Processing index of barley grain and dietary undigested neutral detergent fiber concentration affected chewing behavior, ruminal pH, and total tract nutrient digestibility of heifers fed a high-grain diet

Abstract

The objective of this study was to investigate the effects of processing index (PI) of barley grain and dietary undigested neutral detergent fiber (uNDF) concentration on dry matter (DM) intake, chewing activity, ruminal pH and fermentation characteristics, total tract digestibility, gastrointestinal barrier function, and blood metabolites of finishing beef heifers. The PI was measured as the density after processing expressed as a percentage of the density before processing, and a smaller PI equates to a more extensively processed. Six ruminally cannulated heifers (average body weight, 715 ± 29 kg) were used in a 6 × 6 Latin square design with three PI (65%, 75%, and 85%) × 2 uNDF concentration (low and high; 4.6% vs. 5.6% of DM) factorial arrangement. The heifers were fed ad libitum a total mixed ration consisting of 10% barley silage (low uNDF), or 5% silage and 5% straw (high uNDF), 87% dry-rolled barley grain, and 3% mineral and vitamin supplements. Interactions (P < 0.01) of PI × uNDF were observed for DM intake, ruminating and total chewing time, and DM digestibility in the total digestive tract. Intake of DM, organic matter (OM), starch, and crude protein (CP) did not differ (P > 0.14) between low and high uNDF diets, but intakes of NDF and acid detergent fiber were greater (P = 0.01) for high uNDF diets regardless of barley PI. Heifers fed high uNDF diets had longer (P = 0.05) eating times (min/d or min/kg DM) and tended (P = 0.10) to have longer total chewing times (min/kg DM) than those fed low uNDF diets. Additionally, heifers sorted (P = 0.01) against long particles (>19 mm) for high uNDF diets but not for low uNDF diets. Altering PI of barley grain did not affect (P > 0.12) total volatile fatty acid (VFA) concentration, molar percentages of individual VFA, or duration of ruminal pH < 5.8 and <5.6. Total VFA concentration was less (P = 0.01), acetate percentage was greater (P = 0.01), and duration of ruminal pH < 5.8 and <5.6 was less (P = 0.05) for high compared with low uNDF diets. Digestibility of DM, OM, and CP was greater (P = 0.02) for low vs. high uNDF diets with PI of 65% and 75%, with no difference between low and high uNDF diets at PI of 85%. Blood metabolites and gastrointestinal tract barrier function were not affected (P ≥ 0.10) by the treatments. These results suggest that increasing dietary uNDF concentration is an effective strategy to improve ruminal pH status in finishing cattle, regardless of the extent of grain processing, whereas manipulating the extent of barley processing did not reduce the risk of ruminal acidosis.

Introduction

Finishing beef cattle in North America are typically fed high-concentrate diets (>90% concentrate, dry matter [DM] basis) to achieve maximum productivity in a cost-effective manner and minimize marketing time (Koenig and Beauchemin, 2011). A limitation of feeding high-concentrate diets with low concentrations of fiber and large amounts of rapidly fermentable carbohydrates is the increased risk of ruminal and hind-gut acidosis and liver abscesses (Humer et al., 2018). The role of fiber in high-grain feedlot diets is primarily to maintain rumen health by stimulating chewing and rumen motility, which helps to regulate ruminal pH and maintain a stable ruminal environment (Allen, 1997). Providing a sufficient but not excessive quantity of roughage can help to avoid digestive disorders without having negative effects on growth performance and feed efficiency.

Many of the previous studies have evaluated the effect of increasing the forage to concentrate ratio of finishing diets and report negative effects on feed intake, average daily gain (ADG), and feed efficiency (Galyean and Defoor, 2003; Koenig and Beauchemin, 2011). Swanson et al. (2017) reported that growth performance decreased linearly as the inclusion of a mixture of bromegrass hay and corn silage increased from 5% to 20% of DM. Likewise, Koenig et al. (2020) reported that increasing the barley silage concentration from 0%, 3%, 6%, 9%, and 12% of dietary DM linearly reduced gain:feed ratio of finishing steers. Currently, there are no guidelines for the minimal forage requirement of beef cattle because the effect of forage inclusion level can vary with forage source (Swanson et al., 2017). Additionally, the minimum amount of forage required may also depend on the extent of grain processing (Koenig and Beauchemin, 2011). It was suggested that less roughage may be required in diets containing less extensively processed grain due to reduced risk of ruminal acidosis (Koenig and Beauchemin, 2011). Numerous studies have proven that extensive processing not only maximizes grain digestibility but also increases the chances of digestive and health upsets, including bloat, acidosis, laminitis, and liver abscesses (Beauchemin et al., 2001; Yang et al., 2014).

Physically effective neutral detergent fiber (peNDF), which is a combination of chemical fiber concentration and particle size of feedstuffs (Mertens, 1997), has been intensively studied in dairy cattle to assess the adequacy of dietary fiber (Zebeli et al., 2012). Although peNDF is a good predictor of chewing activity (Yang and Beauchemin, 2007), the contribution of chewing toward the regulation of ruminal pH has been over-emphasized (Yang and Beauchemin, 2009; Zebeli et al., 2012). In fact, Allen (2000) estimated that salivary bicarbonate only removed approximately 30% of the total acid produced in the rumen, and Dijkstra et al. (2012) calculated that with high-forage or high-grain diets, ruminal epithelial bicarbonate secretion exceeds that provided by saliva. Thus, strategies to improve rumen health should increase chewing time, improve rumen motility and mixing of rumen contents, and stimulate volatile fatty acid (VFA) absorption (Storm and Kristensen, 2010), thereby providing bicarbonate to the rumen contents to buffer the rumen environment.

Undigestible NDF (uNDF) has been identified as an indicator of the physical filling effect of feed, and its use in dairy cattle diet formulation has improved the prediction of feed intake (Harper and McNeill, 2015; Fustini et al., 2017). It has also been reported that low peNDF concentration of dairy diets may be at least partially compensated for by increasing the uNDF concentration of the diet (Grant et al., 2020). We are not aware of any research evaluating the uNDF concentration in diets for beef cattle, but it is possible that increasing the uNDF concentration of finishing diets may promote ruminal motility and VFA absorption, thereby improving the regulation of ruminal pH.

We hypothesized that increasing dietary uNDF concentration would increase ruminal pH and decrease the incidence of ruminal acidosis of cattle fed high-grain diets, and that the effect of dietary uNDF concentration would depend upon the ruminal fermentability of carbohydrates (i.e., achieved through the extent of grain processing). Therefore, the objective was to investigate the effects of dietary uNDF concentrations in diets that varied in the extent of barley grain processing on feed intake, chewing activity, ruminal pH and fermentation parameters, total tract digestibility, gastrointestinal barrier function, and blood metabolites of beef cattle.

Materials and Methods

The study was reviewed and received the approval of the Institutional Animal Care and Use Committee at the Lethbridge Research and Development Centre, and all animals were cared for in accordance with the guidelines of the Canadian Council on Animal Care (2009).

