Antimicrobial Resistance Surveillance in Wildlife: Methods and Gaps

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-08-11

Key Takeaways

• Wildlife AMR surveillance fundamentally differs from livestock or human surveillance by focusing on how resistance moves through ecological systems and persists outside clinical settings, rather than how antimicrobial use drives resistance within a treated population.
• Effective wildlife AMR surveillance requires deliberate species selection based on habitat overlap with human/agricultural activity, mobility, and sampling feasibility, alongside careful consideration of specimen types (feces, swabs, tissues) which bias toward different bacterial taxa.
• Data interpretation must acknowledge that absence of detection is not absence of resistance, necessitating reporting of sample size and detection limits, and that findings are most informative when interpreted within an ecological gradient model comparing contaminated and reference sites.
• Key limitations include the lack of harmonized international standards for wildlife AMR surveillance, unlike WOAH standards for domestic animals, and the need to adapt frameworks from other sectors while acknowledging inherent wildlife sampling challenges.
• Laboratory methods should integrate culture-based techniques for phenotypic testing and whole-genome sequencing for genotypic resolution, with careful selection of target bacteria (e.g., E. coli, Enterococcus spp. as indicators, or clinically relevant pathogens) and appropriate interpretive criteria (human vs. veterinary breakpoints).
• Common failure modes include silent selection bias from convenience sampling, specimen degradation due to poor cold chain management, and cross-contamination, all of which require rigorous checks and corrective actions to ensure data integrity.

Wildlife occupies a distinctive position in the epidemiology of antimicrobial resistance (AMR). Free-ranging animals move across landscapes, encounter human and domestic animal waste, and carry bacterial communities that reflect environmental contamination. Surveillance in wildlife therefore answers a different question than surveillance in livestock or humans: not how antimicrobial use drives resistance within a treated population, but how resistance moves through ecological systems and persists outside clinical settings. This article provides a methodological reference for veterinary researchers designing or evaluating wildlife AMR surveillance programs. It covers sampling design, specimen selection, laboratory approaches, data interpretation, and the structural gaps that currently limit the utility of wildlife-derived resistance data.

The reader is assumed to be familiar with antimicrobial susceptibility testing, resistance mechanisms, and basic epidemiological principles. The focus is exclusively on free-ranging wildlife and, where relevant, semi-free-ranging captive wildlife that interface with wild populations. Livestock and human surveillance are discussed only as sources of contamination or as comparative frameworks. The article does not address clinical treatment of wildlife patients, nor does it review antimicrobial stewardship in production animal systems.

Wildlife AMR surveillance is best understood as an ecological monitoring activity instead of a clinical diagnostic service. Its outputs inform risk assessment for zoonotic transmission, track environmental contamination gradients, and identify wildlife species that may serve as sentinels for resistance emerging in human or agricultural sectors. The One Health framework endorsed by the World Health Organization explicitly includes wildlife in the collaborative effort to address AMR, and the Centers for Disease Control and Prevention's zoonotic disease resources similarly recognize wildlife as a component of cross-sector surveillance. The scientific rationale rests on a simple premise: resistant bacteria and resistance genes do not respect species boundaries, and wildlife can acquire, carry, and disseminate both.

At a Glance

Conceptual Foundations of Wildlife AMR Surveillance

What Wildlife Surveillance Measures

Wildlife AMR surveillance detects three related but distinct phenomena. The first is colonisation: a wild animal carries resistant bacteria in its gastrointestinal tract or on mucosal surfaces without clinical disease. The second is infection: resistant bacteria cause disease in the animal itself. The third is environmental carriage: the animal serves as a vehicle for resistant bacteria or resistance genes acquired from contaminated water, soil, or feed. Most wildlife surveillance programs target colonisation, because it is more common than clinical infection and can be detected through non-invasive sampling.

The ecological meaning of a positive finding depends on the bacterial species and the resistance phenotype. Escherichia coli and Enterococcus species are the most frequently used indicator organizms because they are commensal gut bacteria, readily cultured, and frequently carry transferable resistance determinants. Detection of extended-spectrum beta-lactamase-producing E. coli in a migratory bird suggests recent acquisition from a contaminated environment. Detection of the same organizm in a burrowing rodent that never leaves a protected reserve suggests either local environmental persistence or carriage acquired before the animal's isolation. The global burden of AMR, as reviewed in the context of developing countries, demonstrates that resistance reservoirs exist in healthy human and animal populations, and wildlife surveillance must account for the possibility that wild animals are passive reflectors of environmental contamination instead of independent reservoirs.

