Agreed about the cheeks looking odd being inflated after the puff of air.
I find the whole animation a bit hard to follow because you have two distinct, important actions happening at once (the back-creature sprouting, and the puffing), instead of merely overlapping the end of one with the beginning of the other. This means the viewer has to watch the animation at least twice to fully register everything. I think it would read better if the big creature exhaled after the back-creature was fully sprouted. It would mean extending the animation by a few frames, however.