Sure, you can demux/remux and add your audio, but you can cause more problems than solve. Your 'new' audio bitrate will be limited too depending on the video bitrate and the audio compression used in the first place (if any). AC3 and mp2 compress the hell out of audio...I always use PCM...it hogs the DVD bandwidth (1500+kbps), but at least it sounds good and I can control the video bitrate as well. Your new audio will have to be the same as what it already on the DVD, whether it be mp2, ac3, pcm, etc. Me thinks you will be forced to use mp2 unless you reauthor which will then recompress the video...much like going from mp3 > wav > mp3. That is why I recommended starting fresh from the DV tapes (if that is an option). Less headaches and you have 100% control of everything whether you want to maximize the audio or the video. Also, I have always been told to never use TMPGEnc to do audio (unless using PCM). Use it strictly for video.
Here is the lineage I use for video: DV > Vegas (Capture DV-AVI) > Vegas (sync new audio with old) > Remove Old Audio > Render to new DV-AVI (with remastered audio) > Print to tape for archive > TMPGEnc > Manually calculate video and audio bitrates for best results (I always use PCM audio) > Render to m2v and wav > Author.
Just make sure you are prepared and have plenty of HD space and LOTS of time. Rendering in TMPGEnc should take about 10 hours per hour of video using custom bitrate doing 2 passes. You will fill 100gigs before you know it with a 2 set show when working with DV-AVI's.
Good Luck. There are many different ways to do it. This is what I have found to produce the best results and also be the best for me. YMMV.