On the macroscopic equivalence of descriptions of fluctuations for chemical reactions