I wrote a Python script, gluing together a bunch of open source tools, which converts DJVU files with OCR'd text and bookmarks like Tables of Contents, into PDFs with these same features. Everything else I could find just seemed to dump to flat PDFs, sometimes with enormous filesizes. If you've got a shell and the requisite tools installed, maybe you'll find it useful. Its great for textbooks!
https://github.com/kcroker/dpsprep
Let me know if it converts more detailed bookmarks within the document too (it should...)
Word.
-k