I'm sure there are other explanations, but I came across this first while looking for something.
A PSG will also use oxygen levels for part of the formula to determine an AH event. A CPAP can't do that, so it has to estimate. This probably can lead to brand to brand differences. Also, a PSG scores only the AH events that occur after sleep begins, and a CPAP scores events that occur after the machine is turned on.
Automatic Event Detection (AED) from CPAPs was compared to manually scored events during a sleep study.
"The results of our study show that the AED tended to overestimate the AHI when the manually scored AHI was low and underestimate the AHI when the manually scored AHI was high."