Products of parturition are the predominant source of Brucella abortus for transmission in bison (Bison bison). Our objective was to assess whether preventing pregnancy in Brucella-seropositive bison reduced B. abortus shedding. Brucella-seropositive and -seronegative bison from Yellowstone National Park, Wyoming, USA were used in a replicated experiment. Each of two replicates (rep1, rep2) included a group of seropositive females treated with a single dose of gonadotropin-releasing hormone-based immunocontraceptive (Treatment rep1, n=15; Treatment rep2, n=20) and an untreated group (Control rep1, n=14; Control rep2, n=16) housed separately. Seronegative sentinel females were placed in each group to monitor horizontal transmission. Seronegative males were co-mingled for breeding each year. Pregnant females were removed from treatment groups in the first year, but not thereafter. Each January–June we monitored for B. abortus shedding events—any parturition associated with culture-positive fluids or tissues. We analyzed probability of shedding events using a negative binomial generalized linear mixed model fit by maximum likelihood using Laplace approximation. Over 5 yr, we observed zero shedding events in Treatment rep1 vs. 12 in Control rep1. All five Control rep1 sentinels but zero (0/5) Treatment rep1 sentinels seroconverted. In the second replicate, Treatment rep2 had two shedding events over 3 yr and Control rep2 had five events over 2 yr. Sentinels in both Control rep2 (3/6) and Treatment rep2 (5/6) seroconverted by trial endpoint. Treatment rep1 showed a reduced shedding probability relative to Control rep1, Treatment rep2, and Control rep2 (log odds value −25.36 vs. −1.71, −1.39, and −0.23, respectively). Fixed effect predictor covariates, year and age, had no explanatory value. These data suggest that successful contraception of brucellosis-seropositive female bison prevents shedding of B. abortus by individual animals. However, contraceptive treatment may or may not sufficiently reduce disease transmission to reduce brucellosis prevalence in an affected herd.

Bovine brucellosis, caused by Brucella abortus, primarily affects bovids, although species from other families may be affected as well, including humans. Brucella abortus infection causes reproductive failures such as abortion, stillbirths, and weak neonates (Rhyan et al. 2001; Olsen and Tatum 2010). The predominant mode of transmission is via infectious products of parturition, as aborted or nonviable fetuses, neonates, and associated tissues and fluids may be heavily colonized by bacteria. Contact with these materials by herd members facilitates a mucosal route of introduction of B. abortus into susceptible animals (Rhyan and Nol 2019). Other, less frequent, modes of transmission of B. abortus include ingestion of infected milk, or contact with infected genitalia or feces (Cheville et al. 1998).

Brucellosis, probably introduced by cattle (Bos taurus), was first detected in bison of the Greater Yellowstone Area (GYA) of the US in the early twentieth century, and had emerged in the GYA elk (Cervus canadensis) population by the 1930s (US Department of the Interior 2000; Rhyan and Nol 2019). Since the 1950s, a cooperative state–federal brucellosis eradication program has largely eliminated B. abortus from US domestic cattle and bison herds. Nevertheless, disease reservoirs in GYA free-ranging bison and elk populations still persist, and spillovers to domestic livestock occur (Ragan 2002; Olsen 2010). Annually, managers expend great efforts to maintain temporal and spatial separation between wild bison herds and domestic livestock, although data indicate that infected elk are the primary source of B. abortus cases in domestic cattle and bison in the GYA (Rhyan et al. 2013b; Kamath et al. 2016; National Academies of Sciences, Engineering, and Medicine 2020).

Test-and-remove is the most common management tool applied to brucellosis-affected livestock herds in the US (Ragan 2002). However, traditional lethal approaches to disease management of livestock are not necessarily compatible with what is considered feasible or socially acceptable when it comes to managing and conserving wildlife (Bienen and Tabor 2006). Research suggests that immunocontraception could be used as a nonlethal method to decrease B. abortus transmission in GYA bison herds (Miller et al. 2004; Rhyan et al. 2013a). We describe here an experiment designed to assess whether preventing pregnancy, using a gonadotropin-releasing hormone (GnRH) –based immunocontraceptive, in B. abortus–seropositive female bison would reduce probability of B. abortus shedding within a treated herd.

Animals

Animals were captured at the National Park Service’s Stephens Creek, Montana, US (45°02′56.8″N, 110°45′04.8″W) bison facility, as they migrated out of Yellowstone National Park (YNP) in late winter and early spring of 2011 and 2014. Eighty-three nonpregnant female bison, approximately 1−2 yr old, brucellosis seropositive and seronegative, were captured and used for the study. Between 2012 and 2014, ten 2−3-yr-old seronegative males were also captured. Additionally, four seronegative 2-yr-old males were purchased from a commercial source in 2012. Bison were collected in a series of traps and corrals, approximately 9.7 ha, using passive feeding or horses. Animals were sorted and manually restrained in a chute for examination and selection by age, sex, and pregnancy status. We determined age of animals by incisor eruption and wear (Dimmick and Pelton 1996; Fuller et al. 2007). We used rectal palpation to select nonpregnant females. Blood was collected from the jugular vein of eligible animals, using a 30-mL syringe with a 16-gauge needle, for chute-side brucellosis serological testing using a fluorescence polarization assay (FPA) and a standard card test (US Department of Agriculture, Animal and Plant Health Inspection Service [USDA-APHIS] 2003, 2006a). Blood was also transferred to serum separator tubes (Vacutainer, Becton, Dickinson and Company, Franklin Lakes, New Jersey, USA), allowed to clot at approximately 4 C, and centrifuged at 2,000 × G for 10 min. Serum was aliquoted into 2–3-mL cryovials and submitted to the Montana Veterinary Diagnostic Laboratory (MVDL), Bozeman, Montana, US, for additional brucellosis serology. Animals received unique identifications in the form of numbered ear tags (Y-Tex, Cody, Wyoming, USA) in both ears and a radio-frequency identification button tag (AllFlex USA, Dallas Fort Worth Airport, Texas, USA) inserted at the base of the left ear. We transported the bison 11 km in a stock trailer to the USDA-APHIS Veterinary Services (VS) Bison Facilities in Corwin Springs, Montana, US. The facilities comprised seven pastures totaling approximately 36 hectares, including holding corrals and bison working areas.

Experimental design

We established two study replicates, in 2012 and 2014−2015. The first replicate (Replicate 1) was made up of 29 seropositive females and 10 seronegative females. The second replicate (Replicate 2) was assembled over 2 yr and included animals captured in 2011 and 2014, and offspring of Replicate 1 (36 seropositive females, 12 seronegative females).

