You might check that you are not losing coolant. It might be pooling in a cylinder and cause a misfire until it is burnt out.
This is the exact same thing I dealt with on the 3400 in the Montana after I did the LIM on it. Start it and it would miss like it had a dead hole for 10-15 seconds. Never caused a CEL for misfire, just a "EGR Stuck" code. After that it was smooth as new. Kinda thought that might have been it from the extra 'air' coming through the EGR. I had done the ignition with the LIM so that was all new. Compression was within tolerance. Actually, all the cylinders were really close to each other.
In the mean time I had the oil tested again and it had twice the coolant intrusion than the test that made me do the LIM. In half the miles.
btw, I never saw coolant in the oil either. Since I had been testing oil that is how I caught it.
Something else I noticed got me suspicious. The coolant
smelled like gas. It was pretty potent. It had also changed color from [censored] yellow to a kind of burnt orange color.
Finally I threw my hands up and bought a combustion gas detector setup that you put on the radiator. Popped positive for combustion leak almost instantly.
When I finally pulled the rear head, I seen what was going on. The gasket material had been eaten starting at a coolant port and once it got to the fire ring of the back passenger hole it rusted that out. Over a period of time after shutoff due to the pressure in the coolant system it was seeping into the cylinder and pooling. Causing a misfire at initial startup. Still dont know how it passed a compression test so well. That whole setup made me question the validity of compression test results as manner of grading an engine. They only seem to show obvious defects as a sort of "Go/No Go" system. More like "Maybe Go/No Go."
Didnt mean to write a book, or to scare you unnecessarily. Just wanted to explain my experience.
Sorry.