Effect of physically effective fiber on animal growth performance, carcass characteristics, digestibility, and rumination in finishing beef cattle

Abstract

The objective of this experiment was to evaluate the effects of altering dietary peNDF concentration by modifying forage particle size on digestibility, rumination, and performance of finishing beef steers. Forty crossbred beef steers (initial BW = 387 ± 38.2 kg) were blocked by BW and assigned to 10 pens with 4 steers per pen. Pens were then randomly assigned to one of two dietary treatments (n = 5 pens per treatment; 4 steers per pen). Diets were formulated identically, differing only in forage particle size. Prairie hay was processed through a commercial tub grinder equipped with a 17.8- or 10.2-cm screen, resulting in mean particle sizes of 11.54 and 7.68 mm for long-cut and short-cut, respectively. Treatments were based on dry-rolled corn and included 8% long-cut or 8% short-cut prairie hay (DM basis). Particle distribution and peNDF were estimated using the Penn State Particle Separator. Data were analyzed using mixed models. Treatment was included as a fixed effect and block as a random effect for pen-level variables, with pen serving as the experimental unit. For variables measured on individual animals, animal nested within pen was included as a random effect, and day was included as a repeated measure with animal nested within pen as the subject. Decreasing forage particle size reduced dietary peNDF from 14.23% to 12.33% of DM (P = 0.014). Neutral detergent fiber digestibility was reduced in short-cut TMR (P < 0.001), whereas DM and starch digestibility were unaffected (P ≥ 0.62). Feedlot performance and carcass characteristics did not differ between treatments (P ≥ 0.18). Fecal pH differed between treatments at specific sampling days, resulting in a treatment × day interaction (P = 0.001). Serum amyloid A concentrations tended to be greater for long-cut TMR (P = 0.09). Rumination time did not differ between treatments (P = 0.53). These results suggest that a 1.90 percentage-point decrease in dietary peNDF via forage particle size did not affect growth performance, rumination behavior, or carcass characteristics in finishing beef cattle, but may reduce NDF digestibility.

Feeding finishing diets with reduced forage particle size did not affect growth performance, carcass characteristics, or rumination. However, reduced fiber digestibility and a potential increase in liver adhesion incidence were observed.

Introduction

Finishing beef cattle diets are formulated to maximize growth performance and carcass yield by increasing dietary energy density through the inclusion of rapidly fermentable carbohydrates. Although greater concentrate inclusion improves growth rate and feed efficiency, it also increases the risk of digestive disorders that can negatively affect animal growth performance and health (Owens et al. 1998). Effective fiber, therefore, remains a necessary component of finishing diets because it stimulates chewing activity, promotes saliva secretion, and contributes to ruminal buffering capacity and stability (Stock and Britton 1991). Forage sources are generally more expensive per unit of energy, lower in digestibility, and often require additional processing (Mertens 2002). To maintain dietary energy density while supporting normal rumen function, forage inclusion in finishing diets is typically limited to 8%–10% of diet dry matter (DM) (Samuelson et al. 2016). Current nutrient recommendations indicate that a minimum dietary peNDF concentration of 5%–8% of DM for cattle fed high-concentrate finishing diets (National Academies of Sciences and Medicine 2016).

Fiber concentration in finishing beef cattle diets is measured using neutral detergent fiber (NDF), which quantifies the concentration of structural carbohydrates within the diet (Van Soest et al. 1991). Although NDF quantifies the chemical nature of the fiber, it does not capture the physical attributes of fiber, such as particle size and structural rigidity that influence chewing activity and rumination. To address this limitation, physically effective neutral detergent fiber (peNDF) was developed to integrate fiber concentration with particle size distribution (Mertens 1997). Physically effective fiber represents the fraction of dietary fiber that stimulates chewing activity and saliva production, thereby supporting normal rumen function. Estimation of peNDF relies on particle separation techniques, such as the Penn State Particle Separator, to quantify the proportion of feed particles exceeding a defined threshold size that are assumed to contribute to ruminal fiber mat formation, in combination with NDF concentration (Lammers et al. 1996; Maulfair et al. 2011).

In finishing beef cattle diets, where forage inclusion is minimal, the applicability and biological relevance of peNDF estimates remain poorly defined. Most research evaluating peNDF in finishing diets has focused on altering forage inclusion or forage source rather than isolating the effects of forage particle size alone. Forage particle size can be manipulated to improve diet mixing and uniformity, yet its effects on effective fiber supply, rumination behavior, and growth performance in finishing beef cattle remain poorly understood. Therefore, the objective of this experiment was to evaluate the effects of altering dietary peNDF concentration via forage particle size on digestibility, rumination, and growth performance of finishing beef steers.

Materials and methods

All animal procedures were approved by the Oklahoma State University Institutional Animal Care and Use Committee (IACUC-22-39).

Animals

Forty-nine crossbred steers were received at the Oklahoma State University Willard Sparks Beef Research Center, weighed on a pen scale, and held in a dry lot pen overnight with ad libitum access to prairie hay and water before processing. Initial processing was conducted 24 h after arrival and included ear tagging for individual animal identification, weighing, and vaccination against bacterial and viral respiratory pathogens (Titanium 5; Elanco Animal Health, Greenfield, IL, USA), clostridial pathogens (Vision with SPUR; Merck Animal Health, Madison, NJ, USA), and other bacterial pathogens (Fusogard and Nuplura PH; Elanco Animal Health). Additionally, steers were implanted with Synovex Choice (Zoetis, Florham Park, NJ, USA), administered an oral anthelmintic (Safeguard; Merck Animal Health), and treated with a pour-on insecticide (Permectrin; Elanco Animal Health). Following processing, steers were placed on a receiving diet for 14 d [DM basis: 15% dry-rolled corn, 51% Sweet Bran (Cargill Inc., Dalhart, TX, USA), 29% prairie hay, and 5% vitamin/mineral supplement]. Steers were then transitioned to a finishing diet over a 28-d period using a 4-step ration transition. All steers were fed a finishing diet [DM basis: 64% dry-rolled corn, 18% Sweet Bran (Cargill Inc.), 8% prairie hay, 5% vitamin/mineral supplement, and 5% liquid supplement] for approximately 40 d until trial initiation. The diet was formulated to meet or safely exceed nutrient requirements (National Academies of Sciences and Medicine 2016).

Experimental design and treatment

Steers were individually weighed on d −7 to obtain initial BW (n = 49; 387 ± 38.2 kg). Steers were stratified by initial BW and assigned to 1 of 2 weight blocks (6 pens per block) in a generalized randomized block design. Within each block, pens were randomly assigned to one of two dietary treatments, containing either long-cut or short-cut prairie hay (3 pens per treatment per block). Two pens in the light BW block (one short-cut and one long-cut) were removed from the trial due to differences in finishing time. Consequently, forty steers remained on trial and were housed in 10 pens. The final design consisted of 6 pens in the heavy BW block (3 pens per treatment) and 4 pens in the light BW block (2 pens per treatment), with 4 steers per pen and pen serving as the experimental unit. Steers were housed in 4.57 × 13.24 m partially covered feedlot pens with shared 76-L concrete water tanks between 2 adjacent pens (model J 360-F; Johnson Concrete, Hastings, NE, USA). On d 0, steers were weighed, administered a terminal implant (Synovex One Feedlot, Zoetis), and fitted with an ear tag (CowManager SensOor; Agis Automatisering, Netherlands) to continuously measure individual animal behaviors, including rumination. Individual rumination measurements were averaged within pen for statistical analysis. Pens contained a 4.57 × 4.42 m covered concrete pad, with the remainder being soil-surfaced.

Different particle sizes of prairie hay were achieved by passing the hay through a commercial tub grinder (Hydrafork Model 5012; Mighty Giant, Jones Manufacturing Co., Beemer, NE, USA) equipped with a 17.8 cm round hole screen for long-cut hay or a 10.2 cm round hole screen for short-cut hay. Particle size was quantified for both individual ingredients and dietary treatments using the Penn State Particle Separator (PSPS; The Pennsylvania State University, University Park, PA, USA). The treatment diets (Table 1) were dry-rolled corn-based and included 18% Sweet Bran (SB; Cargill Inc.) with either 8% short-cut prairie hay (Short-cut TMR) or 8% long-cut prairie hay (Long-cut TMR) on a DM basis.