In April 2012, Replicate 1 was sorted into two 10-ha pastures. One pasture (Control replicate [rep]1) contained seropositive females (n=14) and seronegative sentinel females (SentinelControl rep1; n=5) to detect transmission through seroconversion. A second pasture (Treatment rep1) contained seropositive females (n=15) and seronegative sentinel females (SentinelTreatment rep1; n=5). Males were housed separately from females until breeding. We allowed comingling of males with females during July to October of each year, two males per pasture. For first breeding of Replicate 1, we used commercially sourced males; in subsequent years we used YNP males.

In April 2014, we assigned 20 seropositive females to Treatment rep2 (n=20), accompanied by six seronegative females (SentinelTreatment rep2). In 2015, we assembled Control rep2 (n=16), along with six seronegative females (SentinelControl rep2). Both groups were housed in 12-ha pastures. Males comingled with females (two per pasture) from July to October each year.

Calves remained with their dams until approximately 9 mo old, at which time they were transferred to a separate pen or to another facility to enter into a brucellosis quarantine protocol (Clarke et al. 2014), retained to supplement Replicate 2, or were shipped to slaughter.

Immunocontraceptive treatment

Gonadotropin-releasing hormone-based immunocontraceptive (GonaCon; USDA-APHIS, Wildlife Services, National Wildlife Research Center, Fort Collins, Colorado, USA) was prepared as described by Miller et al. (2008) and Frey et al. (2024). Seropositive females assigned to treatment groups were manually restrained and injected intramuscularly (IM) 2.5 cm ventral from the hip, bilaterally, with 3,000 μg GnRH in 2 mL adjuvant, divided into 1 mL (1,500 μg GnRH) in each hip, 3 mo before exposure to males. Treatment group animals were given one dose only on initiation of each replicate. Treatment rep1 received immunocontraceptive in May 2012 and Treatment rep2 in May 2014. None of the seropositive bison in control groups, nor any sentinels, received immunocontraceptive treatment. The first year of each replicate occurred on January 1 following immunocontraception treatment–control group assembly. Figure 1 summarizes the design and timeline of this two-replicate study.

Figure 1.

Timeline of an experimental evaluation of immunocontraception as a brucellosis management tool. The experiment included two replicates (rep1, rep2). Each replicate included a group of immunocontraceptive-treated seropositive females (Treatment) and a group of untreated seropositive females (Control) housed in separate paddocks. Treatment groups received a one-time intramuscular administration of 3,000 μg gonadotropin-releasing hormone-based immunocontraceptive in May of the year before onset of the replicate. Smaller numbers of seronegative female sentinels (denoted as Sentinelgroup) were housed with each of the four primary experimental groups (Treatment rep1, Treatment rep2, Control rep1, Control rep2) to monitor for transmission. Study years were defined as January−December. The first replicate trial ran for 5 yr (2013−2017); the second replicate trial had a staggered start and ended after 2 or 3 yr of observing Control rep2 (2016−2017) or Treatment rep2 (2015−2017), respectively.

Figure 1.

Timeline of an experimental evaluation of immunocontraception as a brucellosis management tool. The experiment included two replicates (rep1, rep2). Each replicate included a group of immunocontraceptive-treated seropositive females (Treatment) and a group of untreated seropositive females (Control) housed in separate paddocks. Treatment groups received a one-time intramuscular administration of 3,000 μg gonadotropin-releasing hormone-based immunocontraceptive in May of the year before onset of the replicate. Smaller numbers of seronegative female sentinels (denoted as Sentinelgroup) were housed with each of the four primary experimental groups (Treatment rep1, Treatment rep2, Control rep1, Control rep2) to monitor for transmission. Study years were defined as January−December. The first replicate trial ran for 5 yr (2013−2017); the second replicate trial had a staggered start and ended after 2 or 3 yr of observing Control rep2 (2016−2017) or Treatment rep2 (2015−2017), respectively.

Close modal

Monitoring and sample collection

In January each year, under manual restraint, we collected blood as described above and vaginal swabs (polyester-tipped, polyester stick applicator swabs; Fisher Scientific, Hampton, New Hampshire, USA), placed in World Health Organization media (supplied by MVDL) from females for brucellosis serology and Brucella culture. We checked pregnancy status through rectal palpation, and placed vaginal transmitters in pregnant animals (Advanced Telemetry Systems Inc., Isanti, Minnesota, USA) as described (Rhyan et al. 2009). In order to begin each replicate trial with maximum treatment efficacy, immunocontraceptive-treated animals identified as pregnant in the January handling before the first calving season (year 1) were removed from the study. In subsequent years, however, treated animals that became pregnant (i.e., those with “breakthrough pregnancies”) remained in the treatment group paddocks to allow for waning efficacy of the treatment over time.

During January–May each year, we monitored pregnant animals daily via observation and VHF radio telemetry to document transmitter expulsion events. If no live calf was present after the expulsion signal occurred, to assess pregnancy status we chemically immobilized the dam approximately 24 h after onset of the expulsion signal. We immobilized animals using etorphine (0.01 mg/kg) or thiafentanil (0.015–0.02 mg/kg; Wildlife Pharmaceuticals, Windsor, Colorado, USA) plus xylazine hydrochloride (0.05–0.07 mg/kg; AnaSed, Lloyd Inc., Shenandoah, Iowa, USA), or using premixed butorphanol–azaperone–medetomidine (0.06–0.09 mg medetomidine/kg; BAM, Wildlife Pharmaceuticals). All immobilizing drugs were delivered IM via dart (Pneudart, Williamsport, Pennsylvania, USA). Immobilizing drugs were antagonized with naltrexone (Wildlife Pharmaceuticals; 50 mg IM per mg thiafentanil or 25 mg IM per mg etorphine) and 300 mg tolazoline IM (Tolazine, Akorn Animal Health, Lake Forest, Illinois, USA). In cases where BAM was used, animals were reversed with atipamazole (Wildlife Pharmaceuticals) at 5 mg per 1 mg medetomidine, naltrexone (0.05–0.125 mg/kg IM), and tolazoline (1 mg/kg IM). In animals deemed still pregnant, we obtained a blood sample and replaced the vaginal transmitter. On evidence of abortion or stillbirth, we collected blood, vaginal swabs, feces, and milk, if present, from the dam, and we collected the placenta and fetus when possible. We collected milk from each quarter of the udder into 50-mL polyethylene centrifuge tubes (Fisher Scientific). Feces were placed in Whirl-Pak bags (Nasco, Fort Atkinson, Wisconsin, USA). Placentas were collected into 1-gallon resealable plastic storage bags. Swabs, milk, feces, and placenta were stored at –80 C until shipment to USDA-APHIS VS, National Veterinary Services Laboratories (NVSL) for Brucella culture (culture methods described under Bacteriology). Fetuses/dead calves were collected in doubled large plastic bags and transported to MVDL for necropsy and culture. Personnel handling any potentially infectious materials took all necessary precautions, including using appropriate personal protective equipment.

