character conversion - horrible characters from MS Word

One of my applications is being heavily used by people who are cutting and pasting text from MS word into text areas within a form. I dutifully store what they send, and display it later using htmlentities and nl2br to sort out funny characters. Sadly, quite a few oddballs for which there are no html entity equivalents sneak through the system - for example the different width dashes at chr 150 and 151. Does anyone have a comprehensive system for catching such characters, or at least a suitable character mapping so I don't have to do the research to get an exhaustive list :) Cheers -- Phil Driscoll