Estimation of glucose kinetics in fetal-maternal studies: Potential errors, solutions, and limitations

We investigated whether errors occur in the estimation of ovine maternal-fetal glucose (Glc) kinetics using the isotope dilution technique when the Glc pool is rapidly expanded by exogenous (protocol A) or endogenous (protocol C) Glc entry and sought possible solutions (protocol B). In protocol A (n = 8), after attaining steady-state Glc specific activity (SA) by [U-¹⁴C]glucose (period 1), infusion of Glc (period 2) predictably decreased Glc SA, whereas [U-¹⁴C]glucose concentration unexpectedly rose from 7,208 ± 367 (means ± SE) in period 1 to 8,558 ± 308 disintegrations/min (dpm) per ml in period 2 (P < 0.01). Fetal endogenous Glc production (EGP) was negligible during period 1 (0.44 ± 1.0), but yielded a physiologically impossible 'negative' value of -2.1 ± 0.72 mg · kg^-1 · min^-1 during period 2. When the fall in Glc SA during Glc infusion was prevented by addition of [U-¹⁴C]glucose admixed with the exogenous Glc (protocol B; n = 7), EGP was no longer 'negative'. In protocol C (n = 6), sequential infusions of four increasing doses of epinephrine serially decreased SA, whereas tracer Glc increased from 7,483 ± 608 to 11,525 ± 992 dpm/ml plasma (P < 0.05), imposing an obligatory underestimation of EGP. Thus a tracer 'mixing' problem leads to erroneous estimations of fetal Glc utilization and Glc production via the three-compartment model in sheep when the Glc pool is expanded exogenously or endogenously. These errors can be minimized by maintaining the Glc SA relatively constant.