Normalisation of SARS-CoV-2 concentrations in wastewater: The use of flow, electrical conductivity and crAssphage

Abstract

Over the course of the Corona Virus Disease-19 (COVID-19) pandemic in 2020-2022, monitoring of the severe acute respiratory syndrome coronavirus 2 ribonucleic acid (SARS-CoV-2 RNA) in wastewater has rapidly evolved into a supplementary surveillance instrument for public health. Short term trends (2 weeks) are used as a basis for policy and decision making on measures for dealing with the pandemic. Normalisation is required to account for the dilution rate of the domestic wastewater that can strongly vary due to time- and location-dependent sewer inflow of runoff, industrial discharges and extraneous waters. The standard approach in sewage surveillance is normalisation using flow measurements, although flow based normalisation is not effective in case the wastewater volume sampled does not match the wastewater volume produced. In this paper, two alternative normalisation methods, using electrical conductivity and crAssphage have been studied and compared with the standard approach using flow measurements. For this, a total of 1116 24-h flow-proportional samples have been collected between September 2020 and August 2021 at nine monitoring locations. In addition, 221 stool samples have been analysed to determine the daily crAssphage load per person. Results show that, although crAssphage shedding rates per person vary greatly, on a population-level crAssphage loads per person per day were constant over time and similar for all catchments. Consequently, crAssphage can be used as a quantitative biomarker for populations above 5595 persons. Electrical conductivity is particularly suitable to determine dilution rates relative to dry weather flow concentrations. The overall conclusion is that flow normalisation is necessary to reliably determine short-term trends in virus circulation, and can be enhanced using crAssphage and/or electrical conductivity measurement as a quality check.

Keywords: COVID-19; Normalisation; Public health; Sewage surveillance.

Graphical abstract

Fig. 1

Catchment areas in the Rotterdam Rijnmond area used for wastewater surveillance with locations of wastewater treatment plants (WWTP) and sewer pumping stations (ps).

Fig. 2

Boxplot with Q-based multiplication factors across catchment areas. Each boxplot gives the minimum, 25th percentile, the median, the 75th percentile and the maximum value by the whiskers.

Fig. 3

Boxplot with Q-based (left), EC-based (middle) and crAssphage-based (right) multiplication factors per catchment area. Each boxplot gives the minimum, 25th percentile, the median, the 75th percentile and the maximum value by the whiskers.

Fig. 4

Relations between multiplication factors derived from Q-, EC- and crAssphage-based normalisation for catchment WWTP Dokhaven INF4.

Fig. 5

SARS-CoV-2 RNA concentrations in wastewater (raw values and Q-, EC- and crAssphage- normalised values) and wastewater volumes for catchment WWTP Dokhaven INF3.

Fig. 6

Differences in percentage domestic wastewater (reciprocal of multiplication factors) between Q-based and EC-based normalisation for catchment area Katendrecht.

Fig. 7

CrAssphage daily load per capita for catchment ps Pretorialaan: chronologically (upper graph) and comparison with 24 h wastewater volumes (lower graph).

Fig. 8

Distribution of crAssphage concentration in human stool samples (n = 221) from the Rotterdam Rijnmond area. Data below the detection limit (n = 10³ gc/ml) is not shown in the graph. Normal distributions are fitted through the log¹⁰-transformed concentrations in stool samples of low shedder (<10^7.4 gc/ml; mean: 10^4.5; standard deviation 10^0.82; in black) and high shedder (≥10^7.4 gc/ml; mean: 10^9.3; standard deviation 10^0.96; in red).

Fig. 9

EC values of wastewater samples from catchment area WWTP Dokhaven INF2 and dry weather reference values with (EC_DWF,ref) and without (EC_DWF,ref,noQ) flow information.

Fig. 10

Measured and estimated 24 h wastewater volumes of catchment WWTP Dokhaven INF2 (upper graph) and their relative differences (lower graph).

Fig. 11

Boxplot with Q-based (left), EC-based (middle) and noQ-based (right) multiplication factors per catchment area. Each boxplot gives the minimum, 25th percentile, the median, the 75th percentile and the maximum value.