I’m surprised no one has yet mentioned this potential cause, but I had the same problem as you have, a couple of years ago. Kept losing a centered target from my STL-11000 FOV ... after any meridian flip. As per your comment, it took quite some time, and a lot of persistence, very slowly jogging the scope, to get the target into the frame again – it was way off, even after a perfect centering on the previous side of the pier.
That was until I realized, after a VERY long period of “denial”, that the entire problem was scope “Parallax Error”. After all, I told myself, I have an excellent top class AP mount, and an extremely expensive scope RC-14.5”, so it couldn’t possibly be due to a “mechanical problem”. I spent a large amount of time and effort using PemPro to get the polar alignment very near perfect, thinking PA was to blame, and yet I lost the CdC target immediately after any meridian flip, no matter which side of the pier I started my tests from. It was frustrating, even with a target star right at Dec = 0 and less than about 15 minutes east/west of the PM as the initial framing point for the tests. It gets worse at higher DEC’s toward Zenith.
Once I calculated an approximate required shim thickness and slipped it under the appropriate end of the OTA’s cradle, the problem was solved quite to my satisfaction, and indeed ultimate relief.
Definitely, read the section on Parallax correction in your AP mount’s User Guide. One of your saddle ring’s “riser” may need to be shimmed to nullify the (possible) Parallax Error.