Hi
In general rendering volume, and also SSS will require you to increase the sampling about 4x to get an equivalent amount of noise sampling because of its volumetric based calculation.
This is a very much for all pathtracer unless they use some extra tricks.
So attenuation "could" introduce more noise because light is absent earlier and/or with less power. However it is not adding any !
If you are for a "spotlight" situation you could remove the second bounce which will trade noise for no GI, but works well with thin medium.
Thick mediums like clouds will benefit from some very dim ambient light plus a lower quality of sampling for the volume settings.
Cheers
Kai