1st Page

Journal of Animal Science - Article



This article in JAS

  1. Vol. 88 No. 2, p. 802-810
    Received: Dec 17, 2008
    Accepted: Oct 05, 2009
    Published: December 4, 2014

    2 Corresponding author(s):


Pain mitigation after band castration of beef calves and its effects on performance, behavior, Escherichia coli, and salivary cortisol1

  1. L. A. González*,
  2. K. S. Schwartzkopf-Genswein 2,
  3. N. A. Caulkett,
  4. E. Janzen,
  5. T. A. McAllister*,
  6. E. Fierheller,
  7. A. L. Schaefer,
  8. D. B. Haley§,
  9. J. M. Stookey# and
  10. S. Hendrick#
  1. Agriculture and Agri-Food Canada, Lethbridge Research Centre, Lethbridge, Alberta T1J 4B1, Canada;
    University of Calgary, Calgary, Alberta T2N 4N1, Canada;
    Agriculture and Agri-Food Canada, Lacombe Research Centre, Lacombe, Alberta T4L 1W1, Canada;
    University of Alberta, Edmonton, Alberta T6G 2P5, Canada; and
    University of Saskatchewan, Saskatoon, Saskatchewan S7N 5B4, Canada


Eighty-nine steer and bull calves (210 ± 20 d of age) were randomly assigned to a 2 × 2 factorial arrangement of treatments to study the effect of sham or band castration (CAST) with or without pain medication (MED) on performance, behavior, cortisol, and Escherichia coli shedding. Steers (serving as controls; C) were previously castrated at 34 ± 10 d of age, whereas bulls (B) remained intact until band application on d 20 of the experiment. One-half of the calves in each group received pain medication (M; xylazine epidural and intravenous flunixin meglumine) or saline (NM) solutions at 0.5 h before castration. Feed intake, ADG, and feeding behavior were measured over a 3-wk period before (used as covariates) and a 6-wk period after castration. Fecal and saliva samples were collected from 32 calves for E. coli enumeration at −7 (covariate), 2, 5, and 7 d, and for cortisol concentration at −0.5 (covariate), 0, 0.5, 1, 2, 4, 24, 48, 168, and 336 h relative to castration, respectively. Lying behavior was measured in the home pen and step length at exit from the chute. There were no significant CAST × MED interactions on growth rate or feed intake. Overall ADG was less in B compared with C (P < 0.001), but feed intake was less during wk 4 only (P = 0.01). On average, M had less intake (P = 0.02) and eating rate (P = 0.04) than NM calves. The BM had less fecal E. coli counts than CM (P = 0.05) and castrated and nonmedicated calves (P = 0.08). Salivary cortisol was less in BM compared with the rest of treatments at 1 and 2 h after banding (P ≤ 0.05). Castration reduced time spent lying (P = 0.06) and step length (P = 0.01) until d 28 after banding, whereas M reduced lying and step length on day of banding only (P ≤ 0.06). Significant CAST × MED interactions were observed for frequency of visits to feeders and meal size, length, and frequency (P < 0.05) independent of the time after castration. This was due to greater values in the noncastrated and non-pain-medicated group compared with the rest, except for reduced meal frequency, indicating that pain medication and castration both affected feeding behavior. The medication protocol used in the present study eliminated the acute cortisol surge, which may be linked to reduced fecal E. coli counts observed during wk 1 after banding. However, feed intake, growth, and behavior suggested that pain was greatest during wk 3 and 4 after band application, which highlights the need for chronic pain mitigation strategies.


