H. S. Hussein, H. Han, and H. A. Glimp
Department of Animal Biotechnology
College of Agriculture, Biotechnology and Natural Resources
University of Nevada, Reno

The synthetic antioxidant ethoxyquin (6-ethoxyl-1, 2-dihydro-2, 2, 4-trimethylquinoline) has been used in feedstuffs since 1950s (Coelho, 1995). Recently, ethoxyquin supplementation of feedlot diets increased average daily gain (ADG; Krumsiek and Owens, 1998a; McBride, 2000), decreased morbidity, mortality, and medication cost (Stovall et al., 1999; Kegley et al., 2000), improved beef shelf life (Krumsiek and Owens, 1998b), and reduced beef rancidity (Walenciak et al., 1999). Beef cattle response to ethoxyquin, however, has not been always consistent. For example, feedlot performance was not affected when ethoxyquin was added to the starting diets of beef steers (Kegley et al., 2000) or heifers (Stovall et al., 1999). The variable response could have been due to animal, dietary, or environmental factors.

The positive feedlot responses to ethoxyquin with beef cattle could be explained by its potential antioxidant role at the gastrointestinal tract level and(or) at the post-absorption level. At the gastrointestinal tract, ethoxyquin may improve fermentation patterns (e.g., production or composition of volatile fatty acids [VFA]) in the rumen and the colon and(or) enhance the antioxidant capacity and health of the intestinal mucosa. This study was designed to assess the effects of ethoxyquin at the ruminal level. Because a large number of finishing diets contain monensin and tylosin, it was important to examine the possible interactions between ethoxyquin and these feed additives, especially monensin. Therefore, the objective was to determine the effects of ethoxyquin alone or in combination with monensin and tylosin on ruminal fermentation characteristics of a finishing beef diet. This was accomplished by using the dual-flow continuous culture fermenter system (i.e., artificial rumen) as the closest model to simulate the ruminal microbial fermentation.

Two ruminally-cannulated Angus steers were used as donors of rumen fluid to be used as the inoculum for the dual-flow continuous culture fermenter system. The steers had libitum access to a high-concentrate finishing diet (containing 90% corn and 10% alfalfa hay on a dry matter [DM] basis) for 2 wk before and throughout the study. This diet was formulated to meet or exceed the nutrient requirements of the steers (NRC, 1996). The rumen fluid was collected from each steer approximately 2 h post-feeding and strained immediately through four layers of cheesecloth into a pre-warmed, insulated thermos.

The dual-flow continuous culture fermenter system was developed (Hoover et al., 1976) and modified (Hannah et al., 1986) to simulate differential solid-liquid removal rates occurring in the rumen environment. Eight fermenters (1,020 mL working volume each) were equipped with an automated feeding system and were continuously infused with a mineral buffer solution (artificial saliva; Weller and Pilgrim, 1974) containing urea (0.5 g/L) at a rate of 1.5 mL/min to obtain a liquid dilution rate of 0.082 h^-1. Solid (overflow) dilution rate was maintained at 0.041 h^-1 by removing liquid through a filter at 0.75 mL/min. A pH of 6.0 ± 0.05 was maintained by automated infusion of 3 N HCl or 5 N NaOH regulated by a pH controller (Cole-Parmer, Vernon Hills, IL). Anaerobic conditions were achieved by continuous infusion of N₂ at a rate of 40 mL/min. Maintaining the fermenters’ temperature at 39^oC and mixing of their contents were achieved by using VirTis Omni-Culture fermenter base units (The Virtis Company, Gardiner, NY).

Upon arrival to the laboratory, the rumen fluids from both steers were combined and used to inoculate the eight fermenters. Each fermenter was supplied daily with 75 g DM of a ground (2-mm screen) diet by an automated feeding mechanism adjusted to deliver the diet in 12 equal portions over a 24-h period to establish steady-state conditions.

