I had the same request and found the following solution:
Loading the pdf in PdfViewer a extracted each page as a single file and stored it in a temporaryy folder (I think a memory stream could olso be used).
the files ar stored in a dictionary with page number as key and filename as value.
Thencapturing the mouse position(starting in mouse move event), get the page number and searched for the single pages file name.
In the mouse up event I created a new pdf file )merging all single page files with the new order in a 'result pdf' and loaded this new created 'result pdf' in the PdfViewer.
Because the files are stored an undo and redo functionality is easy to implement.
Hope that idea helps.
Regards
Klaus.