Antimicrobial Metaphylaxis and its Impact on Health, Performance, Antimicrobial Resistance, and Contextual Antimicrobial Use in High-Risk Beef Stocker Calves

Abstract

The objective of this blinded, cluster-randomized, complete block trial was to evaluate the impact of metaphylaxis on health, performance, antimicrobial resistance, and contextual antimicrobial use (AMU) in high-risk beef stocker calves. Calves (n = 155) were randomly assigned to receive either saline or tulathromycin at the time of arrival processing. Deep nasopharyngeal swabs were collected from each calf at arrival and 14 d later. Calves were monitored for bovine respiratory disease (BRD) for 42 d. Body weights were obtained at arrival, days 14, 28, and 42. Contextual antimicrobial use (AMU) was calculated using dose and mass-based metrics. Calves given tulathromycin had a greater average daily gain (0.96 ± 0.07 kg vs. 0.82 ± 0.07 kg; P = 0.034) and lower prevalence of BRD than controls (17% vs. 40%; P = 0.008). Proportions of calves with BRD pathogens identified at arrival were similar between treatment groups [17%; P = 0.94]. Proportions of calves with BRD pathogens identified at day 14 were lower for calves receiving tulathromycin compared to controls (15% vs. 60%, P < 0.001). Overall, 81% of Pastuerella multocida isolates and 47% of Mannheimia haemolytica isolates were pansusceptible. When measured as regimens per head in, AMU in calves receiving tulathromycin was higher than calves receiving saline (P = 0.01). Under the conditions of this study, metaphylaxis had positive impacts on the health and performance of high-risk beef stocker calves, did not contribute to the selection of resistant bacterial isolates in the nasopharynx of treated cattle, and increased AMU.

Under the conditions of this study, metaphylaxis had positive impacts on the health and performance of high-risk beef stocker calves, did not contribute to the selection of resistant bacterial isolates in the nasopharynx of treated cattle, and increased antimicrobial use.

Introduction

Bovine respiratory disease (BRD) is the most common cause of morbidity and mortality in North American beef cattle (Tennant et al., 2014; Woolums et al., 2018). Current estimates suggest that BRD costs the beef industry more than $1 billion each year with most economic losses coming from reductions in feed efficiency, poorer carcass quality, and costs of treatment (Griffin, 1997, 2014; United States Department of Agriculture (USDA)., 2013). Therefore, prevention is of the utmost importance as significant losses occur in the face of apparently successful treatment. A segment of the beef industry particularly affected by BRD is the stocker segment (Groves, 2020). Cattle entering this phase of the beef production cycle often have little to no previous health history, have recently been weaned, are transported long distances, and are highly commingled (Groves, 2020; Credille, 2022). These factors, combined with other stressful events, interact to dramatically increase the risk of disease and death in these animals (Smith, 2009).

One of the most effective tools to mitigate the risk of BRD in stocker cattle is metaphylaxis (Nickell and White, 2010; Ives and Richeson, 2015). Defined as the mass administration of an antimicrobial to cattle at the time of initial processing, metaphylaxis has been shown to reduce the risk of morbidity and mortality in cattle at high risk of developing BRD by more than 50% (Tennant et al., 2014; Apley, 2015; DeDonder and Apley, 2015; Coppin et al., 2022). Unfortunately, multiple recent studies have found that the proportion of bacterial isolates classified as multi-, or extensively drug resistant (MDR, XDR), increases significantly between the time of arrival processing and resampling 2 to 4 wk later (Lubbers and Hanzlicek, 2013; Snyder et al., 2017; Crosby et al., 2018; Woolums et al., 2018; Sarchet et al., 2022). One of the major concerns with the increasing prevalence of resistant bacterial isolates is that some cattle given a metaphylactic antimicrobial will go on to develop clinical BRD and require additional antimicrobial treatment (Snyder and Credille, 2020). Cattle producers use metaphylaxis to protect cattle health and welfare, as well as preserve animal productivity and farm profitability (Dennis et al., 2020; Horton et al., 2023). However, cattle harboring resistant bacterial isolates in their airway could become refractory to future treatment (Snyder and Credille, 2020). Nevertheless, little is known about the true role of metaphylactic antimicrobial administration on the selection of resistant bacterial populations in the airway of beef stocker calves because these previous studies did not include untreated control groups.

In recent years, the use of antimicrobials in the production of animal agriculture has come under intense scrutiny (Lhermie et al., 2020). Driven by concerns that overuse of antimicrobials in animal agriculture contributes to antimicrobial resistance challenges being faced in human medicine, various regulatory, industry, and consumer groups have asked producers in different agricultural segments to begin to quantify antimicrobial use (Brault et al., 2019a, b). In addition to being necessary for meeting regulatory and marketing requirements, accurate measurements of antimicrobial use can serve to promote animal welfare, improve animal health management, enhance animal productivity, and preserve antimicrobial efficacy (Lhermie et al., 2020). The availability of accurate and detailed antimicrobial use (AMU) measurements that are tied to reasons for use, as well as the production class antimicrobials, are most often used in, allow for the investigation of the various factors that distinguish high from low AMU producers (Apley, 2018). Ultimately, this information can be used to improve antimicrobial stewardship and animal health management practices on high AMU production operations (Apley, 2018). Therefore, the objective of this study was to evaluate the impact of metaphylaxis on health, performance, and selection of antimicrobial-resistant bacteria in the upper airway, as well as measure contextual AMU in high-risk beef stocker calves that received metaphylaxis, and those that did not.

Material and Methods

This study and all procedures were approved by the University of Georgia Institutional Animal Care and Use Committee (AUP # A2020 01-022-Y3-A0). The experiment was conducted from February 2021 to May 2022 at the University of Georgia Riverbend Stocker Research Facility in Athens, GA.

Cattle description and processing

Crossbred bull, steer, and heifer calves (n = 155) were sourced from a single livestock auction market in Winterville, GA, and randomly assigned to 1 of 2 arrival treatments in a cluster-randomized complete block design. Cattle were shipped 8 km from the auction market to the study facility on February 3, 2021 (block 1), April 1, 2021 (block 2), June 2, 2021 (block 3), October 13, 2021 (block 4), January 19, 2022 (block 5), and March 23, 2022 (block 6). Shipment to the research facility occurred on the day of purchase. Following arrival, calves were inspected for visible signs of illness, as well as noticeable deformities or abnormalities, and penned together overnight in a dirt-floored, open-air receiving pen. Cattle were allowed ad libitum access to water and coastal Bermuda grass hay (Cynodon dactylon, 12% crude protein [CP], 55.7% total digestible nutrients [TDN], 0.275 Megacalories [MC]/lb net energy of gain [NEg], 64.5% neutral detergent fiber [NDF], 35.7% acid detergent fiber [ADF]).

