Effects of bismuth subsalicylate and encapsulated calcium-ammonium nitrate on feedlot beef cattle production

Abstract

Two experiments were performed to evaluate the effects of bismuth subsalicylate (BSS) and calcium-ammonium nitrate (CAN) on in vitro ruminal fermentation, growth, apparent total tract digestibility of nutrients, liver mineral concentration, and carcass quality of beef cattle. In Exp. 1, four ruminally cannulated steers (520 ± 30 kg body weight [BW]) were used as donors to perform a batch culture and an in vitro organic matter digestibility (IVOMD) procedure. Treatments were arranged in a 2 × 2 factorial with factors being BSS (0 or 0.33% of substrate dry matter [DM]) and CAN (0 or 2.22% of substrate DM). In Exp. 2, 200 Angus-crossbred steers (385 ± 27 kg BW) were blocked by BW and allocated to 50 pens (4 steers/pen) in a randomized complete block design with a 2 × 2 + 1 factorial arrangement of treatments. Factors included BSS (0 or 0.33% of the diet DM) and nonprotein nitrogen (NPN) source (urea or encapsulated CAN [eCAN] included at 0.68% or 2.0% of the diet, respectively) with 0.28% ruminally available S (RAS). A low S diet was included as a positive control containing urea (0.68% of DM) and 0.14% RAS. For Exp. 1, data were analyzed using the MIXED procedure of SAS with the fixed effects of BSS, CAN, BSS × CAN, and the random effect of donor. For Exp. 2, the MIXED procedure of SAS was used for continuous variables and the GLIMMIX procedure for categorical data. For Exp. 1, no differences (P > 0.230) were observed for IVOMD. There was a tendency (P = 0.055) for an interaction regarding H₂S production. Acetate:propionate increased (P = 0.003) with the addition of CAN. In Exp. 2, there was a NPN source effect (P = 0.032) where steers consuming urea had greater carcass-adjusted final shrunk BW than those consuming eCAN. Intake of DM (P < 0.001) and carcass-adjusted average daily gain (P = 0.024) were reduced by eCAN; however, it did not affect (P = 0.650) carcass-adjusted feed efficiency. Steers consuming urea had greater (P = 0.032) hot carcass weight, and a BSS × NPN interaction (P = 0.019) was observed on calculated yield grade. Apparent absorption of S decreased (P < 0.001) with the addition of BSS. Final liver Cu concentration was reduced (P = 0.042) by 58% in cattle fed BSS, indicating that BSS may decrease Cu absorption and storage in the liver. The results observed in this experiment indicate that BSS does not have negative effects on feedlot steer performance, whereas CAN may hinder performance of steers fed finishing diets.

Introduction

Sulfur is an important element for the ruminant animal due to its role in bacterial growth and metabolism (Drewnoski et al., 2014). Despite its importance, concentrations of dietary S above NASEM (2016) recommendations [0.15% dry matter (DM) basis] are related to reductions in growth performance, trace mineral absorption, and the incidence of polioencephalomalacia, with the latter being primarily caused by the reduction of S to H₂S in the rumen and its effect at a cellular level within the brain (Gould, 1998; Drewnoski et al., 2014). Bismuth subsalicylate (BSS) has been extensively used in humans as a H₂S mitigator to alleviate gastrointestinal disorders (Suarez et al., 1998). Using BSS as a ruminal H₂S mitigator only occurred recently when Ruiz-Moreno et al. (2015) studied the effects of various concentrations of BSS on in vitro fermentation of a high concentrate substrate.

Nitrate was used in ruminants primarily as a potential compound to mitigate enteric CH₄ due to a H⁺-sinking mechanism that occurs in the rumen (Newbold et al., 2014; van Zijderveld et al., 2010); however, data have indicated that CAN inhibits H₂S production, theoretically in the same fashion as CH₄ (Henry et al., 2021). The combination of BSS and nitrate to alleviate negative effects caused by S and H₂S has not been extensively investigated (Henry et al., 2021). It was hypothesized that, by adding encapsulated calcium-ammonium nitrate (eCAN) and BSS to the diet of beef cattle, any negative effects observed when cattle are consuming a diet with greater ruminal available S (RAS) would be reduced, improving growth performance, apparent total tract digestibility of nutrients, liver mineral absorption, and carcass quality. Therefore, two experiments were conducted to determine the effects of calcium ammonium nitrate (CAN) and BSS, on in vitro ruminal fermentation and growth performance, apparent total tract digestibility of nutrients, liver mineral concentration, and carcass quality of feedlot beef steers.

Materials and Methods

All experimental procedures involving live animals were conducted at the Texas Tech University—New Deal Research and Education Center located in Idalou, TX, and were approved by the Texas Tech Institutional Animal Care and Use Committee (IACUC-19067/07 and IACUC-T18079).

Experiment 1

Experimental design, substrate, and treatments

In vitro batch culture incubations were conducted as a randomized complete block design with a 2 × 2 factorial arrangement of treatments. Ruminal fluid donors were considered as block, and block was included as a random effect. Factors included BSS (0 or 0.33% of substrate DM) and CAN (0 or 2.22% of substrate DM; Henry et al., 2021). The concentration of CAN was chosen to provide the same amount of N as 0.74% of urea in the substrate DM. Treatments were labeled as follows: CTL (urea at 0.74%), CAN (CAN at 2.22%), CTLB (urea at 0.74% and BSS at 0.33%), and CANB (CAN at 2.22% and BSS at 0.33%). The source of nitrate utilized in this experiment was double salt of calcium-ammonium nitrate decahydrate [5Ca (NO₃)₂·NH₄NO₃·10H₂O; 65.1% NO₃; 15.5% N; Calcinit, Yara, Oslo, Norway]. All treatments were formulated to be isonitrogenous by adding urea.

The substrate was formulated as a commercial high-concentrate diet with 87% concentrate on a DM basis (Table 1). Prior to mixing of the substrate, all ingredients were dried separately in a forced-air oven at 55 °C for 24 h. Urea and CAN were dissolved in reverse osmosis filtered water and the nonprotein nitrogen (NPN) solutions were added to the incubations at 100 μL per bottle/tube at the moment of incubation. Bismuth subsalicylate was incorporated into the substrate during mixing.

Table 1.

Ingredient and nutrient composition of substrate and treatments used in Exp. 1

Item	Treatment1
	CTL	CAN	CTLB	CANB
Ingredient, % DM
Steam-flaked corn	62.35	60.87	62.03	60.55
Wet corn gluten feed	28.45	28.45	28.44	28.44
Cottonseed hulls	7.16	7.16	7.16	7.16
Limestone	1.30	1.30	1.30	1.30
Urea	0.74	–	0.74	–
CAN	–	2.22	–	2.22
BSS	–	–	0.33	0.33
Nutrient composition2
DM, % as fed	88.70
CP, % DM	13.52
OM, % DM	96.20
NDF, % DM	17.70
ADF, % DM	7.20
TDN, % DM	81.00
NEm, Mcal/kg	2.01
NEg, Mcal/kg	1.36
Ca, % DM	0.66
P, % DM	0.37
Mg, % DM	0.23
K, % DM	0.76
S, % DM	0.21
Mo, ppm	<1.00

¹CTL, Control (urea at 0.74%); CAN, Nitrate (CAN at 2.22%); CTLB, Urea-Bismuth (urea at 0.74% and BSS at 0.33%); CANB, Nitrate-Bismuth (CAN at 2.22% and BSS at 0.33%).

²Nutrient composition analysis was performed on the basal substrate only, since CAN and urea were added to the in vitro system upon incubation time. Analyzed by a commercial laboratory using a wet chemistry package (Dairy One, Ithaca, NY).

In vitro incubations

In vitro organic matter digestibility (IVOMD) and batch culture procedures were performed simultaneously using the same substrate, treatments, and inoculum. Four ruminally cannulated Angus-crossbred steers (520 ± 30 kg body weight [BW]; average BW ± SD) were used as donors. Ruminal fluid was extracted from each donor and kept in separate prewarmed containers. After collection, the containers with ruminal fluid were kept at 39 °C and promptly transported to the Sustainable Ruminant Nutrition Laboratory located at Texas Tech University Animal and Food Sciences Building, where they were separately combined with McDougall’s buffer to form a 2:1 buffer to ruminal fluid mixture to be used as inoculum (1 inoculum per steer; McDougall, 1948).

