The bovine practitioner

Generative artificial intelligence (AI), in the form of large language models (LLM), is rapidly emerging as an influential tool, yet most veterinarians have limited training in its effec­tive and responsible use. This paper provides an introductory overview of how LLM chatbots can support mastitis man­agement by synthesizing information, generating treatment protocols and assisting with clinical decision making. While LLMs often produce clinical recommendations that align with accepted mastitis control principles, their performance is inconsistent and highly dependent on prompt quality, user expertise and the completeness of case information. As illus­trated through 3 mastitis case scenarios, LLMs may overlook key clinical details, misidentify pathogens, recommend out­dated practices such as frequent stripping or suggest non-FDA approved therapies. These errors highlight the need for veter­inary oversight and critical appraisal of AI-generated output. Despite these limitations, LLMs are valuable for creating edu­cational materials, bilingual protocols and decision support tools, especially when grounded in user provided documents. As producers increasingly adopt AI tools independently, vet­erinarians must become literate in AI usage to guide appropri­ate use, safeguard antimicrobial stewardship, and maintain clinical standards. Understanding both the capabilities and shortcomings of current models is essential as AI becomes more integrated into dairy herd health management.

The beef cattle industry is the leading sector of the agricul­ture economy in Kansas. The objectives of this study were to evaluate the water quality of Kansas grazing and feedyard water samples, as well as quantify feedyard water use by month for 10 feedyards in western Kansas. Sixteen grazing sites in western Kansas were selected to collect forage and water samples. Forage samples from each grazing site were collected in August 2024. Forage clippings from each opera­tion, which together constituted a representative sample of the pasture, were collected and sent to a feed analysis labora­tory. Monthly water well meter readings were evaluated for nine of the feedyards as one feedyard did not have accurate historical information. Data were evaluated from January 2020 through June 2025. Water pH (P < 0.01) and sulfate (P = 0.06) were significantly associated with sampling month for the grazing operations. For the feedyard water quality analy­ses, water pH (P < 0.01), sulfate (P < 0.01), phosphorus (P = 0.01), calcium (P < 0.01), chromium (P < 0.01), iron (P < 0.01), cobalt (P < 0.01), selenium (P < 0.01), and molybdenum (P < 0.01) were all associated with sampling month; however, magnitude of differences was small by month. Maximum average water use was 14.69 gallons/head/day. Average monthly high tempera­ture was positively associated (P < 0.01) and ration dry matter negatively associated (P = 0.02) with feedyard water use. These results provide a baseline for benchmarking water conserva­tion efforts.

Our objective was to compare the relative effectiveness of three macrolide antimicrobials for the initial treatment of bo­vine respiratory disease (BRD) in medium-risk feedlot steers and heifers that did not receive metaphylaxis. Steers and heif­ers (n = 704; 734 ± 111 lb [333 ± 50 kg]) were enrolled at a com­mercial Kansas feedlot. Animals were eligible for inclusion if they exhibited clinical signs of BRD from natural exposure and had not previously received any antimicrobial treatment, including metaphylaxis, at the feedlot. At the time of first BRD diagnosis, individual animals were randomly allocated to be administered treatment with either tulathromycin, tulathro­mycin + ketoprofen, or tildipirosin. General and generalized linear mixed models, accounting for time and lot, were used for statistical analyses. At enrollment, initial body weight (P = 0.40), rectal temperature (P = 0.50), and days on feed (P = 0.24) did not differ significantly between treatments. There was no evidence of differences in any morbidity outcome during the first 60 days after enrollment, the first 100 days after enroll­ment, or through closeout (P ≥ 0.14). Likewise, there was no evidence of differences in any mortality outcome across these same evaluation periods (P ≥ 0.41). Treatment success, defined as animals that were not retreated and did not die, ranged from 74.46% to 75.45% through closeout (P = 0.96). In medium-risk feedlot cattle managed without on-arrival metaphylaxis, no differences in morbidity or mortality outcomes were ob­served among the three macrolide treatments – tulathromy­cin, tulathromycin + ketoprofen, or tildipirosin.

Nitrate toxicosis from contaminated drinking water is a com­mon cause of acute death in cattle. The objective of this study was to determine if commercially available semi-quantitative colorimetric test strips designed for human drinking water can be used as a screening test to identify elevated nitrate con­centrations in cattle drinking water. Six test strip brands were used to measure nitrate concentrations in water samples with known values. Sample concentrations were prepared and veri­fied via ion chromatography. Ten volunteers were instructed to test water samples and record nitrate levels according to the test strip manufacturer’s instructions. There was no sig­nificant effect of evaluator or evaluator sex. Results were di­chotomized where targeted nitrate concentrations ≥ 100 ppm were considered positive and < 100 ppm negative, resulting in a 50% proportion of positive samples. Sensitivities of the tests ranged from 22.5% to 97.5%. Two brands resulted in few false negatives when screening water samples with sensitivities of 97.5% and 90%, two brands resulted in a moderate number of false negatives with sensitivities of 77.5% and 72.5%, and two brands resulted in numerous false negatives with sensi­tivities of 47.5% and 22.5%. These findings suggest that while some commercially available test strips can be an effective and economical option for producers and veterinarians to utilize for screening water sources for nitrate contamination, others may pose a risk of underestimating harmful nitrate levels. Water samples with nitrate concentrations reaching or exceeding 100 ppm on test strips should undergo confirmatory testing via quantitative analysis.

Intranasal vaccines are critical for early prevention of re­spiratory diseases in neonatal calves; however, concerns remain regarding potential transmission of vaccine virus strains to unvaccinated commingled animals, particularly in high-stress environments where vaccination history may be unknown. This study evaluated the presence, duration and shedding of infectious bovine rhinotracheitis (IBR), parainflu­enza virus type 3 (PI3), and bovine respiratory syncytial virus (BRSV) in nasal secretions of intranasally vaccinated calves and transmission to commingled unvaccinated controls. Three commercially available intranasal vaccines, Inforce 3® (Zoetis), Nasalgen® 3-PMH (Merck), and PrimeLine® IN5 (Huvepharma), were evaluated. Holstein-cross, mixed-sex, colostrum-deprived calves, seronegative and PCR-negative for IBR, PI3 and BRSV, were enrolled. Blood samples were col­lected on days 0 and 21 to measure antibody titers using viral serum neutralization assays. Nasal swabs were collected daily from days 0 to 21 and analyzed by PCR for IBR, PI3 and BRSV. Nasal swabs were also tested for Histophilus somni by PCR to assess prevalence. Calves were randomly assigned to 4 groups housed in separate biosafety level 2 rooms. Each vaccine group contained commingled, inoculated and non-inoculated calves. A fourth sentinel group was monitored for wild-type virus or biosecurity breaches. Intranasal vaccination resulted in detectable virus in nasal secretions and transmission to commingled non-inoculated calves. Seroconversion in inocu­lated and non-inoculated calves confirmed vaccine virus shed­ding and pen-level spread. These findings demonstrate that intranasal vaccine strains can complicate diagnostic interpre­tation during respiratory disease investigations, underscoring the importance of vaccination history when distinguishing vaccine-derived from wild-type viruses in clinical and research settings.