On observation of an apparently normal full-term calf, the dam was immobilized within 5 d of parturition and the samples listed above were collected. We also collected conjunctival swabs (as described above for vaginal swabs), blood, and feces from the calf. In June or July of each year, we manually restrained nonpregnant animals for blood and vaginal swab collection, and handled (either manually or through chemical immobilization) the males to collect blood for brucellosis testing to verify seronegativity before breeding.

Aborted fetuses, stillborn calves, and adult animals that died or were humanely euthanized before the end of the trial were transported to the MVDL for necropsy and tissue collection for Brucella culture. At conclusion of the study, we humanely euthanized the remaining Treatment animals on-site and collected blood and tissues. Nontreated Control and Sentinel animals were shipped to slaughter, where blood and tissues were collected at the slaughter establishment. Postmortem tissues collected from mortalities, aborted fetuses, stillbirths, and animals euthanized at study termination included mandibular, parotid, medial retropharyngeal, prescapular, internal iliac, tracheobronchial, hepatic, mesenteric, prefemoral, mammary and popliteal lymph nodes, mammary gland, uterus, ovary, placenta, ileum, lung, liver, spleen, kidney. We used appropriate personal protective equipment and safety procedures when working with all carcasses and tissues. We sent samples collected at the end of the study to NVSL for serology and culture.

Serological testing

We tested for antibodies against Brucella. On initial collection of the animals, standard card and FPA tube tests were carried out chute-side. Throughout the remainder of the study, standard card, FPA tube, buffered acidified plate antigen (BAPA), complement fixation (CF), and serological tests, as described in the 2001 USDA Brucellosis Uniform Methods and Rules and in NVSL protocols, were performed at the MVDL or NVSL using standard procedures (USDA-APHIS 2003, 2006a, 2006b, 2006c, 2011). We defined positive animals as those testing positive on either CF or FPA. Animals considered suspect produced results in the suspect range of CF or FPA, or were positive on standard card or BAPA without a positive result on either CF or FPA.

Bacteriology

Swabs, feces, milk, and tissues shipped to the NVSL for bacteriologic culture underwent standard isolation protocols (Alton et al. 1988) with a modification to use a blender to homogenize tissues. All isolation procedures were performed using appropriate biosafety level 3 facilities and procedures. Isolates and DNA were inactivated before removal for molecular procedures and viability testing was performed to ensure inactivation on all samples. Tissues obtained from necropsy at MVDL also were cultured as described in Alton et al. (1988); and Brucella-suspect isolates were shipped to NVSL for confirmation. At NVSL, isolates were identified using a modified AMOS PCR, which includes primers to differentiate vaccine strains (Bricker and Halling 1994a, 1994b; Bricker et al. 2003; Ewalt and Bricker 2003). Tissues with confirmed growth of B. abortus were considered culture positive, as was the animal from which the tissues came.

Statistical analysis

We analyzed the probability of shedding events (defined as parturition associated with culture-positive fluids or tissues; positive milk excluded) produced by any nonsentinel female using a generalized linear mixed model fit by maximum likelihood using Laplace approximation. The GLMER model used is a function within the various linear mixed models of the R programming environment packages ‘lme4’ and ‘lmerTest’ (R Core Team 2022). We used a negative binomial approach given the large number of nonshedding events (n=160; defined as any nonsentinel animal that did not produce a shedding event regardless of pregnancy status), versus shedding events (n=19). This created an overdispersed and clumped data set that seemed best suited to a negative binomial distribution for modelling purposes (Bolker 2015). We opted to keep the four groups separate because of differences in the research practices between the two replicates, which seemed large enough to warrant their separation. It would have been difficult to account for these differences in the model. These differences included variation in immunocontraceptive mixing methods (Frey et al. 2024), removal of animals, delayed start of Control Rep2, and early cessation of Replicate 2. In addition, because of the year delay in establishing Control Rep2, we did not include data from the first year of Treatment Rep2 in the analysis, because it lacked a control during that time. Keeping the groups separate resulted in treatment status by replicate becoming one of the fixed effect variables. We acknowledge that this is a somewhat less standard than combining the data and testing only for the effect of treatment, but as described above, the introduced biases in the data required a different approach.

We selected three fixed effect variables to inform or predict shedding. These included treatment status with four factor levels, treatment or control separated by replicate, year (scaled) and animal age (scaled). Scaling was done to avoid giving these variables undue weight as numeric values, allowing the model more numeric stability and better convergence (Fox 2015). Each year in the study represents the calendar year following a breeding season (again, year 1 began in the January following an animal’s first exposure to males in the study). We included unique animal ID as a random effect to account for the repeated measure in this longitudinal study design.

Ethics approval

This study was approved by the Institutional Animal Care and Use Committee (IACUC) of the USDA-APHIS, Wildlife Services, National Wildlife Research Center (Protocol QA-1858) and the IACUC of the USDA-APHIS, VS, Bison Quarantine Research Facility (Title: Evaluation of GonaCon, an immunocontraceptive vaccine as a means of decreasing shedding of Brucella abortus in bison). In addition, this study was made available for public comment via an environmental impact statement in 2012, in which “no significant impact” was the final determination (USDA-APHIS 2012a, 2012b).

Treatment failures, losses, and censored animals

Within year 1 of Replicates 1 and 2, treatment failure (i.e., breakthrough pregnancy) occurred in 3/15 Treatment rep1 bison and 1/20 Treatment rep2 bison. These individuals were removed from the study. In addition, 10 bison from the first replicate and three cows from the second died over the course of the study due to trauma, peritonitis, dystocia, or pneumonia. Ultimately, eight animals were censored from the statistical analysis due to immunocontraceptive failure in year 1 (four), mortality before calving in year 1 of both replicates (three), and one Control Rep1 animal that never bred and was removed after year 3.

