Lactational performance, nutrient digestibility and fermentation characteristics of Holstein dairy cows in response to feeding wheat factory sewage

Abstract

The objective of this study was to evaluate whether wheat factory sewage (WFS) could partially replace barley grain in the diet of dairy cows without adversely affecting dry matter intake (DMI), ruminal fermentation, digestibility and milk production of dairy cows. Eight multiparous (60 ± 3 days-in-milk) Holstein dairy cows were used in a replicated 4 × 4 Latin square design experiment with four 21-d periods. The four diets (treatment) differed by partially substituting WFS for barley grain at the rate of 0% (WFS0), 4% (WFS4), 6% (WFS6) or 8% of dietary dry matter (DM; WFS8). DM content of diets decreased from 65%, 59%, and 57% to 54% by increasing the inclusion of WFS due to high water content of WFS (80%). DMI changed quadratically and tended (P = 0.08) to be higher for WFS4 (23.1 kg/d) than for other three diets which were similar (21.5 kg/d). Ruminal fluid pH linearly increased (P = 0.01) with increasing WFS in the diet. Ruminal concentrations of total volatile fatty acid (VFA; 100–103 mM), acetate (65–67 mM), propionate (24–25 mM) and ratio of acetate:propionate (2.68:2.89) were not affected by increasing the replacement of barley grain with WFS. Apparent total tract digestibilities of DM (67%), crude protein (CP; 68%) and neutral detergent fibre (NDF; 54%) were not different among treatments. Milk yield (averaged 40 kg/d) and milk composition were not affected by the dietary treatments. These results showed that the inclusion of WFS at rates of 0, 4, 6, and 8% in place of barley grain in dairy cow diets decreased DM content of diets, but had no adverse effects on DMI, rumen fermentation, digestibility and milk production responses. In conclusion, WFS can be a candidate by-product as an alternative feedstuff for the partial replacement of barley grain in lactating dairy cow diets and also considered as an alternative energy source in a cost-effective manner when the price is competitive.

Keywords:

• barley grain
• wheat by-product
• dairy cow

1. Introduction

Global demand-led fluctuations in the price of cereal grains (wheat, barley, corn, etc.) have generated growing interest in feeding home-produced by-products rich in energy and/or protein in many countries. Feeding the by-products in livestock production has recently received great attention by dairy and beef producers and nutritionists to reduce the dependence of livestock on grains that can be consumed by humans as well as to diminish environmental pollution. In the world, there are large available by-products derived from agri-food manufacture, such as wheat gluten meal, wheat factory sewage (WFS), pasta wastes, citrus pulp, rendered fats, etc., to dairy and beef cattle feeding. WFS, a by-product of starch and wheat gluten meal-producing factories consisted of [dry matter (DM) basis] high non-fibre carbohydrates (NFC; 61.9%) and sugar content (≥50%; mainly containing fructose and glucose) but very low starch (≤10%) and low neutral detergent fibre (NDF; 3%) with moderate crude protein (CP; 15%), seems to be a good source of rumen fermentable energy and protein source for ruminal microbes (Kamalian 2011).

Although starch and fibre are primary carbohydrates fed to dairy cows, sugars can be good alternative energy sources. The effects of sugars on ruminal fermentation and dairy productivity are of interest by dairy nutritionists (Oba 2011). Modern dairy cow-feeding systems in the world aim to maximise milk production through high starch diets, which can be led to sub-acute ruminal acidosis and a decrease in acetate:propionate ratio, fibre digestibility, milk fat content and dry matter intake (DMI; Plaizier et al. 2009). Partial replacement of the dietary starch with sucrose (0%, 2.5%, 5.0%, and 7.5%) was shown to linearly increase DMI and milk fat percentage and yield (Broderick et al. 2008). It has been recommended that 2.4% added sugar from molasses or totally 5% sugar as optimal point for lactating cows was based on DMI and yield of milk and milk components (Broderick & Radloff 2004). The effect of carbohydrate source on rumen pH is inconsistent. It was reported that rumen pH increased (Heldt et al. 1999) or tended to increase (Penner et al. 2009; Penner & Oba 2009) with the partial substitution of dietary starch sources with sugar. In contrast, rumen pH is not affected when dietary starch sources are partially replaced with sucrose (McCormick et al. 2001; Broderick et al. 2008). Hence, there is potential to alleviate ruminal acidosis by feeding non-starch sources of NFC.