Male cattle are castrated to facilitate management by reducing aggressive and sexual behaviors and to improve meat quality (Appleby, 1986; Baker and Gonyou, 1986; Faulkner et al., 1992). However, castration may reduce performance and is considered as one of the most painful experiences of cattle leading to physiological stress, suppression of immune response, and pain-related behaviors (AVMA-AWD, 2007; Stafford, 2007). Band castration is widely used and consists of placing a rubber ring on the neck of the scrotum to interrupt blood supply. This results in chronic ischemia and coagulation necrosis until final detachment of the scrotum and its contents after 35 to 65 d, and results in acute and chronic pain and stress (Chase et al., 1995; Molony et al., 1995; Knight et al., 2000). Consequently, pain mitigation strategies need to be developed to avoid the negative effects of castration. Anesthesia, such as xylazine epidural, has effectively been used to avoid pain during the castration procedure, providing sedation and blocking the nerves (Caulkett et al., 1993; Ting et al., 2003). Analgesics, such as nonsteroidal anti-inflammatory drugs, have also been used to mitigate pain secondary to castration. For instance, ketoprofen eliminated the acute cortisol response in banded calves (Stafford et al., 2002); however, carprofen did not (Pang et al., 2006). As opposed to these latter drugs, flunixin meglumine is approved in the United States for use in cattle, has anti-pyretic properties, and its accumulation in inflammatory tissues may prolong its effect (Landoni et al., 1995; Smith et al., 2008). This combination of anesthesia and analgesia may help to mitigate pain during and after the banding procedure. The hypothesis of the present study was that band castration or the combination of anesthesia with caudal xylazine epidural and analgesia with intravenous (i.v.) flunixin meglumine does not affect performance, cortisol, fecal Escherichia coli shedding, and behavior in weaned beef calves.


All procedures described within this study were approved by the Animal Care and Use Committee of the Lethbridge Research Centre and according to the guidelines established by the Canadian Council on Animal Care (1993).

Animals, Experimental Design, and Facilities

All Angus calves, composed of 43 steers and 46 bulls (BW = 277 ± 23 kg; age = 210 ± 20 d), came from the same calf crop of the Agriculture and Agri-Food Canada One-Four experimental ranch located in southeastern corner of Alberta, Canada. Steers were castrated at the ranch when they were 34 ± 10 d of age, whereas bulls remained intact until they were transported to the Lethbridge Research Centre experimental feedlot. Calves were left to adapt to their pens and feed for 1 mo before commencing the study. At the start of the study, calves were blocked by BW and age and then randomly assigned to 1 of 4 treatments within each block according to a 2 × 2 factorial arrangement of treatments. Calves were then assigned randomly to 1 of 4 pens (20 to 24 calves/pen) to achieve approximately the same number of animals per treatment and pen because all treatments were represented within each pen. Main factors were castration (CAST) and pain medication (MED). All bulls were castrated with the application of a rubber band (B; Callicrate Bander, No-Bull Enterprises Inc., St. Francis, KS) on d 20 after the start of the experiment. Steers were sham castrated (C) by holding them in the chute for a similar amount of time and subjecting them to the same procedures used to handle and castrate bulls with the exception that they were not banded. One-half of the calves within each castration group received pain medication drugs (M) or saline solutions (NM) at 30 min before band application. Medication consisted of combined anesthesia through xylazine epidural at a dose of 0.07 mg/kg of BW (Rompun 20 mg/mL injectable, Bayer Health, Toronto, Canada) and systemic analgesia through an i.v. jugular injection of flunixin meglumine at a dose of 1.1 mg/kg (Banamine, Schering-Plough Animal Health, Kirkland, Québec, Canada). In all calves, epidural injection was administered first, and i.v. flunixin followed. Each calf was examined before castration for completeness of anesthesia by a lack of tail tone and response to needle pricking around the perineum. All focal animals for which detailed physiological and behavioral measures were required were castrated the same day, whereas the remaining animals were castrated the day after the focal group to accommodate sampling times. In addition, calves were subdivided into subgroups of 6 to allow the approximate time required (30 min) between the administration of drugs and the actual castration procedure. Thus, all 6 calves were given injections sequentially at 5-min intervals and then started to enter the chute for castrations in the same order every 5 min. The number of animals per treatment for the noncastrated and non-pain-medicated (CNM), noncastrated and medicated (CM), castrated and nonmedicated (BNM), and castrated and medicated (BM) groups was 23, 20, 22, and 24, respectively.

