Guys,
Mirrors do not have problem with IR (CA), because they reflect IR (and UV) the same way they reflect visible light (reflection coefficient may be lower if appropriate coating is not used). Of course, resolution is lower for longer wavelengths, but the saturated image of the bright (and overexposed) stars obtained by sensor is actually way larger than the theoretical size of the Airy disk, partly also due to scattering inside the sensor.
However, lenses are different from mirrors, due to different refraction indexes for different wavelength. Therefore the focal length of refractor lens for IR may differ quite a lot from that for visible light. That is called CA and it is the main cause of star bloating, and this where the IR filter is sometimes a must.
IR emitted from stars is emitted from their photospheres just like the visible light (and UV, and X and gamma and so on) - so stars are point light sources for all practical purposes.
If IR is emitted from spread objects, then capturing that light is not the problem per se, you are capturing something that is really there.
Of course, if you are into NB, IR (or UV) filter may help to eliminate spurious pass bands of NB filters.
BTW, have a look at this discussion on CN on the subject - it may be interesting for some of you.
As for lenses (Canon 400mm f/2.8L in particular) and star bloating caused by excessive IR (CA in principle), have a
look at what I have done in terms of IR filter.
And, to reiterate one more time - my Newt (mirror system) and my C11 (SCT) do not show bloated stars even when used with modified camera (which is expected).