Saturday, November 27, 2010

The anacetrapib giggle

Stan has a link up to the full text but here is the fun table:

My summary is that a much larger study is desperately needed to confirm that the 300% increase found in cardiovascular mortality is a direct effect of anacetrapib. An even larger study is also urgently required to demonstrate statistical significance for the more modest increase in all cause mortality caused by this drug.

Volunteers should join the queue marked "idiot", unless they are certain that they will get in to the placebo group.


Still no blogging but this was too fun to skip.


Anonymous said...

"Giggle"; "fun". Ah yes -- people diseased and dying, and the lives of others in the balance. Party time!

Peter said...

Certainly if you have shares in anacetrapib. Laugh on.


David Moss said...

That's one way to reduce non-fatal myocardial infarction I suppose, make sure it's a fatal one...

Ned Kock said...

Hi Peter.

Did they include those who underwent revasc. in the other counts?

It seems that the placebo takers underwent dramatically more revasc. procedures than those taking the drug.

JamesSteeleII said...

Giggle, and giggle again.

Dr. B G said...

You and Stan crack me the h*ck up!!! *BIG GIGGGGGLE!!!!~*

who the f*ck does Big Pharma hire for statistians...?! Obviously they are paid off enough...

Anonymous said...

Yes the 40% increase in deaths was somehow not mentioned in MSM press releases.

When I read them, I instantly knew they must be counting interventions as "cardiac end-points" as there is absolutely no way THEORETICALLY that anyone could think this drug could work - unless you are either stupid or a cardiologist....

And the chart confirms that and that it is more likely to kill you as well - even if not as effective a killer as torcetrapib!

I am going to short their stock with long-dated options on monday morning.. seriously

Peter said...

Hmmm, there is a long queue to play Russian Roulette it seems...

Ned, I've no idea why there were more revascs in the placebo group. One day we may find out, if the body count is not so high as to have anacetrapib dropped after the 30,000 have been played with (unlikely). Should anacetrapib survive it will turn out to have a non lipid effect, along the lines of statins increasing nitric oxide levels which decreases that chest pain which is the automatic ticket to a revascularisation procedure, soft end point...

Ultimately brec, if you can look at a group of people with sky high HDL and rock bottom LDL who derive absolutely ZERO total mortality or CV mortality benefit and not find it hysterical, you are ABSOLUTELY in the wrong place on the net. Or on a statin. Or both.


Peter said...

Hi Kurt,

If I could remotely afford it I would join you there. Stan comments that Merck's stock is down 5% already...


Ed said...

Mortality at 76 weeks went against anacetrapib both from cardiovascular causes (four versus one death, or 0.5% versus 0.1%) and from all causes (11 versus eight, 1.4% versus 1.0%).

With additional follow-up to 88 weeks, though, the placebo group caught up for a total of 12 deaths in both groups.

Let's see what a higher-powered study can do :-)

Moreover, he pointed to the revascularization rates, which favored the CEPT inhibitor (eight to 28, or 1.0% versus 3.5%).

This surprising magnitude of difference in revascularizations in a stable outpatient population was highly significant despite the small numbers, Cannon noted.

"That's a marker potentially of progression ... a little hint that it (anacetrapib) may be impacting on the arteries progression," he said at the press conference. "That was the moment for me that said this looks like it's real."

Why don't they do something silly like actually MEASURE IMT or something direct, to see if there is progression, rather than trying to use revascularization rates as an intermediate? Gah.

Brinton cautioned, though, that revascularizations should be considered a "soft" endpoint as physicians outside the study would not have been blinded to cholesterol changes and may have put off revascularization if they saw evidence of a drug response.

So the placebo group may have undergone more revascularizations due to uncooperative lipids. Do revascularizations result in lower (or higher) mortality?

Anacetrapib's effect on LDL on its own likely wouldn't get a nod from the FDA, which -- in part because of ezetimibe -- has made clear it is looking for hard clinical endpoints rather than surrogate ones, noted Nissen.

Thank god!

"We're all guessing it's [REVEAL] going to stop early," he said in an interview.

Aren't there FDA regulations which invalidate a study that is stopped early? It seems obvious that you can game your results if you can dynamically choose your ending date. There were a lot of complaints recently about drug companies doing this, and these guys are boldly planning it already? Disgusting.

. said...

Hi Peter. According to my calculations, its a 300% increase in cardio mortality, and 37.5% increase in total mortality. Maybe these are not statistically significant but, anyway, I will not volunteer to the next Anacetrapib study.

Anonymous said...

you beat me to it..

anacetrapib apparently causes silent ischemia - more infarcts but fewer trips to the cath lab!

Stan Bleszynski said...

