To achieve file size reduction MP3 codecs discard audio information. While some actions, splitting a file, trimming a file, etc. can be done on some MP3 editors without reencoding the data anything that effects audio attributes necessitates re-encoding . . . that re-encoding alters, discards additional audio data
that is simply the nature of a lossy codec
the way around it is work with, edit with a full spectrum file (i.e. 'wav') save that and convert to MP3 only once for distribution. If additional edits are required you return to the 'wav' file and encode that to MP3, not MP3 to new MP3.
and generally speaking, with regard to 44.1 Vs 48, there is no reason to not record @ the sampling rate (e.g. 44.1) of the distribution product, though currently it does not, on a practical basis make all that much difference (to the end listener). I regularly record music @ 48k hz for video projects that is then converted to 44.1 for CD and vice versa. I typically make some effort to have as few conversions in final product as possible but don't go out of my way to track multiple takes at different sampling rates.