I was curious as to whether this was true, so to test it out, I looked at Hilbert's Problems[0], which are widely considered some of the most important problems in Mathematics in the 20th century. Of the solvers whose ages I could find on Wikipedia, the median age was 30.5 and the mean was 29.4. (For Hilbert's 10th problem the solution is listed as the joint work of 4 people. I treated them as one person with an average age of 42.5)
This obviously isn't statistically significant, but it lends some weight to the myth that Mathematicians "die young."
You see the effect in math because math requires the sharpest most complex pure thinking. Other major creative efforts (business, movies) involve collaboration and social/organizational structures that benefit from years of relationship-building and reputation-building. Or years of drudgery, if your innovation is something like the world's first dictionary.
This obviously isn't statistically significant, but it lends some weight to the myth that Mathematicians "die young."