For the IVOMD determination, a modified Tilley and Terry (1963) procedure was used where 0.2 g of substrate and 20 mL of inoculum were incubated in 100-mL plastic centrifuge tubes at 39 °C under constant agitation (60 rpm) for 24 h. Two blank tubes were also incubated without added substrate to correct for organic matter (OM) from ruminal fluid, and all tubes were capped with a rubber stopper equipped with an 18-G needle that allowed the gases of fermentation to escape. After 24 h of incubation, abomasal digestion was simulated by the addition of 6 mL of 20% HCl and 2 mL of a 5% pepsin solution into the tubes and an additional incubation for 48 h at 39 °C under constant agitation (60 rpm). Once incubation was completed, tubes were removed from the incubator and contents were filtered through previously labeled and weighed P8 filters (Fisher Scientific, Pittsburgh, PA). Filters were then folded and placed in a previously weighed ceramic crucible before being dried at 100 °C for 24 h to record hot weights. Dry filters were combusted in a muffle furnace at 600 °C for 6 h before placing them back at 100 °C for at least 16 h to record hot weights to calculate IVOMD.

The batch culture consisted of a similar procedure as described for IVOMD with the difference that batch culture bottles did not undergo a postruminal digestion step. For the batch culture incubations, 0.2 g of substrate plus 20 mL of a 2:1 buffer to ruminal fluid were added into a 125-mL serum bottle. The bottles were capped and incubated for 24 h at 39 °C under constant agitation (60 rpm). After 24 h, bottles were taken out of the incubator and placed on ice for 15 min to stop microbial fermentation. Bottles were placed in ice for 15 min to stop fermentation and allowed to reach room temperature before measurement of total gas production (TGP). A 10-mL gas sample for CH₄ and another 5 mL for H₂S production were taken before the bottle content was extracted (inoculum plus substrate). After gas measurement and collection, bottles were opened, contents were poured into a plastic cup, final pH of the inoculum was recorded, and 0.2 mL of H₂SO₄ solution were added into each cup to preserve volatile fatty acids (VFA). Finally, two aliquots were stored at −20 °C in 12 × 75 mm polypropylene tubes (Fisherbrand; Thermo Fisher Scientific Inc., Waltham, MA) for further NH₃-N and volatile fatty acids analyses.

Total gas production, CH₄ production, H₂S production, and NH₃-N concentration

Total gas production was measured by connecting the sealed bottle to an inverted burette filled with water allowing the headspace of the bottle to equilibrate to atmospheric pressure. The amount of water displaced by the gas was recorded as TGP. After pressure was released, using syringes fitted with a one-way valve, 10 and 5 mL of gas was extracted from the sealed bottles for CH₄ and H₂S analyses, respectively. To measure concentrations of CH₄, gas samples from the 10 mL syringe were injected into a gas chromatograph (Trace 1310 Gas Chromatograph, Thermo Scientific, Waltham, MA) equipped with a flame ionization detector and packed columns (HayeSep Q and Porapak Q 0.5 m × 2.0 mm, 80/100, SilcoSteel, Restek Corp., Bellefonte, PA). Temperatures of the injector, columns, and detector were 80, 60, and 250 °C, respectively. Nitrogen was the carrier gas flowing at 10 mL/min.

To determine concentrations of H₂S, gas samples from the 5-mL syringe were slowly bubbled into 5 mL of alkaline water (pH 8.5 to 9.0), which was contained into 15-mL red top evacuated tubes (BD Vacutainer, Franklin Lakes, NJ). Alkaline water was prepared as described by Smith et al. (2010). Tubes were vigorously shaken to allow proper dispersion of the gas in the alkaline water. Subsequently, 0.5 mL of an N, N dimethyl-p-phenylenediamine sulfate solution was injected into the tubes followed by 0.5 mL of a ferric chloride solution. Tubes were shaken vigorously and allowed to rest for 30 min for the reaction to occur. Absorbance was read in a 96-well, flat-bottom plate at 665 nm using a spectrophotometer plate reader (SynergyHT Microplate Reader, BioTek Instruments Inc., Winooski, VT).

To measure concentrations of NH₃-N, the phenol-hypochlorite assay was performed as described by Broderick and Kang (1980) and modified by Henry et al. (2015). The only modification to the original protocol was that samples were transferred (200 µL) into a 96-well flat-bottom plate, and absorbance was read at 665 nm using a plate reader.

Volatile fatty acid analysis

Concentrations of VFA were determined in a water-based solution using an ethyl acetate extraction as described by Ciriaco et al. (2016). A gas chromatograph (Trace 1310 Gas Chromatograph, Thermo Scientific, Waltham, MA) equipped with a flame ionization detector and a capillary column (CP Wax 58 FFAP 25 m × 0.32 mm, CP7767, Agilent Technologies, Santa Clara, CA) was used to analyze samples. Column temperature was maintained at 110 °C, and the injector and detector temperatures were 200 and 350 °C, respectively.

Experiment 2

Experimental design, animals, and treatments

Two hundred thirty-five Angus-crossbreed steers [385 ± 27 kg BW; average BW ± SD] were received on December 28, 2018 at the Texas Tech University Burnett Center Feedlot within the New Deal Research and Education Center. After arrival, steers were allocated to earthen-surfaced receiving pens and were provided with ad libitum access to water, WW B-dahl old world bluestem hay (Bothriochloa bladhii), and a 65% concentrate receiving diet delivered at 2.5% of incoming BW. Three days after arrival (day 21), all steers were identified individually with a unique ear tag and treated in accordance with the Texas Tech University protocols. Briefly, the initial processing included the application of an antiviral (Bovi-Shield Gold 5; Zoetis, Florham Park, NJ), anti-clostridial (One-Shot Ultra 7; Zoetis, Florham Park, NJ), anti-mycoplasma bovis (Myco-Vac B; Texas Vet Lab Inc., San Angelo, TX), and a treatment for internal parasites (Dectomax Pour-On; Zoetis, Florham Park, NJ).

All 235 steers were weighed individually (Silencer Chute; Moly Manufacturing, Lorraine, KS; mounted on Avery Weigh-Tronix load cells, Fairmount, MN, accuracy ± 0.45 kg) on day 14 and were blocked by BW. Out of all 235 steers received, only 200 were included in this trial (10 blocks; 20 steers/block). After blocking, steers returned to the receiving pens and were sorted into their respective blocks 2 d later (day 12). On day 8, cattle were sorted into 50, partially slotted, concrete floor pens (4 steers/pen; 2.9 × 5.5 with 2.4 m of linear bunk space). Liver biopsies were collected on day 5 from one steer randomly selected per pen. On the same day, all steers received an implant with a combination of trenbolone acetate + estradiol benzoate (Synovex Choice; Zoetis, Parsippany, NJ). Four days after liver biopsy extractions, and implant insertion, steers were weighed to record initial BW (day 0) and started a 4-wk adaptation period.

The experiment was conducted as a randomized complete block design with a 2 × 2 + 1 factorial arrangement of treatments. The factors were dietary concentrations of BSS (0 and 0.33% of the diet DM) and NPN source (NPNS; urea and eCAN at 0.68% and 2.0% of the diet DM, respectively). The + 1 treatment, as described below, served as a positive control. For Exp. 2, an encapsulated form of CAN (eCAN) was used to ensure a slow release in the rumen. The concentration of RAS (Sarturi et al., 2013a) was formulated to be 0.28% by adding calcium sulfate (CaSO₄) for the 2 × 2 factorial arrangement of treatments, and 0.14% RAS for the + 1 treatment, which did not receive CaSO₄. The RAS was calculated based on the potential availability of dietary S sources (Sarturi et al., 2013a). In this experiment, the S in CaSO₄ was considered to be 100% available in the rumen, whereas availability of organic sources of S (i.e., S containing amino acids) was calculated by multiplying the percent S in an ingredient by the NASEM (2016) ruminal degradability values of the ingredient protein (Sarturi et al., 2013a). All treatments were isonitrogenous. Treatments are labeled as follows: U (urea at 0.68% with 0.28% RAS), NIT (eCAN at 2.0% with 0.28% RAS), UBS (urea at 0.68% and BSS at 0.33% with 0.28% RAS), NITBS (eCAN at 2.0% and BSS at 0.33% with 0.28% RAS), and LOW-S (urea at 0.68% with 0.14% RAS; Table 2). The eCAN used in the current experiment contained 15.5% N and 75% NO₃⁻ on the DM basis (GRASP Ind. & Com. LYDA, Curitiba, Paraná, Brazil/EW|Nutrition GmbH, Visbek, Germany).