Measure of uNDF of feed ingredients

The concentration of uNDF of barley straw, silage, and grain was determined in situ using three ruminally cannulated beef heifers fed a high forage diet as described by Beauchemin et al. (2019). The barley straw and silage were ground to pass a 2-mm sieve, and the barley grain was ground to pass a 4-mm sieve using a Wiley mill (standard model 4, Arthur H. Thomas Co., Philadelphia, PA). The ground substrates were sieved through a 53-μm sieve to remove the fine particles that could potentially wash out from the bags immediately upon ruminal incubation. All substrates were weighed into 10 × 20 cm polyester bags (R1020, ANKOM Technology, Macedon, NY; 50-μm porosity) with approximately 10 ± 0.05 g of barley straw and silage DM, and 20 ± 0.05 g of barley grain DM per bag. The amount of grain in each bag was doubled to ensure a sufficient amount of residue to enable the measurement of the NDF at the end of incubation, given the low NDF concentration and high fermentability of barley starch. The tops of the bags were folded inside the bag and sealed (Impulse heat sealer, ANKOM Technology, Macedon NY). For each heifer, triplicate bags of each substrate were placed inside large mesh bags (30 × 30 cm) and incubated in the rumen for 240 h. Upon removal from the rumen, the bags were immediately submerged in ice water to stop the microbial activity and washed under cold running water until the water ran clear. After gently squeezing to remove excess water, the bags were dried in a forced-air oven at 55 °C for 72 h. The residues in each bag were ground to pass through a 1-mm sieve and analyzed for NDF in a fiber analyzer (F57 Fiber Filter Bags, 200 Fiber Analyzer, ANKOM Technology, Macedon, NY) using sodium sulfite and heat-stable amylase. The uNDF concentration in feed ingredients was the amount of NDF remaining in the residue after incubation, expressed as percentage of ingredient DM before incubation. The uNDF concentration of the diet was then estimated according to the respective proportion of each ingredient in the diet and its uNDF concentration. The samples from the three animals were measured independently and averaged to obtain the uNDF concentration for each feed ingredient. The potentially digestible NDF (pdNDF) was estimated as the difference between total NDF and uNDF concentrations. The measured uNDF concentration was 27.4%, 10.6%, and 4.2% of DM, respectively, for barley straw, silage, and grain (Table 1).

Table 1.

Chemical composition and in situ rumen disappearance of barley silage, barley straw, and barley grain

Item1	Barley silage	Barley straw	Barley grain
DM, % as-fed	38.5 ± 0.65	91.0 ± 1.31	90.8 ± 1.10
OM, % of DM	92.2 ± 0.56	92.4 ± 0.32	97.8 ± 0.28
NDF, % of DM	41.0 ± 0.74	78.0 ± 1.45	14.7 ± 0.75
ADF, % of DM	25.5 ± 1.28	49.2 ± 1.15	4.9 ± 0.41
CP, % of DM	11.3 ± 0.84	4.9 ± 0.37	13.5 ± 0.70
Starch, % of DM	22.6 ± 1.48	1.4 ± 0.29	58.8 ± 1.43
dry matter digestibility (DMD)240, %	85.6 ± 0.80	65.4 ± 0.45	95.0 ± 0.28
uNDF2, % of DM	10.6 ± 0.13	27.4 ± 0.85	4.2 ± 0.11
pdNDF3, % of DM	30.4 ± 0.71	50.6 ± 1.12	10.5 ± 0.72
pef84	0.59 ± 0.021	0.61 ± 0.056	–
peNDF85, %	24.19 ± 2.19	47.58 ± 5.23	–

¹Measured based on six samples collected, respectively, from six experimental periods.

²uNDF, undigestible NDF, measured in situ after 240 h of incubation.

³pdNDF, potentially digestible NDF = NDF – uNDF.

⁴pef₈, physical effectiveness factor determined as the sum of particles retained on 19- and 8-mm sieves.

⁵peNDF₈, physically effective NDF estimated as the NDF concentration of forage multiplied by its pef₈.

Animals, treatments, and experimental design

Six ruminally cannulated beef heifers (initial body weight [BW] ± SD = 715 ± 29 kg) were randomly assigned to a 6 × 6 Latin square with a 3 × 2 factorial arrangement of treatments. Each experimental period was 24 d, including 14 d for adaptation to new diets and 10 d for sampling and data collection. The main factors were three extents of barley grain processing based on processing index (65%, 75%, or 85% of PI), and two uNDF concentrations (low and high; averaging 4.6% and 5.6% of dietary DM). Barley grain was dry rolled, and PI was measured as the density (or volume weight, g/L) after processing expressed as a percentage of the density before processing (Beauchemin et al., 2001). As grain is rolled more finely, its density decreases, thus a smaller PI equates to a more extensively rolled cereal grain. The low and high uNDF concentrations of diets corresponded, respectively, to barley silage as the sole forage and a 50:50 combination of barley silage and barley straw (DM basis, Table 2). The barley straw was chopped to obtain a particle size distribution similar to that of barley silage so that the effect of uNDF would not be confounded with that of particle size. The diets were formulated with 10% roughage, 87% barley grain, and 3% vitamin and mineral supplements (DM basis) to meet or exceed the nutrient requirements of beef cattle (NASEM, 2016), with no monensin, tylosin, or other in-feed antimicrobials.

Table 2.

Ingredients and chemical composition of experimental diets

	PI, %
	65		75		85
Items	Low1	High1	Low	High	Low	High
Ingredients, % of DM
Barley silage	10	5	10	5	10	5
Barley straw	—	5	—	5	—	5
Barley grain	87	87	87	87	87	87
Supplement2	3	3	3	3	3	3
Chemical composition3, % of DM
DM, %	74.6 ± 1.64	79.9 ± 1.22	75.4 ± 1.54	79.8 ± 1.21	75.3 ± 1.43	79.9 ± 1.13
OM	95.5 ± 0.96	96.0 ± 0.60	95.9 ± 0.58	96.2 ± 0.38	95.8 ± 0.67	96.1 ± 0.33
NDF	17.0 ± 0.69	18.9 ± 0.67	17.2 ± 0.73	19.1 ± 0.78	16.5 ± 0.46	18.4 ± 0.41
uNDF	4.6 ± 0.51	5.6 ± 0.54	4.6 ± 0.45	5.6 ± 0.48	4.7 ± 0.53	5.6 ± 0.56
ADF	6.8 ± 0.53	8.0 ± 0.60	6.7 ± 0.27	7.9 ± 0.22	6.8 ± 0.37	8.0 ± 0.36
Starch	53.2 ± 1.68	52.2 ± 1.66	53.3 ± 1.05	52.2 ± 1.02	53.8 ± 1.76	52.8 ± 1.72
CP	13.6 ± 0.53	13.2 ± 0.54	13.3 ± 0.55	12.9 ± 0.57	13.5 ± 0.64	13.2 ± 0.64

¹Low, low concentration of uNDF diet; High, high concentration of uNDF diet.

²Composition: 54.67% ground barley, 9.67% canola meal, 24.33% calcium carbonate, 2.33% molasses, 5% salt, 2% urea, 0.07% vitamin E (500,000 IU/kg) and premix, which supplied: 15 mg Cu, 65 mg Zn, 28 mg Mn, 0.7 mg I, 0.2 mg Co, 0.3 mg Se, 6,000 IU vitamin A, 600 IU vitamin D, and 47 IU vitamin E per kilogram of dietary DM.

