The City That Reports Its Illness Through Sewage

In February 2022, an unnamed person somewhere in London shed poliovirus into a toilet. There was no clinical alarm. No patient with paralysis led investigators to a household or a travel history. The signal travelled through pipes to the Beckton sewage treatment works, where routine sampling turned a city's waste into evidence. Over the next five months, laboratories found 118 genetically linked poliovirus isolates in 21 of 52 sequential samples. Sewage had revealed transmission that medicine had not yet seen [1].

The episode captured the promise of wastewater surveillance in its purest form. A city reports illness through the material it discards. People do not have to book an appointment, afford a test or recognise a symptom. A sample taken from a treatment works pools traces from a large population, including people whose infections are mild or silent. For poliovirus, silence is common: paralysis occurs in only a small fraction of infections. Waiting for a paralysed patient means waiting for a rare and devastating outcome.

Wastewater surveillance is the systematic testing of sewage or other human-impacted water for signs of infection or health-related exposures. It is not new. Researchers isolated poliovirus from sewage in the twentieth century, and environmental sampling became a central tool in global eradication work. What changed during COVID-19 was scale. Laboratories around the world adapted the approach to track SARS-CoV-2, the virus that causes COVID-19, and public-health officials learned how to turn repeated measurements into a trend.

The process begins with a catchment, often called a sewershed: the area whose wastewater flows to a sampling point. A sampler may collect small amounts over 24 hours, creating a composite that smooths some hourly variation. In the laboratory, technicians concentrate the material because the biological target is dilute. For an RNA virus, they may use reverse-transcription quantitative polymerase chain reaction, or RT-qPCR. This method converts viral RNA into DNA, copies selected genetic targets and measures how much target material was present.

A positive signal is not the same as finding live, infectious virus. Genetic fragments can persist after the virus has lost the ability to infect cells, and people may continue shedding material after their most infectious period. Laboratories can sometimes grow a virus in cell culture, which provides different evidence, or sequence its genome to identify a lineage. Each technique answers a narrower question. Detection shows that someone contributing to that sewage shed the target. It does not show who, where within a large catchment or whether that person was ill.

London's polio investigation used several of these layers. The samples were concentrated, virus was isolated in cell culture and genomes were characterised. The isolates were related to the weakened type 2 strain once used in oral polio vaccine. As the virus passed between people, it accumulated mutations and recombined with another enterovirus. Twenty isolates met the genetic definition of vaccine-derived poliovirus, and the sequences showed that the detections belonged to one transmission lineage [1].

The phrase vaccine-derived requires context. Oral polio vaccine contains weakened live virus and has been enormously important in controlling polio because it generates strong immunity in the gut. Vaccinated people can briefly shed the weakened virus. In places with very low immunisation coverage, that virus can occasionally circulate for long enough to change genetically and recover the capacity to cause paralysis. The United Kingdom switched from oral to inactivated polio vaccine in 2004. The injected vaccine cannot cause vaccine-derived virus and gives excellent protection against paralysis, but it is less effective at stopping gut infection and shedding [1].

Genomes allowed investigators to go beyond a single surprising sample. The related sequences, recovered repeatedly from February to July, supported community transmission and ruled against one recent vaccine recipient passing through London as the explanation. Expanded sampling helped localise circulation to several boroughs in north and east London. Public-health teams intensified surveillance and offered an additional inactivated polio vaccine dose to children aged one to nine. No associated paralytic case was identified in the United Kingdom during the response [1].

That last fact is the achievement and the interpretive challenge. Wastewater found risk before a severe sentinel event. It cannot prove which individual cases were prevented by the vaccination campaign, because there was no untreated London for comparison. The circulation might have ended for several reasons. The appropriate claim is narrower: environmental surveillance detected importation and sustained transmission, supplied genomic evidence and enabled a response before recognised paralysis.

The network has since grown. As of July 2026, untreated sewage in England was being collected monthly from 28 treatment sites for routine polio surveillance, with sampling designed to cover places at higher risk and provide geographical spread [2]. A treatment works can represent hundreds of thousands of people with a single specimen. That economy is compelling, especially for a rare infection. Yet coverage follows pipes. People using septic tanks, informal sanitation or sewer systems outside the selected catchments are absent from the sample.

COVID-19 showed what repeated sewage measurements could do during widespread transmission. An English analysis used 45 large wastewater sites covering about 31 per cent of the population. After accounting for factors such as flow, the researchers estimated community prevalence close to estimates from representative infection surveys. Wastewater signals appeared four to five days earlier than clinical testing data, although they coincided with the survey-based estimate of infection prevalence [3]. In other words, sewage led reported symptomatic testing, not necessarily infection itself.

