Is the "pool-fraction" paradigm a valid model for assessment of in vivo turnover in non-steady state?