Shedding events and seroconversion of sentinels

Control rep1 animals produced 12 shedding events out of 47 birth events that occurred and 57 possible birth events over five reproductive seasons (Table 1 and Supplementary Material Table S1). The first Control rep1 shedding event occurred late January, preceding detection of the first seroconversion and abortion (albeit culture-negative) in a SentinelControl rep1 animal 15 d later. Two more Control rep1 animals produced shedding events and two more SentinelControl rep1 animals seroconverted and produced shedding events in year 1. In contrast, Treatment rep1 animals gave birth to seven culture-negative live calves out of 50 possible births. The five SentinelTreatment rep1 animals produced 22 culture-negative parturitions during the first trial. No sentinels seroconverted in the Treatment rep1 throughout the study.

Table 1. Brucella abortus shedding events and seroconversion observed in captive bison (Bison bison) during experimental evaluation of immunocontraception as a disease management tool. Shedding events were the Brucella abortus culture–positive births detected within each group (below) in a year. New seropositive females were the number of female bison in each sentinel group (below) that newly tested seropositive for B. abortus exposure in January, at calving, or in June–July of the same study year. Annual and overall totals are shown. The experiment included two replicates (rep1, rep2). Each replicate included a group of immunocontraceptive-treated seropositive females (Treatment) and a group of untreated seropositive females (Control) housed in separate paddocks. Treatment groups received a one-time intramuscular administration of 3,000 μg gonadotropin-releasing hormone-based immunocontraceptive in May of the year before onset of the replicate. Smaller numbers of seronegative female sentinels (denoted as Sentinelgroup) were housed with each of the four primary experimental groups (Treatment rep1, Treatment rep2, Control rep1, Control rep2) to monitor for transmission. Study years were defined as January−December. The first replicate trial ran for 5 yr (2013−2017); the second replicate trial had a staggered start and ended after 2 or 3 yr of observing Control rep2 (2016−2017) or Treatment rep2 (2015−2017), respectively. See text for additional details of experimental design and study methods.

Table 1. Brucella abortus shedding events and seroconversion observed in captive bison (Bison bison) during experimental evaluation of immunocontraception as a disease management tool. Shedding events were the Brucella abortus culture–positive births detected within each group (below) in a year. New seropositive females were the number of female bison in each sentinel group (below) that newly tested seropositive for B. abortus exposure in January, at calving, or in June–July of the same study year. Annual and overall totals are shown. The experiment included two replicates (rep1, rep2). Each replicate included a group of immunocontraceptive-treated seropositive females (Treatment) and a group of untreated seropositive females (Control) housed in separate paddocks. Treatment groups received a one-time intramuscular administration of 3,000 μg gonadotropin-releasing hormone-based immunocontraceptive in May of the year before onset of the replicate. Smaller numbers of seronegative female sentinels (denoted as Sentinelgroup) were housed with each of the four primary experimental groups (Treatment rep1, Treatment rep2, Control rep1, Control rep2) to monitor for transmission. Study years were defined as January−December. The first replicate trial ran for 5 yr (2013−2017); the second replicate trial had a staggered start and ended after 2 or 3 yr of observing Control rep2 (2016−2017) or Treatment rep2 (2015−2017), respectively. See text for additional details of experimental design and study methods.
Table 1. Brucella abortus shedding events and seroconversion observed in captive bison (Bison bison) during experimental evaluation of immunocontraception as a disease management tool. Shedding events were the Brucella abortus culture–positive births detected within each group (below) in a year. New seropositive females were the number of female bison in each sentinel group (below) that newly tested seropositive for B. abortus exposure in January, at calving, or in June–July of the same study year. Annual and overall totals are shown. The experiment included two replicates (rep1, rep2). Each replicate included a group of immunocontraceptive-treated seropositive females (Treatment) and a group of untreated seropositive females (Control) housed in separate paddocks. Treatment groups received a one-time intramuscular administration of 3,000 μg gonadotropin-releasing hormone-based immunocontraceptive in May of the year before onset of the replicate. Smaller numbers of seronegative female sentinels (denoted as Sentinelgroup) were housed with each of the four primary experimental groups (Treatment rep1, Treatment rep2, Control rep1, Control rep2) to monitor for transmission. Study years were defined as January−December. The first replicate trial ran for 5 yr (2013−2017); the second replicate trial had a staggered start and ended after 2 or 3 yr of observing Control rep2 (2016−2017) or Treatment rep2 (2015−2017), respectively. See text for additional details of experimental design and study methods.

Control rep2 animals produced five B. abortus shedding events out of 17 birth events and 29 possible births, and SentinelControl rep2 sustained four shedding events out of 10 birth events over 2 yr (Table 1 and Supplementary Material Table S1). Two of the three Control rep2 shedding events in year 1 preceded the first two seroconversions or shedding events in SentinelControl rep2. In Treatment rep2, one of six females with breakthrough pregnancies in the second or third years produced two B. abortus shedding events. No shedding events were detected among the SentinelTreatment rep2 females, but one animal was detected as seropositive in January of year 3. The five remaining SentinelTreatment rep2 animals remained seronegative through the third calving season, but four of these animals were identified as seropositive in December of year 3.

Only the fixed effect variable, “treatment status,” which equated to the four group assignments, provided predictive value for shedding events (P≤0.001; Table 2). Control rep1, Control rep2, and Treatment rep2 had similar values (log odds −1.71, −0.23, and −1.39, respectively; Table 2a). Treatment rep1 had a log odds value of −25.36 indicating a much lower shedding probability relative to the other three groups. Fixed effect predictor covariates, year (scaled) and age (scaled), had no explanatory value, and were ultimately removed from the model. The random effect of unique animal ID, represented by the interclass correlation coefficient value output, was very low and indicates no influence on the model outcome (Table 2b).

Table 2. (a) and (b). Model output for generalized linear mixed model fitted by maximum-likelihood methods using a negative binomial distribution to determine which variables predicted shedding of Brucella abortus in bison in a replicated experiment evaluating immunocontraception in Brucella-seropositive female bison as a disease management tool. The experiment included two replicates (rep1, rep2). Each replicate included a group of immunocontraceptive-treated seropositive females (Treatment) and a group of untreated seropositive females (Control) housed in separate paddocks. Treatment groups received a one-time intramuscular administration of 3,000 μg gonadotropin-releasing hormone-based immunocontraceptive in May of the year before onset of the replicate. Smaller numbers of seronegative female sentinels were housed with each of the four primary experimental groups (Treatment rep1, Treatment rep2, Control rep1, Control rep2) to monitor for transmission. Study years were defined as January−December. The first replicate trial ran for 5 yr (2013 − 2017); the second replicate trial had a staggered start and ended after 2 or 3 yr of observing Control rep2 (2016 − 2017) or Treatment rep2 (2015 − 2017), respectively. We selected three fixed effect variables to inform or predict shedding. These included treatment status with four factor levels: treatment or control separated by replicate, project year (scaled), and animal age (scaled). Only treatment status provided predictive value in regards to shedding. The difference in log odd values between the Control rep1 and Treatment rep1 is large. Treatment rep1 had no shedding events. Though Treatment rep2 indicates a value that could be interpreted within the realm of statistical significance, this is more likely an artifact of small sample size and a single individual being positive two consecutive years. The random effect of unique animal ID, represented by the ICC value output, was very low and indicates no influence on the model outcome.