Table 1.

Ingredient and nutrient composition of diets.

	Dietary treatment1
Ingredient, % of DM	Long-cut TMR	Short-cut TMR
Dry-rolled corn	64.0	64.0
Long-cut prairie hay2	8.19	–
Short-cut prairie hay2	–	8.19
Sweet bran3	18.15	18.15
Liquid supplement4	4.72	4.72
Dry supplement5	4.94	4.94
Nutrient composition, DM basis
Dry matter, %	83.53	83.63
Crude protein, %	13.10	13.50
Neutral detergent fiber, %	22.79	20.97
Acid detergent fiber, %	6.50	6.20
Ether extract, %	3.13	3.21
TDN, %	77.00	77.00
Ca, %	0.54	0.66
P, %	0.48	0.50
Mg, %	0.20	0.22
K, %	0.78	0.78
Na, %	0.08	0.08
NEm, Mcal/kg	1.90	1.87
NEg, Mcal/kg	1.27	1.26

¹Long-cut TMR = 8% long-cut prairie hay; Short-cut TMR = 8% short-cut prairie hay.

²Long-cut prairie hay was passed through a commercial tub grinder equipped with a 17.8 cm round hole screen. Short-cut prairie hay was passed through a commercial tub grinder equipped with a 10.2 cm round hole screen.

³Sweet Bran (Cargill Inc., Dalhart, TX).

⁴Liquid supplement was formulated to contain (% DM basis) 45.86% corn steep, 36.17% cane molasses, 6% hydrolyzed vegetable oil, 5.46% 80/20 vegetable oil blend, 5.2% water, 1.23% urea (55% solution), and 0.10% xanthan gum.

⁵Dry supplement was formulated to contain (% DM basis) 42.6% ground corn, 27.1% limestone, 20.6% wheat middlings, 6.5% urea, 0.9% salt, 0.49% magnesium oxide, 0.12% copper sulfate, 0.15% manganous oxide, 0.08% selenium, 0.47% zinc sulfate, 0.29% vitamin A (30,000 IU/g), 0.09% vitamin E (500 IU/g), 0.008% vitamin D (30,000 IU/g), 0.19% tylosin (Tylan-40, Elanco Animal Health, Greenfield, IN), and 0.30% monensin (Rumensin-90, Elanco Animal Health).

Feed management

Feed bunks were visually evaluated daily at 0600 h to estimate orts remaining from the previous day and at 1700 h to assess orts remaining from the morning feeding. The amount of feed delivered was adjusted based on these evaluations to target ≤0.45 kg of orts remaining in the feed bunk. Cattle were fed once daily at 0830 h. Diets were mixed for 3 minutes at 2000 rpm and delivered using a trailer-mounted feed mixer (274-12B feed mixer; Roto-mix, Dodge City, KS, USA).

Sample collection and calculations

Ingredient and diet samples were collected weekly throughout the trial for particle separation analysis and used to calculate peNDF concentration (n = 12 samples per ingredient and dietary treatment). Diet samples were collected weekly from each bunk to obtain DM values. Dry matter was determined after the samples were dried in a forced-air oven at 55 °C for 48 h. Diet samples were ground through a 1-mm screen (Puverisette 19, Fritsch, Pittsbro, NC, USA) and composited at the conclusion of the trial for nutrient analysis. Particle separation of ingredients and mixed diet samples were conducted prior to sample drying using the PSPS as described by Kononoff et al. (2003). Briefly, the particle separator was placed on a flat surface with the sample on the top sieve, then shaken 5 times and rotated 90°. This was repeated for 8 sets, or for a total of 40 shakes. Physically effective neutral detergent fiber was calculated by multiplying the percentage of sample larger than 4 mm in particle size (as a percent) by the percent NDF concentration (see below for method) of the sample before separation. Additionally, peNDF was estimated using fiber-source ingredients only (pefNDF). The ingredients used in this calculation were the prairie hay (Long-cut or Short-cut) and SB, and total-ingredient peNDF values were multiplied by their respective dietary inclusion on a DM basis and summed to obtain dietary pefNDF.

Body weights (BW) were collected on d 0, 21, 42, 63, 84, 105, and at shipping. Due to a mechanical malfunction of the squeeze chute, all measurements scheduled for d 63 were conducted on d 64. All BW measurements were obtained before feeding at 0630 h, with no withdrawal from water or feed, and the scale was validated before measurement. All BW measurements were adjusted using a 4% pencil shrink (BW × 0.96), averaged within pen, and used to calculate performance measurements. Daily dry matter intake (DMI) was calculated from the amount of feed delivered to each pen and the DM concentration of the diet, divided by the number of steers in the pen. Daily NDF intake was estimated by multiplying DMI by the NDF concentration of the offered diet. The NDF concentration of orts was not determined. Pen average daily BW gain (ADG) was calculated by averaging the individual ADG of steers within a pen for each 21-d period and for the entire trial. All pens contained 4 steers throughout the experiment, and period length was consistent across treatments. Gain to feed (G:F) ratio was calculated by dividing the pen ADG (kg/d) by the pen DMI (kg/d).

Steers were shipped to a commercial abattoir on d 124. Trained personnel from Oklahoma State University collected liver abscess scores on the day of slaughter, using the Elanco Liver Check scoring system (Brink et al. 1990). Livers without abscesses were classified as normal; A- livers contained 1 or 2 small abscesses; A livers contained 2–4 small, active abscesses; and A + livers contained multiple small abscesses or 1 or more large, active abscesses. Livers classified as A + adhesion had adhesion to the diaphragm, gastrointestinal tract, or both, indicative of current or previous liver abscesses. Carcass data were collected by plant personnel following a 48-h chill. Carcass-adjusted final shrunk BW (SBW) was calculated from hot carcass weight (HCW) divided by the average dressing percent (DP) across all animals (62.75%). Carcass-adjusted ADG was calculated from days on feed, initial BW (kg), and carcass-adjusted final SBW (kg). Carcass-adjusted G:F was calculated by dividing carcass-adjusted ADG (kg/d) by average DMI (kg/d).

Fecal and blood samples were collected on d 0, 21, 42, 64, 84, and 105 before feeding. Blood was collected via jugular venipuncture into tubes (9 mL neutral Sarstedt Monovette, Sarstedt AG & Co. KG, Nümbrecht, Germany) containing K₂EDTA (1.5 mg/mL), inverted, and immediately placed on ice. Samples were then centrifuged for 25 minutes at 3,500 × g at 4 °C. Plasma was stored at −20 °C until further analysis. Fecal samples were collected via rectal stimulation from all animals at each of the six sampling time points. Fecal pH was immediately measured chute-side using a portable pH meter (Oakton pH 5+ portable meter, Thermo Fisher, MA, USA). Additionally, fecal samples were visually assessed and assigned a fecal consistency score (FCS). This scoring method used a 1–5 scale and characterized samples as follows: 1 = firm; 2 = slightly firm; 3 = high moisture and soft; 4 = runny and loose; and 5 = watery consistency and very loose (Woolsoncroft et al. 2018). The same evaluator assigned FCS every collection day. Fecal samples were dried in a forced-air oven at 55 °C for 72 h, and DM values were determined. Fecal samples were ground using a commercial blender and composited within each pen for each day. Fecal composites were analyzed for NDF, starch, and indigestible NDF (iNDF) content (described below).

Rumination

Steers were fitted with a CowManager SensOor (Agis Automatisering) on d 0 to continuously measure individual animal behavior. Rumination data used for analyses were collected beginning on d 13 of the feeding trial. Behavioral data were collected every minute and classified as 1 of 5 behaviors: not active, active, highly active, ruminating, or eating. This experiment focused exclusively on rumination time. Individual animal rumination times were summed daily and averaged within pen to obtain the average time spent ruminating per pen.

