WHI - Unblinding a study
Joan Livingston
happened with the WHI findings or over 1% of the hormone drug takers in this
double blinded study showing adverse cardio-vascular events and death.
This unblinding of the study and reporting on the difference that were found
among the hormone drug takers and the placbo group was claimed to have been
done as a result of a "contractual relation" the study had with the women.
Bullocks. This study like all studies is bound by the guidelines and
protocol on human experimentation. I recently posted the California law
version of these human drug experimentation protocols.
One can learn more about these types of trials and the seriousness of
unblinding them prematurely in the book "Deadly Medicine" by medical
investigative journalist Thomas Moore who tracked all the public records and
FDA hearing process on a heart arrythymia drug that was taken off the market
a few years back, because it was found the drug takers were dying in larger
numbers than the placebo group.
How this came about is probably instructional on what happened here. A
double blind study such as the WHI means that both the researchers and the
subjects do not know who belongs to either group: the placebo group or the
drug group (though this is hard when the HRT group of post meno women start
bleeding)
The statistics are tracked without the researchers knowing where the numbers
came from. In this other study, the researchers started seeing a trend
between the two group with this increased death rate. The assumed that the
higher deaths were among the placbo takers, which I suspect was also the
suspicion in this HRT study. And because the rates of difference were so
significant, it was decided to "unblind" the study to see what was
happening, or to bring the study to a close with "good news" that the drug
was doing so much better than the placebo. But in the case with the heart
group, there was "surprise" after it was unblinded, that it was the drug
that was causing the increase in deaths, not the placebo group.
Somehow I think this was the same scenario for these HRT findings, and why
there was some a common reaction of "surprise" mentioned in all the early
reports (but absent in the W/A spin a few days later - W/A did not say they
were surprised at the findings because this would give the findings
significance, they only said they were small to possibly disappearing, but
did not base this on any actual numbers or proof).
I think the researchers started seeing this increase in adverse C-V
events and death among the three groups. When it got to sigificant of
numbers, they then unblinded, and I bet you dollars to donuts they expected
to find that the placebo group had the highest CV events, the HRT group
fewer and the ERT group the least C-V events.
Instead after the unblinding came the "surprise" that over one percent of
the hormone drug takers and NOT the placebo group were having these adverse
C-V events and deaths.
The actual numbers and breakdown among the groups remains undisclosed. And
this scary spin on the actual numbers is now taking place. But keep in mind
no matter what you start hearing later is that OVER ONE PERCENT OF THE
HORMONE DRUG TAKERS HAD ADVERSE CARDIO-VASCULAR EVENTS AND DEATHS.
What we do not know, is how much over one percent. We do not know how this
broke down between the HRT users and the ERT users. We do not know how this
broke down among the various forms reported of (1) heart disease, (2)
strokes, (3) blood clots, and (4) deaths from any of these causes.
As taxpayers, this should be available information under public records.
The intentional coyness of the "researchers" in inexcusable right now as if
they had not been alarmed at the original findings, they would not have
'unblinded" this study.
There is a conflict in loyalties right now between the need to continue
this study and the importance of full disclosure of the actual numbers and
facts. The "unblinding" of this study is a significant event and should not
be discounted the way it is now being done in the media spin articles.
J
Chris Malcolm
[This is an *excellent* analysis by Joan. I usually edit posts down,
but this one deserves repetition, so I interpolate a few comments in
the whole posting.]
>Unblinding a double blind study is a Big Deal, such as what has recently
>happened with the WHI findings or over 1% of the hormone drug takers in this
>double blinded study showing adverse cardio-vascular events and death.
>This unblinding of the study and reporting on the difference that were found
>among the hormone drug takers and the placbo group was claimed to have been
>done as a result of a "contractual relation" the study had with the women.
>Bullocks. This study like all studies is bound by the guidelines and
>protocol on human experimentation. I recently posted the California law
>version of these human drug experimentation protocols.
You are quite right. However, it could be that part of the contract
with the subjects mentioned these protocols.
All too plausible.
> I think the researchers started seeing this increase in adverse C-V
>events and death among the three groups. When it got to sigificant of
>numbers, they then unblinded, and I bet you dollars to donuts they expected
>to find that the placebo group had the highest CV events, the HRT group
>fewer and the ERT group the least C-V events.
>Instead after the unblinding came the "surprise" that over one percent of
>the hormone drug takers and NOT the placebo group were having these adverse
>C-V events and deaths.
> The actual numbers and breakdown among the groups remains undisclosed. And
>this scary spin on the actual numbers is now taking place. But keep in mind
>no matter what you start hearing later is that OVER ONE PERCENT OF THE
>HORMONE DRUG TAKERS HAD ADVERSE CARDIO-VASCULAR EVENTS AND DEATHS.
> What we do not know, is how much over one percent. We do not know how this
>broke down between the HRT users and the ERT users. We do not know how this
>broke down among the various forms reported of (1) heart disease, (2)
>strokes, (3) blood clots, and (4) deaths from any of these causes.