Table 2. (a) and (b). Model output for generalized linear mixed model fitted by maximum-likelihood methods using a negative binomial distribution to determine which variables predicted shedding of Brucella abortus in bison in a replicated experiment evaluating immunocontraception in Brucella-seropositive female bison as a disease management tool. The experiment included two replicates (rep1, rep2). Each replicate included a group of immunocontraceptive-treated seropositive females (Treatment) and a group of untreated seropositive females (Control) housed in separate paddocks. Treatment groups received a one-time intramuscular administration of 3,000 μg gonadotropin-releasing hormone-based immunocontraceptive in May of the year before onset of the replicate. Smaller numbers of seronegative female sentinels were housed with each of the four primary experimental groups (Treatment rep1, Treatment rep2, Control rep1, Control rep2) to monitor for transmission. Study years were defined as January−December. The first replicate trial ran for 5 yr (2013 − 2017); the second replicate trial had a staggered start and ended after 2 or 3 yr of observing Control rep2 (2016 − 2017) or Treatment rep2 (2015 − 2017), respectively. We selected three fixed effect variables to inform or predict shedding. These included treatment status with four factor levels: treatment or control separated by replicate, project year (scaled), and animal age (scaled). Only treatment status provided predictive value in regards to shedding. The difference in log odd values between the Control rep1 and Treatment rep1 is large. Treatment rep1 had no shedding events. Though Treatment rep2 indicates a value that could be interpreted within the realm of statistical significance, this is more likely an artifact of small sample size and a single individual being positive two consecutive years. The random effect of unique animal ID, represented by the ICC value output, was very low and indicates no influence on the model outcome.
Table 2. (a) and (b). Model output for generalized linear mixed model fitted by maximum-likelihood methods using a negative binomial distribution to determine which variables predicted shedding of Brucella abortus in bison in a replicated experiment evaluating immunocontraception in Brucella-seropositive female bison as a disease management tool. The experiment included two replicates (rep1, rep2). Each replicate included a group of immunocontraceptive-treated seropositive females (Treatment) and a group of untreated seropositive females (Control) housed in separate paddocks. Treatment groups received a one-time intramuscular administration of 3,000 μg gonadotropin-releasing hormone-based immunocontraceptive in May of the year before onset of the replicate. Smaller numbers of seronegative female sentinels were housed with each of the four primary experimental groups (Treatment rep1, Treatment rep2, Control rep1, Control rep2) to monitor for transmission. Study years were defined as January−December. The first replicate trial ran for 5 yr (2013 − 2017); the second replicate trial had a staggered start and ended after 2 or 3 yr of observing Control rep2 (2016 − 2017) or Treatment rep2 (2015 − 2017), respectively. We selected three fixed effect variables to inform or predict shedding. These included treatment status with four factor levels: treatment or control separated by replicate, project year (scaled), and animal age (scaled). Only treatment status provided predictive value in regards to shedding. The difference in log odd values between the Control rep1 and Treatment rep1 is large. Treatment rep1 had no shedding events. Though Treatment rep2 indicates a value that could be interpreted within the realm of statistical significance, this is more likely an artifact of small sample size and a single individual being positive two consecutive years. The random effect of unique animal ID, represented by the ICC value output, was very low and indicates no influence on the model outcome.

Final Brucella culture and serostatus

Supplementary Material Table S1 presents individual animal data on culture status at final disposition. Table 3 summarizes the initial serologic status and final serologic and culture status of all study animals. Four originally seropositive animals became seronegative at some time during the study and remained so until final disposition. Four animals originally seropositive were intermittently seronegative or suspect during the study but were ultimately seropositive at final disposition.

Table 3. Number of Brucella-seropositive female bison in each group out of total animals in each group at study initiation, and number and percent of seropositive and culture-positive female bison in each group out of total animals in each group at final disposition. Study is a replicated experiment evaluating immunocontraception in Brucella-seropositive female bison as a disease management tool. The experiment included two replicates (rep1, rep2). Each replicate included a group of immunocontraceptive-treated seropositive females (Treatment) and a group of untreated seropositive females (Control) housed in separate paddocks. Treatment groups received a one-time intramuscular administration of 3,000 μg gonadotropin-releasing hormone-based immunocontraceptive in May of the year before onset of the replicate. Smaller numbers of seronegative female sentinels (denoted as Sentinelgroup) were housed with each of the four primary experimental groups (Treatment rep1, Treatment rep2, Control rep1, Control rep2) to monitor for transmission. Study years were defined as January–December. The first replicate trial ran for 5 yr (2013−2017); the second replicate trial had a staggered start and ended after 2 or 3 yr of observing Control rep2 (2016−2017) or Treatment rep2 (2015−2017), respectively. See text for additional details of experimental design and study methods.

