The decoupling theorem of Appelquist-Carazzone says that if you want to decouple a particle, the low energy resulting theory need to be renormalizable. You can't do that for the top, because you break $SU(2)$ gauge symmetry, in fact, top is the bottom's partner in the doublet.
So when I use decoupling theory for $W$ and $Z$. Does the top remain in my theory, or I can substitute it with point-like vertex? If not, how do I handle a particle heavier than gauge boson $W$, in my effective theory?