Antimicrobial resistance (AMR) occurs when microorganisms withstand medicines that would ordinarily inhibit their growth or kill them. It affects bacteria, fungi, viruses, and parasites, and therefore encompasses resistance to antibacterial, antifungal, antiviral, and antiparasitic drugs. Antibiotic resistance is the bacterial component of AMR, although the two terms are sometimes used interchangeably. Resistance is a property of the microorganism or microbial population, not of the person or animal receiving treatment.
Resistance arises through biological evolution, but its emergence and spread are strongly influenced by antimicrobial use, infection transmission, sanitation, agricultural practices, and access to effective healthcare. Its consequences extend beyond otherwise treatable infections: reliable antimicrobial treatment also supports surgery, cancer treatment, intensive care, and organ transplantation. A major global analysis estimated that bacterial AMR directly accounted for approximately 1.27 million deaths in 2019 and was associated with approximately 4.95 million deaths.
AMR is simultaneously a molecular phenomenon, a clinical problem, and a challenge for public health. Understanding it requires distinguishing different resistance mechanisms, tracing how organisms and resistance genes circulate, interpreting laboratory measurements, and evaluating interventions across human, animal, and environmental systems.
Scope, terminology, and biological boundaries
Antimicrobials act on particular biological processes. Antibacterial drugs may inhibit cell-wall synthesis, protein production, nucleic-acid metabolism, or membrane function. Antifungals target structures such as fungal membranes and cell walls; antivirals interfere with stages of viral replication; antiparasitic agents act on diverse metabolic and cellular processes. Because these targets differ, resistance is not a single biochemical condition. It is a family of outcomes in which a drug’s activity becomes insufficient against a particular organism under defined conditions.
An important distinction is between intrinsic and acquired resistance. Intrinsic resistance is characteristic of an organism because of its normal structure or physiology. For example, bacteria lacking a conventional peptidoglycan cell wall are intrinsically unaffected by drugs whose action requires that structure. Acquired resistance appears in a previously susceptible lineage through genetic change or acquisition of genetic material. The distinction is clinically useful, although natural populations contain substantial variation and the boundary can become less clear as new mechanisms are characterized.
Resistance also differs from tolerance and persistence. Resistance generally increases the concentration required to prevent microbial growth. Tolerance describes slower killing without necessarily changing that concentration, while persistence commonly refers to a small subpopulation that survives exposure through a reversible physiological state. A persistent infection is not automatically an infection caused by a resistant organism: inaccessible infection sites, foreign material, impaired host defenses, or inadequate drug exposure can also prevent clearance.
The terms multidrug-resistant, extensively drug-resistant, and pandrug-resistant describe increasingly restricted susceptibility profiles, but their formal definitions depend on the organism and applicable framework. Widely used definitions for certain bacteria classify multidrug resistance as acquired nonsusceptibility to at least one agent in three or more antimicrobial categories. These definitions should not be transferred uncritically to viruses, fungi, or tuberculosis, which have their own classifications.
Colonization must also be distinguished from infection. A person may carry a resistant organism in the intestine, nose, or skin without illness, yet contribute to its transmission or later develop an infection from that reservoir. Conversely, detecting a resistance gene in an environmental sample does not establish that it is present in a viable human pathogen. These distinctions determine how laboratory findings, surveillance counts, and disease-burden estimates are interpreted.
Historical development and recognition
Resistance predates modern medicine. Microorganisms have long produced antimicrobial compounds and evolved mechanisms for surviving exposure to competing organisms. Resistance determinants recovered from ancient environmental material demonstrate that parts of the microbial resistance repertoire are much older than industrial drug production. Modern antimicrobial use did not create this evolutionary capacity; it greatly increased the selective opportunities through which particular resistant organisms and genes could become common.