Dietary profile of NFC has the potential to alter the supply of metabolisable nutrients to the animal as its fermentability varies with profiles. Balance of carbohydrates in the diet influences milk yield as it affects the amount and profiles of ruminal volatile fatty acid (VFA) produced, which in turn changes metabolism and partitioning of nutrients (Mertens 1992). Carbohydrates indirectly influence milk yield by changing microbial protein yield and amino acid supply (Mertens 1992; Hall & Herejk 2001). However, the effects of feeding sugar as part of NFC on rumen fermentation characteristics in vivo have been considerably variable; the molar proportion of butyrate in the rumen fluid increased (Kellogg & Owen 1969), did not change (Broderick et al. 2008; Penner & Oba 2009) or tended to be decreased (McCormick et al. 2001) by partially replacing grain with sucrose. Furthermore, it has been shown that when sugar partially replaced dietary starch, the molar proportion of propionate either decreased (Heldt et al. 1999) or did not change (Kellogg & Owen 1969; Vallimont et al. 2004).

The information on feeding WFS in dairy cows is limited. Hence, the objective of this study was to determine whether WFS can effectively replace part of grain in dairy cow diets by measuring ruminal fermentation characteristics, nutrient digestibility, milk yield, milk composition and net energy balance.

2. Materials and methods

2.1. Animals, experimental design and treatments

The experiment was conducted in Lavark, the Farm Animal Research and Teaching Unit of Isfahan University of Technology (IUT). Animals were cared for according to the guidelines of the Iranian Council of Animal Care (1995). WFS was provided by the Shahdineh-Aran Corporation (Jay Industrial Town, Isfahan, Iran). WFS is a by-product of starch and gluten-producing factories from wheat. It is produced at the final step of wheat manufacturing after precipitation of protein with salt. For this, ash content and salt of WFS were relatively high (14 of 20%). The rest of the ash composition is related to higher amounts of calcium, potassium, magnesium and sulphated components of used water (Kamalian 2011). Chemical composition of WFS is presented in Table 1. Eight multiparous Holstein cows (60 ± 3 days-in-milk; 616.1 ± 21.5 body weight, mean ± standard deviation) were used in a replicated 4 × 4 Latin square design during four 3-week periods to evaluate the effects of partial replacement of barley grain with WFS in lactating dairy cow diet. Barley grain was replaced with 0% (WFS0), 4% (WFS4), 6% (WFS6) and/or 8% WFS (WFS8). Salt was removed from the ingredient composition of WFS-supplemented diets due to its high salt content. Each experimental period consisted of a 15-d adaptation to new diet and 6 d for data collection. Cows within a square were randomly assigned to dietary treatments. Treatment sequences were balanced for carry over effects. Cows were housed individually in box stalls (4 × 4 m) that were equipped with concrete feed bunkers and automatic waterers. Clean sand was used for bedding and refreshed once daily. Cows were allowed to graze in an outdoor lot daily from 1200 to 1400 h. Diets were formulated to be isonitrogenous and isocaloric to meet or exceed the nutrient requirement for a 600-kg cow producing 45 kg/d of milk containing 3.0% protein and 3.5% fat and consuming 25 kg of DM using Cornell Net Carbohydrate and Protein System software (version 5.0) (Table 2). Alfalfa hay was crushed (4 mm mean particle size on the basis of Penn State Particle Separator box values) by a machine conventionally used for the separation of cereal grains from straw (Golchin Trasher Hay Co., Isfahan, Iran). The diets were prepared daily and fed as a total mixed ration (TMR) ad libitum. WFS was mixed into the concentrate fraction prior to adding the TMR. Cows were fed twice daily at 0730 and 1630 h, permitting 7.5–15% orts with free access to fresh water.

Table 1. Chemical composition of wheat factory sewage used in the experiment.

Chemical compositions
DM^a, %	20.0
CP^b, % of DM	15.0
NFC^c, % of DM	61.9
Starch, % of DM	10.0
Total sugar, % of DM	51.9
Glucose, % of sugar DM	18.4
Fructose, % of sugar DM	77.9
NDF^d, % of DM	3.0
ADF^e, % of DM	1.0
Ether-extract, % of DM	0.1
Ash	20.0
Salt (NaCl)	14.0

^a Dry matter.

^b Crude protein.

^c Non-fibrous carbohydrate.

^d Neutral detergent fibre.

^e Acid detergent fibre.

