Prediction of liver vitamin A in rats by an oral isotope dilution technique.

Male Sprague-Dawley rats (n = 41) that had liver vitamin A levels ranging from 5 nmol to 22 mumol (0.48 to 1629 nmol/g) were intubated with [3H]retinol in oil (167 kBq/rat). Plasma samples collected at 3, 4, 4.4, 5, 5.4 and 6 d were analyzed for radioactivity, and liver vitamin A mass was determined at 8 d. Nonlinear regression analysis was used to develop a biexponential equation describing the relationship at each sampling time between fraction of the dose remaining in plasma (FDp) and liver vitamin A. The best equation, developed with 3-d data, was liver vitamin A (nmol) = 58,577 x exp(-2715 x FDp) + 1810 x exp(-127 x FDp); Spearman's rank correlation coefficient = 0.982. This equation showed good predictive ability as evidenced by the ratio of the calculated to the observed liver vitamin A (1.23 +/- 0.65) and by its ability to distinguish animals with deficient, marginal, adequate or high vitamin A status. We conclude that this sensitive, noninvasive procedure can be used to predict rat liver vitamin A over a wide range.