The antibiotic era transformed the practical significance of resistance. Alexander Fleming observed the antibacterial effect of the mold associated with penicillin in 1928, and penicillin became widely available during the 1940s. Penicillin-destroying activity was described by Edward Abraham and Ernst Chain in 1940, before the drug’s broad clinical deployment. In his 1945 Nobel lecture, Fleming warned that inadequate exposure could select resistant bacteria. Penicillin-resistant Staphylococcus aureus subsequently became widespread.
The introduction of successive drug classes was repeatedly followed by resistance. Streptomycin, introduced into clinical use in the 1940s, initially produced striking improvements in tuberculosis, but resistance frequently emerged during single-drug treatment. Combination treatment became central to tuberculosis control. Methicillin was introduced in 1959, and methicillin-resistant S. aureus was reported in 1961. Later decades brought increasing resistance to extended-spectrum cephalosporins, fluoroquinolones, vancomycin, and carbapenems in different bacterial groups.
Research also changed the understanding of how resistance spreads. Work on transferable resistance in enteric bacteria during the 1950s and 1960s established that resistance could move between organisms rather than arise independently in each lineage. The discovery of mobile genetic elements helped explain why resistance to several unrelated drugs could appear together and disseminate rapidly. Hospital outbreaks revealed the additional importance of successful bacterial clones, contaminated environments, and transmission through healthcare networks.
International recognition broadened as resistance became established across community, hospital, agricultural, and environmental settings. The World Health Assembly adopted a Global Action Plan on Antimicrobial Resistance in 2015, and the United Nations General Assembly held its first high-level meeting on AMR in 2016. These developments reframed resistance from a problem addressed mainly by discovering replacement drugs into one requiring sustained prevention, surveillance, access, stewardship, and coordinated policy.
Evolution, selection, and genetic exchange
Resistance evolves when inherited differences affect survival and reproduction under antimicrobial exposure. A resistant variant may already exist before treatment or arise during microbial replication. Exposure suppresses susceptible competitors, allowing the variant to increase in frequency. This is natural selection, not a purposeful adaptation by the organism. Drugs can also affect stress responses and mutation processes, but the central evolutionary principle is differential survival and reproduction among variants.
New resistance can originate through mutation in chromosomal DNA. A single nucleotide change may alter a drug-binding site, reduce activation of a drug, or increase expression of an efflux system. Other phenotypes require several changes accumulated over time. The probability that resistance emerges depends on microbial population size, mutation rates, the number of genetic steps required, and the pattern of exposure. Large populations and prolonged infection can supply many opportunities for resistant variants to arise.
Bacteria also acquire resistance through horizontal gene transfer. Conjugation transfers genetic material through cell-to-cell contact, often using plasmids. Transformation involves uptake of DNA from the environment, while transduction transfers DNA through bacterial viruses. Transposons and insertion sequences move genetic material within or between DNA molecules. Integrons capture and express gene cassettes, many of which encode resistance, although integrons generally depend on other elements for movement between cells.
These processes enable co-selection. A plasmid may carry resistance determinants for several antimicrobial classes, so exposure to one drug can preserve resistance to others. Genes conferring tolerance to certain metals or biocides may occur on the same mobile elements, creating additional routes of selection under particular conditions. Cross-resistance is different: a single mechanism, such as an efflux pump or modified target, reduces susceptibility to multiple agents.
Resistance may impose a fitness cost when the selecting drug is absent, but this does not ensure that resistance will disappear after use declines. Compensatory mutations can reduce that cost, and some determinants impose little detectable disadvantage in relevant environments. Continued transmission, repeated introduction, and selection by other agents can maintain resistance. Consequently, resistance trends cannot be predicted from antimicrobial consumption alone: evolutionary history, microbial ecology, and transmission networks also matter.
Molecular mechanisms of bacterial resistance
Bacterial resistance mechanisms can be grouped into drug destruction or modification, target alteration, reduced intracellular drug accumulation, and metabolic bypass. Individual organisms often combine several mechanisms. Their effects depend on the drug, bacterial species, expression level, and other features of the cell. A resistance gene therefore does not always produce the same susceptibility profile in every organism carrying it.
