Antibiotics: How They Work, Why Resistance Happens, and What to Do

Antibiotics are drugs that kill bacteria or stop them from multiplying, and they work only against bacteria, not viruses. Most common antibiotics attack one of four bacterial targets: the cell wall, the ribosome, DNA replication, or folate synthesis. Penicillin, invented in 1928 and mass-produced by 1945, cut pneumonia and wound-infection deaths by more than 70 percent in a decade. The catch is evolution: every course exerts selection pressure, and resistant bacteria survive to pass on their defenses. That is why the same drug that cured a strep throat in 1990 may fail today in roughly 1 in 5 cases for some organisms.

The mechanism: four ways antibiotics break a bacterium

Bacteria are single cells with a rigid outer wall made of peptidoglycan, a mesh of sugar chains cross-linked by short peptides. That wall is the pressure vessel holding the cell together; internal osmotic pressure runs roughly 2 to 20 atmospheres depending on the species. Penicillin and its relatives, the beta-lactams, block the enzyme that stitches those cross-links, so the wall weakens and the cell bursts. This is why beta-lactams kill actively dividing bacteria: a dormant cell builds no new wall and escapes.

Ribosome-targeting drugs work differently. Macrolides and tetracyclines bind the bacterial 70S ribosome and jam protein synthesis, so the cell cannot make the enzymes it needs to grow. Aminoglycosides go further and cause the ribosome to misread genetic code, producing toxic junk proteins. Because the bacterial ribosome differs structurally from the human 80S version, these drugs hit the invader without shutting down your own protein factories, though side effects like tinnitus with aminoglycosides show the overlap is not perfect.

A third class attacks DNA. Fluoroquinolones trap the enzyme DNA gyrase mid-cut, which shatters the chromosome. Metronidazole, used for anaerobic infections like C. difficile, is activated only inside low-oxygen bacteria, where it fragments DNA. A fourth class, the sulfonamides, blocks folate synthesis, a pathway humans do not have at all, making them selectively toxic by design.

In plain terms: each antibiotic is a key that fits one bacterial lock, and resistance is the bacterium changing the lock.

Bactericidal drugs kill outright; bacteriostatic drugs only pause growth and let your immune system finish the job. The distinction matters clinically: in a neutropenic patient with no white cells to mop up, a static drug may fail where a cidal one succeeds. Most cell-wall agents are cidal; most protein-synthesis inhibitors are static.

The resistance timeline: how a wonder drug became a race

Resistance is not new. Alexander Fleming warned about it in his 1945 Nobel lecture. The pattern since has been remarkably consistent: a drug launches, resistance appears within years, and the clock restarts.

The median gap from launch to first clinical resistance report for major classes has been roughly 5 to 10 years in hospital settings. For some organisms the gap has collapsed to months. That compression is the core problem: the drug pipeline cannot outrun a one-generation-per-20-minutes evolutionary engine.

How resistance actually spreads: the horizontal gene problem

Here is the non-obvious part most explainers skip. Bacteria do not only pass resistance to their offspring vertically. They trade resistance genes horizontally, across species, via small circular DNA molecules called plasmids. A harmless gut E. coli can hand a carbapenem-resistance gene to a deadly Klebsiella in the same patient's intestine, even though the two never competed directly.

This is why antibiotic use in one setting echoes in another. Livestock receives an estimated 70 percent of global antibiotic tonnage, mostly for growth promotion and disease prevention in crowded pens. Those drugs select resistant bacteria in animals, and resistant strains or their genes reach humans through food, water, and direct contact.

Four genetic mechanisms drive resistance: enzymatic destruction of the drug (beta-lactamases, including the notorious NDM-1 and KPC enzymes), target modification (MRSA's altered penicillin-binding protein), reduced drug entry or increased pumping out through efflux pumps, and bypass pathways that route around the blocked step. A single bacterium can carry several at once, which is how "superbugs" become resistant to five or six drug families simultaneously.

The horizontal transfer rate is not trivial: in dense gut communities, plasmid transfer can occur in 1 in 100 to 1 in 10,000 cell contacts, enough to spread a gene through a population within days.

What the numbers say: how big the problem is

Antimicrobial resistance caused an estimated 1.27 million direct deaths in 2019 and was associated with 4.95 million, per the landmark GRAM study published in The Lancet in 2022. By 2050, projections place direct deaths at 1.9 million and associated deaths at 10 million annually if trends hold. Those are projections, not certainties, and they depend heavily on how fast new drugs arrive.

The clinical picture varies sharply by organism. MRSA causes roughly 120,000 bloodstream infections in the US each year and kills about 20,000. Carbapenem-resistant Enterobacteriaceae, the hardest to treat, carry mortality of 40 to 50 percent in bloodstream infections because the fallback drugs are older and more toxic. C. difficile, often triggered by antibiotics themselves, causes about 500,000 US infections and 29,000 deaths annually.

