If your scope is relatively new, last 20 years or so, many have
an internal cal routine that tests the probe and stores a correction
data set for use in interpolation of scope results.
So no, not much need to do that unless you suspect there is a
real problem. Then based on cal costs might be just better to
retire probe and buy another.
Other possibility is if you need some form of metrology trace-ability
for the probe.....
There's not much that can change in a passive probe expect the value of the attenuation resistor so, unless you need calibration verification for a customer, I see no reason for an outside calibration.
In the older datasheets I have looked at I saw no mention of need to do this.
You can always go to the Agilent or Tektronix websites, for newer probes, and look at their
datasheets.
The passive elements in a probe, the resistor elements come to mind, I am sure could shift
over time. Pinched cable causing reflections, compromised ground return losses....But again
in newer scopes this is taken care of by a internal scope probe cal routine.