Table 2. Ingredients and chemical composition of experimental diets.^a

Ingredients, % DM	WFS0	WFS4	WFS6	WFS8
Alfalfa hay	22.0	22.0	22.0	22.0
Corn silage	22.0	22.0	22.0	22.0
Whole cottonseed	6.4	6.4	6.4	6.4
Ground barley	15.3	11.8	9.8	7.8
Ground corn	13.2	13.2	13.2	13.2
Fish meal	1.2	1.2	1.2	1.2
Soybean meal	12.6	12.6	12.6	12.6
Canola meal	3.0	3.0	3.0	3.0
Hydrogenated palm oil	1.4	1.4	1.4	1.4
Wheat factory sewage	–	4.0	6.0	8.0
Vitamin-mineral mix^b	0.92	0.92	0.92	0.92
Sodium bicarbonate	0.92	0.92	0.92	0.92
Calcium carbonate	0.24	0.24	0.24	0.24
Di-calcium phosphate	0.24	0.24	0.24	0.24
Salt	0.56	–	–	–
Chemical compositions
DM^c, %	65.3	60.0	57.2	54.4
CP^d, % of DM	18.9	19.1	19.1	19.2
NFC^e, % of DM	40.2	40.4	40.4	40.4
NSC^f^, ^g, % of DM	32.8	33.2	33.4	33.6
Starch^g, % of DM	27.6	26.0	25.1	24.3
Total sugar^g, % of DM	5.2	7.2	8.3	9.3
Glucose	0.83	1.30	1.49	1.67
Fructose	0.78	5.62	6.47	7.25
NDF^h, % of DM	30.1	29.6	29.3	29.0
ADF^i, % of DM	21.0	21.0	20.7	20.4
Calcium, % of DM	0.79	0.77	0.78	0.80
Phosphorous, % of DM	0.44	0.42	0.45	0.44
DCAD^j, meq/kg of DM	246	250	252	254

^a WFS0 = no added WFS + 15.3% barley grain; WFS4 = 4% WFS + 11.8% barley grain; WFS6 = 6% WFS + 9.8% barley grain; and WFS8 = 8% WFS + 7.8% barley grain. All diets contained a similar forage-to-concentrate ratio (44:56) on a DM basis.

^b Vitamin-mineral mix contained (DM basis) 250,000 IU/kg vitamin A; 50,000 IU/kg vitamin D₃; 1500 IU/kg vitamin E; 2.25 g/kg Mn; 120 g/kg Ca; 7.7 g/kg Zn; 20 g/kg P; 20.5 g/kg Mg; 186 g/kg Na; 1.25 g/kg Fe; 3 g/kg S; 1.25 g/kg Cu; 14 mg/kg Co; 56 mg/kg I; and 10 mg/kg Se.

^c Dry matter.

^d Crude protein.

^e Non-fibrous carbohydrate.

^f Non-structural carbohydrate.

^g Based on non-structural carbohydrate (NSC) determinations conducted by T.K.M. Webster of West Virginia University (Morgantown).

^h Neutral detergent fibre.

ⁱ Acid detergent fibre.

^j Dietary cation–anion difference was calculated according to NRC (2001).

2.2. Feed and faecal sampling

Feed offered and refusals were recorded daily for each cow to determine DMI. Feeds and refusals were sampled daily before the evening feeding through the data collection period. Collected diet and refusal samples for each cow were combined into three subsamples to represent 6 d for digestibility determination (d 16–21). The DM content of pooled diets and orts samples was determined by drying at 60°C in an air-forced oven for 48 h and DM results were adjusted to 100°C according to AOAC (2002). Feed and refusal samples were ground using a Wiley mill through a 1-mm screen (Arthur H. Thomas, Philadelphia, PA). Two faecal grab samples were taken from the rectum twice daily at 1000 and 1500 h across d 16–21 of each period and frozen at −20°C. The frozen samples were thawed at room temperature (24°C) and dried at 60°C for 72 h, ground and pooled for individual cows at each period. The TMR, refusals and faecal samples were analysed in triplicate for CP using Kjeldahl method (Kjeltec 1030 Auto Analyzer, Tecator, Höganäs, Sweden), calcium, phosphorous, ash (AOAC 2002), NDF using heat stable α-amylase (100 µl/0.5 g of sample), sodium sulphite and acid detergent fibre (ADF; Van Soest et al. 1991). The TMR composites were also analysed for soluble sugars using sucrose as the standard and for starch (Hall et al. 1999; T.K.M. Webster, West Virginia University, Morgantown). Furthermore, liquid samples of TMR's were filtered through 0.2 µm nylon filter vials (Alltech Associates Inc., Deerfield, IL), pipetted into 500 µl polyethylene high performance liquid chromatography (HPLC) vials (Alltech Associates Inc., Deerfield, IL) and kept refrigerated at 4°C until analysed. Liquid samples together with the calibration sugar standards were run on a Waters Alliance HPLC system (Model 2695, Waters Corporation, Milford, MA) employing an Aminex HPX-87P column (Bio-Rad Laboratories, Hercules, CA) and a refractive index detector (Waters 2414). Concentrations of monomeric glucose and fructose were calculated based on the calibration sugar standards. The non-fibrous carbohydrate component was calculated as 100 – (CP + NDF + ether extract (EE) + Ash). Acid detergent insoluble ash was used as an internal marker to determine the apparent total tract nutrient digestibility (Van Keulen & Young 1977).