After the overnight rest period, calves were processed and this consisted of recording of individual body weights, placement of a uniquely numbered ear tag, as well as vaccination with a Mannheimia haemolytica leukotoxoid (Presponse SQ, Boehringer Ingelheim Animal Health, Duluth, GA, USA) and 7-way clostridial bacterin (Caliber 7, Boehringer Ingelheim Animal Health). Calves were also treated for internal and external parasites with topical eprinomectin (Eprinex, Boehringer Ingelheim Animal Health) and oral oxfendazole (Synanthic 22.5% Bovine Dewormer Suspension, Boehringer Ingelheim Animal Health). Bulls were surgically castrated by first incising the scrotum with a Newberry castrating knife followed by emasculation with a Serra emasculator. Prior to castration, all bull calves (n = 88) were given meloxicam (1 mg/kg of BW orally, Meloxicam 15 mg tablets, Zydus Pharmaceuticals, Pennington, NJ, USA) and a scrotal block was performed by infiltrating 10 mL of 2% lidocaine (Lidocaine 2% solution, VetOne, Boise, ID, USA) subcutaneously into the neck of the scrotum. A growth-promoting implant containing 36 mg of zeranol (Ralgro, Merck Animal Health, Rahway, NJ, USA) was also administered according to label directions. Any calves with horns (n = 3) had their horns removed with a Barnes dehorner. Prior to dehorning, each calf was given meloxicam (1 mg/kg of BW orally, Meloxicam 15 mg tablets, Zydus Pharmaceuticals) and each horn was anesthetized by performing a cornual nerve block with 5 mL of 2% lidocaine (Lidocaine 2% solution, VetOne). All calves were ear notched to test for persistent infection with bovine viral diarrhea virus (BVDV) via antigen-capture ELISA at the Athens Veterinary Diagnostic Laboratory in Athens, GA. Two weeks after initial processing, all cattle were vaccinated with a pentavalent inactivated viral respiratory vaccine to provide protection against bovine herpesvirus 1, parainfluenza virus, bovine respiratory syncytial virus, and BVDV types 1 and 2 (Triangle 5, Boehringer Ingelheim Animal Health). Body weights were obtained on each calf every 2 wk during the 42-d study period.

Treatment allocation

At the time of initial processing, calves were randomly assigned to receive either tulathromycin (Draxxin, Zoetis, Parsippany, NJ) at 1.1 mL/45.4 kg (2.5 mg/kg of BW) or saline at 1.1 mL/45.4 kg of BW using the random number generator function in a commercial spreadsheet program (Excel, Microsoft Corporation, Redmond, WA). Each injection was given subcutaneously in the neck according to label directions and beef quality assurance guidelines. All study personnel were blinded to treatment assignment.

For each arrival date, 6 pens were used with 3 pens available for each arrival treatment group. Calves were housed in pens specific to treatment and different sexes were mixed within a single pen. Including each block of cattle, a total of 36 pens with up to 6 calves per pen were used for a total of 18 replicates per experimental treatment. Each pen was 0.25 hectares (ha) in size and standing forage in the pens consisted of predominantly endophyte-infected tall fescue (Festuca arundinacea) with a smaller amount of Bermuda grass (Cynodon dactylon) intermixed. Pens were separated by a 4.9 m lane such that calves receiving separate arrival treatments were not housed with fence-line contact.

Feeding procedures

Once in their assigned pens, all calves were fed a common commodity blend ration made up of 50% pelleted corn gluten feed, 35% pelleted soy hulls, and 15% ground corn (19.2% CP, 73.8% TDN, 0.513 MC/lb NEg, 20.1% crude fiber) that is representative of rations fed by southeastern beef stocker producers. The ration was formulated to provide 200 mg of lasalocid (Bovatec, Zoetis) per calf per day. Calves were fed at a rate of 3.2 kg per calf per day. The amount of feed presented to each calf daily remained constant for the duration of the study and calves were fed once daily by hand at 0730 hours from 3 m galvanized bunks at the front of each pen. Refusals of feed were not measured and if feed remained in the bunk the following day, the feed was removed and new feed was added. Calves also had ad libitum access to Bermuda grass hay (Cynodon dactylon, 12% CP, 55.7% TDN, 0.275 MC/lb NEg, 64.5% NDF, 35.7% ADF) and water from automatic waterers. The size of the bunks and waterers ensured that each calf had at least 0.61 m of linear feed bunk space and 5.1 cm of linear water through space.

Assessment for clinical signs of BRD

Calves were visually monitored once daily at the time of feeding by a single-blinded evaluator for clinical signs consistent with BRD. Blinding was maintained until the conclusion of the trial. To maintain blinding and ensure that cattle receiving saline were treated in a timely manner if clinical signs of BRD were detected, no postmetaphylactic interval (PMI) was observed for cattle assigned to either treatment group in this study.

Daily evaluation employed criteria based on the DART system with modifications as described by Step et al. (Wilson et al., 2015). Briefly, the subjective criteria used for the removal of calves from the pen included depression (depressed attitude, separation from penmates, lowered head, glazed or sunken eyes, slow movement, arched back, or difficult standing/walking), abnormal appetite (completely off feed, eating less than expected or slowly, lack of gut fill), and respiratory signs (labored breathing, extended head and neck, audible noise when breathing). A severity score of 0 to 4 was assigned to each calf based on clinical signs and the severity of observed signs. A score of 0 was observed for a clinically normal calf. A score of 1 was assigned for mild clinical signs, 2 for moderate clinical signs, 3 for severe clinical signs, and 4 for a moribund animal (Perino and Apley, 1998; Ives and Richeson, 2015).

Objective criteria used to determine if antimicrobial therapy was necessary, was rectal temperature (Perino and Apley, 1998). Any calf removed from the pen due to clinical signs consistent with BRD and having a severity score of 1 to 4 was taken to the processing chute for rectal temperature measurement. Any animal with a severity score of 1 or 2 and a rectal temperature of ≥40 °C received an antimicrobial (Perino and Apley, 1998). Any calf with a severity score of 1 or 2 and a rectal temperature of <40 °C, was not treated and returned to its home pen following evaluation (Perino and Apley, 1998). Any animal with severe clinical signs (severity score = 3 or 4) received an antimicrobial regardless of rectal temperature (Perino and Apley, 1998).

