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

2.2. Nucleic acid extraction

Extraction of the nucleic acids from the frozen and shattered filters followed previously published manual RNA extraction methods (Griffiths et al., 2000) or the RNEasy Power Water extraction kit (Qiagen, USA). Lab 1 and lab 2 used the manual extraction method, while lab 3 used the RNEasy kit. To assess the mass of virus RNA on solids compared to that suspended in solution, RNA from the wastewater solids recovered from the centrifugation step were also extracted using the same methods. Resulting RNA concentrations were quantified by a plate reader with a Take3 plate (BioTek, USA), nanodrop (ThermoScientific, USA) or fluorometer (Qubit, Invitrogen) and were diluted to working concentrations of 25 to 50 ng/μl to prevent RT-qPCR reaction inhibition due to excess template or inhibitors.

2.3. RT-qPCR

Determination of the number of viral gene copies per mL of wastewater was determined by reverse transcriptase quantitative polymerase chain reaction (RT-qPCR). Primers and probes used for this study included the N1 and N2 primers and probe mix (2019-nCoV RUO, Integrated DNA Technologies, USA). Each 20 μl RT-qPCR reaction included 1× mastermix (either TaqPath™ 1-step RT-qPCR from Thermofisher or qScript XLT One-step RT-qPCR from Quantabio), 1.5 μM N1 primer/probe mix, 1.5 μM N2 primer/probe mix, 5 μl of template RNA at 25 to 50 ng/μl and PCR grade water. Thermocyclers used for the RT-qPCR included a QuantStudio 3 (ThermoFisher Scientific, USA) at lab 1 and lab 2 and a Quantabio (USA) at lab 3. The thermocycler conditions were used without modification from the CDC guidance (CDC, 2020). Briefly, at lab 1 and lab 2 the thermocycler conditions were: an initial step of 25 °C for 2 min; 50 °C for 15 min; 95 °C for 2 min; and 45 cycles of denaturation at 95 °C for 3 s and annealing at 55 °C for 30 s. At lab 3 the thermocycler conditions were: an initial step of 50 °C for 10 min; 95 °C for 3 min; and 45 cycles of denaturation at 95 °C for 3 s and annealing at 55 °C for 30 s. Each RT-qPCR run included positive controls consisting of 2019-nCoV_N_Positive Control (Integrated DNA Technologies, USA, hereafter positive control) and negative amplification controls consisting of 5 μl of PCR grade water. RT-qPCR assays were run in singlet (week 1 and 2) or triplicate (week 3 to 9). Virus concentrations were determined by comparing Ct values of samples against an assay-specific standard curve from a dilution of the positive control. Standard curves were made using a six-fold dilution of the positive control and the minimum detection limits (MDL) per RT-qPCR reaction and reaction efficiencies were as follows: Lab 1 MDL of 10 gene copies/5 μL RT-qPCR reaction and 93% efficiency, Lab 2 MDL of 10 gene copies/5 μL RT-qPCR reaction and 97% efficiency, Lab 3 MDL of 10 gene copies/5 μL RT-qPCR reaction and 98% efficiency. Dilution factors from the filtered sample volume, the RNA extraction procedure, and the RNA-containing sample volume per well in the RT-qPCR assay resulted in the calculated gene copy/mL of wastewater. Method detection limits of SARS-CoV-2 from wastewater, based on the recovery of spiked human beta coronavirus OC43 reported by two labs herein in other studies (Pecson et al., 2021) were <3 GC/mL or 1.2 ± 0.5 MVGC/cap/d for all facilities monitored.

2.4. Virus RNA signal decay during storage

To assess the influence of sewer travel time and sample storage conditions on the decay of SARS-CoV-2 RNA signal in wastewater several decay studies were conducted. Specifically, influent wastewater to CVWRF was collected in replicate non-sterile 250 mL PP bottles and transported to the laboratory. Upon arrival at the laboratory they were immediately incubated at 4, 10 and 35 °C for 1 to 22 h. Five bottles were incubated at each temperature. After the specified incubation period, one bottle at each temperature was sacrificially tested for SARS-CoV-2 using the methods presented above. Replicate samples collected at the same time as the 250 mL samples were processed immediately and assumed to represent the initial concentration of virus in the incubated samples. To compare effects of refrigeration and freezing on the virus in wastewater, influent to LCCWWTP and HCWWTP was incubated in sterile centrifuge tubes at 4 °C and -80 °C for 6 h, 24 h and 7 d, respectively. After storage for the required time, frozen samples were then thawed and processed by the methods above. The virus RNA abundance in the incubated samples was compared to samples that were processed immediately upon arrival at the laboratory.