Drug-inactivating enzymes are especially important. Beta-lactamases hydrolyze the beta-lactam ring of beta-lactam antibiotics, preventing effective inhibition of bacterial cell-wall synthesis. Extended-spectrum beta-lactamases commonly affect third-generation cephalosporins, while carbapenemases can compromise carbapenems. Examples include KPC enzymes, NDM metallo-beta-lactamases, and OXA-48-like enzymes. Their substrate profiles differ, and inhibitors effective against one enzyme family may not inhibit another. Aminoglycoside-modifying enzymes provide a separate example, chemically modifying drugs so they bind less effectively to their targets.
Target alteration prevents otherwise active drugs from binding effectively. Methicillin resistance in S. aureus is usually mediated by mecA or, less commonly, mecC, which encode an alternative penicillin-binding protein with reduced affinity for many beta-lactams. Changes in DNA gyrase or topoisomerase IV can contribute to fluoroquinolone resistance. Ribosomal modifications can reduce susceptibility to macrolides, while changes in cell-wall precursor chemistry can produce vancomycin resistance in enterococci.
Reduced entry and increased export can lower intracellular drug concentrations. The outer membrane of Gram-negative bacteria is an important permeability barrier, and loss or alteration of porin channels can restrict drug entry. Efflux pumps actively export compounds from bacterial cells; some have broad substrate ranges and affect several drug classes. Reduced permeability combined with beta-lactamase production may produce clinically significant resistance even when neither mechanism alone would do so.
Other mechanisms protect a target, replace an inhibited pathway, or prevent conversion of a prodrug into its active form. The transferable mcr family, for example, encodes enzymes that modify lipid A and reduce susceptibility to colistin. Resistance to trimethoprim or sulfonamides may involve alternative enzymes in folate metabolism. Such diversity explains why a label such as “carbapenem-resistant” describes an observed phenotype but does not by itself identify the mechanism, transmission risk, or activity of every possible alternative treatment.
Major resistant bacterial pathogens
The clinical significance of bacterial resistance depends on both the resistance profile and the organism’s ability to cause disease. Resistant bloodstream infections, pneumonia, urinary tract infections, and surgical-site infections are major concerns because they are common, potentially severe, and frequently require prompt treatment. A globally important resistant organism may be uncommon in one region yet endemic in another, making local epidemiology essential to interpretation.
Among Gram-negative bacteria, Escherichia coli and Klebsiella pneumoniae are prominent causes of urinary, abdominal, and bloodstream infections. Extended-spectrum beta-lactamase production can compromise commonly used cephalosporins. Carbapenem-resistant Enterobacterales are particularly concerning because carbapenems have been important treatment options for serious infections caused by organisms resistant to other beta-lactams. Carbapenem resistance may arise through carbapenemases or combinations of reduced permeability and other beta-lactamases.
Acinetobacter baumannii and Pseudomonas aeruginosa cause important healthcare-associated infections, especially among critically ill patients. Their resistance can combine low permeability, efflux, target changes, and acquired enzymes. Environmental persistence is particularly important for A. baumannii, while P. aeruginosa is associated with moist environments and chronic infection in conditions such as cystic fibrosis. These organisms illustrate why resistance control involves environmental and clinical practices as well as drug selection.
Among Gram-positive bacteria, methicillin-resistant S. aureus causes skin infections, bloodstream infection, and other invasive disease in both healthcare and community settings. Vancomycin-resistant enterococci are especially important in hospitalized patients with substantial underlying illness or prolonged antimicrobial exposure. Resistant Streptococcus pneumoniae can complicate respiratory and invasive infections, although susceptibility interpretation depends on the drug and infection site. Neisseria gonorrhoeae has successively developed resistance to several historically effective therapies, making sustained susceptibility monitoring important.
