I think you mean isn't very Muslim, not Arab. The Christians in Beirut are still Arabs.

I once had to convert a large number of pages generated by Word into something that was at least close to validating and I used Tidy HTML. It took a little bit of poking around with all the arguments to get it to do what I wanted, but once I had I just ran it on all the Word exports and it popped out clean code. It even had a special flag (though I don't remember it off the top of my head) to specifically deal with Word exports.