2.5. Cross lab validation study on split samples

To evaluate the reproducibility of SARS-CoV-2 RNA detection and quantitation in wastewater between labs, split samples from four utilities were shared among testing laboratories on three different sampling events. Each sample was filtered in singlet (n = 1) or triplicate (n = 10), RNA from each filter was extracted and then triplicate RT-qPCR was performed. RNA quality as measured by A260/280 ratios were suitable, although the RNA yields from the RNEasy Power Water kit were lower (96.3 ± 50 ng/ul) than the Griffiths et al. (2000) manual extraction method (1674 ± 432 ng/ul). Analysis of variance (ANOVA) or general linear modeling (GLM) of the reproducibility of the analysis was performed in SAS (ver. 9.4; SAS Institute, Inc., Cary, NC).

2.6. GIS census population overlay with sewershed maps and COVID cases

To determine the area served by the individual treatment plants, sewershed polygons were either provided by the individual utilities or were extracted from city boundaries. Polygons of adjacent cities were clipped as necessary to prevent overlapping boundaries in the GIS shapefiles. These shapefiles were used to define the sewersheds served by each utility. Populations served by each utility were estimated by geocoding addresses for 3.2 M current residents of Utah and summing the number of individuals whose residence fell within each sewershed. These data were provided by the Utah Population Database, a collection of administrative data compiled from vital statistics, driver license, voter registration and healthcare claims provided by the State of Utah (UPDB, 2020). Over 97% of the provided addresses were geocoded with high confidence to the street segment or address point. COVID-19 daily and weekly counts of new cases were provided by the Utah Department of Health within the specific sewershed polygons provided to the Department of Health. If less than 5 cases were present in a sewershed during the specified time period, the data were suppressed and listed as <5 cases to protect privacy and avoid identifiability. For these time periods and sewersheds, it was assumed that one individual in the sewershed was ill with COVID-19 and the case load per 100,000 individuals was estimated.

3.2. Considerations on virus survival in sewersheds and sample handling

The data herein suggests that wastewater monitoring is useful for identifying new outbreaks of COVID-19 and confirming declining trends in infections. However, additional information is needed before SARS-CoV-2 RNA wastewater loads may be directly correlated with disease burden. Specifically, information is needed on: the rate and mass of virus RNA shedding in feces pre-, during and post-symptomatic COVID-19 phases; the virus survival and persistence in the sewer; the influence of facility and sewershed-specific factors such as runoff or groundwater infiltration or the presence of hospitals caring for COVID-19 patients; and the effect of sample handling on the virus abundance estimation.

The quantity of SARS-CoV-2 introduced into a sewershed is generally expected to be proportional to the actual number of infected individuals in that area – those that have and have not been clinically identified. To more accurately assess the number of infected individuals in a sewershed, an estimate of the SARS-CoV-2 RNA gene copies per unit weight of feces is needed during all stages of the disease. Using the literature reported values for SARS-CoV-2 in feces (median 4.7 log RNA GC/mL feces) (Cheung et al., 2020), the number of COVID-19 ill individuals within a sewershed was estimated herein over the study period by converting from GC/mL wastewater to SARS-CoV-2 shedding individuals and compared to the COVID-19 caseloads in the sewersheds. Overall, the estimated number of SARS-CoV-2 shedders in each sewershed was found to be linearly correlated with the cumulative diagnosed COVID-19 cases in a sewershed (linear regression, R² = 0.81, n = 10, P < 0.001) (Fig. 4 ). However, the daily estimated number of SARS-CoV-2 shedding individuals did not correlate with daily COVID-19 cases (Spearman correlation, P > 0.05). This lack of correlation at a finer temporal scale may be due to the variability in daily case counts reported to the Department of Health that are influenced by reporting lags due to weekends or holidays, test kit availability and processing rates, etc. Further, it is important to note that the Department of Health case counts are tied to a residential address, yet ill individuals may be residing or working in locations outside the expected sewershed or may be present in hospitals not in the same sewershed as their residence. In our study area's the number of hospital beds accounted for only 0.8 ± 0.5% of the total sewershed population. Further, only a subset of these beds would have been occupied by patients hospitalized for COVID-19. Conversely, other biological and non-biological factors must be influencing the SARS-CoV-2 persistence or detection in the wastewater when the data is evaluated at a finer temporal scale. Therefore, rolling case counts are likely a better metric to compare to the SARS-CoV-2 RNA in wastewater.

