In my opinion, people are not giving the OP enough credit for already having a PhD in a related field. This means s/he has a better grasp of what it means to get a PhD than 99% of the general population, has already shown considerable technical skill, the drive to complete a task which requires years of hard work, and so forth.
So far as I know, by far the most common reason for someone to drop out of a doctoral program in math (or any discipline) is that they just didn't really know what they were getting into: they had no way of properly gauging the scope and amount of work involved.
I am currently involved in graduate admissions at UGA. I have only been doing this for a little while, so my ideas may change, but at the moment if I saw a candidate who had a decent undergraduate background, test scores within the normal range of our successful applicants (e.g., greater than 50% on the math subject GRE) and already had a PhD, I would be tempted to put them towards the top of the list. Thinking it through as a hypothetical admissions decision, my only major concern would be that the candidate's undergraduate background may not be as fresh or relevant as that of our other strong applicants. Thus I would recommend taking a few math classes at the advanced undergraduate level. With good performance on those, I see no reason not to count the candidate's previous experience as an advantage, rather than his/her age as a disadvantage.
