What I used to do before I discovered this forum, and still do for some books, is run the HTML through something called "Detagger" by Jafsoft, and then use Makedoc to make the Ereader PDB file from text. It is actually really good at keeping formatting though the HTML > text process. It supports so many things I haven't found 1/2 of them yet and have been a registered user for years.
http://www.jafsoft.com/detagger/