Kaygee - there are some theories floating around for the early failures. The most likely to me is the gear mesh is on tight side of manf. tolerences. and generates extra heat. The ptu is located next to rear cat, so it gets hot already, regardless of whats happening inside the unit. This added "gear induced" heat puts the ptu over some binary failure mode for the oil. (other forums have mentioned 340 degrees). The oil is damaged and turns to sludge and stops lubricating. It may spill out of vent also, some catch it, some don't. Personally, if this is true, its best to let it fail early (under warranty) and roll dice with a new unit that hopefully has looser gear mesh.
I work in the auto industry and we lubricate 30,000 parts a day. The equipment that dispenses grease and oil is so accurate, I cannot see many failures due to underfill. Not saying it isn't possible, just does not seem likely to me.
From very limited data points, it does seem that if you get a good ptu, it can last up to 150k with original oil and you don't have a thermal event (like driving in sand, towing up a mountain where cat is generating a ton of heat). For those that change the oil, it should be more I would think. I know I have 100k on my 11 now. I changed it at 60k and just now, at 100k. Oil was normal at both changes IMHO.
I've worked in auto for over 20 years. I've seen and heard a lot of crazy ****. Even with the best laid plans and ISO/TS/audit trails, **** still happens.
As far as gear mesh, I would consider that a defect if it wasn't correct. If there is a spec and the nominal is fine, but a part at min/max is subject to early and explosive failure, than there is a problem with the spec. Typically, parts of an assembly should be gauged/matched to ensure all component parts combine to a workable and in spec assembly. For example, with a bearing, your major component parts are an outer race, inner race, and balls - all three major components can't be at their min or max spec, so you may have an outer race at max, an inner race at min, and that is combined with balls at nominal to make an in spec assembly. Typically, testing is done on assemblies at nominal, min and max spec to ensure any and all parts coming off the line are good. If the component parts of the ptus aren't being gauged/matched or the specs aren't ideal, that would explain the total randomness of some being okay and others being ticking time bombs. As I said before, something like this, coupled with the known issues and revisions and the low fluid fill, there are a lot of things conspiring against it.
A far as fluid fill quantity, I can't say why some are underfilled. For all I know, maybe 12 oz was the spec at one point and it was changed, but somehow ptus were still filled with 12 oz after the change was made?
As far as other heat sources, like overloading from driving in sand (getting stuck), that would seem to indicate the ptu isn't sized for the application, likely due to packaging/cost concerns. As far as the excess heat from the cat, that too, would seem to be an issue with the fluid fill not being adequate to shed all the direct and indirect heat making it's way to the unit - again, a spec issue, likely due to packaging concerns. The only way I can see to combat these issues is to be cautious with situations like these and check/change the fluid even more frequently. Unfortunately, not all ptus have the drain plug and are as easy to service as changing engine oil.