Lab analysis

Composited treatment diet samples were shipped to a commercial laboratory for nutrient analysis (Dairy One Laboratories, Ithaca, NY, USA), aside from NDF and ether extract (EE), which were analyzed separately. NDF was analyzed in duplicate using an ANKOM 2000 automated fiber analyzer (ANKOM Technology; Macedon, NY, USA) according to the manufacturer’s instructions, with the addition of 20 g of sodium sulfite and 4.0 mL of alpha-amylase (ANKOM Technology). Diet samples were analyzed in duplicate for EE using ANKOM XT15 extract (ANKOM Technology) according to the manufacturer’s instructions. Gravimetric analyses were adjusted for laboratory DM, determined by drying samples in a convection oven at 105 °C for 24 h.

Plasma samples were thawed at ambient temperature and analyzed for glucose and serum amyloid A (SAA). Plasma glucose was analyzed using a commercially available kit (FujiFilm Wako Pure Chemical Corporation, Chuo-Ku, Osaka, Japan). Samples were analyzed in duplicate in flat-bottom 96-well polystyrene plates using a microplate reader (BioTek EPOCH, BioTek Instruments, Inc., Winooski, VT, USA) at 505 nm. The intra- and inter-assay coefficients of variation for glucose were 3.08% and 8.70%, respectively. SAA concentrations were determined in plasma using a commercially available sandwich ELISA assay (Phase Range SAA assay, Tridelta Development Limited, Maynooth, Ireland), as described by McDonald et al. (1991). Samples were analyzed in duplicate in flat-bottom SAA antibody-coated 96-well plates. The intra- and inter-assay coefficients of variation for SAA were 2.31% and 8.37%, respectively.

Fecal samples were ground using a commercial blender and composited by pen for each day. Fecal composites were analyzed for DM, NDF, starch, and iNDF content. Laboratory DM for fecal composites was determined by drying samples at 105 °C for 24 h and was calculated as the weight after drying divided by the weight before drying. Fecal NDF was analyzed using an ANKOM 2000 automated fiber analyzer (ANKOM Technology). Starch concentrations for fecal composites and diet composites were obtained using methods adapted from Knudsen (1997). Briefly, 100 mg of ground diet samples and 200 mg of ground fecal samples were analyzed using heat-stable alpha-amylase (ANKOM Technology) and amyloglucosidase (ANKOM Technology). Glucose concentrations were analyzed as described above for plasma samples. The percent starch of the sample was calculated by dividing the amount of glucose (mg), corrected for glucose recovery from glucan chains by dividing by 1.11, by the sample weight, multiplied by the sample DM, and multiplying by 100 to convert to a percentage.

Indigestible NDF (iNDF) was analyzed by incubating duplicate fecal and composite diet samples in the rumen of two cannulated steers (two bags per sample in each steer) for 576 h (Adams et al. 2020). Bags were removed from the cannulated steers, placed in ice-cold water to stop microbial activity, and rinsed with water until the rinse water was clear, using a washing machine on a delicate cycle. Bags were dried for 48 h at 55 °C, followed by 24 h at 105 °C in a forced-draft oven. After drying, bags were subjected to the NDF protocol without sodium sulfite (Van Soest et al. 1991) and weighed after 3 h at 105 °C. Dry matter, starch, and NDF digestibility were determined using equations previously described (Adams et al. 2020).

Statistical analysis

Particle size distribution and peNDF estimates were analyzed using the MIXED procedure of SAS (version 9.4; SAS Institute Inc., Cary, NC, USA). Analyses were conducted separately for ingredients and diets. Particle size distribution data were analyzed independently for each sieve, with ingredient or diet included as the fixed effect and the percentage of particles retained as the response variable. The experimental unit was the individual ingredient or diet sample collected on each sampling date. Estimated peNDF values were analyzed, similarly, with ingredient or diet included as the fixed effect. Least squares means were generated for all models, and post-hoc pairwise comparisons were performed using Tukey’s adjustment. Normality was assessed using residual plots.

Animal performance and carcass data were analyzed as a randomized complete block design with block as a random effect and dietary treatment as a fixed effect using the MIXED procedure of SAS (version 9.4; SAS Institute Inc.). For ADG and SBW, pen nested within block and treatment were included as random effects to account for animals within pen. Carcass-adjusted final SBW and carcass-adjusted ADG were analyzed using the same model structure. The denominator degrees of freedom were adjusted using the Kenward-Roger method. Pen was the experimental unit, and normality was assessed using residual plots. Digestibility variables were analyzed with fixed effects of dietary treatment, sampling day, and their interaction, with block included as a random effect. Sampling day was included as a repeated measure, with pen serving as the experimental unit and subject for repeated measures. The first-order autoregressive covariance structure provided the lowest Akaike information criterion (AIC) and Bayesian information criterion (BIC) for plasma and digestibility variables. Fecal variables (fecal pH and FCS) and plasma variables were analyzed with fixed effects of block, dietary treatment, sampling day, and the treatment × day interaction. Animal nested within pen was included as a random effect. Sampling day was included as a repeated measure, with animal nested within pen serving as the subject for repeated measures. The heterogeneous autoregressive covariance structure provided the lowest Akaike information criterion (AIC) and Bayesian information criterion (BIC) for fecal variables. Means from the mixed model analyses were separated using a Tukey post-hoc adjustment. Least squares means ± the standard error of the mean (SEM) are reported. The GLIMMIX procedure of SAS (version 9.4; SAS Institute Inc.) was used to analyze the effect of treatment on the prevalence of liver adhesion. Models were fit using a binomial distribution. Rumination behavior variables were analyzed using the MIXED procedure of SAS (version 9.4; SAS Institute Inc.). Fixed effects included dietary treatment, experimental day, and their interaction, with block included as a random effect. Experimental day was included as a repeated measure, with pen serving as both the experimental unit and the subject for repeated measures. Rumination variables were expressed as minutes per day (min/d) and relative to intake as minutes per kilogram of DM, NDF, peNDF, and pefNDF. The compound symmetry covariance structure provided the lowest Akaike information criterion (AIC) and Bayesian information criterion (BIC). Denominator degrees of freedom were estimated using the Kenward-Roger method. Model assumptions were evaluated using residual diagnostics. For all analyses, statistical significance was declared when P ≤ 0.05 and a tendency when 0.05 < P ≤ 0.10. P-values are reported with two decimal places unless P ≤ 0.05, and they are reported to three decimal places.

Results

Particle separation

Long-cut hay had a greater proportion of particles larger than 19 mm compared with short-cut hay (Table 2; P < 0.001), with 41% greater particle retention on the largest screen. DRC, Sweet Bran, and the dry vitamin/mineral supplement had no particles retained on the 19 mm sieve (P < 0.001). Long-cut hay and short-cut hay had similar proportions retained on the 8 mm sieve (P = 0.33) and exhibited the greatest proportion retained among the ingredients. For the 4 mm sieve, DRC had the greatest proportion of particles retained (P < 0.001). Particles greater than 4 mm (sum of particles retained on the top three sieves) were greatest for DRC and long-cut hay, which did not differ from one another (P = 0.99), followed by short-cut hay, then Sweet Bran (P < 0.001). Estimated peNDF using particle separation was greater for long-cut hay than short-cut hay (60.23, 51.06%, respectively; P < 0.001). Estimated mean particle sizes calculated using the PSPS were 11.43 and 7.68 mm for long-cut hay and short-cut hay, respectively (P < 0.001). However, these values should be interpreted with caution, as the calculation uses a default length for particles retained on the top sieve of 30 mm, which is likely not equal between hay sources.

Table 2.

Particle separation analysis and estimated physically effective NDF (peNDF) of feed ingredients.

