I don't get the logic behind this reasoning, and I've seen it a few times on this forum. If you are working in a field that is related to what you want to do a PhD in, I don't think you are going to lose knowledge of relevant topics. In fact one could argue that, after being in the industry for a few year, you gain an understanding of where there is need for better research, and choose to do your thesis on that topic.
Also, I came across this interesting part-time phd at Imperial, designed for those who have had a few years of working experience in quant finance.
PhD in Mathematical Finance | Imperial College London
^Scroll toward the bottom of that link. Anyone familiar with this program, or have any thoughts on this?