A virus that hunts the bacteria making you ill has been used as medicine for a century. It still cannot be sold like one

Every antibiotic works by finding something bacteria have and we do not — a cell wall, a bacterial ribosome — and poisoning it. That strategy has a fixed lifespan, because bacteria evolve, and the pipeline of genuinely new antibiotic classes has been thin for decades.

There is an older answer that never went away. Bacteria have their own predators: bacteriophages, viruses that infect bacteria and nothing else. They are thought to be the most numerous biological entities on the planet, they have been killing bacteria for billions of years, and they were given to human patients in the 1920s, before anyone had heard of penicillin. The question worth asking is not whether they work. It is why, a century later, you still cannot buy one.

A predator with a very narrow menu

A phage attaches to a specific molecule on a bacterial surface — a protein, a sugar, a piece of the cell's outer machinery — and injects its genetic material. The cell's own machinery is hijacked to build new phages, and within minutes to an hour the bacterium bursts, releasing dozens or hundreds of copies to find the next one.

The critical property is specificity. An antibiotic is a blunt instrument that hits broad groups of bacteria, which is why a course of it disrupts the gut for weeks. A phage is usually specific not merely to a species but to particular strains within it. One phage may kill one hospital's isolate of Klebsiella and completely ignore another isolate of the same species from down the corridor.

That is simultaneously the best and the worst thing about phage therapy. Best, because it leaves the rest of the microbiome untouched — the collateral damage that defines antibiotic treatment simply does not occur. Worst, because there is no such thing as a phage that treats "a urinary tract infection". You must first grow the patient's own bacteria, then test candidate phages against that isolate to find one that lyses it. Treatment begins with matching, not prescribing.

Only some phages are usable

Phages split into two lifestyles, and only one is safe to use as medicine.

Lytic phages do what was described above: infect, replicate, burst the cell, move on. Temperate phages can instead insert their genome into the bacterial chromosome and sit there quietly, replicating along with the host, sometimes for generations, before excising and going lytic. A dormant phage does not kill anything, and worse, when it eventually leaves it can carry fragments of bacterial DNA with it to the next cell — the process called transduction. That is one of the routes by which resistance and toxin genes move between bacteria.

So therapeutic phages must be strictly lytic, and characterising them properly means sequencing the genome and confirming it carries no integration machinery, no toxin genes and no antibiotic-resistance genes. This is also where engineering enters: a promising temperate phage can be converted into an obligately lytic one by deleting the gene that maintains dormancy, which is precisely what was done in the widely reported case of a teenager with a disseminated mycobacterial infection after a lung transplant, treated in 2019 with an engineered phage cocktail when nothing else remained.

A medicine that makes more of itself

Phage therapy breaks the pharmacological logic that every dosing regimen is built on.

A conventional drug is administered, distributed, metabolised and cleared; concentration falls predictably, so you dose again. A phage does the opposite. Where its target bacteria are abundant it replicates, so the effective dose rises at the site of infection — and once the bacteria are gone it stops replicating and is cleared. The active agent is amplified by the disease it is treating and fades with it.

This has real consequences. It means a small dose delivered to the right place can be enough. It also means delivery is everything: phages must physically reach the bacteria, so route matters enormously — direct application to a wound or a joint, inhalation for a lung infection, intravenous administration for a systemic one, where the immune system will also start clearing the phages as foreign particles. And preparations must be purified carefully, because a phage lysate is by definition full of burst bacteria, and the endotoxin in those fragments is dangerous in its own right.

An antibiotic is a chemical that is consumed as it works. A phage is an organism that is reproducing as it works, and the bacterium is reproducing too. This is not pharmacology, it is ecology — and both sides are evolving during treatment.

Resistance arrives quickly, and can be used

Bacteria are not defenceless against phages; they have been fighting them for billions of years, and the defences are formidable. They modify or lose the surface receptor the phage binds to. They cut incoming phage DNA with restriction enzymes. And they use CRISPR — the gene-editing tool now famous in laboratories is, in its native role, a bacterial immune system that stores fragments of previously encountered phage DNA in order to recognise and destroy it on return.

