Sewage Microbes Reveal Antibiotic Resistance Signals That Trail Prescriptions by Months

Every flush tells a story. For years, scientists have mined sewage for genetic traces of pathogens, turning wastewater treatment plants into community-wide health sentinels. The approach, known as wastewater-based surveillance, proved its worth during the COVID-19 pandemic, when measurements of viral RNA in influent tracked infection waves days before clinical case counts caught up. Extending the same logic to antibiotic resistance, however, has proven far more difficult. Resistance genes are not simply shed by sick individuals; they are carried by entire bacterial communities that live, reproduce, and swap DNA inside both human guts and the sewer pipes themselves. A new study published in Nature Communications now untangles this complexity, showing that the antibiotic resistance genes detectable in sewage do mirror real patterns of clinical antibiotic use in the surrounding community, but only when researchers account for the ecology of the microbes that carry them.

The research team, led by C. L. Brown of Virginia Tech’s Department of Civil and Environmental Engineering together with collaborators at Aalborg University, Carilion Roanoke Memorial Hospital, and the U.S. Environmental Protection Agency, assembled an unusually rich pair of datasets. On the clinical side, they obtained five years of outpatient antibiotic prescription records for the community served by a single wastewater treatment plant. On the environmental side, they conducted a yearlong metagenomic survey of that plant’s influent, sampling sewage twice a week and sequencing the full complement of microbial genomes present. This pairing of granular prescription data with high-frequency genomic monitoring is rare, and it allowed the researchers to ask a question that has long frustrated the field: what does the abundance of an antibiotic resistance gene in sewage actually mean, epidemiologically speaking?

The central challenge is that sewage is not a passive snapshot of what humans excrete. Once resistance genes and the bacteria bearing them enter the sewer system, they enter a dynamic microbial ecosystem. Bacteria from different hosts mix, nutrients and temperature fluctuate, and organisms that thrive under sewer conditions can bloom regardless of what the community above has been prescribed. A spike in a resistance gene could therefore reflect a genuine rise in resistant infections, or simply an environmental population of bacteria expanding on its own. Without disentangling these two sources of signal, public health officials cannot know whether a wastewater measurement of resistance is a meaningful indicator of clinical reality or an ecological artifact.

The answer, it turns out, depends on which bacteria carry the genes. The researchers found that antibiotic resistance genes associated with Enterobacteriaceae, the family of bacteria that includes Escherichia coli, Salmonella, and Klebsiella, tracked closely and contemporaneously with community antibiotic use. When prescriptions for a given antibiotic class rose, the corresponding resistance genes in sewage rose with them, with no discernible delay. This tight coupling makes intuitive sense: Enterobacteriaceae are dominant members of the human gut microbiota, so the resistance genes they carry are shed directly into wastewater by the very patients taking the drugs. The signal passes from prescription to gut to sewer with minimal ecological distortion in between.

Genes associated with Pseudomonadaceae told a strikingly different story. Pseudomonas species are environmental organisms, opportunistic pathogens that colonize soils, water, biofilms, and hospital plumbing rather than dominating the healthy human gut. When the team analyzed resistance genes linked to this family, they found lagged correlations of one to three months between sewage signals and antibiotic prescriptions. Rather than reflecting direct shedding by treated patients, these lagged patterns suggest that Pseudomonas populations in the sewer system respond to antibiotic residues and selective pressures over time, with their resistance gene profiles shifting weeks after the community’s prescribing patterns change. The host organism’s ecology, in other words, imprints a characteristic temporal signature on the resistance signal.

To test this interpretation rigorously, the researchers turned to nested statistical modeling. They built models that accounted for sewage characteristics such as flow and composition, as well as seasonal variation, and then asked whether adding antibiotic prescription data improved the explanation of the genomic observations. When they examined metagenome-assembled genomes, or MAGs, reconstructed from the Pseudomonas populations in the sewage, the models linked the dynamics of these organisms to lagged use of sulfamethoxazole/trimethoprim, a commonly prescribed antibiotic combination, even after controlling for the confounding influence of sewage conditions and seasonality. The persistence of this association under such controls strengthens the case that the lagged relationship is genuine rather than a statistical mirage produced by weather, holidays, or treatment plant operations.

