We Keep Treating Resistant Infections. We Rarely Ask Where the Resistance Came From.

Every clinician who has cultured a resistant organism in the last few years has felt the same small chill: the list of drugs that will still work is shorter than it used to be. We manage it the way we manage most things in medicine, one patient, one culture, one narrowed regimen at a time. What we rarely do, in the exam room or at the bedside, is ask where that resistance is coming from.

Most of it is not coming from a hospital. About 70% of the antibiotics that matter most to human medicine in the U.S. are sold for use in food animals (species raised for human consumption), and most of that is not treating sick livestock. It is keeping healthy animals alive in the crowded, stressful confinement that industrial meat production depends on. Antibiotics have become a structural input to that system, as ordinary as feed, the thing that makes it possible to pack thousands of animals together and have most of them survive until slaughter. We prescribe carefully and still inherit the consequences of a use pattern we have almost no visibility into and no say over.

The trajectory should concern anyone who treats infection for a living. New projections show livestock antibiotic use, already the largest single use on the planet, is on track to rise by nearly a third by 2040. Drug-resistant infections already kill an estimated 35,000 Americans annually, and a Lancet forecast projects roughly 39 million deaths globally from resistant infections between now and 2050. These numbers show up in our own mortality and morbidity conferences far more than most of us track them back to their source. Resistant organisms and the mobile genes that carry resistance between them do not stay on the farm. They travel in meat, water, dust, on workers, and in manure, and they show up, eventually, in the culture results we read.

We have direct evidence that the farm-to-clinic pathway is not theoretical, and that closing it works. Colistin is a drug most of us hold in reserve for infections that have already defeated everything else. When it was used as a cheap growth promoter on Chinese pig farms, a transferable resistance gene called mcr-1 emerged and circled the globe within a couple of years, showing up in patient isolates on multiple continents. After China banned colistin as a feed additive in 2017, resistance to it measurably fell in animals and in people. That is not a modeling exercise. It is a natural experiment with a control arm, and the result tells us something we do not say often enough to patients: a drug we still needed became more usable again because of a change in food-animal policy, not a change in clinical practice.

The World Health Organization has urged farmers worldwide to stop routine, preventive antibiotic use in animals that are not sick; this is a targeted ask, not a call to eliminate veterinary antibiotics altogether. Genuinely sick animals still need treatment. But most current use isn’t that, and stewardship efforts aimed only at prescribing behavior in clinics and hospitals are managing a downstream symptom of an upstream practice we have no clinical control over.

We do have one lever most of us underuse: what we tell patients about diet, and what our own institutions serve. Reducing demand for cheap, confinement-raised animal protein lowers the pressure on the system that requires routine antibiotic use to keep it running. It is also one of the few stewardship interventions available to us, and does not require waiting on a regulatory fight.

When we counsel a patient toward a more plant-forward diet for cardiometabolic reasons, we are also quietly making a small dent in the demand driving this problem. It costs nothing extra to mention that connection. Hospital food-service contracts and formulary-adjacent purchasing decisions are a second, underused lever inside our own institutions, one that stewardship committees rarely put on their agenda alongside dosing protocols and restricted-drug lists.

None of this replaces farm-level regulation, and I am not arguing that it should. The colistin case shows targeted policy works, and clinicians should be pushing for it through our professional societies, the same way many of us already push for vaccine policy or opioid-prescribing reform. But regulation is fighting a demand curve it did not create, and diet is the one piece of that curve we can actually influence from an exam room.

We built an extraordinarily cheap meat supply chain in part by spending down a shared resource that took nearly a century to build and that no amount of stewardship can quickly rebuild: the effectiveness of the antibiotics we still have. We will keep paying that bill in resistant cultures unless we start treating what happens upstream of the farm gate as part of our own stewardship responsibility.