The Ecological Gradient Model

A useful organizing concept is the ecological gradient. Wildlife populations can be arrayed along a continuum of anthropogenic exposure. At one end are synanthropic species such as urban gulls, rats, and raccoons that feed on human refuse and wade through sewage outfalls. At the other end are species in remote protected areas with minimal direct human contact. Surveillance design should place target species along this gradient deliberately, because the interpretation of resistance detection changes with position. Detection in a synanthropic species confirms environmental contamination. Detection in a remote species suggests either long-range transmission, often through migratory birds, or persistence of resistance determinants in environmental reservoirs.

This gradient logic also informs the selection of control populations. A well-designed study compares resistance prevalence in wildlife from contaminated sites with matched species from reference sites. Without such comparison, a single prevalence figure from an urban wildlife population cannot be distinguished from background environmental resistance. The review of antimicrobial resistance as a One Health problem emphasizes that resistant bacteria and resistance determinants spread within and between human, animal, and environmental sectors, which means wildlife surveillance data are most informative when interpreted as part of a multi-sector dataset instead of in isolation.

Study Design Logic

Defining the Surveillance Objective

The first design decision is the surveillance objective, and it determines nearly everything that follows. Three objectives are common. The first is trend monitoring: repeated sampling of the same wildlife population over time to detect changes in resistance prevalence. The second is outbreak or hotspot detection: targeted sampling of wildlife near a known contamination source, such as a hospital wastewater outfall or a concentrated animal feeding operation. The third is sentinel surveillance: using wildlife as an early warning system for resistance that may later appear in humans or domestic animals. Each objective implies different sampling frequency, sample size, and statistical analysis.

Trend monitoring requires standardized, repeatable methods and a sampling frame that can be reproduced across seasons and years. Hotspot detection requires spatial sampling around the suspected source, with distance gradients built into the design. Sentinel surveillance requires selection of wildlife species with documented exposure pathways to humans or domestic animals, and it requires bacterial targets that are clinically relevant to the species at risk. The global trends in antimicrobial use in aquaculture illustrate the value of this logic in another sector: the authors estimated use intensity by species group and production system specifically to target future surveillance efforts, an approach that wildlife programs can mirror by identifying which species and habitats carry the highest exposure risk.

Sampling Frame and Species Selection

Wildlife species differ dramatically in their utility as AMR sentinels. The ideal sentinel species has a large home range that overlaps with human activity, is easy to sample non-invasively, and has a gut microbiome that supports the bacterial targets of interest. Gulls and other Laridae meet most of these criteria and are among the most frequently sampled wildlife taxa in AMR studies. Rodents are useful for localized spatial analysis because their small home ranges allow resistance findings to be attributed to a specific site. Large carnivores occupy high trophic positions and may accumulate resistance through prey consumption, but they are difficult to sample in sufficient numbers.

Sampling method introduces its own biases. Fecal sampling from the environment is non-invasive and permits large sample sizes, but the species origin of each sample must be confirmed, and environmental degradation of the sample can reduce bacterial viability. Trapping and swabbing live animals provides species certainty and allows repeated sampling of marked individuals, but capture stress, trap avoidance, and permit requirements limit sample size. Necropsy sampling of animals found dead or culled provides access to multiple tissue sites, including the lower gastrointestinal tract, but the sample is biased toward diseased or otherwise compromised animals. The MSD Veterinary Manual provides species-specific guidance on handling and specimen collection that applies to wildlife presented for examination, and the American Veterinary Medical Association's practice resources include professional standards relevant to the ethical conduct of wildlife sampling.

Sample Size and Detection Limits

Wildlife AMR surveillance is chronically underpowered. A study that collects 30 fecal samples from a gull colony and detects no resistant E. coli cannot conclude that resistance is absent, the 95% confidence interval for a zero numerator out of 30 is approximately 0% to 12%. Reporting detection limits is therefore mandatory. The minimum detectable prevalence for a given sample size should be calculated before fieldwork begins, and the calculation should be stated in the methods. For rare resistance phenotypes, targeted enrichment culture is often necessary, because direct plating of fecal samples may fail to detect resistant clones present at low density.

The WHO One Health initiative and the WOAH terrestrial animal health standards both recognize that surveillance systems must be designed with explicit objectives and statistical rigour, even though neither organization currently provides species-specific guidance for wildlife AMR surveillance. Researchers must therefore adapt frameworks developed for domestic animal surveillance, while acknowledging that wildlife sampling rarely achieves the same level of control over population size, detection probability, and sampling effort.

