistically significant, and again this is unsurprising given that only five people fell into this category.
We could say, well gee, let's conduct a larger study so this massive increase in cardiovascular deaths can be proven with statistical significance.  But the layers of uninterpretability just keep piling on. 

First, they took 142 out of the 801 people who received anacetrapib off treatment because their LDL-cholesterol fell to less than 25 mg/dL.  All of these people were included in the analysis, but it's unclear whether their cardiovascular risk would better reflect the fact that they initially received the drug or the fact that they were taken off of it.  This appraoch is called "intention to treat analysis" and it's very useful for making policy decisions (if I tell 100 people to take the drug, assuming some stop taking it, how many die?), but it doesn't help us understand whether the drug saves people or kills them.

Second, 3.5 times as many people in the placebo group underwent a revascularization procedure:

This might indicate that anacetrapib reduced the need for revascularization procedures.  On the other hand, the report gives no description of who ordered the revascularization procedure and what criteria they used to order it.  Furthermore, it's unclear whether the revascularization procedures helped or hurt the patients, since even the study protocol provides guidelines for assessing heart attacks and deaths that are caused by revascularization procedures.  They don't tell us whether anyone died as a result of a revascularization procedure.

Thus, this report makes it quite unclear whether anacetrapib is saving people or killing them, and makes it even less clear whether anacetrapib will save people or kill them when it's tested in a much larger trial.

One little experiment that could be done now would be to take the plasma samples from this trial out of the freezer, isolate the HDL, and measure the vitamin E in it.  That would very quickly obviate the authors' concerns expressed in response to my letter last year that "to the best of our knowledge, there are no clinical data indicating a definitive role for CETP in vitamin E exchange," since they could test that hypothesis and either show it true or bury it for good in the course of several days, and would provide some potentially relevant safety data, and would also contribute important knowledge to the body of basic science about vitamin E transport.

But as for saving or killing people, we will have to wait a few years for the final word.

Read more about Chris Masterjohn, PhD, here.


  1. Cetp drugs will be DOA because large trials will show their metabolic effects and hopefully, for this clinician, this uncovers the fraud of the hypothesis of heart disease to begin with. Great work Chris. And Peter at hyperlipid.

    Dr. K

  2. What blows my mind is a need for a high powered study. Given the absolutely gigantic effect anacetrapib has on lipids, shouldn't the results be evident in a low powered study? I mean, how small of a percentage change are they expecting?

  3. Hi Chris.

    On vitamin E, what do you make of the study by Sesso et al. suggesting that not only is supplementation worthless, but also that "vitamin E was associated with an increased risk of hemorrhagic stroke".

    They claimed in that study that "compared with placebo, vitamin E had no effect on the incidence of major cardiovascular events".

  4. Most (all?) Vitamin E supplements on the market are dl-tocopherol, which is a racemic mixture of a molecule with 3 chiral centers? That means that only 1/8 of it is identical to regular tocopherol, while the other 7/8ths are other enantiomers, sort of like how trans fats are isomers of natural unsaturated fats.

    Maybe this has something to do with the fact that Vitamin E has miserably failed in clinical trials. If so, that would be another reason to think that knocking off natural Vitamin E could be a bad thing.

  5. Ned Kock,
    According to that study, 400 IU synthetic α-tocopherol was used. This would be the dl form. You can readily buy the natural d form over the internet. Also, it appears that the mixed natural form of vitamin e that includes gamma-tocopherol is preferable.

  6. Responses to Anonymous, Ed, Ned, sr, and Richard.

    Anonymous, thanks! I hope this provides something fruitful besides innocent death as well.

    Ned, sr, and Richard...

    I agree with the comments of sr and Richard to a certain extent. However, I do not think the major issue is that it is a synthetic racemic mixture. It is possible that the unnatural isomers are harmful but I am not aware of evidence of this and in fact they have similar activity as antioxidants but lower affinity for the tocopherol transport protein, which is responsible for keeping them in circulation.

    However, as little as 100 mg RRR-alpha-tocopherol suppresses gamma-tocopherol levels, and this study used double this amount. Gamma-tocopherol, but not alpha-tocopherol, is capable of undergoing nitration, and peroxynitrite is probably the primary oxidant in plasma even in the absence of inflammatory conditions. Thus, it is fully expected that alpha-tocopherol should be harmful under conditions that mixed tocopherols would prove useful.

    Nevertheless, the effect of supplementation is completely irrelevant here. 400 IU vitamin E is so mind-boggling vastly outside of normal dietary intakes that a study using this amount cannot be used to generalize to the effects of physiological doses. The issue is whether directly interfering with the transport of normal physiological levels will cause problems.

    If, say, normal dietary intakes are 5-10 mg, then I humbly submit that causing a ~50% decrease of the transport of that 5-10 mg to endothelial cells in the brain, blood vessels, and elsewhere will have a much greater impact on physiological function than trying to supplement with 20 times the normal dose.