Drug-resistant tuberculosis has distinct biology and treatment classifications. Multidrug-resistant tuberculosis is resistant to at least isoniazid and rifampicin. Under WHO definitions introduced in 2021, pre-extensively drug-resistant tuberculosis is multidrug- or rifampicin-resistant disease with additional resistance to any fluoroquinolone. Extensively drug-resistant tuberculosis additionally involves resistance to at least one of bedaquiline or linezolid. Resistance may emerge during inadequate effective treatment or be transmitted directly, so a resistant case does not establish that the affected person previously received inappropriate therapy.
Resistance in fungi, viruses, and parasites
Fungal resistance has become increasingly important as more people survive with intensive medical support or impaired immune defenses. Candida species can cause bloodstream and deep-tissue infections, while Aspergillus species cause invasive disease particularly in susceptible hosts. Resistance may be intrinsic to a species or acquired through changes in drug targets, increased efflux, altered sterol synthesis, or other mechanisms. Species identification is therefore central to interpreting antifungal susceptibility.
Candida auris is notable for healthcare-associated transmission, persistence in healthcare environments, and frequent resistance to one or more major antifungal classes. It was first described in 2009, although retrospective identification has found earlier isolates. Echinocandin resistance in Candida often involves changes in FKS genes associated with cell-wall glucan synthesis. In Aspergillus fumigatus, azole resistance can arise during treatment or be associated with environmental selection involving agricultural azole fungicides. The relative importance of these pathways varies geographically.
Viral resistance commonly results from changes in enzymes or other proteins targeted by an antiviral drug. In HIV infection, mutations affecting reverse transcriptase, protease, or integrase can reduce the activity of particular medicines. Combination antiretroviral therapy suppresses replication and reduces opportunities for resistance when the regimen remains active and drug exposure is sufficient. Resistance can nevertheless arise during treatment or be transmitted to someone who has never received antiretrovirals. Hepatitis B and influenza viruses provide other clinically important examples.
Parasites display equally diverse resistance mechanisms. In malaria, resistance in Plasmodium falciparum has affected chloroquine, sulfadoxine–pyrimethamine, and other agents. Artemisinin partial resistance is associated with delayed parasite clearance and particular mutations in the parasite’s kelch13 gene. It is not synonymous with complete failure of an artemisinin-based combination: outcomes also depend on the partner drug, parasite susceptibility, drug exposure, and host factors. Partial artemisinin resistance has been documented in Southeast Asia and parts of Africa.
Resistance also affects medicines used against other protozoa and parasitic worms, with patterns differing between human and veterinary medicine. These examples demonstrate why bacterial resistance data cannot represent the whole AMR problem. Surveillance methods, clinical definitions, available drug classes, and evolutionary pathways must be adapted to each organism group. At the same time, common principles—selection, transmission, reliable diagnosis, and effective combination strategies where appropriate—connect these otherwise distinct fields.
Transmission across people, animals, and environments
Resistance spreads through two related processes: transmission of resistant organisms and movement of resistance genes between organisms. A successful bacterial lineage may disseminate across hospitals, communities, and countries, while a mobile genetic element may move through several species. These processes can occur simultaneously. Identifying a shared resistance gene therefore does not necessarily establish that two infections came from the same immediate source.
In healthcare, transmission can occur through contaminated hands, medical devices, shared equipment, surfaces, and environmental reservoirs. Invasive devices can bypass normal barriers, while antimicrobial exposure can alter the microbiota and favor resistant colonization. Patient transfers connect wards, hospitals, rehabilitation facilities, and long-term care institutions. A resistant organism introduced into one facility may consequently become a regional problem rather than remain an isolated outbreak.
Community transmission involves households, sexual networks, food, water, and other pathways, depending on the organism. Travel and population movement can redistribute resistant strains, but transmission risk is shaped by exposure and infrastructure rather than nationality itself. Resistant intestinal bacteria may be carried without symptoms, and carriage can persist after an exposure has ended. Colonization can subsequently lead to infection when host defenses or anatomical barriers are disrupted.
