Further reasoning on the choice of primary and secondary...

Further reasoning on the choice of primary and secondary endpoints in this study.

Endpoints are frequently ordered by clinical importance, with the most important being
86 designated as primary (e.g., mortality or irreversible morbidity). This is not always done,
87 however, for a variety of reasons. The most common reasons not to order endpoints by clinical
88 importance are if there are likely to be too few of the more clinically important endpoint events
89 to provide adequate power for the study, or if the effect on a clinically less important endpoint is
90 expected to be larger. In these cases, endpoints are often ordered by the likelihood of
91 demonstrating an effect. For example, time-to-disease progression is often selected as the
92 primary endpoint in oncology trials even though survival is almost always the most important
93 endpoint; the reasons being that an effect on disease progression may be more readily
94 demonstrable, may be detected earlier, and often has a larger effect size because the observed
95 effect on survival can be diluted by subsequent treatment post-progression. Section III.A
96 includes further discussion of the primary and secondary endpoint families. The determination
97 of which endpoints are primary, secondary, or exploratory, regardless of the reasons for the
98 determination, should always be made prospectively (see ICH E9).

https://www.fda.gov/media/102657/download