All animals were identified with radio frequency ear tags (Allflex Canada, St-Hyacinthe, Canada) and trained and became accustomed to the facilities during the month before the start of the study by handling them gently 5 times through the facility in which castration took place. Each outdoor pen measured 40.2 × 27.4 m and was protected with windbreak fencing to the west and north, contained a centrally located water system (Bolhmann Inc., Denison, IA), and had a concrete apron (2.4 × 24.5 m) directly in front of the feeders. Straw bedding was added as needed in one corner of all pens. A digital video camera (Panasonic WV-CP474, Mississauga, Canada) was set up in each pen to record behavior of the animals in their home pen. Video was captured using Omnicast, Genetec (Dorval, Canada) software and stored on a personal computer. Another video camera was placed 4 m in front and at a 90° angle to a 15-m-long dirt alley located directly off the headgate and squeeze chute where castrations took place. This camera allowed for the recording of individual animal step length (front and back) as the calves exited the chute (Currah et al., 2009).

Five electronic feeders per pen allowed automatic recording of individual feed intake and time at the feeder (GrowSafe Systems Ltd., Airdrie, Alberta, Canada). Each feeder measured 0.91 × 0.38 × 0.53 m (height × depth × width), was mounted onto 2 load cells, and allowed only 1 calf to eat at a time. Feeders were equipped with an antenna placed on the rim that detected electromagnetic signals from ear transponders when they were within approximately 0.5 m. A reader panel recorded 1 reading every 2 s and identified the transponder number present above the feeder, feeder number, feed weight, and time of day. All data were stored into a personal computer for later analysis. Cattle were fed a total mixed back-grounding ration (57.6% DM) for ad libitum intake consisting of 61.5% barley silage, 16.4% rolled barley, 17.1% rolled oats, and 5% supplement containing minerals and vitamins (DM basis) for the first 45 d of the experiment. During the last 28 d of the experiment, calves were fed a ration (57.1% DM) containing 62, 33.1, and 4.9% (DM basis) of barley silage, steam rolled barley, and supplement, respectively. Feed was delivered twice daily at 0900 and 1300 h, and fresh water was available at all times.

Sampling and Procedures

The experiment was conducted for a total of 9 wk with data being collected 3 wk before and 6 wk after castration. All animals were weighed at 0800 h for 2 consecutive days at the beginning (d −21) and end of the experiment, with intermediate BW taken weekly beginning the day of castration (d 0) until the end of the study (d 42). Individual feed intake and feeding behavior was registered daily throughout the experiment using GrowSafe Systems Ltd. A subsample of 32 focal calves (2 calves per treatment per pen) was randomly selected to obtain more detailed behavioral and physiological measures. These calves were identified with tape marks on their backs and flanks, and their resting behavior in their home pen was measured at d 2, 7, and 14 after castration. To achieve this, start and end times (to the nearest second) of lying were recorded from 0740 to 1640 h. Saliva samples were taken with a cotton swab from this subsample of calves and immediately frozen at −20°C for later cortisol analysis as described by Cook and Schaefer (2002). The intra- and interassay CV were 12.3 and 15.6%, respectively, for samples containing 1.23 ng of cortisol/mL of saliva. Saliva samples were taken immediately before epidural and i.v. injections, which was at −0.5 h, and then at 0 (immediately before castration), 0.5, 1, 2, 4, 24, and 48 h, and then at 7 and 14 d, relative to castration. Mean cortisol concentration during 4 and 24 h, and 14 d after castration were calculated with the area under the curve (assuming the change between times was linear) divided by the number of hours. Finally, fecal samples also were taken from this subset at d −7, 2, 5, and 7 by rectal palpation for enumeration of total E. coli as described by Bach et al. (2004).

