I am not implying earthquakes, suppose one of the cohorts had a condition not fully documented on the H&P, or died of an opportunistic infection it could skew results. That is why some of the other trials changed the primary endpoint away from all cause mortality.