> As taxpayers, this should be available information under public records.
>The intentional coyness of the "researchers" in inexcusable right now as if
>they had not been alarmed at the original findings, they would not have
>'unblinded" this study.
> There is a conflict in loyalties right now between the need to continue
>this study and the importance of full disclosure of the actual numbers and
>facts. The "unblinding" of this study is a significant event and should not
>be discounted the way it is now being done in the media spin articles.
It will be interesting to see what is said when they get round to
reporting this properly in a peer-reviewed journal where the editors
will insist on numbers and statistically justified comment.
Congratulations, Joan, this is a first class piece of scientific
journalism.
It is fascinating to see how much commercial pressures can foster a
culture where the scientists and research managers involved actually
genuinely believed the outcome would be favourable to them, despite
the ease with which pointers to the contrary can be found. Either
that, or company and profit loyalty so blinded the decision makers
that they simply refused to listen to their own scientists.
--
Chris Malcolm c...@dai.ed.ac.uk +44 (0)131 650 3085
School of Artificial Intelligence, Division of Informatics
Edinburgh University, 5 Forrest Hill, Edinburgh, EH1 2QL, UK
<http://www.dai.ed.ac.uk/daidb/people/homes/cam/> DoD #205
Joan Livingston
know who is taking what. All funded studies like this have an oversight
committee to insure it is following its guidelines.
Reading "Deadly Medicine" (Moore) gives you the best background on how these
types of studies operate. Yes, the researcher evaluate the data points to
see any developing trends, but only in extreme measures do they go in and
"unblind" the study and reveal this early information about the different
groups.
The differences between the two groups, the treated and the untreated had
to be alarming enough to need to intervene. And as I explained, the
decision to "unblind" the heart rythym drug was made because the researchers
assumed it was the drug that was benefiting the subjects and not the
placebo. They felt it was important enough of a difference to rethink the
continuation of the study. But to their surprise when it was unblinded, it
was the heart drug that was killing the subjects, not the placebo.
I would not be surprised to see that the same "conclusions" were reached
in this study because of this consistent report of "surprise." Now that it
appears the patients on hormone drugs were the ones harmed, this difference
that came to their attention to unblind the study is now "small - to
miniscule." If this were in fact the case, there would have been no reason
to unblind the study.
J
<ti...@cheerful.com> wrote in message
news:fs11fscp23qo9sdkf...@4ax.com...
> On 9 Apr 2000 13:27:16 GMT, c...@holyrood.ed.ac.uk (Chris Malcolm)
> wrote:
> Joan said:
> >> As taxpayers, this should be available information under public
records.
> >>The intentional coyness of the "researchers" in inexcusable right now as
if
> >>they had not been alarmed at the original findings, they would not have
> >>'unblinded" this study.
>
> I am missing something here. My understanding of "unblinding" is that
> the participants (subjects and researchers) in a study learn "who"
> specifically is taking "what". Is this a wrong understanding? I have
> not found any evidence that this has occurred.
>
> Somewhere in my surfing I read that there is a committee "overseeing"
> the study, said committee being totally uninvolved with the actual
> operation of the study. Obviously *somebody* has to have access to the
> computerized randomization or the results could never be known, and
> judging from the "interim" reports which are regularly made during the
> course of long term studies (eg the (in)famous Nurses' Study) it is
> not unusual to have scheduled examinations of the current situation.
> We don't of course know whether this was scheduled, but one at two
> years seems plausible.
>
> I don't in any way intend to play down what was found - it serves to
> solidify my view that HRT should not be used as a "preventative"
> measure, and agree with the narrow range of usefulness for symptom
> treatment which Joan proposed in an earlier post. The spin being put
> on it is certainly fishy and I intend to comment on this on my
> website, but I'm not convinced that the initial examination of the
> numbers had to have been precipitated by alarm.
>
> Tishy
>
Terri
> On 9 Apr 2000 13:27:16 GMT, c...@holyrood.ed.ac.uk (Chris Malcolm)
> wrote:
> Joan said:
> >> As taxpayers, this should be available information under public records.
> >>The intentional coyness of the "researchers" in inexcusable right now as if
> >>they had not been alarmed at the original findings, they would not have
> >>'unblinded" this study.
>
> I am missing something here. My understanding of "unblinding" is that
> the participants (subjects and researchers) in a study learn "who"
> specifically is taking "what". Is this a wrong understanding? I have
> not found any evidence that this has occurred.
>
> Somewhere in my surfing I read that there is a committee "overseeing"
> the study, said committee being totally uninvolved with the actual
> operation of the study.
