The issue I've noticed from training at abnormal cadences is that as cadence increases at the same power, oxygen consumption, and therefore HR, increases. Momentary max power doesn't seem to be an issue as pro sprinters use normal crank lengths. Of course they have to get to the last kilometer near the front to even contest the sprint, so that might be a factor too. One might consider what one's max cadence OOS is, which would certainly affect max sprinting speed.