17.12 — Antibiotics and Resistance

Alexander Fleming warned about resistance in his 1945 Nobel lecture, before penicillin had been in widespread use for a decade.

"The time may come when penicillin can be bought by anyone in the shops. Then there is the danger that the ignorant man may easily underdose himself and by exposing his microbes to non-lethal quantities of the drug make them resistant."

He was right, and the mechanism he described is only one of several.

Resistance is not a future threat. It is present, measurable, and it kills an estimated 1.3 million people a year directly, with a further several million deaths associated with it.

And it is one of the few medical problems where what individuals do genuinely aggregates into a global outcome.

Where antibiotics came from

Almost all of them come from soil microorganisms (Chapter 3.7).

Bacteria and fungi have been fighting each other chemically for billions of years, and we largely took their weapons.

Which has an implication people find surprising: resistance genes are ancient. They have been found in 30,000-year-old permafrost and in bacteria isolated from caves sealed for millions of years. Resistance did not evolve in response to medicine; it was already there, and medicine selected for it.

The classes, and each attacks something bacteria have and we do not:

Cell wall synthesis — penicillins, cephalosporins, carbapenems, vancomycin. We have no cell wall, which is why these are among the safest antibiotics.

Protein synthesis — macrolides, tetracyclines, aminoglycosides, clindamycin, linezolid. Targeting the 70S bacterial ribosome rather than our 80S (Chapter 1.5).

DNA replication — fluoroquinolones, targeting bacterial topoisomerase (Chapter 2.2).

Folate synthesis — trimethoprim and sulfonamides. Bacteria must make folate; we absorb it from food, which is the difference exploited.

Cell membrane — colistin, polymyxins. Toxic, and reserved for organisms resistant to everything else.

How resistance happens

Four mechanisms.

Destroying the drug — beta-lactamases, which cleave the ring in penicillins. Extended-spectrum beta-lactamases and carbapenemases are the versions that defeat successive generations of drug.

Pumping it out — efflux pumps, which frequently confer resistance to several classes at once.

Changing the target — MRSA works this way: an altered penicillin-binding protein that the drug does not fit.

Preventing entry — altering the outer membrane porins, which is one reason Gram-negative organisms are harder to treat (Chapter 3.7).

And the genes move.

Vertically, to daughter cells — and horizontally, between bacteria of different species, on plasmids (Chapter 3.7).

Which is why resistance spreads far faster than ordinary evolution would allow. A resistance gene arising in a harmless gut organism can be transferred to a pathogen, and plasmids frequently carry several resistance genes together, so exposure to one antibiotic selects for resistance to others.

Why it is accelerating

Overuse in humans — and the largest single source is antibiotics for viral illness. Coughs, colds, sore throats, bronchitis (Chapter 17.5).

Also: too long a course, too broad a spectrum, and use without diagnosis.

Over-the-counter availability without prescription in many countries.

Agricultural use, which exceeds human use in many countries. Antibiotics given for growth promotion and routine prophylaxis in intensive farming, rather than to treat sick animals.

The European Union banned growth promotion in 2006 and several countries have followed, and the evidence that agricultural use contributes to human resistance is now strong enough that this is no longer seriously contested.

Poor infection control, which allows resistant organisms to spread.

Global travel, which moves them.

And a dry antibiotic pipeline. Most large pharmaceutical companies have left the field.

The economics explain it, and they are perverse. A new antibiotic is used for a few days per patient, is held in reserve to preserve its effectiveness, and is expected to be cheap. A drug for a chronic condition is taken daily for decades. So the drug the world most needs is the one least worth developing, and various push-and-pull incentive schemes are being trialled to correct it.

The organisms to know

MRSA — meticillin-resistant Staphylococcus aureus. The same organism with a changed target. Rates in hospitals have fallen substantially in several countries following intensive infection control, which shows it is not inevitable.

VRE — vancomycin-resistant enterococci.

ESBL and carbapenemase-producing Enterobacteriaceae — Gram-negative organisms resistant to most beta-lactams. Carbapenem resistance is the more serious, because carbapenems were the reliable last resort.