³Measured based on six samples of each diet collected by experimental period.

The diets were prepared daily using a feed mixer (Data Ranger, American Calan Inc., Northwood, NH), and heifers were fed once daily (0900 hours) ad libitum to ensure at least 5% refusals. Fresh water was freely available throughout the entire experiment. The heifers were exercised daily for 3 h in an outdoor pen, as the measurement and sampling schedule permitted. Heifers were weighed before feeding at the beginning and at the end of each period, and heifers were housed in individual tie stalls on rubber mats and bedded with wood shavings.

Feed intake, sampling, and chemical analyses

The amount of feed offered and refused were recorded daily and daily DM intake (DMI) was calculated as feed offered minus refusals (DM basis). Barley silage, straw, grain, and the total mixed rations (TMR) were sampled weekly, composited by period, and stored at −20 °C. Samples of refusals were collected daily during the data collection period, pooled in large plastic bags for each heifer, and stored at −20 °C. At the end of each period, the composite TMR, ingredient, and refusal samples were mixed and partitioned with a splitter into two portions. One portion was dried in an oven at 55 °C for 48 h (AOAC, 2005; method 930.15) and ground through a 1-mm sieve using a Wiley mill (standard model 4; Arthur Thomas Co., Philadelphia, PA) for chemical analysis. The other portion was used fresh for particle size distribution determination.

Chemical analyses of TMR, ingredients (silage, straw, and grain), and refusals were determined. The DM concentration was determined by oven drying at 55 °C for 48 h, while analytical DM was determined by drying at 135 °C for 2 h (AOAC, 2005; method 930.15). Ash concentration was determined by combustion at 550 °C for 5 h, and organic matter (OM) concentration was calculated as 100 minus ash concentration (AOAC, 2005; method 942.05). Samples were also analyzed for NDF and acid detergent fiber (ADF), with amylase and sodium sulfite used during NDF determination (AOAC, 2005; method 2002.04). Starch was analyzed by enzymatic hydrolysis of α-linked glucose polymers as described by Rode et al. (1999). The concentration of total nitrogen (N) was determined using the flash combustion and thermal conductivity detection technique (model 1500, Carlo Erba Instruments, Milan, Italy), and the crude protein (CP) concentration was calculated as N × 6.25.

Particle distribution and sorting index

The particle size distribution of the TMR and refusal samples was determined using a 3-sieve (19, 8, and 1.18 mm) Penn State Particle Separator (PSPS; Kononoff et al., 2003). The physical effectiveness factor (pef₈) was calculated as the sum of proportions of particle weight retained on the 19- and 8-mm sieves (Lammers et al., 1996). Physically effective NDF was calculated as the NDF concentration of the sample (DM basis) multiplied by the pef₈ (Kononoff et al., 2003). The peNDF was calculated using the particles retained on 19- and 8-mm screens because chewing activity and ruminal pH are primarily correlated with these particles (Yang and Beauchemin, 2007). Furthermore, using a 4- or 1.18-mm screen would have inflated the peNDF concentration, due to the retention of grain particles on these sieves. Sorting index was calculated as the actual DMI of each PSPS fraction expressed as a percentage of the predicted DMI of the fraction (Leonardi and Armentano, 2003). The actual DMI of each individual fraction was determined as the difference between the DM amount of each fraction in the offered feed and that in the refusal. The predicted DMI for each individual fraction was determined by the DMI of the total diet multiplied by the DM percentage of that fraction in the offered diet. A sorting index value equal to 100% indicates no sorting, <100% indicates sorting against, and >100% indicates sorting for particles.

Chewing activity

The activity of the animals was continuously recorded for 3 d (day 15 to 17) using color CCTV cameras (model WV-CP484, Panasonic Corp., Kadoma, Osaka, Japan) placed on a wall shelf installed in the barn. Three 1,400 lumen lights were used to facilitate video recording during the night. Memory cards of cameras were changed daily at 1000 hours, and the videos were copied to the computer and reviewed by three trained observers. The beginning and end of behaviors including eating and ruminating were recorded. The activities of eating and ruminating were defined as reported by Dong et al. (2018).

Rumen pH and fermentation characteristics

Ruminal pH was monitored continuously for 4 d from day 19 to 23 of each period using the Lethbridge Research Centre Ruminal pH Measurement System (LRCpH; Dascor, Escondido, CA) as described by Penner et al. (2006). The pH data loggers (model M1b-pH-1KRTD, Dascor, Escondido, CA) were placed in the rumen at 0900 hours on day 19 and were taken out at 0900 hours on day 23, with ruminal pH readings taken every minute. Ruminal pH data were summarized for each heifer as mean pH, minimum pH, maximum pH, and the duration (h) of pH under 5.8, 5.6, or 5.2. The minimum and maximum pH values for each treatment were obtained from the raw input data using PROC MEANS (version 16.0.0, SAS Inst. Inc. Cary, NC). Ruminal pH of 5.6 and 5.2 was chosen as benchmarks because these threshold values are associated with moderate and severe subclinical rumen acidosis, respectively (Penner et al., 2007).

Ruminal fermentation characteristics were measured on days 18 and 24 at 1, 3, 5, and 7 h after the morning feeding. Ruminal fluid (approximately 300 mL) was collected from different sites (reticulum, dorsal, and ventral sac) and immediately squeezed through four layers of cloth with a mesh size of 250 µm. Five milliliters of filtrate were preserved in a tube that contained 1 mL of 25% HPO₃ for the analysis of VFA; 5 mL of filtrate was persevered with 1 mL of 1% (wt/vol) H₂SO₄ for NH₃-N analysis; another 5 mL of filtrate was preserved in methyl green-formalin-saline for enumeration of total protozoa. All samples for VFA and NH₃-N analysis were stored at −20 °C until analyzed, while samples for protozoa enumeration were stored at room temperature in the dark until counting.

Total tract digestibility and gastrointestinal tract barrier function

From day 19 to 23, total excretions of urine and feces were collected from each heifer for the determination of total tract digestibility as reported previously (Ran et al., 2019). Urine was collected directly from the bladder using indwelling catheters (26 French, 75-cc ballon; C. R. Bard, Inc., Covington, GA) and directed through tubing into closed collection containers containing a sufficient quantity of acid (0.7 liter of 4 N H₂SO₄) to maintain urine pH < 2.5 to prevent volatilization of NH₃. Feces were collected using pans placed behind the animals. The total volume of excreted urine and weight of feces was recorded every 24 h. Representative fecal samples (5% wet weight) were collected and composited by period for each heifer, dried at 55 °C for 48 h, ground through a 1-mm sieve (standard model 4, Arthur Thomas Co., Philadelphia, PA), and retained for chemical analysis.