Table 2.

Description of diets fed to steers in Exp. 2

Item	Treatment1
	U	NIT	UBS	NITBS	LOW-S
Ingredient, % DM
Steam-flaked corn	61.27	61.00	61.27	61.27	61.52
Wet corn gluten feed	20.00	20.00	20.00	20.00	20.00
Cottonseed hulls	5.00	5.00	5.00	5.00	5.00
Alfalfa hay, chopped	5.00	5.00	5.00	5.00	5.00
Fat (yellow grease)	3.00	3.00	3.00	3.00	3.00
Mineral supplement	2.00	2.00	2.00	2.00	2.00
Limestone	1.30	0.20	1.30	1.30	1.80
Ground corn	1.00	1.00	1.00	1.00	1.00
Urea	0.68	–	0.68	–	0.68
eCAN	–	2.00	–	2.00	–
BSS	–	–	0.33	0.33	–
Calcium sulfate	0.75	0.75	0.75	0.75	–
Nutrient composition2
DM, % as fed	77.88	77.29	77.75	77.63	77.88
CP, % DM	13.40	13.20	13.50	13.60	13.40
OM, % DM	94.42	94.81	94.16	94.26	94.53
NDF, % DM	20.74	21.61	21.71	21.37	21.77
ADF, % DM	10.43	10.80	11.00	10.07	11.00
TDN, % DM	75.14	74.71	74.86	75.00	75.00
NEm, Mcal/kg	2.01	2.03	2.03	2.04	2.04
NEg, Mcal/kg	1.37	1.39	1.39	1.40	1.40
Ca, % DM	0.68	0.65	0.69	0.66	0.65
P, % DM	0.48	0.52	0.51	0.53	0.49
Mg, % DM	0.20	0.21	0.21	0.20	0.21
K, % DM	1.01	1.06	1.03	1.06	1.07
Na, % DM	0.15	0.17	0.16	0.17	0.18
S, % DM	0.32	0.32	0.34	0.35	0.24
Fe, ppm	150.29	146.14	149.71	147.71	193.29
Zn, ppm	86.71	93.29	84.86	93.29	96.57
Cu, ppm	11.71	12.00	12.14	11.29	10.86
Mn, ppm	31.43	32.71	30.71	33.00	31.86
Mo, ppm	1.07	1.13	1.00	1.09	1.06
RAS3, % DM	0.28	0.28	0.28	0.28	0.14

Diet was formulated to meet or exceed NASEM (2016) requirements for finishing beef cattle.

¹U, Control (Urea at 0.68% with 0.28% RAS); NIT, Nitrate (eCAN at 2.0% with 0.28% RAS); UBS, Urea-Bismuth (Urea at 0.68% and BSS at 0.33% with 0.28% RAS); NITBS, Nitrate-Bismuth (eCAN at 2.0% and BSS at 0.33% with 0.28% RAS); LOW-S, Low sulfur (Urea at 0.68% with 0.14% RAS).

²Analyzed by a commercial laboratory using a wet chemistry package (Dairy One, Ithaca, NY); Net Energy values were calculated based on observed cattle growth performance using the Beef Cattle Nutrient Requirement Model (NASEM, 2016).

³Ruminally available soluble (RAS) was calculated using the formula published by Sarturi et al. (2013a).

All steerson trials were adapted to eCAN and urea for 4 wk through a “4-step” protocol based on a modified version published by Newbold et al. (2014). Throughout the four steps (7 d/step), the NPN sources were gradually added starting from 25% of the total inclusion (0.68% urea and 2.0% eCAN) to 50%, 75%, and 100% for steps 1, 2, 3, and 4, respectively. During the adaptation, all of the animals were closely monitored for any negative health signs. The addition of BSS started at step 1 of the NPN adaptation where it was added at full inclusion (0.33% of the diet DM). The “step-up” protocol also included the adaptation to concentrate in the diet to reach the typical level of a commercial 90% concentrate ration (Table 2). Diets were balanced to meet or exceed the requirements specified by NASEM (2016). The NPN sources were added directly into the feed during mixing; however, a pre-mix (33.33% BSS and 66.66% finely ground corn carrier, DM basis) was added at 1.0% (diet DM) to deliver BSS.

Diets were delivered to each pen once per day. All steers were fed every morning at 0800 h using the clean bunk management technique. Each treatment was mixed with the diet in a paddle type mixer, transferred to a tractor pulled mixer (Roto-mix wagon 274-12; Roto-Mix, Dodge City, KS), and delivered to their respective pen.

Sampling procedures

Body weight

The total duration of thisfeeding trial was 171 d. Body weight data were collected unshrunk prior to feeding (weighing time = 0700 h). All steers were weighed, individually (readability ± 0.45 kg), for initial and final BW on days 0 and 171, respectively, and on a pen basis on days 37, 72, 107, and 142. For intermediate weights, each pen of steers was brought up to a pen scale (readability ± 2.3 kg; validated with 454 to 907 kg of certified weights) where total pen weight was taken. After weighing, steers were returned to their respective pen to be fed at normal routine time (feeding time = 0800 h). Similar management was implemented for individual BW measurements. All scales were calibrated with standard weights before measurements, and all BW data were adjusted with a 4% shrink to account for gut fill.

Dry matter intake

Feed samples were collected for each treatment diet, weekly, immediately after mixing. Subsamples were weighed and placed in a forced-air drying oven at 100 °C for 24 h to determine DM. Another subsample of the diet was collected and kept at −20 °C for further nutrient analysis. Samples for nutrient analysis were composited within month. Orts were collected weekly (every Saturday prior to feeding at 0700 h). All orts were weighed and a sample was placed in a forced-air drying oven at 100 °C for 24 h to calculate weekly DMI.

Apparent total tract digestibility of nutrients

Feed and fecalsamples were collected for a 7-d period from days 44 to 51 to determine apparent total tract digestibility of nutrients. Throughout the 7 d, fecal samples were collected from the pen surface immediately after defecation, prior to feeding (0700 h) and in the afternoon (1600 h). Additionally, feed samples were collected by hand immediately after feeding (sampling time = 0800 h) from each bunk and were placed into plastic bags. Feed samples and orts were stored at 4 °C, and all fecal samples were stored at −20 °C for further analysis.

Liver tissue collection and rectal temperature

Liver samples were extracted via biopsy from one randomly selected steer per pen on days 5 and 124 to obtain initial and final liver mineral concentrations. Both initial and final liver biopsies were collected from the same animal. During the procedure, each selected steer was restrained with a hydraulic chute (Silencer Chute; Moly Manufacturing, Lorraine, KS).

The biopsy site was located in the 11th intercostal space where a line from where the tuber coxae and the point of the shoulder intersect, at which point a 1-cm incision was made using a sterile scalpel. Before performing the incision, the biopsy area was shaved using a disposable razor and sanitized using a solution of water and 7.5% povidone iodine surgical scrub (Betadine, Purdue Products, L.P., Stamford, CT), and 70% ethanol. After sanitation, a local anesthetic (lidocaine HCl, 20 mg/mL, 8 mL per biopsy) was administered subcutaneously around the biopsy area in a square-shaped pattern (4 injection sites, 2 mL of lidocaine HCl per site). To extract the liver tissue, a sterile 16.5 cm long, 6-mm diameter custom designed trocar and sharpened cannula biopsy instrument (SurgiPro, Minneapolis, MN) were used to collect approximately 0.4 g (as-is) of liver tissue (Chapman et al., 1963). The tissue sample was promptly placed in a properly labeled micro-centrifuge tube and into dry ice for further storage.

Immediately after completing the liver tissue extraction, the incision was cleaned with 70% ethanol, closed with a veterinary tissue adhesive (VetBond, 3M Animal Care Products, St. Paul, MN), and coated with a protective aerosol bandage (AluShield, Neogen Corp, Lexington, KY) to prevent any infection during the healing process. Once the biopsy procedure was performed, the samples were transported to the laboratory and were kept at −20 °C for further analysis. All steers were closely monitored for 2 to 3 d after the procedure to ensure health and wellness. Initial and final liver mineral concentrations were determined using ICP-MS (Animal Health Diagnostic Laboratory, Lansing, MI). Rectal temperature was also measured in all steers on day 124 using a rectal thermometer (GLA Agricultural Electronics, San Luis Obispo, CA).