And colistin resistance is the one that alarms specialists, because colistin is the drug used when carbapenems fail. A transferable colistin resistance gene, mcr-1, was identified in 2015 and has since spread globally.

Multidrug-resistant tuberculosis (Chapter 17.8).

Drug-resistant gonorrhoea (Chapter 15.13) — the organism has defeated every class used against it in sequence, and it is now on the WHO's high-priority list.

Candida auris (Chapter 17.3).

And the WHO priority list groups organisms by urgency, which is a useful thing to know exists — it directs research funding and surveillance.

What actually slows it

Antimicrobial stewardship, and the components are specific.

Do not prescribe for viral illness. The largest single lever.

The right drug — narrow spectrum where the organism is known.

The right dose — and Fleming's warning was about underdosing. Too low a dose selects for resistance without clearing the infection.

The right duration.

And here is where advice has genuinely changed. "Always finish the course" has been substantially revised.

The reasoning behind it was that stopping early leaves resistant survivors — which is true for a small number of infections, most importantly tuberculosis.

For most common infections, the evidence supports shorter courses, and unnecessarily long courses select for resistance in the body's other bacteria without benefit.

Trials have shown non-inferiority for 5 days rather than 10 in community pneumonia, 3 days for uncomplicated urinary infection, and shorter courses across several other conditions.

So the current advice is: take the course you were prescribed, and expect that course to be shorter than it used to be — and for some conditions, guidance now explicitly permits stopping when symptoms resolve.

IV to oral switch as soon as the person can swallow.

Rapid diagnostics — molecular tests giving an organism and its resistance genes in hours rather than days (Chapter 17.1).

Delayed prescribing — a prescription to be used only if not improving, which reduces antibiotic use substantially without worse outcomes (Chapter 17.5).

Infection control — because preventing an infection prevents the antibiotic course entirely.

And vaccination, which is the most under-appreciated antibiotic-sparing measure there is. Pneumococcal conjugate vaccination reduced antibiotic prescribing measurably by preventing the infections that would have been treated.

What individuals can do

Because this is one of the rare cases where it genuinely matters.

Do not ask for antibiotics for a cold, flu, or most sore throats and coughs.

And do not interpret a doctor declining them as being fobbed off. Declining an antibiotic for a viral illness is the correct treatment, and the conversation is easier if both parties know that.

Do not use leftover antibiotics, or someone else's.

Take them as prescribed.

Get vaccinated.

Wash your hands.

And in countries where antibiotics are sold without prescription, self-treating is a substantial driver — with the added problem that the drug is frequently the wrong one, at the wrong dose, for the wrong duration.

Where hope lies

And the picture is not uniformly bleak.

New classes are emerging. Teixobactin, discovered by culturing previously unculturable soil bacteria in situ, was the first genuinely new class in decades, and the technique that found it — the iChip — opens up an enormous unexplored reservoir, since around 99 percent of soil bacteria had never been grown in a laboratory.

Bacteriophage therapy — using viruses that infect bacteria. Developed in the Soviet Union and largely ignored in the West, and now being revisited seriously. Phages are exquisitely specific, self-amplifying at the site of infection, and there have been striking individual case successes — including a patient with a multidrug-resistant infection after a lung transplant. Trials are underway; it is promising rather than established.

Antimicrobial peptides, based on the body's own defences (Chapter 13.1).

Anti-virulence approaches — disarming rather than killing bacteria, which imposes less selection pressure.

Microbiome-based approaches — faecal transplantation for C. difficile is the proven example (Chapter 9.6).

Better diagnostics, which reduce empirical prescribing.

And the important structural point: resistance is at least partly reversible. When the selection pressure is removed, susceptible strains — which are frequently fitter without the resistance machinery — recover ground. Several countries have reduced MRSA rates substantially through concerted action.

Which means this is a problem of collective behaviour rather than an inexorable natural process, and behaviour can change.

What the next page fixes

Chapter 17.13 closes this Part with what happens when an infection spreads faster than it can be contained — epidemics and pandemics, what determines their course, and what the last one taught.