Food-producing animals and aquaculture form additional antimicrobial-use and transmission systems. Antimicrobials can select resistant organisms in treated animals; bacteria and genes may then move through animal contact, food production, manure, water, or occupational exposure. The strength and direction of these connections vary by organism and setting. Detecting the same resistance determinant in livestock and humans does not alone quantify how much human disease is attributable to agricultural use.
The One Health approach examines these connections across human, animal, and ecosystem health. Wastewater, pharmaceutical manufacturing discharges, hospital effluent, and agricultural runoff can introduce antimicrobial residues and resistant organisms into environmental compartments. The resistome comprises the resistance genes present in a microbial community, including genes in organisms that rarely cause disease. Environmental detection is valuable for understanding reservoirs and pathways, but assessing health risk also requires information about gene mobility, viable hosts, exposure, and the ability of those hosts to infect humans.
Laboratory detection and susceptibility interpretation
Laboratory assessment links the biological phenomenon of resistance to clinical decision-making and surveillance. Conventional investigation generally involves collecting a suitable specimen, recovering and identifying the organism, and testing its response to selected drugs. The quality of each stage matters. Contamination, poor specimen handling, previous antimicrobial exposure, and failure to distinguish colonization from infection can all distort interpretation before susceptibility testing begins.
A central measurement is the minimum inhibitory concentration (MIC), the lowest tested concentration that prevents visible growth under specified conditions. Broth microdilution estimates MICs using serial drug concentrations. Disk diffusion measures the zone of growth inhibition around a drug-containing disk on agar. Automated systems and gradient methods provide other approaches, but performance varies by organism–drug combination, and some results require confirmation using reference methods.
Interpretive breakpoints translate measurements into categories relevant to treatment. Organizations including the Clinical and Laboratory Standards Institute and the European Committee on Antimicrobial Susceptibility Testing develop standards using susceptibility distributions, dosing, drug exposure, and clinical evidence. A breakpoint is not a universal biological constant: it can change when dosing assumptions or evidence change. Under current EUCAST terminology, “I” means “susceptible, increased exposure,” rather than simply representing an indeterminate middle category.
Molecular tests detect resistance-associated genes or mutations, sometimes directly from clinical samples or positive blood cultures. They can rapidly identify selected mechanisms, such as particular carbapenemases, but a negative result excludes only the targets actually tested. Conversely, detecting a gene does not always establish its expression, the identity of its host in a mixed sample, or the full susceptibility profile. Whole-genome sequencing can support resistance prediction and outbreak analysis, although accuracy depends on organism-specific knowledge and validated analytical methods.
Phenotypic and molecular methods are therefore complementary. Some resistance mechanisms are well predicted from known genetic markers; others involve poorly characterized combinations of gene expression, permeability, and regulatory changes. Rapid results also have limited value if specimen transport is slow or findings are not communicated and acted upon. Diagnostic performance consequently includes the entire pathway from specimen collection to interpretation, not merely the analytical speed of the instrument.
Epidemiology and measurement of global burden
The burden of AMR can be expressed as resistant infections, deaths, disability, healthcare use, or economic losses. Each measure answers a different question. The proportion of tested isolates that are resistant describes susceptibility among a selected group, whereas the population incidence of resistant infection depends on both resistance frequency and the total occurrence of infection. A high resistance percentage in a small, highly selected sample does not necessarily imply a high population burden.
A major analysis published in 2022 estimated the global burden of bacterial AMR in 2019 across 204 countries and territories. It estimated approximately 1.27 million deaths attributable to resistance and 4.95 million deaths associated with resistance. The estimates covered specified bacterial pathogens and organism–drug combinations, not all fungal, viral, and parasitic resistance. They were model-based estimates with uncertainty intervals rather than a direct worldwide count of laboratory-confirmed deaths.