Feeding behavior was recorded as visits to the feeders. A feeding visit ended when a calf was last recorded in a feeder before being recorded in a different feeder (change of feeder), or in the same feeder after 5 min of absence, or when a different transponder was recorded in the same feeder (replacement by another animal). The length of each visit was calculated as the time since the animal was first registered until the time the visit ended due to the reasons mentioned above. Daily feeding time was the sum of the length of all visits within a day (min/d) and feeding rate was calculated as daily DMI divided by feeding time (g of DM/min). To pool feeding visits into meals, the meal criterion was calculated as the longest interval between visits that was still accepted as part of a meal. This was determined to be 18.2 min from the frequency distribution of the nonfeeding interval length pooled across all animals (Yeates et al., 2001). The meal criterion allowed pooling of feeding visits by meals to calculate meal size (g of DM/meal) and duration (min/meal), which was the sum of feed eaten, number of visits, and length of visits within a meal, respectively. Meal frequency was the number of times per day that a nonfeeding interval length exceeded the meal criterion. Meal length was calculated as the elapsed time between the start of the first visit until the end of last visit before the meal criterion occurred (including time within meals when calves remained out of the feeders). Visit frequency was calculated as the numbers of feeding visits per day (number/d) and per meal (number/meal), which were used as indicators of the activity of the animal at the feeding area. All these calculations concerning feeding behavior were determined for each animal using SAS (SAS Inst. Inc., Cary, NC).

Step length was measured at d −1 (before castration), 0 (at exit from castration), 7, 14, and 28 after being castrated. Two photographs for each calf were taken from the videos (Adobe Premiere Elements 3.0, Adobe Systems Inc., San Jose, CA) as the calves traversed a portion of the alley located at the end of the squeeze chute. The alley was 1.5 m wide and 15 m long; pictures were taken between the 10th and 14th m. Photographs were taken only while both back (or front) hooves were contacting the ground, which was at the time the calf placed one hoof on the ground and before it lifted the opposite hoof while walking. Step length was the distance measured from the middle of 1 hoof to the middle of the opposite with the help of a software application (ImageJ, Bethesda, MD) after calculating the distance between 2 reference points in the photograph. The reference points were the fence posts of the alley.

Statistical Analysis

All data collected before castration were used as covariates for each variable analyzed during the postcastration period. Average daily gain, feed intake, and feeding behavior values of the 3-wk period before castration were used as covariates on each out of the 6-wk postcastration period, whereas salivary cortisol concentration at −0.5 h, step length on d −1, and fecal E. coli counts at d −7 were the covariates for the postcastration analysis. A mixed-effects model with repeated measures was used (PROC MIXED; SAS) considering the fixed effects of covariate, CAST, MED, CAST × MED, week, CAST × week, MED × week, and CAST × MED × week interactions. The random effects considered were pen to account for the relationship among animals sharing the same pen: pen × CAST × MED interaction, which was the error term for testing the effect of treatments (n = 4 per treatment); pen × week × CAST × MED interaction, which was the error term for testing the effect of treatments at a given point in time; and animal within pen to which the repeated measures of time were subjected (St. Pierre, 2007). Degrees of freedom were calculated using the Kenward-Rogers method, and the covariance structure was chosen accordingly to fit the statistics and characteristics of the variable tested. The spatial power function, unstructured covariance structure was used under unequal distances among repeated measurements (e.g., cortisol) or heterogeneous covariance when needed. Least squares means differences were corrected by a Bonferroni test for multiple comparisons.


Average daily gain was not affected by pain medication, but it was reduced by CAST (P < 0.001; Table 1). In addition, the CAST effect on growth was different by week over the experimental period after banding (CAST × week; P = 0.07) due to decreased growth rate in B compared with C calves during wk 1 (P = 0.03), 2 (P = 0.04), 3 (P = 0.001), and 5 (P = 0.07; Figure 1A), whereas both groups grew at similar rates during wk 4 and 6 (P > 0.10). Pain medication reduced DMI independent of the time after its administration (7.88 vs. 7.65 ± 0.120 kg of DM/d for NM and M, respectively; P = 0.02; Table 1), but CAST had no effect (P = 0.40). However, the CAST × week interaction (P = 0.10) indicated that CAST groups did not have the same DMI over the data collection period because B ate less than C calves on wk 4 (P = 0.01; Figure 1B). The precastration values (covariate) affected DMI after CAST (P < 0.001), but not ADG (P = 0.34; data not shown). Precastration values for DMI were 7.10 and 7.09 ± 0.15 kg of DM/d, and ADG were 0.79 and 0.96 ± 0.067 kg/d for C and B, respectively. It is also important to note that removing the covariate from the model did not change the mean values, differences among means, or significance of fixed effects.