The finishing diet was formulated to contain 12.5% crude protein (CP) on a DM basis. This diet was previously evaluated (McBride, 2000) without (control) or with ethoxyquin (150 ppm). The only exception was replacing cottonseed hulls with corn cobs (similar in their nutritional value) to avoid the delivery problems of cottonseed hulls in the automated feeding mechanism used in our continuous culture system. Table 1 shows the ingredient composition of the four experimental diets evaluated in this study. Treatments (Table 1) were arranged as a 2 × 2 factorial with the main factors being two ethoxyquin treatments (without or with 150 ppm [DM basis] from AGRADO^®; Solutia Inc., St. Louis, MO) and two monensin/tylosin treatments (without or with monensin and tylosin at 0.0028 and 0.0014% of dietary DM, respectively). The levels of monensin and tylosin used in this study were those commonly fed to finishing cattle. The maximum concentration of ethoxyquin in feeds permitted by the FDA is 150 ppm. Each dietary ingredient was ground through a 2-mm screen before mixing each diet. Because the chemical analysis of the four diets was similar, the average values are presented in Table 2.

The experimental design was a randomized complete block design (Steel et al., 1997) and consisted of two experimental periods (blocks) of 8 d each. The first 5 d of each period were used for stabilization (e.g., adaptation to the diet) and the last 3 d were used for sample collection. The four diets (Table

1) were allocated randomly to fermenters, giving two replications for each diet per period.

Solid and liquid fractions of the effluent were collected in two vessels submerged in a refrigerated water bath (2^oC) to retard microbial metabolism. On each sampling day, both overflow and the filtered fraction of the effluent for each fermenter were combined and homogenized (T25 Basic Homogenizer; IKA Labortechnik, Wilmington, NC) for 5 min. A 600-mL sample was removed via vacuum aspiration during homogenization. At the end of each period, the samples of each fermenter from each of the 3 d were composited and lyophilized to a constant weight. Dried samples were ground through a 1-mm screen and used for all subsequent analyses. During homogenization of the effluent, subsamples were taken for VFA analysis.

Bacterial samples were collected from fermenter contents on the last sampling day of each experimental period (2 h post feeding) by straining the total content of each fermenter through eight layers of cheesecloth. This strained fluid was centrifuged at 500 × g for 20 min to remove feed particles and protozoal cells. Bacteria were separated from the supernatant by centrifuging at 26,000 × g for 20 min. Bacterial samples then were lyophilized and ground by using a mortar and pestle.

Absolute DM determination was conducted on freeze-dried digesta, bacterial, and diet samples by drying at 105^oC for 24 h, followed by ashing at 500^oC for 16 h in a muffle furnace to determine OM. The neutral detergent fiber (NDF; Jeraci et al., 1988) and ether extract (EE; AOAC, 2000) concentrations of the diets were determined. Total nonstructural carbohydrates (TNC) in the diet and digesta samples were determined as described by Smith (1969). Effluent samples were prepared for VFA analysis by the procedure of Erwin et al. (1961). Concentrations of VFA were determined by gas chromatography (Varian Model 3800, Varian Inc., Walnut Creek, CA). The CP content of the diets was determined by the macro-Kjeldahl procedure (AOAC, 2000). Purine concentrations in effluent and bacterial samples were determined by using the method of Zinn and Owens (1986). Digestibilities were calculated as described by Hannah and Stern (1985).

The data were analyzed as a randomized complete block design by using the GLM procedures of SAS (SAS Inst. Inc., Cary, NC). Because treatments were arranged as a 2 × 2 factorial, treatment sums of squares were separated into the main factors (ethoxyquin and monensin/tylosin) and their interactions. Because no interactions (P > 0.05) were detected for most of the measurements evaluated, means of the main factors were separated by ANOVA.

The diet examined in this study was previously evaluated in the feedlot (McBride, 2000). During a 182-d finishing period, beef steers receiving ethoxyquin had 10.9% greater (P < 0.05) ADG and 18% more (P < 0.05) DM intake than those fed the control diet. During the last 14 d of the trial, steers receiving ethoxyquin gained more (P < 0.05) weight than those fed the control diet (1.53 vs 1.06 kg/d). Because of the significant and consistent feedlot benefits observed in the Texas A&M trial (McBride, 2000), it was important to evaluate this diet in the current study.