On the human cost side, a resistant infection roughly doubles hospital length of stay and adds USD 1,000 to USD 2,000 per case in direct costs for common infections, and far more for resistant ICU cases. The WHO estimates antibiotic resistance could push 24 million people into extreme poverty by 2030 through health costs and lost income.

Outpatient prescribing remains a major lever. Studies in the US show roughly 30 percent of antibiotic prescriptions written in doctors' offices are unnecessary, and another large share are for the wrong drug or duration. Each unnecessary course is pure selection pressure with no benefit.

Common misconceptions, corrected

The biggest myth is that antibiotics work on colds and flu. They do not. Viral infections have no cell wall, no bacterial ribosome, and no folate pathway for the drug to touch. Taking antibiotics for a viral illness does not shorten it; it only selects resistant bacteria in your gut and risks side effects. Roughly 1 in 5 patients on antibiotics develops diarrhea, driven partly by disruption of the gut microbiome that can persist for months.

A second misconception: "finish the whole course" is now more nuanced than it was. The old advice to always complete every pill was designed to prevent relapse in a less-resistant era. Modern evidence, reflected in updated guidance, suggests that for many common infections a shorter course (5 days instead of 10) works equally well and exerts less selection pressure. The precise duration depends on the infection: do not stop early on your own for a bloodstream infection or tuberculosis, where interruption breeds resistance fast.

A third error is the belief that stronger or broader-spectrum drugs are always better. Broad-spectrum agents like ceftriaxone hit many species at once and disrupt the microbiome more than narrow agents. For a strep throat, plain penicillin is still first-line, not because it is old but because it is narrow, cheap, and S. pyogenes has never developed meaningful resistance to it.

A fourth and counterintuitive point: antibiotics do not just kill the target. They release endotoxin as bacteria die, which can trigger a spike in fever and inflammation within a few hours of the first dose. In severe infections this Herxheimer-like reaction is expected, not a sign the drug is failing.

What to actually do: practical antibiotic use

The single highest-value action is to not ask for antibiotics when the illness is viral. A sore throat with cough and runny nose is viral in roughly 85 to 90 percent of cases in adults; strep is more likely only with sudden fever, no cough, and swollen tonsils. Ask your clinician what the diagnosis is and why the drug is or is not indicated.

When you are prescribed an antibiotic, the concrete steps are these:

• Take it at evenly spaced intervals and complete the prescribed duration unless your clinician adjusts it.
• Never save leftover pills or share them; leftover courses are a leading route to sub-therapeutic dosing and resistance.
• Report severe diarrhea, rash, or difficulty breathing immediately, since these are red flags for C. difficile infection or allergy.
• Ask whether a narrower drug or shorter course is appropriate for your specific infection.
• Finish the course, but ask at the outset whether a shorter duration is supported for your condition.

Prevention matters as much as treatment. Vaccination against pneumococcus, Haemophilus influenzae type b, and typhoid reduces the infections that drive antibiotic use. Hand hygiene cuts transmission of resistant organisms in households and hospitals. In food, cooking meat thoroughly and washing produce reduces exposure to resistant bacteria, though it cannot eliminate the gene-transfer route.

For clinicians and institutions, the evidence-backed tool is antibiotic stewardship: a program that reviews prescriptions, de-escalates broad drugs once culture results arrive, and sets local guidelines. Hospitals with active stewardship programs cut inappropriate antibiotic use by 20 to 40 percent and reduce C. difficile rates by roughly a third.

Limitations: what antibiotics cannot fix and what is coming

Antibiotics cannot reach all infections equally. Biofilms, the slimy communities bacteria build on catheters, heart valves, and prosthetic joints, resist drug penetration so well that a biofilm infection may need surgical removal rather than more drugs. Abscesses need drainage, because no antibiotic penetrates a walled-off pool of pus. Dormant persister cells survive cidal drugs and seed relapses, which is why tuberculosis needs 6 months of treatment, not 6 days.

The drug pipeline is thin. Only about 40 new antibiotics were in clinical development as of the mid-2020s, and few represent genuinely new mechanisms against the hardest gram-negative bugs. The economics work against developers: a truly novel antibiotic must be used sparingly to preserve its power, which undercuts sales.

Alternatives are advancing but unproven at scale. Bacteriophage therapy, which uses viruses that infect bacteria, has case reports of success against resistant infections and is in clinical trials. Monoclonal antibodies and antimicrobial peptides are in early stages. Fecal microbiota transplant restores gut bacteria after recurrent C. difficile and works in roughly 85 to 90 percent of refractory cases. None yet replaces the antibiotic.

The honest bottom line: antibiotics remain among the most consequential drugs ever made, and their future depends less on the next molecule than on how carefully the current ones are used. Every course is a trade between curing one patient and preserving the drug for the next.