Effect of welfare standards and biosecurity practices on antimicrobial use in beef cattle

The study addresses how farm-level welfare standards and biosecurity practices relate to antimicrobial use (AMU) in beef cattle. Animal welfare—encompassing both biological functioning and behavioral needs—is closely linked to health and productivity, while biosecurity (external and internal) aims to prevent pathogen introduction and spread. Although extensive evidence in pigs and dairy cattle links better welfare and biosecurity to improved health and reduced AMU, similar evidence in beef systems is limited. AMU remains a key management tool in intensive livestock systems but contributes to antimicrobial resistance, prompting calls for judicious use and better AMU monitoring. In Italy, a large proportion of beef cattle are finished in specialized fattening farms with relatively standardized intensive systems, where animals face welfare challenges (e.g., respiratory and digestive disorders, locomotor problems). The study’s purpose was to quantify welfare, biosecurity, and emergency management standards on beef farms and test their association with AMU. The hypothesis was that higher welfare and biosecurity standards would be associated with lower AMU.

Prior studies have documented benefits of high biosecurity and welfare for animal health and performance across species. In pigs, higher biosecurity is associated with lower disease prevalence and improved performance, and interventions targeting welfare-friendly procedures have reduced AMU. Reviews emphasize biosecurity’s role in preventing respiratory diseases in cattle. Despite general recognition of welfare and biosecurity as management pillars, beef-specific evidence on their influence on AMU is scarce. Some studies in veal calves and beef systems reported reduced AMU following improvements in management (e.g., crowding reduction, barn climate). Farmer attitudes can hinder implementation, including beliefs that antimicrobials are essential for welfare or are already used judiciously. Previous work in Italian beef systems found breed-related differences and substantial between-farm variability in AMU, suggesting farm-specific factors (welfare, biosecurity, vet–farmer dynamics) may play roles. The European Innovation Partnership (EIP-AGRI) has highlighted the need for research on welfare’s impact on AMU.

Design and data sources: Retrospective observational study using farm-recorded data from specialized Italian beef fattening farms (Veneto region) covering January 2016 to April 2019. Animals were sourced mainly from France (Charolaise, Limousine, and others) and finished intensively in Italy.

Dataset: Initial dataset included 1,487 batches from 35 farms (87,902 animals). After exclusions (missing welfare/biosecurity or body weights; breeds with <20 batches; single-breed farms), the final dataset comprised 1,294 batches from 27 farms across seven breed groups (Charolaise, Blonde d’Aquitaine, Limousine, Salers, Italian crossbred, French crossbred, Irish crossbred). A total of 175,124 parenteral treatments with antimicrobial-containing veterinary medicinal products (VMP; n=33) were recorded.

AMU quantification: Treatment incidence per 100 animal-days (TI100) was calculated using defined daily doses for animals. Two indices were computed per batch:

• TI100it: based on Italian defined daily dose for animals (DDDAit) from ClassyFarm (www.classyfarm.it), using standard body weight 400 kg and days at risk 230 days. TI100it sums across VMPs: (mg active ingredient administered)/(DDDAit mg/kg/day × animals at risk × standard weight × days at risk) × 100.
• TI100vet: analogous, using EMA’s DDDvet; some actives lacked DDDvet and were excluded from TI100vet. Additionally, HPCIA TI100it and HPCIA TI100vet were calculated using only Highest Priority Critically Important Antimicrobials.

Welfare, biosecurity, and emergency management assessment: Annual on-farm assessments used a modified ClassyFarm-based protocol (56 items) covering:

• Total welfare (items 1–42), subdivided into Area A (management and staff training; items 1–13), Area B (housing; items 14–29), and Area C (animal-based measures; items 30–42).
• Biosecurity (items 43–52; mainly external biosecurity: visitor control, quarantine, rodent control, lorry cleaning).
• Emergency management (items 53–56: fire/ventilation alarms, noise risk, drinking water source). Six trained veterinarians scored items (1=high risk/poor status to 3=low risk/good status) with item-specific weights. Section/area scores were summed, converted to percentage 0–100, and categorized: low (<60%), medium (60–80%), high (>80%). Total welfare combined Areas A and B (50%) and Area C (50%) per protocol. Farm-level scores were matched to farm batches for analysis.

Statistical analysis: SAS 9.4 was used. Distributions were checked (Shapiro-Wilk, skewness/kurtosis) and normal plots inspected. Pearson correlations quantified relationships among total welfare, biosecurity, emergency management, and among Areas A–C. ANOVA (GLM) tested differences among years and breeds for section scores and among categories for performance traits and fattening length. TI100 indices were non-normal; thus, generalized linear mixed models (GLIMMIX) with gamma distribution and log link tested effects of welfare and biosecurity on TI100 indices. Fixed effects: sex, season, total welfare (medium, high), biosecurity (low, medium), emergency management (low, high), and welfare-by-sex interaction; random effects: farm, breed. A second model replaced total welfare with Areas A (low, medium, high), B (low, medium), and C (medium, high). Least squares means (LSM) ± SEM with Tukey-Kramer adjustment were reported. Significance at P<0.05; trends 0.05<P<0.10. Ethics approval was obtained (University of Padova, no. 74/2018), and procedures complied with Italian and EU directives.

