I would be very suprised to see that work. Even with UTF8 it will not work unless the doxypress web pages are for HTML 5. The escape sequence ± for the plus-or-minus symbol should work everywhere.
I'd like to confirm what encoding your input files are in, and what encoding your browser has detected for the HTML output. We strive to support UTF-8 on input and output however there may still be some code that is not unicode aware.
Andrew is correct that the ± HTML entity is more widely supported and will also work on all backends, even ones which do not support UTF-8 encoding.