Specimen Collection and Handling

Sample type determines which resistance questions can be answered. Fecal samples are the standard for enteric bacteria such as Escherichia coli, Salmonella spp., and Campylobacter spp. They are non-invasive, collectable from live animals, and suitable for repeated sampling of marked individuals. Cloacal swabs serve the same purpose in birds and reptiles. Oropharyngeal or nasal swabs target respiratory pathogens and colonising species such as staphylococci. Skin swabs and fur or feather samples detect environmental contamination and surface colonisation. Tissue samples from dead animals permit culture of invasive pathogens and paired histopathology, but they represent a single time point and cannot be repeated.

Freshness is the dominant pre-analytical variable. Fecal samples degrade rapidly, and overgrowth by environmental bacteria can obscure target organizms. Samples should be collected within hours of defecation, placed in transport medium, and refrigerated or frozen according to the target organizm. Campylobacter spp. require microaerophilic conditions and transport media such as Cary-Blair. E. coli tolerates standard transport buffers. For wildlife sampled in the field, a cold chain may be unavailable, and the choice of transport medium must anticipate delays of 24 to 72 hours.

Pooling samples reduces cost but sacrifices individual-level data. Pooled fecal samples from a social group estimate group-level carriage prevalence but cannot identify individual shedders or measure within-group transmission. Individual samples are required when the objective is to track resistance dynamics in marked animals or to correlate resistance carriage with individual health outcomes. The MSD Veterinary Manual provides species-specific guidance on sample collection and handling that should be consulted before fieldwork begins.

Laboratory Methods and Target Selection

Culture-based methods remain the backbone of wildlife AMR surveillance because they yield viable isolates suitable for phenotypic testing and whole-genome sequencing. Selective media containing antimicrobials at breakpoint concentrations allow direct detection of resistant subpopulations, but the choice of antimicrobial and concentration must match the surveillance objective. Screening for extended-spectrum beta-lactamase (ESBL) producers requires media supplemented with cefotaxime or ceftazidime. Screening for carbapenemase producers requires meropenem-supplemented media. These approaches detect resistant organizms regardless of species, which is useful for sentinel surveillance but provides no information on the susceptible background population.

Quantitative culture adds a dimension that qualitative methods miss. Most probable number (MPN) techniques estimate the density of resistant bacteria per gram of feces, which is a more sensitive indicator of selection pressure than binary presence or absence. The trade-off is labor and cost. For large-scale screening, direct plating on selective media followed by confirmation of a subset of isolates by matrix-assisted laser desorption ionisation time-of-flight mass spectrometry (MALDI-TOF MS) is a pragmatic workflow.

Phenotypic susceptibility testing follows standardized protocols such as those of the Clinical and Laboratory Standards Institute (CLSI) or the European Committee on Antimicrobial Susceptibility Testing (EUCAST). Wildlife isolates are tested against human and veterinary breakpoints, and the choice matters. A wild bird isolate of E. coli inhibited by ciprofloxacin at 0.5 mg/L is susceptible by human CLSI breakpoints but resistant by veterinary breakpoints for some species. Reporting both interpretations avoids misclassification and supports cross-sectoral comparison. The One Health perspective on antimicrobial resistance emphasizes that resistance determinants move between human, domestic animal, and wildlife compartments, so harmonised interpretive criteria are a prerequisite for meaningful comparison.

Genotypic methods add resolution. Whole-genome sequencing identifies resistance genes, mobile genetic elements, and strain relationships. It distinguishes acquired resistance genes from chromosomal mutations and can infer the likely origin of a resistance determinant through phylogenetic comparison. Sequencing is not a substitute for phenotypic testing, because a resistance gene may be present but not expressed. The two approaches are complementary, and the choice depends on the question. Phenotypic testing answers the clinical question, what will this isolate resist. Genotypic testing answers the ecological question, what resistance determinants are circulating and where did they come from.

Sampling Strategy Selection

The table below links surveillance objectives to recommended sampling strategies and target bacteria.

The correct strategy changes with the species. Colonial roosting birds such as gulls concentrate feces at predictable sites, making composite sampling efficient. Solitary carnivores such as lynx defecate at low density, and scat detection dogs or GPS-collared individuals may be required to obtain adequate sample sizes. For aquatic species, water samples filtered and cultured for indicator organizms can supplement direct animal sampling, but they cannot attribute resistance to a particular species.