Yes, but it would be unusual to provide unblinded numbers of adverse events to
the researchers unless these numbers were significant. At least two studies
which seemed to be favorable to the drugs being tested were unblinded and halted
when these results seemed apparent. It was done with a statin (mevachor) used in
a study of military personnel and with the tamoxifen study. The argument was
that it waa no longer ethical to keep the "benefits" from the placebo group. In
neither case were the numbers compelling once they were published. I agree with
Joan that the researchers asked for this analysis because they *expected* to
find evidence of benefits in the hormone group and evidence of harm in the
placebo group. When it didn't work out that way and the opposite proved to be
true, they tried to spin the results and make them seem almost insigfnicant. The
study may not "unblinded" yet, but the researchers do know that the harm is in
the hormone group. W/A in particular does not ant this study to continue until
the year 2005. The numbers of breast cancers their drug causes will then be
apparent, and the drug may be forced off the market because there are specific
rules about the medical use of carcinogens
Terri
Joan Livingston
difference, and they had -hoped it- was because of the drug, but they did
not know this ahead of time.. The book explains this whole process so well
and in great day by day detail. It reads like any great fiction thriller,
but it is non-fiction and the author sifted through a lot of hearing
transcripts and correspondence to piece this whole thing together.
Bottomline: unlbinding a study is a very Big Deal.
And the figures apparently were significant enough to unblind the WHI
study. Or I wonder what exactly this little "contract" with the subjects
actually said? Did it say we will unblind this study if we find even the
"minutest, most miniscule, most insignificant" difference between the two
study groups (or 3study groups ERT.HRT and placebo)?
In the interests of public integrity here they need to give more
specifics. How much more than 1% of the drug takers were affected? How many
died? How did it break down between the the HRT takers and the ERT takers?
How did it break down between the various conditions: heart attack; stroke,
blood clots and death.
J
<ti...@cheerful.com> wrote in message
news:0cc1fsk4l2mv9elho...@4ax.com...
> On Sun, 09 Apr 2000 15:14:03 GMT, "Joan Livingston"
> <joan.liv...@gte.net> wrote:
>
> > And as I explained, the
> >decision to "unblind" the heart rythym drug was made because the
researchers
> >assumed it was the drug that was benefiting the subjects and not the
> >placebo.
>
> So this means that they must have known that "these" subjects fell
> into group A, while "those" subjects fell into group B - but they
> didn't know which group was the placebo one? If so, then what you
> were/are saying makes sense to me. I hadn't been contemplating this
> semi-knowledge of who was what.
>
> > They felt it was important enough of a difference to rethink the
> >continuation of the study. But to their surprise when it was unblinded,
it
> >was the heart drug that was killing the subjects, not the placebo.
>
> Your surmise only, I take it - but one which I find eminently
> reasonable.
>
> I found it interesting that the official update made *no* mention of
> any numbers at all - in the sense of less/more than 1% or whatever.
> Neither was there any mention of death.
>
> Tishy
>
Terri
Livingston" <joan.liv...@gte.net> wrote:
> Not clearly stated. My apologies, the researchers noticed there was a
> difference, and they had -hoped it- was because of the drug, but they did
> not know this ahead of time.. The book explains this whole process so
> well
> and in great day by day detail. It reads like any great fiction thriller,
> but it is non-fiction and the author sifted through a lot of hearing
> transcripts and correspondence to piece this whole thing together.
> Bottomline: unlbinding a study is a very Big Deal.
>
> And the figures apparently were significant enough to unblind the WHI
> study. Or I wonder what exactly this little "contract" with the subjects
> actually said? Did it say we will unblind this study if we find even the
> "minutest, most miniscule, most insignificant" difference between the
> two
> study groups (or 3study groups ERT.HRT and placebo)?
>
> In the interests of public integrity here they need to give more
> specifics. How much more than 1% of the drug takers were affected? How
> many
> died? How did it break down between the the HRT takers and the ERT
> takers?
> How did it break down between the various conditions: heart attack;
> stroke,
> blood clots and death.
>
> J
I agree. I also question the assertion that the rate was lower than that
in the general population, unless they loaded the general population
sample by including more older women in it. if the deaths and adverse
events were actually lower than in a general population sample *with the
same age distribution* why would this cause alarm? Why would there be
any need to report this to anyone? After all, you are apparently still
safer taking hormones in the WHI than taking nothing in the general
population. better to take placebo, but still no need to worry if you
are taking homrones. Does being in the study provide some measure of
protection?
I think the Post got it right including the mention of deaths because
someone in the NIH leaked the figures to the press. Once the cat was out
of the bag, all that was left was to spin the numbers to make them look
as innocent as possible.
Terri
Joan Livingston
For an example of a very well-researched study of heart arrythymia drugs
trial that led to a mid study "unblinding" please read "Deadly Medicine" by
Thomas Moore.
I was using that model to -speculate- about this early "unblinding" of the
WHI study. And unblinding means that the researchers went in an
intentionally found out who was taking what drug, because of a concern that
something was happening to one group that was not happening in the other.