Open in a separate window

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.

In evaluating Fig. 4 it can be seen that the estimated wastewater SARS-CoV-2 concentration compared to case counts is 0.78:1. This suggests that the estimated sum of SARS-CoV-2 shedders is less than the sum of confirmed cases during the study period. While theoretically this would be a 1:1 ratio, there are several possible explanations for the discrepancy including (1) not all COVID-19 individuals shed the virus in their feces and the length and duration of shedding may vary between individuals, (2) weekly or bi-weekly wastewater sampling compared to daily case counts underestimating RNA abundance, (3) decay of the RNA signal in wastewater, or (4) inefficiencies in the sample processing method. Further, research in fecal shedding of the SARS-CoV-2 RNA in symptomatic and asymptomatic COVID-19 infected individuals is required to address some of these uncertainties.

Numerous physical, chemical and biological factors may influence the persistence of viral RNA in wastewater. These factors include temperature, sunlight, ionic strength, presence of antiviral chemical constituents (Nannou et al., 2020), solids content, residence time in the sewer, microbial antagonism (Boehm et al., 2019; Tennant et al., 1994) and sampling methodology. In this study, we evaluated the effect of incubation of SARS-CoV-2 RNA containing wastewater at different temperatures on the loss of RNA over time (Fig. 5 ). Specifically, wastewater from three plants was evaluated to determine the loss of RNA during storage at 4 °C and -80 °C, and during transport in a sewer system at 10 °C and 35 °C. Initial concentrations of the virus RNA in the wastewater used for the decay studies ranged from 135 to 953 gene copies per mL of wastewater. Overall the results indicate a first order decay rate of the viral RNA ranging from 0.09 to 0.12 h⁻¹ over the 22 to 24 h at 4, 10 and 35 °C. While the RNA was not detectable after 6 h at 35 °C, the RNA was still detectable after 22 h of incubation at 4 and 10 °C and after one week at -80 °C. The overall reduction in viral RNA during the storage or incubation periods was 67% at 10 °C over 22 h, 86.5 ± 0.5% at 4 °C over 24 h and 92.4 ± 10.3% at -80 °C over one week. These results presented herein on the effect of temperature, indicate that the SARS-CoV-2 RNA is quite labile in wastewater as suggested by others based on studies of SARS-CoV and surrogate coronaviruses. Reported decay rates for SARS-CoV and surrogate coronaviruses in unpasteurized wastewater at 23 °C range from 0.02 to 0.143 h⁻¹ (Gundy et al., 2008; Hart and Halden, 2020; Ye et al., 2016), well within the range reported herein. Similarly, the inactivation of MHV was reported to be 0.048 h⁻¹ at 4 °C, 0.059 h⁻¹ at 10 °C and 0.142 h⁻¹ at 25 °C. Others reported that SARS-CoV RNA could be measured by RT-PCR in domestic sewage for up to 14 days at 4 °C but only 3 days at 20 °C (Wang et al., 2005b). These prior studies corroborate our findings of loss of SARS-CoV-2 RNA in wastewater. Recent reports of the decay rates of gamma irradiated SARS-CoV-2 spiked into untreated wastewater were 0.0035 h⁻¹ at 4 °C and 0.012 h⁻¹ at 37 °C (Ahmed et al., 2020c). Use of gamma irradiated SARS-CoV-2 may not necessarily represent the mixture of SARS-CoV-2 likely present in wastewater that could be detected by RT-qPCR. Specifically, SARS-CoV-2 shed with feces may contain intact virus, capsid compromised virus and free nucleic acids. Therefore, it is likely the more rapid decay of SARS-CoV-2 RNA measured by RT-qPCR in the studies herein reflect the decay of all three of these expected viral RNA sources which would more typically be found in naturally contaminated wastewater. Given that the wastewater residence time in some larger cities may be up to 13 h (Kapo et al., 2017), whereas smaller cities such as ECWRF had a 1 to 3-h residence time, the potential for decay of the virus RNA should be considered when assessing virus loads. Before an accurate model of the likely number of COVID-19 infected individuals in a sewershed can be made, an understanding of the decay rate of the virus RNA in each sewer system is needed.

Open in a separate window

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%.

