When we are dealing with baro-vnav, we are very aware of the cold temp corrections, and I have always wondered why this has not been an issue before.
Within the approach procedures, the FAA has a very slim margin for cold temperatures, 0.917, and this coupled with the ISA of the airport for the effective glidepath and the ROC, doesnt leave much room for colder temperatures. (it doesnt take temperatures a whole lot to get a 3 degree GPA to an effective 2.5 degree GPA)
Given all of the recent NPA constant descent conversions, it seems to me that the temp limitations should be shown on the plate...
Additionally, the different boxes handle to prox warn differently, especially important in the latency transition areas from initial to intermediate 1000' ROC to 500' ROC.
One trick we use in Alaska, is the 3 degree GPA is typical down to -20C, then the same procedure, with a 3.5 GPA is used from -45C to -21C....