In the heart drug study, more men were dying of heart problems in one
group so the decision was made to "unblind" the groups to find out who was
taking what to see why there was this emerging difference. The book talked
about the researchers' "surprise" because they had assumed that it would be
the placbo group that was showing more deaths, because these drugs had been
on the market already being sold as heart benefit drugs. (sound familiar).
The surprise was that it was a certain class of the heart drugs that showed
the higher deaths, not the placebo group.
So based on the history of that study which was set forth in great detail in
the book, with cites, names, hearing dates and correspondence quoted, I was
speculating about why this WHI study may have been unblinded at this point.
The scary reality after finding these heart drugs were killing more men
than a placebo, was that the doctors after they found this out did not stop
prescribing these drugs. They did not want to have to tell their patients
they had been giving them bad drugs. All the details are in the book.
The same author also wrote an excellent book years ago about all of the
NIH and FDA hearings on the "cholesterol issue" and the mega-billion dollar
heart disease industry, along the same lines of investigating hearing
transcripts and sorting out all the incestuous relationships between the
industry and "expert review" panels. This one was called "Heart Disease" by
Thomas Moore.
Hope you can read these books .They are both educational and eyeopening.
And they are very readable. They read like a mystery story, looking for
clues, looking for answer and trying to sort out all the possible guilty
parties.
J
Eva D. Struction <EvaD...@aol.com> wrote in message
news:EdaI4.6533$YB4.4...@bgtnsc05-news.ops.worldnet.att.net...
> I'm sorry, but this whole discussion has lost me. Can someone please
> explain what happened? I mean the "unblinding" part--I don't understand
it
> at all--who did what? Where does the "heart drug" come into it?
>
> Thanks Eva
>
>
>
>
mro...@wwdc.com
where the WHI is concerned.
This study was not "unblinded". (you could have a good discussion as
to the possibilities of truly blinding any study, particularly one
where the treatment arm(s) use estrogen)
Today, all protocols of this nature are overseen by an independent
Data Safety and Monitoring Board (DSMB).
This board has the authority to stop a protocol should their
examination of the data reveal either an excess of risk, or an
unequivocal benefit, in any trial arm.
The DSMB is in place to protect the interests of the human subjects
participating in the trial. To avoid the unfortunate past instances
where protocols proceeded against the interests of the participants,
for example, the often cited anti-arrhythmia studies.
A more current example is the termination of the doxazosin arm of the
ALLHAT study, due to an excess risk in the doxazosin arm. The rest of
the study proceeds.
In the case of the WHI, the DSMB made some observations which did not
reach the level of statistical certainty required to stop what many
understand to be a very important protocol. So, the study goes on.
With the caveat that "what I believe" is the lowest form of human
thought....
I believe that the WHI will fail to validate oral estrogen with or
without MPA as an important risk management strategy for these
reasons;
1. You can not prevent something that is not going to happen. The
WHI DSMB commented on the low level of risk the WHI population
represents. If there is nothing to be prevented, all you do is expose
the patient to the risk of the therapy. All risk, no benefit.
This is not unusual for a hormone study. If you run the baseline
patient characteristics from the often cited PEPI study through the
Framingham Risk Assessment Tool, you will appreciate that these women
had a level of risk lower than the population. The alterations in
lipids at the end of the study do not change the assessment.
This basically supports another finding you have all heard about, that
users of HRT/ERT are healthier than non-users.
2. The WHI initiative uses an oral estrogen (a particularly archaic
and inappropriate one, I believe).
Oral Estrogens;
-negatively effect the Renin Angiotensin System.
-negatively effect the Haemostatic System.
-negatively effect the quality of lipid.
-negatively effect the Somatotropic axis.
-negatively effect the handling of glucose.
3. The WHI uses daily MPA, a progestin that has blunted the putative
beneficial effects of estrogen on the cardiovascular system in most
evaluations.
If the risk for cardiovascular disease is real, or if CVD is present,
then there are proven strategies to use; exercise, stop tobacco,
control blood pressure (ACE Inhibitors, diuretics and Beta blockers),
manage lipids with a statin, use ASA.
HRT/ERT is no substitute for these proven measures, and should not be
taken in their place.
Wyeth/Ayerst knows the gig is up on horse urine, if their recently
approved submission for a patch is any indication.
regards,
Mark.
On Sun, 09 Apr 2000 02:22:01 GMT, "Joan Livingston"
<joan.liv...@gte.net> wrote:
>Unblinding a double blind study is a Big Deal, such as what has recently
>happened with the WHI findings or over 1% of the hormone drug takers in this
>double blinded study showing adverse cardio-vascular events and death.
>
>This unblinding of the study and reporting on the difference that were found
>among the hormone drug takers and the placbo group was claimed to have been
>done as a result of a "contractual relation" the study had with the women.
>Bullocks. This study like all studies is bound by the guidelines and
>protocol on human experimentation. I recently posted the California law
>version of these human drug experimentation protocols.