    Hope that makes some sense,

  7. I realized I didn't reply to Ed.

    Ed, I agree. You'd think that with such a massive effect, if the hypothesis is correct, you'd see at least a non-significant trend in the correct direction.


  8. This makes sense Chris. Indeed, many cause-effect relationships between health variables have a J-curve shape.

  9. okay, so you are just totally awesome. I am so so heartened to find scientific minds tackling this lipid debate. I have tried to follow Sally Fallen Morell, but the lack of scientific information to back things up has troubled me. NOW I've found it. For me, the bottom line is, "How have human beings done things for the past 10's of thousands of years?" Well, they sure didn't have the technology to extract oil from a soy bean! So if they ate animal fats and coconut and olive and sesame oil, that's good enough for me. But to see the research back this up, to understand the biochemical mechanisms behind the practice is everything I've been searching for! Our grandchildren are going to look back at our generation and marvel at our scientific stupidity the way we marvel at the stupidity of our grandparents who promoted smoking! I hope my grandchildren are healthier than all the rest as my legacy to them....Thank you so so you and Steve at Whole Health Source.

    former biologigist, turned SAHM,

  10. Chris,

    What about eating virgin red palm oil? It is very high in natural tocopherols? The vitamin E content is super rich. I ask because I eat about one spoon per day, and want to make sure that's an intelligent thing to do.

    Jack K

  11. Hi Jack,

    Yes, vitamin E is present in palm oil as mixed tocopherols as well as mixed tocotrienols. I think virgin palm oil is a good food.


  12. Hi Shelley, thank you so much for your kind words!

  13. Um, not to be a cynic, but I wouldn't hold my breath in terms of expecting the researchers to look into this. In my opinion, there is no way that the researchers would run a test which could conceivably kill the drugs chances of being marketed. They would never receive funding again from their benefactors. I believe, that the goal of the company which is funding this (and of every drug company), is to concoct some test, somehow, in some way, that will show some statistically significant improvement, which they can then shop to their friends at the FDA (who are waiting for a consulting/lobbying job once they retire from 'public service'). Look, maybe I'm just cynical, so God bless you sciencey types for assuming the best motivations in people. It balances out the cynics like me.

  14. Hi Dune,

    They might not investigate it, but someone else might. As I pointed out, a recent review on vitamin E transport cited my letter, so researchers involved in the basic science of vitamin E transport might test it just to define the role of CETP in that process. Also, if the drug does wind up killing people, it's important to have a published warning on record.


  15. Doesn't it bother you that no-one knows what vitamin E does that's so essential? It's the great unknown of nutrition - why is this even a vitamin?
    It does lots of things, sure, but most are interchangeable with other non-essential AOs. It's a midwife for carotene conversion to retinol, but probably not the only one. It modulates enzymes but has no receptor.
    No receptor yet.
    "Vitamin E is the last of all vitamins whose essentiality is not yet understood. Its widely accepted role as a lipophilic antioxidant has been questioned, since proof of its in vivo relevance remained scarce. The influence of vitamin E on biomarkers of oxidative stress in vivo is inconsistent and metabolites of vitamin E having reacted as an antioxidant are hardly detectable. Novel functions of vitamin E include the regulation of enzymes, most of which are membrane bound or activated by membrane recruitment. Also, expression of genes responds to vitamin E. The search for a transcription factor common to all regulated genes failed so far and a receptor that specifically binds vitamin E has not yet been identified. According to microarray data, pathways preferentially affected by the vitamin E status are the inflammatory response and cellular traffic. A role of vitamin E in cellular trafficking could best explain the neurological symptoms seen in vitamin E deficiency" - 2009

    1. Hi George,

      Thanks for your comment. I don't have the time to read this paper now, but there is good evidence, in my opinion, that vitamin E acts as an antioxidant in vivo, and I don't think it is interchangeable with other "non-essential antioxidants," most of which probably don't act as direct antioxidants in vivo. If it regulates gene expression but there is no receptor, why should this take mechanistic precedence over its role as an antioxidant? Virtually everything affects cell signaling. Oxidative stress is a major regulatory node of cell signaling.


  16. I love theseblogs, I hope you continue creating them! Thanks and good luck

  17. Folks entered and taken the new technologies. Check this out will help the many lives.

  18. they draw the line at forcing their colleagues, let alone corporate CEO's who benefit from big dollar special interest tax breaks, to do the same.

  19. There's a growing number of physicians who are very concerned about the detrimental effects of combining medications, both prescription and over-the-counter. And you should be concerned too. If you want to get more information's then just visit my blog

  20. your system is actually mostly consumed intended for producing smoke free fuels. Coal Carbonization towards various other hand is usually a method obtained for converting organic materials straight into different forms connected with carbon.