My simplified understanding is that you have two classes of 'things' that need to be considered for short circuit currents:
1)
passive things that carry current such as cables, bus bars, non-automatic switches, and other components that don't change state during the short circuit event.
2)
dynamic impedances such as circuit breakers, which actively respond and change state during the short circuit event.
Short circuit events can create huge magnetic/mechanical stresses, and components must survive these stresses. See, for example:
Continuing my simplified understanding:
Current limiting fuses operate so quickly that they limit the total I^2T that components will be subject to. For
passive components such as cables this reduces the total impulse that the components must tolerate. Thus for
passive components a current limit fuse can be treated as reducing the available fault current.
For
dynamic impedances the device is actively changing the circuit characteristics during the time that the fuse is supposed to be operating. This _might_ increase the clearing time of the fuse, or the opening contacts of the dynamic impedance might have a much shorter time constant for damage than passive devices. For these devices, the current limiting fuse cannot be treated as reducing the available fault current without actual testing.
This is my simplified understanding of the situation. I've never had to actually go through the engineering documentation to really understand what is going on with current limiting fuses; just casual reading and reading of discussions here.