Antibiotic Resistance

TL;DR. Antibiotics do not create resistance; they select for it. In any large bacterial population, a few organisms are already resistant by chance mutation or because they picked up a resistance gene from a neighbour. Kill everything else and the resistant ones inherit the space. This is evolution operating on a timescale of days, and it has been running at industrial scale since the 1940s, in hospitals, in homes, and in the billions of food animals given antibiotics routinely. The result is that surgery, chemotherapy, transplantation, and intensive care, all of which depend on reliable antibiotics, are quietly becoming riskier. In 2019, drug-resistant bacterial infections were directly responsible for about 1.27 million deaths and associated with 4.95 million.

Key takeaways

• Resistance is not new or surprising. Fleming warned about it in his 1945 Nobel lecture, and penicillin resistance in Staphylococcus aureus appeared within a few years of its introduction.
• Bacteria share resistance genes horizontally, across species, on plasmids. A harmless gut bacterium can hand a resistance gene to a dangerous one.
• The biggest driver by volume is agriculture: a large share of all antibiotics produced are given to food animals, much of it for growth promotion and disease prevention rather than treating sick animals.
• The pipeline is broken for economic reasons. A new antibiotic is meant to be used sparingly and briefly, which is the opposite of a profitable drug, so most large pharmaceutical companies left the field.
• Individual action matters less than people are told and more than nothing: the main levers are prescribing practice, infection control, vaccination, sanitation, and agricultural policy.
• Without change, projections suggest millions of deaths a year attributable to resistance by 2050, with the largest burden in South Asia and sub-Saharan Africa.

What it is

In short: Antibiotics attack something bacteria have and we do not, and resistance is any change that defeats that attack.

An antibiotic kills or stops bacteria by attacking something bacteria have and human cells do not: a cell wall, a distinct ribosome, a bacterial enzyme.

Resistance is any bacterial change that defeats one of these attacks.

Don't be confused: it is the bacteria that become resistant, not you. People commonly believe their body becomes "immune to antibiotics." It does not. The resistant population lives in and on you, and it can be passed to others and into the environment. This distinction changes what the problem is: it is ecological, not personal.

The history

In short: Fleming warned about resistance in his 1945 Nobel lecture, and every new class since has been followed by resistance within a few years.

1928: Alexander Fleming notices a mould contaminating a Staphylococcus plate, with a clear ring where bacteria failed to grow. 1940 to 1942: Florey, Chain, and Heatley turn it into a usable drug. Penicillin transforms the treatment of pneumonia, sepsis, syphilis, and wound infection.

1945: Fleming's Nobel lecture warns explicitly that under-dosing and careless exposure would select for resistant organisms, and that "the thoughtless person playing with penicillin treatment is morally responsible for the death of the man who succumbs to infection with the penicillin-resistant organism." He was describing a mechanism that was already underway.

1940s to 1960s: the golden age. Most antibiotic classes still in use were discovered in these two decades, mostly from soil organisms.

1960s onward: resistance follows each new drug, usually within a few years. Penicillin-resistant staphylococci, then methicillin-resistant S. aureus (MRSA) in 1961, then vancomycin-resistant enterococci, then extended-spectrum beta-lactamase producers, then carbapenem-resistant Enterobacterales, then colistin resistance carried on a plasmid (the mcr-1 gene, reported in 2015) affecting the drug held in reserve as the last resort.

1970s to 2000s: discovery slows to almost nothing. Most "new" antibiotics since the 1980s are modifications of existing classes. Two genuinely new classes reached the clinic (oxazolidinones and lipopeptides), and resistance to both has been reported.

How resistance actually works

In short: Four mechanisms, plus the ability to trade resistance genes directly between species, which is why it spreads faster than inheritance allows.

Four mechanisms, and knowing them explains why some combinations of drugs exist.

1. Destroy the drug. Beta-lactamases are enzymes that cut the beta-lactam ring at the heart of penicillins. Thousands of variants exist. The counter-move is a beta-lactamase inhibitor given alongside the antibiotic (as in amoxicillin-clavulanate), which sacrifices itself to the enzyme so the antibiotic survives. Extended-spectrum beta-lactamases (ESBLs) destroy a wider range, and carbapenemases destroy even the reserve drugs.

2. Change the target. MRSA acquired a gene (mecA) encoding an alternative cell-wall building enzyme that beta-lactams cannot bind. Fluoroquinolone resistance comes from point mutations in DNA gyrase. Rifampicin resistance in TB is a change in RNA polymerase.

3. Pump the drug out. Efflux pumps in the bacterial membrane export antibiotics before they reach a lethal concentration. These are often broad, conferring resistance to several unrelated drugs at once.