Table 3. Number of Brucella-seropositive female bison in each group out of total animals in each group at study initiation, and number and percent of seropositive and culture-positive female bison in each group out of total animals in each group at final disposition. Study is a replicated experiment evaluating immunocontraception in Brucella-seropositive female bison as a disease management tool. The experiment included two replicates (rep1, rep2). Each replicate included a group of immunocontraceptive-treated seropositive females (Treatment) and a group of untreated seropositive females (Control) housed in separate paddocks. Treatment groups received a one-time intramuscular administration of 3,000 μg gonadotropin-releasing hormone-based immunocontraceptive in May of the year before onset of the replicate. Smaller numbers of seronegative female sentinels (denoted as Sentinelgroup) were housed with each of the four primary experimental groups (Treatment rep1, Treatment rep2, Control rep1, Control rep2) to monitor for transmission. Study years were defined as January–December. The first replicate trial ran for 5 yr (2013−2017); the second replicate trial had a staggered start and ended after 2 or 3 yr of observing Control rep2 (2016−2017) or Treatment rep2 (2015−2017), respectively. See text for additional details of experimental design and study methods.
Table 3. Number of Brucella-seropositive female bison in each group out of total animals in each group at study initiation, and number and percent of seropositive and culture-positive female bison in each group out of total animals in each group at final disposition. Study is a replicated experiment evaluating immunocontraception in Brucella-seropositive female bison as a disease management tool. The experiment included two replicates (rep1, rep2). Each replicate included a group of immunocontraceptive-treated seropositive females (Treatment) and a group of untreated seropositive females (Control) housed in separate paddocks. Treatment groups received a one-time intramuscular administration of 3,000 μg gonadotropin-releasing hormone-based immunocontraceptive in May of the year before onset of the replicate. Smaller numbers of seronegative female sentinels (denoted as Sentinelgroup) were housed with each of the four primary experimental groups (Treatment rep1, Treatment rep2, Control rep1, Control rep2) to monitor for transmission. Study years were defined as January–December. The first replicate trial ran for 5 yr (2013−2017); the second replicate trial had a staggered start and ended after 2 or 3 yr of observing Control rep2 (2016−2017) or Treatment rep2 (2015−2017), respectively. See text for additional details of experimental design and study methods.

Our findings suggest that halting reproduction in Brucella-seropositive bison can reduce B. abortus shedding events, if not prevent them entirely. In both replicates, no shedding events occurred as long as pregnant animals were removed from the groups. As we allowed breakthrough pregnancies to remain in the treatment groups in subsequent years, our intent was to observe whether shedding, and thus transmission, might continue to be reduced over time, despite a gradual return to fertility. This concept appeared to be strongly supported by Replicate 1, wherein Treatment rep1 did not produce any shedding events in seven parturitions among four individuals with breakthrough pregnancies; this is in contrast to 12 shedding events produced by Control rep1. Low probability of shedding in Treatment rep1 was also reflected in the data analysis. However, the very different outcome observed in Replicate 2 emphasizes that the production of shedding events in subsequent years, and the impacts of those shedding events in terms of disease transmission in a herd, are highly unpredictable. This is especially apparent in that, although only 1/6 Treatment rep2 animals with breakthrough pregnancies after year 1 produced shedding events, these two events apparently led to seroconversion in all but one of the sentinels; an outcome effectively indistinguishable from that observed in Control rep2. Therefore, use of tools such as immunocontraception for disease control will likely need nearly perfect efficacy in preventing pregnancy, lasting several years, to reduce disease prevalence in an affected herd successfully.

Three of five SentinelControl rep1 seroconversions and two of three SentinelControl rep2 seroconversions could almost certainly be attributed to direct exposure to parturition-associated shedding events produced by their respective control groups, as these seroconversions were detected shortly after Brucella-positive birth events occurred in both control groups. Unfortunately, we were unable to determine a more precise time frame regarding when transmission occurred in the remainder of sentinels that seroconverted, as detections in those animals were not made until December and January. This would have been important to document more precisely, because using a contraceptive tool relies on the premise that the vast majority of horizontal Brucella spp. transmissions are associated with infectious parturition events. In retrospect, more frequent testing between August and November should have occurred each year to estimate timing of seroconversion in relation to known shedding events within their groups.

Although our data do suggest that overall shedding events do diminish as animals age, age was not important as an explanatory variable in our analysis. Age, in the context of targeting younger animals for reproductive removal from a brucellosis-affected bison herd, is considered an important strategy in managing brucellosis. Models have indicated that sterilization or removal of probably infectious seropositive females at or under 5 yr of age, can potentially aid in reducing disease prevalence in the GYA bison population (Ebinger et al. 2011; Treanor et al. 2011; Hobbs et al. 2015). However, in our study, 6-yr-old and 8-yr-old recently seroconverted sentinels produced shedding events. This was probably because they had only recently been infected with B. abortus, suggesting that timing of infection is perhaps the more influential factor, rather than age.

Brucella abortus was recovered from only a few animals in the first replicate at termination of the study. In contrast, the Replicate 2 had many more culture-positive animals at end of the study. As discussed above, this was probably because the second replicate animals had been more recently infected than the animals in the Replicate 1 at time of necropsy. This outcome is consistent with findings by Treanor et al. (2011), where older female bison, over 5 yr of age, had reduced probability of harboring active infection, probably because their infections were acquired at a young age. In the case of this study, however; we must also consider the possibility that occurrence of fewer culture-positive animals in Replicate 1 were due to sample handling or laboratory issues, such as variable durations of storage time at −70 C before culturing, but we regard this as a less likely explanation.

In both replicates, none of the healthy offspring from seropositive dams that were associated with Brucella-negative births, were identified as seropositive when tested at 9 mo of age, or throughout the remainder of the study, unless they were purposefully exposed to infectious birth events in following years. We therefore did not see any evidence of vertical transmission in those calves, despite being born to and having nursed from seropositive females. These offspring did initially have maternal antibodies to Brucella, which waned by approximately 6 mo. This is consistent with the report from Rhyan et al. (2009), where only a few (3/20) calves born to seropositive bison were seropositive at 5–6 mo. Although there is a possibility that some of these calves could be latently infected, latency is considered an uncommon occurrence (Ray et al. 1998).

It is interesting to note that 1/5 Control rep1 sentinels did not seroconvert until the fourth reproductive season, despite the occurrence of 16 shedding events in the 3 yr prior, through which the other four sentinels became positive on serology during the first two seasons. In contrast, one Treatment rep2 animal produced two shedding events that seem to have led to the eventual seroconversion of 5/6 associated sentinels. Differences in occurrence of transmission of B. abortus within free-ranging herds may be attributed to a variety of animal and environmental conditions, including herd dynamics, animal-specific bacterial shedding levels, plane of nutrition, disease resistance, ambient temperatures, severity of winters, and other environmental factors (Aune et al. 2012; Treanor 2013; Treanor et al. 2015). Controlling for such variables can be challenging; therefore, strategically deploying the most effective possible reproductive intervention would be very important to reduce disease prevalence in a bison herd successfully.

None of the male bison seroconverted during the study, despite multiple exposures to Brucella-infected females. Venereal transmission of B. abortus is thought unlikely to occur in cattle and bison. Previous studies exploring male to female transmission have been unsuccessful; however, seroconversion did occur in female bison artificially inseminated with attenuated B. abortus Strain 19, and virulent B. abortus has been isolated from semen collected from male bison in the GYA (Crawford et al. 1990; Robison et al. 1998; Frey et al. 2013; Uhrig et al. 2013). More work is needed to determine if venereal transmission does occur, as that will influence decisions regarding the need for disease or reproductive intervention in males.