That distinction explains why early warning is a relational concept. Wastewater is early compared with a system that waits for symptoms, care seeking and reporting. It may be contemporaneous with infection, because shedding begins around the same biological period. Its advantage expands when home testing declines or reporting becomes inconsistent. It shrinks when clinical surveillance is fast and comprehensive. A sewer does not predict the future. It offers a different view of the present.

Turning concentration into case numbers is difficult. Rain can dilute a combined sewer. Industrial discharge changes chemistry. Travel time, temperature and substances in wastewater can degrade genetic material. People vary greatly in how much virus they shed and for how long. The number of people in a catchment changes with commuting and tourism. Researchers can normalise measurements using flow and biological markers of human waste, then fit statistical models, but every conversion carries uncertainty [3]. Trends within the same well-characterised site are often more defensible than comparisons between raw numbers from different cities.

Quality control begins before any model. Field teams need consistent sampling times and containers; laboratories need negative controls to reveal contamination and positive controls to show that an assay worked. Results may be adjusted for recovery efficiency, the fraction of a known target that survives the method. Repeated samples help distinguish a genuine rise from one noisy bottle. This unglamorous standardisation determines whether a line on a public dashboard represents viral change or a change in how water was collected and processed.

Sequencing introduces another level of inference. A sewage sample mixes genomes from many people, producing a molecular crowd rather than one clean sequence. Computational methods use combinations of mutations to estimate the proportions of known variants. During the pandemic, researchers showed that this approach could detect SARS-CoV-2 variants and reveal cryptic transmission that clinical sequencing had missed [4]. It can also be confounded by low concentrations, uneven genome coverage and lineages that share mutations. A new signal needs confirmation before it earns a dramatic name.

The best surveillance systems therefore combine streams. Wastewater can show that viral material is rising across a catchment. Clinical tests reveal symptoms and patient characteristics. Hospital admissions measure severe burden after a delay. Genomic data help distinguish lineages, while outbreak investigation reconstructs contacts and settings. The World Health Organization describes wastewater as part of multimodal surveillance, meaning a system in which different methods compensate for one another's blind spots [5].

The list of possible targets extends beyond polio and SARS-CoV-2. Sewage can contain influenza, respiratory syncytial virus, norovirus, antimicrobial-resistance genes and many other biological traces. Technical detectability is only the first test. A useful programme also needs a defined public-health question, a reliable sampling and laboratory method, an interpretation that decision-makers understand and an action that could follow. WHO guidance asks programmes to prioritise targets and resist assuming that everything measurable deserves permanent monitoring [5].

That discipline matters because surveillance has costs. Samples need collection through weekends and bad weather. Assays require quality controls, trained staff and reagents. Data systems must join laboratory results to catchment maps and population estimates. A pilot may demonstrate that a pathogen is detectable without showing that the result changes a decision. Sustainable monitoring needs an answer to a practical question: what would public health do differently if the line went up tomorrow?

There is an ethical question too. City-scale wastewater is usually described as anonymous because it pools material from many people and cannot identify an individual. Sampling a prison wing, school residence or small neighbourhood can make a group recognisable even when no person is named. A signal associated with drug use or a stigmatised infection could harm that group if presented carelessly. A 2024 ethics analysis argued that governance should include transparency, oversight and enforceable limits on how wastewater data are used [6]. Resolution should be proportional to a legitimate health purpose.

Equity cuts in both directions. Wastewater can reduce dependence on individual testing and health-care access, making infection visible where clinical data are sparse. At the same time, sewer-based programmes systematically miss communities without central sanitation. Urban treatment works are easier to sample than remote settlements. A map of monitored sewage can therefore look comprehensive while reproducing an infrastructure map. WHO's framework includes non-sewered settings, where environmental waters affected by human waste may sometimes be sampled, but the methods and interpretation must fit local conditions [5].

Public communication determines whether a signal becomes knowledge or noise. A tenfold change in viral concentration does not mean ten times as many sick people. A nondetection does not prove absence, because the virus may be below the assay's limit. One anomalous result may reflect laboratory or sewer conditions. Reports are clearest when they show sustained trends, explain the population represented and place sewage beside other indicators. The method earns trust by displaying its uncertainty openly.

London's poliovirus episode remains powerful because the chain from sample to action was unusually clear. Repeated detections established circulation. Genomes linked the isolates. Expanded sites narrowed the area of concern. Vaccination strengthened protection against paralysis while clinical surveillance remained alert. The sewage did not diagnose a Londoner, reconstruct every transmission event or count infections. It told the city something urgent that no other system had yet managed to say.

A sewer is an imperfect observatory. Its view is blurred by water, weather, infrastructure and human biology. It also includes people whom conventional surveillance routinely overlooks. Wastewater complements medicine and has proved more durable than a pandemic novelty. It is a population sample waiting beneath the street. In 2022, London listened before the rare sound of paralysis. The enduring question for every city is whether it has built the laboratory capacity, analytical judgment, governance and public trust to understand what its wastewater says next.