Chromium-EDTA (Cr-EDTA) was used as a paracellular permeability marker to evaluate total gastrointestinal tract barrier function (Saunders et al., 1994; Zhang et al., 2013). One liter of 180 mM Cr-EDTA solution was pulse dosed into the rumen at 0900 hours on day 19, and urine was subsampled daily to measure urinary Cr output. Briefly, 15 mL of urine was diluted with 60 mL of double-distilled water (ddH₂O) and stored at −20 °C until analysis. The concentration of Cr was measured using inductively coupled plasma emission spectrometry (SpectoCirosCCD, Specto Analytical Instruments GmbH, Kleve, Germany) after dry ashing and extraction with H₃PO₄, MnSO₄, and KBrO₃ (AOAC, 2005; method 968.08).

Blood metabolites and acute phase proteins

Blood samples were collected from each heifer via jugular venipuncture on days 18 and 24 of each period 2 h after feeding. During sampling, blood samples were collected into two 10-mL vacuum tubes containing Na heparin and one 10-mL vacuum tube without additive. Plasma samples were centrifuged at 3,000 × g for 20 min at 4 °C, while serum samples were centrifuged at 2,000 × g for 15 min at 4 °C. Plasma and serum were frozen at −20 °C until analyzed. One milliliter of the plasma was centrifuged at 16,000 × g for 2 min at 4 °C to remove fibrinogen, and the supernatant was analyzed for glucose with a commercial kit (Genzyme Diagnostics, BioPacific Diagnostics Inc., Vancouver, BC, Canada). Serum amyloid A (SAA), plasma haptoglobin (Hp), and lipopolysaccharide-binding protein (LBP) were measured using bovine ELISA kits (Cloud-Clone Corp., Katy, TX; Product No.: SEA885Bo, SEA817Bo, and SEB406Bo) according to the manufacturer’s instructions. The detection limits for SAA, Hp, and LBP were 0.067, 5.9, and 0.287 ng/mL, respectively.

Statistical analysis

Data for feed intake, digestibility, activity, and Cr recovery were analyzed using the PROC MIXED procedure of SAS (version 16.0.0, SAS Institute Inc.) with the PI of barley grain, dietary uNDF concentration, and their interaction considered as fixed effects, and period and heifer considered as random effects. Data for ruminal pH, VFA, NH₃-N, protozoa, and blood metabolites were analyzed using the mixed model procedure of SAS with sampling day or sampling time within day considered as repeated measurements. For repeated measures, various covariance structures were tested and AR(1) was selected based on the lowest value for Akaike’s information criteria. Orthogonal contrasts were conducted to test linear responses to increasing PI of barley grain. Effects were declared significant at P ≤ 0.05, and tendencies are discussed at 0.05 < P ≤ 0.10 unless otherwise stated.

Results

Particle size distribution

There was no interaction of grain PI with dietary uNDF concentration on particle size distribution, pef₈, or peNDF concentration of diets (Table 3). Proportions of particles (DM basis) retained on 19- or 8-mm sieves were not affected by grain PI or uNDF concentration; therefore, pef₈ and peNDF were not affected by PI. Increasing PI (i.e., decreased extent of rolling) linearly (P = 0.01) increased the proportion of particles on the 1.18-mm sieve but linearly (P = 0.01) decreased that of particles retained on the pan. The proportion of particles on the 1.18-mm sieve was greater (P = 0.01), and the proportion retained on the pan was less (P = 0.01), for low uNDF vs. high uNDF diets. Although the pef₈ did not differ between low (0.099) and high (0.095) uNDF diets, the peNDF concentration was greater (P = 0.01) for high uNDF (5.6% DM) vs. low uNDF (4.1% DM) diets due to the inclusion of straw in the high uNDF diets.

Table 3.

Effects of PI of barley grain and dietary concentration of uNDF on particle size distribution of the experimental diets

	PI, %
	65		75		85			P-value
Items1	Low2	High	Low	High	Low	High	SEM	PI	uNDF	PI × uNDF
Particle size, % of DM
19 mm	1.0	0.9	1.0	1.0	1.0	1.0	0.09	0.63	0.31	0.93
8 mm	8.6	8.9	9.1	8.2	9.2	8.6	0.35	0.67	0.18	0.21
1.18 mm*	38.3	34.2	55.0	50.0	68.1	63.7	2.04	0.01	0.01	0.96
Pan*	52.1	56.0	34.9	40.9	21.7	26.7	2.14	0.01	0.01	0.84
pef83	0.096	0.098	0.101	0.091	0.102	0.096	0.004	0.08	0.19	0.76
peNDF84, %	3.9	5.8	4.2	5.4	4.2	5.7	0.30	0.58	0.01	0.92

¹Measured based on six samples of each diet collected by experimental period.

²Low, low concentration of uNDF diet; High, high concentration of uNDF diet.

³pef₈, physical effectiveness factor determined as the sum of particles retained on 19- and 8-mm sieves.

⁴peNDF₈, physically effective NDF estimated as the NDF concentrations of forage multiplied by the pef₈ of the diets.

*Particle weight retained on the 1.18-mm sieve linearly (P = 0.01) increased with increasing PI, whereas the particles passing through the 1.18-mm sieve linearly (P = 0.01) decreased with increasing PI.

Chewing and sorting activity

There was no PI × uNDF interaction for daily eating time, regardless of how it was expressed (Table 4). Altering the PI of barley grain did not affect eating times. However, heifers spent more (P = 0.05) time eating (min/d and min/kg DM; P = 0.05) when fed high uNDF diets although eating time per unit of NDF intake was not affected. Interactions (P = 0.05) between PI and uNDF for ruminating and total chewing times (min/d, min/kg DM, and min/kg NDF) were observed. When the PI was 65%, heifers fed high uNDF diets had increased (P = 0.04) time spent ruminating and chewing compared with heifers fed low uNDF diets, yet uNDF had no effect for 75% and 85% PI diets. When the ruminating activity was represented as min/kg DM or min/kg NDF, heifers fed low uNDF diets spent more time ruminating (P = 0.05) than those fed high uNDF diets for 85% PI, while there were no differences (P > 0.12) among the other treatments (interaction, P = 0.02). For total chewing activity, heifers fed the 75% PI diet had greater chewing time (min/kg DM) when fed high uNDF compared with low uNDF diets, while heifers fed barley processed to 85% PI spent less (interaction, P = 0.03) time chewing (min/kg NDF) when fed high vs. low uNDF diets.

Table 4.

Effects of PI of barley grain and dietary concentration of uNDF on chewing activity and sorting index of heifers fed high-concentrate diets

	PI, %
	65		75		85			P-value
Items	Low1	High1	Low	High	Low	High	SEM	PI	uNDF	PI × uNDF
Eating
min/d	95.2	101.5	94.5	110.5	94.8	107.4	7.57	0.67	0.03	0.77
min/kg DM	8.3	8.8	7.8	10.2	8.1	9.0	1.23	0.81	0.02	0.32
min/kg NDF	49.1	47.9	45.5	55.2	48.2	50.5	4.24	0.73	0.41	0.37
Ruminating
min/d	258.2b	305.0a	284.9ab	296.8a	316.5a	277.8ab	13.49	0.20	0.40	0.04
min/kg DM	22.5b	26.4ab	23.1ab	26.9ab	27.1a	22.8b	2.24	0.21	0.24	0.02
min/kg NDF	133.7b	141.2ab	133.6b	144.2ab	163.8a	127.0b	13.44	0.11	0.83	0.02
Total chewing
min/d	356.7a	405.9b	380.0ab	406.2b	409.2b	384.9ab	18.87	0.18	0.19	0.05
min/kg DM	31.2b	35.1ab	30.9b	37.1a	34.8ab	31.9b	3.11	0.23	0.10	0.02
min/kg NDF	184.4ab	188.6ab	178.5ab	199.1ab	210.0a	176.9b	12.87	0.12	0.98	0.03
Sorting index2, %
19 mm	100.7	84.2	100.2	81.6	102.4	54.1	11.24	0.40	0.01	0.16
8 mm	100.9	98.8	101.0	101.0	102.2	95.3	3.43	0.69	0.17	0.41
1.18 mm	102.8	102.3	100.7	102.8	100.3	102.3	1.26	0.53	0.22	0.44
pan	96.9	99.4	98.9	100.1	98.0	99.8	1.86	0.68	0.16	0.93

¹Low, low concentration of uNDF diet; High, high concentration of uNDF diet.

