"Thanks.  Is it odd in your mind that it was ineffective for stroke?"

No this is not odd. It is quite common in CVOTs with 3-point MACE endpoint for not all 3 to show improved hazard ratio and/or statistical significance. Only 34 of the 274 3-point MACE events in BETonMACE were stroke; most were CVD death and non-fatal MI. If we had to chose the 2 out of 3 to have reduction, it is great that it was in the 2 events (CVD death and non-fatal MI) that comprised the majority of events. And it is especially great that it included CVD death. 

BDAZ