Preventing reproduction in brucellosis-positive bison might be an effective, nonlethal way to exclude these animals from participating in the epizootiology of this disease within a herd (Miller et al. 2004; Ebinger et al. 2011; Rhyan et al. 2013a). However, we observed in our study that an imperfect contraceptive tool can still lead to high-consequence shedding events that have the potential to transmit B. abortus to many susceptible animals in the herd. Success of immunocontraception in reducing disease prevalence would probably require a higher and longer-lasting efficacy than we observed in these bison. In addition, tools such as this should be implemented in conjunction with other tools such as vaccination against brucellosis, with adaptive modifications as prevalence changes over time. Further research is needed, such as building upon existing models in which simulations include application of pregnancy prevention at different efficacy levels and durations, to assess whether reduction in brucellosis prevalence can be attained using this strategy.

We are very grateful to J. Lombard, J. Powers, and R. Wallen for their advice on study design and analysis. Many thanks to P. Hoppe and D. Tilton for their efforts in the care of the animals and to B. Thompson for his expertise in bison pregnancy evaluations. Thanks to H. Michaelbrink and F. Rigler for hosting the animals. We appreciate M. Zaluski, T. Linfield, J. Mount, D. Knopp, and A. Fuentes-Sanchez for their operations and laboratory support throughout the study. Thank you to M. Islam for his review of and assistance with statistical analysis. Many thanks to S. Cox, J. Lombard, and G. Murphy for their excellent reviews of the manuscript. Finally, much appreciation goes to L. Miller in his role as co-inventor of GonaCon and supporter of the use of contraception for disease management in wildlife. The findings and conclusions in this publication are those of the authors and should not be construed to represent any official USDA or US Government determination or policy.

Supplementary material for this article is online at http://dx.doi.org/10.7589/JWD-D-21-00167.

