It is 7.2 events per 100 patient years, which is the same as 7.2% annual rate. As the trial goes on, more patient years are accumulated. But it is always normalized to 100 patient years. Some patients have been dosed greater than 36 months. Other patients only ~ 9 months. Expressing the event rate per patient years normalizes this variability in dosing. Once BETonMACE reports full data, then the event rates will likely be officially reported at the time corresponding to the median dosing or median follow up period. Initially, this was projected to be 18 months. But due to the change in the protocol to allow dosing beyond 104 weeks and until the trial ends, the median dosing/follow up period is going to be at a time point beyond 18 months. At 18 months in EXAMINE, the event rates were about 11.5 percent. 

Also, typo in my previous post. "if" should have been "of"

"Take care not to confuse projected event rates after 18 months of dosing with observed event annual rates or events per 100 patient years."