Table 1.

Please see the pdf to view this table.

Figure 1.
Figure 1.

Performance (A) and feed DMI (B) of beef calves after sham (not castrated) or band castration (castrated). Within sampling week, means differ at P ≤ 0.001 (***), P ≤ 0.01 (**), P ≤ 0.05 (*), and P ≤ 0.10 (†; trend). Data reported are based on approximately 22 animals/treatment.


Total fecal E. coli showed a CAST × MED interaction (P = 0.02) because BM calves shed less than CM calves (P = 0.05) and tended to shed less than BNM (P = 0.08; Table 1). Both band castration (P = 0.05) and the lack of pain medication (P = 0.03) resulted in greater mean salivary cortisol concentration during the 4-h period after castration, but there were no effects on 24-h and 14-d mean values (Table 1). In addition, all the interactions with time tended to be significant (CAST × time, P = 0.07; MED × time, P = 0.06; CAST × MED × time, P = 0.14). The 3-way interaction indicated that BNM had greater salivary cortisol concentrations compared with all groups at 1 (P ≤ 0.05) and 2 h after castration (P ≤ 0.01; Figure 2). Band-castrated calves exhibited greater salivary cortisol concentrations than C calves only at 2 h (P < 0.02; data not shown), whereas NM calves had greater salivary cortisol concentrations at 0.5, 1, and 2 h than M (P < 0.05; data not shown). Finally, the initial cortisol concentration as indicated by the covariate affected all mean values as well as the postcastration kinetic at the individual level (P ≤ 0.02; data not shown). However, mean cortisol concentrations at 0.5 h before CAST did not differ among CAST groups, 1.60, 1.30, 1.21, and 1.70 ± 0.45 ng/mL for CNM, CM, BNM, BM, respectively (P > 0.10; data not shown).

Figure 2.
Figure 2.

Salivary cortisol concentrations of calves subjected to sham (C) or band castration (B) and receiving epidural plus intravenous solutions of saline (NM) or xylazine plus flunixin meglumine (M). Hour zero corresponds to saliva sampling immediately before castration and 30 min after drug administration. Within sampling time, mean of band castrated calves not receiving pain medication (BNM) differs from the rest at P ≤ 0.01 (**) and P ≤ 0.05 (*). Data reported are based on 8 animals/treatment.


Calves tended to spend less time lying down if they were castrated (31.4 vs. 24.2 ± 3.00% for C and B, respectively; P = 0.06) or if they were not medicated for pain (24.7 vs. 30.8 ± 3.01% for NM and M, respectively; P = 0.10; Table 1). Length of steps measured for the back legs were reduced by castration (53.9 vs. 51.5 ± 0.67 cm/step for C and B, respectively; P = 0.01) or pain medication (53.6 vs. 51.8 ± 0.66 cm/step for NM and M, respectively; Table 1). However, the greatest difference in step length due to castration was observed on d 0 (P = 0.002), followed by d 14 and 28 (P = 0.02), but no differences were observed at d 7 (P = 0.41; Figure 3A). Similarly, the MED × day interaction (P < 0.001) indicated that MED reduced back step length on d 0 only (P < 0.001; Figure 3B). Step length in the front legs tended (P = 0.06) to be reduced by castration (54.4 vs. 52.2 ± 1.22 cm/step for C and B, respectively) but was not affected by pain medication (P = 0.38). However, the MED × day interaction resulted from a significant reduction in front leg step length on d 0 when exiting the chute after CAST, with mean values of 52.7 vs. 48.5 ± 1.38 cm/step (P = 0.004; data not shown).