²Sorting index (%) was calculated as the actual DMI of each PSPS fraction expressed as a percentage of the predicted DMI of that fraction (Leonardi and Armentano, 2003).

^a,bMeans within a row with different superscripts differ (P < 0.05).

No PI × uNDF interaction or effect of barley grain PI on the sorting index was observed (Table 4). Heifers fed high uNDF diets selected (P = 0.01) against particles retained on the 19-mm sieve, whereas no selection of particles occurred for heifers fed low uNDF diets.

Feed intake and apparent digestibility

Interactions between PI × uNDF were observed for DMI and intakes of OM, starch, NDF, ADF, and CP (P < 0.05; Table 5). When fed barley with PI of 85%, intakes of DM and OM were greater (P = 0.01) for high uNDF than low uNDF without effects (P = 0.38) of uNDF for the other PI treatments. When barley was processed at 75% PI, starch intake was less (P = 0.01) when feeding high uNDF compared with low uNDF. Intakes of NDF and ADF were greater (P = 0.05) for high vs. low uNDF diets, with differences being greatest (P < 0.01) for 85% PI. Intake of uNDF was greater (P = 0.01) for high than low uNDF diets as expected, with no impact of barley PI or interaction between PI and uNDF. Although there was no effect of uNDF or PI on CP intake, a PI × uNDF interaction (P = 0.01) occurred where the intake was greater (P < 0.05) when fed high uNDF with 85% PI than the other 65% or 75% PI treatments.

Table 5.

Effects of PI of barley grain and dietary concentration of uNDF on feed intake and digestibility in the total digestive tract of heifers fed high-concentrate diets

	PI, %
	65		75		85			P-value
Items	Low1	High	Low	High	Low	High	SEM	PI	uNDF	PI × uNDF
Feed intake, kg/d
DM	11.87c	12.06bc	12.44ab	12.20bc	12.05bc	12.72a	0.270	0.15	0.97	0.01
OM	11.39b	11.58b	11.88ab	11.70ab	11.56b	12.19a	0.262	0.18	0.92	0.01
Starch	6.32b	6.31b	6.65a	6.33b	6.49ab	6.74a	0.160	0.06	0.25	0.01
NDF	2.02c	2.26ab	2.15b	2.29a	1.98c	2.29a	0.056	0.21	0.01	0.02
uNDF	0.54c	0.68b	0.57c	0.70ab	0.56c	0.73a	0.017	0.19	0.01	0.16
ADF	0.81c	0.95a	0.76d	0.87b	0.82c	0.99a	0.024	0.01	0.01	0.05
CP	1.61ab	1.59b	1.65ab	1.58b	1.62ab	1.68a	0.036	0.12	0.14	0.01
Total tract digestibility, %
DM	83.5a	81.0b	83.3a	80.7b	78.6c	78.1c	0.66	0.01	0.01	0.02
OM	84.9a	82.4b	84.6a	82.1b	80.0c	79.6c	0.62	0.01	0.01	0.02
Starch	98.0	97.6	97.9	97.1	94.8	95.0	0.57	0.01	0.30	0.34
NDF	44.8ab	41.9b	47.0a	44.9ab	40.3b	44.4ab	1.79	0.01	0.76	0.01
ADF	28.8	25.8	20.9	20.7	22.3	27.4	2.76	0.01	0.69	0.06
CP2	79.1	75.2	77.7	73.8	71.6	70.1	0.99	0.01	0.01	0.14

¹Low, low concentration of uNDF diet; High, high concentration of uNDF diet.

²Digestibility of CP linearly (P < 0.05) decreased with decreasing PI of barley grain.

^a–dMeans within a row with different superscripts differ (P < 0.05).

Interactions between PI × uNDF were detected for digestibility of DM (P = 0.02), OM (P = 0.02), NDF (P = 0.01), and ADF (P = 0.06; Table 5). Generally, these interactions were the result of high uNDF diets reducing digestibility when barley was processed to 65% or 75% PI, with no effect when processed to 85% PI. The digestibility of starch was less (P = 0.01) when barley was processed to a PI of 85% than 65% and 75%. The digestibility of ADF was less (P = 0.01) for 75% PI diets than 65% and 85% PI diets without a difference between low and high uNDF diets. The digestibility of CP linearly (P = 0.01) decreased with increasing PI (from 65% to 85%) of barley grain. Feeding high vs. low uNDF diets also reduced (P = 0.01) CP digestibility.

Ruminal pH and fermentation characteristics

There was no PI × uNDF interaction (P = 0.39) for any of the pH variables (Table 6). Furthermore, ruminal pH was not affected (P = 0.30) by PI of barley grain, whereas increasing dietary uNDF concentration decreased (P = 0.05) the duration that ruminal pH was below 5.8 (low vs. high; 14.6 vs. 13.3 h/d) and 5.6 (10.8 vs. 8.6 h/d). Rumen NH₃-N concentration was not affected (P = 0.38) by PI of barley grain, but it was greater (P = 0.02) with high uNDF (average, 6.90 mM) than low uNDF diets (average, 6.14 mM). Ruminal protozoa counts tended to be linearly increased (P = 0.10) with increasing barley PI, and counts were greater (P = 0.03) for high vs. low uNDF.

Table 6.

