There can be no guarantee (or expectation) that these will remain in sync without using QLab 5 and Timecode, as the timebase for the audio track will be derived from the audio device clock and the video won't.
What I don't understand is that by saying MTC is off the cards and implying that the video can't send its soundtrack to the audio inputs of the installation, you seem to be saying that one of the machines is completely electrically isolated from the other. If this is the case how are you sending OSC between the machines? If copper or wireless transmissions can breach the barrier, then they can carry audio or timecode.
If however there is some strange force field which means any reasonable engineering solution is off the table, then you will need to unloop and follow on from the end of the audio to start cues for the audio and video so they are at least close to sync at the beginning if the cue. As you can't predict whether the video willl be behind or ahead of the audio at the end, you will have to set hard stop and restart, on the videocue on the remote machine when it receives a second trigger To try and avoid any missing video the best thing to do is to run the last frame of the video on a lower layer so that when the video stops and restarts it shows this content.