	Ingredient
Item	Long-cut Hay1	Short-cut Hay2	DRC	Sweet Bran3	Supplement4	SEM5	P-value
NDF, % DM	75.30	73.79	9.51	35.78	8.57	–	–
	Retained/sieve, %
Sieve size, mm
19.0	43.47a	25.49b	0.00c	0.00c	0.00c	1.588	<0.001
8.0	23.53a	27.51a	5.15b	5.73bcx	0.18bcy	1.475	<0.001
4.0	12.99c	14.20c	75.17a	34.06b	4.62d	1.195	<0.001
Particles less than 4 mm	20.01d	32.80c	19.68d	60.21b	95.20a	2.016	<0.001
Particles greater than 4 mm	79.99a	67.20b	80.32a	39.79c	4.80d	2.016	<0.001
Estimated peNDF6, % DM	60.23a	51.06b	7.60y	18.16c	0.41x	1.196	<0.001
Mean particle size, mm	11.54a	7.68b	–	–	–	0.477	<0.001

Abbreviation: DRC = dry-rolled corn.

¹Long-cut prairie hay was passed through a commercial tub grinder equipped with 17.8 cm round hole screen.

²Short-cut prairie hay was passed through a commercial tub grinder equipped with 10.2 cm round hole screen.

³Sweet Bran (Cargill Inc., Dalhart, TX).

⁴Dry supplement was formulated to contain (% DM basis) 42.6% ground corn, 27.1% limestone, 20.6% wheat middlings, 6.5% urea, 0.9% salt, 0.49% magnesium oxide, 0.12% copper sulfate, 0.15% manganous oxide, 0.08% selenium, 0.47% zinc sulfate, 0.29% vitamin A (30,000 IU/g), 0.09% vitamin E (500 IU/g), 0.008% vitamin D (30,000 IU/g), 0.19% tylosin (Tylan-40, Elanco Animal Health, Greenfield, IN), and 0.30% monensin (Rumensin-90, Elanco Animal Health).

⁵Standard error of the mean. The most conservative SEM estimate is reported for each item (n = 12).

⁶Percent peNDF was estimated by multiplying the percent NDF (as a decimal) by the percentage of particles greater than 4 mm of the ingredient.

^a-eMeans within a row without a common superscript differ (P ≤ 0.05).

^x-yMeans within a row without a common superscript tend to differ (0.05 < P ≤ 0.10).

When mixed diets were separated using the PSPS, the long-cut TMR had a greater proportion of particles retained on the 19 mm sieve (Table 3; P < 0.001). For particles less than 4 mm, the short-cut TMR had the greatest proportion retained (P = 0.044). Particles greater than 4 mm were greatest for long-cut TMR (P = 0.044). Estimated peNDF using particle separation was greater for long-cut TMR than for short-cut TMR (P = 0.014), with estimates of 14.23% and 12.33%, respectively. However, when fiber-source ingredients were used to estimate peNDF, long-cut TMR tended to be greater than short-cut TMR (P = 0.07), with estimates of 8.38% and 7.48%, respectively. Estimates of peNDF using total-sample values were 1.7 and 1.6 times greater than those derived from fiber-sources alone for long-cut TMR and short-cut TMR, respectively.

Table 3.

Particle separation analysis and estimated physically effective NDF (peNDF) of dietary treatments.

	Dietary treatment1
Item	Long-cut TMR	Short-cut TMR	SEM2	P-value
NDF, % DM	22.79	20.97	–	–
	Retained/sieve, %
Sieve size, mm
19.0	5.48a	2.66b	0.484	<0.001
8.0	7.71	7.33	0.671	0.69
4.0	45.22	45.01	0.879	0.87
Particles less than 4 mm	41.59b	45.00a	1.126	0.044
Particles greater than 4 mm	58.41a	55.00b	1.126	0.044
Estimated pefNDF3, % DM	8.38	7.48	0.340	0.07
Estimated peNDF4, % DM	14.23a	12.33b	0.504	0.014

¹Long-cut TMR = 8% long-cut prairie hay; Short-cut TMR = 8% short-cut prairie hay.

²Standard error of the mean. The most conservative SEM estimate is reported for each item (n = 12).

³Percent pefNDF was estimated from fiber sources by multiplying the proportion of particles greater than 4 mm for each ingredient by the sample’s NDF concentration. This value was then multiplied by the ingredient’s dietary inclusion (DM basis). For long-cut TMR, pefNDF was calculated using long-cut hay (8% DM) and Sweet Bran (18% DM); short-cut TMR pefNDF was calculated using short-cut hay (8% DM) and Sweet Bran (18% DM).

⁴Percent peNDF was estimated by multiplying the percent NDF (as a decimal) by the percentage of particles greater than 4 mm of the diet.

^a-bMeans within a row without a common superscript differ (P ≤ 0.05).

Animal performance and carcass characteristics

Dietary treatment had no effect on SBW at any time point or on final SBW (P ≥ 0.18; Table 4). Average daily gain did not differ between dietary treatments during any period or overall (P ≥ 0.26). Steers fed the long-cut TMR tended to have greater DMI during d 0–21 compared with those fed the short-cut TMR (P = 0.09), but treatments did not differ during the other periods or overall (P ≥ 0.18). No differences in feed efficiency (G:F) were observed between dietary treatments during any period or overall (P ≥ 0.25). Dietary treatment had no effect on carcass-adjusted final SBW, ADG, or G:F (P ≥ 0.73; Table 4).

Table 4.

Effect of physically effective fiber on performance of crossbred beef steers.

	Dietary treatment1
Item	Long-cut TMR	Short-cut TMR2	SEM3	P-value
SBW4, kg
Day 0	398	392	7.34	0.53
Day 21	434	424	4.80	0.19
Day 42	477	467	5.30	0.19
Day 63	500	485	6.83	0.18
Day 84	533	526	7.31	0.49
Day 105	564	557	8.73	0.58
Final	589	578	8.56	0.39
ADG, kg
Days 0–21	1.69	1.55	0.178	0.58
Days 21–42	2.06	2.03	0.129	0.85
Days 42–63	1.07	0.89	0.176	0.49
Days 63–84	1.60	1.93	0.192	0.26
Days 84–105	1.46	1.48	0.115	0.90
Days 105–final	1.30	1.09	0.200	0.48
DMI, kg
Days 0–21	12.2	11.5	0.24	0.09
Days 21–42	12.1	11.9	0.58	0.63
Days 42–63	11.5	11.0	0.33	0.28
Days 63–84	10.8	11.5	0.34	0.18
Days 84–105	11.3	11.4	0.51	0.94
Days 105–final	11.6	12.1	0.51	0.53
G:F5, kg/kg
Days 0–21	0.138	0.132	0.0195	0.79
Days 21–42	0.169	0.172	0.0103	0.85
Days 42–63	0.093	0.079	0.0211	0.56
Days 63–84	0.147	0.168	0.0142	0.33
Days 84–105	0.129	0.130	0.0089	0.91
Days 105–final	0.115	0.092	0.0132	0.25
Final SBW6, kg	586	581	9.88	0.73
ADG7, kg/d	1.51	1.52	0.071	0.88
G:F8, kg/kg	0.130	0.132	0.0067	0.89

Abbreviations: ADG = average daily gain; DMI = dry matter intake; SBW = shrunk body weight.

¹Long-cut TMR = 8% long-cut prairie hay; Short-cut TMR = 8% short-cut prairie hay.

²d0 = short-cut TMR treatment started.

³Standard error of the mean. The most conservative SEM estimate is reported for each item (n = 5 pens per treatment).

⁴4% shrink applied.

⁵G:F = ratio of ADG to DMI.

⁶Carcass-adjusted final BW was calculated as HCW divided by the average dressing percent of all steers (62.75%).

⁷Carcass-adjusted ADG was calculated from carcass-adjusted final SBW, initial SBW, and days on feed (DOF = 124).

⁸Carcass-adjusted G:F was calculated from carcass-adjusted ADG divided by average DMI.

Dietary treatment had no effect on HCW, DP, marbling score, 12th-rib fat, longissimus muscle (LM) area, yield grade, USDA quality grade, or liver score (P ≥ 0.18; Table 5).

Table 5.

Effects of physically effective fiber on carcass characteristics.

