Oculus implements a feature called Asynchronous Time Warp, where if a new rendered simulation state is not available when it's time to present a new frame to the user, the previous rendering is re-projected using the latest head tracking information, to minimize the apparent judder.
This is not quite the scenario you describe, but it is a situation in which VR experts have found it's worthwhile to display an old simulation state from a new perspective - breaking the 1:1 match between simulation time and presentation time for better perceived responsiveness.
A similar logic may well apply to your case: try it and see if it improves your game.
Note that it's actually very common for games to run their simulation at a different rate than the rendering, and interpolate the result to get smooth gameplay even on frames intermediate between simulation steps. So I don't see reason to doubt this tactic could also be useful in VR.