2.3. Milking and milk sampling

Cows were milked three times daily at 0400, 1200 and 2000 h in a milking parlour. Milk yield was recorded using special graduated jars (Agri & SD Co., Frankfurt, Germany) and sampled in pre-labelled 50-ml plastic vials, at each milking during the last 6 d of each period. Before each milking, cows were monitored for udder inflammation and presence of milk clots in nipples to ensure that milk yield and composition were not affected by different forms of mastitis. Milk samples were composited in proportion to milk yield, preserved with potassium dichromate and stored at 4°C until analysed. Milk for the composite samples was warmed to aid in fat dispersion and inverted before subsampling. One aliquot of the composite sample was analysed for the concentration of fat, protein and lactose by Milk-o-Scan (134 BN Foss Electric, Hillerød, Denmark; AOAC 2002).

2.4. Rumen fluid sampling and VFA analyses

Rumen fluid samples were collected at least three times from each cow using the rumenocentesis technique from the ventral sac 4 h after morning feeding (Kowsar et al. 2008). The pH of rumen fluid was measured using a mobile pH metre (HI 8314 membrane pH metre, Hanna Instruments, Villafranca, Italy). The samples were preserved with 50% H₂SO₄ and frozen at −20°C for VFA analyses using gas chromatography (0.25 × 0.32, 0.3 µm i.d. fused silica capillary, model no. CP-9002 Vulcanusweg 259 a.m., Chrompack, Delft, the Netherlands) and for NH₃-N, as described by Broderick et al. (2008).

2.5. Calculations and statistical analyses

The NE_L intake was calculated from apparent total tract DM digestibility according to NRC (2001), with the modifications used by Penner and Oba (2009). Thus, the equation used to estimate dietary total digestible nutrient (TDN) was:

The TDN was then used to calculate dietary NE_L as described in NRC (2001). The energy required for maintenance was calculated as NE_M = 0.080 Mcal/kg of BW^0.75, and NE_L (Mcal/d) was calculated according to NRC (2001) with the observed milk yield and concentrations of milk fat, milk CP and milk lactose.

Data were composited within the period and subjected to Mixed Model procedure of SAS Institute (2003) according to the following model:

where Y_ijkl is the dependent variable, µ is overall mean, P_i is fixed effect of period, S_j is fixed effect of square, T_k is fixed effect of treatment, C (S)_k(l) is the random effect of cows within squares and e_ijkl is the residual. The REML method was used to estimate least square means and the containment method was used to calculate denominator degrees of freedom. The effect of increasing levels of WFS (0%, 4%, 6% and 8% of dietary DM) in the diet was examined through linear and quadratic orthogonal contrasts using the CONTRAST statement of SAS. Least square means are reported throughout. Differences between treatments were declared significant at P ≤ 0.05 unless otherwise noted (differences of P > 0.05 to P < 0.10 are discussed as trends).

3. Results and discussion

3.1. WFS and diet compositions, intake and digestibility

Ingredient and nutrient composition of the WFS and dietary treatments are presented in Tables 1 and 2, respectively. Moisture content of WFS was 80% and on the DM basis consisted of moderate level of CP (15%), a higher level of NFC (61.9%) and sugar (51.9%), with fructose and glucose as main constitutive monosaccharaides (77.9% and 18.4% of sugar DM, respectively). CP content was similar across diets averaging 19% of DM. Dietary sugar content and thereby glucose and fructose contents were increased with increasing amounts of WFS in the diets. Diets were formulated to be higher than NRC (2001) recommendations for CP to ensure that responses due to WFS supplementation (containing more soluble carbohydrate up to 52% of DM) would not be limited by ruminal NH₃. Sannes et al. (2002) suggested that beneficial responses to sugar as soluble carbohydrate might require increased dietary rumen degradable protein (RDP) levels to avoid a ruminal NH₃ limitation. Averaged across treatments, dietary NFC (40.4%), NDF (29.5%) and ADF (20.8%) contents were close to the high end of NRC (2001) minimum recommendations of 35–42%, 25–33% and 17–21%, respectively, for the maintenance of proper ruminal function and milk yield. Increased replacement of barley grain with WFS in diets decreased dietary DM content as a result of high moisture of wet WFS.

