The abstract should be self-contained, so any abbreviations used therein should be explained at first occurence (if a term occurs only once, then there is no need to introduce an abbreviation), and citations should be given in its full form (i.e., [Name, Journal, Volume, Pages, Year]). Then starting from the Introduction, one introduces abbreviations at their first occurence again, even if they were introduced already in the abstract.
And while I somewhat understand the desire to keep the Conclusions self-contained, too, at times when I wanted to do so either the reviewer or the language editor insisted that there should be only abbreviations once they are introduced. Personally, I don't like starting a sentence from an acronym, or a number, but the editors don't have such dillemas; so I usually just write so that there are none at the beginning of a sentence.
I once argued about it with the reviewer, but the language editor then again said to stick to abbreviations. I guess there are more chances to get it one's way in case of a very long paper (>50, or even >100 pages) for the sake of clarity.