Effects of PI of barley grain and dietary concentrations of uNDF on ruminal pH and fermentation characteristics in heifers fed high-concentrate diets

	PI, %
	65		75		85			P-value
Items	Low1	High1	Low	High	Low	High	SEM	PI	uNDF	PI × uNDF
Ruminal pH
Minimum	5.16	5.17	5.20	5.14	5.23	5.26	0.027	0.66	0.78	0.90
Mean	5.64	5.73	5.74	5.68	5.79	5.85	0.047	0.30	0.33	0.74
Maximum	6.48	6.48	6.56	6.48	6.51	6.61	0.073	0.41	0.49	0.57
pH < 5.8, h/d	15.34	12.95	13.99	15.33	14.54	11.56	0.545	0.70	0.05	0.75
pH < 5.6, h/d	12.33	8.38	10.00	10.18	9.99	7.33	0.601	0.35	0.05	0.39
pH < 5.2, h/d	2.11	1.35	1.89	1.58	1.62	1.33	0.361	0.96	0.60	0.88
NH3-N, mM	6.53	6.41	5.80	6.68	6.10	6.62	0.984	0.38	0.02	0.57
Protozoa, × 104/mL	1.21	1.36	1.30	1.50	1.46	1.71	0.221	0.10	0.03	0.78
VFA
Total, mM	177.6	168.3	170.9	157.2	168.7	159.5	5.45	0.30	0.01	0.93
Acetate, %	45.8	47.7	45.3	48.5	47.0	48.6	1.66	0.24	0.01	0.83
Propionate, %	39.9ab	39.1ab	40.4a	35.0b	34.8b	36.9ab	2.05	0.21	0.45	0.04
Butyrate, %	9.5	8.2	8.7	10.3	11.0	9.6	1.54	0.23	0.07	0.11
BCVFA2, %	4.9	5.0	5.6	6.2	7.24	5.0	0.91	0.12	0.32	0.08
Acetate:propionate	1.19	1.23	1.15	1.54	1.39	1.42	0.121	0.20	0.11	0.20

¹Low, low concentration of uNDF diet; High, high concentration of uNDF diet.

²BCVFA, branched chain VFA.

^a,bMeans within a row with different superscripts differ (P < 0.05); period × treatment interactions (P < 0.01) for all pH measurements except for minimum pH.

Total VFA concentration was not affected (P = 0.30) by PI but it was less (P = 0.01) with high than low uNDF diets (Table 6). The molar percentages of individual VFA were not affected (P = 0.12) by PI of barley. However, percentage of acetate was greater (P = 0.01) for high uNDF than low uNDF diets. A PI × uNDF interaction occurred for percentage of propionate (P = 0.04) because it was less (P = 0.04) for high uNDF with barley processed to 75% PI with no effect (P = 0.60) of uNDF for 65% and 85% PI. A tendency (P = 0.07) for less molar percentage of butyrate was observed when heifers were fed high uNDF than low uNDF diets. The tendency for interaction for branched-chain VFA concentration (P = 0.08) occurred because with high uNDF, the concentration increased (P < 0.05) when combined with 75% PI, but it decreased (P < 0.05) with 85% PI. Acetate-to-propionate ratio was not affected (P = 0.54) by treatment.

Blood metabolites and gastrointestinal tract barrier function

Blood acute phase proteins (SAA, LBP, and Hp) and blood glucose concentrations were not affected (P > 0.12) by PI of barley grain or dietary uNDF concentration, with no interactions (Table 7). Concentrations of Cr in urine did not differ (P = 0.17) among treatments, and approximately 72% and 93% of Cr that was excreted in urine were recovered within 24 and 48 h, respectively, following ruminal dosing of Cr-EDTA.

Table 7.

Effects of PI of barley grain and dietary concentration of uNDF on blood biochemistry and urinary Cr recovery of heifers fed high-concentrate diets

	PI, %
	65		75		85			P-value
Items	Low1	High	Low	High	Low	High	SEM	PI	uNDF	PI × uNDF
Blood biochemistry
SAA, μg/mL	41.4	36.9	36.4	31.1	43.4	36.2	6.08	0.88	0.40	0.50
LBP, μg/mL	70.5	86.2	97.1	65.0	107.4	93.0	17.42	0.24	0.37	0.23
Hp, μg/mL	237.0	258.6	248.4	247.6	269.5	296.9	20.11	0.12	0.31	0.74
Glucose, mg/dL	118.6	116.4	120.0	124.6	122.2	119.1	7.86	0.28	0.92	0.37
Daily Cr recovery, mg/d
Day 1	251.0	277.2	271.1	283.0	235.3	276.0	36.24	0.73	0.24	0.87
Day 2	69.3	77.6	87.9	74.7	71.2	83.9	7.92	0.50	0.63	0.13
Day 3	17.6	15.6	29.6	17.8	20.4	18.5	4.14	0.18	0.10	0.33
Day 4	5.2	7.2	5.6	5.1	7.3	4.9	1.40	0.67	0.73	0.13
% of Cr recovery
Day 1	73.1	73.4	68.8	74.3	70.4	72.0	9.83	0.73	0.24	0.87
Day 2	93.3	94.0	91.1	94.0	91.7	93.7	9.44	0.60	0.19	0.58
Day 3	98.5	98.1	98.7	98.6	97.8	98.8	9.25	0.48	0.29	0.48
Day 4	100	100	100	100	100	100	0	0.50	0.30	0.49

¹Low, low concentration uNDF diet; High, high concentration uNDF diet.

Discussion

Particle size distribution of TMR

The lack of PI × uNDF interaction on particle size distribution of the diets indicates that processing of barley grain and dietary uNDF concentration acted independently in terms of the physical characteristics of the diets. Changes in particle size distribution were mainly due to a shift in proportion of particle weight retained on the 1.18-mm sieve vs. pan, because the dry-rolled barley grain was small enough to pass through the 8-mm sieve regardless of its PI. The particles retained on 19- and 8-mm sieves of the PSPS have been shown to be positively correlated with chewing activity of cattle (Yang and Beauchemin, 2007). The linear increase in particles on the 1.18-mm sieve with increasing PI of dry-rolled barley grain is consistent with a previous report (Zhao et al., 2015) that indicated less fine particles (<1.18 mm) were generated with coarse dry rolling (i.e., higher PI). Moreover, the decrease in particles on the 1.18-mm sieve when silage was partially replaced with straw (high vs. low uNDF diets), without altering the percentage of particles on 19- and 8-mm sieves, suggests aggressive chopping of straw generated more fines < 1.18 mm. The greater peNDF concentration of high uNDF vs. low uNDF diets, therefore, resulted mainly from the greater NDF concentration of straw because the pef₈ did not differ between high and low uNDF diets. Unlike most of the previous studies that focused on the effects of peNDF concentration of diets without considering uNDF concentration, the present study focused on the effects of dietary uNDF concentration in addition to peNDF concentration. Although both peNDF and uNDF may stimulate chewing and rumen motility, thereby promoting a more regulated ruminal fermentation system (Grant et al., 2018), by definition, the physical effectiveness of peNDF relies on longer particle size of feeds, which may increase sorting of the diet. In contrast, uNDF is associated with the chemical characteristics of NDF, and increasing the uNDF concentration of the diet does not require increasing forage proportion or particle length. Moreover, the uNDF may affect rumen fill, passage rate, digestion kinetics, OM digestibility (Fustini et al., 2017), and prediction of DMI (Grant et al., 2020) to a greater extent than peNDF.

Effects on sorting index and chewing activity

Ruminants can preferentially select components within their TMR and sorting should not be overlooked in intensive beef production systems, in which cattle are fed high-concentrate diets (Madruga et al., 2017). Sorting can result in excessive intake of starch leading to digestive upset, and it can alter the nutrient concentration of the feed remaining in the bunk, which affects the composition of the feed available to pen mates (Miller-Cushon and DeVries, 2017; Coon et al., 2018). In some situations, beef cattle fed high-concentrate diets will intentionally select for long forage particles, assumed to be a feedback response to low ruminal pH (DeVries et al., 2014; Dykier et al., 2020). Increased extent of barley processing in the present study did not lead to feed sorting, possibly because ruminal pH was not affected, whereas the inclusion of barley straw to increase the dietary uNDF concentration caused sorting against long particles, even though the proportion of long particles in the diets did not increase with increasing uNDF concentration. Increased sorting with greater uNDF diets may indicate low palatability of straw or may have been due to the small increase in DM content (from 75.1% to 79.9%, on average) of the diets. Leonardi et al. (2005) reported that cows sorted more when the dietary DM content increased from 64% to 81%.

