For temperature dependent material properties being evaluated outside the ranges defined in the user-inputted data curves, the software will take the last two data points of the curve and linearly extrapolate the material properties at the required temperature values. This is discussed in KB solution S-056372, as well as S-035100 for a more specific property case.
I believe that the SPR you are referring to relates to Nonlinear studies specifically (SPR 647071). The workaround in such cases is to add additional data points to temperature dependent material properties curves to fully encompass the temperature range of the study.
Linear Static studies should not have a problem linearly extrapolating material data beyond the defined temperature ranges. That being said, I would generally recommend inputting material data that fully encompasses the temperature range in the study regardless, as this ensures what properties are being used and will also catch if the material properties would become suspect if linear extrapolation is assumed for the temperatures applied.