This study aimed at evaluating the accuracy, precision and robustness of in vivo dry matter apparent digestibility estimates (DMAD), using as markers indigestible acid detergent fiber, indigestible neutral detergent fiber, acid detergent lignin, indigestible acid detergent lignin and chromic oxide in comparison to the fecal total collection. Eighteen wethers (56.5 ± 4.6 kg BW) were randomly assigned to diets containing 25, 50 or 75% of concentrate and Coast cross hay for 25 days. Feces were collected during the five days to determine apparent digestibility of dry matter and samples of feed and feces were incubated for 144 hours in bovine rumen to obtain the indigestible fractions. Chromic oxide (4.0 g/animal) was offered daily. Accuracy was evaluated by comparing mean bias of estimates (predicted dry matter apparent digestibility - observed dry matter apparent digestibility) among markers. Precision was assessed by the root mean square prediction error and the residual error; robustness was studied by the regression between bias and dry matter intake, diet concentrate level and animal body weight. Fecal recovery and accuracy of dry matter apparent digestibility estimates were higher for indigestible acid detergent fiber, followed by indigestible neutral detergent fiber, indigestible acid detergent lignin and chromic oxide, and at last for acid detergent lignin. Linear bias was significant only for indigestible acid detergent fiber, indigestible neutral detergent fiber and indigestible acid detergent lignin. By using chromic oxide it was possible to estimate dry matter apparent digestibility more precisely. All markers were robust regarding to variation of dry matter intake while only indigestible acid detergent lignin and chromic oxide were robust regarding to concentrate levels in the diet. Chromic oxide was not robust when animal body weight varied. In this experimental condition, indigestible acid detergent fiber is the most recommended marker to estimate dry matter apparent digestibility in ovine when the objective is to compare the results with the ones found in literature. On the other hand, chromic oxide is the most recommended marker when the objective is to compare treatments within the same experiment.
bias; fecal recovery; fiber; lignin
