When a0=0, the condition p∣0 is satisfied by every integer, so the theorem filters nothing. Factor out the largest power of x first.
x3−4x2+4x=x(x2−4x+4)=x(x−2)2
The roots are x=0 and x=2 (a double root). Testing only ±1,±2,±4 on the original polynomial would find 2 but miss the root 0.