The first antimicrobial treatment for all calves meeting treatment criteria was danofloxacin (Advocin, Zoetis Animal Health, Parsippany, NJ). Danofloxacin was administered at a dose of 8 mg/kg subcutaneously in the neck once. Following treatment, animals were returned to their home pen. A 7-d posttreatment interval (PTI) was observed after danofloxacin administration before a second antimicrobial could be administered (Apley, 2015). If antimicrobial treatment criteria were met a second time, ceftiofur crystalline free acid (CCFA, Zoetis Animal Health) was administered subcutaneously in the fat pad at the base of the ear at a dose of 6.6 mg/kg. After CCFA administration, a 7-d PTI was observed before a third antimicrobial treatment could be administered (Apley, 2015). If antimicrobial treatment criteria were met a third time, CCFA was administered as previously described. Gross postmortem examinations were performed on all mortalities by the same study veterinarian to determine the cause of death.

Sample collection and processing

Deep nasopharyngeal (DNP) swabs were obtained from each calf at the time of arrival processing and then again 14 d later. Additionally, calves diagnosed with BRD had a DNP obtained at the time of BRD diagnosis. Deep nasopharyngeal swabs were collected as described previously with some modifications (Snyder et al., 2017; Crosby et al., 2018). Briefly, calves were restrained with a rope halter and the head was secured to the side of the chute. A double-guarded swab (Double Guarded Uterine Culture Swab, Jorgensen Laboratories, Inc, Loveland, CO) was inserted ventromedially into one nostril. Upon entrance into the nasopharynx, the swab was advanced through both sheaths. The cotton tip was swirled for 20 s, at which point it was retracted back into the sheaths prior to removal. The external sheaths were discarded, the swabs placed into liquid Amies transport media without charcoal, labeled with the calf’s unique identification number and placed in a cooler on ice. Samples were transported on ice to the University of Georgia Tifton Diagnostic and Investigational Laboratory for aerobic culture and antimicrobial susceptibility testing. Any calves diagnosed with BRD prior to collection of the 2nd DNP on day 14 were sampled prior to treatment to ensure each calf had received only one antimicrobial exposure before susceptibility testing was performed.

Bacterial culture and antimicrobial susceptibility testing

Bacterial culture for the isolation and identification of bacteria from the DNP samples was performed as follows: The DNP swab samples were removed from the transport media and directly inoculated onto 5% sheep blood agar and MacConkey agar according to standard laboratory procedures (Quinn and Markey, 1994). The MacConkey agar inoculated plate and one blood agar inoculated plate was incubated aerobically at 35 ± 2 °C, and another blood agar inoculated plate was also incubated at 35 ± 2 °C in a CO₂ incubator. After an 18 to 24-h incubation, culture plates were removed from the incubator and the presence or absence of growth on each media, colony morphology, and their hemolytic patterns were recorded. If needed, colonies were re-isolated and subjected to further testing for identification as per the standard bacterial isolation and identification procedures (Quinn and Markey, 1994; Isenburg, 1998; Scientific, 2017). Plates were incubated for an additional 18 to 24 h, if necessary, for identification. Plates showing no growth were incubated for up to 5 d before being discarded as “no growth”. Bacterial identification of isolated colonies was made using one or multiple identification methods: 1) conventional biochemical methods (oxidase, indole, and trehalose); 2) the Gram-negative organism bacterial auto-identification system (Sensititre Gram Negative GNID; Thermo Fisher Scientific, Waltham, MA) according to the manufacturer’s procedure; 3) API NE (Biomerieux, Durham, NC) and RapID NF Plus (Thermo Fisher Scientific) commercial test panels according to the manufacturer’s instructions; 4) MALDI-TOF (Bruker Scientific Instruments, Billerica, MA); and 5) 16S sequencing (Eurofins Genomics, Louisville, KY) as needed.

Minimum inhibitory concentrations (MIC) were obtained for bacterial isolates using the broth microdilution method with commercial plates (BOPO6F, TREK Diagnostic Systems, Cleveland, OH) as per the manufacturer’s instructions. The MIC of each isolate was read using the BIOMIC V3 reader and defined as the minimum concentration of antimicrobial where no growth was recorded (Scientific, 2017). For each antimicrobial agent, MIC₅₀, MIC₉₀ (minimum inhibitory concentration of 50% and 90% of isolates), and the range of MICs obtained were calculated, and the results of testing were converted to qualitative categories (susceptible, intermediate, and resistant) per breakpoints established by the Clinical and Laboratory Standards Institute (CLSI, 2023).

Contextual antimicrobial use

Contextual AMU was calculated for each cohort and treatment group using both a dose (regimens per head in) and mass-based metric (milligrams of active substance per kilogram liveweight sold) as described previously (Brault et al., 2019a, 2019b). Briefly, a treatment regimen was defined as a single antimicrobial administration or a series of consecutive antimicrobial administrations associated with one antimicrobial product, one animal, and use indication. To meet this criterion, the time gap between administrations belonging to the same regimen had to never be greater than 2 d (Apley et al.). Milligrams (mg) of active substance was calculated by multiplying the weight of the animal at the time of treatment (in kg) by the label dose of active substance in mg. Average daily gain (ADG) was calculated for each arrival block by subtracting the total animal weight on day 0 from the total animal weight on day 42. This value was then divided by the total number of head days fed to calculate the ADG for each arrival block. The number of head days fed for each calf was calculated by adding the number of days a given calf was present in the study. Each calf alive at the end of the study was assigned a value of 42 head days fed, while calves that died were assigned the number of calendar days between their arrival date and date of death.

