In studies of glucose metabolism in neonates, a commonly used approach is based on constant rate infusion of glucose labelled with a stable isotope in order to reach an approximate steady state with regard to isotopic enrichment in plasma. Under presumed conditions of a glucose steady state, the rates of appearance and disappearance of glucose can, in principle, be calculated with use of a simple steady-state formula. However, in the neonate, steady-state conditions do not generally prevail and results derived on such assumptions are questionable. In the present study, we have taken a pragmatic approach and compared estimates obtained from the conventional formula with values calculated with Steele's nonsteady-state method. The results show that the estimated glucose appearance and disappearance rates change more or less over the observed time period in all the typical cases studied, and that "steady-state estimates" may differ from the corresponding non-steady-state values by up to 37%. In a sensitivity analysis, the value of the distribution volume factor was found to be non-critical, a circumstance that supports the use of Steele's method. Thus, even though the classical Steele pool-fraction method for computation of rates of appearance and disappearance under non-steady-state conditions has been criticized, it is still the most realistic alternative to the frequently used simple steady-state formula in applications to newborns.