>
> I think the researchers started seeing this increase in adverse C-V
>events and death among the three groups. When it got to sigificant of
>numbers, they then unblinded, and I bet you dollars to donuts they expected
>to find that the placebo group had the highest CV events, the HRT group
>fewer and the ERT group the least C-V events.
>
>Instead after the unblinding came the "surprise" that over one percent of
>the hormone drug takers and NOT the placebo group were having these adverse
>C-V events and deaths.
>
> The actual numbers and breakdown among the groups remains undisclosed. And
>this scary spin on the actual numbers is now taking place. But keep in mind
>no matter what you start hearing later is that OVER ONE PERCENT OF THE
>HORMONE DRUG TAKERS HAD ADVERSE CARDIO-VASCULAR EVENTS AND DEATHS.
>
> What we do not know, is how much over one percent. We do not know how this
>broke down between the HRT users and the ERT users. We do not know how this
>broke down among the various forms reported of (1) heart disease, (2)
>strokes, (3) blood clots, and (4) deaths from any of these causes.
>
> As taxpayers, this should be available information under public records.
>The intentional coyness of the "researchers" in inexcusable right now as if
>they had not been alarmed at the original findings, they would not have
>'unblinded" this study.
>
> There is a conflict in loyalties right now between the need to continue
>this study and the importance of full disclosure of the actual numbers and
>facts. The "unblinding" of this study is a significant event and should not
>be discounted the way it is now being done in the media spin articles.
>
>J
>
Terri
> There seems to be some confusion over just what has happened recently
> where the WHI is concerned.
>
> This study was not "unblinded". (you could have a good discussion as
> to the possibilities of truly blinding any study, particularly one
> where the treatment arm(s) use estrogen)
We've had this discussion. But this might be a good time to have it again. I
have no idea how they are hiding the nature of the drug from these
participants. Since they are using prempro we know that large numbers of them
will have some degree of vaginal bleeding for some period of time. For me this
would be the tipoff. Are these women subjected to endometrial biopsy even though
the cause of the bleeding is most assuredly the drug in an attempt to prevent
them from learning that they are taking the study drug? Are they subjected to
unnecessary worry about postmenopausal bleeding and its connection to
endometrial cancer? Are all required to have regular endometrial biopsies to
find out what the drug actually does to the endometrium and whether or not it
actually does offer protection from endometrial cancer? Those are a few of the
questions we posed in this discussion a couple of years ago. I don't know the
answers.
>
>
> Today, all protocols of this nature are overseen by an independent
> Data Safety and Monitoring Board (DSMB).
>
> This board has the authority to stop a protocol should their
> examination of the data reveal either an excess of risk, or an
> unequivocal benefit, in any trial arm.
So tell us how this works. At what level of benefit or harm is the data deemed
significant? Why inform participants if the level of harm does not reach the
level of "significance?" If women in the study are actually at lower risk than a
similar segement of the general population, then even those on hormones are not
being 'harmed." They are simply getting less benefit than the ones on placebo.
What politics and drug company input is there into the decisions made by this
DSMB. Independent doesn't always mean what the dictionary says it means.
>
>
> The DSMB is in place to protect the interests of the human subjects
> participating in the trial. To avoid the unfortunate past instances
> where protocols proceeded against the interests of the participants,
> for example, the often cited anti-arrhythmia studies.
>
> A more current example is the termination of the doxazosin arm of the
> ALLHAT study, due to an excess risk in the doxazosin arm. The rest of
> the study proceeds.
>
> In the case of the WHI, the DSMB made some observations which did not
> reach the level of statistical certainty required to stop what many
> understand to be a very important protocol. So, the study goes on.
Who decides that this is a "very important protocol?" If it's deemed to be "very
important" is a higher level of harm required to stop the study than if it's
deemed to be of less importance? The tamoxifen numbers for example were less
than compelling, especially when the study population was broken down by age of
study subjects. Yet the study was halted and press releases about a new miracle
for women's health, which Novartis has exploited in both print and other media
ads ( and twice been cautioned by the FDA for the content of those ads) abounded
in every newspaper, and on all news broadcasts. other studies done in other
countries which were carried through to completion showed little or no benefit.
I've left the rest of your post because I think it's important. Can you direct
me to large scale studies on transdermal estrogen?
Terri
dro...@home.com
>mro...@wwdc.com wrote:
>
>> 3. The WHI uses daily MPA, a progestin that has blunted the putative
>> beneficial effects of estrogen on the cardiovascular system in most
>> evaluations.
>>
>> If the risk for cardiovascular disease is real, or if CVD is present,
>> then there are proven strategies to use; exercise, stop tobacco,
>> control blood pressure (ACE Inhibitors, diuretics and Beta blockers),
>> manage lipids with a statin, use ASA.
>>
>> HRT/ERT is no substitute for these proven measures, and should not be
>> taken in their place.
>>
>> Wyeth/Ayerst knows the gig is up on horse urine, if their recently
>> approved submission for a patch is any indication.
>>
>> regards,
>>
>> Mark.