Figure 3.
Figure 3.

Step length of back legs in beef calves after sham (not castrated) or band (castrated) castration (panel A), and after administration of epidural plus intravenous solutions (panel B) of saline (not medicated) or xylazine plus flunixin meglumine (medicated). Day 0 was at exit from the chute upon castration and medication. Within sampling day, means differ at P ≤ 0.001 (***), P ≤ 0.01 (**), and P ≤ 0.05 (*). Data reported are based on approximately 22 animals/treatment.


The covariate used for all feeding behavior measurements presented in Table 2 was significant (P < 0.001; data not shown). Daily feeding time was not affected by CAST or by MED (P ≥ 0.45). Feeding rate was not affected by castration (P = 0.52); however, MED reduced feeding rate 63.6 and 60.1 ± 1.03 g of DM/min for NM and M calves, respectively (P = 0.03). The frequency of feeding visits per day and per meal showed significant CAST × MED interactions because CNM calves visited the feeders more frequently than the rest of treatments (P ≤ 0.05; Table 2). Therefore, MED in C calves reduced the frequency of visits to the feeders to the same extent as CAST did (P < 0.05). In contrast, meal frequency showed a CAST × MED interaction (P = 0.04) due to CNM calves having fewer meals per day than the rest of treatments (P < 0.05; Table 2). As a result, meals were larger and lasted longer (P < 0.05) in CNM compared with the rest of the treatments, although only a tendency was found between CNM and BM for meal length (P = 0.06). However, meal duration, which accounts for time actually spent at the feeders per meal, was not affected (P > 0.10; Table 2).

Table 2.

Please see the pdf to view this table.



Band castration of 6- to 8-mo-old weaned intact calves in the present study reduced their growth rate by 31% throughout the 6-wk postcastration period with no effects on feed intake. In agreement with results of the present study, band castration of 400-kg bulls reduced ADG by 55% compared with intact bulls (Chase et al., 1995) and 43% compared with steers during the first month (Knight et al., 2000). In contrast, other studies reported that castration with rubber rings or banding did not significantly affect ADG in 1- to 20-wk-old male cattle (Molony et al., 1995; Stafford et al., 2002; Pang et al., 2006) and even increased ADG of Angus bulls during the first week (Chase et al., 1995). These differences in growth among studies appear to be related to age at the time of banding (the older the calves, the greater the impact of banding on growth performance), hormonal status of the control group, and length of the experimental period after banding. Feed intake was not affected in any of the previously cited studies, which agree with the present study, indicating that banding reduces the availability of nutrients or energy for growth toward other functions with greater priority (Gabler and Spurlock, 2008). For example, castrated calves may have greater energy expenditure due to the stressors (pain) associated with banding, or energy may be reallocated to other functions such as immune response and tissue repair (inflammation and tissue healing). It should be noted that the effect observed on ADG in the present study cannot be attributed to differences in hormonal profiles, as in other studies, because the control groups were steers.

Body weight did not differ among groups at the start of the study (d −21); however, on d 0, B calves were heavier than C. This is most likely a result of the anabolic effects of hormones such as testosterone during the 3-wk precastration period (covariate), which led to faster growth of bulls (data not shown). Nevertheless, B were lighter than C from d 21 until the end of the experiment when BW were 331.6 vs. 323.9 ± 1.5 kg for C vs. B, respectively. In the present study, differences in growth rate among castration groups were largest during wk 3 (B grew 0.68 kg/d slower than C), and feed intake was less in B compared with C calves during wk 4 only. Weeks 3 and 4 after banding were observed to be the time that testicles dropped off and the time of greatest inflammation in banded bulls (Molony et al., 1995; Knight et al., 2000; Pang et al., 2006). Therefore, growth and appetite appears to be most affected when inflammation and tissue and nerve trauma are greatest. By the end of the study (d 42), only 60% of banded calves had completely lost their scrotum and contents, whereas 84% lost them by d 50. However, the scrotum and its contents had dried up some time before final detachment and a large open wound may develop afterward (Stafford et al., 2002). More research is needed to determine the relationship between tissue trauma and inflammation, pain, feed intake, growth rate, and feed efficiency.