4. Keep the drug out. Reduced permeability, by losing or altering the porin channels that antibiotics use to cross the outer membrane. This is one reason gram-negative bacteria, which have an extra outer membrane, are so much harder to treat than gram-positives, and why the gram-negative pipeline is the emptiest.

Why it spreads so fast: horizontal gene transfer

Human resistance genes would spread only by inheritance, over generations. Bacteria do something far more efficient. Plasmids, small circular DNA molecules separate from the chromosome, carry resistance genes and can be transferred directly between bacteria, including between different species, by conjugation: one cell builds a bridge to another and copies the plasmid across. Bacteria also pick up free DNA from their environment (transformation) and receive genes delivered by viruses (transduction).

The consequences are severe. A single plasmid can carry resistance to several antibiotic classes at once, so using one drug selects for resistance to others. Resistance genes move between harmless gut bacteria and pathogens. And they move between animal and human bacterial populations, through food, water, farm workers, and manure spread on fields.

What it does

In short: Beyond failed treatment, it undermines surgery, chemotherapy, transplantation, and neonatal care, all of which assume working antibiotics.

Clinically, a resistant infection means: treatment fails while the patient gets worse, second-line drugs are more toxic and less effective, hospital stays lengthen, and mortality rises. For a bloodstream infection, the delay between presentation and effective therapy is one of the strongest predictors of death.

Systemically, it undermines everything that depends on controlling infection:

The organisms of most concern are often summarised as ESKAPE: Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacter species, because they escape the effects of common antibiotics. Add drug-resistant TB (Chapter 29), resistant gonorrhoea (Chapter 35), extensively drug-resistant typhoid, and Candida auris, a resistant fungus that spreads in hospitals and is difficult to eradicate from surfaces.

Clostridioides difficile deserves separate mention because it is caused by antibiotics rather than resisting them. Broad-spectrum antibiotics wipe out the normal gut flora, allowing C. difficile to overgrow and produce toxins causing severe colitis. It is a leading healthcare-associated infection, it recurs in a fifth of cases, and the most effective treatment for recurrent cases is faecal microbiota transplantation, restoring a normal bacterial community from a healthy donor.

Is it deadly?

Yes, at a scale comparable with the largest infectious diseases.

• The Global Research on Antimicrobial Resistance (GRAM) study estimated that in 2019, 1.27 million deaths were directly attributable to bacterial antimicrobial resistance and 4.95 million deaths were associated with it. That places direct attributable mortality above malaria and above HIV.
• The highest death rates were in sub-Saharan Africa and South Asia.
• Six pathogens accounted for the majority: E. coli, S. aureus, K. pneumoniae, S. pneumoniae, A. baumannii, and P. aeruginosa.
• Forecasts published in 2024 projected on the order of 39 million cumulative deaths attributable to resistance between 2025 and 2050 on current trends, with the sharpest rises among older adults.

Those forecasts carry real uncertainty. The direction does not.

Is it contagious?

Yes, and this is the point people most often miss. Resistant bacteria spread exactly like their susceptible relatives: hands, surfaces, food, water, and contact. And the resistance genes themselves spread between bacteria.

Practical consequences:

• Hospitals are amplifiers, concentrating sick patients, invasive devices, and heavy antibiotic use in one building. Hand hygiene, isolation of colonised patients, device care bundles, and environmental cleaning are the countermeasures, and compliance with hand hygiene in hospitals is chronically below target.
• Travel spreads resistance. Travellers to regions with high resistance frequently return colonised with ESBL-producing bacteria in their gut, usually without any illness.
• Farms and food. Resistant organisms move from animals to humans through meat, farm contact, and environmental contamination.

What drives it

In short: Over-prescribing, agriculture, poor infection control, substandard medicines, and an empty pipeline caused by an economic model that punishes reserved drugs.

The economics deserve explaining, because it is the crux of why this is not being solved by the market. A new antibiotic that works against a resistant organism should be reserved, used rarely and for short courses, to preserve it. That is the correct clinical policy and the worst possible business case. Compare a drug for a chronic disease taken daily for thirty years. Several small companies that successfully brought new antibiotics to market went bankrupt shortly afterwards. Proposed fixes involve delinking revenue from volume: subscription models where a health system pays a fixed annual sum for access regardless of use (the UK and Sweden have piloted this), market entry rewards, and push funding for early research such as CARB-X and GARDP.

Treatment, and how it works

In short: Stewardship, shorter courses, rapid diagnostics, infection prevention, and vaccination, which is an underrated antibiotic-sparing measure.

