Opinion: Challenges Quantifying COVID-19 Cases Using Wastewater

Editor's Note: The views expressed in this post are the sole opinion of the author and not those of WaterOperator.org, our sponsors, or the University of Illinois.

In the May 5, 2020 edition of the WaterOperator.org newsletter, we highlighted ongoing research that uses wastewater-based epidemiology to monitor the spread of SARS-CoV-2, the virus that causes COVID-19. Especially in locations where no confirmed cases have been identified, any samples positive for SARS-CoV-2 viral RNA implies that there are people infected in that community excreting it. For that purpose, wastewater monitoring shows real promise as an approach to early detection. By monitoring wastewater influent, scientists hope we can develop an advanced warning system for outbreaks.

There has been significant buzz about using wastewater to quantify the actual number of people infected within a given service area, but there are some issues with quantifying cases I want to discuss. In our newsletter we highlighted MIT research aiming to quantify the number of infected from a large area in Massachusetts. In that article, the researchers point to concerns about meeting the litmus test of sound science.

The wastewater system they studied had 450 confirmed cases at the time of sampling. Results from this monitoring suggested the number of people infected could be much higher. They estimated somewhere between 2,300 and 115,000 infected people. A range this wide does little to help planners or health officials prepare for what might be coming during a pandemic.

Quantifying the number of people infected with COVID-19 using wastewater samples requires a much more comprehensive data set that we cannot gather today in any cost effective way. Here are a few of the problems I see in quantifying the positive COVID-19 population within a given wastewater system:

• Not everyone excretes the viral RNA:
  A recent study published March 30 in the American Journal of Gastroenterology found that some COVID-19 patients exhibit gastrointestinal symptoms, with those patients more likely to produce a positive stool test. In other words, COVID-19 positive patients may not have ANY viral RNA in their stool. How do we identify those people?
• Wastewater varies throughout the day and throughout the week:
  The influent coming through a plant varies based on the discharges from the users. A lot of variables can affect wastewater characteristics at the specific time a sample is collected. The time of day, time of week, and even the time of year can affect the flow into a plant depending on the types of users in the system.
• Every system has a variety of sources for their wastewater:
  What percentage of the wastewater is residential? Are there commercial or industrial facilities that are discharging to the community system? If so, how much, and what types of businesses? In some communities, commercial and industrial users could make up a significant portion of the wastewater treated. In a rural area, the regional hospital may be in a smaller community making it a significant source and contributor. Other communities could be almost completely residential.
• Sampling time and frequency can skew the results:
  Sampling time matters, as do the number of samples collected. How do we decide what is representative? Once an hour? Once a day? Sampling may need to be continuous to really understand the variability.
• Wastewater collection systems leak:
  Leaking can occur both ways. Some wastewater leaks into the environment through the collection system while, at other times, a high groundwater table may be leaking groundwater into the collection system. I looked at approximately 50 smaller systems in Illinois to compare the amount of wastewater discharge to the amount of groundwater they withdraw from drinking water wells. (You would expect the amount withdrawn from wells to be more than that treated at the wastewater plant because of consumptive use.) In many cases systems were treating more wastewater than the raw water being used for their community supply and, in some cases, it was 2-3 times a much. This would be significant factor when using any volumetric approach to evaluating COVID-19 sampling results.
• We have no benchmark to compare results:
  Without having data for a number of communities where the total number of residents with active COVID-19 infection is known, there is no way to validate assumptions and calibrate estimates built into the method. This would not be possible without a consensus understanding about the rate of asymptomatic cases.

If researchers must accept such a high degree of uncertainty, how can this method ever be accurate or useful? Many factors would have to be considered to quantify the number of positive cases for a given community and these would be unique to the individual system. That said, these are not likely new considerations for the talented researchers working on this effort. 

In the future I hope an approach to accurately quantify an infected population using wastewater-based epidemiology becomes a reality. It would be a tremendous asset. In the meantime, however, I believe our focus should be on evaluating the pitfalls mentioned above and working toward technologies/protocols needed within a wastewater plant to reduce uncertainty and move us closer to our common goal of protecting public health.