Theoretical ICM adjustments in early stages are generally pretty subtle. At 75% of the field, you might be slightly tighter opening because you do have a risk premium, although it's usually pretty small (like less than 2%).
The other thing you're supposed to from an ICM perspective is call less pre and do more 3-betting. You'll see a lot of ICM sims, even at 75% of the field remaining, where you'll be more inclined to pure 3-bet hands that you might otherwise flat. For example, look at these two ranges in GTOWizard - they're both basically 60BB CO vs. LJ open ranges, except one is an iCM sim and the other one is cEV:
The thing you'll notice is that you roughly 3-bet the same amount (7.3% at cEV, 8.2% in ICM), but you flat significantly less in an ICM sim. You'll see this in a lot of ICM sims, whether it be ICMIZER, HRC, etc.
The argument is that ICM doesn't take into account future game play, post-flop edge, etc. Because of that, I wouldn't say that making significant adjustments for ICM - especially deeper stacked - is a particularly good idea. But I do think that some subtle adjustments, like finding more 3-bets and doing a little less calling, definitely has merit.