DMI was quadratically changed and tended (P = 0.08) to be higher for WFS4 (23.1 kg/d) than for other three dietary treatments which were similar (Table 3). The cows of the WFS4, WFS6 and WFS8 groups consumed 0.924, 1.314 and 1.688 kg DM of WFS per day or received 0.009, 0.064 and 0.116 kg NaCl more than the cows of the control group, respectively. Although the water intake of cows was not measured, greater intake of NaCl in cows fed WFS especially in cows receiving WFS8 diet might increase water intake and nutrient rate of passage and thereby mediate similar DMI among treatments. Intakes of other nutrients changed in a similar manner of DMI, reflecting nutrient content of diets. These results are somewhat in agreement with other reports (Cherney et al. 2003; Penner et al. 2009; Zhang et al. 2010), which substituted sugar or dried distillers’ grains with solubles (DDGS) for part of the concentrate or barley grain. These authors observed a linear increase in DMI with increasing dietary molasses or sugar in dairy diets, and they attributed the effect, in part, to improved palatability. The higher DMI with WFS4 might be due to improved palatability by adding WFS, whereas failure to increase DMI with increasing WFS feeding may have resulted from increased moisture content of the diet. Birkelo et al. (2004) reported that DMI was reduced by 11% with dietary moisture which decreased from 68% to 47% due to substitution of wet Distillers Grain (DG) for grain and soybean meal (SBM). Lahr et al. (1983) concluded that the reduction in DMI would be associated with diet DM contents below 60%. In the present study, DM content was below 60% for WFS6 and WFS8 diets.

Table 3. Intake and nutrient digestibility of dairy cows as influenced by partially replacing barley grain with wheat factory sewage.

	Treatment^a LSM^b					Contrast P-value
Item	WFS0	WFS4	WFS6	WFS8	SE^c	Linear	Quadratic
Intake, kg/d
DM^d	21.4	23.1	21.9	21.1	0.58	0.39	0.04
CP^e	4.0	4.4	4.1	4.0	0.11	0.80	0.03
NDF^f	6.5	6.9	6.4	6.1	0.17	0.08	0.06
ADF^g	4.5	4.8	4.5	4.3	0.12	0.09	0.02
NFC^h	8.5	9.3	8.9	8.6	0.23	0.98	0.02
Digestibility, %
DM	71.1	67.6	65.2	66.7	2.61	0.13	0.31
CP	68.7	70.7	69.9	65.9	3.45	0.52	0.38
NDF	56.8	54.3	50.8	54.0	2.14	0.30	0.30
ADF	60.2	56.9	53.4	52.6	2.91	0.01	0.62

^a WFS0 = no added WFS + 15.3% barley grain; WFS4 = 4% WFS + 11.8% barley grain; WFS6 = 6% WFS + 9.8% barley grain; and WFS8 = 8% WFS + 7.8% barley grain. All diets contained a similar forage-to-concentrate ratio (44:56) on a DM basis.

^b Least square means.

^c Standard error.

^d Dry matter.

^e Crude protein.

^f Neutral detergent fibre.

^g Acid detergent fibre.

^h Non-fibrous carbohydrate.

