<"What matters in death rate, not late thrombosis rate.">
Agreed - although I would add MI to that, and the vast majority of late thrombosis do result in either MI or death.
<"And if 0.75% of BMS patients are dying per year from silent restenosis then the DES is the better device.">
If this were true your point would be well taken - but, except for the setting of acute coronary syndrome, angioplasty / BMS have been shown NOT to reduce death or MI compared to medical (drug only) management. They simply help deal with angina.
<"Also, if the thrombosis is preventable, but the restenosis is not.">
Not everyone can afford clopidogrel, or tolerate it (allergies are common). You may find shortly after the stent is placed that you can't take the prophylaxsis.
<"Or if patients getting restenosis fixed by CABG die on the operating table at 2%, ... .">
Well, I suppose I might buy that one if there was data behind it. However I assume that death post-CABG would already be incorporated into the stats that the stent trials report. i.e. I would have assumed that if you have restenosis, then CABG, and then death, that you are still counted as a death for the relevant arm of the trial - yet no one is suggesting that the stent trials show lower death rates for DES compared to BMS.
Market Data 
Markets 