Chemical analyses

Feed and fecal samples were placed in a forced-air drying oven at 55 °C for 48 and 72 h, respectively. As feed refusals represented less than 3% of the amount fed daily (DM basis), nutritional composition adjustments based on refusals were not necessary. After drying, all samples were ground to pass through a 2-mm screen (Willey mill, Thomas Scientific, Swedesboro, NJ). Samples were composited within pen. Briefly, 10 g of each sample were placed in a new bag and shaken for proper mixing. Weekly feed samples were composited by treatment within month and shipped to a commercial laboratory (Dairy One Laboratory, Ithaca, NY) for nutrient analysis. To determine sample DM 0.5 g of sample (in duplicate) was weighed into ceramic crucibles and placed in a forced-air drying oven at 100 °C for at least 16 h before weighing. The dry samples were placed in a muffle furnace for 6 h at 600 °C before returning to a forced-air drying oven at 100 °C for at least 16 h. Hot ashed samples were weighed and used to calculate OM.

For crude protein (CP) analysis, 0.3 g of sample was analyzed in duplicate in a LECO nitrogen/protein analyzer (LECO, St. Joseph, MI). To calculate CP, the concentration of N in each sample was multiplied by 6.25.

Feed and fecal samples were shipped to Dairy One Laboratory (Ithaca, NY) where they were analyzed for S. Treating S as a nutrient, apparent S absorption was calculated the same as “apparent total tract digestibility of S.”

Samples of feed and feces were weighed (0.5 g) in duplicate into previously weighed and labeled F57 bags (Ankom Technology Corp., Macedon, NY) and analyzed for NDF, using heat stable α-amylase and sodium sulfite. Acid detergent fiber analysis was performed sequentially on the same samples as described by Van Soest et al. (1991). Both analyses were performed in an Ankom 200 Fiber Analyzer (Ankom Technology Corp.).

The concentration of indigestible NDF in feed and feces was determined as described by Gregorini et al. (2008), Cole et al. (2011), and Krizsan and Huhtanen (2013). Briefly, 0.5 g of sample was weighed into F57 bags and then incubated in the rumen of a cannulated steer consuming WW B-Dahl (Bothriochloa bladhii) hay for 288 h to ensure complete digestion of potentially digestible NDF. After incubation, samples were rinsed twice with hot tap water followed by 4 rinses with water filtered through a reverse osmosis system. The rinsed samples were then analyzed for NDF as previously described, without the use of α-amylase.

Harvest and carcass data collection

All steers were harvested on the same day with the same days on feed. Steers were transported 169 km to a commercial abattoir (Cargill Meat Solutions, Friona, TX) where trained personnel from Texas Tech University collected carcass measurements. Individual carcass measurements included hot carcass weight, fat thickness, longissimus muscle area, quality grade, marbling scores, dressing percentage, calculated yield grade, and liver scores following the Eli Lilly Check System (Elanco Animal Health, Greenfield, IN) as described by Brown and Lawrence (2010). Dressing percentage was calculated using HCW divided by the nonshrunk individual final BW. Carcass-adjusted final BW was calculated from HCW divided by the average dressing percentage (62.5%) from all five dietary treatments and adjusting for 4% shrink [i.e., (HCW ÷ 0.625) × 0.96)]. The carcass-adjusted BW data were then used to determine carcass-adjusted ADG and G:F. The interim BW data were used to calculate ADG and G:F during each period.

Statistical analysis

All data from Exp. 1 were analyzed as a randomized complete block design with a 2 × 2 factorial arrangement of treatments using the MIXED procedure of SAS (SAS Inst. Inc., Cary, NC). The model included the fixed effects of BSS, NPNS, and the BSS × NPNS interaction. The donor steer was considered the block, which was included in the model as a random effect. The average of 2 bottles or tubes was considered the experimental unit (n = 4/treatment).

In Exp. 2, data were analyzed as a randomized complete block design with a 2 × 2 + 1 factorial arrangement of treatments. Pen was considered the experimental unit (n = 10/treatment). Performance data, rectal temperature, apparent total tract digestibility of nutrients, liver mineral concentration, and continuous carcass data were analyzed using the MIXED procedure of SAS. Initial liver mineral concentration was used as a covariate in order to evaluate final liver mineral concentration. All categorical carcass data were analyzed using the GLIMMIX procedure of SAS. For both continuous and categorical data, the model included the fixed effect of treatment and the random effect of block. To aid in the interpretation of data, the following contrasts were used: the main effect of NPNS = the mean of U, UBS vs. the mean of NIT, NITBS; the main effect of BSS = the mean of U, NIT vs. the mean of UBS, NITBS; the main effect of S (SUL) = the mean of U vs. the mean of LOW-S; and NPNS × BSS = the mean of U and NITB vs. the mean of UB and NIT. In both experiments, the Tukey–Kramer adjustment was used to account for multiple comparison, significance was declared at P < 0.05, and tendencies were discussed at 0.05 < P < 0.10.

Results and Discussion

Experiment 1

In vitro ruminal fermentation measurements are presented in Table 3. A BSS × NPNS interaction (P = 0.005) was observed for TGP, where CTLB increased (P = 0.005) TGP when compared with CTL. There were no interactions (P ≥ 0.661) observed for IVOMD, final pH, CH₄ production, or concentration of NH₃-N. When considering IVOMD as well as pH, there were no effects of NPNS (P = 0.230) or BSS (P = 0.241). The lack of effect on final pH (P > 0.754) may be attributed to the lack of effect on IVOMD. Reductions in IVOMD, as well as an increase in pH were observed when BSS was included above 0.33% of the substrate DM (Ruiz-Moreno et al., 2015; Henry et al., 2021). Moreover, in agreement with the current experiment, data from Henry et al. (2021) indicated that when BSS is added in vitro at 0.33% of the substrate DM, IVOMD was not affected when compared to control. Generally, effects on IVOMD are accompanied by changes in pH that can be attributed primarily to alterations in microbial fermentation (i.e., VFA accumulation), or lack thereof.

Table 3.

Effect of calcium-ammonium nitrate and bismuth subsalicylate on in vitro fermentation parameters (Exp.1)

Item	Treatments1				SEM2	P-value3
	CTL	CAN	CTLB	CANB		NPNS	BSS	BSS × NPNS
TGP4, mL/g incubated OM	159a	220b	223b	221b	32.6	0.029	0.016	0.005
IVOMD, %	70.4	72.5	66.1	70.5	5.20	0.230	0.241	0.661
pH	5.95	5.95	5.95	5.97	0.118	0.754	0.624	0.788
CH4, mmol/g incubated OM	0.13	0.19	0.18	0.20	0.082	0.477	0.644	0.753
NH3-N, mM	5.31	5.86	5.39	6.25	0.968	0.092	0.551	0.687
H2S, µmol/g fermented OM	3.76	6.47	5.67	3.99	1.085	0.617	0.781	0.055
VFA, mol/100 mol
Acetate	50.50	53.47	52.12	53.26	3.278	0.052	0.464	0.346
Propionate	29.48	30.03	30.98	30.29	5.330	0.908	0.179	0.326
Butyrate	13.88	10.48	10.88	10.44	2.572	0.215	0.317	0.331
BCVFA5	2.83	3.20	3.17	3.19	0.708	0.318	0.401	0.372
Valerate	2.37	2.44	2.47	2.42	0.904	0.762	0.292	0.178
Total VFA	115.7	125.9	120.2	125.5	8.31	0.293	0.773	0.728
A:P	1.91	2.02	1.89	1.99	0.451	0.003	0.456	0.758

¹CTL, Control (Urea at 0.74%); CAN, Nitrate (CAN at 2.22%); CTLB, Urea-Bismuth (Urea at 0.74% and BSS at 0.33%); CANB, Nitrate-Bismuth (CAN at 2.22% and BSS at 0.33%).

²n = 4 experimental units/treatment mean.

³NPNS, Effect of non-protein nitrogen source; BSS, Effect of BSS; BSSS × NPNS, Interaction between BSS and NPNS;

⁴TGP after 24 h.

⁵BCVFA, Branched-chain volatile fatty acids: isobutyrate + isovalerate + 2 methylbutyrate.

The lack of effect on CH₄ production by CAN was unexpected as other data from our laboratory and previous literature indicated reductions in in vitro and in vivo CH₄ production when CAN is added to the diet (Olijhoek et al., 2016; Henry et al., 2020b; Henry et al., 2021). Additionally, Henry et al. (2021) observed a linear decrease in in vitro CH₄ production when both CAN and BSS were added, which is in disagreement with the results observed in Exp. 1. It is possible that 24 h of incubation were not enough to observe an effect of CAN on CH₄ production, as donors were not adapted to the treatments. The steers were not adapted, as in Henry et al. (2021), to ensure that any changes in the ruminal microbiome due to adaptation to CAN would not impact in vitro fermentation of substrates with urea as it has been reported that CAN increases populations of nitrate reducers and lactate utilizers (Latham et al., 2016).