Data Interpretation and Bias

Wildlife surveillance data carry biases that domestic animal surveillance does not. Detectability varies by species, habitat, and season. A fecal sample from a forest floor may be hours or weeks old, and the bacterial community changes with age. Samples collected from road-killed animals over-represent species that cross roads and individuals that are behaviorally predisposed to vehicle collisions. Samples from rehabilitation centers over-represent sick, injured, or juvenile animals. Each of these biases must be stated in the methods and considered in the interpretation.

Prevalence estimates require confidence intervals that account for clustering. Animals from the same social group are not independent, and treating them as such inflates precision. Hierarchical models that include group as a random effect are appropriate when social structure is known. When it is not known, the analysis should acknowledge the potential for clustering and its effect on variance.

The absence of a resistant isolate does not establish the absence of resistance. Detection limits are a function of sample size, culture method, and the prevalence of the target organizm. A negative result from 20 fecal samples can exclude a carriage prevalence of 14% with 95% confidence, but it cannot exclude a prevalence of 5%. Reporting detection limits alongside results is standard practice in international animal health surveillance standards and should be applied to wildlife AMR work.

Integration with One Health Surveillance

Wildlife AMR data gain value when interpreted alongside human and domestic animal data. The WHO One Health initiative and the CDC One Health resources both frame antimicrobial resistance as a cross-sectoral problem requiring coordinated surveillance. In practice, this means harmonising sampling protocols, laboratory methods, and interpretive criteria across sectors. A wild bird isolate tested by the same protocol as a poultry isolate can be compared directly. Isolates tested by different protocols cannot.

Data sharing is the limiting step. Wildlife agencies, veterinary diagnostic laboratories, public health laboratories, and academic groups each hold data in different formats with different metadata standards. A minimum metadata set should include host species, sampling date, geographic coordinates, sample type, animal age and sex, and the reason for sampling. Without these fields, isolates cannot be placed in an ecological context, and their value for One Health analysis is substantially reduced.

The global problem of antimicrobial resistance in developing countries is particularly relevant to wildlife surveillance, because many regions with high biodiversity and high antimicrobial use in humans and livestock have the least laboratory capacity. Wildlife surveillance in these settings may need to rely on simplified protocols, selective media, and transport of isolates to reference laboratories. These constraints should be anticipated in study design instead of treated as post hoc limitations.

Recognized Complications and Failure Modes

Wildlife AMR surveillance programs fail in characteriztic patterns. The most common is silent selection bias, where sampling convenience replaces ecological reasoning. A program that samples birds at a single landfill site will detect resistance phenotypes associated with human waste streams, not the background resistance profile of the regional avifauna. The discriminating check is to compare the sampling frame against the stated objective: if the objective references a species or habitat and the sampling plan does not, the design is already compromised.

Specimen degradation is the second major failure mode. Wildlife samples travel unpredictable distances and face variable ambient temperatures before reaching the laboratory. Delayed culture reduces recovery of fastidious organizms and allows overgrowth of environmental contaminants. Detection of this failure requires monitoring the interval from collection to plating and tracking the proportion of samples yielding no growth or mixed cultures. A rising no-growth rate across a season usually indicates a cold-chain problem, not a true absence of target bacteria.

Cross-contamination between samples produces false-positive resistance findings. This occurs when sampling equipment is reused without adequate decontamination or when multiple carcasses are transported in shared containers. The discriminating check is the resistance phenotype itself: unexpected patterns, such as identical multilocus sequence types across unrelated host species sampled at different sites, should trigger an audit of field protocols before the data are interpreted as ecology.

The table below summarizes these failure modes and their early detection.

Common Errors and Corrective Action

Less experienced investigators frequently confuse detection of a resistance gene with evidence of clinical resistance. A bacterial isolate carrying a beta-lactamase gene may still be susceptible to the drug if the gene is not expressed or if the enzyme does not hydrolyse the specific compound. The corrective action is to pair genotypic detection with phenotypic susceptibility testing and to report both results separately.

A second recurring error is the assumption that wildlife isolates are independent. Animals from the same social group, den, or migratory flock share bacteria and resistance determinants. Treating them as independent observations inflates precision and produces confidence intervals that are falsely narrow. The corrective action is to include a clustering variable in the analysis, such as capture site or social group, and to use mixed-effects models or generalized estimating equations.