So phage resistance emerges fast, often within days, and any serious treatment uses a cocktail of several phages targeting different receptors, so that escaping one does not mean escaping all.

But there is an elegant twist that pure antibiotics cannot offer. Resistance always costs the bacterium something, and if you choose a phage that binds to a structure the bacterium needs for its own virulence or drug resistance, then the mutants that escape the phage are the ones that have damaged that structure. There are documented cases of bacteria evolving phage resistance by altering an efflux pump — and becoming re-sensitised to antibiotics they had previously defeated. The phage does not have to win. It only has to make winning expensive, a strategy sometimes called phage steering.

Why you cannot buy one

The obstacle is not biology. It is that drug regulation was built around a fixed chemical entity: you define the molecule, prove it is consistently manufactured, run trials on a defined population, and receive an approval for that exact thing.

Phage therapy fits none of that. The treatment may be assembled per patient from a library, the composition changes when the bacterium changes, and the point of the approach is that it is matched rather than standardised. A trial testing one fixed cocktail against a broad patient group is testing the therapy at its weakest — many participants will be infected with strains the cocktail does not cover — and several trials have returned ambiguous results for exactly this reason. Most use in Europe and North America has therefore proceeded case by case under compassionate-use provisions, while sustained clinical experience has accumulated in a few centres that never abandoned the approach.

There is movement — magistral and hospital-preparation frameworks in some countries, phage banks, and regulators consulting on how to license a process and a library rather than a product. But the gap between "this saved a patient" and "this is an approved medicine" remains regulatory, not scientific.

Why it matters for students and researchers

India has one of the world's heaviest burdens of drug-resistant infection, driven by high infection rates, widespread over-the-counter antibiotic availability and crowded hospital settings. That makes the alternatives to antibiotics a practical national question rather than a speculative one, and it makes India a place where phage work has unusually direct relevance — abundant clinical isolates, real need, and the fermentation and biologics manufacturing base to produce phage preparations to standard.

The research questions are open and unusually varied: rapid matching of phage to isolate, host-range prediction from genome sequence, engineering phages to broaden or retarget their binding, formulating them to survive storage and delivery, and understanding phage–antibiotic combinations, which are frequently synergistic in ways that are not yet well predicted. Underlying all of it is the conceptual shift worth teaching — that treating an infection with a living, evolving agent is not a better antibiotic but a different kind of intervention, and the discipline it demands is ecological rather than purely chemical.

Frequently asked questions

What is phage therapy?

It is the treatment of bacterial infection using bacteriophages — viruses that infect and destroy bacteria. The phage is matched to the bacterial strain causing the infection, replicates at the site, and is cleared once the bacteria are gone.

Why is phage therapy not widely available?

Mainly for regulatory rather than scientific reasons. Approval systems are designed for fixed chemical products, whereas a phage treatment is often assembled for an individual patient and changes with the bacterial strain, which does not fit standard trial and licensing frameworks.

How is a phage different from an antibiotic?

An antibiotic is a chemical that affects broad groups of bacteria and is consumed as it acts. A phage is a self-replicating organism specific to particular strains, so it spares the rest of the microbiome and increases in quantity where the infection is.

Can bacteria become resistant to phages?

Yes, often within days, by altering the surface receptor the phage binds to or by using defences such as restriction enzymes and CRISPR. Cocktails of several phages are used to slow this, and resistance can sometimes be steered so that escaping the phage weakens the bacterium in other ways.

Are phages safe for humans?

Phages infect only bacteria and cannot enter human cells, and the approach has a long clinical record in some countries. The practical safety concerns are purification — removing bacterial debris and endotoxin from preparations — and ensuring only strictly lytic phages carrying no harmful genes are used.