Heavy and wet: evaluating the validity and implications of assumptions made when measuring growth efficiency using 18O water