Modification of a rumen fluid priming technique for measuring in vitro neutral detergent fiber digestibility

Recently, we developed an alternate method to measure in vitro neutral detergent fiber (NDF) digestibility (ivNDFD) based on a primed rumen fluid inoculum. Pretreating rumen fluid inoculum with cellulose and holding the inoculum until it generated 0.3 mL of gas/mL of rumen fluid before inoculating forage samples improved ivNDFD assay repeatability but depressed ivNDFD means. Our objective in this study was to determine if pretreating rumen fluid with a mixture of carbohydrates and urea would affect the ivNDFD mean and variance. We also used the modified procedure as a reference assay to calibrate near-infrared reflectance spectroscopy (NIRS) to predict 24-, 30-, and 48-h ivNDFD. Two experiments were completed. In experiment A, 3 ivNDFD assays modified from the method of Goering and Van Soest were evaluated over 24, 28, 48, 54, and 72 h by using dried, ground alfalfa (1 mm) or wheat straw (0.5 g) sealed in Ankom F57 forage fiber bags. Bags were placed individually in 125-mL Erlenmeyer flasks and incubated with Goering and Van Soest media and 10 mL of rumen fluid. Rumen fluid was collected before feeding from 2 cannulated cows fed a high-forage diet and was prepared in 1 of 3 ways: 1) pooled rumen fluid was strained and used immediately to inoculate flasks (modified Goering and Van Soest method); 2) strained, pooled fluid was combined with buffer, reducing solution, and 1.25 mg of primer/mL of rumen fluid and allowed to produce 0.12 mL of gas/mL of rumen fluid before sample inoculation [Combs-Goeser (CG) method]; or 3) the CG method was used without the primer mixture (unprimed method). The assay was repeated 5 times, with 5 time points ( 24, 28, 48, 54, and 72 h) and 2 subsamples per time point for each method. Neutral detergent fiber was analyzed using an Ankom(200) forage fiber analyzer and ivNDFD was determined as follows: ivNDFD (% of NDF) = 100 x [(NDF0h - NDFresidue)/(NDF0h)]. Results were analyzed using a mixed model procedure, and data were blocked by method to obtain repetition sums of squares, which were compared by an F-test to assess interassay error. Repetition sums of squares were reduced with the CG method compared with the Goering and Van Soest method (19 vs. 228), and mean ivNDFD estimates were similar at 28, 48, and 54 h. In experiment B, 24-, 30-, and 48-h ivNDFD data for 54 feeds were determined in triplicate using the CG method, and corresponding samples were then scanned with an NIRS instrument. Calibrations were computed using partial least squares regression techniques. The NIRS calibration equation R-2 values were 0.93, 0.93, and 0.89 for 24-, 30-, and 48-h ivNDFD. Results suggest that the modified ivNDFD method using rumen fluid primed with a mixture of carbohydrate and urea (CG method) reduced interassay error.