	Dietary treatment1
Item	Long-cut TMR	Short-cut TMR	SEM2	P-value
HCW, kg	368	365	10.5	0.73
DP3	62.4	63.1	0.41	0.30
Marbling score4	440	482	21.5	0.18
12th-rib fat, cm	1.44	1.41	0.079	0.79
LM area, cm2	90.0	88.1	2.03	0.50
Calculated yield grade	2.93	2.98	0.172	0.84
Choice or greater, %	80.0	80.0	0.09	1.00
Select or less, %	20.0	20.0	0.09	1.00
Liver Score5, %
Normal	95.0	75.0	4.87	0.35
A + Adhesion	5.0	25.0	9.68	0.35
Abscessed	0.0	0.0	–	–

¹Long-cut TMR = 8% long-cut prairie hay; Short-cut TMR = 8% short-cut prairie hay.

²Standard error of the mean. The most conservative SEM estimate is reported for each item. Animal = experimental unit (n = 20 steers per treatment).

³DP = dressing percent: HCW/Final SBW.

⁴Slight0 = 300; Small0 = 400; Modest0 = 500; Moderate0 = 600.

⁵Trained personnel from Oklahoma State University collected liver abscess scores on the day of slaughter, using the Elanco Liver Check scoring system (Brink et al. 1990). Livers without abscesses were classified as normal. Livers classified as A + adhesion had adhesion to the diaphragm, gastrointestinal tract, or both, indicative of current or previous liver abscesses. A − and A + were combined and classified as Abscessed due to the lack of active abscesses.

Fecal characteristics

No treatment × day interactions were observed for FCS (P ≥ 0.25; Table 6), while sampling day affected FCS (P = 0.044). A treatment × day interaction was observed for fecal pH (P = 0.001; Figure 1). Fecal pH was greater in steers fed long-cut TMR on day 21, but on day 42, fecal pH was greater in short-cut TMR.

Table 6.

Effects of physically effective neutral detergent fiber on fecal characteristics and plasma metabolites.

Item	Dietary treatment1			P-value
	Long-cut TMR	Short-cut TMR	SEM2	Treatment	Day	Treatment × Day
Fecal characteristics
FCS	3.11	3.21	0.060	0.22	0.044	0.25
pH3	6.65	6.69	0.042	0.57	< 0.001	0.001
Plasma metabolites
Glucose, mg/dL	122.96	119.74	3.754	0.54	0.003	0.21
SAA4, µg/mL	39.15	32.03	3.011	0.09	< 0.001	0.89

Abbreviations: FCS = fecal consistency score; SAA = serum amyloid A.

¹Long-cut TMR = 8% long-cut prairie hay; Short-cut TMR = 8% short-cut prairie hay.

²Standard error of the mean. The most conservative SEM estimate is reported for each item. Animal = experimental unit (n = 5 pens per treatment).

³Interaction shown in Figure 1.

⁴Results for the day effect are displayed in Figure 2.

Figure 1.

Effect of physically effective neutral detergent fiber on fecal pH of crossbred beef steers. A treatment × day interaction (P = 0.001) and a main effect of day (P < 0.001) were observed, whereas dietary treatment alone did not affect fecal pH (P = 0.57).

Plasma metabolites

Dietary treatment had no effect on plasma glucose concentration, and no treatment × day interaction was observed (P ≥ 0.21; Table 6), while sampling day affected plasma glucose concentrations (P = 0.003). No treatment × day interactions were observed for SAA concentrations (P = 0.89). However, SAA increased from day 42 to day 63, then declined to values similar to those on day 0 (Day effect, P < 0.001). SAA concentrations tended to be greater in long-cut TMR (P = 0.09; Figure 2).

Figure 2.

Effect of physically effective neutral detergent fiber on plasma serum amyloid A (SAA) concentrations of crossbred beef steers. A main effect of day (P < 0.001) was observed, and treatments tended to differ (P = 0.09). No treatment × day interaction was observed (P = 0.89).

Digestibility

No treatment × day interactions were observed for fecal nutrient concentrations (P ≥ 0.64; Table 7), and no day effects were observed (P ≥ 0.15). Dietary treatment had no effect on fecal NDF or starch concentrations (P ≥ 0.59). No treatment × day interactions were observed for DMD, starch digestibility, or NDF digestibility (P ≥ 0.41). Dietary treatment had no effect on dry matter digestibility (DMD) or starch digestibility (P ≥ 0.62). However, NDF digestibility was greater for steers fed long-cut TMR (P < 0.001). A day effect was also observed for NDF digestibility (P = 0.013; Figure 3).

Table 7.

Effects of physically effective neutral detergent fiber on fecal NDF and starch concentrations and apparent total-tract digestibility of beef steers.

	Dietary treatment1			P-value
Item	Long-cut TMR	Short-cut TMR	SEM2	Treatment	Day	Treatment × Day
Fecal NDF, %	31.30	31.15	0.561	0.86	0.15	0.99
Fecal starch, %	17.65	18.61	1.458	0.59	0.30	0.64
Digestibility3, %
DM	48.07	47.58	1.233	0.78	0.13	0.41
NDF4	31.83	19.62	1.883	<0.001	0.013	0.46
Starch	83.36	82.40	1.624	0.62	0.40	0.70

Abbreviation: NDF = neutral detergent fiber digestibility.

¹Long-cut TMR = 8% long-cut prairie hay; Short-cut TMR = 8% short-cut prairie hay.

²Standard error of the mean. The most conservative SEM estimate is reported for each item (n = 5 pens per treatment).

³Total-tract digestibility determined using internal indigestible NDF marker concentration.

⁴Results for the effect of day are displayed in Figure 3.

Figure 3.

Effect of sampling day on neutral detergent fiber (NDF) digestibility of crossbred beef steers. NDF digestibility differed across sampling day (P = 0.013). Values represent least squares means ± SEM averaged across dietary treatments.

Rumination behavior

Daily rumination behavior of the steers consuming the two dietary treatments over the feeding period is shown in Figure 4. Average rumination time was 273 and 259 min/d for long-cut TMR and short-cut TMR, respectively, and was not affected by dietary treatment (P = 0.56; Table 8). Similarly, rumination expressed relative to intake of DM, NDF, peNDF, or pefNDF did not differ between treatments (P ≥ 0.28). Tendencies for treatment × day interactions were observed for rumination expressed relative to DM, NDF, peNDF, and pefNDF intake (P ≤ 0.08; Figure 4). Because temporal response patterns were similar across denominators, suggesting that the observed interaction likely reflected normal variation over time rather than a consistent treatment effect.

Figure 4.

Effect of physically effective neutral detergent fiber (peNDF) on rumination behavior in finishing beef steers expressed as (A) minutes per day, (B) minutes per kg of dry matter (DM) consumed, and (C) minutes per kg of peNDF consumed. Long-cut TMR = 8% long-cut prairie hay (-●-); short-cut TMR = 8% short-cut prairie hay (-◼-).

Table 8.

Effect of physically effective neutral detergent fiber on rumination time per day and per kilogram of DM, NDF, and peNDF consumed.

	Dietary treatment1			P-value
Item	Long-cut TMR	Short-cut TMR	SEM2	Treatment	Day	Treatment × Day
Rumination, min/d	272.5	258.9	16.11	0.56	<0.001	0.46
Rumination, min/kg
DM	23.6	22.3	1.21	0.49	<0.001	0.08
NDF	103.5	106.5	5.41	0.69	<0.001	0.07
peNDF3	165.8	181.2	8.78	0.24	<0.001	0.06
pefNDF4	281.5	298.7	14.81	0.43	<0.001	0.07

¹Long-cut TMR = 8% long-cut prairie hay; Short-cut TMR = 8% short-cut prairie hay.

²Standard error of the mean. The most conservative SEM estimate is reported for each item (n = 12).

³Percent peNDF was estimated by multiplying the percent NDF (as a decimal) by the percentage of particles greater than 4 mm of the diet.

⁴Percent pefNDF was estimated from fiber sources by multiplying the proportion of particles greater than 4 mm for each ingredient by the sample’s NDF concentration. This value was then multiplied by the ingredient’s dietary inclusion (DM basis). For long-cut TMR, pefNDF was calculated using long-cut hay (8% DM) and Sweet Bran (18% DM); Short-cut TMR pefNDF was calculated using short-cut hay (8% DM) and Sweet Bran (18% DM).

