No, it is a theorem that the Newton divided difference $f[x_0,x_1,\dots,x_n]$ is invariant under permutation. This is tedious but mostly straightforward to prove by induction.
I think you have actually misunderstood Wikipedia's statement. I think they are probably referring to the algebraic form. For example, if you have a Newton interpolant $a+b(x-x_0)$ for two points and you add a third point $(x_2,y_2)$ (irrespective of how $x_2$ compares to $x_0$ and $x_1$), then you get $a+b(x-x_0)+c(x-x_0)(x-x_1)$. The first two terms stay the same! In the Lagrange form this does not hold, because each of the Lagrange basis elements depends on all of the points.