Statistical analysis

Sex and arrival culture results were compared between treatments using Chi square tests of homogeneity. Arrival body weights were compared using a two-sample t test. Changes in body weight and ADGs were compared between treatments using linear mixed models, and morbidity and mortality percentages were compared using mixed logistic regression models. All mixed models included nested random effects for block and pen to account for the clustering of responses for calves that were housed together. The model for changes in body weight, which included repeated observations from the same calves, also included a random calf effect. The time from arrival until the first BRD treatment was graphically evaluated using Kaplan–Meier curves, and statistically evaluated using a Cox proportional hazards model with a shared frailty for calves from the same pen. The treatment variable did not meet the proportional hazards assumption, so an extended Cox model with an interaction between treatment and a dichotomous function of time was used to evaluate the effect of treatment from days 0 to 14 and from days 15 to 42. For contextual AMU, a two-sample t test was used to evaluate the significance of differences in means of AMU metrics between treatment groups, while the relationship between treatment groups across metrics was further explored by plotting arrival cohort ordered by rank. Antimicrobial use was reported in scatter plots with the metric of interest (regimens per head in or milligrams per kilogram liveweight sold) on the x-axis and ADG on the y-axis. Mean (solid line) and median (dashed line) values for each axis were calculated across all cohorts and graphically represented as previously described (Schrag et al., 2022). Total values for each treatment group were calculated and presented as bar graphs, with corresponding cohort values represented by box plots. All testing assumed a two-sided alternative hypothesis, and values of P < 0.05 were considered statistically significant. Analyses were performed using commercially available statistical software (Stata version 17.0, StataCorp LLC, College Station, TX and R version 4.2.3, Richmond Hills, ON, Canada).

Results

Description of the study population

A total of 155 calves were randomly allocated to the two treatments, with 77 assigned to the saline group and 78 assigned to the tulathromycin group. Calves were received in 6 blocks between February 2021 and March 2022, and block sizes ranged from 23 to 30 calves. Within each block, calves were allocated to one of 6 group pens, each of which contained 3 to 6 animals. Within each pen, all calves belonged to the same treatment group. At the time of enrollment, calves assigned to the two treatments were similar with respect to sex (P = 0.29), arrival culture results (P = 0.97), and body weight (P = 0.94; Table 1). In block 1, there were 8 bulls, 1 steer, and 18 heifers. In block 2 there were 7 bulls, 5 steers, and 11 heifers. In block 3, there were 21 bulls, 3 steers, and 4 heifers. In block 4, there were 27 bulls, 3 steers, and 0 heifers. In block 5, there were 7 bulls, 0 steers, and 16 heifers. In block 6, there were 18 bulls, 0 steers, and 6 heifers. No calves were removed from the study due to the presence of visible signs of illness or noticeable deformities/abnormalities detected at the time of delivery. Additionally, all calves were negative for BVDV on ear notch screening.

Table 1.

Characteristics of stocker calves at the time of arrival. Values are reported as the frequency (%) or mean ± SD

	Saline (n = 77)	Tulathromycin (n = 78)	P †
Sex
Bull	48 (62)	40 (51)
Steer	4 (5)	8 (10)	0.29
Heifer	25 (32)	30 (38)
Arrival culture
Positive	13 (17)	13 (17)	0.97
Negative	64 (83)	65 (83)
Body weight (kg)	235.4 ± 23.5	235.1 ± 22.9	0.94

^† P value for a Chi square test (sex, arrival culture), or two-sample t test (body weight).

Performance

In a longitudinal mixed-model analysis of body weights, there was a significant interaction between the effects of treatment and day postarrival (P < 0.001). There was no difference between treatments on day 0 (P = 0.92), but there was a difference between treatments on day 14, with calves that received tulathromycin having a higher mean ± SE body weight than calves receiving saline (245.3 ± 3.9 kg vs. 235.3 ± 3.9 kg; P = 0.035; Figure 1). There was no difference between treatment groups with respect to body weight on day 28 (P = 0.11), or day 42 (P = 0.18).

Figure 1.

Mean body weight in kilograms (95% CI) for stocker calves that received either saline (untreated, n = 77) or tulathromycin (treated, n = 78) at the time of arrival. *P < 0.05.

Compared to the saline group, the tulathromycin group had higher mean ± SE ADGs for the periods from days 0 to 14 (0.74 ± 0.17 kg vs. 0.01 ± 0.17 kg; P = 0.001), days 0 to 28 (0.92 ± 0.12 kg vs. 0.64 ± 0.12 kg; P = 0.007), and days 0 to 42 (0.96 ± 0.07 kg vs. 0.82 ± 0.07 kg; P = 0.034).

Morbidity and mortality

The proportion of calves that required one or more BRD treatments during the 42-d follow-up period was lower for calves receiving tulathromycin compared to those receiving saline (13/78 [17%] vs. 31/77 [40%]; P = 0.008). Kaplan–Meier curves for the time to the first BRD treatment are shown in Figure 2. In a Cox proportional hazards survival analysis, the effect of treatment was dependent on the number of days postarrival. The tulathromycin group had a lower hazard rate for BRD treatment compared to the saline group from days 0 to 14 after arrival (Hazard Ratio [95% CI]: 0.22 [0.08, 0.56]; P = 0.001), but there was no difference between groups with respect to the hazard rate for BRD treatment from days 15 to 42 (Hazard Ratio [95% CI]: 0.94 [0.24, 3.6]; P = 0.92). First treatment success was higher for the calves that received tulathromycin compared to those that received saline (12/13 [92.3%] vs. 15/31 [48.4%]; P = 0.004). The proportion of calves that died (n = 6) or were euthanized (n = 3) during the follow-up period was similar for the calves that received tulathromycin and those that received saline (5/78 [6%] vs. 4/77 [5%], respectively; P = 0.77). Of the calves that died or were euthanized, 4/5 (80%) of those that received saline and 3/4 (75%) of those that received tulathromycin succumbed to the effects of BRD. For the two calves that died or were euthanized for reasons other than BRD, one received saline, and one received tulathromycin. The calf that received saline was euthanized for severe lameness and ultimately diagnosed as a ruptured cranial cruciate ligament, while the calf that received tulathromycin was euthanized for severe myositis.

Figure 2.

Kaplan–Meier curves showing the time to first treatment for BRD in calves that received either saline (untreated, n = 77) or tulathromycin (treated, n = 78). Each step in the curve represents the proportion of cattle remaining untreated on that day.

Culture results

At the time of arrival, the proportions of calves that had one or more bacterial BRD pathogens identified on nasal swab culture was similar between the treatment groups (tulathromycin 13/78 [17%] vs. saline 13/77 [17%]; P = 0.94). At the time of the follow-up sampling, however, the proportion of calves that had one or more BRD pathogens identified on nasal swab culture was lower for calves that received tulathromycin compared to those that received saline (12/78 [15%] versus 46/77 [60%], respectively; P < 0.001). The prevalence of specific bacterial BRD pathogens identified by culture of nasal swabs is summarized by treatment group and sampling occasion in Table 2.