Hi everybody,

The data are not statistically significant! That was the main "success" of the study. 8-:)

The difference between a case of 4 deatch and a case od 1 death is 3. Statistical error (one sigma) is sqrt(N) that makes it 4+/-2 and 1+/-1 therefore both results are within error bar from each other.

It is the same about 11 versus 8, that is 11+/-3.3 and 8+/-2.8 which is only one sigma apart.

Re: shorting Merck

I would recomend not to or be careful, unless you are fully aware what you are up against!

You may find yourself fighting against a big crowd of wealthy idiots bidding this POS up and up and up. In the current situation we also have to deal with the corrupt and criminal elements in the government especially Mr. O who spoiled my bank and insurance shorts in 2008 by first banning shorting and then by bailing those parasites out. (I predicted the financial crisis in September 2007 using certain unusal methods, and begun shorting US banks back then but they forced me to cover too soon!). Governments are capable nowadays of spoiling any good shorting play by selectively bailing out some corporations or by propping up the entire market through their money printing called "Quantitative Easing". It is no longer a capitalism. I call what is happening now in the US "Back to the USSR". :(


woly said...

I tend to agree more with Stan, the results are not statistically significant and therefore the most you can say is that it didn't do any better than the placebo. This is also pretty damning but not as severe.

Unknown said...

I'm disappointed in the study that it did not make more mention of the causes of death for those without cardiac mortality. As you know, Peter, non-statin cholesterol lowering drugs especially seem to keep killing people more than placebo via suicide, accidents, or other violent deaths Statins do it too but less robustly - and violent death and suicide are relatively rare so the data is often not significant. (I covered the murky link in my blog here:

Also interesting - Merck funded study also seems to blame torcetrapib increasing the adrenal glands' production of aldosterone and cortisol for it being so effective at killing people. Especially the heart-diseased people it was studied on. Anacetrapib did not seem to have the increased BP and sodium regulation issues that torcetrapib does but still... What the heck are these drugs *really* doing.

Stan Bleszynski said...

It is quite damning and severly so for 2 reasons:

1) because if more than doubling of HDL by drugs to over 100mg/dl was in any way correlated with cardio protection then there should have been NO heart event whatsoever! There is no stat significant difference between 11:8 but it would have been if there were 0:8 cases! We should keep in mind that the patients had already high cardio risk status.

2) I suspect that what is even more damning was not what was published but what wasn't! Did you notice the lack fasting blood glucose levels and no HbA1c figures?!

IOW the giggle factor is in the fact that artificially increasing doubling or tripling the HDL and halving LDL does not significantly change the risk. Medical mainstream and most of their vic - pardon 'patients' do not seem to be even aware that they have the problem because they still believe religiously in the cholesterol-cause-heart disease paradigm. That is funny indeed. Whether we ought to laught of this all? I don't know. I personally would submit every single one of the trial patients and all Lipitor users (collectively) up for Darwin award.


. said...

DEFINE: Video Interview with Christopher P. Cannon -

Unknown said...

O Primitivo - I'm hurt as I am a teacher at HMS and no doubt my students do rounds with those teachers who promote anacetrapib... they are usually thoughtful. I can't explain it.

Dr. Curmudgeon Gee said...

wow. thanks.

interesting observation by Stan the Heretic on the omissions of other important "markers"


Anonymous said...


Yes, we are all aware that the results (lucky for them) were did not reach statistical significance. Of course, I am sure you realize that inadequate power to prove that a 300% increase in cardiac deaths does NOT MEAN THE RESULTS ARE NOT REAL.

We should not be surprised as this is only the phase I trial. The same thing happened with trocetrapib - no excess mortality with phase one, then the big trial causes 50% excess mortality.

The possibility of type II error here cannot be excluded, especially as the result is in the direction one would predict based on the results with torcetrapib. My money is on a real difference showing up in the big trial just like with the first drug.

As far as shorting the common, I was actually being rhetorical, but I am quite comfortable with short selling as a strategy. I too predicted and shorted the crash in the S and P from 1520 all the way to 900 via futures. That was pretty easy to predict and was the trade of the decade.

. said...