In vitro ruminal concentration of NH₃-N of the treatments containing CAN tended (P = 0.092) to be 13% greater than those with urea. This response was expected as NH₃ can be synthesized as an intermediary in the reduction of nitrate to ammonium (NO₂⁻ + 3H₂ + 2H⁺ → NH₃⁺ → NH₄⁺ + 2H₂O; Olijhoek et al., 2016). Zhou et al. (2012) also observed an increase in concentrations of NH₃-N when nitrate was added; however, Henry et al. (2021) reported that CAN linearly decreased NH₃-N concentrations in vitro. Ruiz-Moreno et al. (2015) observed an 8% reduction in NH₃-N when BSS was added at 0.5% or higher, which indicates that concentrations of BSS greater than 0.33% (DM basis) may affect NH₃-N concentration. Data observed in Exp.1 indicate that the addition of BSS at 0.33% DM basis does not affect (P = 0.551) concentrations of NH₃-N which is in agreement with data observed previously by Henry et al. (2021).

There was a tendency (P = 0.055) for a BSS × NPNS interaction for H₂S production; however, no differences (P > 0.283) were observed among treatments. These results are not in accordance with previous data, where, independent of NPNS, linear reductions on in vitro H₂S production were observed when BSS was added at concentrations of 0.33% of the substrate DM or greater (Ruiz-Moreno et al., 2015; Henry et al., 2021). Suarez et al. (1998) also observed a significant reduction of H₂S in human stool and rat cecum when BSS was used. In the current experiment, BSS was utilized to mitigate H₂S by binding to S; therefore, a decrease in H₂S by BSS was expected. The NASEM (2016) recommends at least 0.15% levels of dietary S in the diet of beef cattle to support adequate growth. Ruiz-Moreno et al. (2015) utilized a substrate with a 0.44% of dietary S, whereas the substrate used in the current experiment only contained 0.21% S. Although dietary S in the substrate used exceeds the recommendations of NASEM (2016), perhaps a greater concentration or availability (0.14% RAS in Exp.1, data not shown) of S was necessary in order to observe the effects expected from BSS on H₂S.

Volatile fatty acid profile is presented in Table 3. No interactions (P ≥ 0.178) were observed for any VFA profile variable. There was an effect of NPNS (P = 0.052) on acetate molar proportions and acetate to propionate ratio (A:P), where both were greater when CAN was used compared with urea. Similar results have been reported in the literature where CAN increased acetate molar proportions (Zhou et al., 2012; Guyader et al., 2017). Greater molar proportion of acetate when CAN was used was attributed to a lesser availability of H⁺ due to nitrate acting as a H+ sink (Guyader et al., 2017). This leads to ruminal microorganisms to replenish the milieu with ions by using pathways that yield more H⁺, such as acetate production (Guyader et al., 2017; Janssen, 2010), which may explain what occurred during Exp. 1. No effects of NPNS (P ≥ 0.215) or BSS (P ≥ 0.179) were observed for molar proportions of propionate, butyrate, branched chain VFA, or valerate. Moreover, total VFA concentration was not affected by NPNS (P = 0.293) or BSS (P = 0.773). Previous reports indicated a similar response when CAN was provided in a forage-based substrate up to 2.4%; however, when BSS has been added in the substrate up to 1%, total VFA production decreased linearly (Henry et al., 2021).

Experiment 2

Effects on growth performance

All growth and performance data are presented in Table 4. There was no effect of SUL (P = 0.429) on carcass-adjusted final shrunk BW. There was no BSS × NPNS interaction (P = 0.344) observed for carcass-adjusted final shrunk BW. Final BW was affected (P = 0.045) by NPNS where cattle consuming eCAN were 13 kg lighter than those consuming urea; however, after calculating carcass-adjusted final shrunk BW had a tendency for a NPNS effect (P = 0.055) where steers consuming eCAN only tended to be lighter than those consuming urea. There was no BSS effect (P = 0.841) on carcass-adjusted final shrunk BW. There was a NPNS effect (P = 0.001) observed for DMI from days 72 to 107, where steers receiving eCAN had lesser intake than steers consuming urea, and these steers maintained a reduced DMI (P < 0.001) until the culmination of the trial. The reduction of DMI when steers were fed eCAN can be attributed to a number of factors. When nitrate enters the rumen, nitrite (NO₂⁻) is one of the nitrate-reduction intermediaries (Olijhoek et al., 2016). It is possible that when NO₂⁻ builds up in the ruminal environment, to the point that it surpasses the reduction capacity of the microorganisms, it can be absorbed in the blood and increase the levels of methemoglobin, which can trigger a reduction in DMI (Lee et al., 2017). It has also been suggested that nitrate could change the organoleptic properties of the diet, especially palatability, which can explain the reduction in DMI (Lee et al., 2017). Data collected during the current experiment is not sufficient to suggest that the reduction in DMI was due to NO₂⁻ buildup in the rumen or methemoglobin levels. Therefore, the more appropriate explanation for the decrease in DMI in the current experiment is the change in palatability.

Table 4.

Effect of bismuth subsalicylate and calcium ammonium nitrate on feedlot growth performance

Item1	Treatment2					SEM3	P-value4
	U	NIT	UBS	NITBS	LOW-S		SUL	NPNS	BSS	BSS × NPNS
IBW5, kg	369	368	368	371	371	8.0	0.211	0.338	0.342	0.126
FBW5, kg	637	617	626	620	629	9.6	0.380	0.045	0.531	0.310
CAFSBW6, kg	636	618	626	620	629	11.2	0.429	0.055	0.555	0.344
Days 37–72
ADG, kg	1.58	1.60	1.59	1.80	1.66	0.082	0.495	0.185	0.223	0.297
DMI, kg	8.76	8.52	8.74	8.57	8.72	0.285	0.921	0.496	0.970	0.906
G:F	0.184	0.191	0.184	0.210	0.195	0.0127	0.572	0.216	0.486	0.462
Days 72–107
ADG, kg	1.85	1.77	1.69	1.62	1.73	0.113	0.423	0.473	0.148	0.965
DMI, kg	10.47	9.69	9.99	9.54	10.35	0.213	0.643	0.001	0.076	0.353
G:F	0.176	0.184	0.168	0.170	0.167	0.0107	0.493	0.637	0.256	0.733
Days 107–142
ADG, kg	1.36	1.14	1.34	1.14	1.11	0.071	0.020	0.007	0.950	0.889
DMI, kg	10.44	9.33	10.00	9.10	9.75	0.228	0.030	< 0.001	0.127	0.623
G:F	0.131	0.122	0.135	0.125	0.113	0.0072	0.088	0.193	0.629	0.965
Days 0–171
CAADG6, kg	1.47	1.37	1.42	1.36	1.42	0.037	0.309	0.041	0.446	0.537
DMI, kg	9.41	8.77	9.10	8.69	9.06	0.122	0.036	< 0.001	0.095	0.313
CAG:F6	0.156	0.156	0.156	0.157	0.157	0.0030	0.993	0.978	0.880	0.856

¹Individual body weight data were recoded for initial and final weights (days 0 and 171), whereas pen weights were recorded for interim weights (days 37, 72, 107, and 142).

²U, Control (Urea at 0.68% with 0.28% RAS); NIT, Nitrate (eCAN at 2.0% with 0.28% RAS); UBS, Urea-Bismuth (Urea at 0.68% and BSS at 0.33% with 0.28% RAS); NITBS, Nitrate-Bismuth (eCAN at 2.0% and BSS at 0.33% with 0.28% RAS); LOW-S, Low sulfur (Urea at 0.68% with 0.14% RAS).

³n = 10 experimental units/treatment mean.

⁴Significance is declared at P < 0.05, tendencies are discussed at 0.05 < P < 0.10. SUL, Effect of sulfur; NPNS, Effect of non-protein nitrogen source; BSS, Effect of BSS; BSS × NPNS, Interaction between NPNS and BSS.

⁵IBW, Initial body weight with a common shrink (4%); FBW, Final body weight with a common shrink (4%).

