To my understanding, sample rate can be changed without destruction. Whenever you work in a DAW all files need to be of the same sample rate in order to play together. 48kHz is preferable to 44.1kHz as the highest audible frequencies can be better reproduced as per Nyquist Theorem.
Bit rate is different in that once audio is dropped from one rate to a lower one, e.g. 32 to 24 or 16, the conversion throws out data that is no longer in the resulting file. If you convert from 24 to 16 then from 16 to 24, the data discarded when going down to 16 no longer exists and cannot be brought back.
When converting to 16 bit, there is an article written elsewhere here (SRC then Dither or Dither then SRC?) by Paul Fawkes that discusses in detail how to dither the conversion so that the discarded data can be replaced with noise that accounts for it in the resulting file.