A third error involves extrapolating from sentinel species to the broader ecosystem. Gulls and other synanthropic birds are valuable indicators of environmental contamination, but their resistance carriage reflects their foraging ecology, not the resistance status of all wildlife. The corrective action is to state the sentinel function explicitly in the interpretation and to avoid claims about species that were not sampled.

Limitations of the Current Evidence

The evidence base for wildlife AMR surveillance remains uneven. Most published data come from high-income countries and from a narrow range of taxa, particularly gulls, waterfowl, and urban rodents. Data from tropical regions, where antimicrobial use in human medicine and agriculture is often less regulated, are sparse. The global problem of antimicrobial resistance is particularly pressing in developing countries, where the infectious disease burden is high and cost constraints limit the application of newer agents, and reservoirs for resistance may be present in healthy human and animal populations (Okeke et al., 2005). Wildlife surveillance in these settings is correspondingly scarce.

Expert opinion differs on the value of routine wildlife surveillance as an early-warning system. Some argue that wildlife data provide an integrated measure of environmental contamination that precedes clinical detection in humans or domestic animals. Others contend that the lag between environmental contamination and detectable wildlife carriage is too long and too variable to be operationally useful. The evidence does not yet resolve this dispute, and program designers should be explicit about which rationale they are adopting.

A further limitation is the absence of standardized interpretive criteria for wildlife isolates. Clinical breakpoints derived from human or domestic animal pharmacokinetic data may not apply to bacteria from free-ranging animals, where the selective pressure is environmental instead of therapeutic. Reporting minimum inhibitory concentrations as raw distributions, instead of as susceptible or resistant categories, preserves information and allows future re-interpretation as standards evolve.

Referral, Consultation, and Reporting

Referral to a specialist laboratory is warranted when a surveillance program detects a resistance phenotype of public health significance, such as carbapenem resistance, vancomycin-resistant enterococci, or methicillin-resistant staphylococci in a species with zoonotic potential. The role of companion animals in the dissemination of antimicrobial resistance has historically received less attention than food animals, but species with a potential for zoonotic transmission and resistance phenotypes of clinical interest have been documented in pet populations (Guardabassi et al., 2004). Wildlife findings of the same phenotypes warrant confirmatory testing and notification.

Regulatory reporting obligations vary by jurisdiction and by the organizm detected. International standards for animal health surveillance and trade-related disease control are set by the World Organization for Animal Health, and national veterinary authorities should be consulted when a notifiable pathogen is identified (WOAH terrestrial animal health standards). The One Health framework links human, animal, and environmental health for zoonotic disease and antimicrobial resistance control, and wildlife surveillance findings should be shared across sectors when they indicate a shared risk (WHO One Health initiative).

Consultation with a veterinary epidemiologist is advisable before the surveillance design is finalised, not after data collection begins. Sample size calculations, clustering adjustments, and bias mitigation are far easier to implement at the design stage. Laboratory involvement should begin equally early, to confirm that the proposed specimen types and transport conditions are compatible with the target organizms and that the laboratory can provide the required phenotypic and genotypic capacity.

Frequently Asked Questions

How Much Does Wildlife AMR Surveillance Cost, and How Can We Prioritize When Funding Is Limited?

Costs scale with sample size, laboratory methods, and the number of species sampled. Culture-based surveillance of a single indicator organizm, such as Escherichia coli, is the least expensive option. Whole-genome sequencing and metagenomic approaches cost substantially more per sample. When funding is constrained, prioritize a narrow objective, for example detecting third-generation cephalosporin resistance in gulls at a landfill site, and sample fewer species with adequate numbers per species. Opportunistic sampling of animals already handled for other purposes reduces capture costs. Archived diagnostic specimens can provide retrospective data. The WHO One Health Initiative and CDC One Health resources describe how wildlife surveillance can be embedded within existing national programs to share laboratory infrastructure and data platforms.

What Should We Do When Ideal Equipment for Specimen Collection Is Not Available?

The essential requirement is a sterile swab or fecal sample that reaches the laboratory without contamination or excessive delay. When transport medium is unavailable, use sterile containers and keep samples cool, not frozen, if processing will occur within 24 to 48 hours. Freezing at minus 20 degrees Celsius is acceptable for culture-based work when delays exceed two days, but repeated freeze-thaw cycles reduce recovery of some organizms. For carcass sampling, use sterile instruments for each animal and collect from the intestinal lumen or target organ before decomposition advances. If selective media are unavailable, request that the laboratory plate samples on MacConkey agar for Gram-negative screening. The MSD Veterinary Manual provides species-specific guidance on specimen handling that applies to wildlife samples.