Discussion

This study evaluated the effects of reducing forage particle size, thereby altering peNDF estimates, on digestion, rumination, and performance of finishing beef steers. Although forage processing modified the particle size distribution and reduced dietary peNDF, performance, carcass characteristics, and rumination time were not affected by dietary treatment. However, apparent total tract NDF digestibility was reduced with a decrease in forage particle size. Additionally, fecal characteristics and plasma metabolites varied across sampling days or were associated with environmental conditions rather than dietary treatment. The following discussion integrates particle separation and peNDF estimation with observed responses in digestion, rumination, fecal characteristics, and metabolic indicators to determine how forage particle size influences biological outcomes in high-concentrate finishing diets.

Particle separation and peNDF estimation

Differences in forage processing altered estimates of peNDF despite similar chemical composition and forage inclusion across treatments. Long-cut hay contributed greater peNDF than short-cut hay, indicating that forage particle size alone influenced peNDF supply at the ingredient level. However, these differences occurred in a high-concentrate finishing diet with limited forage inclusion, which is important when assessing the extent to which forage particle size differences influence dietary peNDF.

At the ingredient level, forage sources contributed greater peNDF than non-forage ingredients, due to their greater NDF concentration and a higher proportion of particles greater than 4.0 mm. Although DRC retained a substantial proportion of particles exceeding the 4.0 mm threshold, its contribution to peNDF is limited because of its low NDF concentration. Therefore, estimating dietary peNDF requires considering both NDF concentration and the physical characteristics of individual ingredients (Mertens 1997). Estimation of non-fibrous ingredients from the complete mixed diet may also include non-fibrous components retained with larger particles that have limited capacity to support ruminal mat formation or stimulate chewing activity (Woodford and Murphy 1988).

Processing-induced differences in forage particles were sufficient to alter particle size distribution between hay types, but the magnitude of these differences was attenuated at the dietary level. In the present study, 80% and 67% of particles exceeded the 4.0 mm threshold for long-cut and short-cut prairie hay, respectively, indicating a detectable reduction in physically effective particles in the short-cut hay. However, because forage accounted for only 8% of dietary DM, the particle size distribution of the mixed diets differed only slightly. As a result, differences in dietary peNDF were minimal, with only a 1.9% difference (DM basis) when evaluated at the diet level. Although the peNDF concentration differed statistically, the numerical differences were small and likely insufficient to produce biologically meaningful changes in chewing activity.

Previous studies evaluating the effects of forage particle size on dietary peNDF in finishing cattle have reported variable responses, likely reflecting differences in forage source and inclusion. Reductions in forage particle size decreased dietary peNDF when forage inclusion was 15% DM (Llonch et al. 2020) and 10% DM (Weiss et al. 2017). However, when forage was included at 5% DM, particle size manipulation did not alter dietary peNDF (Gentry et al. 2016). At the diet level, non-fibrous ingredients contributed substantially to particle size distributions due to high inclusion of DRC, potentially inflating peNDF estimates derived from the total-diet particle distribution and NDF concentrations. In the present study, peNDF estimates derived from total-diet values were 1.7 and 1.6 times greater than those derived from fiber-source ingredients alone for long-cut and short-cut TMR, respectively. Comparable inflation of peNDF estimates has been reported in other finishing diets, where total-diet estimates exceeded fiber-source estimates by 1.9- to 3.9-fold (Pereira et al. 2023). In the present study, dietary treatments were similar when peNDF estimates were derived from fiber-source ingredients alone. The discrepancy between analytical estimates may reflect the absence of differences between treatments in animal-derived data. These findings suggest that total-diet peNDF values may overestimate biologically effective fiber supply in high-concentrate finishing diets.

Taken together, the present peNDF results provide mechanistic support for the lack of dietary treatment effects observed in rumination, performance, and carcass characteristics. Although forage particle size altered total-diet peNDF estimates, the magnitude of these differences was insufficient to meaningfully affect feedlot performance in finishing beef cattle. These findings suggest that total-diet peNDF estimation methods may overrepresent biologically effective fiber in diets with minimal forage inclusion. Consequently, peNDF estimation approaches that rely solely on fiber-source ingredients may more accurately reflect biologically meaningful fiber supply in high-concentrate finishing diets.

Animal performance and carcass characteristics

Reducing forage particle size from 11.43 to 7.68 mm, without altering forage inclusion did not affect growth performance or carcass characteristics of finishing beef steers in the present study. Although steers fed the long-cut TMR tended to exhibit greater DMI early in the feeding period, this difference did not persist over time or result in cumulative differences in growth performance. The tendency observed in DMI is likely attributed to the dietary transition experienced by steers fed the short-cut TMR at d 0, whereas steers fed the long-cut TMR remained on the same diet. All steers were adapted to the long-cut TMR and consumed this diet for approximately 40 d prior to treatment initiation. Consequently, steers assigned to the short-cut TMR underwent an additional adaptation period. The transient tendency toward altered DMI in short-cut TMR may therefore reflect short-term dietary adaptation rather than a sustained treatment effect. Despite this, the similarity between live-weight and carcass-adjusted performance metrics suggests forage particle size manipulation alone was insufficient to alter intake or growth performance over the finishing period.

The lack of performance differences observed in the present study is consistent with previous research in which manipulation of forage particle size did not alter growth performance or carcass characteristics in finishing beef cattle. No differences in ADG, DMI, or carcass traits were observed when corn stalks differing in particle size were included at 5% DM (Gentry et al. 2016). Similarly, the coarseness of sudangrass hay in finishing diets did not affect steer performance (Calderon-Cortes and Zinn 1996). Collectively, these findings suggest that when forage inclusion is limited, reductions in forage particle size may not strongly affect growth performance.

Carcass characteristics were likewise unaffected by dietary treatment in the present study. Similar marbling scores and quality grade distributions indicate comparable energy availability and partitioning toward intramuscular fat deposition between treatments. These findings align with studies reporting minimal carcass responses to forage particle size manipulation in high-concentrate finishing diets (Calderon-Cortes and Zinn 1996; Gentry et al. 2016). In contrast, increased marbling and dressing percentage have been reported in steers fed diets with greater peNDF concentration (Pereira et al. 2021). However, those diets were barley-based and differed in forage inclusion and source, including varying proportions of barley silage, wheat silage, and wheat straw. Differences in diet composition and forage inclusion may therefore account for the divergent carcass response observed between studies.

Liver abscess incidence in the present study did not differ between treatments, indicating comparable ruminal health. Adequate fiber supply in both diets likely contributed to the absence of differences in liver scores, despite the reduction in forage particle size. However, liver adhesion incidence was numerically greater in steers fed short-cut TMR compared to those fed long-cut TMR, which may warrant further investigation. Interpretation of this finding should be made cautiously, as there was no statistical evidence of a treatment effect, and adhesions may have resulted from processing-related factors rather than underlying liver pathology. Although the sample size of the present study may have limited the ability to detect small differences in liver abscess prevalence, the lack of accompanying changes in rumination, feedlot performance, or systemic inflammatory markers suggests that biologically meaningful differences were unlikely.

Overall, these findings indicate that reducing forage particle size does not compromise growth performance, carcass characteristics, or liver abscess incidence in finishing beef steers when forage inclusion is held constant. The absence of performance responses supports the claim that peNDF estimation from fiber-source ingredients may more accurately reflect meaningful effective dietary fiber concentrations in finishing beef cattle diets. These results align with feedlot studies conducted at similarly low forage inclusion rates and support the conclusion that forage particle-size reductions of the magnitude evaluated in the present study can be implemented without negative effects in high-concentrate finishing diets.

Fecal characteristics

Indicators of gut function remained within physiologically normal ranges despite differences in forage particle size. FCS was not affected by dietary treatment, and no treatment × day interaction was observed. However, FCS varied across sampling days. Despite this temporal variation, FCS values remained within a narrow and physiologically normal range throughout the finishing period. An FCS of 3 is considered the ideal consistency in finishing beef cattle (Woolsoncroft et al. 2018). An optimal FCS was generally maintained across all sampling days, indicating both diets provided sufficient effective fiber to support adequate digestion. The absence of a treatment × day interaction further indicates that fecal consistency did not diverge between treatments at specific time points.

