Communities around the world have embraced the trend of wastewater monitoring to determine whether COVID-19 is circulating. From the start, the approach promised an inexpensive way to identify an outbreak early, giving public health authorities time to take action. Viral particles are shed in feces even before infected people show symptoms, and samples drawn from the sewer are a snapshot of the health of everyone served by that system. Researchers hoped regularly testing those samples would provide officials data to better allocate resources.
Now, some 18 months into the pandemic, researchers have a better idea of what sewage surveillance can and can’t tell us about a local outbreak—and how the emergence of variants complicates the measurement and interpretation of wastewater data.
The clearest takeaway is that widespread efforts to measure the virus in wastewater have proven the method to be a reliable way of monitoring for COVID-19 outbreaks. “There’s plenty of evidence that we have early-warning capacity to see what’s going to happen with respect to recorded infections, morbidity, and mortality,” says Rolf Halden, an environmental engineer at Arizona State University and cofounder of AquaVitas, a wastewater-based epidemiology start-up that performed the first nationwide study in over 100 US cities for the Department of Health and Human Services.
In the early days of the pandemic, people hoped that wastewater measurements could be a proxy for widespread clinical testing and reveal the number of COVID-19 infections in a community. The reality is that extracting case numbers from wastewater data continues to be a challenge, says Amy E. Kirby, program lead for the National Wastewater Surveillance System (NWSS) at the US Centers for Disease Control and Prevention. Researchers can calculate case numbers from sewage samples, but their wide confidence intervals make it hard to say anything definitive. “Yeah, you get a number, but the actual true range is so big that it’s encompassing any kind of changes over time,” Kirby says.
Part of the disconnect is due to a fundamental difference between clinical and wastewater data. The results from individual tests, such as nasal swabs or saliva samples, “go back to a single person, and wastewater surveillance never will,” Kirby says. Wastewater data are a community metric, she says. “It has a lot of noise in it that we understand, mostly, and can account for. But we have to think about how we use that data a little differently because it’s not going to ever be about a single person’s activity.”
The founders of Biobot Analytics, which focuses on acquiring public health data from sewage, say the pandemic has caused their thinking about the role of wastewater monitoring to evolve. “When we started, we thought that wastewater epidemiology data needed to be better than clinical data for it to get traction and get adopted,” says Mariana Matus, Biobot’s CEO. Biobot was part of another pilot study funded by the US Department of Health and Human Services to assess the feasibility of commercial wastewater testing. For that project, Biobot worked with utility companies to get samples and then provided the data to the CDC, which in turn shared the data with public health departments.
Matus and her colleagues have come to realize that it’s too early for them to calculate the number of infected individuals from wastewater data. Instead, the technology is best at providing a big-picture view of what’s happening in communities. “If you see the same trend in both types of data, you are more confident to make a high-stakes decision like imposing curfews in your city, like sending children to learn from home, like closing your office or closing a dorm,” she says.
The CDC established the NWSS, a partnership with state health departments, to fill a need that emerged early in the pandemic. “In 2020, we saw that testing capacity was not available everywhere,” Kirby says. “One of the real drivers for establishing wastewater surveillance is that it gave us a lab-based surveillance method that was totally independent of health-care access.”
Yet just as officials are gaining confidence in using wastewater data to monitor infection trends in communities, they now face a new challenge: the emergence of viral variants. Wastewater analyses aren’t in danger of missing infections—the PCR primers being used target regions of genes that have so far been unaltered in the variants. But people infected with different variants shed the virus at different rates, which can widen the error bars on the trends that researchers monitor.
And it is harder for researchers to spot an emerging variant in wastewater. Researchers can easily determine which variant is making someone sick by sequencing their nasal swab or spit sample and looking for changes in the viral genome. But sequencing an entire viral genome out of sewage samples is not as straightforward, Halden says.
In sewage, “most of what we’re detecting is damaged and degraded viruses,” the CDC’s Kirby says. “The genomes are already in pieces, and then the sequencing process breaks them up into even more pieces.”
Nonetheless, researchers are figuring out how to monitor variants in wastewater. Niko Beerenwinkel, a computational biologist at the Swiss Federal Institute of Technology (ETH), Zurich, and the SIB Swiss Institute of Bioinformatics, is part of a team that is performing such DNA sequencing to detect genomic variants in wastewater collected at treatment plants. The team sequences the samples using the same protocols used for clinical samples, then uses a computational method called V-pipe to determine whether variants are present.
Beerenwinkel argues that wastewater has a better chance than clinical samples of alerting a community to a new variant. “Most of the time we are able to detect variants earlier in the wastewater” than from the nasal swabs collected by a lab, Beerenwinkel says. When a new variant first arrives in a city, many clinical samples must be sequenced to detect even one case. “The advantage of wastewater is that you’re sampling a large population, whatever the wastewater treatment plant covers.” For example, Zurich’s wastewater treatment plant serves a population of approximately 400,000. “You’re taking a pooled sample from these 400,000,” Beerenwinkel says. “The chance that you will find [a variant] is higher as compared with sequencing only a subset of individuals who are infected.”
Once a new variant is identified from sequencing, whether that genome came from a wastewater or clinical sample, researchers can develop new primers that specifically target the variant’s characteristic mutations. That should allow the many sites using quantitative PCR to monitor the presence of those variants. Biobot has been working on methods to detect known variants since the beginning of this year, Matus says. It’s an “interesting challenge,” she says, because a variant of concern can emerge and disappear or become dominant in a matter of weeks. That’s not much time to develop a PCR test that targets the multiple mutations that differentiate variants. So “we have a fully dedicated team that could crank out a new assay in a week or two,” she says.
Even as researchers work through the ever-evolving challenges of sewage surveillance, they acknowledge the approach still has inherent limitations—an obvious one is when a location doesn’t have a sewer system but instead uses septic systems. “It’s really important to work closely with utilities and understand what they’re actually capturing,” Kirby says. Just as important is understanding what’s not included. For instance, some institutions like universities or prisons may be on their own systems. In addition, she says, the CDC is working on pilot projects looking at intermediate systems between full sewer systems and individual septic systems.
One of those pilot projects involves monitoring wastewater from prison systems. The project is modeled on the success that many universities had monitoring SARS-CoV-2 in wastewater from individual dormitories. Those universities tended to have labs on-site that could perform the analyses quickly. “We’re trying to replicate that in correctional facilities with a field-portable, on-site test kit that allows you to go from wastewater to results within 4 hours,” Kirby says. “I’m really excited to see if it works because a lot of our correctional facilities are remote. The few that have tried [wastewater monitoring] have had a turnaround time of 5–7 days or even longer. At that point, it’s hard to know what the data mean.”
Everyone in the field believes that wastewater monitoring is here to stay and is looking ahead to how it might be useful after this pandemic. The CDC’s Kirby says the NWSS project has been funded through 2025, and she anticipates that it will expand into monitoring wastewater for things such as foodborne diseases and genes that encode antibiotic resistance.
“We’re heading for a new normal where wastewater-based surveillance is part of the infrastructure of wastewater utilities and environmental agencies,” says Janelle Thompson, an environmental microbiologist at Nanyang Technological University who has been involved with wastewater surveillance in Singapore. She originally intended to use wastewater monitoring for viruses of concern in Singapore, such as dengue and other arboviruses, and plans to pivot back to them after the pandemic. “We really have built up the infrastructure over the past year,” Thompson says. “I bet we’re going to be using that infrastructure for years and years to come.”