⁶CAFSBW, Carcass-adjusted final shrunk body weight; CAADG, Carcass-adjusted average daily gain; CAGF, Carcass-adjusted gain to feed ration. Carcass-adjusted final body weight calculated as HCW divided by overall dressing percent and then multiplied by a common shrink (4%).

An effect of SUL was observed (P = 0.036) where LOW-S cattle had lesser DMI than U over the entire experimental period (days 0 to 171). These results were unexpected as increases in dietary S have been observed to reduce DMI in feedlot cattle (Drewnoski et al., 2014); however, Loneragan et al. (2001) reported a quadratic effect on DMI of steers provided high levels of sulfate (up to 2,360 mg/L) in their drinking water. Moreover, Sarturi et al. (2013b) observed a linear decrease in DMI as levels of RAS in beef cattle finishing diets reached 0.38% (0.55% total S; DM basis). It is possible that the level of dietary S in treatments containing CaSO₄ was not high enough to observe the decrease in DMI that was expected, since the greatest amount of RAS used in the current experiment was 0.28% (DM basis). Overall performance data (days 0 to 171) indicated that DMI (P < 0.001) and carcass-adjusted ADG (P = 0.041) were affected by NPNS. Adding eCAN as a NPN source induced a reduction in DMI by 0.50 kg per day when compared to urea treatments. This reduction in carcass-adjusted ADG by eCAN translated into a lesser carcass-adjusted final shrunk BW in steers that consumed eCAN; however, carcass-adjusted G:F was not affected (P = 0.978). There was a tendency (P = 0.095) for BSS to reduce DMI by the end of the trial, but no effect on carcass-adjusted ADG (P = 0.446) or carcass-adjusted G:F (P = 0.880) was observed. Nellore bulls fed nitrate concentrations up to 2.4% of DM had reduced DMI with increasing nitrate without affecting ADG resulting in improved feed efficiency (Newbold et al., 2014). Similarly, Lee et al. (2017) observed reductions in DMI from cattle consuming eCAN at 2.5% of diet DM with tendencies to increase ADG, which consequently improved G:F. The data presented by previous researchers indicate that reductions in DMI from nitrate are expected with improvements in G:F; however, this was not observed in the current experiment. It is likely that the reduction in DMI was so extensive that gain was subsequently reduced, thereby preventing an improvement in feed conversion efficiency.

Effects on carcass characteristics

Continuous data for carcass characteristics are presented in Table 5. No interactions (P = 0.832) were observed for hot carcass weight; however, there was an effect of NPNS on hot carcass weight (P = 0.033) where steers receiving eCAN had lesser hot carcass weight than those receiving urea. Previous data have indicated a lack of effect of nitrate on hot carcass weight (Newbold et al., 2014). Additionally, other authors have reported data where eCAN has not affected carcass characteristics (Lee et al., 2017). Although previous literature indicates that eCAN does not affect hot carcass weight, it is clear that in Exp. 2, the reduction in this variable by eCAN is related to the response in carcass-adjusted final shrunk BW where steers consuming eCAN were nearly 13 kg lighter than steers consuming urea, likely driven by the reduction in DMI by eCAN. Longissimus muscle area tended (P = 0.080) to be reduced when BSS was added to the diet. A SUL effect (P = 0.029) was observed on fat thickness, where it was lesser in U than in LOW-S steers. Loneragan et al. (2001) observed a quadratic effect on fat thickness when S was added up to 2,360.4 mg/L in the form of sulfate to the drinking water of cattle; however, the authors did not speculate as to why this may occur. There was a tendency for an interaction (P = 0.097) on dressing percentage where cattle consuming UBS tended to have greater dressing percentage than those consuming U. There was an interaction of BSS × NPNS (P = 0.019) on calculated yield grade where UBS tended (P = 0.060) to be greater when compared with U but NIT and NITBS were not different (P = 0.960). There was also a SUL effect (P = 0.008) where LOW-S had greater calculated yield grade than U, which has been reported in the literature (Loneragan et al., 2001).

Table 5.

Effect of bismuth subsalicylate and calcium ammonium nitrate on continuous carcass characteristics

Item1	Treatment2					SEM3	P-value4
	U	NIT	UBS	NITBS	LOW-S		SUL	NPNS	BSS	BSS × NPNS
Hot carcass weight, kg	412.8	402.9	411.1	402.9	410.5	7.27	0.694	0.033	0.841	0.832
Dressing, %	62.3	62.6	63.0	62.4	62.7	0.32	0.366	0.575	0.379	0.097
LM area, cm2	88.97	86.11	83.88	85.42	86.18	1.949	0.226	0.685	0.080	0.179
Fat thickness, cm	1.71	1.67	1.78	1.66	1.96	0.076	0.029	0.282	0.712	0.637
Marbling score	579.1	578.8	583.5	566.1	589.8	17.54	0.676	0.622	0.816	0.636
Calculated yield grade	3.24	3.28	3.68	3.17	3.58	0.121	0.036	0.044	0.143	0.019

¹LM, Longissimus muscle area (cm2). Marbling score, 400 =Small00; 500, Modest00; 600, Moderate00.

²U, Control (Urea at 0.68% with 0.28% RAS); NIT, Nitrate (eCAN at 2.0% with 0.28% RAS); UBS, Urea-Bismuth (Urea at 0.68% and BSS at 0.33% with 0.28% RAS); NITBS, Nitrate-Bismuth (eCAN at 2.0% and BSS at 0.33% with 0.28% RAS); LOW-S, Low sulfur (Urea at 0.68% with 0.14% RAS).

³n = 10 experimental units/treatment mean.

⁴Significance is declared at P< 0.05, tendencies are discussed at 0.05 < P < 0.10. SUL, Effect of sulfur; NPNS, Effect of non-protein nitrogen source; BSS, Effect of BSS; BSS × NPNS, Interaction between NPNS and BSS.

Categorical data for carcass characteristics are presented in Table 6. There were no BSS × NPNS interactions (P > 0.177) observed on any quality grade. There was a tendency for a BSS effect (P = 0.099) on quality grade for Choice where less cattle tended to grade in this category when BSS was added to the diet. No effects (P > 0.137) were observed on the other quality grades. There was a BSS × NPNS interaction (P = 0.022) on yield grade 3 cattle where less steers tended (P = 0.095) to score in this grade when consuming UBS in comparison to cattle consuming U. A tendency for a BSS × NPNS interaction (P = 0.068) was observed on yield grade 4; however, there were no differences (P > 0.458) observed among treatments. Only a few researchers have looked at the effects of eCAN in combination with BSS on cattle performance and carcass characteristics (Henry et al., 2020b). Henry et al. (2020b) evaluated some carcass variables using ultrasound on growing cattle receiving high-forage diets with eCAN and BSS. The authors did not observe any effects on longissimus muscle area; however, there was an interaction of BSS × NPNS where animals consuming eCAN and BSS had less back fat compared to those consuming urea and BSS, which is in disagreement with the data observed in Exp. 2.

Table 6.

Effect of bismuth subsalicylate and calcium ammonium nitrate on categorical carcass traits

Item	Treatment1					SEM2	P-value3
	U	NIT	UBS	NITBS	LOW-S		SUL	NPNS	BSS	BSS × NPNS
Quality grade, %
Prime	7.70	14.29	17.95	10.26	19.44	6.301	0.137	0.964	0.551	0.177
Premium Choice	61.54	60.00	56.41	56.41	66.67	8.046	0.409	0.571	0.908	0.446
Choice	29.03	25.64	15.38	15.38	19.44	8.469	0.385	0.843	0.099	0.843
Select	0.00	0.00	6.45	7.69	5.56	3.767	0.979	1.000	0.970	1.000
Yield grade, %
1	0.00	0.00	0.00	0.00	0.00	0.000	0.960	0.989	0.994	0.999
2	5.13	14.29	5.13	17.95	11.11	5.968	0.342	0.038	0.818	0.818
3	64.10	45.71	38.46	53.85	44.44	7.233	0.059	0.827	0.217	0.022
4	23.08	38.46	41.03	28.21	36.11	8.017	0.226	0.817	0.592	0.068
5	0.00	0.00	11.43	0.00	10.00	3.470	0.971	0.983	0.983	0.983
Liver score4, %
Edible	82.05	89.74	87.18	82.05	95.00	6.449	0.112	0.796	0.796	0.287
Abnormal	17.95	10.26	12.82	17.95	5.00	6.449	0.112	0.796	0.796	0.287

¹U, Control (Urea at 0.68% with 0.28% RAS); NIT, Nitrate (eCAN at 2.0% with 0.28% RAS); UBS, Urea-Bismuth (Urea at 0.68% and BSS at 0.33% with 0.28% RAS); NITBS, Nitrate-Bismuth (eCAN at 2.0% and BSS at 0.33% with 0.28% RAS); LOW-S, Low sulfur (Urea at 0.68% with 0.14% RAS).

