Validation of the main modeling methods for the estimation of marker mean retention times in the different compartments of the gastrointestinal tract in sheep.

Four Texel wethers (60 to 64 kg) fitted with rumen and duodenal cannulae were used to compare methods for estimating mean retention times (MRT) of digesta markers. They were fed, in eight equal meals, 1,200 g of a mixture of a chopped and ground (8-mm screen) and pelleted orchardgrass hay in the ratios 90/10, 50/50, 30/70, or 10/90 according to a 4 x 4 Latin square design. Mean retention time in the stomach and in the whole tract were estimated by orally delivering pulse doses of 170Tm-labeled chopped hay, 169Yb-labeled ground hay, and [51Cr]EDTA, followed by duodenal spot-sampling and total collection of feces; calculations were done using an algebraic method. Mean retention time in the abomasum was estimated following administration of a pulse dose of the markers (169Yb-labeled duodenal particles and [51Cr]EDTA) in the abomasum via the ruminal cannula through the reticulo-omasal orifice and collecting duodenal samples. The reference value for MRT in the reticulorumen (MRTRR) was calculated by subtracting MRT in the abomasum from MRT in the stomach. For all markers, fecal excretion curves were fitted to a two-compartment age-independent model, a gamma-2 age dependent-age independent two-compartmental model, and a multicompartmental model. Comparisons were made among parameters of the models and their anatomical or physiological attributes when these were clearly defined or easily calculated. The mean retention time in the reticulorumen (MRTRR) obtained from the multicompartment model was similar to those calculated with the algebraic method whatever the sampling site and the marker used.