Comparison of direct and indirect estimates of apparent total tract digestibility in swine with effort to reduce variation by pooling of multiple day fecal samples.

The intent of this study was to establish a fecal sampling procedure for the indicator method (IM) to provide digestibility values similar to those obtained by the total collection (TC) method. A total of 24 pigs (52.6 ± 1.5 kg) were fed 1 of 4 diets with a 2 × 2 factorial arrangement of virginiamycin and phytase (PHY) added to a corn-soybean meal diet with no inorganic P supplement. Pigs were housed in metabolism crates for a 5-d TC period after 7 d of adaptation. Immediately after the TC, a fecal collection period followed, using the IM by including 0.25% of Cr2O3 in the feed for 10 d. Fecal collection for the IM started the day after diets containing Cr2O3 were first fed, and continued for 9 consecutive days with a single grab sample per day. Similar portions of feces from d 5 to 9 were also composited into 4 samples to evaluate multi-day pooling combinations. Highly variable means and CV among samples for apparent total tract digestibility (ATTD) were observed at d 1 and 2 using the IM. The mean ATTD for DM, GE, and nutrients appeared to be stabilized by d 5 or 6 in all dietary treatments. The TC data seemed to have lower CV than the IM data for many components. Based on the linear broken-line analysis, fecal Cr concentration plateaued at d 3.75 (P < 0.001) after the first feeding of Cr. Mean ATTD values by the IM were lower than those by the TC method for DM (P < 0.05), GE (P < 0.01), P (P < 0.01), and Ca (P < 0.001). The PHY supplementation improved ATTD of P (P < 0.001) and Ca (P < 0.001) in both collection methods, whereas the PHY effect on ATTD of DM was observed only for the IM (P < 0.05). Differences related to PHY effect on ATTD were detected from d 4 to 9 in a single grab sample for P and DM but the ATTD of DM had inconsistent P-values by day. Fecal sampling after 4 d of initial feeding of marker always allowed detection of treatment effects on ATTD of P but not on ATTD of DM. Results indicated that the IM results in lower digestibility values than the TC method and does not provide the same treatment difference as the TC digestibility for energy and nutrients that are not highly impacted by the dietary treatment. For the IM, ATTD values and fecal Cr concentration stabilize at least on d 5 after initial feeding of diets containing Cr2O3. At least 2-d pooling of feces for the IM appears to be needed to provide greater accuracy and lower variations than a single grab sample.