USD573M in retained losses and a less than great 2020!. 2021 a year that cannot get worse, page-90

You cannot change a primary endpoint in any trial during or after because of the data. Its self-defeating - the whole point of trials is too have verification via independent unbias data. Think about it, you are making the trial fit the data for your purposes everyone would just manipulate their trials to fit the data to get a positive result. Defeats the whole notion of unbiased empirical assessment. Trial end points may sometimes be changed or added during trials, but this should be independent of the data, for example new discoveries made in other trials that shed new light on a disease - discovery of a critical bio-marker for example.

The secondary endpoints are not without merit, but ultimately are weaker than the same success in primary endpoints given the nature of statistical trials. That we had 60% mortality improvements in CHF during the secondary endpoints does not directly compare to a 60% improvement if it had been the primary endpoint in the same data/trial. It may be less - it may be more. Secondary endpoints ultimately should only be used to help interpret the primary result of the trial or to provide information, or prompts, for future research. Clearly CHF requires a further trial with mortality as the primary endpoint. GvHD does not.