Medication for pain mitigation did not affect growth rate, but reduced daily DMI and feeding activity regardless of the time after its administration, as observed by the reduced frequency of feeding visits and feeding rate. There is no explanation for this long-term effect, and it is difficult to assess whether the cause of this was the xylazine epidural, the i.v. flunixin, or both. Nevertheless, we speculate this is a negative side effect arising from the xylazine epidural as stated by Stafford and Mellor (2005). For instance, calves may have suffered distress after the analgesic and anesthetic effect of drugs was gone or experienced permanent changes in social status that lead to permanent changes in behavior. Faulkner et al. (1992) also reported that i.v. administration of butorphanol and xylazine upon castration reduced feed intake over a 27-d period after administration. However, other studies have shown that feed intake and ADG were not affected after several pain mitigation strategies using xylazine and lidocaine caudal epidural anesthesia after burdizzo (Ting et al., 2003) or i.v. carprofen after banding (Pang et al., 2006). It is worthy to note that only the CM group ate 0.6 kg/d less DMI during wk 1 after banding, with values being 7.59, 6.87, 7.35, and 7.42 ± 0.208 kg of DM/d for CNM, CM, BNM, and BM, respectively. This observation may help to explain the results on fecal E. coli shedding. The greatest fecal E. coli counts were observed in CM calves that had the least feed intake of all groups during wk 1. This may be due to the fact that shorter transit time and turnover rate of feed through the gastrointestinal tract results in greater fecal E. coli. In contrast, BM showed the least fecal E. coli shedding during the first 7-d after castration and a lack of the acute cortisol response. Medication upon castration reduced glucocorticoids, pain, and stress, contributing to maintain reduced E. coli populations in the gut related to stronger immune status. Glucocorticoids play a major role in the reduction of the immune status and response (Dhabhar, 2002; Webster et al., 2002). The stress caused by long transport times and the administration of exogenous corticoids have also increased fecal E. coli shedding (Bach et al., 2004; Sreerama et al., 2008). Other research groups working with castration and pain medication did not find support for an alteration of the immune status as a result of greater pain (Chase et al., 1995; Pang et al., 2006). It is worthy to notice that all treatments were mixed within each pen in the present experiment. Although E. coli cross-contamination can occur between pen-mates, the statistical model considered the relationships between all animals in a pen (the experimental unit). In addition, we measured changes over time in E. coli shedding as a result of castration, and the initial E. coli count measured before castration was used as covariate at the individual level, which was significant. Therefore, all precautions were taken to avoid confounding factors.

Cortisol concentration has extensively been used as an indicator of the degree of pain and physiological stress suffered by castrated cattle (AVMA, 2007; Stafford, 2007). The increase in salivary cortisol concentrations of nonmedicated banded bulls in the present study indicates that acute pain and physiologic stress were experienced between 1 and 4 h after band application. This is in agreement with studies conducted by Fell et al. (1986), Stafford et al. (2002), and Pang et al. (2006). However, salivary cortisol after 4 h and up to 14 d postbanding revealed no effects in the present study, which is in contrast to the increase in blood cortisol observed by Chase et al. (1995) at 48 h postbanding. The lack of treatment differences in salivary cortisol at 0.5 h after band application suggest that this medication may be administered at the time of banding, a management practice that would reduce handling of the animal and associated labor. However, evidence of acute pain was observed in banded bulls taking shorter steps when exiting the chute after band application, indicating calves were already feeling discomfort at this time. In conclusion, behavior as measures of pain indicated that banded bulls exhibited chronic signs of pain despite the fact that salivary cortisol had returned to basal concentrations at 4 h after banding. This indicates that salivary cortisol may be ineffective as an indicator of immediate (procedural) and chronic pain, but a good indicator of acute pain experienced in the first 1 to 2 h after CAST.