Table 2.

Prevalence (%) of specific bacterial BRD pathogens identified in stocker calves that were treated with saline (n = 77) or tulathromycin (n = 78) at the time of arrival processing and follow-up

		Sampling Occasion
Bacteria	Treatment	Arrival	Follow-up
Pasteurella multocida	Saline	12/77 (16)	34/77 (44)
	Tulathromycin	10/78 (13)	6/78 (8)
Mannheimia haemolytica	Saline	1/77 (1)	15/77 (19)
	Tulathromycin	1/78 (1)	1/78 (1)
Histophilus somni	Saline	0/77 (0)	2/77 (3)
	Tulathromycin	0/78 (0)	4/78 (5)
Trueperella pyogenes	Saline	0/77 (0)	0/77 (0)
	Tulathromycin	0/78 (0)	1/78 (1)
Mannheimia pernigra	Saline	1/77 (1)	0/77 (0)
	Tulathromycin	2/78 (3)	0/78 (0)

Antimicrobial susceptibility

The MIC₅₀, MIC₉₀, range of MICs, and proportion of Pastuerella multocida and Mannheimia haemolytica isolates collected at each sampling occasion classified as susceptible to specific antimicrobials are presented in Tables 3 and 4. The number of P. multocida and M. haemolytica isolates with different MIC values are summarized by treatment and sampling occasion in Figure 3, and the number of isolates with different antimicrobial resistance patterns at the different sampling time points (arrival and postmetaphylaxis) is summarized in Table 5. Overall, 52/68 (81%) P. multocida isolates, and 8/17 (47%) M. haemolytica isolates, were classified as pansusceptible.

Table 3.

MIC_50, MIC₉₀, range and percent of P. multocida and M. haemolytica isolates collected from high-risk beef stocker calves prior to and 14 d after administration of saline susceptible to 12 antimicrobials at the time of arrival processing

		Pasteurella multocida				Mannheimia haemolytica
Antimicrobial	Sampling occasion	MIC50 (μg/mL)	MIC90 (μg/mL)	Range (μg/mL)	% Susceptible	MIC50 (μg/mL)	MIC90 (μg/mL)	Range (μg/mL)	% Susceptible
Ampicillin	1	≤0.25	≤0.25	≤0.25	N/A	N/A	N/A	N/A	N/A
	2	≤0.25	≤0.25	≤0.25	N/A	≤0.25	0.5	≤0.25 to 0.5	N/A
Ceftiofur	1	≤0.25	≤0.25	≤0.25	100	N/A	N/A	N/A	N/A
	2	≤0.25	≤0.25	≤0.25	100	≤0.25	2	≤0.25 to 2	100
Chlortetracycline	1	≤0.5	2	≤0.5 to 2	N/A	N/A	N/A	N/A	N/A
	2	≤0.5	2	≤0.5 to 8	N/A	>8	>8	0.5 to >8	N/A
Danofloxacin	1	≤0.125	≤0.125	≤0.125	100	N/A	N/A	N/A	N/A
	2	≤0.125	≤0.125	≤0.125	100	≤0.125	>1	≤0.125 to >1	86
Enrofloxacin	1	≤0.125	≤0.125	≤0.125	100	N/A	N/A	N/A	N/A
	2	≤0.125	≤0.125	≤0.125	100	≤0.125	>2	≤0.125 to >2	86
Florfenicol	1	0.5	1	0.25 to 1	100	N/A	N/A	N/A	N/A
	2	0.5	1	0.25 to 1	100	1	>8	≤0.25 to >8	86
Oxytetracycline	1	≤0.5	4	≤0.5 to 8	93	N/A	N/A	N/A	N/A
	2	≤0.5	8	≤0.5 to >8	80	>8	>8	0.5 to >8	46
Penicillin	1	≤ 0.125	0.25	≤ 0.125 to 0.25	100	N/A	N/A	N/A	N/A
	2	≤ 0.125	0.25	≤ 0.125 to 0.25	100	0.5	1	≤0.125 to 1	50
Spectinomycin	1	16	32	8 to >64	93	N/A	N/A	N/A	N/A
	2	16	>64	8 to >64	83	32	>64	≤8 to >64	86
Tilmicosin	1	≤4	32	≤4 to >64	93	N/A	N/A	N/A	N/A
	2	≤4	64	≤4 to >64	86	8	>64	8 to >64	79
Tulathromycin	1	≤1	32	1 to >64	93	N/A	N/A	N/A	N/A
	2	≤1	>64	1 to >64	86	8	>64	≤1 to >64	86
Tylosin	1	16	>32	8 to >32	N/A	N/A	N/A	N/A	N/A
	2	16	>32	16 to >32	N/A	>32	>32	8 to >32	N/A

N/A—breakpoints not available or too few isolates to summarize.

Table 4.

MIC_50, MIC₉₀, range and percent of P. multocida isolates collected from high-risk beef stocker calves prior to and 14 d after administration of tulathromycin susceptible to 12 antimicrobials at the time of arrival processing

		Pasteurella multocida
Antimicrobial	Sampling occasion	MIC50 (μg/mL)	MIC90 (μg/mL)	Range (μg/mL)	% Susceptible
Ampicillin	1	≤0.25	≤0.25	≤0.25	N/A
	2	≤0.25	≤0.25	≤0.25	N/A
Ceftiofur	1	≤0.25	≤0.25	≤0.25	100
	2	≤0.25	≤0.25	≤0.25	100
Chlortetracycline	1	≤0.5	>8	≤0.5 to >8	N/A
	2	2	>8	≤0.5 to 8	N/A
Danofloxacin	1	≤0.125	≤0.125	≤0.125	100
	2	≤0.125	≤0.125	≤0.125	100
Enrofloxacin	1	≤0.125	≤0.25	≤0.125 to 0.25	100
	2	≤0.125	≤0.125	≤0.125	100
Florfenicol	1	0.5	1	≤0.25 to 1	100
	2	1	1	≤0.25 to 1	100
Oxytetracycline	1	≤0.5	>8	≤0.5 to >8	90
	2	1	>8	≤0.5 to >8	60
Penicillin	1	≤ 0.125	0.25	≤ 0.125 to 0.25	100
	2	0.25	0.5	≤ 0.125 to 0.5	80
Spectinomycin	1	16	>64	8 to >64	90
	2	16	>64	8 to >64	60
Tilmicosin	1	≤4	>64	≤4 to >64	90
	2	8	>64	≤4 to >64	60
Tulathromycin	1	≤1	>64	≤1 to >64	90
	2	4	>64	≤1 to >64	60
Tylosin	1	16	>32	8 to >32	N/A
	2	16	>32	16 to >32	N/A

N/A—breakpoints not available or too few isolates to summarize.