Stewardship, meaning using the right drug, at the right dose, for the right duration, and not at all when it is not needed:

• Do not treat viral infections. Most sore throats, colds, coughs, and sinus congestion are viral. Delayed prescriptions (given with instructions to fill only if not improving) reduce antibiotic use substantially without worsening outcomes.
• Narrow the spectrum as soon as culture results allow. Starting broad in a septic patient is correct; staying broad on day four is not.
• Shorter courses. A large body of trial evidence now shows that shorter courses are as effective as longer ones for community-acquired pneumonia, urinary infections, intra-abdominal infection, and several others. The old instruction to "always finish the course" was based on reasoning rather than evidence, and current advice for many infections is to use the shortest evidence-based duration. The instruction still holds where it was properly established, notably tuberculosis.
• Rapid diagnostics, distinguishing bacterial from viral infection and identifying the organism and its sensitivities in hours rather than days. Procalcitonin-guided protocols reduce antibiotic use in respiratory infection.
• Infection prevention: hand hygiene, catheter and line care, surgical prophylaxis timing, isolation, and cleaning.
• Vaccination, which is an underrated antibiotic-sparing intervention. Pneumococcal conjugate vaccine reduced both pneumococcal disease and the resistant strains specifically. Influenza vaccination reduces the secondary bacterial infections that prompt antibiotic use. Typhoid conjugate vaccine directly addresses drug-resistant typhoid.

New and revived treatments for resistant infections: newer beta-lactam and beta-lactamase-inhibitor combinations (ceftazidime-avibactam, meropenem-vaborbactam, ceftolozane-tazobactam), cefiderocol (which smuggles itself into gram-negative bacteria through their iron uptake system, a genuinely clever mechanism), and the return of older toxic drugs such as colistin because nothing else works.

What the person can do

Honestly scaled, because individual behaviour is not the main driver:

• Do not ask for antibiotics for viral illness, and do not treat a prescription as the measure of a good consultation.
• Do not use leftover antibiotics or someone else's.
• Take them as prescribed, at the intervals given, for the duration your clinician specified for that infection.
• Get vaccinated, which prevents both the infections and the antibiotic courses.
• Wash hands and prepare food safely, particularly around raw poultry and meat.
• Ask two questions when antibiotics are offered: "Is this likely to be bacterial?" and "How long a course does the evidence support for this?"
• As a consumer and voter, agricultural antibiotic policy and hospital infection control funding are where the large levers are.

What's next

In short: Phage therapy, antivirulence drugs, machine-learning discovery, and payment models that pay for access rather than for volume.

• Bacteriophage therapy: viruses that infect and kill specific bacteria. Used in Georgia and the former Soviet Union for decades, largely ignored in the West, and now returning through compassionate-use cases and formal trials. Phages are extremely specific, which is both their strength (they spare the microbiome) and their difficulty (each infection needs a matched phage).
• Antibodies, antivirulence drugs, and anti-biofilm agents, which disarm bacteria rather than killing them, in principle producing less selection pressure.
• CRISPR-based antimicrobials engineered to cut resistance genes specifically.
• Microbiome restoration, extending the C. difficile faecal transplant success to decolonising patients carrying resistant organisms.
• Machine learning in discovery: models trained to predict antibacterial activity have identified new candidate compounds from large chemical libraries, including structurally novel candidates. Early, promising, and not yet in clinical use.
• Policy: national action plans, surveillance networks (GLASS), delinked payment models, and agricultural restrictions. The 2024 UN General Assembly high-level meeting on antimicrobial resistance set political targets; whether financing follows is the open question.

Sources and notes

Mortality figures: Antimicrobial Resistance Collaborators, The Lancet, 2022 (1.27 million deaths attributable and 4.95 million associated with bacterial AMR in 2019). Forecasts to 2050: GRAM Project, The Lancet, 2024 (approximately 39 million cumulative attributable deaths, 2025 to 2050). Fleming's warning: Nobel Lecture, 11 December 1945. MRSA first reported 1961. mcr-1 plasmid-borne colistin resistance: Liu et al., Lancet Infectious Diseases, 2016. ESKAPE designation: Rice, Journal of Infectious Diseases, 2008. Short-course therapy evidence: multiple randomised trials summarised in current infectious disease guidelines. Antibiotic use in agriculture: WOAH and FAO reporting; the exact human/animal split is contested and varies by country. Machine learning antibiotic discovery: Stokes et al., Cell, 2020, and subsequent work.

Open questions. How much agricultural antibiotic use contributes to human resistance relative to human prescribing is genuinely disputed, and the answer differs by organism and country. Whether phage therapy can be manufactured and regulated at scale is unresolved.

Next: the organ that makes you who you are, and the diseases that take it apart. 👉