The reason I started thinking about this is simple: the outbreak is moving fast, but the numbers we can actually see are lagging behind reality. Several family members in my hometown, a third- or fourth-tier city, tested positive this week, far more than the official counts suggest. Yet there is still no reliable way to understand how the epidemic is progressing in time.
Search-engine trends, Weibo polls, and the kind of anecdotal “statistics” people build from their own circle are all too easily distorted by emotion. They also only reflect internet users, which leaves out a large part of the population. In China, many older people are concentrated in rural areas; they may not have felt the full impact yet, but once the Spring Festival travel period begins, a significant gap in medical resources is easy to imagine. Without data, that problem may never become visible in the online conversation. As a researcher, I would rather work with a hard indicator.
My background is in environmental chemistry, and over the past three years I have seen plenty of papers using municipal wastewater treatment plants for epidemiological surveillance. A city’s wastewater plant is where water-soluble pollutants from that city converge, so shifts in measured concentrations can reflect changes at the urban scale. Sampling wastewater is also already part of routine environmental monitoring work in many places.
If we can use those measurements to track changes in SARS-CoV-2 concentration, then we can roughly understand how an outbreak is evolving locally. Because infection and symptom onset usually do not happen at the same moment, a peak in viral load in environmental samples can also help predict clinical case counts one to two weeks ahead. That matters a lot when medical resources are limited.
Wastewater epidemiology has already been accepted by many countries as a tool for public-health decisions. In the United States, Biobot works with the CDC and publishes wastewater SARS-CoV-2 data at the county level on a regular basis. Their data showed that during the Omicron surge at the end of last year, wastewater concentration rose about one to two weeks earlier than clinical cases. The company is now focusing more on opioids in wastewater.
Europe has had similar experiences. In the Netherlands, SARS-CoV-2 was detected in wastewater three weeks before the first confirmed COVID case; in Spain, the lead time was about one month; in Milan, Italy, it was almost two months. Several EU countries have launched similar monitoring programs. According to EU estimates, daily monitoring of one wastewater treatment plant for a full year costs about 25,000 euros; if monitoring is reduced to twice a week, which is currently the minimum for obtaining reliable data, the cost is lower still.
China already has many research groups working on this topic, but we still do not have monitoring-grade datasets. In fact, building a full workflow for SARS-CoV-2 concentration surveillance in wastewater, from sample collection to data display, is not difficult. A paper published in Nature in 2020 provides a process that can actually be implemented.
Sampling and analysis
Collect 40 mL of sludge from the primary settling tank of a wastewater treatment plant, with a solids content of 2.6%–5%, and store it at -80°C before analysis. Take 2.5 mL of the sludge and extract total RNA using a commercial soil total RNA kit, such as the RNeasy PowerSoil Total RNA Kit (Qiagen). After RNA extraction, dissolve the RNA fraction in 50 µL nuclease-free water and measure total RNA concentration with a spectrophotometer.
Then quantify SARS-CoV-2 in the sample using the CDC-recommended one-step RT-PCR method. The protocol is available online. The principle is to measure viral abundance with the N1 and N2 primer/probe sets. These primers can be bought online for less than 200 dollars, and because the sequences are already known, they can also be synthesized directly using information released by the WHO. A human RP gene control should be included at the same time; only samples containing the RP gene should be considered positive for SARS-CoV-2.
The exact quantification depends on your instrument. In the paper, the authors used the Bio-Rad iTaq Universal Probes One-Step Kit, with the following conditions:
20-µl reactions run at 50°C for 10 min and 90°C for 1 min, followed by 40 cycles of 95°C for 10s and 60°C for 30s per the manufacturer’s recommendations
SARS-CoV-2 RNA concentration must be corrected with a standard curve. The basic idea is to synthesize a DNA template corresponding to SARS-CoV-2 RNA, amplify the N gene from it, and then use the MEGAscript T7 Kit to generate single-stranded RNA. Because the copy number of that RNA is known from the amplification setup, it can be serially diluted and measured with the RT-PCR fluorometer to build a standard curve. That allows you to estimate viral concentration in the sample.
The first-step total RNA measurement should then be used for normalization. Under a five-fold dilution, no qRT-PCR inhibition was observed.
For quality control, use sludge samples from 2018 as a negative control.
This workflow does not require a P4 laboratory. Once the viral RNA is extracted from sludge, it no longer has meaningful activity, so a standard bioscience lab should be enough for routine measurements. The only obvious drawback I can think of is the smell of the sludge, but anyone working in environmental chemistry is probably already used to that.
Working with the data
Once the viral concentration in each sample has been measured, a scatter plot with a trend line is enough to produce a quick view of how the concentration is changing over time.
If sequencing is also possible, then different variants can be tracked at the same time. At that point, the data-processing mirror provided by UCSD and a paper published in Nature this year can be used directly.
There is nothing especially difficult about the experimental side. The workflow relies on commercial kits and established instruments. In a research lab, it should be possible to set everything up in a day once the materials are ready. For a monitoring station, a more detailed standard operating procedure may be needed, but even that should take no more than a week before it can be used in practice. People read papers not just to publish more papers, but to make them useful. Today, automation and standardized kits have shortened the gap considerably.
What I hope is that labs with the necessary conditions can move quickly to build their own SOPs and begin monitoring. If city and regional viral-peak information were released through social media, it might buy vulnerable groups one or two weeks of breathing room to adjust resources. Sampling does not even need to be limited to urban wastewater treatment plants; rural domestic wastewater can also be monitored.
In the long run, China should deploy this kind of capability on top of its existing national environmental monitoring network so that public-health events can be flagged earlier. Realistically, that kind of warning may be needed several times every year.