Digestibility of OM and TNC are summarized in Table 3. No interactions (P > 0.05) between ethoxyquin and monensin/tylosin supplementations were detected and, therefore, results of the main factors are presented. Apparent OM digestibility tended to increase (P = 0.05) with feeding ethoxyquin. Feeding ethoxyquin improved (P < 0.05) true OM digestibility (corrected for bacterial OM in the effluent) by 16.0%. No studies on site and extent of digestion of diets containing ethoxyquin were found. However, performance studies with feedlot cattle have shown improvement in ADG and DM intake (McBride, 2000) with feeding ethoxyquin at the 150 ppm level used in our study. Our data (Table 3) suggest that the improved DM intake (McBride, 2000) was a direct result of increased (P < 0.05) true digestibility of OM.

Apparent OM digestibility tended to increase (P = 0.14) by 11.8% with monensin/tylosin supplementation (Table 3). True OM digestibility also tended to increase (P = 0.12) by 9.5%. The effects of monensin/tylosin on OM digestibility were investigated in vivo (Zinn, 1987), and results indicated that apparent OM digestibility was numerically increased (from 54.1 to 55.9% of intake; P > 0.05) by such supplementation. These data and ours (Table 3) are in agreement with the conclusion of Schelling (1984), who indicated that monensin does result in a slight to moderate improvement in OM digestibility, and the magnitude of the response was attributed to ruminal factors such as DM intake, fill, and rate of passage.

Digestibility of TNC (Table 3) was not affected (P > 0.05) by ethoxyquin or monensin/tylosin supplementation and averaged 86%. No data were found on the effects of ethoxyquin on ruminal TNC digestion. Because TNC is composed mainly of starch and a small amount of sugars, our data (Table 3) are in agreement with those of Zinn (1987). He indicated that monenisn/tylosin supplementation did not alter (P > 0.05) starch digestion (averaging 77.9%) in the rumen of cattle consuming a corn-based finishing diet.

Concentrations of VFA in the effluent are presented in Table 4. With the exception of butyrate and isobutyrate, no interactions (P > 0.05) between ethoxyquin and monensin/tylosin supplementation were detected for concentrations of total or individual VFA. Concentrations of butyrate and isobutyrate were highest (P < 0.05) for the diet containing ethoxyquin without monensin/tylosin (30.6 and 0.33 mM, respectively) and were lowest (P < 0.05) for the diet containing monensin/tylosin without ethoxyquin (15.5 and 0.12 mM, respectively). Table 4 summarizes the effects of the main factors on concentrations of VFA in the effluent. Concentrations of total VFA and acetate were not affected (P > 0.05) by ethoxyquin or monensin/tylosin supplementation and averaged 131 and 59 mM, respectively. Feeding ethoxyquin decreased (P < 0.05) propionate concentration by 17% and increased (P < 0.05) butyrate concentration by 25%. Monensin/tylosin supplementation increased (P < 0.05) propionate concentration by 23% and decreased (P < 0.05) butyrate concentration by 37%.

No effects were detected (P > 0.05) for ethoxyquin or monensin/tylosin supplementations on concentrations of isovalerate or valerate (Table 4). The average values were 0.47 and 3.77 mM, respectively. Table 4 also shows that isobutyrate concentration was increased (P < 0.05) by 85% with feeding ethoxyquin, and it was decreased (P < 0.05) by 39% with monensin/tylosin supplementation. Because isobutyrate is produced by the deamination of the branched-chain amino acid valine (Harwood and Canale-Parola, 1981), ethoxyquin appeared to enhance the ruminal degradation of dietary valine without affecting the rate of degradation of other branched-chain amino acids. It is also possible that ethoxyquin reduced isobutyrate utilization. Based on the limited information available on the mode of ethoxyquin action at the rumen level, no explanation to this effect can be provided. It should be noted, however, that the reduced (P < 0.05) degradation of valine to isobutyrate when monensin/tylosin were supplemented is consistent with the role of monensin in decreasing protein degradation in the rumen environment (Whetstone et al., 1981).

Results of this study illustrate the positive effects of dietary supplementation of the antioxidant ethoxyquin (150 ppm) on ruminal digestion of OM. This may explain why feedlot cattle in several investigations gained faster or consumed more feed when ethoxyquin was fed. The modes of such actions, however, remain to be elucidated.