Re: Lawsuit questions need for Lipitor: The ASCOT trial's 2000 women


David Rind wrote:
> fresh~horses wrote:
> > The study refererenced in Roni Rabin's Newsday article at the head of
> > this thread ("...2000 women...") is ASCOT, published in Lancet in April
> > 2003.
> >
> > And here is an snippet regarding the lawsuit, with the lawsuit website
> > linked below.
> >
> > David if you want the table for the ASCOT trial I can e-mail it as
> > attachment.
> >
> > MEM do you have a workable address? Anyone wants it say so on group,
> > please. Despammed is defunct.
> >
> >
> > Zee
> >
> > "Although Lipitor has been on the market since 1996,
> > it wasn't approved for the prevention of heart attacks
> > until 2004. The FDA approved it for this because a
> > large study (called the "ASCOT" study) showed that there
> > was a reduction in heart attacks overall for the
> > approximately 10,000 patients in that study. However, the
> > FDA looked at this study as a whole, rather than
> > looking at the different types of patients in the study. In
> > fact, the women in that study who had no prior history
> > of heart disease who took Lipitor actually had 10%
> > more heart attacks than the 1,000 women taking a placebo.
> > Despite the fact that the FDA approved Lipitor for
> > prevention of heart attacks, there is still no reliable
> > medical evidence that Lipitor or any other statin is
> > effective at preventing heart attacks for women and
> > people over 65, who have no history of heart disease."
>
> Okay, it probably should have occurred to me that this was from ASCOT-LLA.
>
> First a philosophical comment about looking at subgroup analyses: they
> are usually a bad idea. This applies both in negative studies where
> pharmaceutical companies try to find a subgroup the drug worked in, and
> positive studies where doubters try to find a subgroup the drug failed in.
>
> In general, a therapy that works in one group has simlar relative
> effects in all groups. There are exceptions, but this should be the
> default expectation. If I show you a study demonstrating that statins
> work in people, and you are short, left-handed, and have brown hair,
> it's not generally reasonable to expect that I prove to you that the
> study also showed benefit in short, left-handed, brunettes. In the
> absence of strong evidence to the contrary, the relative benefit should
> be assumed to be the same as in everyone else. By chance alone, when we
> start performing subset analyses we will find groups where a proven
> treatment appears not to work and where a useless treatment appears to work.
>
> The default assumption, in the absence of strong evidence to the
> contrary, should be that statins (and blood pressure medicines, and
> antibiotics, and vitamins) have the same relative effects in men and
> women. Note, by the way, that this does not apply to differences in
> absolute benefit -- women, on average, are at lower risk for coronary
> heart disease than men, and so, on average, will get less absolute
> benefit from statins than men. But we should expect that women have the
> same relative benefit and that a woman who happened to be at baseline
> high risk would have the same absolute benefit as a man with the same
> baseline high risk.
>
> This should even be the default assumption when a subgroup analysis
> appears to show a statistically different effect in one subgroup from
> another -- the likelihood of this being due to chance remains high if
> the study performed lots of subgroup analyses. However, this was not the
> case in ASCOT. ASCOT found no statistically significant difference in
> the effect of statins in men and women and so looking at the individual
> subgroups is particularly suspect.
>
> Second, ASCOT actually provided almost no data at all on the effects of
> statins in women. The comment quoted above about a 10% increase in
> events in women is just the sort of disingenous language that one might
> expect from a pharmaceutical company trashing a competitor's product or
> from someone with a stake in the outcome ignoring the actual evidence
> (not that one could ever imagine a litigator doing such a thing).



The litigator suggests that you read the whole press release and case
outline before you make a judgement about the case or the litigator. I
provided the link, and say to you; this is no ordinary litigator.



>
> There were thirty-six primary outcome events in women in ASCOT --
> THIRTY-SIX! Nineteen occurred in women taking atorvastatin and 17 in
> women taking placebo. Those two events out of 17 represent the 10%
> increase the article is apparently referring to. To say that this was
> not statistically significant and that the confidence intervals were
> wide doesn't begin to address how silly it is to make any conclusions
> about a differential effect of statins in men and women based on ASCOT.
>
> We have evidence from other studies with more events, by the way, that
> statins really do seem to have the same relative benefits in men and
> women just as we would have expected had we never bothered to look.

Of course I can't dispute you in any scientific way, but I offer you
this quote from someone who could (but doesn't have the time or
inclination). You'd respect his creds. If it's HPS you are referring
to; "HPS is crap".


>
> So, when the author of the quote above writes that the FDA "looked at
> the study as a whole, rather than looking at the different types of
> patients in the study", the response should be that indeed that's what
> the FDA did, and they did exactly the right thing.

That doesn't deny that the process is flawed. People are getting hurt
with this flawed outdated and, I would say, careless way of the FDA and
Health Canada carrying out their mandate to protect the consumer.



>
> --
> David Rind
> drind@xxxxxxxxxxxxxxxxxxxxx

.

Relevant Pages

  • Re: Lawsuit questions need for Lipitor: The ASCOT trials 2000 women
    ... >> it wasn't approved for the prevention of heart attacks ... >First a philosophical comment about looking at subgroup analyses: ... If I show you a study demonstrating that statins ... >absence of strong evidence to the contrary, the relative benefit should ...
    (sci.med.cardiology)
  • Re: Lawsuit questions need for Lipitor: The ASCOT trials 2000 women
    ... >>> it wasn't approved for the prevention of heart attacks ... >>First a philosophical comment about looking at subgroup analyses: ... If I show you a study demonstrating that statins ... >>absence of strong evidence to the contrary, the relative benefit should ...
    (sci.med.cardiology)
  • Re: Lawsuit questions need for Lipitor: The ASCOT trials 2000 women
    ... The FDA approved it for this because a large study (called the "ASCOT" study) showed that there was a reduction in heart attacks overall for the approximately 10,000 patients in that study. ... This applies both in negative studies where pharmaceutical companies try to find a subgroup the drug worked in, and positive studies where doubters try to find a subgroup the drug failed in. ... The default assumption, in the absence of strong evidence to the contrary, should be that statins have the same relative effects in men and women. ...
    (sci.med.cardiology)
  • Re: Lawsuit questions need for Lipitor: The ASCOT trials 2000 women
    ... >> it wasn't approved for the prevention of heart attacks ... If I show you a study demonstrating that statins ... > absence of strong evidence to the contrary, ... evidence that statins are beneficial in primary prevention of heart ...
    (sci.med.cardiology)
  • Re: Cholesterol Drug Has No Benefit in Trial
    ... Many cholesterol drugs do that, ... A clinical trial of Zetia, ... know men suffer more heart attacks, often at an early age, ... some few people have livers that react badly to statins. ...
    (soc.retirement)