Figure 3.

Number of P. multocida and M. haemolytica isolates with different MIC values by treatment group and sampling occasion. Sa1 (saline, arrival), Tu1 (tulathromycin, arrival), Sa2 (saline, follow-up), Tu2 (tulathromycin, follow-up). Only one isolate was included from each calf unless duplicate isolates from the same calf had different MIC values. Horizontal lines indicate susceptibility breakpoints established by the Clinical Laboratory and Standards Institute (CLSI).

Table 5.

Number (%) of P. multocida and M. haemolytica isolates with different antimicrobial resistance patterns by treatment (saline or tulathromycin) and sampling occasion (arrival or follow-up)

Resistance pattern	Saline		Tulathromycin		Total
	Arrival	Follow-up	Arrival	Follow-up
Pasteurella multocida
None	11 (90.9)	27 (74.1)	9 (90)	2 (33.3)	52 (81.3)
OTC SPT TIL TUL		5 (18.5)	1 (10)	2 (33.3)	8 (12.5)
OTC		1 (3.7)			1 (1.6)
OTC SPT		1 (3.7)			1 (1.6)
PEN				1 (16.6)	1 (1.6)
SPT TIL TUL	1 (9.1)				1 (1.6)
Mannheimia haemolytica
None	1 (100)	5 (33.3)	1 (100)	1 (100)	8 (47.1)
OTC PEN		4 (26.7)			4 (23.5)
DAN ENR FLO OTC PEN SPT TIL TUL		1 (6.7)			1 (5.9)
FLO OTC PEN SPT TIL TUL		1 (6.7)			1 (5.9)
OTC PEN SPT		1 (6.7)			1 (5.9)
OTC PEN TIL		1 (6.7)			1 (5.9)
PEN		1 (6.7)			1 (5.9)

DAN, danofloxacin; ENR, enrofloxacin; FLO, florfenicol; OTC, oxytetracycline; PEN, penicillin; SPT, spectinomycin; TIO, ceftiofur; TIL, tilmicosin; TUL, tulathromycin

The MIC₅₀, MIC₉₀, range, and aggregate antimicrobial susceptibility profile of isolates collected from saline-treated cattle at the time of the first BRD treatment are summarized in Supplementary Table S1 and Supplementary Figure S1. The analysis provided here is descriptive in nature due to the small sample size. Pastuerella multocida was isolated from 32.7% (n = 14) of calves, M. haemolytica from 34.6% (16) of calves, Trueperella pyogenes from 2.7% (n = 1) of calves, and Histophilus somni from 11.5% (n = 6) of calves, and 19.2% (n = 10) yielded no significant growth. More than 80% of all M. haemolytica (n = 15) and P. multocida (n = 13) isolates were susceptible to ceftiofur, danofloxacin, enrofloxacin, florfenicol, spectinomycin, tilmicosin, and tulathromycin. In contrast, only 40% and 53% of M. haemolytica isolates were susceptible to oxytetraycline and penicillin, respectively.

In the tulathromycin-treated cattle, too few susceptibilities were available to provide summary data of the first BRD treatment aggregate profiles. In this group, P. multocida was isolated from 20% of treated calves (n = 3), M. haemolytica from 13.3% (n = 2) of treated calves, and H. somni from 26.7% (n = 4) of treated calves. More than 40% (n = 6) of calves treated for BRD in this group had no significant growth reported. Nevertheless, of the P. multocida (n = 3) and M. haemolytica (n = 1) isolates that yielded complete susceptibility data, 100% were susceptible to ceftiofur, danofloxacin, enrofloxacin, florfenicol, and penicillin.

The MIC₅₀, MIC₉₀, range, and aggregate antimicrobial susceptibility profile of isolates collected from saline-treated cattle at the time of the second BRD treatment are summarized in Supplementary Table S2 and Supplementary Figure S2. As with the first BRD treatment data, the analysis provided here is descriptive in nature due to the small sample size. Overall, P. multocida was isolated from 60% (n = 9) of calves, M. haemolytica from 20% (3) of calves, and H. somni from 13.3% (n = 2) of calves. Only 6.7% (n = 1) yielded no significant growth. Pastuerella multocida isolates were routinely susceptible to almost all antimicrobials tested. In contrast, M. haemolytica isolates demonstrated inconsistent susceptibility to all antimicrobials except ceftiofur.

Contextual antimicrobial use

The relationship between AMU and ADG varied across cohorts and AMU metrics and is illustrated in Figure 4a and b. When measured in milligrams per kilogram of liveweight sold, AMU of the different treatment groups within the same cohort fell in the same AMU quadrant (i.e., high AMU, high ADG, Figure 4a). Calves receiving tulathromycin had higher total ADG from days 0 to 42 than calves receiving saline in 4 of 6 cohorts but also demonstrated higher AMU in 3 of 6 cohorts. When evaluating the relationship between regimens per head in and ADG, calves receiving tulathromycin were classified as high AMU, high ADG in 3 of 6 cohorts and high AMU, low ADG in 3 of 6 cohorts. In contrast, calves receiving saline were classified as low AMU, high ADG in 4 of 6 cohorts and high AMU, low ADG in only 2 of 6 cohorts (Figure 4b).

Figure 4.

Performance of each enrolled cohort (as assessed by determination of ADG) as a function of total antimicrobial use (AMU) expressed as milligrams per kilogram of liveweight sold (a) or antimicrobial regimens per head in (b) by treatment group (saline or tulathromycin). ADG is plotted on the y-axis, while AMU metric is plotted on the x-axis. The upper left quadrant represents low AMU, high ADG, while the upper right quadrant represents high AMU, high ADG. The lower left quadrant represents low AMU, low ADG, while the lower right quadrant represents high AMU, low ADG. Solid black lines represent the intersection of the mean values of the specific AMU metric and ADG for the entire dataset, while the dashed lines represent the intersection of the median values of the specific AMU metric and ADG for the entire dataset.

Total AMU and ADG by treatment group, as well as the range across cohort for each measure of use, are reported in Figure 5. When measured as regimens per head in, AMU was higher in calves receiving tulathromycin than in calves receiving saline (P = 0.01).