The methodological machinery behind these findings deserves attention. Metagenomic sequencing of wastewater yields billions of short DNA reads drawn from thousands of species. The team used genome-resolved approaches to reconstruct MAGs, which are composite genomes representing individual microbial populations, and then tracked the abundance of resistance genes both in aggregate and within specific genomic contexts. This distinction matters because a resistance gene found on a plasmid circulating among many species carries different epidemiological information than the same gene embedded in the chromosome of a single gut bacterium. By tying genes to their host organisms, the researchers transformed a noisy aggregate signal into a set of interpretable indicators, each with its own expected relationship to clinical practice.

The implications for public health surveillance are substantial. Antimicrobial resistance is responsible for well over a million deaths annually worldwide, and stewardship programs that aim to reduce inappropriate antibiotic prescribing need timely feedback on whether their interventions work. Traditional surveillance relies on clinical isolates, which capture only patients who seek care and undergo testing, leaving large blind spots in the community. Wastewater-based surveillance, by contrast, samples everyone who uses the sewer system, including people with untreated infections and asymptomatic carriers. The new study demonstrates that this population-level window can capture epidemiologically relevant patterns of resistance, but it also delivers a caution: naive interpretation of aggregate resistance gene abundances, without attention to the microbial hosts carrying them, could mislead. A gene blooming in environmental Pseudomonas is not the same event as a gene surging in gut Enterobacteriaceae, and the two require different responses.

The temporal lags themselves may prove to be a feature rather than a complication. If resistance signals in environmental organisms consistently trail prescribing patterns by one to three months, then wastewater data could serve as a leading or trailing indicator depending on the gene family examined, offering stewardship programs a way to evaluate whether changes in prescribing propagate through the community’s resistance reservoir. The five-year prescription dataset provided the temporal depth needed to detect these relationships, and the twice-weekly sampling cadence provided the resolution to resolve lags on the scale of weeks. Few surveillance programs currently operate at this intensity, but the study suggests that the investment yields information that coarser monitoring cannot.

Limitations remain. The study followed a single wastewater treatment plant serving a single community, and the specific relationships between prescriptions and sewage signals may vary with local sewer geometry, climate, industrial inputs, and population demographics. The authors also note that the research was not performed or funded by the EPA and does not represent the agency’s policies. Scaling the approach will require testing whether the host-linked lag structure holds across different plants and regions, and whether prescription records of comparable quality can be obtained elsewhere. Still, the core message is clear and actionable: wastewater can indeed mirror clinical antibiotic use, but the mirror is curved by microbial ecology, and reading it correctly requires knowing whose genes are being measured. As antimicrobial resistance continues its global rise, tools that translate the genetic chatter of sewage into epidemiologically meaningful intelligence are likely to become an increasingly important part of the public health arsenal.

Subject of Research: Wastewater-based surveillance of antibiotic resistance genes and their relationship to community antibiotic prescribing patterns

Article Title: Resistance indicators in sewage microbiota mirror clinical antibiotic use with temporal lag linked to host

Article References: Brown, C. L., Rumi, M., McDaniel, L., Maile-Moskowitz, A., Sein, J., Nguyen, L., Choi, M., Hindi, F., Mullet, J. I., Emon, M., Moumi, N., Blair, M. F., Davis, B. C., Rao, J., Baffoe-Bonnie, A., Vikesland, P., Pruden, A., & Zhang, L. (2026). Resistance indicators in sewage microbiota mirror clinical antibiotic use with temporal lag linked to host. Nature Communications. https://doi.org/10.1038/s41467-026-78142-3

Image Credits: AI Generated

DOI: 10.1038/s41467-026-78142-3

Keywords: antimicrobial resistance, wastewater-based surveillance, antibiotic resistance genes, metagenomics, Enterobacteriaceae, Pseudomonas, antibiotic stewardship, sewer microbiology, public health surveillance, sulfamethoxazole-trimethoprim, microbial ecology, epidemiology