Traduction assurée par un de mes aimables visiteurs, mais aussi vérifiée “by myself” …
What is wrong with ARBITER-6 HALTS?
The first essential rule in medicine clinical tests – besides lot drawing (or randomisation) – is that they must be led double blind, i.e. neither the investigators nor the patients know who takes the medicine and who is present as a witness (and takes the placebo).
This is not the case for ARBITER-6 HALTS and there’s no excuse for not doing it.
The second essential rule is to respect the terms of the procedure carefully: sample size and test duration.
If this is not done, we do not respect the principle of a priori assumption and all sorts of false ways are therefore possible.
The investigators of ARBITER-6 HALTS interrupted their test sharply with no particular reason.
They cut its head off.
These two violations of the clinical tests principles make the ARBITER-6 HALTS trial totally open to many bias.
Although the authors report a decrease of LDL (“bad” cholesterol) and an important increase of HDL (“good” cholesterol) due to the tested medicine (nicotinic acid), these results cannot be recognized as being protective for the heart unless playing dumb!
It seems important to specify that the technique used (measuring the thickness of the inner surface of the carotid artery through ultrasound) to estimate the artery disease that was supposed to be stopped in this test was severely criticized in recent surveys (particularly the ENHANCE test I described in a previous note) and a lot of experts had sworn they would never trust this technique to test a medicine.