Replies to post #157705 on Biotech Values

[iwfal]

RVX -

Rosu and Ator won't get IVUs reductions in six months of dosing.

Why not? - I cited hard data that they DO create IVUS reductions even at 1/2 max dose. It may not be as much at 6 months as at 12 - but even if it is only 1/2 the amount seen in that cite it is still enough that even a clinically useless RVX-208 can nudge it over the primary endpoint. And, more importantly, be the basis for a claim that the trial is a success even though the drug is essentially worthless.

The primary is change from baseline, but there is a control arm to compare the changes in baseline.

I would suggest you have missed the fact that the primary endpoint in the RVX trial DOES NOT COMPARE TO PLACEBO. (I truly apologize for shouting - but I don't think you are reading what I am writing).

As to your criticisms of the 2004 Asztalos paper and preference for the 1989 Gordon paper, there are a couple of things you should understand.

FWIW - Apologies for not reposting the link to my "preferred" paper in my last post - but will do so as part of my response below:

1) My preferred paper is NOT the Gordon paper. As I noted in my last post my preferred paper is the 2005 Asztalos paper (quote from my last post was "the later... better paper done by the same authors"). And it showed no such wild correlation. Essentially alpha1 largely disappears when you actually do a methodologically sound datamining. See link in #msg-81742249

2) My "preference" - this really isn't about preference. This is about a paper that should never have been published. And should certainly never be cited by anyone in the field. See #msg-85006464 for detailed critique of the 2004 paper. It numbers among its problems self inconsistency and being a truly classic example of way too much analysis on way too little data.

They concluded a 1% increase in a1-HDL generated a 26% decrease in CHD. This was supported internally by a1-HDL to HDL-C ratios having an even higher benefit -- which, to some extent, addresses the point that HDL-C measured by weight (which isn't how they measure it in modern studies) is largely a1-HDL.

The authors who concluded that alpha1 was the font of cardiac beneficence subsequently un-concluded it within 1 year. And, effectively, dropped the subject. (Nothing more current from them on the same topic). More generally I would strongly recommend that any paper trying to conclude anything about 20+ parameters with a 169 event database, especially with no checkset, should be thrown in the garbage. No need to bother even opening it. Especially if it is self inconsistent - which the 2004 paper is (see the Standard Deviations).

Recall the ETC-216 data didn't report a 4% difference from baseline in ETC-216. It reported a 4% difference between the growth in the placebo arm and the reduction in the control arm.

Ok - but not sure how that is pertinent to the fact that RVX have made their primary endpoint a comparison to baseline. NOT a comparison to placebo.

In conclusion - and just repeating a summary of my view. RVX has a history of reporting only data favorable to them, regardless of the datamining necessary, and entirely hiding the unfavorable. And the upcoming IVUS trial will allow them to continue that if the primary endpoint hits even while the comparison to placebo shows it as a dud.

BTW - I know you have been a strong proponent of placebo controlled trials in onc. What makes this any different?