I am piggybacking on Terri's response, my newserver us acting up
today and I have not yet received the original, so who know when this
will get out come to think of that.
I have been thinking about W/A's new patch in the past few days, it
was approved last August, yet looking at the FDA website they haven't
even selected a name for it let alone started to market this. Why?
Was it because they wanted the furore over these new estrogen
/progestin findings to settle down first?
Was it because they did not like the required labelling for this
product on cardiovascular risk?
See the list of new drug approvals on the FDA website, W/A's
(17-beta-estradiol transdermal) comes up on the top in the
alphabetizing of the list. click on 9/24/99 in the label column to
read the professional and patient insert information.
http://www.fda.gov/cder/approval/index.htm
Kathryn
dro...@home.com
Joan Livingston
the term more loosely to refer to the exact situation you stated: that they
went in to find out what was happening to the different group. Real
unblinding woul let both the subjects and researchers know, and this has
not happened (exept for this bleeding occurrance potential, but who knows
maybe a placebo expectation of bleeding will lead to some incidence of
bleeding among that group too.)
Good to have your far more educated input on this, Mark. Stick around.
J
<mro...@wwdc.com> wrote in message
news:v3n3fs4dls961md36...@4ax.com...
mro...@wwdc.com
>mro...@wwdc.com wrote:
>
>> There seems to be some confusion over just what has happened recently
>> where the WHI is concerned.
>>
>> This study was not "unblinded". (you could have a good discussion as
>> to the possibilities of truly blinding any study, particularly one
>> where the treatment arm(s) use estrogen)
>
>We've had this discussion. But this might be a good time to have it again. I
>have no idea how they are hiding the nature of the drug from these
>participants. Since they are using prempro we know that large numbers of them
>will have some degree of vaginal bleeding for some period of time. For me this
>would be the tipoff. Are these women subjected to endometrial biopsy even though
>the cause of the bleeding is most assuredly the drug in an attempt to prevent
>them from learning that they are taking the study drug? Are they subjected to
>unnecessary worry about postmenopausal bleeding and its connection to
>endometrial cancer? Are all required to have regular endometrial biopsies to
>find out what the drug actually does to the endometrium and whether or not it
>actually does offer protection from endometrial cancer? Those are a few of the
>questions we posed in this discussion a couple of years ago. I don't know the
>answers.
Sorry, but I have not been able to keep up, and had not realised this
had been discussed.
Randomised, double blind placebo controlled protocols are kind of like
the proverbial saying regarding capitalism, not perfect, but the best
so far.
I am only familiar with the protocol in the most superficial of ways.
But, I doubt very much that biopsies would be done to maintain
blinding. They are uncomfortable, and would likely serve as a
deterrent to recruitment, or ongoing participation. More likely,
biopsies would be ordered during the protocol for the usually
understood reasons of medical need (generally bleeding).
Bleeding does not equate to cancer.
>So tell us how this works. At what level of benefit or harm is the data deemed
>significant? Why inform participants if the level of harm does not reach the
>level of "significance?" If women in the study are actually at lower risk than a
>similar segement of the general population, then even those on hormones are not
>being 'harmed." They are simply getting less benefit than the ones on placebo.
>What politics and drug company input is there into the decisions made by this
>DSMB. Independent doesn't always mean what the dictionary says it means.
>
The generally accepted level of significance in medicine is a p value
of less than 0.05. This means that the observation may be explained
by chance less than 5% of the time, or, less than 1 time in 20.
p<0.05 is not a very rigorous standard, and there are many instances
of hypothesis' generated in small numbers at this level of
significance which do not pan out in larger studies.
While I do not know, my suspicion would be that the DSMB would use the
0.05 level in their decisions.
That they did not terminate the study suggests that the excess of
events in the treated groups could be explained by chance.
DSMB's are entirely independent, and function autonomously from a
study's sponsor(s). If an excess of events was present in the WHI
treatment arms, and this excess could not be explained by chance, you
can be sure that the study would have been stopped.
The closer medicine looks at current prescribing practice (horse urine
and continuous MPA), the more glaring the deficiencies become. You
could argue that the premature termination of this protocol, before
solid answers (to the extent that any one study can deliver broadly
applicable answers) come available would suit the supplier of this
product more than any other interest [group]. Then they could fall
back on a reliable position supported by an absence of data and
innuendo.
>Who decides that this is a "very important protocol?" If it's deemed to be "very
>important" is a higher level of harm required to stop the study than if it's
>deemed to be of less importance? The tamoxifen numbers for example were less
>than compelling, especially when the study population was broken down by age of
>study subjects. Yet the study was halted and press releases about a new miracle
>for women's health, which Novartis has exploited in both print and other media
>ads ( and twice been cautioned by the FDA for the content of those ads) abounded
>in every newspaper, and on all news broadcasts. other studies done in other
>countries which were carried through to completion showed little or no benefit.
