Print/True peaks

Hey everyone, kind of a weird question but here we go.

When I master, I like to print my tracks with the processing onto another track, and export/dither from that track (so I basically record in another track).

If I just directly render from my track with the processing, without printing it to another track and analyse it after, my true peaks are fine (if I put my limiter to -1dB true peaks, it’s going to be around that in the analysis). But if I take the track that I printed, same bit depth, same sample rate, same processing, the analysis says that my true peaks are 0,5 to 1db louder than they should be. Both tracks are at the same level, same processing, so I don’t understand why it’s doing this. Should I put my final limiter on the printed track?

Hope that it’s clear, thank you!!

Félix