The distinction between “attributable” and “associated” is essential. Attributable deaths compare observed resistant infections with a counterfactual in which those infections are susceptible. Associated deaths compare them with a counterfactual in which the resistant infections do not occur. The associated estimate is consequently broader and should not be described as deaths caused by resistance alone. These quantities are alternative analytical perspectives, not figures to be added together.
In that analysis, E. coli, S. aureus, K. pneumoniae, S. pneumoniae, A. baumannii, and P. aeruginosa were leading contributors to the estimated bacterial AMR burden. Lower respiratory infections accounted for a substantial share. Estimated attributable death rates were particularly high in western sub-Saharan Africa, illustrating how resistance interacts with infection incidence, access to care, laboratory capacity, and treatment availability rather than reflecting antimicrobial consumption alone.
Burden estimates face major data limitations. Blood cultures may be unavailable, obtained only in severe cases, or collected after treatment has begun. Laboratories may disproportionately serve urban referral hospitals. Deaths following treatment failure can be difficult to distinguish from deaths driven primarily by underlying disease. Changes in testing intensity, case mix, and reporting can also resemble changes in resistance. Reliable interpretation therefore requires uncertainty estimates, clear denominators, and explicit assumptions about missing data and causal comparisons.
Clinical consequences and antimicrobial exposure
Resistance can delay effective treatment, restrict available medicines, and increase the complexity of infection management. Consequences depend on the organism, infection site, severity, host condition, and availability of active alternatives. A resistant isolate does not inevitably cause a worse outcome, particularly when an effective drug is promptly available. Conversely, resistance can be especially consequential when infection progresses rapidly, as in sepsis, or when access to alternatives is limited.
AMR affects procedures that rely on infection prevention and treatment. Patients receiving chemotherapy, undergoing major surgery, or living with transplanted organs may have disrupted barriers or reduced immune defenses. Effective prophylaxis and treatment help make these interventions possible. However, prophylaxis itself creates antimicrobial exposure, illustrating why clinical benefit and selection pressure must be considered together rather than treating all antimicrobial use as avoidable.
Laboratory susceptibility is only one determinant of treatment effect. Pharmacokinetics describes how a drug is absorbed, distributed, metabolized, and eliminated; pharmacodynamics describes relationships between exposure and effect. Depending on the drug, antibacterial activity may relate most closely to time above the MIC, peak concentration relative to MIC, or total exposure relative to MIC. Drug penetration into cerebrospinal fluid, bone, abscesses, or intracellular compartments can differ substantially.
Some resistant infections require medicines with greater toxicity, less predictable exposure, restricted routes of administration, or limited availability. Others can be treated with newer agents whose activity depends on the precise resistance mechanism. Drug combinations may prevent resistance emergence in particular diseases or provide complementary activity, but combination therapy is not inherently superior in every infection. Its effects depend on organism, drugs, dose, and clinical context.
Treatment failure can also occur without genetic resistance. An undrained abscess, infected prosthesis, obstructed urinary tract, or poorly perfused tissue may sustain infection despite in-vitro susceptibility. Biofilms—structured microbial communities embedded in a matrix—can show reduced antimicrobial susceptibility through physiological heterogeneity, limited penetration, and slow growth. These conditions overlap with tolerance and persistence and help explain why laboratory findings must be interpreted alongside anatomy, host factors, and the physical source of infection.
Preventing infections and interrupting spread
Preventing infection reduces both the number of resistant infections and the antimicrobial exposure that can select resistance. This makes infection prevention a central AMR intervention rather than a separate supporting activity. Its effects are not limited to organisms already known to be resistant: reducing susceptible infections can also reduce treatment demand and opportunities for resistance to emerge or spread.
Reliable sanitation, safe water, and hygiene reduce transmission of many intestinal and other pathogens. Their importance is especially apparent where contaminated water and inadequate waste disposal create repeated exposure. In healthcare facilities, reliable water and sanitation support hand hygiene, environmental cleaning, sterilization, and safe clinical procedures. Infrastructure failures can undermine these processes even when formal infection-control policies exist.
