FINDINGS OF A
STUDY THAT EVALUATED PRIMARY PREVENTION CLINICAL STATIN TRIALS
A basic tentative modern Cardiology says that elevated
cholesterol increases the risk of heart attacks, significantly lowering
cholesterol should therefore reduce heart attack risk, statins reduce
cholesterol and in some context adverse cardiac outcomes, but meta-analyses of
primary prevention clinical statin trials for women with elevated cholesterols
have found no statistically significant cardioprotective effects for women, so
is it still appropriate or perhaps even illegal to market statins to women
given the nonsignificant effect seen in women. Welcome to the Business of
Medicine. I am Dr. Larry Kaskel, your host. My guest today is Theodore
Eisenberg, Henry Allen Mark Professor of Law and adjunct professor of
statistical sciences at Cornell University and he is here to discuss the
results of the findings of a study that evaluated primary prevention clinical
statin trials, which was recently published in Empirical Legal Study journal.
DR. LARRY KASKEL:
Counselor Eisenberg, welcome to the show.
THEODORE EISENBERG:
Hi, how are you?
DR. LARRY KASKEL:
I am good. Before we get to the results of the study, tell
me why you even became interested in this and conducted the study in the first
place?
THEODORE EISENBERG:
Sure, my wife went to a doctor as we advance in age,
apparently some times the female cholesterol goes up. She went to the doctor
who noticed the cholesterol was up and the doctor suggested that perhaps you
should go on Lipitor. What gave her pause was the doctor said, of course
before you do that, we will have to test your liver before and after and so
therefore she came home and said why they do want to put me on a drug where
they have to worry about my liver, does this work? and I had done some medical
research in connection with litigation and tried to find evidence that not only
did Lipitor reduce cholesterol, but it also better reduced, what you are
interested in which is bad heart outcomes and to my surprise, I couldn’t find
any evidence for any statin that there was a significant reduction in heart
outcomes for reasonably healthy women.
DR. LARRY KASKEL:
Can you repeat that last sentence?
THEODORE EISENBERG:
I couldn’t find any evidence for any statin that there was a
significant reduction in adverse heart outcomes for reasonably healthy women.
DR. LARRY KASKEL:
And you said any study that includes studies done by the
pharmaceutical giants.
THEODORE EISENBERG:
It includes the studies, I think, that they primary rely on
to get FDA approval such as the ASCOT study for Lipitor where in fact the
result for women was insignificant, but it was an increase in heart attack
risk, not a decrease.
DR. LARRY KASKEL:
All right so, can you explain to our audience what is a
meta-analyses study and are they better than regular studies or are they only
better if they prove your thesis?
THEODORE EISENBERG:
A traditional meta-analyses, the analyst will not have
access to the individual studies underlying data and the traditional
meta-analyses works only from the published versions of the data and what the
meta-analyses can allow you to do is to combine studies and perhaps support
findings that trend in individual studies, but aren’t significant in any
individual study, but then when you combine the analyses into a single study
that accounts for the uncertainty within each study, you get a single result
that can point you in a reasonable direction. So, for example, suppose you had
10 small or moderate-sized studies, none of which showed a significant effect,
but all of which went in the same direction, when you combine those to a
meta-analyses, you might well find that there is a statistically significant
effect and you might make policy based on that rather than dismiss each of the
studies individually.
DR. LARRY KASKEL:
And counselor, can you tell me what the results in your
findings were in your study?
THEODORE EISENBERG:
Well, I think there were 2 major findings. One, I think,
was known before. When we combined the 5 major high-quality clinical trials of
statins and these are different statins, but when we combined them, we found 2
things and this was looking at the primary prevention studies, one was that
there were consistently statistically significant benefits in heart outcomes
for men and what we found was there are equally consistently no significant benefits
in heart outcomes for women. When you combine the 2 studies, you get an added
result.
1. Significant protection for men.
2. No significant protection for women.
3. You get a statistically significant difference between
men and women, which I do not think has been observed before in this context.
I mean there is lot of context where we know men and women
are at different risk and I think we kind of knew, though not everyone
acknowledged it, that there was no real evidence of protection for women in
this context, but what there wasn’t before was evidence of a significant
difference between men and women. I think it is important because the National
Cholesterol Education Program acknowledges that the benefits of statins for
women are primarily based on extrapolation for men and so now that we had
evidence that there is a statistically significant difference between the men
and women, that position, I believe, is no longer so clearly scientifically
supportable. You can't extract from 1 group to another when there is serious
evidence that the groups are heterogeneous.