The effect of sample collection and handling on viral RNA concentrations is also needed to develop comparable relationships between wastewater samples and disease prevalence. An interlaboratory replicate analysis of spilt samples indicated comparability between the sample processing at the different labs (Fig. 6 ). This interlaboratory analysis suggests the mean among labs was 210 gene copies/mL wastewater (10 and 90% confidence intervals [CI] of 76 to 309 gene copies/mL) for CVWRF, 104 gene copies/mL wastewater (16 and 230 CI) for OWRF and 98 gene copies/mL wastewater (41 and 178 CI) for SLCWRF. There was no significant difference in means among the three labs in the CVWRF and SLCWRF samples, but one of three filters from lab 1 and lab 2 did differ from the lab 3 results for the OWRF sample. This variation between filters was used to refine sample processing to include multiple filters for all wastewater samples processed in each of the labs. These results suggest the data when using replicate filters and RT-qPCR assays is comparable between labs with this sample handling and processing method.

Open in a separate window

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).

Finally, the amount of SARS-CoV-2 RNA on wastewater influent solids, compared to the liquid was determined by quantifying the gene copies per g of solids and per g of liquid in eight samples. Overall, more viral SARS-CoV-2 RNA was detectable in the liquid phase (91 ± 12% by mass) of the wastewater influent compared to the RNA sorbed on the influent solids (9 ± 12% by mass). This finding is similar to that reported for MHV, where 70% of the enveloped virus was associated with the wastewater liquids at equilibrium (Ye et al., 2016). A recent study showed that the removal of solids prior to SARS-CoV-2 quantitation did not systematically impact recovery corrected SARS-CoV-2 estimated concentrations. Specifically, as reported in Pecson et al., (2021), the methods used herein on five replicates of two influent samples from treatment plants in California, showed that solids removal reduced the average SARS-CoV-2 concentrations by 15%. Interestingly, some methods that removed solids prior to ultrafiltration showed a higher concentration of recovery corrected SARS-CoV-2 than ultrafiltration without solids removal (Pecson et al., 2021). Others have reported the detection of the SARS-CoV-2 RNA on activated sludge in treatment plants (Kocamemi et al., 2020; Peccia et al., 2020). Given that most wastewater is near neutral pH and the SARS-CoV-2 spike proteins have estimated isoelectric points around 5.4 and 5.3 (Mallik, 2020), the virion positive core is likely surrounded by negatively charged envelopes and spikes in these wastewaters. Moreover, RNA bases adenosine and cytidine can be protonated on N1 and N3 atoms, respectively, with 3.8 and 4.3 solution pK_a (Bink et al., 2002; Izatt et al., 1971). Therefore, the virus and viral RNA are likely to adsorb to activated sludge. However, due to variable return activated sludge wasting rates at facilities and periodic sludge bulking, using the virus RNA abundance in sludge to correlate with COVID-19 case rates may be difficult. In this study, we focused on wastewater influent, as it was the most comparable sample type between the ten facilities sampled, which varied from advanced mechanical plants to lagoon systems.

The authors are grateful for the voluntary participation of the wastewater utilities, in particular: Tiffini Adams, Snyderville Basin WRD; Giles Demke, Orem WRF; Paul Fulgham, Tremonton WWTP; Issa Hamud, Logan City WWTP; Kevin Maughan, Hyrum City WWTP; Rich Mickelsen, Timpanogos SSD; Jeff Richens, Price River WID; Obe Tejada, Moab City WWTP; Jamey West, Salt Lake City WRF. The authors are grateful to Danielle Zebelean, Katrina Brown, and Alison Barnett for aid in sample processing; to Marissa Taddie, Lenora Sullivan, and Matt McCord for help with GIS of the sewersheds, population sets, and sewershed case counts; and to Benjamin Brown, Alex Anderson, and Dan English for sample collection and transport. D.K.R. acknowledges financial support from the Utah State University Department of Biological Engineering. R. Jamal was funded under the US National Science Foundation (Award No. 1650098). A portion of this work was supported by the Immunology, Inflammation, and Infectious Disease Initiative at the University of Utah. Partial funding for the project was provided by the Utah Department of Environmental Quality.Any opinions, findings and conclusions or recommendations expressed in this manuscript are those of the authors and do not necessarily reflect the views of the State of Utah, Utah Department of Environmental Quality, Utah Department of Health, or the National Science Foundation. DEQ regulates wastewater in the state of Utah, however the data collected for this project do not serve a regulatory purpose. DOH participated in this work as part of the ongoing response to the COVID-19 pandemic.