Figure 5.

Total antimicrobial use (AMU) evaluated on milligrams per kilogram of liveweight sold or regimens per head in by treatment group (saline or tulathromycin) in high-risk beef stocker calves. Treatment group is on the x-axis and AMU on the y-axis. Solid vertical bars represent the mean total AMU, while the box-and-whisker plots superimposed on the solid bars represent the median, 25th and 75th percentiles, and range of AMU values within a group.

The relationship between treatment groups across AMU metrics is presented in Figure 6. No definable relationship between treatment groups was observed for ADG or milligrams per kilogram liveweight sold. However, for regimens per head in, AMU in cattle receiving metaphylaxis is consistently higher than the control group when disease pressure is low (cohorts with positive ADG in both treatment groups), and nearly identical to the control group when disease pressure is high (cohorts with very low ADG in at least one treatment group).

Figure 6.

Relationship between ADG and antimicrobial use (milligrams per kilogram of liveweight sold or regimens per head in) by cohort. Cohort is plotted on the x-axis, while ADG and AMU are plotted on the y-axis. Solid bars represent average ADG or total AMU by treatment group (saline or tulathromycin) while the solid lines represent trends in measured parameter across cohorts.

Discussion

Stocker calves experience a period of adaptation as they transition from commercial cow–calf operations to feed yards (Peel, 2003). Marketing and shipment of this class of cattle leads to exposure to multiple stressors that increase the risk of BRD, and it is not unusual to see the risk of morbidity due to BRD exceed 60% in some populations (Richeson et al., 2008, 2009; Smith, 2009). Stocker calves diagnosed with BRD have a reduction in ADG and a greater likelihood of death than unaffected calves, resulting in a decrease in per animal return of up to 20% (Pinchak et al., 2004; Tennant et al., 2014). To reduce the negative impact that BRD has on stocker calves, metaphylaxis is commonly used (Nickell and White, 2010; Ives and Richeson, 2015). The administration of an effective antimicrobial for metaphylaxis is associated with a reduction in morbidity and mortality, as well as an increase in performance relative to untreated controls (Tennant et al., 2014; Coppin et al., 2022). Economic analyses have shown that the use of metaphylaxis in high-risk cattle entering feedlots is associated with a direct net return of $532 million (Dennis et al., 2020).

In this study, calves given tulathromycin gained 0.14 kg/d more than saline controls over the 42-d observation period. This effect was most pronounced over the first 14 d following arrival with calves receiving tulathromycin weighing, on average, 10 kg more than saline controls. In total, calves receiving tulathromycin outperformed the saline controls over the entirety of the observation period and this translated to nearly 6 kg of additional saleable weight for calves given metaphylaxis. At current market prices ($234/45.4 kg for a 295 kg steer), the value of added weight alone represents a nearly $30 increase in gross return per calf marketed.

In this study, the hazard of BRD in calves receiving metaphylactic tulathromycin was 78% lower than that of saline controls, which corresponds with the results of previous studies (Tennant et al., 2014; Coppin et al., 2022; Pollreisz et al., 2022). Interestingly, this effect was most profound over the first 14 d after arrival, as our study found that the hazard of BRD between treated and untreated cattle was not statistically different compared to days 15 and 42. Although the exact mechanisms underlying the beneficial effects of metaphylactic antimicrobial use are not yet completely understood, metaphylaxis is thought to modify the epidemiologic parameters and pathogen transmission dynamics associated with BRD outbreaks; in other words, metaphlyaxis results in both delayed disease onset and reduced level of morbidity (Ley, 2019). It is possible that metaphylactic antimicrobial use enables the development of adaptive immune responses by reducing the bacterial load of the respiratory tract during times of encountered stressors, resulting in an enhanced immune response during exposure to viral and bacterial BRD pathogens (Ley, 2019). This effect was observed in this study as treated cattle not only had a reduction in morbidity due to BRD, but also a delay in the onset of cases as seen in the Kaplan–Meier survival curve in Figure 2. In addition, our study showed that calves given tulathromycin at arrival processing were significantly less likely to be shedding recognized bacterial pathogens than calves given saline at resampling on day 14 (15% vs. 60%, respectively). Moreover, 40% of calves given treated with tulathromycin at arrival and subsequently treated for BRD had no significant bacterial growth reported on DNP culture. Calves given tulathromycin at arrival also had a higher first treatment success than calves given saline. Although morbidity was reduced significantly, the risk of mortality due to all causes was no different between treatment groups, and 6% of cattle in each group died during the 42-d observation period used in this study. This is not surprising as other recent studies have shown similar levels of mortality in treated and untreated cattle (Horton et al., 2023).

Multiple recent studies have found that the proportion of bacterial isolates classified as multi-, or extensively-drug resistant (MDR, XDR), increases significantly between the time of arrival processing and resampling 2 to 4 wk later (Snyder et al., 2017; Crosby et al., 2018; Woolums et al., 2018; Sarchet et al., 2022). This is important because cattle with MDR/XDR bacteria in their airway might fail to respond to treatment should they be diagnosed with BRD. These studies also found that the prevalence of shedding M. haemolytica increased between the time of arrival and resampling (Snyder et al., 2017; Crosby et al., 2018; Woolums et al., 2018; Sarchet et al., 2022). Unfortunately, this earlier work did not include untreated control groups so a direct association between the administration of a metaphylactic antimicrobial and the selection of MDR/XDR pathogens could not be made. Conversely, the results of this study contrast with previous work as the same levels and patterns of resistance were not seen. Overall, 81% of P. multocida isolates and 47% of M. haemolytica isolates were classified as pansusceptible. Additionally, lower levels of pathogen shedding at resampling were found, with P. multocida instead being the primary organism isolated from calves with and without BRD at each sampling time point (days 0 and 14). While reasons for the discrepancy in resistance prevalence and pathogen shedding between this study and others are not clear, cattle in this study were managed in a slightly different fashion than cattle evaluated in previous work. For example, calves in this study were transported directly to the research facility on the day of purchase and were sourced from a single auction market. Moreover, the auction market was only 8 km from the research facility. Thus, calves were not in transit for a prolonged period and the commingling of cattle from different origins was reduced. Additionally, in commercial settings, individual calves might spend time at order buyer facilities or be transported through multiple different marketing channels prior to reaching their destination. A truckload of calves might contain animals that have been procured from multiple auction markets rather than just one so that specific orders can be adequately filled. These factors inherent to the stocker calf production system increase opportunities for pathogen exposure to occur and, when compounded with prolonged stress and alterations in social structure, disruption of the normal bacterial communities in the upper and lower respiratory tract can take place (Stroebel et al., 2018; Credille, 2022). It has been shown that spending less than 24 h in an auction market setting does not affect the diversity or composition of the nasopharyngeal microbiota to the same extent that longer periods of time do (Stroebel et al., 2018). Thus, future work directed at determining the extent to which traditional marketing methods combined with metaphylaxis contribute to pathogen shedding and antimicrobial resistance is needed.