There is no prospective, end point driven information on the
usefulness of estrogen as primary prevention of cardiovascular
disease. It is an important protocol. Unfortunately, it is only
testing one way of doing HRT/ERT.
Far as I know, tamoxifen is a product of Astra/Zenica, and various
generic companies, not Novartis. But, I could be wrong.
>
>I've left the rest of your post because I think it's important. Can you direct
>me to large scale studies on transdermal estrogen?
There are no large scale, end point driven, studies on transdermal
estrogen in place (or contemplated) that I am aware of. In my
opinion, the various suppliers of transdermal products should
collaborate on a large scale study. Before the final results of WHI
(which I believe will finish off orals) spoil the possibilty of
recruiting.
On the other hand, the manufacturers would only need to do them if
they are interested in a product claim of "lowering CV risk".
Lots of estrogen was sold before W/A began throwing this spurious
claim around, and lots will be sold with out it.
Industry can do quite well without the risk and expense. Ultimately,
it is in the patients interest that these studies get done.
regards,
Mark.
Terri
> On Mon, 10 Apr 2000 11:33:08 -0400, Terri <vl-h...@erols.com> wrote:
>
> >mro...@wwdc.com wrote:
> >
> >> There seems to be some confusion over just what has happened recently
> >> where the WHI is concerned.
> >>
> >> This study was not "unblinded". (you could have a good discussion as
> >> to the possibilities of truly blinding any study, particularly one
> >> where the treatment arm(s) use estrogen)
> >
> >We've had this discussion. But this might be a good time to have it again. I
> >have no idea how they are hiding the nature of the drug from these
> >participants. Since they are using prempro we know that large numbers of them
> >will have some degree of vaginal bleeding for some period of time. For me this
> >would be the tipoff. Are these women subjected to endometrial biopsy even though
> >the cause of the bleeding is most assuredly the drug in an attempt to prevent
> >them from learning that they are taking the study drug? Are they subjected to
> >unnecessary worry about postmenopausal bleeding and its connection to
> >endometrial cancer? Are all required to have regular endometrial biopsies to
> >find out what the drug actually does to the endometrium and whether or not it
> >actually does offer protection from endometrial cancer? Those are a few of the
> >questions we posed in this discussion a couple of years ago. I don't know the
> >answers.
>
> Sorry, but I have not been able to keep up, and had not realised this
> had been discussed.
No need to be sorry. I'm glad you brought it up since it re-opens the discussion.
>
>
> Randomised, double blind placebo controlled protocols are kind of like
> the proverbial saying regarding capitalism, not perfect, but the best
> so far.
Understood. But some drugs are harder to do this kind of study with than others, and
I think this is especially true of estrogen.
>
>
> I am only familiar with the protocol in the most superficial of ways.
> But, I doubt very much that biopsies would be done to maintain
> blinding. They are uncomfortable, and would likely serve as a
> deterrent to recruitment, or ongoing participation. More likely,
> biopsies would be ordered during the protocol for the usually
> understood reasons of medical need (generally bleeding).
>
> Bleeding does not equate to cancer.
Of course not but postmenopausal women are warned of the connection and generally
advised to seek medical attention should they have bleeding. I think this is a wise
precaution. But it does lead to anxiety and concern which would not be justfied when
the bleeding is the expected result of drug use. As to the discomfort etc., There
have been small studies showing that this protocol (continuous combined estroen/MPA)
does not fully protect the endometrium. I suspect that this may be something else
this study may be trying to get some accurate data on.
>
>
> >So tell us how this works. At what level of benefit or harm is the data deemed
> >significant? Why inform participants if the level of harm does not reach the
> >level of "significance?" If women in the study are actually at lower risk than a
> >similar segement of the general population, then even those on hormones are not
> >being 'harmed." They are simply getting less benefit than the ones on placebo.
> >What politics and drug company input is there into the decisions made by this
> >DSMB. Independent doesn't always mean what the dictionary says it means.
> >
>
> The generally accepted level of significance in medicine is a p value
> of less than 0.05. This means that the observation may be explained
> by chance less than 5% of the time, or, less than 1 time in 20.
>
> p<0.05 is not a very rigorous standard, and there are many instances
> of hypothesis' generated in small numbers at this level of
> significance which do not pan out in larger studies.
>
> While I do not know, my suspicion would be that the DSMB would use the
> 0.05 level in their decisions.
Thank you. I understand.
>
>
> That they did not terminate the study suggests that the excess of
> events in the treated groups could be explained by chance.
>
> DSMB's are entirely independent, and function autonomously from a
> study's sponsor(s). If an excess of events was present in the WHI
> treatment arms, and this excess could not be explained by chance, you
> can be sure that the study would have been stopped.
Hmm... I'm wondering why the studies on the safety of viagra were not stopped then?
>
>
> The closer medicine looks at current prescribing practice (horse urine
> and continuous MPA), the more glaring the deficiencies become. You
> could argue that the premature termination of this protocol, before
> solid answers (to the extent that any one study can deliver broadly
> applicable answers) come available would suit the supplier of this
> product more than any other interest [group]. Then they could fall
> back on a reliable position supported by an absence of data and
> innuendo.
