Correlation of SARS-CoV-2 RNA in wastewater with COVID-19 disease burden in sewersheds

Abstract

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), which causes the coronavirus disease (COVID-19), is shed in feces and the viral ribonucleic acid (RNA) is detectable in wastewater. A nine-week wastewater epidemiology study of ten wastewater facilities, serving 39% of the state of Utah or 1.26 M individuals was conducted in April and May of 2020. COVID-19 cases were tabulated from within each sewershed boundary. RNA from SARS-CoV-2 was detectable in 61% of 126 wastewater samples. Urban sewersheds serving >100,000 individuals and tourist communities had higher detection frequencies. An outbreak of COVID-19 across two communities positively correlated with an increase in wastewater SARS-CoV-2 RNA, while a decline in COVID-19 cases preceded a decline in RNA. SARS-CoV-2 RNA followed a first order decay rate in wastewater, while 90% of the RNA was present in the liquid phase of the influent. Infiltration and inflow, virus decay and sewershed characteristics should be considered during correlation analysis of SAR-CoV-2 with COVID-19 cases. These results provide evidence of the utility of wastewater epidemiology to assist in public health responses to COVID-19.

Keywords: COVID-19; Disease burden; Interceptor; SARS-CoV-2 RNA; Sewershed; Wastewater epidemiology.

Graphical abstract

Fig. 1

Location of wastewater treatment plants sampled during this study, representing 1.26 M individuals or 39% of Utah's population.

Fig. 2

Average and standard deviation of SARS-CoV-2 million viral GC/capita/day in wastewater (grey bars) compared to weekly COVID-19 case rate per 100,000 (red lines). Vertical dashed lines indicate the first week of sampling. Estimated detection limits for SARS-CoV-2 is 3 GC/mL or 1.2 ± 0.5 MVGC/cap/d for all facilities monitored. Note that the cases/100 K do not include data from asymptomatic or pre-symptomatic SARS-CoV-2 shedders. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 3

Correlations between daily (plot A) or weekly (plots B, C, D) COVID-19 cases or cases/100 K and SARS-CoV-2 million viral gene copies/capita/day in a sewershed. Plot D shows the 1-week lag in COVID-19 cases in ECWRF compared to the prior week SARS-CoV-2 RNA. Spearman correlations and 95% prediction intervals (dashed line) on the linear regressions (solid line) are shown in each figure.

Fig. 4

Estimated sum of SARS-CoV-2 shedding individuals in sewersheds over the study period compared to the sum of confirmed COVID-19 cases.

Fig. 5

Observed decay (symbols) of SARS-CoV-2 RNA in wastewater during storage at 4, 10 or 35 °C and predicted first order decay rates (lines). Virus signal was not detectable after 12 h of storage at 35 °C. The grey area indicates the data used to estimate the first order decay rates. Correlation coefficients exceeded 95%.

Fig. 6

Intra laboratory comparison of replicate filter extractions and triplicate qPCR assays. Bonferroni grouping of least square means (alpha = 0.05). Mean estimated gene copies/mL by filter, with the same letter are not significantly different. OWRF: significant difference in means among filters (GLM, P < 0.001, F = 9.52). CVWRF: No significant difference in means among filters (GLM, P = 0.023, F = 3.1). SLCWRF: No significant difference in means among filters (GLM, P = 0.024, F = 4.51).

Fig. 7

SARS-CoV-2 RNA in the sewer interceptors feeding CVWRF influenced by flow rates (plot A) and without considering flow rates (plot B), and correlation between SARS-CoV-2 million viral gene copies/capita/day in wastewater as compared to rolling 3-day average COVID-19 case counts in the cities contributing to CVWRF (plot C).

Fig. 8

Effect of flow rates on relationship between SARS-CoV-2 million viral gene copies/capita/day or gene copies/mL wastewater. Plot A: eight facilities sampled with more than two detections of the virus RNA. Inset B: variation in the ECWRF flows. Plot C: sub-sewersheds feeding CVWRF as compared to the influent into the main plant (bold black line).