Fecal pH exhibited a strong effect of sampling day and a treatment × day interaction, indicating time-dependent variation in hindgut acidity. Differences between treatments were detected on d 21 and d 42, whereas fecal pH did not differ on other sampling days. Therefore, a sustained treatment effect of forage particle size on fecal pH was not observed. The direction of treatment differences varied over time, with steers fed the long-cut TMR exhibiting greater fecal pH on d 21 and steers fed the short-cut TMR exhibiting greater fecal pH on d 42. The lower fecal pH observed in short-cut TMR on d 21 may reflect short-term physiological responses associated with dietary adaptation following treatment initiation, as these steers changed diets while steers fed the long-cut TMR remained on the same diet. This interpretation is consistent with the tendency for differences in DMI observed during the early feeding period, further supporting the conclusion that early fecal pH differences were transient and related to diet adaptation rather than a persistent effect of forage particle size. In addition to dietary adaptation, within-day feed sorting behavior may have contributed to the transient differences observed in fecal pH and fecal consistency. Reduction in forage particle size increases diet uniformity and is assumed to reduce sorting ability (Devries et al. 2008). Therefore, cattle fed long-cut TMR may have been able to selectively consume larger fibrous particles to a greater extent than those fed short-cut TMR, where particle size reduction may have limited sorting opportunities. Such selective intake could lead to greater within-day variability in nutrient intake, potentially influencing both metabolic and inflammatory responses. Additionally, under group-feeding conditions, cattle with greater or earlier access to the feed bunk may preferentially consume the limited forage fraction, resulting in heterogeneity in diet composition within individual animals across the same day and among animals within a pen (DeVries et al. 2014). Although the current study used a slick bunk protocol, this practice does not eliminate the potential for within-day sorting behavior. Sorting may be particularly relevant in high-grain finishing diets with a limited proportion of long or medium particle fractions (Yang and Beauchemin 2006; DeVries et al. 2014). Previous work in dairy cattle has demonstrated that sorting behavior increases as forage inclusion decreases (Leonardi and Armentano 2003; Devries et al. 2008). Selective consumption of fibrous particles, which may influence post-ruminal nutrient flow and hindgut patterns, could result in lower fecal pH in the long-cut TMR following diet adaptation. Although feed sorting behaviors were not directly measured in the present study, differential sorting between treatments could contribute to short-term variation in hindgut indicators without resulting in sustained digestive disruption. It is important to note that fecal pH integrates multiple physiological processes, including ruminal digestion, post-ruminal nutrient flow, and hindgut fermentation, and therefore cannot be used to isolate the specific site of digestive response.

Taken together, FCS and fecal pH indicate that differences in forage particle size resulted in limited, time-specific variation in hindgut indicators that remained within normal physiological ranges, with no evidence of persistent digestive disturbance or adverse effects on animal health.

Digestibility

Forage particle size did not alter starch or NDF concentrations of the feces, indicating that forage particle size did not affect fecal excretion of undigested starch or fiber. The lack of treatment differences in fecal starch concentration, together with similar total-tract starch digestibility, suggests that reducing peNDF through forage particle size reduction did not compromise starch digestion. Total-tract starch digestibility observed in the present study was lower than values commonly reported for DRC-based finishing diets. In steers fed DRC-based diets, starch digestibility frequently exceeded 90% (Owens and Soderlund 2006; Luebbe et al. 2012; Coulson et al. 2023). However, starch digestibility values comparable to those observed in the present study have also been reported, including values of 85.6% and 89.7% in finishing beef cattle fed DRC-based diets (Zinn et al. 1998; Corona et al. 2006). Differences among studies likely reflect variation in diet composition, extent of grain processing, and methodological approaches used to estimate digestibility. Importantly, dietary treatment did not affect starch digestibility, fecal starch concentration, or DM digestibility, indicating that differences in forage particle size were insufficient to produce a biological difference in starch and DM digestibility in steers fed a DRC-based finishing diet.

In contrast, total-tract NDF digestibility was greater for steers fed the long-cut TMR compared to the short-cut TMR. The difference in NDF digestibility was consistent throughout the finishing period and was not driven by transient sampling effects. Because apparent total-tract DM digestibility and starch digestibility were unaffected by treatment, improvements in fiber digestibility occurred independently. Despite greater NDF digestibility, NDF concentration of the feces did not differ between treatments. Therefore, increased forage particle size improved the extent of fiber digestion without decreasing fecal excretion of undigested fiber. Improved NDF digestibility in long-cut TMR is consistent with increased dietary peNDF concentrations. Increasing forage particle size is often expected to support ruminal fiber mat formation and conditions favorable to fibrolytic activity, thereby improving fiber degradation. However, rumination time did not differ between treatments in the present study, suggesting that the observed improvement in NDF digestibility occurred without a detectable change in rumination behavior. In addition, hindgut indicators did not indicate persistent digestive disruption, as FCS remained within a narrow physiological range and fecal pH differences were transient and time-specific. Therefore, larger forage particle size improved total-tract NDF digestibility but did not affect DMI, growth performance, or indicators of gut function. As expected in high-concentrate diets, total-tract NDF digestibility values were relatively low. Previous studies have reported NDF digestibility values ranging from approximately 20%–40% in cattle fed high-concentrate diets (Owens and Soderlund 2006; Luebbe et al. 2012). In this context, the greater NDF digestibility observed with the long-cut TMR indicates a detectable improvement in fiber digestion, although it did not translate into performance differences.

In the present study, forage inclusion was held constant between dietary treatments, allowing the effects of forage particle size to be evaluated independently. Under these conditions, reducing peNDF by decreasing forage particle size decreased fiber digestibility without altering starch digestion, intake, or growth performance. Thus, forage particle size influenced fiber digestion, but the magnitude of the response under low forage inclusion was insufficient to produce consistent changes in other biological endpoints measured in this experiment.

Plasma metabolites

Mean glucose concentrations in the present study consistently ranged from 115 to 130 mg/dL, which is greater than values commonly reported for finishing beef steers. In finishing beef cattle, plasma glucose concentrations typically range between 65 and 101 mg/dL (Evans et al. 1975; Hancock et al. 1988; Kolath et al. 2006), although a broader range (62–279 mg/dL) has been reported (LaBrune et al. 2000). Thus, while glucose concentrations in the present study were elevated relative to many reports, they remained within the broader physiological range observed in beef cattle. Elevated plasma glucose concentrations may reflect the metabolic consequences of high-energy, starch-rich diets, which increase ruminal propionate production and hepatic gluconeogenesis. In finishing diets, insulin resistance has been reported, with increased days on feed associated with reduced glucose efficiency and declines in gain efficiency (Joy et al. 2017; de Sousa et al. 2022). Insulin was not measured in the present study, but this metabolic state may contribute to elevated circulating glucose concentrations. However, a decline in gain efficiency was not observed in the present study.

Aside from dietary and metabolic factors, elevated plasma glucose concentrations may also result from stress-related physiological responses. Activation of the sympathetic nervous system and the hypothalamic-pituitary-adrenal axis results in the release of cortisol and catecholamines, which stimulate gluconeogenesis and glycolysis while inhibiting glycogenesis, ultimately increasing circulating glucose availability (Minka and Ayo 2009). Transportation and handling stress have been shown to elevate plasma glucose concentrations in cattle by increasing hepatic glycogenolysis or by depleting skeletal muscle glycogen reserves (Averós et al. 2008). The cattle in the present study exhibited an excitable disposition, making it plausible that handling during sampling induced a stress response, reflected by elevated SAA and plasma glucose concentrations outside the expected range.

Serum amyloid A concentrations differed over time but followed a similar pattern across treatments, suggesting that forage particle size did not meaningfully influence systemic inflammatory responses. Although steers fed long-cut TMR tended to exhibit greater SAA concentrations, this difference was small and not accompanied by changes in growth performance.