Behavior is likely to be the most practical tool for assessing pain and is noninvasive and very specific to different types of pain when quantified in detail (Lester et al., 1996; Rutherford, 2002). In the present study, we used quantifiable and objective measurements to assess acute and chronic pain secondary to band application. Banding increased standing time, likely as a result of the restlessness caused by the discomfort and pain, as reported by Molony et al. (1995) in 1-wk-old calves. Molony et al. (1995) also advocated for the development of improved techniques to measure chronic pain between 20 to 36 d after ring application, which were addressed in the present study. In addition to increased standing time, banded calves in the present study also had reduced feeding activity and made shorter steps while walking throughout the 6-wk study period. These results suggest that banded calves may have been more reluctant to walk and change from one feeder to another because of pain. Dairy cows suffering from lameness also made fewer feeding visits per day to reduce the painful activity of walking (González et al., 2008). However, those lame cows also showed shorter daily feeding duration, whereas eating rate increased markedly as a compensating mechanism because standing was also painful. Calves in the present study appeared to experience pain while walking and moving from one feeder to another rather than while standing. Meal length of castrated calves was shortened although they returned to feed at other times to compensate daily feed intake through more meals per day. The possibility exists that pain reduced the number of visits per meal and, therefore, the short-term appetite within the course of the meal because meals ended sooner in castrated animals.

Banding reduced lying time, frequency of visits (feeding activity), and step length regardless of time postbanding, suggesting it causes chronic pain (6 wk). In contrast, pain medication tended to increase the time spent lying and further reduced the length of steps during the first few hours after drug administration in C and B groups. This suggests that drugs used in the present study had behavioral effects unrelated to pain and nociception, which may be a result of the sedative and depressive effect of the xylazine epidural on cardio-respiratory function and activity (Caulkett et al., 1993; Rutherford, 2002; Meyer et al., 2007). In contrast to the results of the present study, Currah et al. (2009) reported bulls receiving lidocaine epidural anesthesia plus i.v. flunixin meglumine had longer strides during the first 12 h after surgical castration compared with nonmedicated counterparts, but no differences were observed after the drugs wore off. Such discrepancy with our results is likely due to the lack of depressive effect of the lidocaine epidural on cardio-respiratory function (Meyer et al., 2007). The pain medication protocol used in the present study was also effective in reducing the acute cortisol response observed at 1 and 2 h after band application. To date there are no published studies using this same protocol upon band castration, and therefore, comparisons are difficult. However, local anesthetic (lignocaine 2%) or i.v. ketoprofen eliminated the rise in plasma cortisol after rubber ring or band application in 100-kg bulls (Stafford et al. 2002), whereas carprofen did not in 5-mo-old bulls (Pang et al., 2006).

Band castration and pain medication with xylazine epidural and i.v. flunixin have reduced the activity of calves under intensive housing conditions of the present experiment. However, it would be interesting to know the extent of such effects under extensive conditions where animals are required to walk for longer distances and times searching for food and water. In addition, medications used in the present study were administered at standard or manufacturer-recommended doses. More research is needed to look for the optimum dose to administer, although it may be specific for each castration method (e.g., surgical and band castration).

In conclusion, the greatest benefits of pain medication with xylazine epidural and i.v. flunixin upon castration of 6-mo-old bulls were 1) reduction of fecal E. coli counts during the first week after banding and 2) elimination of the acute physiological stress and pain, as indicated by salivary cortisol during the first 4 h after band application. However, such medication may slightly reduce appetite and activity as a consequence of the sedative effect of the xylazine epidural. Long-term effects in performance and behavior secondary to castration were not reflected in salivary cortisol, which emphasizes the need of alternative quantifiable and objective indicators of pain. These effects also indicate that calves experience chronic pain lasting for at least 6 wk and emphasize the need of long-term pain mitigation strategies.