How Does Surveillance Design Change When Working With Endangered or Protected Species?

Non-invasive sampling becomes the primary option. Fresh fecal samples collected from the environment avoid capture stress and do not require handling permits. Scat age affects bacterial recovery, so prioritize samples with visible moisture and intact structure. When capture is justified for other purposes, combine AMR sampling with routine health assessments to minimize additional handling events. Sample size targets may be unattainable for small populations, so interpret results as presence or absence data instead of prevalence estimates. Archived samples from rehabilitation centers and mortality investigations provide an alternative source. The WOAH terrestrial animal health standards outline how surveillance in protected species must balance disease monitoring with conservation obligations.

What Records Must We Keep for a Wildlife AMR Surveillance Program to Be Scientifically Useful?

Each sample requires a unique identifier linked to species, age class, sex, capture location with coordinates, date, sampling method, and any observed health status. Record the specimen type, transport conditions, and time from collection to processing. Laboratory records must include the isolation method, media used, and antimicrobial susceptibility testing panel with the interpretive standard applied. Store isolates in a viable collection with linked metadata. Field notes should document environmental context, such as proximity to agricultural operations, wastewater outfalls, or urban areas, because these factors influence interpretation. The AVMA practice resources emphasize that complete records support both data validation and future re-analysis when interpretive criteria change.

How Do We Explain Wildlife AMR Surveillance Findings to a Client or Agency Supervisor?

Frame the findings in terms of ecosystem health and human health risk, not as a wildlife disease outbreak. Explain that resistant bacteria in wildlife usually reflect environmental contamination instead of illness in the animals themselves. Use the One Health framework to show how wildlife samples serve as sentinels for resistance circulating in the shared environment. Present prevalence data with confidence intervals and avoid overinterpreting single positive samples. If clinically relevant resistance is detected, recommend follow-up sampling to confirm the finding and identify potential sources. The CDC One Health and zoonotic disease resources provide language for communicating cross-species resistance risks to non-specialist audiences.

When Is It Appropriate to Publish or Formally Report Wildlife AMR Surveillance Data?

Report data when the sampling design supports the claims being made. A convenience sample from a single season supports a descriptive report of resistance detection but not a prevalence estimate. Formal reporting is appropriate when sample sizes meet calculated detection limits, laboratory methods follow recognized standards, and the data fill a documented geographic or taxonomic gap. Negative results are publishable when the sampling effort was sufficient to detect resistance at a meaningful prevalence. Before publication, confirm that sampling and testing complied with relevant permits and animal welfare standards. The WOAH terrestrial animal health standards describe reporting obligations for notifiable diseases, while routine AMR findings should be shared through national surveillance networks and peer-reviewed channels.

Related Clinical & Scientific Guides

• Wildlife Disease Surveillance: Designing and Implementing a One Health Program
• Biosecurity Risk Assessment for Livestock Operations: A Practical Framework
• Rabies Post-Exposure Prophylaxis in Veterinary Personnel

References and Further Reading

• Antimicrobial resistance in developing countries. Part I: recent trends and current status.. 2005.
• Pet animals as reservoirs of antimicrobial-resistant bacteria.. 2004.
• Global trends in antimicrobial use in aquaculture.. 2020.
• Antimicrobial Resistance: A Growing Serious Threat for Global Public Health.. 2023.
• Antimicrobial Resistance: a One Health Perspective.. 2018.
• Antimicrobial resistance: A global emerging threat to public health systems.. 2017.
• WHO One Health Initiative. WHO.
• CDC One Health and Zoonotic Disease Resources. CDC.
• MSD Veterinary Manual, Professional Edition. MSD Veterinary Manual.

Related Articles

• Antimicrobial Resistance Surveillance in Food Animals: Sampling Strategies and Data Interpretation
• Antimicrobial Resistance in Wildlife: Environmental Reservoir and Public Health Risk
• One Health Approach to Antimicrobial Resistance: A Veterinary Perspective
• Zoonotic Disease Surveillance in Animal Shelters: Methods and Challenges
• Rabies Control in Endemic Regions: Vaccination Strategies and Surveillance Gaps

This article is educational professional reference material for veterinary audiences. It is not a substitute for veterinary diagnosis, individual clinical judgment, current product labeling, or applicable regulatory requirements.