• Section scores (per batch, mean ± SD): Total welfare 76% ± 5 (range 66–84), Biosecurity 24% ± 12 (range 9–66), Emergency management 39% ± 20 (range 14–83). Areas of total welfare: Area A 74% ± 6 (59–86), Area B 56% ± 8 (40–75), Area C 87% ± 9 (63–100).
• Correlations: Total welfare positively correlated with biosecurity (Rho=0.31, P<0.001). Correlations were low between total welfare and emergency management (Rho=−0.17) and between biosecurity and emergency management (Rho=0.20).
• Variation by year and breed: Total welfare, biosecurity, emergency management, and Areas A–C differed among years (P<0.001) and generally among breeds (P≤0.01; biosecurity P=0.079 trend).
• Association with AMU (GLIMMIX): Higher total welfare (>80%) was associated with significantly lower AMU compared with medium (60–80%): • TI100it: 1.25 vs 1.61 (P=0.008) • TI100vet: 1.02 vs 1.28 (P=0.021) • HPCIA TI100it and HPCIA TI100vet: similar numerical trend but not significant (P>0.05).
• No significant differences in AMU were detected across biosecurity (low vs medium) or emergency management (low vs high) categories.
• Areas A–C analyzed separately did not show significant differences for TI100it or TI100vet (P>0.05); however, HPCIA TI100vet was lower with higher Area A (P=0.034).
• Breed effects: Breeds tended to differ for all TI100 indices (e.g., TI100it P=0.052; TI100vet P=0.061), indicating substantial variability in AMU among breeds.
• Performance descriptors: Batches in the high total welfare category had slightly lower ADG (1.37 vs 1.40 kg/d) and shorter fattening cycles (200.3 vs 205.1 days) than medium; descriptive data suggested lower percentages treated in high emergency management (30.9%) vs low (65.2%), though modeling did not show significant effects for emergency management.

The study confirms the hypothesis that improved animal welfare is associated with reduced antimicrobial use in intensive beef finishing systems. The significant reduction in TI100it and TI100vet with higher total welfare likely reflects better overall health status, consistent with the inclusion of animal-based health indicators (e.g., lameness, lesions, respiratory disease) in the welfare assessment. Analyses of welfare sub-areas (management/staff training, housing, animal-based measures) suggested that significant effects emerge when welfare is considered comprehensively rather than via single subdomains, aligning with the multifactorial nature of welfare and analogous findings in biosecurity research. Biosecurity and emergency management did not show significant associations with AMU in this dataset. The authors attribute this mainly to low and skewed scores (few or no farms in high categories, limited medium category), potential category imbalance, and the focus on external biosecurity items. Literature generally supports that robust biosecurity, especially internal biosecurity (e.g., isolation of sick animals, dedicated equipment/boots for hospital pens), can reduce disease transmission and AMU. Therefore, combining internal with external biosecurity metrics may better capture effects on AMU in future studies. Breed-related differences in AMU and in welfare scores were observed, but high welfare scores did not uniformly translate to lower AMU across breeds, suggesting that the effectiveness of specific welfare measures may be breed-dependent (e.g., flooring solutions benefiting Limousine but not heavier Charolaise). This underscores the need for tailored welfare and management interventions by genotype. Although high welfare was associated with slightly lower ADG and body weights, welfare encompasses more than growth, and improved welfare can still yield net benefits through reduced AMU and potentially lower treatment costs. Overall, the findings support prioritizing comprehensive welfare improvements as a cornerstone of antimicrobial stewardship in beef systems, while highlighting the need to elevate and accurately measure biosecurity to detect its impact on AMU.

There was wide between-farm variation in biosecurity and emergency management, with average scores low (24% and 39%, respectively), while total welfare averaged high (76%). Higher total welfare was significantly associated with lower antimicrobial use (TI100it and TI100vet). No significant effects of biosecurity or emergency management on AMU were detected, likely due to skewed, generally low scores and limited representation of higher categories. These benchmarks can guide farm-level improvements and stewardship programs. The authors recommend targeted strategies to raise both external and internal biosecurity and to implement tailored welfare programs suited to each farm and breed. Future longitudinal, controlled studies are needed to validate these associations, incorporate comprehensive biosecurity (internal + external), and control for confounding factors.

• Retrospective observational design with potential missing values and lack of controlled allocation; data were compiled from multiple sources.
• Measurement mismatch: AMU data at batch level versus welfare/biosecurity assessments at farm level may dilute associations.
• Biosecurity assessment focused on external measures; internal biosecurity was not captured, possibly underestimating biosecurity’s impact on AMU.
• Skewed category distribution: very few or no farms in the high categories for biosecurity and emergency management; limited medium category representation, introducing bias and reducing power to detect effects.
• Some DDDvet values were unavailable for certain active ingredients, leading to exclusions in TI100vet calculations.
• Potential confounding by breed, season, and management not fully resolvable despite model adjustments.