Great question. I use the 5.1 panners in Pro Tools for every project, whether I'm designing backgrounds or FX. It's really the only way to go for maximum flexibility and efficiency.
To address your question specifically, a good approach is to design the track as if you are not only assembling the backgrounds but also predubbing them, taking responsibility for everything from leveling to panning, EQ (mostly corrective) and even reverbs (virtual or rendered). When you're done, you have a finely honed bed of backgrounds that are ready to be incorporated into the final mix.
This technique becomes really powerful when you take control of your layers and pan them thoughtfully. Example: A nighttime forest scene may have:
- 3 stereo tracks of air; 1 pair LR, 1
pair LsRs, 1 pair somewhere in the
- 2 stereo tracks of wind; 1 pair LR, 1
- 3 stereo tracks of insects; 1 pair
L+Ls, 1 pair R+Rs, 1 pair C+Rear C
- 5 mono tracks of spotted BGs (owls,
nightbirds, cicada, etc); discretely
panned wherever you want them
When grouped together in "food groups" and bussed out to the appropriate auxes, you've just created your first virtual predub. This method, by the way, is fast becoming the norm in feature post-sound here in LA and has been successful for many years now.
Regarding issues with LtRt fold downs, I'd just be aware of any potential phase problems in the recordings you are using. Other than that, you should be good to go.