The increased substitution of WFS for barley grain did not affect the total tract digestibilities of DM, CP and NDF, but linearly decreased the digestibility of ADF, expressed either as percentage of ADF intake or kilogram of digested ADF (Table 3). Amounts of digestible DM and NDF linearly decreased with increasing WFS, which reflected linearly decreased intakes of DM and NDF. The present results are in agreement with Broderick and Radloff (2004), who observed a cubic response of nutrient digestibilities to feeding liquid molasses. Others (Broderick et al. 2008; Penner et al. 2009; Penner & Oba 2009) reported that feeding sugar as a replacement of dietary starch had no effect on nutrient digestibility; in contrast, supplementing dry molasses (Broderick & Radloff 2004) or DDGS (Zhang et al. 2010) increased digestibility. Decreased apparent total tract digestibility of ADF might be due to the increased rate of passage with increasing WFS in the diet (Sutoh et al. 1996). Greater digestibility of ADF than NDF implies that hemicellulose was less digestible than cellulose, possibly due to faster passage of hemicellulose from the rumen. Similarly, Broderick and Radloff (2004) and Broderick et al. (2008) have shown greater apparent digestibility of ADF than NDF when two-thirds (Broderick & Radloff 2004; Broderick et al. 2008) or all (Broderick et al. 2001) dietary forage was fed as alfalfa silage. Nevertheless, the digestibility of ADF was not higher when red clover silage was fed (Broderick et al. 2001). In the current experiment, alfalfa hay and corn silage were fed in equal proportion. It suggests that the digestibility of NDF and ADF vary with type of roughage.

3.2. Ruminal pH and fermentation characteristics

Ruminal fluid pH linearly increased with increasing inclusion of WFS in the diet (Table 4). Improved ruminal pH with increasing WFS inclusion is in agreement with some reports (Heldt et al. 1999; Penner & Oba 2009; Penner et al. 2009) that ruminal pH increases or tends to increase with increased amounts of sugar in diets. In contrast, ruminal pH was unchanged when liquid or dry molasses were fed (Broderick & Radloff 2004) or when sugar (McCormick et al. 2001; Broderick et al. 2008) was substituted partly for dietary starch. Discrepancy in the effects of feeding sugar on ruminal pH among experiments would be expected since other dietary factors, such as sugar source, amount of sugar in diets, may be involved. The increased ruminal pH with feeding sugar or feeds high in sugar, such as WFS, seems to contrast to the expectation as sugar is rapidly fermented in the rumen. Oba (2011) described several theories that may explain the phenomenon: sugars provide less carbon compared with starch for VFA production per unit of mass (Hall & Herejk 2001); disappearance of carbohydrates from the rumen may not necessarily result in acid fermentation production if organic matter (OM) is converted to microbial N (Allen 1997) or stored as glycogen (Hall & Weimer 2007). It is believed that sugar stimulates ruminal microbial protein synthesis. Ribeiro et al. (2005) showed in vitro that sucrose supplementation results in an increase in the efficiency of microbial N yield per kilogram of OM which coincides with our result on a linear decrease in NH₃–N of rumen. Results from this experiment suggested that the substitution of barley grain with WFS would be beneficial to alleviate the risk of ruminal acidosis, and therefore may be a valuable nutritional strategy in managing early lactation cow feeding. Increased ruminal fluid pH might be because of increased NaCl and also water intake and greater nutrient rate of passage with increasing levels of WFS.

Table 4. Rumen fluid pH and VFA concentrations of dairy cows as influenced by partially replacing barley grain with wheat factory sewage.

	Treatment^a LSM^b					Contrast P-value
Item	WFS0	WFS4	WFS6	WFS8	SE^c	Linear	Quadratic
Rumen pH	6.13	6.07	6.24	6.33	0.06	0.01	0.23
Total VFA, mM	101.3	100.4	103.4	102.4	3.95	0.38	0.65
Acetate, mM	65.0	65.2	66.8	67.4	1.96	0.62	0.65
Propionate, mM	24.0	23.6	24.9	23.3	1.77	0.39	0.43
Butyrate, mM	10.4	10.8	10.4	11.1	0.73	0.81	0.98
Isobutyrate, mM	0.08	0.07	0.04	0.08	0.01	0.31	0.04
Valerate, mM	0.21	0.18	0.17	0.19	0.03	0.52	0.44
Isovalerate, mM	0.15	0.14	0.13	0.16	0.02	0.88	0.23
Caproate, mM	0.16	0.07	0.08	0.10	0.02	0.25	0.10
Acetate:propionate	2.79	2.79	2.68	2.89	0.26	0.31	0.32
Rumen NH3–N, mg/dL	16.56	15.96	14.66	14.62	0.37	<0.001	0.43

^a WFS0 = no added WFS + 15.3% barley grain; WFS4 = 4% WFS + 11.8% barley grain; WFS6 = 6% WFS + 9.8% barley grain; and WFS8 = 8% WFS + 7.8% barley grain. All diets contained a similar forage-to-concentrate ratio (44:56) on a DM basis.

^b Least square means.

^c Standard error.

