We came together for EDAR8 in Brisbane at the moment the environment stopped being optional. Within months, European law will require antimicrobial resistance monitoring in urban wastewater, and a revised pharmaceutical regimen will extend environmental risk assessment across the whole life cycle of an antimicrobial. An updated Global Action Plan carries environmental considerations through its text for the first time.

The methods, the evidence and the institutions to deliver on the plan are still being built. For five days, 430 delegates from more than 40 countries worked together on building the science, evidence and policy platforms needed to fill that gap.

“The environmental dimension of tackling AMR is not optional. It must be at the very core of our response,” UNEP Executive Director Inger Andersen told the opening. It has taken 98 years since Fleming's penicillin paper for environmental AMR dimensions to have global institutional weight behind them. UNEP joined the other global agencies working on AMR only four years ago.

The science opened with three clinical cases: a fit 75-year-old who broke her ankle on holiday and came home facing potential amputation, because colistin was the only drug her infection was still susceptible to; a three-year-old with no travel or antibiotic history whose urinary tract infection had no oral treatment options; and a multidrug-resistant Shigella outbreak at a rural music festival that forced hospitals to reach straight for a last-line drug. None of these people did anything wrong.

Europe will soon require AMR monitoring in wastewater from communities of more than 100,000 people, and environmental risk assessment across an antimicrobial’s entire life cycle, including manufacturing and waste management, wherever in the world it is made. The science here is complex and still moving but we need the insights that come from iteration, rather than waiting till current drugs fail. Delegates heard that the wastewater monitoring proposal has narrowed through negotiation from metagenomics to two species and six genes — without a clear statement of what the surveillance is for. Further research and development to support that regulatory science is already underway.

For the first time the program carried dedicated sessions pairing regulators, industry and scientists, alongside a workshop on risk assessment across the antimicrobial life cycle that will feed a collaborative peer-reviewed publication. Industry arrived with a decade’s head start: the AMR Industry Alliance set initial manufacturing discharge limits 10 years ago and has since built an independent certification scheme covering 150 products. They’re actively supporting the call for new science to fill knowledge gaps and support effective evidence-based regulation. Researchers noted separately that current European Medicines Agency guidelines do not consider selection risk at all, and that only 24 usable studies exist of the concentrations at which resistance is selected.

Carbapenems, a crucial class of antibiotics for human health, have never been licensed for food-producing animals in Europe, yet carbapenemase-producing organisms have now been reported in animals or food in around 16 countries. Worldwide carbapenem resistance is increasingly observed and a likely route runs from humans, through the environment, into livestock — and delegates heard the regulatory response has been minimal, because no legislation compels one.

The largest strand of the program spanned sewage networks in Cyprus, post-pandemic Japan, informal settlements in South Africa and national resistome analysis in the United States. A decade of Dutch sampling has found carbapenemase producers in around 90 per cent of treatment plants, in a country where fewer than one person in 10,000 carries these bacteria. But we were also reminded of the limits: wastewater will never give patient-level information or guide an individual's treatment. Its value is population-level intelligence on rare and emerging resistance and the complexity can’t be underestimated.

Different bioinformatic pipelines produce markedly different resistome estimates from identical data. A European consortium spent two and a half years negotiating what 20 countries should measure first. Two decades of food sector experience points to a blunter answer: we need to start somewhere. Comparability across 28 legal systems came only from regulating the sampling, the methods and the reference laboratories. Voluntary genotyping, offered for five years, was taken up consistently by 10 member states out of 28.

European work on reclaimed water for irrigation and food processing found bacterial numbers fall after treatment but resistance genes remain widespread and persistent across multiple scenarios. Commercial monitoring of one of Europe’s best treatment plants showed the same. Separate work found clinically relevant resistant E. coli below the limit of detection in wastewater-irrigated produce, which proliferated once incubated under conditions mimicking the human gut.

A dedicated session examined chemical mixtures in treatment plants, on-site selection in hospital sewers driven by ciprofloxacin and beta-lactams, methods for prioritising which of thousands of aquatic pollutants to test, and evidence that antibiotic transformation products exert selective pressure comparable to their parent compounds. A masterclass on AMR chemistry for non-chemists addressed what its convenors called a critical but underrepresented discipline in the field.

Plant surfaces are gene transfer hotspots by nature and agricultural inputs may make them worse: agrochemicals can trigger the bacterial stress responses that mobilise resistance elements and metals used on some crops co-select for it. Sampling from Australian soils to Antarctica has recovered thousands of gene cassettes whose function is unknown — genetic dark matter that contains resistance we cannot yet see.

A European multi-agency review found more than 119,000 tonnes of azoles had been sold as crop protection across the EU in 12 years, much more than human therapeutic consumption. Critically, it located the hotspots for resistance selection in stockpiled agricultural waste and freshly cut wood rather than in the sprayed crop, because the fungus of concern is not the spray’s target and grows on decaying plant material. The review’s recommendations centre on waste management and good practice, not prohibition. A panel on fungicide resistance discussed a study of Australian vineyard regions under the highest disease pressure that had lower resistance — evidence that better grower advice and stewardship can hold it down.

A survey of Thai shrimp farms — 509 samples, more than 36,000 assays — found feed and shrimp tissue almost clean but resistance genes in water nearly everywhere sampled. Three exceeded the global baseline, including a mobilised colistin resistance gene. It’s a reminder that the food production environment is critical to the quality of the output – it’s not just a catchment problem or a farm problem, it’s a whole of system issue.

Europe’s two summary indicators for food-producing animals are both moving the right way. The proportion of fully susceptible indicator E. coli has risen significantly in around 45 per cent of reporting countries, and the prevalence of ESBL-producing E. coli has fallen in roughly half of them. Both are attributed to a sustained push to reduce antimicrobial use in food production animals — evidence that stewardship works.

We don’t really know what is happening worldwide. More than 215,000 AMR papers appeared between 2003 and 2023 and reading them all would take 19 years. Machine analysis of that literature shows hydrological-cycle research concentrated on treatment plants rather than natural environments, dominated by high-income countries, and sitting at the margin of global AMR funding. A separate speaker made the same point from the policy side: national action plan spending remains heavily focused on humans and animals, with crop protection and environment the blind spots. Globally coordinated action is required to make the most of constrained resources to understand the environmental dimensions of AMR.

The conference opened with reference to the original One Health philosophy, one that has existed since time immemorial. Drawing on two principles articulated by Kombumerri elder Aunty Mary Graham — that the land is the law, and that you are not alone in the world — Associate Professor Veronica Matthews framed AMR as a relational problem as much as a technical one. She left delegates with three questions.

What responsibility do we have:

• To the ecosystems receiving antimicrobial pollution?
• To future generations who will inherit the consequences of today’s decisions?
• To act before relationships become damaged beyond repair?