In recent years, the use of antimicrobials in the production of animal agriculture has come under intense scrutiny (Apley, 2018; Lhermie et al., 2020). Today, there is pressure to reduce the use of antimicrobials considered important for human medicine that is driven in large part by concerns that the overuse of antimicrobials in animal agriculture contributes to antimicrobial resistance challenges being faced in human medicine (Brault et al., 2019a, b). In this study, a macrolide (tulathromycin), fluoroquinolone (danofloxacin), and the third generation cephalosporin (ceftiofur) were used for metaphylaxis and therapeutic treatments, and all of these drugs are classified as critically important in human medicine. Antimicrobial use characterization is an important part of global antimicrobial stewardship efforts and is essential for assessing the risk of resistance associated with AMU in the production settings, as well as variability in AMU among different animal populations and different production operations (Brault et al., 2019a). Nevertheless, there is a lack of standardization of the various metrics that are available (Schrag et al., 2022).

In this study, contextual AMU was calculated for each cohort and treatment group using dose (regimens per head in) and mass-based metrics (milligrams of active substance per kilogram liveweight sold) because they provided two different numerators and denominators by which AMU can be assessed (Apley et al., 2022; Schrag et al., 2022). In addition, both metrics are highly correlated to each other, as well as other metrics commonly used to evaluate AMU in the context of BRD (Apley et al., 2022). In this study population, the use of tulathromycin was associated with improvements in performance in 4 of 6 cohorts while also demonstrating higher AMU in 3 of 6 cohorts when AMU has measured on a mg/kg of liveweight sold basis. This relationship, however, is confounded by differences in potency between drugs, and differences in the number of animals treated with each drug across cohorts when AMU is measured with this metric. When AMU was evaluated on regimens per head in basis, calves receiving tulathromycin were classified as high AMU, high ADG in 3 of 6 cohorts and high AMU, low ADG in 3 of 6 cohorts, while calves receiving saline were classified as low AMU, high ADG in 4 of 6 cohorts and high AMU, low ADG in only 2 of 6 cohorts. When evaluated as a function of disease pressure, AMU is consistently higher in calves receiving tulathromycin when it is evaluated on regimens per head in basis. This should be interpreted with caution as these cohorts were numerically small for deads in calculations of ADG and further study in larger groups is necessary to confirm this relationship. These results are consistent with previous work using different metrics in medium-risk feeder calf populations (Horton et al., 2023). It is important to note, however; that the different metrics evaluated in this study gave different results for total AMU and, consistent with other work, the choice of AMU indicator can have a notable effect on the results obtained (Apley et al., 2022, 2023). No PMI was observed in this study so as to maintain blinding. This could have led to animals receiving tulathromycin being treated earlier than they would normally be treated in commercial lots that adhere to a strict 7-d PMI and artificially increase AMU in treated animals. Nevertheless, examination of the survival curve in Figure 2 reveals that few, if any animals, receiving tulathromycin were treated prior to day 7. Thus, if any effect was present, it is likely to be small and insignificant.

The major weakness of the present study is the low number of animals shedding bacterial pathogens following metaphylaxis. This made determining the true effect of metaphylaxis on the selection of MDR/XDR bacteria in treated cattle difficult. This also precluded the making of meaningful conclusions as to how resistance might impact the outcome of antimicrobial therapy in cattle diagnosed with and treated for BRD. As stated previously, cattle in this study were managed differently than cattle in other studies and this could have impacted the results of our work. Thus, future studies evaluating stocker calf populations managed in ways more representative of commercial stocker operations are needed. Nevertheless, the results of this study do show that antimicrobial administration alone is not likely to be the only factor responsible for the high prevalence of MDR/XDR found in previous work. Commingling, stress, disruption of the upper and lower respiratory tract microbiota, and other factors, combined with antimicrobial administration, likely interact to contribute to previously reported outcomes (Stroebel et al., 2018; Credille, 2022).

In conclusion, the results of this study show that the use of metaphylaxis is associated with improved health, enhanced production, and an increased total AMU (when measured as regimens per head in) in high-risk beef stocker calves relative to the use of pull-and-treat strategy in untreated controls. Interestingly, the high prevalence of pathogen shedding and high rates of MDR/XDR previously reported in other studies was not seen in the cattle evaluated here (Snyder et al., 2017; Crosby et al., 2018; Woolums et al., 2018; Sarchet et al., 2022). In the end, using a pull-and-treat strategy might result in the use of fewer antimicrobial doses. Unfortunately, decreased calf performance, poorer health outcomes, and increased costs are likely to occur and must be considered in the overall equation (Tennant et al., 2014; Dennis et al., 2020; Coppin et al., 2022; Horton et al., 2023). The decision on whether to use a metaphylactic antimicrobial will need to be based on outcomes deemed to be most important to a given production system. Thus, veterinarians and producers tasked with managing the health and well-being of beef stocker calves should consider the entirety of the stocker production segment of the beef industry (marketing, procurement, nutrition, husbandry, etc.) so that the best decisions can be made for a given operation (Groves, 2020).

Glossary

Abbreviations

ACE

antigen-capture ELISA

ADF

acid detergent fiber

ADG

average daily gain

AMU

antimicrobial use

BRD

Bovine respiratory disease

BQA

beef quality assurance

BVDV

bovine viral diarrhea virus

CCFA

ceftiofur crystalline free acid

CF

crude fiber

CP

crude protein

DNP

deep nasopharyngeal swab

MC

megacalories

MIC

minimum inhibitory concentration

MDR

multidrug resistance

NDF

neutral detergent fiber

NEg

net energy of gain

PMI

post-metaphylactic interval

PTI

post-treatment interval

TDN

total digestible nutrients

XDR

extensive drug resistance

Conflict of interest statement

The authors have nothing to disclose.