Why do individuals who are not dead have death age in endpoint data?

DEATH_AGE variable is the same as FU_END_AGE variable; age of death if individual has died, age of emigration if a person has moved abroad, or age at the time when register follow-up ends in registers (a.k.a FU_END_DATE).

As endpoint age (eg. DEATH_AGE/FU_END_AGE/T2D_AGE) in endpoint data is the first event age in cases and age at censoring from the data in controls, endpoint age variable can directly be used in survival analyses, for example using cox-models.

Take a look at the Complete follow-up time of the FinnGen registries – primary endpoint data section.