Abstract 568 Poster Session II, Sunday, 5/2 (poster 201)

Background: Evaluating the power of metabolic studies in children requires estimating the variability introduced by diet, exercise and maturity. We are unaware of intra- and intersubject reproducibility data in children on energy intake (EI) and expenditure (EE), glucose production (GPR), gluconeogenesis (GNG), lipolysis, insulin sensitivity (Si) and glucose effectiveness (Sg) to be used in power calculations. The present studies were undertaken to determine the intrasubject variability regarding the above parameters when dietary and experimental conditions are controlled.

Subjects: Eight children (4 boys, 4 girls) (Age 8-16 y; 27-65 kg) were studied at two occasions separated by 2-8 weeks.

Methods: Pre-packed meals (60% CHO, 25% fat and 15% protein) were delivered daily for 7 days preceding each study and unconsumed food was returned and analyzed. EE was measured over 24 hrs in a room calorimeter. GPR, lipolysis, GNG, Si and Sg were determined using [13C] glucose, [2H5] glycerol, 2H2O, and a stable-labeled IVGTT, respectively.

Results: EI at home and in the room calorimeter was nearly identical on both study occasions and EE reflected both the caloric intake and the distribution of macronutrients. The average of the two studies were for [glucose] 4.8 mM; [insulin] 6.8 µU/mL; [C-peptide] 1.6 ng/mL; total glucose Ra 62.9 µmol/min, GNG 51% of glucose Ra; glycerol Ra 3.3 µmol/kg min; Si 6.6 10-4min-1/µU mL-1, and Sg 1.1 min-1. Intrasubject parameter variations between studies (% of mean±SD) were for [glucose] 3.5±2.4%; [Insulin] 14.5±11.5%; [C-peptide] 2.0±1.3%; total glucose Ra 8.7±4.0%; GNG 11.1±5.3%; glycerol Ra 29.7±21.7%; Si 20.9±16.7% and Sg 7.6±10.2%. Thus, using paired analyses (two tailed) with a power of 0.8 and a type 1 error of 0.05; and n<5 would be sufficient to detect a 10% difference for [glucose], [C-peptide], glucose Ra and GNG, and for [insulin] and Sg, 12 and 10 subjects, respectively, while for detection of a 20% difference, 11 and 8 subjects, respectively, would be sufficient.

Conclusions: Using a pre-pack meal strategy, controlled dietary and experimental conditions and a paired study design, glucose parameters can be measured with high reproducibility and the impact of various diet compositions can be evaluated with a small number of subjects, while to detect an equivalent relative difference in lipolysis and insulin sensitivity requires a larger number of subjects.