The increase in SAA concentrations observed on d 63 likely reflects a transient, nonspecific acute phase response to handling rather than diet-induced inflammation. Reported SAA concentrations associated with subacute ruminal acidosis (SARA) in dairy cattle range from 160 to 500 µg/mL (Gozho et al. 2005, 2006). In cattle with acute inflammation, the average SAA concentration was 74.3 µg/mL (Horadagoda et al. 1999). In beef cattle adapting to a finishing diet, peak SAA concentrations were 20 µg/mL (Ametaj et al. 2009). The SAA concentrations in the present study were substantially lower than those reported in cattle experiencing SARA, but greater than those in steers adapting to a finishing diet, with peak concentrations of 54.2 and 48.6 µg/mL observed in steers fed long-cut TMR and short-cut TMR, respectively. On d 53 through d 57, two steers from each pen were selected for additional measures not included in this manuscript, which included repeated animal restraint and handling. Additionally, on d 63 several pens were pulled up, but due to a mechanical malfunction of the chute, all samples were moved and conducted on d 64, adding an additional handling event. Acute-phase proteins, such as SAA, are sensitive to a range of stressors, including handling and immune stimulation, and are not necessarily indicative of diet-induced inflammation. The transient increase in SAA was observed in both treatments and does not suggest a treatment-driver inflammatory response. Therefore, it is plausible that the additional handling events prior to sample collection on d 64 triggered an acute-phase response, resulting in elevated SAA concentrations. Although SAA concentrations varied over time, a similar temporal pattern was observed in both treatments, with only a small treatment effect. Despite this, concentrations in the present study remained well below values commonly reported in cattle experiencing SARA or major inflammatory responses. Furthermore, average SAA concentrations did not differ between steers with liver adhesions and non-affected steers, supporting the interpretation that the observed adhesions were not associated with a measurable systemic inflammatory response and may have arisen from processing-related factors. Collectively, these findings indicate that reducing forage particle size did not elicit a sustained systemic inflammatory response and that temporal changes in SAA likely reflect short-term, nonspecific physiological responses rather than diet-induced inflammation.

Although dietary treatment did not significantly affect plasma glucose or SAA concentrations, steers fed the long-cut TMR tended to have greater SAA concentrations than those fed the short-cut TMR throughout the feeding period. This pattern was unexpected, as it was hypothesized that the reduced forage particle size in short-cut TMR would increase gastrointestinal irritation and thereby elevate systemic inflammatory markers. One potential explanation for this pattern is differential feed-sorting behavior across treatments. Reduction of forage particle size improves diet uniformity and is generally assumed to reduce the opportunity for selective consumption of fibrous particles (Devries et al. 2008). Feed sorting behavior has been documented in finishing beef cattle, with heifers reported to sort against fractions low in starch and high in fiber while preferentially consuming fractions with intermediate starch and fiber concentrations (DeVries et al. 2014). In high-concentrate finishing diets with minimal forage inclusion, even small differences in selective intake may result in greater short-term variation in nutrient patterns, potentially influencing inflammatory and metabolic markers. Although feed bunks were assessed using a slick bunk protocol, this management strategy does not eliminate within-day sorting under group-feeding conditions. Cattle with greater or earlier access to the feed bunk may preferentially consume the fraction with less forage, resulting in heterogeneity in nutrient intake both within individual animals throughout the day and among animals within a pen (DeVries et al. 2014). Such variability in intake could contribute to numerically higher plasma glucose and SAA concentrations in steers fed the long-cut TMR, along with reduced fecal pH, without producing sustained changes in performance. However, because sorting behavior was not directly measured in the present study, this interpretation remains speculative.

Collectively, differences in feed sorting behavior and handling-related stress may contribute to subtle differences in metabolic and inflammatory status between treatments, despite the absence of statistically significant treatment effects. However, because sorting behavior was not directly measured and the limited statistical power of the present study, this interpretation remains speculative.

Rumination behavior

The lack of dietary treatment effect on rumination as measured by ear-tag accelerometers suggests that reducing forage particle size, while maintaining identical nutrient composition, did not meaningfully alter the physical stimulation required to drive rumination in finishing beef cattle. Rumination is primarily driven by dietary physical characteristics, particularly particle size and structural fiber content, and time spent ruminating is generally proportional to cell wall intake (Van Soest 1994). Based on rumination measurements derived from ear-tag accelerometers, the magnitude of forage particle size reduction evaluated was insufficient to produce a detectable difference in rumination time.

Average rumination time observed in the present study was 266 and 251 min/d for long-cut TMR and short-cut TMR, respectively, which is consistent with values in beef cattle fed finishing diets using different measurement approaches. Average rumination time reported in finishing beef cattle has been reported to range from 225 to 313 min/d when measured using collars equipped with sensory microphones (Gentry et al. 2016; Weiss et al. 2017). Studies using video observation reported comparable rumination times of 246–316 min/d in beef heifers (Chibisa et al. 2020). Greater rumination times, ranging from 319 to 477 min/d, have also been reported in beef heifers fed finishing diets, using video observation (Llonch et al. 2020). However, the diets in the study contained higher forage inclusion and peNDF concentrations than those in other studies. The ear-tag accelerometers in the present study indirectly measured rumination, but the observed rumination time aligns with that reported in previous studies using both collar-based sensors and video-based observation methods.

Although total-diet peNDF concentrations differed between treatments, these differences did not translate into behavioral differences. This disconnect supports the interpretation that non-fibrous ingredients can substantially contribute to peNDF estimates derived from total-diet values in finishing beef cattle diets, potentially inflating estimates of biologically effective fiber. In contrast, estimating peNDF from fiber-source ingredients yielded smaller treatment differences, which were consistent with the absence of detectable differences in rumination, intake, and performance. These findings support the interpretation that fiber-source peNDF estimates may better reflect biologically meaningful physically effective fiber in high-concentrate finishing diets. Taken together, these results indicate that rumination behavior was not affected by forage particle size in finishing beef cattle.

Conclusion

These findings indicate that reducing mean forage particle size from 11.43 mm to 7.68 mm in a DRC-based finishing diet containing peNDF concentrations above the minimum recommended by the National Academies of Sciences and Medicine (2016) decreased apparent total tract NDF digestibility but did not affect growth performance, rumination behavior, or carcass characteristics. In high-concentrate finishing diets with minimal forage inclusion, peNDF estimates derived from total-diet metrics may overrepresent biologically effective fiber because of substantial contributions from non-fibrous ingredients. Instead, peNDF estimation based on fiber-source ingredients may be more closely aligned with observed animal responses and provide greater biological relevance under these feeding conditions. Collectively, these findings suggest that reducing forage particle size to improve diet uniformity may be a viable strategy in finishing beef cattle diets without markedly compromising performance or carcass outcomes.

Glossary

Abbreviations

ADG

average daily gain

AIC

Akaike information criterion

BIC

Bayesian information criterion

BW

body weight

DM

dry matter

DMD

dry matter digestibility

DMI

dry matter intake

DP

dressing percent

DRC

dry rolled corn

DS

dry supplement

EE

ether extract

FCS

fecal consistency score

G:F

gain to feed

HCW

hot carcass weight

iNDF

indigestible neutral detergent fiber

LM

longissimus muscle

NDF

neutral detergent fiber

NEg

net energy for gain

NEm

net energy for maintenance

pefNDF

physically effective neutral detergent fiber from forage sources

peNDF

physically effective neutral detergent fiber

PSPS

Penn State Particle Separator

SAA

serum amyloid A

SARA

subacute ruminal acidosis

SB

Sweet Bran

SBW

shrunk body weight

TDN

total digestible nutrients

Author contributions

Mallory E. LeDoux (Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Writing—original draft, Writing—review & editing), Ryan Reuter (Formal analysis, Methodology, Resources, Visualization, Writing—review & editing), David Lalman (Formal analysis, Investigation, Methodology, Resources, Writing—review & editing), and Andrew Foote (Conceptualization, Data curation, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Writing—review & editing)

Conflict of interest statement. The authors declare no conflicts of interest.

Data availability

Data are available from the corresponding author on request.