Yes. I think by the time 2005 rolls aroung W/A is going to wish they had never heard
of the WHI, since we already know with reasonable certainty that there *will* be an
excess of breast cancer in the hormone group and that it will show up by then.
>
>
> >Who decides that this is a "very important protocol?" If it's deemed to be "very
> >important" is a higher level of harm required to stop the study than if it's
> >deemed to be of less importance? The tamoxifen numbers for example were less
> >than compelling, especially when the study population was broken down by age of
> >study subjects. Yet the study was halted and press releases about a new miracle
> >for women's health, which Novartis has exploited in both print and other media
> >ads ( and twice been cautioned by the FDA for the content of those ads) abounded
> >in every newspaper, and on all news broadcasts. other studies done in other
> >countries which were carried through to completion showed little or no benefit.
>
> There is no prospective, end point driven information on the
> usefulness of estrogen as primary prevention of cardiovascular
> disease. It is an important protocol. Unfortunately, it is only
> testing one way of doing HRT/ERT.
>
> Far as I know, tamoxifen is a product of Astra/Zenica, and various
> generic companies, not Novartis. But, I could be wrong.
Yes it's sold under the name Novartis. I misspoke.
>
> >
> >I've left the rest of your post because I think it's important. Can you direct
> >me to large scale studies on transdermal estrogen?
>
> There are no large scale, end point driven, studies on transdermal
> estrogen in place (or contemplated) that I am aware of. In my
> opinion, the various suppliers of transdermal products should
> collaborate on a large scale study. Before the final results of WHI
> (which I believe will finish off orals) spoil the possibilty of
> recruiting.
>
> On the other hand, the manufacturers would only need to do them if
> they are interested in a product claim of "lowering CV risk".
Selling this stuff on a needed for a lifetime basis will be much harder without it
though. We see women coming here frequently saying they have been given a
prescription because of a family history of heart disease. And the AHA has been
pushing its use for this purpose for a couple of years now. Of course the
"unrestricted educational grant" they got from W/A may have something to do with
that.
>
>
> Lots of estrogen was sold before W/A began throwing this spurious
> claim around, and lots will be sold with out it.
>
> Industry can do quite well without the risk and expense. Ultimately,
> it is in the patients interest that these studies get done.
>
> regards,
>
My thanks for your detailed and informative response. I join Joan in asking you to
stick around.
Terri
>
Terri
> >
> > Far as I know, tamoxifen is a product of Astra/Zenica, and various
> > generic companies, not Novartis. But, I could be wrong.
>
> Yes it's sold under the name Novartis. I misspoke.
I misspoke again. It's sold under the name of Nolvadex. I'll get it right yet.
Terri
>
>
Joan Livingston
to give their patients the "Gail Model" to find out their breast cancer risk
factor ( series of 6 questions) and then they can be "counseled" about
taking preventive tamoxifin, I gather. BE SURE TO READ THE SIDE EFFECTS FOR
TAMOXIFEN
More on the "Gail Model" in J. National Cancer Institute 1998-90: 1371
J
Terri <vl-h...@erols.com> wrote in message
news:38F25926...@erols.com...
> Terri wrote:
>
> > >
> > > Far as I know, tamoxifen is a product of Astra/Zenica, and various
> > > generic companies, not Novartis. But, I could be wrong.
> >
> > Yes it's sold under the name Novartis. I misspoke.
>
dro...@home.com
<joan.liv...@gte.net> wrote:
>, but who knows
>maybe a placebo expectation of bleeding will lead to some incidence of
>bleeding among that group too.)
>
There was some bleeding in the placebo group in the PEPI trial just as
there are cases of post menopause bleeding in women now who are not
taking any hormones, it does happen and it is good to find out what
the natural rate of this occurance is IMO.
The fact that post menopause bleeding occurs shows me that some women
are definetely better off staying away from HRT, that is for sure. ( I
had some minor spotting last fall )
>Good to have your far more educated input on this, Mark. Stick around.
>
>J
Yes I too am glad to have your comments on this study Mark.
Kathryn
dro...@home.com
Robert Ames
>Oral Estrogens;
>
>-negatively effect the Renin Angiotensin System.
>-negatively effect the Haemostatic System.
>-negatively effect the quality of lipid.
>-negatively effect the Somatotropic axis.
>-negatively effect the handling of glucose.
Would you comment on the last sentence above, please? Is the effect
on glucose transporters? Doesn't oral estrogen reduce the effect of
TNF, which would increase insulin sensitivity?
It's a matter of some debate here whether HRT results in increase of
body fat. Of course, progestins can cause increase in appetite, but
what is the contribution of estrogen, and can this be negated if
transdermal estrogen is used instead?
--
"If dual standards are good enough for the FDA
they're good enough for me" -- Terri