²The highest SEM was included in this table (n = 10 experimental units/treatment mean).

³Significance is declared at P< 0.05, tendencies are discussed at 0.05 < P < 0.10. SUL, Effect of sulfur; NPNS, Effect of non-protein nitrogen source; BSS, Effect of BSS; BSS × NPNS, Interaction between NPNS and BSS.

⁴Liver scores: Edible, No presence of abscess; Abnormal, at least one abscess.

Previous research on the effects of S on carcass characteristics has been performed by increasing levels of dietary S with distiller grains (Klopfenstein et al., 2008; Richter et al., 2012). It is important to consider that when the effect of S is being evaluated, the source of S plays an important role in the outcome of a given experiment, meaning that when the same variables are evaluated, the effect from organic sources of S is less pronounced compared to inorganic sources of S (Drewnoski et al., 2014). Therefore, the results observed in the current experiment on carcass quality are still unclear, and more data are warranted for a better understanding of these effects.

No differences (P > 0.112) were observed on liver scores (Table 6); however, it should be noted that, in the current experiment, there was only a 5% incidence of liver abscesses when steers consumed the LOW-S treatment compared to a 17.95% from steers receiving U. The incidence of liver abscesses is generally considered to be related to ruminal acidosis and the lesions that a continual high-acid load can create on the ruminal wall, which, in consequence, leads to the entry of microorganisms into the portal system causing liver abscesses (Galyean and Rivera, 2003). An increase in ruminal H₂S concentration has been related to reductions in ruminal pH, which is an indication that cattle consuming a finishing ration are more susceptible to greater ruminal concentrations of H₂S due to a lower ruminal pH (Gould et al., 1997). The main microorganism thought to be responsible for bovine liver abscesses is Fusobacterium necrophorum (Nagaraja et al., 2005; Nagaraja and Lechtenberg, 2007). Additionally, F. necrophorum and members of the Fusobacterium genus have been observed to have H₂S enzymatic activity, indicating that these bacteria may have S reducing activity (Yoshida et al., 2010; Basic et al., 2017). It is possible that a combination of low ruminal pH and a greater availability of S may have created conditions for larger populations of F. necrophorum. This may explain the numerically greater incidence of liver abscesses in steers consuming U compared with those consuming LOW-S. It is also important to mention that Exp. 2 was not designed to evaluate the effects of S on liver abscess incidence; therefore, further research should be performed to better understand this response.

Effects on apparent total tract nutrient digestibility

Apparent total tract digestibility of nutrients, nutrient intake, and S absorption are presented in Table 7. There were no BSS × NPNS interactions (P > 0.472), or effects of NPNS (P > 0.516) or BSS (P > 0.346) on intake of nutrients. A SUL effect (P < 0.001) was observed where cattle consuming U had greater (P < 0.001) S intake compared to LOW-S. This was expected as steers receiving the LOW-S treatment did not receive added CaSO₄. There were no interactions (P = 0.176) observed on DM digestibility; however, DM digestibility tended (P = 0.056) to be greater when steers received diets containing BSS. Similarly, OM digestibility was greater (P = 0.045) when steers were consuming BSS. Henry et al. (2021) observed that addition of BSS above 0.33% of the substrate DM reduced IVOMD; however, the authors did not observe a difference in IVOMD between 0 and 0.33% of the substrate DM. Moreover, it was also reported that BSS did not affect apparent total tract digestibility of OM when provided in a forage-based diet (Henry et al., 2020b). In the current experiment, steers receiving BSS tended to have lesser DMI than steers receiving the non-BSS treatments. It is possible that the lesser DMI induced by BSS may have caused a greater retention time in the rumen. Greater retention time would allow the rumen microorganisms to further digest the diet, and therefore, increase DM and OM digestibility.

Table 7.

Effect of bismuth subsalicylate and calcium ammonium nitrate on nutrient intake, apparent total tract digestibility, and sulfur absorption

Item	Treatment1					SEM2	P-value3
	U	NIT	UBS	NITBS	LOW-S		SUL	NPNS	BSS	BSS × NPNS
Intake4, kg/d
DM	8.47	8.05	8.19	8.28	8.24	0.362	0.674	0.661	0.962	0.499
OM	8.02	7.63	7.73	7.83	7.81	0.342	0.687	0.694	0.907	0.499
CP	1.23	1.20	1.21	1.24	1.18	0.064	0.512	0.988	0.839	0.546
NDF	1.70	1.68	1.61	1.60	1.61	0.082	0.499	0.867	0.346	0.967
ADF	0.62	0.62	0.61	0.59	0.61	0.031	0.717	0.664	0.549	0.769
S	0.031	0.029	0.030	0.030	0.021	0.0013	< 0.001	0.516	0.996	0.472
Digestibility, %
DM	66.0	67.7	72.4	68.9	64.5	2.53	0.590	0.657	0.056	0.176
OM	68.1	69.4	74.1	71.0	66.7	2.42	0.593	0.611	0.045	0.225
CP	53.7	55.9	64.0	55.7	51.7	3.46	0.593	0.253	0.065	0.055
NDF	26.3	32.8	31.7	27.7	21.2	6.01	0.436	0.770	0.975	0.227
ADF	18.7	26.3	27.9	19.3	15.9	6.43	0.692	0.914	0.812	0.083
S absorption, %	60.8	63.6	55.1	48.9	44.6	2.75	< 0.001	0.4705	< 0.001	0.056

¹U, Control (Urea at 0.68% with 0.28% RAS); NIT, Nitrate (eCAN at 2.0% with 0.28% RAS); UBS, Urea-Bismuth (Urea at 0.68% and BSS at 0.33% with 0.28% RAS); NITBS, Nitrate-Bismuth (eCAN at 2.0% and BSS at 0.33% with 0.28% RAS); LOW-S, Low sulfur (Urea at 0.68% with 0.14% RAS).

²n = 10 experimental units/treatment mean.

³Significance is declared at P< 0.05, tendencies are discussed at 0.05 < P < 0.10. SUL, Effect of sulfur; NPNS, Effect of non-protein nitrogen source; BSS, Effect of BSS; BSS × NPNS, Interaction between NPNS and BSS.

⁴Nutrient intake was measured (on the per steer basis) over a 7-d period.

A tendency for a BSS × NPNS interaction (P = 0.055) was observed for CP digestibility, where UBS tended (P = 0.065) to have greater CP digestibility than U; however, NIT and NITBS were not different (P = 1.000). Likewise, there was a tendency for a BSS × NPNS interaction (P = 0.083) on ADF digestibility; however, no differences (P > 0.400) were observed among treatments. There was also a tendency (P = 0.065) for an effect of BSS on CP digestibility where BSS tended to increase digestibility of CP by 9.2% compared to the non-BSS treatments. Ruiz-Moreno et al. (2015), using a dual-flow continuous culture system, evaluated the addition of BSS on in vitro digestibility of OM, NDF, and ADF. The data indicated reductions in OM, NDF, and ADF digestibility when BSS was added at 1.0% of the substrate DM (Ruiz Moreno et al., 2015). Nevertheless, the concentration of BSS used in the current experiment was 0.33% of the diet DM, indicating that BSS at such concentration fed to feedlot steers is not detrimental to apparent total tract digestibility of nutrients.

There was a tendency (P = 0.056) for a BSS × NPNS interaction where there appears to be an additive effect of the two mitigating strategies on reducing apparent S absorption. There was an effect of SUL (P < 0.001) and BSS (P < 0.001) on apparent S absorption, where cattle consuming LOW-S absorbed less S, when compared with U steers, as well as steers consuming BSS. The SUL effect was expected, as steers receiving less S in their feed should present less apparent absorption of the mineral. Additionally, these data indicate that S absorption is reduced by BSS as S and bismuth bind and become bismuth sulfide, a black salt that is not absorbed through the intestinal wall (Suarez et al., 1998). Henry et al. (2020b) observed similar results where cattle consuming BSS presented lesser S retention. The proposed reaction by which bismuth binds to S is the following:

$${\displaystyle \begin{array}{l}{\displaystyle 2\mathrm{B}{\mathrm{i}}^{3+}+3{\mathrm{H}}_2\mathrm{S}\leftrightarrow \mathrm{B}{\mathrm{i}}_2{\mathrm{S}}_3+6{\mathrm{H}}^{+}.}\end{array}}$$

The outcome observed in the current experiment, as well as previous research, indicates that BSS is acting as expected; however, the exact mechanisms have not been established. Therefore, further research should be performed for a better understanding of this compound mechanism of action on distinct substrates and other potentials.