Healthcare infection prevention combines surveillance, hand hygiene, cleaning, device management, aseptic technique, and organism-specific precautions. Screening and isolation or cohorting can form part of responses to selected resistant organisms, depending on the setting and available resources. Outbreak investigation examines patient movements, procedures, equipment, environmental reservoirs, and microbiological relationships. Genomic similarity can strengthen an outbreak hypothesis, but epidemiological evidence remains necessary to reconstruct plausible transmission.
Vaccines can reduce AMR through several pathways. They prevent infections that might otherwise require antimicrobial treatment, reduce transmission of resistant strains within the targeted pathogen, and sometimes prevent viral illnesses that lead to antibiotic prescribing or secondary bacterial infections. Pneumococcal conjugate vaccination has reduced disease caused by several resistant vaccine-type pneumococcal lineages. Effects on total resistance nevertheless depend on coverage, circulating serotypes, and replacement by non-vaccine types.
In animal production, vaccination, biosecurity, veterinary services, suitable housing, and husbandry can reduce infection pressure and dependence on antimicrobials. Environmental interventions include appropriate waste treatment and control of avoidable releases from healthcare, agriculture, and manufacturing. Measuring their AMR effects can be difficult because transmission pathways overlap and changes may take time. Evaluations therefore distinguish reductions in antimicrobial residues, resistance genes, colonization, infection, and disease burden rather than treating those outcomes as equivalent.
Antimicrobial stewardship and equitable access
Antimicrobial stewardship comprises coordinated efforts to improve antimicrobial use so that treatment is effective while avoidable harms and unnecessary selection are reduced. It concerns the choice of drug, indication, dose, route, timing, and duration, as well as whether treatment is needed. Stewardship is not synonymous with minimizing prescribing regardless of consequences: delayed or unavailable treatment of a serious bacterial infection can be immediately dangerous.
Healthcare stewardship programs commonly combine local treatment guidance, prescribing review, specialist or pharmacist input, and feedback to clinical teams. Diagnostic information can support revision of initial broad treatment to a more targeted regimen when an organism and susceptibility profile become available. Electronic decision support and prescribing restrictions are additional tools, but their effects depend on implementation, clinical context, and whether effective alternatives are accessible.
The WHO AWaRe classification groups antibiotics into Access, Watch, and Reserve categories. Access antibiotics generally have activity against common susceptible pathogens and a comparatively lower resistance potential; Watch agents require particular attention because of their resistance-selection potential; Reserve agents are intended for selected situations involving highly resistant infections. These categories support monitoring and policy, but they do not replace organism-specific evidence or mean that every Access antibiotic is appropriate for every infection.
Stewardship also addresses avoidable harms beyond resistance. Antibiotics can disrupt the microbiota, cause adverse drug reactions, and increase susceptibility to infection with Clostridioides difficile. For some infections, evidence has supported shorter treatment courses than those historically used, while other diseases require prolonged therapy. Duration is therefore a disease- and evidence-specific question rather than a universal rule that either longer or shorter treatment is always safer.
Equitable access is inseparable from stewardship. Some populations experience excessive or poorly targeted use, while others cannot obtain timely diagnosis, reliable medicines, or complete effective treatment. Stockouts, unaffordable prices, substandard products, and fragmented healthcare can contribute to inadequate exposure and treatment failure. Policy must therefore distinguish necessary access from inappropriate use. Antimicrobial consumption data are most informative when interpreted alongside infection burden, prescribing indications, medicine quality, and unmet treatment need.
Surveillance, governance, and international coordination
AMR surveillance connects laboratory findings with information about patients, antimicrobial use, and health outcomes. Its purposes include detecting emerging mechanisms, identifying outbreaks, informing treatment policy, and evaluating interventions. Effective systems need standardized methods, quality assurance, appropriate sampling, and stable reporting. A large database is not necessarily representative if participating laboratories serve only selected hospitals or patient groups.
