It's rather known that this theorem is "correct" (otherwise it wouldn't be one! :P). I'm sure you can find many proofs for it, the simplest one is by induction. See if you can prove it by yourself.

But why do you think having 0 as an eigenvalue would make this incorrect?