Liver mineral concentration

Liver biopsy samples for initial concentration of minerals (data not shown) were collected before treatments were administered while all steers were consuming the same diet; therefore, any effects observed on initial concentrations of minerals were merely random and will not be discussed.

Final liver mineral data are presented in Table 8. A tendency for a BSS × NPNS interaction was observed (P = 0.103) on final concentration of Mn; however, there were no differences (P > 0.185) among treatments. There was an effect of BSS (P = 0.042) on final concentration of Cu, where Cu was reduced by 58% in BSS treatments when compared with treatments without BSS. There were no main effects (P > 0.142) observed on final concentration of Fe, Zn, and Mn. It was hypothesized that when BSS enters the rumen, it will bind to H₂S forming Bi₂S₃; therefore, reducing the amount of available S that could otherwise bind to trace minerals and, consequently, the absorption of these minerals would be greater in steers receiving BSS. The results observed in this experiment indicate that BSS can reduce the absorption of some minerals such as Cu, which does not support our hypothesis. Henry et al. (2020b) observed the same effect on final liver Cu concentration where BSS reduced the absorption of Cu. It is possible that BSS is not only binding to S, but it is also binding to Cu, forming Bi-Cu-S compounds, and reducing the availability of Cu for absorption.

Table 8.

Effect of bismuth subsalicylate and calcium ammonium nitrate on final liver mineral concentration, liver mineral concentration change, and rectal temperature.

Item	Treatment1					SEM2	P-value3
	U	NIT	UBS	NITBS	LOW-S		SUL	NPNS	BSS	BSS × NPNS
Final liver mineral4, mg/kg DM
Se	1.81	1.27	1.28	1.27	1.68	0.193	0.619	0.149	0.156	0.160
Fe	397.91	298.65	400.27	372.50	407.76	54.277	0.211	0.211	0.455	0.479
Cu	134.35	89.95	61.34	33.26	171.98	30.566	0.384	0.251	0.042	0.794
Zn	133.99	114.17	103.89	105.18	112.87	12.878	0.218	0.454	0.145	0.403
Mn	10.71	7.93	7.96	8.11	9.25	0.923	0.233	0.142	0.153	0.103
Liver mineral change5, mg/kg DM
Se	−0.31	−0.93	−0.81	−1.06	−0.62	0.225	0.322	0.059	0.161	0.421
Fe	62.84	30.64	33.67	−16.00	184.02	67.689	0.207	0.717	0.750	0.916
Cu	−131.81	−180.81	−186.72	−266.32	−125.55	38.894	0.906	0.100	0.073	0.690
Zn	2.00	−13.66	−32.67	−39.41	−22.31	13.010	0.181	0.391	0.025	0.732
Mn	1.42	−1.32	−1.36	−1.89	−0.24	1.123	0.293	0.156	0.148	0.333
Rectal temperature, °C	38.93	39.09	38.91	39.00	39.04	0.081	0.356	0.132	0.536	0.736

¹U, Control (Urea at 0.68% with 0.28% RAS); NIT, Nitrate (eCAN at 2.0% with 0.28% RAS); UBS, Urea-Bismuth (Urea at 0.68% and BSS at 0.33% with 0.28% RAS); NITBS, Nitrate-Bismuth (eCAN at 2.0% and BSS at 0.33% with 0.28% RAS); LOW-S, Low sulfur (Urea at 0.70% with 0.14% RAS).

²The highest SEM was included in this table (n = 10 experimental units/treatment mean).

³Significance is declared at P < 0.05, tendencies are discussed at 0.05 > P < 0.10. SUL = Effect of sulfur; NPNS = Effect of non-protein nitrogen source; BSS = Effect of BSS; BSS × NPNS = Interaction between NPNS and BSS.

⁴Final liver mineral was analyzed using initial liver mineral concentration as a covariate.

⁵Change in liver mineral status was calculated by subtracting final mineral concentration minus initial mineral concentration.

The effect of bismuth on mineral retention of rats has been investigated, where injections of BiCl₃ caused an increase in total body Cu content with no effects on liver Cu concentration (Szymańska and Zelazowski, 1979). The authors observed that while concentration of Cu was not affected in the liver, kidney concentrations of Cu were increased by bismuth (Szymańska and Zelazowski, 1979). It can be speculated that bismuth may bind to Cu, and once this occurs, Cu becomes unavailable for liver storage and it is excreted through the urine; however, urine was not collected in this experiment. Henry et al. (2020b) observed an increase in liver concentration of Fe in cattle consuming BSS. The authors attributed this effect to the decrease in concentration of Cu in the liver since Fe is mobilized by ferroxidase, a Cu-dependent enzyme (Henry et al., 2020b; Suttle, 2010); however, Fe was not affected by BSS in the current experiment.

Changes in liver mineral concentration are presented in Table 8. There was a tendency (P = 0.059) for a NPNS effect on mineral concentration change of Se, where the decrease tended to be more extensive in eCAN treatments. A tendency (P = 0.073) for a BSS effect was also observed on concentration change of Cu, where BSS treatments tended to further the decrease in Cu concentration within the liver. Change in concentration of Zn was greater (P = 0.025) in BSS treatments when compared with non-BSS treatments. Changes in concentration of Fe and Mn were not affected (P ≥ 0.148).

Unfortunately, to the authors’ knowledge, few researchers have reported data where the change in mineral concentration has been evaluated when cattle received nitrate (Henry et al., 2020b). It is unclear why nitrate tended to affect change in concentration of Se. The tendency for an effect of BSS on Cu concentration change is likely a response to the main effect of BSS observed on final Cu concentration. Prados et al. (2017) reported that adequate concentrations of liver Cu range from 100 to 400 mg/kg. During the current experiment, steers consuming BSS had a final concentration of Cu of 47.3 mg/kg within the liver, indicating a deficiency of Cu; however, the performance of these steers was not affected by BSS and there were no symptoms of a Cu deficiency detected.

There were no differences observed for rectal temperature (P ≥ 0.132) during this experiment (Table 8). Henry et al. (2020a) reported a reduction in rectal temperature of cattle receiving a high forage diet consuming BSS compared with those not consuming BSS. The authors attributed the effect of BSS on rectal temperature to the possibility of BSS acting as a nonsteroidal anti-inflammatory drug (NSAID; Henry et al., 2020a). The steers in this experiment were consuming a high-concentrate diet; therefore, it is possible that diet type is influencing the effects of BSS on rectal temperature.

Conclusions

Bismuth subsalicylate is a novel compound in the field of ruminant nutrition. Data have indicated that BSS may have potential as a mitigator of H₂S in the rumen; however, some of the results observed are still unclear and contradict other literature, which calls for more research with this compound to elucidate these impacts. The idea behind combining BSS and eCAN revolves around the ability of these two compounds to decrease H₂S by differing modes of action. During the developing stage of this experiment, it was hypothesized that the combination of BSS and eCAN would create an additive effect, which was observed in some of the results, but not to the extent that was expected. The reduction on liver Cu concentration by BSS indicates that BSS may be binding to Cu in the rumen; however, the absence of steers showing symptoms of a mineral deficiency begs the question of severity and where this Cu is going after it binds to BSS. The results obtained during this experiment show that BSS may be fed to cattle as a ruminal H₂S mitigator and it appears to bind to S as expected; however, the magnitude of the effect and the exact mechanisms of this compound within the animal are still unclear and require further research.

Glossary

Abbreviations

A:P

acetate to propionate ratio

ADF

acid detergent fiber

ADG

average daily gain

BSS

bismuth subsalicylate

BW

body weight

CAN

calcium-ammonium nitrate

CP

crude protein

DM

dry matter

DMI

DM intake

eCAN

encapsulated CAN

G:F

gain to feed ratio

IVOMD

in vitro OM digestibility

NDF

neutral detergent fiber

NPN

nonprotein nitrogen

NPNS

NPN source

NSAID

nonsteroidal anti-inflammatory drug

OM

organic matter

RAS

ruminally available sulfur

S

sulfur

TGP

total gas production

TDN

total digestible nutrients

VFA

volatile fatty acids