The WHO launched the Global Antimicrobial Resistance and Use Surveillance System, commonly known as GLASS, in 2015. It provides standardized approaches for collecting and interpreting national data. National and regional systems differ in coverage and design, and surveillance also includes organism-specific networks for tuberculosis, HIV, malaria, influenza, and other infections. Integrating these activities requires attention to different biological and programmatic requirements rather than assuming one sampling strategy can cover all AMR.
Antimicrobial-use surveillance provides complementary information. Human medicine commonly uses measures such as defined daily doses or days of therapy, each with limitations. A defined daily dose is a standardized technical unit, not necessarily the actual prescribed dose, and can be difficult to interpret in children or populations with altered dosing needs. Veterinary monitoring often reports quantities relative to animal populations or biomass, but such metrics can obscure differences in species, potency, and treatment practices.
International governance increasingly uses a One Health framework. The Food and Agriculture Organization, United Nations Environment Programme, WHO, and World Organisation for Animal Health collaborate as the Quadripartite. The 2015 Global Action Plan established objectives addressing awareness, surveillance and research, infection prevention, optimized antimicrobial use, and sustainable investment. National action plans translate these aims into different legal, financial, laboratory, agricultural, and healthcare arrangements.
At the 2024 United Nations high-level meeting, governments adopted a political declaration including a target to reduce deaths associated with bacterial AMR by 10% by 2030 against the 2019 baseline estimate. Such targets require careful specification of measurement and counterfactual assumptions. Governance also involves difficult allocation questions: responsibility for manufacturing discharges, financing laboratory systems, incentives for drug development, and fair access to new products. Differences in resources and infrastructure mean that identical formal obligations can require very different levels of investment.
Drug development, alternative approaches, and research challenges
New antimicrobials remain important, but discovery alone cannot provide a permanent escape from evolution. Development must address activity against priority pathogens, penetration into relevant tissues, toxicity, formulation, resistance potential, and feasibility of use. A compound active in laboratory culture may fail because adequate exposure cannot be achieved safely or because its performance changes in the complex environment of infection.
New beta-lactam–beta-lactamase inhibitor combinations illustrate mechanism-directed development. Inhibiting selected resistance enzymes can restore the activity of an accompanying antibiotic, but coverage varies by enzyme family. Other strategies include modifying existing scaffolds, identifying new targets, and exploiting bacterial nutrient-uptake systems. Clinical evaluation must establish not merely in-vitro activity but safety and efficacy in defined infections, ideally including patients whose pathogens express the resistance mechanisms of interest.
Alternative approaches include bacteriophages, anti-virulence agents, monoclonal antibodies, antimicrobial peptides, and microbiome-directed interventions. Phages infect bacteria and can be highly host-specific, which creates opportunities for targeted treatment but complicates matching, manufacturing, and trial design. Bacteria can also evolve phage resistance. Anti-virulence approaches attempt to reduce disease-causing functions rather than directly inhibit growth, but their clinical value depends on the infection and host response.
Microbiota-based therapies have established roles in particular recurrent C. difficile settings, while attempts to use microbiome manipulation to eliminate other resistant organisms remain an area of investigation. Vaccines against resistant pathogens or their virulence components could prevent infections without relying on treatment after disease begins. Artificial intelligence and computational screening can assist candidate discovery, but predictions still require experimental validation, toxicology, manufacturing development, and clinical trials.
Economic incentives present a distinct obstacle. Antibiotics are often used for short courses, while stewardship may deliberately restrict sales of valuable new agents. Commercial returns can therefore be poorly aligned with public-health value. Proposed responses include research grants, public–private partnerships, market-entry rewards, and subscription-style payments that partly separate revenue from sales volume. Evaluating these arrangements requires attention to access, appropriate use, and sustained supply. Research also continues on rapid susceptibility testing, resistance prediction, transmission attribution, and combinations that suppress emergence without creating unacceptable toxicity or ecological disruption.