A reasonably plausible model representing the interactions between car
bohydrate, protein and lipid metabolism in the fasted state is present
ed. Using computer solutions, we fit the model to literature data. The
data and model are insufficient for totally rigorous conclusions, but
they suggest tentatively that in 60 hr fasted man, a considerable fra
ction of the gluconeogenic flux may proceed (in the liver or kidney) d
irectly from Krebs cycle intermediates, not through pyruvate. Using th
e [U-C-14]glucose incorporation into protein found by Shipley et al. [
Am. J. Physiol. 213: 1149-1158, 1967], we illustrate the effect that t
he isotopic interactions betwen protein and carbohydrate turnover have
in making difficult the use of simple models, e.g. for the estimation
of gluconeogenesis. Ne also outline a possible approach for the estim
ation of the kidney contribution to gluconeogenesis.