How does antibiotic resistance develop?

How does antibiotic resistance develop?

When bacteria become resistant, did each cell decide to adapt after encountering the drug? Or were some variants already better able to survive, allowing the treatment to change which bacteria remained? The two accounts sound similar and are not. One says the drug creates the trait; the other says the drug reveals a trait that was already present. Everything downstream depends on which is right.

A bacterial population contains genetic variation created by mutation and by exchanging genes. An antibiotic imposes a strong environmental test: susceptible bacteria die or stop reproducing, while bacteria carrying a useful resistance trait are more likely to survive.

Consider the numbers involved, because they are what make this inevitable rather than unlucky. A single infection can hold populations of bacteria far larger than the human population of the planet, dividing on the order of every twenty minutes under good conditions. Even a very rare mutation — one cell in many millions — is therefore not rare in absolute terms; it is present, several times over, before any drug arrives. Selection does not need to wait for the right variant to appear. It only needs to remove everything else.

Those survivors reproduce and pass resistance onward. But descent alone is too slow to explain what is actually observed, which is resistance appearing in a species that has never met the drug, in a hospital, within months. So something else must be happening. Bacteria can move genes sideways, between living cells and even across species boundaries, by three routes worth naming: conjugation, where one cell extends a bridge to another and copies across a plasmid — a small loop of DNA separate from the chromosome, which often carries several resistance genes at once; transformation, where a cell simply takes up loose DNA from a dead neighbour in its surroundings; and transduction, where a virus that infects bacteria carries fragments of one host's DNA into the next. The consequence is that resistance is not confined to a family tree. It circulates. And because plasmids frequently bundle resistance to several unrelated drugs together, using one antibiotic can select for resistance to others the patient never took.

Use matters because every exposure creates selection pressure. Unnecessary antibiotics expose bacteria without providing patient benefit, while incorrect use can leave surviving populations opportunities to recover. Hospitals, farms, communities, travel, sanitation, and prescribing practices all affect how resistant strains spread — and the farms belong on that list rather than as an afterthought. A substantial share of the antibiotics produced worldwide, in some countries the majority by weight, goes to food animals, much of it not to treat sick animals but at low doses across healthy herds. Low, sustained exposure across enormous populations is close to a laboratory design for selecting resistance, and the resulting genes do not stay on the farm.

The human body does not become antibiotic-resistant; bacterial populations do. This matters practically: a resistant infection is not a property you have acquired but a population you have been colonized by, and it can be passed to someone who has never taken an antibiotic in their life.

If resistance is selected rather than created, we should expect resistance genes to exist before the drugs did — and they do, which is the most striking confirmation available. Researchers recovered genes conferring resistance to beta-lactam, tetracycline and glycopeptide antibiotics from 30,000-year-old permafrost sediments in the Yukon, with the vancomycin-resistance element close in structure to modern versions. Bacteria cultured from a region of Lechuguilla Cave in New Mexico, isolated from the surface for over four million years, proved resistant to a wide range of commercially available antibiotics, including daptomycin, a drug of last resort. Nothing there had ever encountered a pharmaceutical. The genes are ancient because most antibiotics are themselves microbial weapons, and the organisms sharing that soil have been defending against them for a very long time. Medicine did not start this contest; it walked into one already in progress.

Two things the popular account gets wrong. The first is dosage intuition: more drug is not automatically better, because unnecessary exposure also increases selection pressure and can cause harm. The second is the instruction to always finish the course. That advice is no longer settled. A 2017 analysis in the BMJ argued that for many common infections there is no evidence that stopping early promotes resistance, while taking antibiotics for longer than needed demonstrably increases selection pressure — and that the traditional course lengths were largely inherited convention rather than the product of trials. This remains genuinely contested, and the correct duration varies by infection: some, such as tuberculosis, absolutely require long completed regimens. The honest statement is not "stop when you feel better" but "the right duration is a clinical question your prescriber should answer, and shorter is more often correct than the slogan implies."

One last piece, which is economic rather than biological. If resistance is inevitable, we need a supply of new drugs — and the pipeline is thin, not because the chemistry is impossible but because the business case is inverted. A genuinely new antibiotic will be held in reserve, used sparingly and briefly, and priced like a generic. Achaogen won FDA approval for plazomicin in 2018, sold well under a million dollars of it, and filed for bankruptcy about a year later. A drug can be a public health success and a commercial failure at the same time, and until that is fixed, stewardship of the antibiotics we already have is doing most of the work.

• Why combination therapies can slow resistance.
• How antibiotic stewardship balances individual and public health.
• Why resistance sometimes fades when a drug is withdrawn, and sometimes does not.

Every term the collection defines is gathered in the glossary.