Total eating time (min/d or min/kg DM) was not affected by PI of barley grain, which was consistent with its lack of effect on DMI. However, information on the effect of dietary uNDF concentration on the eating behavior of finishing beef cattle is scarce. The longer time spent eating (min/d and min/kg DM) when heifers were fed high uNDF diets may have been related to increased sorting against long particles. Greter and DeVries (2010) reported dairy cows that displayed more feed sorting behavior consumed feed at a slower rate. Madruga et al. (2018) reported that sorting and meal length increased with increasing inclusion of alfalfa hay in finishing diets, which presumably corresponded to an increase in uNDF concentration of diets, although not measured in the study. Although NDF intake was greater for heifers fed high uNDF diets, the eating rate of NDF (min/kg) was not affected by dietary uNDF concentration likely due to the low uNDF concentrations of the diets and limited differences in dietary particle size. Furthermore, low correlation between eating time and dietary NDF concentration or forage NDF concentration was reported across studies (Beauchemin, 2018).

Rumination time is highly influenced by NDF intake, particle size, fragility, indigestibility of the fiber, and their complex interactions (Beauchemin, 2018). The observed interactions between the PI of barley grain and dietary uNDF concentration on ruminating time, expressed as minutes per day or per unit of DM or NDF intake, indicate that the effects of dietary uNDF concentration on ruminating depended upon the PI of barley. The effects of dietary uNDF concentration on rumination in finishing beef cattle have not previously been reported to our knowledge. Cotanch (2015) reported that increasing dietary uNDF concentration of dairy cow diets from 7.2% to 8.9% (DM basis) decreased rumination time (min/d or min/kg NDF) but without effects on min/kg DM or min/kg uNDF. Similarly, Fustini et al. (2017) observed decreased rumination time for dairy cows fed high (10.9%) compared with low (9.4% DM) uNDF diets. Those reports are somewhat consistent with our present finding that with less processing the low uNDF was enough to keep rumination activity. This interaction is not easily explained. Krizsan et al. (2010) found that uNDF has a faster passage rate from the rumen compared with pdNDF, which might explain the decreased rumination time with high uNDF and high PI diet. However, in a review, Beauchemin (2018) indicated that less digestible forage fiber led to increased rumination time per unit of DMI because the lower rate of digestion maintains physical fill in the rumen for a longer time. Therefore, the lower DM digestibility of the high uNDF vs. low uNDF diets for PI of 65% and 75% may have caused the increase in rumination time. These results suggest that the decrease in rumination time due to more extensive processing grain (lower PI) was partially offset by providing additional uNDF. Grant et al. (2020) also suggested that adding more uNDF to a diet can adjust for a lack of dietary peNDF. However, the results from our study indicate that increasing dietary uNDF concentration in highly processed feedlot finishing diets may increase ruminating time and decrease the extent of ruminal acidosis, but diet digestibility may decrease as well. The total chewing time (eating and ruminating) showed a pattern similar to that of ruminating time. Thus, for chewing and ruminating activity, the effects of dietary uNDF concentration depended upon grain processing, whereas eating time was consistently affected by uNDF.

Effects on feed intake, apparent digestibility, and rumen fermentation

The interaction between PI of barley and dietary uNDF concentration for DMI indicates that these factors were not independent. Further examination of the interaction revealed that high uNDF stimulated intake only when barley was very coarsely processed (PI = 85%). It has been suggested that uNDF intake can limit the ability of cattle to consume sufficient nutrients when offered high forage diets with a maximum daily uNDF intake of about 13 g/kg BW^0.75 for yearling steers (Lippke, 1986) or 0.3% to 0.4% of BW for dairy cows (Mertens, 1997). In the present study, the uNDF intakes (0.54 to 0.73 kg/d) were far below the threshold maximum uNDF consumption (1.8 kg/d for 715 kg heifers; Lippke, 1986). Therefore, the dietary uNDF concentrations used in the study were not expected to limit DMI. The greater DMI with high vs. low uNDF diets when barley was processed to 85% PI may have been due to increased rumen motility (and passage rate) due to increased rumen fill and improved ruminal pH. The lack of effect of PI of barley on DMI by cattle fed high-grain diets was consistent with some (Beauchemin et al., 2001; Koenig et al., 2003) but not all studies (Bengochea et al., 2005; Koenig and Beauchemin, 2011; Moya et al., 2015). Beauchemin et al. (2001) reported no effect of processing barley to a PI of 82%, 75%, 70%, and 65% on the intake of DM, OM, or NDF by finishing beef heifers. Koenig et al. (2003) also reported no difference in DMI between diets containing a barley PI of 86% and 61% with 20% or 5% of barley silage. In contrast, Moya et al. (2015) reported greater DMI by steers fed finishing diets containing dry-rolled barley with a PI of 85% than with a PI of 75%. The current research indicates the inconsistent effects of varying barley PI on DMI of feedlot cattle among studies are due to complex interactions of chemical composition, and type and concentration of roughage in the diet, as well as other factors that impact ruminal fermentability. Furthermore, barley processed to the same PI can differ in particle distribution and ruminal digestion kinetics (Zhao et al., 2015), and DMI has been shown to be negatively correlated (r = −0.67) to rate of DM disappearance from the slowly degradable fraction (Ramsey et al., 2002).

The interaction of PI of barley grain with dietary uNDF concentration on the total tract digestibility of DM, OM, NDF, and ADF was primarily due to the similar digestibility between low and high uNDF at PI of 85%, whereas the digestibility decreased with high uNDF at PI of 65% or 75%. The results indicate that the impact of dietary uNDF becomes more apparent when feeding diets with barley grain processed to a greater extent. The decreased digestibility by increasing dietary uNDF concentration at PI of 65% and 75% is consistent with studies in dairy cows fed diets containing 45% to 55% of forage (Cotanch, 2015; Fustini et al., 2017), and in sheep fed diets containing 75% forage (Yousefian et al., 2019). In those dairy studies, the digestibility of DM was not measured, but the reduction in the digestibility of NDF appeared mainly due to the added uNDF that is undigested in the total digestive tract. Similarly, in the present study, the lowered digestibility with high uNDF appeared due to the substitution of low-quality straw (i.e., adding uNDF) for relatively highly digestible silage. In addition, the interaction effect may be partially explained by the increase in sorting against long particles when heifers were fed the 85% PI diets. The long particles have high fiber concentration but low digestible NDF concentration. As a result, the digestibilities of NDF and ADF were 4 and 5 percentage units greater with high uNDF than low uNDF diets. These results suggest that the impact of dietary uNDF on the total tract digestibility may primarily be due to its indigestibility, in particular, with high grain and low uNDF diets as were used in the present study.