DR. LARRY KASKEL:
Let's not extrapolate, but let's take the results of your
findings. How can a drug company support a claim that statins reduce the risk
of heart attacks for women with this knowledge and this data out there? How
can they do that?
THEODORE EISENBERG:
I think, at this point, one has to be quite careful in
exactly what is they claim. If you look at through the 1 example I am most
familiar with of course is Lipitor and they are advertising probably most
people who watch TV at all or read the Wall Street Journal or other major media
have seen a Lipitor ad in there and the one that's most striking to me is one
with a picture of Robert Jarvik and a picture of the heart and in big fonts
saying Lipitor reduces risk of heart attack by 36% and that's literally true in
the following sense. In the 1 clinical trial on which their claim is based,
which is known as the ASCOT trial, the combined results for men and women did
reduce heart attacks by 36%. However, the results for men and women were
different. There was a 41% reduction for men and a 10% increase for women, but
the ad does not say reduction for women, but I think what the ad leaves out is
at least as important as what it puts in. If I were a physician prescribing
Lipitor for women, I would kind of want to know that the best point estimate of
what it does is increase the risk even that is not significant.
DR. LARRY KASKEL:
You'd wanted to decrease the risk you mean.
THEODORE EISENBERG:
If I am prescribing it and they haven't told me that it
increases the risk 10%, I think I would be angry.
DR. LARRY KASKEL:
All right, so let's explore that a little bit. Besides
being angry, what else could you do for what is potentially misrepresentation
or false advertising and will you take the case counselor?
THEODORE EISENBERG:
I won't take the case because these are expensive and these
are very difficult cases to win, but I think there are at least 2 ways to
reinforce potential misconduct by advertisers and by drug companies. One is
private losses, the ideal plaintiff would be someone who has seen the ad, say
look this drug reduces our risk of heart attack that plaintiff would be an
otherwise healthy women and she would go into court and say I would like my
money back because the evidence that Lipitor reduces my risk of heart attack is
no greater than the evidence that Coca Cola reduces my risk of heart attack
because Lipitor increased it by 10%, not decreased it for people in my group
and I think that they should give me my money back even if I haven't suffered
an adverse event because I paid for something that really there was no basis
and my doctor may have read their label which says results were inconclusive
for women, but it didn't tell him that actually risk increased for women. It
is technically true results were inconclusive because the increase wasn't
significant, but I think both me and my doctor should have been told a bit more
than they told us and I think we wouldn't allow stocks to be sold on this
basis. We shouldn't allow drugs to be sold on this basis.
DR. LARRY KASKEL:
And I would like to read to you counselor a quote from a
book and this quote was written in the year 1865 – the men who have excessive
faith in their theories or ideas are not only ill prepared for making
discoveries, they also make very poor observations. Of necessity, they observe
with a preconceived idea and when they device an experiment, they can see in
its results only a confirmation of their theory. In this way, they distort
observation and often neglect very important facts because they do not further
their aim. So, I thought of that when I recently looked at the Jupiter trial
and I know you have looked at that and is wondering if you would like to
comment on what we have learnt or not learnt from the Jupiter study,
statistically speaking.
THEODORE EISENBERG:
It's important to understand what the patient’s population
was in Jupiter. They excluded anyone with high cholesterol and they only
included people with high C-reactive protein, which is as I understand it a
marker for inflammation and so in some sense Jupiter is not directly comparable
with the prior primary intervention studies that interest me with respect to
women because in those studies people had on average either mildly or more than
mildly elevated cholesterol and other risk factors. Suppose you treat our
meta-analyses as one finding, so we don't have to talk about each individual
study. The studies being ASCOT, ALLHAT, AFCAPS, and PROSPER, but we combine
those for women and what we found is each of those studies showed a substantial
reduction in cholesterol of women and none of them showed a cardioprotective
effect. So, now you are over to Jupiter where you have taken out people with
even moderately elevated cholesterol by current standards and you give them a
different drug, not Lipitor in this case and they get a clear reduction in
adverse cardiac outcomes. So, I mean you put those 2 next to each other and it
seems to be saying, if you select people based on the risk factors in the 4
previous clinical trials that included some women, you get no result. If you
select people based on C-reactive protein, you get a result, which seems to
suggest that may be you shouldn't be giving these drugs to people selected on
the criteria in the 4 studies, which would include high cholesterol that is we
know from those studies statins do not work even though they lower cholesterol
in those populations. We know from Jupiter, there is a beneficial effect for a
different population. It seems to me that adds up to saying that cholesterol
and perhaps other things have been a giant red herring in getting the right
population on drugs that might help them.
