The use of the in vitro filter bag method for predicting digestibility of forages *

The possibility of predicting the in vivo organic matter digestibility of forage by estimating in vitro digestibility using either rumen fluid or cellulase as incubation media in the filter bag method was studied. The in vivo OM digestibility of 35 forage samples was determined on 6 rams and in vitro digestibility was determined by two methods: rumen fluid (in a DaisyII Incubator)-neutral detergent (RF-ND) or cellulase solution (also in a Daisy incubator)-neutral detergent (C-ND). The coefficients of correlation between in vivo digestibility and that estimated by the RF-ND and C-ND methods were 0.62 and 0.69, respectively. Regression analysis shows that both tested methods have good potential for predicting in vivo OM digestibility.


INTRODUCTION
The filter bag method developed by Ankom (Ankom Co, Fairport, NY) is a good alternative to classical in vitro digestibility methods.Most studies have used buffered rumen fluid as the incubation medium (e.g., Mabjesh et al., 2000;Wilman and Adesogan, 2000), but this method has its limitations related to standardizing rumen fluid.Recently, the use of enzymes (cellulase and xylanase) instead of rumen fluid has been studied (Ludwin and Kowalski, 2004).The aim of the present study was to determine the possibility of predicting in vivo forage organic matter digestibility by in vitro digestibility estimated using either rumen fluid or cellulase solution as the incubation medium in the filter bag method.

MATERIAL AND METHODS
The in vivo organic matter digestibility of 35 forage samples from Denmark (10 maize silages, 5 whole crop barley silages, 10 green whole crop maize, 10 green whole crop barley) was determined on 6 rams, using the total collection method with a 7-day preliminary period and 7-day faecal collection.The chemical composition of feedstuffs was determined by standard methods (AOAC, 1995).The in vitro digestibility (IVD) was determined on the same 35 samples ground to pass through a 1 mm screen.The forages were dried in an air-forced oven for 48 h at 50°C.About 0.5 g of sample of feed was placed in a filter bag (Ankom F57;50 × 55 mm) made from polyester-polyethylene extruded filaments of a pore size about 25 µm.There were 4 bags for each feed.The bags were then placed in glass jars and incubated in a Daisy II Incubator (Ankom Co., Fairport, NY).The IVD was determined by two methods: incubation of the sample for 48 h at 39.5°C in rumen fluid (RF-ND) or in a cellulase solution (C-ND) in a Daisy II Incubator and then (in both methods) boiling the residue in neutral detergent for 1 h in an Ankom Fiber Analyser (for details see Ludwin and Kowalski, 2004).The results of two in vitro digestibility methods were subjected to one-way analysis of variance (SAS, 1996).Linear regression equations were calculated to estimate the precision of predicting in vivo digestibility when using both filter bag methods.

RESULTS
The variation in chemical composition of feedstuffs (Table 1) was sufficient to use this set of samples in the comparison.The average in vivo OM digestibility was 72.0%, and varied from 63.2 to 78.5%.The IVD of all forages estimated by the C-ND method was significantly higher than when estimated by RF-ND.When the set of samples was divided according to types of forages, the differences between in vitro methods were significant for maize and silages but they were not significant for barley and green forages (Table 2).Coefficients of correlation between in vivo OM digestibility and IVD estimated by the RF-ND and C-ND methods were 0.62 (P<0.01) and 0.69 (P<0.01),respectively.The linear regression equations for both methods are: y=0.672x + 26.33 (RF-ND; n=35; r 2 =0.39;SE=3.10) and y=0.6939x + 22.766 (C-ND; n=35; r 2 =0.48;SE=2.86), respectively.

DISCUSSION
This study showed that the filter bag method could be effectively used to predict in vivo digestibility of forages.It allows simultaneous incubation of many samples and different samples can be incubated at the same time in the same medium (Holden, 1999).The incubation medium, however, had a significant effect on IVD, with the C-ND method giving slightly higher digestibility coefficients.Regression 573 LUDWIN J. ET AL.

In vivo digestibility, %
In vivo digestibility, % Rumen fluid-neutral detergent method; digestibility, % Cellulase-neutral detergent method; digestibility, % analysis shows that both tested methods have a good potential for predicting in vivo OM digestibility.However, both the higher correlation coefficient and lower SE suggest that using cellulase instead of rumen fluid can be even more advantageous.Using enzymes also seems to be easier for standardization.

CONCLUSIONS
It can be concluded that the filter bag method is an easy and precise tool for studying in vivo digestibility.Enzymes have potential as an alternative medium to rumen fluid in this technique.
Figure 1.The relationship between in vivo OM digestibility and filter bag in vitro digestibility when measured using rumen fluid-neutral detergent (A) or cellulase-neutral detergent (B)

Table 2 .
Means of in vivo OM and in vitro digestibility (%) using two filter bag methods: rumen fluid-neutral detergent (RF-ND) and cellulase-neutral detergent (C-ND)