The impacts of extent of barley processing on feed digestibility and growth performance of feedlot cattle are well documented (Ramsey et al., 2002; Dehghan-banadaky et al., 2007). Increasing the extent of barley processing from coarse to fine often leads to quadratic responses in total tract digestibility. Beauchemin et al. (2001) reported that the DM digestibility in the total digestive tract of finishing heifers was greater in diets containing barley with PI of 75% than 82%, with no further increase in digestibility as grain was processed below 75%. Our results are in agreement with that study, as digestibility of DM and OM did not differ between PI of 65% and 75% but decreased with PI of 85%. For barley, once the fibrous hull and pericarp are disrupted during processing, barley starch and CP are highly digestible in the rumen (Koenig and Beauchemin, 2011). Therefore, excessive processing may not further improve the digestibility, but it may increase the risk of digestive upsets, especially ruminal acidosis (Dehghan-banadaky et al., 2007). Thus, it is important to identify the optimum extent of processing that maximizes the productivity of cattle (Dehghan-banadaky et al., 2007). In the present study, similar to the recommendation from other studies (Beauchemin et al., 2001; Wang et al., 2003), the optimum PI based on maximizing digestibility and minimizing potential for acidosis was 75%. Although there were no major differences in variables measured between PI of 75% and 65%, less processing energy is required for coarser rolling. The greater ADF digestibility of 65% PI diets is in agreement with Beauchemin et al. (2001), who reported that ADF digestibility in the total digestive tract of finishing heifers tended to increase with increasing extent of barley processing. Our results are consistent with previous reports that show increasing the extent of processing of barley had no negative impact on ruminal fiber digestion because ruminal pH was also not affected (Zinn, 1993; Beauchemin et al., 2001). The increased digestibility of CP with decreasing PI of barley in the current study may be due to an increase of microbial protein flow to the duodenum (Beauchemin et al., 2001).

The decreased duration of pH < 5.8 (14.6 vs. 13.3 h) and <5.6 (10.8 vs. 8.6 h) for high uNDF vs. low uNDF diets is consistent with the corresponding decrease in total VFA concentrations. These results are consistent with the observed decrease in dietary DM and OM digestibility. Therefore, incorporating straw into a finishing diet as a source of uNDF to improve rumen function may decrease animal performance if the decrease in digestibility is not compensated for by an increase in DMI. Overall, there was no increase in DMI with high uNDF diets, except when barley PI was 85%. Alternatively, the decrease in duration of pH < 5.6 with increasing uNDF concentration may suggest that animals fed high uNDF diets had greater ability to remove ruminal VFA from the rumen either by neutralization due to increased salivary buffering attributed to greater eating time or by increased rumen motility that helps move VFA toward the ruminal epithelium (Storm and Kristensen, 2010). Consistent with our results, Fustini et al. (2017) reported a shorter duration of pH < 5.8 and <5.5 in dairy cows fed high uNDF compared with low uNDF diets. These results suggest an important role of dietary uNDF content in stabilizing ruminal pH and demonstrate that even though the uNDF concentration of feedlot finishing diets is relatively low, it can have substantial beneficial impact on ruminal pH status. However, increasing uNDF concentration of diets lowered DM digestibility (high vs. low uNDF; 81.8% vs. 79.9%), although the low cost of straw in high uNDF diets may compensate for lower digestibility, if animal growth performance is not negatively affected. Moreover, the greater ruminal NH₃-N concentration (high vs. low uNDF; 6.57 vs. 6.14 mM) might have resulted from reduced ruminal microbial protein synthesis because the digestibility of OM and CP in the total digestive tract was adversely affected by increasing dietary uNDF concentration. This speculation is also consistent with the enhanced protozoa counts with high vs. low uNDF diets because protozoa can decrease ruminal microbial production efficiency. Lack of barley PI effects on ruminal pH and fermentation characteristics are in line with reports by Beauchemin et al. (2001) and Koenig et al. (2003). The lack of barley processing effect on ruminal pH suggests that manipulating the extent of barley processing did not reduce the risk of rumen acidosis in finishing heifers.

Effects on blood metabolites and gastrointestinal barrier function

The acute phase proteins SAA, LBP, and Hp are commonly attributed to the activation of a systemic immune response (Gruys et al., 2005; Eckersall and Bell, 2010). Serum SAA, LBP, and Hp concentrations in the present study were within physiological ranges but were not affected by treatments. Thus, we conclude that the heifers did not experience inflammation or other diet-related stress. Greater risk for immune system activation and other diseases is expected when gastrointestinal barrier function is compromised (Zhang et al., 2013). Therefore, Cr-EDTA was pulse dosed into the rumen, and urinary Cr recovery was measured to assess total tract barrier function. The use of Cr-EDTA is an indicator of generalized barrier function but it does not directly indicate potential for endotoxin translocation due to the differences of molecular weight, with 340 Da for Cr-EDTA and 2 to 70 kDa for endotoxins (Zhang et al., 2013). Endotoxin release from Gram-negative bacteria increases when cattle are fed highly fermentable grain. It is a strong pro-inflammatory agent that stimulates the acute phase response (Plaizier et al., 2012). In the present study, Cr-EDTA recovery in urine during 48 h was only 3.4% of the amount dosed into the rumen, which was lower than 8% reported by Zhang et al. (2013). The lack of differences of Cr recovery among treatments suggested no treatment effect on gastrointestinal barrier function. The considerable low urinary Cr recovery supports the results of acute phase proteins, suggesting that gastrointestinal barrier function was not compromised by treatments, even though concentrations of VFA in ruminal fluid were high and duration that pH was < 5.6 was extensive (7 to 12 h/d).

Conclusions

Numerous interactions between PI of barley and dietary uNDF concentration were observed (feed intake, rumination, and total tract digestibility of DM, OM, and NDF), suggesting that the effects of these variables were interdependent for beef heifers fed high-grain diets. Overall, increasing extent of barley processing (i.e., decreasing PI) tended to decrease ruminating activity of cattle fed low uNDF diets and increased the total tract digestibility of OM, starch, and CP without impacting ruminal pH and fermentation parameters. We conclude that a PI of 75% is optimum in terms of feed digestibility because further decreasing the PI to 65% did not result in a further improvement in digestibility. Increasing dietary uNDF concentration elevated the intake of fiber regardless of PI of barley grain and stimulated rumination when barley was processed with a PI of 65%. The digestibility of DM and total VFA concentration decreased, whereas ruminal pH status improved with high uNDF compared with low uNDF diets. Thus, increasing the uNDF concentration of diets is an effective strategy to improve ruminal pH status in finishing cattle without increasing the proportion of forage in the diet, regardless of the extent of grain processing. In contrast, minimizing the extent of grain processing did not reduce the risk of rumen acidosis.

Glossary

Abbreviations

ADF

acid detergent fiber

ADG

average daily gain

BW

body weight

CP

crude protein

DM

dry matter

DMI

dry matter intake

Hp

haptoglobin

LBP

lipopolysaccharide-bi

n

ding protein

NDF

neutral detergent fiber

OM

organic matter

PI

processing index

SAA

serum amyloid A

TMR

total mixed rations

VFA

volatile fatty acid

Conflict of interest statement

The authors declare no conflicts of interest regarding the publication of this manuscript.