Dear Dr. Emily Deans, the Lipid Hypothesis is more related to food politics, pharmaceutical

lobbies and scientific careers than it is with the scientific method. Most people fail to aknowledge this simple fact: it doesn't matter how inteligeng a person is, if he studies in the HMS or elsewhere, he is allways human. As a psychiatrist you should be able to explain this better than me. Anyway, here are some hints for the reasons of why these cholesterol myths will live forever:

"Once a belief has become deeply ingrained in our psyche, it takes on a whole new life of its

own--it becomes part of the value structure and knowledge base from which we draw upon to

navigate through an often chaotic and unpredictable world. It becomes, in effect, an integral

part of our psychological make-up, a part of our very identity." -

"Scientists have lost their grants, and even positions, for criticizing ‘accepted truths’ and

particularly so if this ‘rocks the boat’ for self appointed gurus ! or/and threatens profits

e.g. for pharmaceutical industries. Again human emotions: ‘Holy-cows-should-remainsacred’ (and

particularly so if they yield plenty of milk)" -


"If you want to be among the world's great mass of unthinkers whose ignorance and gullibility

allows the incompetent, the unscrupulous, and the corrupt to run amok, then be sure to live by

the following five tips! 1. Fail to verify information for yourself; 2. Believe that because

something is widely held as fact, it must be true; 3. Practice blind obedience to authority; 4.

Place great emphasis on formal credentials; 5. Be swayed by the ad hominem attack" -

"The idea that a high intake of saturated fat and a high cholesterol concentration in the
blood lead to atherosclerosis and cardiovascular disease emanates from a variety of
sources. When considered together, it is understandable that a whole world of doctors
and medical scientists have embraced the diet-heart idea and the cholesterol hypothesis,
in particular because two of the main supporters, Joseph Brown and Michael
Goldstein, have been honored with the Nobel Prize." -

"An almost endless number of observations and experiments have effectively falsified the

hypothesis that dietary cholesterol and fats, and a high cholesterol level play a role in the

causation of atherosclerosis and cardiovascular disease. The hypothesis is maintained because

allegedly supportive, but insignificant findings, are inflated, and because most contradictory

results are misinterpreted, misquoted or ignored." -


Anand Srivastava said...

Like always great find.

I guess the placebo group underwent more procedures because they had a bad lipid profile.

People on anacetrapib would have a perfect lipid profile and so would not have gotten the procedures.

Of course not getting the procedures may also be behind the increasing death rate.

I hope they learn soon that lipid profile does not matter.

blogblog said...

I just downloaded and stasteed to read read the following MSc thesis:

"The effects of red meat consumption and high-intensity resistance training of skeletal muscle strength, muscle mass and functional status in healthy older adults."

What was the definition of high red meat diet?

800g per week!

I would very rarely eat less than 800g/day of meat.

blogblog said...

Emily have you ever considered recommending ketogenic diets for your patients? Stanford have been planning a ketogenic diet trial for bipolar for several years. But they haven't been unable to recruit enough patients.

The Japanese were also treating psychiatric illnesses with prolonged fasting (4-6 weeks with only IV glucose supplementation) and meditation back in the 1970s.

Unknown said...

Hi blog blog -

The most I've been able to convince some people to do is to ditch the grains and vegetable oil. Unfortunately, there are only two case studies in the literature of ketogenic diets and bipolar disorder. One made no difference, and the other (Atkins-based, though, so lots of omega 6 I would imagine) made the guy much, much worse. Without any data I just have the theory, and most people aren't willing to go ketogenic on a theory. Of course there is the bonus fat loss and normalization of diabetes... I can't even convince a diabetic patient to eat butter after being terrorized by his dietitians and primary care doctors into the margarine fold. I sure wish Stanford could convince some people.

. said...

Great cholesterol article -

Stan Bleszynski said...

There is some interesting info about CETP,HDL and health risk:


Now enter cholesteryl ester transfer protein activity (CETP), long been known to correlate with lipid levels. The lower the CETP the higher the HDL and the higher the HDL, the lower the CV risk, presumably. However, an initial clinical trial last year with Torcetrapib, a CETP inhibitor, resulted in a catastrophe of CV disease and had to be stopped prematurely. Something clearly was wrong with the underlying concepts.

Now this concept has been evaluated using the old, reliable Framingham study. The plasma CETP values of 320 study participants were followed as these people went on about their life and, in some cases, death from coronary heart disease, cerebrovascular disease, peripheral vascular disease, or heart failure.

The results were that a robust inverse relationship existed between CETP and CV disease risk. The lower your CETP ( and the higher the HDL ) the greater the incidence of these problems.

karl said...

Re Davis -

I've challenged him on the small LDL bit several times to this point: We know that high blood sugars cause small LDL AND oxLDL ( The sugar in particular is needed for the creation of oxLDL) There is only a correlation - no cause and effect. I've not seen where these confounding variables have been sorted out.

oxLDL is most predictive thing I've seen for CHD - yet is not tested for - is it due to the investments in these advanced LDL tests?

Higher HDL seems to reduce oxLDL - and one would think that keeping this garbage from being reabsorbed would work, but the numbers for such drugs don't impress.