Ruminal concentrations of total VFA, acetate, propionate, butyrate and acetate:propionate ratio were not affected by the replacement of barley grain with WFS. The results are in agreement with some in vivo reports (Broderick & Radloff 2004; Penner et al. 2009) using dry or liquid molasses and sucrose as replacement for dietary starch. There are inconsistent results from in vitro studies on the effects of sugar on fermentating VFA. They reported that sugars increased butyrate production or increased the molar proportion of butyrate when replacing dietary corn starch (Vallimont et al. 2004) or hay (Ribeiro et al. 2005) with sucrose. When sugar partially replaced dietary starch, the molar proportion of propionate decreased (Heldt et al. 1999) or did not change (Kellogg & Owen 1969; Vallimont et al. 2004). Oba (2011) described that the inconsistent effects of sugars on the rumen VFA profile can be attributed to the types of sugars that were included in the diet, feedstuffs that were replaced by sugar and rumen microbial activity.

Concentrations of caproate tended (P = 0.10) to change quadratically to inclusion of WFS in the diet. However, valerate and isovalerate as the main branched-chain VFA were not affected. Unlike our results, ruminal branched-chain VFA was the highest with the diet containing 6% liquid molasses and observed a cubic response (Broderick & Radloff 2004). Broderick et al. (2008) and Penner et al. (2009) also observed a linear increase in molar proportion of isobutyrate with increasing amounts of sugar in diets. However, others reported that it was not affected by feeding sucrose in vivo (Kellogg & Owen 1969) or in vitro (Vallimont et al. 2004). In agreement with Broderick et al. (2008), ruminal ammonia concentration linearly decreased in our study as WFS increased in the diets.

3.3. Milk yield and composition

Least square means for milk yield, milk composition and production efficiency are presented in Table 5. Milk yield, 4% fat-corrected milk (FCM) and energy-corrected milk (ECM) were averaged 40.3, 34.2 and 36.6 kg/d, respectively, and were not affected by increasing WFS in the diet. Feed efficiency, expressed as milk yield/DMI, FCM/DMI, ECM/DMI and ECM/NE_L intake, was not affected by increasing WFS in the diet. The similar milk production and feed efficiency are not consistent with the linearly decreased digestible DM with increasing WFS feeding. The milk production measured during a short period in a Latin square design experiment may explain the inconsistent results. The lack of response to WFS is consistent with other authors when feeding molasses (Broderick & Radloff 2004), sugar (Cherney et al. 2003; Penner et al. 2009; Penner & Oba 2009) or DDGS (Zhang et al. 2010) to lactating dairy cows. The same were true for milk component composition and yield. In contrast, in an experiment performed by Murphy (1999), cows received incremental amount of molasses, produced greater milk and milk component. It was attributed to the indirect influence of molasses inclusion on DMI via increasing palatability and greater fermentability in the rumen. However, in our study, DMI was quadratically changed by treatment. The similar feed efficiency in the present study indicated that energy utilisation was not impaired even at the greater level of WFS inclusion in the diet as a replacement for barley grain. However, Broderick et al. (2008) observed a linear decrease in apparent DM efficiency when cows received either molasses or sugar in the diets. They attributed the effect in part to increased DMI with increasing amounts of sugar without changing milk yield.

Table 5. Milk yield, milk composition and feed efficiency of dairy cows as influenced by partially replacing barley grain with wheat factory sewage.

	Treatment^a LSM^b					Contrast P-value
Item	WFS0	WFS4	WFS6	WFS8	SE^c	Linear	Quadratic
Yield, kg/d
Actual milk	40.7	40.4	40.2	39.8	1.15	0.36	0.99
4% FCM^d	34.6	34.6	33.8	33.7	1.02	0.36	0.90
ECM^e	37.1	37.1	36.3	35.9	1.05	0.26	0.79
Fat	1.22	1.23	1.19	1.18	0.04	0.39	0.87
Protein	1.14	1.13	1.12	1.10	0.03	0.13	0.62
Lactose	2.18	2.20	2.20	2.14	0.06	0.46	0.29
Composition,%
Fat	3.09	3.08	2.91	2.95	0.09	0.14	0.76
Protein	2.80	2.79	2.81	2.82	0.03	0.52	0.79
Lactose	5.45	5.47	5.43	5.41	0.05	0.54	0.77
Feed efficiency
Milk yield/DMI^f	1.91	1.76	1.91	1.90	0.06	0.65	0.23
FCM/DMI	1.63	1.51	1.59	1.60	0.05	0.95	0.18
ECM/DMI	1.75	1.62	1.70	1.71	0.05	0.91	0.20
ECM/NEL intake	1.02	0.95	1.00	1.01	0.03	0.99	0.16

^a WFS0 = no added WFS + 15.3% barley grain; WFS4 = 4% WFS + 11.8% barley grain; WFS6 = 6% WFS + 9.8% barley grain; and WFS8 = 8% WFS + 7.8% barley grain. All diets contained a similar forage-to-concentrate ratio (44:56) on a DM basis.

