Estimating diet digestibility: a comparison of 11 techniques across six different diets fed to rams.

Six digestion trials were conducted using eight Suffolk rams (four/trial; two trials run concurrently) to determine the ability of 11 techniques to estimate in vivo apparent DM digestibility (DMD). Diets (trials) were as follows: 1) ad libitum access to chopped fescue hay, 2) ad libitum access to alfalfa hay, 3) limit-fed fescue hay, 4) limit-fed alfalfa hay, 5) 25% soybean meal and 75% fescue hay and 6) 40% rolled corn and 60% alfalfa hay. Total feces were collected for 7 d following a 21-d adaptation to each diet. Digestibility methods evaluated were the following: 1) in vitro disappearance (INVITRO); 2) 48 h and 3) 72 h in situ nylon bag (48NB, 72NB); 4) 48NB followed by a 48-h acid pepsin digestion; 5) 96 h and 6) 144 h in vitro fermentation followed by NDF analysis (INDF96 and INDF144) 7) 96 h and 8) 144 h in vitro fermentation followed by ADF analysis (IADF96 and IADF144); 9) ADL ratio; 10) alkaline hydrogen peroxide treatment before ADL analysis (APFPRE) and 11) alkaline hydrogen peroxide treatment after ADF extraction of ADL (ADLPST). Apparent in vivo DMD was accurately estimated (P greater than .10) by INVITRO, ADL and APLPST with Diet 1. However no marker accurately estimated in vivo DMD when fescue was limit-fed. Apparent DMD of Diet 2 was accurately predicted (P greater than .10) by INVITRO, but when Diet 2 was limit-fed, only 48NB and 72NB predicted (P greater than .10) DMD accurately. In vivo DMD coefficients for Diet 5 followed a pattern similar to that of Diet 1. Diet 6 was predicted by INVITRO, 48NB, INDF96, INDF144 and IADF96 (P greater than .10). Results suggest that caution should be exercised when using these techniques, because no single technique provided accurate estimates across all diets and feeding conditions.