Alton
GG,
Jones
LM,
Angus
RD,
Verger
JM.
1988
.
Techniques for the brucellosis laboratory
.
INRA Publications
,
Paris, France
.
Aune
K,
Rhyan
JC,
Russell
CR,
Roffe
TJ,
Corso
B.
2012
.
Environmental persistence of Brucella abortus in the Greater Yellowstone Area
.
J Wildl Manage
76
:
253
261
.
Bienen
L,
Tabor
G.
2006
.
Applying an ecosystem approach to brucellosis control: Can an old conflict between wildlife and agriculture be successfully managed
?
Front Ecol Environ
4
:
319
327
.
Bolker
BM.
2015
. Linear and generalized linear mixed models. In:
Ecological statistics: Contemporary theory and application
,
Fox
GA,
Negrete-Yankelevich
S,
Sosa
VJ
, editors.
Oxford University Press
,
Oxford, UK
, pp.
309
333
.
Bricker
BJ,
Ewalt
DR,
Olsen
SC,
Jensen
AE.
2003
.
Evaluation of the Brucella abortus species-specific polymerase chain reaction assay, an improved version of the Brucella AMOS polymerase chain reaction assay for cattle
.
J Vet Diagn Invest
15
:
374
378
.
Bricker
BJ,
Halling
SM.
1994a
.
Differentiation of Brucella abortus (biovars 1, 2, and 4), Brucella melitensis, Brucella ovis, and Brucella suis (biovar 1) by the polymerase chain reaction
.
J Clin Microbiol
32
:
2660
2666
.
Bricker
BJ,
Halling
SM.
1994b
.
Enhancement of the Brucella AMOS PCR assay for differentiation of Brucella abortus vaccine strains S19 and RB51
.
J Clin Microbiol
33
:
1640
1642
.
Cheville
NF,
McCullough
DR,
Paulson
LR.
1998
.
Brucellosis in the Greater Yellowstone Area
.
National Research Council, The National Academies Press
,
Washington, DC
,
204
pp.
Clarke
PR,
Frey
RK,
Rhyan
JC,
McCollum
MP,
Nol
P,
Aune
K.
2014
.
Feasibility of quarantine procedures for bison (Bison bison) calves from Yellowstone National Park for conservation of brucellosis-free bison
.
J Am Vet Med Assoc
244
:
588
591
.
Crawford
RP,
Huber
JD,
Adams
BS.
1990
.
Epidemiology and surveillance
.
CRC Press
,
Boca Raton, Florida, USA
.
Ebinger
M,
Cross
P,
Wallen
R,
White
PJ,
Treanor
J.
2011
.
Simulating sterilization, vaccination, and test-and-remove as brucellosis control measures in bison
.
Ecol Appl
21
:
2944
2959
.
Ewalt
DR,
Bricker
BJ.
2003
. Identification and differentiation of Brucella abortus field and vaccine strains by BaSS-PCR. In:
Methods in molecular biology: PCR detection of microbial pathogens: Methods and protocols
.
Sachse
K,
Frey
J
, editors.
Humana Press
,
Totowa, New Jersey
, pp.
97
108
.
Fox
GA.
2015
. Linear and generalized linear mixed models. In:
Ecological statistics: Contemporary theory and application
.
Fox
GA,
Negrete-Yankelevich
S,
Sosa
VJ
, editors.
Oxford University Press
,
Oxford, UK
, pp.
106
129
.
Frey
RK,
Clarke
PR,
McCollum
MP,
Nol
P,
Johnson
KR,
Thompson
BD,
Ramsey
JM,
Anderson
NJ,
Rhyan
JC.
2013
.
Evaluation of bison (Bison bison) semen from Yellowstone National Park, Montana, USA, bulls for Brucella abortus shedding
.
J Wildl Dis
49
:
714
717
.
Frey
R,
Wehtje
M,
Nol
P,
Clarke
P,
Rhyan
J,
McCollum
M,
Miller
L,
Eckery
D.
2024
.
Brucella abortus infection in bison (Bison bison), and the effects of immunocontraception on pregnancy
.
J Wildl Dis
60
:339–345.
Fuller
JA,
Garrott
RA,
White
PJ,
Aune
KE,
Roffe
TJ,
Rhyan
JC.
2007
.
Reproduction and survival of Yellowstone bison
.
J Wildl Manage
71
:
2365
2372
.
Hobbs
NT,
Geremia
C,
Treanor
J,
Wallen
R,
White
PJ,
Hooten
MB,
Rhyan
JC.
2015
.
State-space modeling to support management of brucellosis in the Yellowstone bison population
.
Ecol Monogr
85
:
525
556
.
Kamath
PL,
Foster
JT,
Drees
KP,
Luikart
G,
Quance
C,
Anderson
NJ,
Clarke
PR,
Cole
EK,
Drew
ML,
Edwards
WH,
Rhyan
JC,
Treanor
JJ,
Wallen
RL,
White
PJ,
Robbe-Austerman
S,
Cross
PC.
2016
.
Genomics reveals historic and contemporary transmission dynamics of a bacterial disease among wildlife and livestock
.
Nat Commun
7
:
11448
.
Miller
LA,
Gionfriddo
JP,
Fagerstone
KA,
Rhyan
JC,
Killian
GK.
2008
.
The single‐shot GnRH immunocontraceptive vaccine (GonaCon) in white‐tailed deer: Comparison of several GnRH preparations
.
Am J Reprod Immunol
60
:
214
223
.
Miller
LA,
Rhyan
JC,
Drew
M.
2004
.
Contraception of bison by GnRH vaccine: A possible means of decreasing transmission of brucellosis in bison
.
J Wildl Dis
40
:
725
730
.
National Academies of Sciences, Engineering, and Medicine
.
2020
.
Revisiting brucellosis in the Greater Yellowstone Area
.
National Academies Press
Washington, DC
. https://doi.org/10.17226/24750.
Olsen
SC.
2010
.
Brucellosis in the United States: Role and significance of wildlife reservoirs
.
Vaccine
28
:
F73
F76
.
Olsen
SC,
Tatum
F.
2010
.
Bovine brucellosis
.
Vet Clin of North Am Food Anim Pract 26:15–27
.
R Core Team.
2022
. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/.
Ragan
VE.
2002
.
The Animal and Plant Health Inspection Service (APHIS) brucellosis eradication program in the United States
.
Vet Microbiol
90
:
11
18
.
Ray
WC,
Brown
RR,
Stringfellow
DA,
Schnurrenberger
PR,
Scanlan
CM,
Swann
AI.
1988
.
Bovine brucellosis: An investigation of latency in progeny of culture-positive cows
.
J Am Vet Med Assoc
192
:
182
186
.
Rhyan
JC,
Aune
K,
Roffe
T,
Ewalt
D,
Hennager
S,
Gidlewski
T,
Olsen
S,
Clarke
R.
2009
.
Pathogenesis and epidemiology of brucellosis in Yellowstone bison: Serologic and culture results from adult females and their progeny
.
J Wildl Dis
45
:
729
739
.
Rhyan
JC,
Gidlewski
T,
Roffe
TJ,
Aune
K,
Philo
LM,
Ewalt
DR.
2001
.
Pathology of brucellosis in bison from Yellowstone National Park
.
J Wildl Dis
37
:
101
109
.
Rhyan
JC,
Miller
LA,
Fagerstone
KA.
2013a
.
The use of contraception as a disease management tool in wildlife
.
J Zoo Wildl Med
44
:
S135
S137
.
Rhyan
JC,
Nol
P.
2019
.
Brucellosis. In Fowler’s zoo and wild animal medicine current therapy
. Vol.
1
.
Mille
RE,
Lamberski
N,
Calle
PP
, editors.
Elsevier
,
Amsterdam, Netherlands
.
Rhyan
JC,
Nol
P,
Quance
C,
Gertonson
A,
Belfrage
J,
Harris
L,
Straka
K,
Robbe-Austerman
S.
2013b
.
Transmission of brucellosis from elk to cattle and bison, Greater Yellowstone Area, U.S.A., 2002–2012
.
Emerging Infect Dis
19
:
1992
1995
.
Robison
CD,
Davis
DS,
Templeton
JW,
Westhusin
M,
Foxworth
WB,
Gilsdorf
MJ,
Adams
LG.
1998
.
Conservation of germ plasm from bison infected with Brucella abortus
.
J Wildl Dis
34
:
582
589
.
Treanor
JJ,
Geramaia
C,
Ballou
MA,
Keisler
DH,
White
PJ,
Cox
JJ,
Crowley
PH.
2015
.
Maintenance of brucellosis in Yellowstone bison: Linking seasonal food resources, host–pathogen interaction, and life-history trade-offs
.
Ecol Evol
5
:
3783
3799
.
Treanor
JJ.
2013
.
Integrating ecology with management to control wildlife brucellosis
.
Rev Sci Tech
32
:
239
247
.
Treanor
JJ,
Geremia
C,
Crowley
PH,
Cox
J,
White
PJ,
Wallen
RL,
Blanton
DW.
2011
.
Estimating probabilities of active brucellosis infection in Yellowstone bison through quantitative serology and tissue culture
.
J Appl Ecol
48
:
1324
1332
.
Uhrig
SR,
Nol
P,
McCollum
P,
Salman
M,
Rhyan
JC.
2013
.
Transmission of Brucella abortus Strain 19 in bison by intravaginal, intrauterine,and intraconjunctival inoculation
.
J Wildl Dis
49
:
522
526
.
USDA-APHIS (United States Department of Agriculture, Animal and Plant Health Inspection Service).
2003
. Brucellosis eradication: Uniform methods and rules. APHIS91-45-013.
USDA-APHIS.
2006a
. Brucella abortus fluorescence polarization assay (FPA) using tubes. National Veterinary Services Laboratories protocol SOP-SERO-1051.03:1–10.
USDA-APHIS.
2006b
. Buffered acidified plate antigen test for detection of antibodies to Brucella abortus. National Veterinary Services Laboratories protocol SOP-SERO-1024.04:1–5.
USDA-APHIS.
2006c
. Card test for detection of antibodies to Brucella abortus. National Veterinary Services Laboratories protocol SOP-SERO-1027.04:1–5.
USDA-APHIS.
2011
. Complement fixation test for detection of antibodies to Brucella abortus and Brucella suis. National Veterinary Services Laboratories protocol SOP-SERO-1022.09:1–68.
USDA-APHIS
.
2012a
. Evaluation of GonaCon, an immunocontraceptive vaccine, as a means of decreasing transmission of Brucella abortus in bison in the Greater Yellowstone Area.
Environmental Assessment
.
USDA-APHIS.
2012b
. Finding of no significant impact for the proposed study: Evaluation of GonaCon, an immunocontraceptive vaccine, as a means of decreasing transmission of Brucella abortus in bison in the Greater Yellowstone Area.
US Department of the Interior, National Park Service and USDA-APHIS.
2000
. Final environmental impact statement for the interagency bison management plan for the State of Montana and Yellowstone National Park. Author,
Washington, DC
.

Supplementary data