^b Least square means.

^c Standard error.

^d FCM yield calculated as 0.399 × [milk yield (kg/d)] + 15.02 × [fat yield (kg/d)].

^e ECM yield calculated as (kg of milk × 0.3246) + (kg of milk fat × 12.96) + (kg of milk protein × 7.04).

^f Dry matter intake.

3.4. Body weight change and energy balance

Body weight change, energy output and energy balance were not different among the treatments, whereas NE_L intake was quadratically changed with the highest for WFS4 (Table 6). The quadratical response of NE_L is consistent with the quadratical change of DMI. The positive BW change and NE_L balance indicated that the energy requirement by dairy cows was met. The results are in agreement with some previous reports (Broderick & Radloff 2004; Broderick et al. 2008; Zhang et al. 2010).

Table 6. Body weight, body weight change and calculated energy intake, output, and balance of dairy cows as influenced by partially replacing barley grain with wheat factory sewage.

	Treatment^a LSM^b					Contrast P-value
Item	WFS0	WFS4	WFS6	WFS8	SE^c	Linear	Quadratic
BW^d, kg	619	610	609	625	19.04	0.86	0.55
BW change, kg/d	0.41	0.53	0.36	0.54	0.04	0.26	0.53
Dietary NEL, Mcal/kg	1.69	1.72	1.71	1.70	0.02	0.87	0.35
NEL intake, Mcal/d	35.9	39.4	36.9	35.7	0.99	0.44	0.03
NEM output, Mcal/d	10.0	9.8	9.9	10.0	0.24	0.87	0.58
NEL output, Mcal/d	25.4	26.1	25.7	24.8	0.76	0.45	0.18
Total NE output, Mcal/d	35.3	36.0	35.5	34.8	0.77	0.49	0.28
NE balance^e, Mcal/d	0.70	1.17	1.36	0.88	0.74	0.83	0.56

^a WFS0 = no added WFS + 15.3% barley grain; WFS4 = 4% WFS + 11.8% barley grain; WFS6 = 6% WFS + 9.8% barley grain; and WFS8 = 8% WFS + 7.8% barley grain. All diets contained a similar forage-to-concentrate ratio (44:56) on a DM basis.

^b Least square means.

^c Standard error.

^d Body weight.

^e NE balance calculated based on NRC (2001). Postpartum NEB = (DMI × NE_L diet) – [(0.08 × BW ^0.75) + (0.0929 × fat + 0.0563 × protein + 0.0395 × lactose) × milk yield)].

4. Conclusion

Overall, increased inclusion of WFS in place of barley grain at 0%, 4%, 6% and 8% (DM basis) in lactation cow diets quadratically changed DMI and intakes of other nutrients. Ruminal pH linearly improved but ruminal VFA concentrations, the digestibility in the total digestive tract, milk production and milk composition and energy balance were not affected with increasing WFS in dairy cow diets. Results indicated that at the level and method of supplemental WFS inclusion in the diets used in the current experiment, feeding WFS to Holstein dairy cows did not adversely impact feed utilisation and lactation performances. Linearly improved ruminal pH and increasing amounts of WFS in cow diet are of interest, suggesting that WFS could be used to alleviate ruminal acidosis. In general, the results of the current experiment indicated that WFS might be developed as an alternative energy source to grain to allow the dairy producers have more choices of feed ingredients or improve profits when the price is competitive. Further studies are needed to determine the long-term effects of feeding diets containing WFS to dairy cows.

Acknowledgements

Partial funding of the study by Shahdineh-Aran Co. and help from Mr Sarraf in grounding for suitable experimental conditions are appreciated. We also express our sincere appreciation to technicians Skandar Bagheri, Morteza Gholshadi and Mohammad Abbaseyan, the farm staffs at Lavark, the farm animal research and teaching unit (IUT, Isfahan, Iran), to Hamid Khoshouei (IUT, Isfahan, Iran), and to several students from the IUT completing course requirements for their input to the research and laboratory analyses.