DR. LARRY KASKEL:
Well, it has been very good for business for the last 20 to
30 years of treating cholesterol. I don't know that it has necessarily
prolonged lives at all, but it has been very good for shareholders.
THEODORE EISENBERG:
Yeah and I don't want to over-claim. We are not saying that
lowering cholesterol is never a good thing or worse thing is the evidence as we
see it is that lowering the cholesterol in all the former primary prevention
women studies didn't do any good and it seems to do some good when you don't
have cholesterol as the filter or cholesterol as one of the filters.
DR. LARRY KASKEL:
Counselor, what would you tell a young physician coming out
of their training that still believes in evidence-based medicine, if they were
going into primary care treating patients with high cholesterol.
THEODORE EISENBERG:
Oh! well, this is really hard because I would say you really
have to skeptically trust no one. You need to develop the skills to be able to
read the studies and to evaluate the studies by yourself without input from
interested parties because often on the face of the study, you will see things
that will lead you to treat one way versus the other and I think the bottom
line is I feel this is way about warriors too, we need more statistically
sophisticated doctors and we need to give them the time to do the reading they
need to do rather than to rely on third party supplying them summary
information.
DR. LARRY KASKEL:
Well, we do unfortunately rely on the third party coming in
for lunch, showing us a very pretty glossy poster showing a 50% reduction in
risk and that is a relative risk reduction which has kind of taken over in the
last 20 years and if you read the small print, they might tell you the absolute
risk reduction and I asked the drug reps in my office, what is the absolute
risk reduction and they usually don't know. They have to look it up and it is
usually a very small number meaning 1% to 2%.
THEODORE EISENBERG:
It is a really bit scary because for example the ASCOT
study, which was the one the FDA based approval of Lipitor with respect to it reducing
heart attacks, so may be a dutiful physician gets to the point of reading the
abstract, doesn't have time to read the whole article and actually think about
each, you know, grass and table in it. The abstract doesn't tell him there is
no result for women. The only way you get this is by, you know, reading the
whole article yourself, which is I imagine rather difficult with respect to the
time of most physicians, especially people with general practices, how can they
read, you know, the heart literature, the lung literature, they must be very
daunting. So, they have got to rely on things like national recommendations,
one hopes perhaps more than, you know, drug company representatives, but the
national recommendation in this case seemed to be not evidence based.
DR. LARRY KASKEL:
Counselor, in the news there has been a lot about preemption
lately and I was wondering if you could explain how legal concept applies to
marketing materials related to pharmaceuticals that have been approved by the
FDA.
THEODORE EISENBERG:
I think 2 different aspects of preemption were separating.
One is preemption with respect towards claims that a drug harmed someone. So,
there wasn't adequate disclosure of the risk. That's not the central focus of
what we were writing about, ours is about whether there is preemption of the
advertising claims of drug companies and there even the FDA in its most extreme
version has said advertising is different and that they have a much less
aggressive view of state law actions being preempted with respect to
advertising and they do with respect to things warned about on the label. My
primary concern with Lipitor is not that it's hurting people necessarily, it's
that of being sold a billion dollars or more a month to people many of whom
believe it would reduce the heart attacks and there is no evidence of that and
I think there is a strong case with those people if fully informed would want
their money back.
DR. LARRY KASKEL:
Well, on that note, I would like to thank our guest, attorney
Theodore Eisenberg, Henry Allen Mark Professor of Law and adjunct professor of
statistical sciences at Cornell University. Thank you very much for coming on
the show.
THEODORE EISENBERG:
Thank you very much.
DR. LARRY KASKEL:
I am Dr. Larry Kaskel, you have been listening to the
Business Of Medicine on ReachMD XM160, The Channel for Medical Professionals
and thanks for listening.
Hello, I am Dr. Susan Love from the Dr. Susan Love
Research Foundation and a clinical professor at the David Geffen School of
Medicine and you are listening to ReachMD, the Channel for Medical
Professional.
