Catching a resurgence: Increase in SARS-CoV-2 viral RNA identified in wastewater 48 h before COVID-19 clinical tests and 96 h before hospitalizations

Abstract

Graphical abstract

Open in a separate window

1. Introduction

Wastewater has been used as a window in the habits and health of populations. A relatively novel epidemiological technique called wastewater-based epidemiology (WBE) has been successfully used by public health units and authorities for decades to track trends in the general population for numerous diverse substances: illicit and recreational drugs (Been et al., 2014; Feng et al., 2018; González-Mariño et al., 2016; Gushgari et al., 2018), toxic and harmful chemicals and pesticides (Du et al., 2018; Markosian and Mirzoyan, 2019; Rousis et al., 2017), as well as prevalence of disease and illness (Nordgren et al., 2009; Yang et al., 2017). An intense interest has developed in looking to sewage to seek out biomarkers, pathogens and viruses to further the understanding of the prevalence of diseases, diets and habits in the general population (Bowes and Halden, 2019; Lorenzo and Picó, 2019). The earliest occurrence of using sewage as a source of information was approximately 25 years ago, where scientists discovered that polio was in fact found and transmissible in sewage and human stool (Metcalf et al., 1995). One of the major advantages of WBE is that with well-designed sampling/testing regimes, it can be utilized to gain valuable, and sometimes otherwise unobtainable information for subsets of population, as small as a single building or neighborhood, or as large as a city or county (Choi et al., 2018; Tscharke et al., 2019).

Most recently, WBE has been applied to the quantification of COVID-19 disease signal. The COVID-19 global pandemic has had profound impacts on day-to-day life through sickness, death and the economic and social effects of lockdowns, closures and curfews put in place by local governments to control and limit the community spread of the disease. At the beginning of the pandemic, COVID-19 surveillance approaches invoked by nations around the world included case ascertainment of individual patients via nucleic acid-based tests, serological tests and contact tracing. As such, WBE has been investigated as a potential surveillance tool to obtain a rapid measure of the relative incidence rate of COVID-19 infections in the population at large, generating an entirely new epidemiological metric for public health units (PHU) to utilize. Fecal shedding of SARS-CoV-2 viral particles has been shown to occur before, during and after active COVID-19 infections, for periods ranging from a few days to several weeks (Chen et al., 2020; Cheung et al., 2020; Xing et al., 2020; Xu et al., 2020), with current literature suggesting that children may shed viral particles longer than adults (Li et al., 2020b; Ma et al., 2020). Additionally, fecal shedding of SARS-CoV-2 viral particles in stool occurs in patients infected with COVID-19 and appears to be independent of the presence of SARS-CoV-2 viral particles in the upper respiratory tract.

WBE has been applied at wastewater resource recovery facilities (WRRFs), with early studies demonstrating that viral signal is detected from both influent and primary clarified sludge, with recent studies identifying that at WRRFs, primary clarified sludge may provide more sensitive measurements than in influent (D'Aoust et al., 2021; Graham et al., 2020; Peccia et al., 2020). Additionally, recent studies outlined that viral particles of SARS-CoV-2 in wastewater could in fact be predominantly solid particle-bound (Canadian Water Network (CWN), 2020; Graham et al., 2020) due to the virus' lipid bilayer, which causes it to be lipophilic (Schoeman and Fielding, 2019). Furthermore, it was observed in several studies that SARS-CoV-2 viral signal in wastewater could be utilized as a tool to track epidemiological metric trends (Ahmed et al., 2020; Medema et al., 2020a; Nemudryi et al., 2020; Thompson et al., 2020). One of the recent developments to the field of WBE is the renewed interest in investigating biomarkers to normalize the SARS-CoV-2 viral signal. Previous WBE work identified several possible normalization biomarkers in wastewater such as 5-Hydroxyindoleacetic acid (Chen et al., 2014), caffeine and cotinine (Rico et al., 2017). However, due to the higher incidence of SARS-CoV-2 viral particles in feces as compared to urine (Collivignarelli et al., 2020; Jeong et al., 2020; Lodder and de Roda Husman, 2020; Zhurakivska et al., 2020), other biomarkers present in wastewater solids, such as pepper mild mottle virus (PMMoV) (Jafferali et al., 2020; Kitajima et al., 2018; Rosario et al., 2009; Symonds et al., 2014) and crAssphage (Crank et al., 2020; Polo et al., 2020) saw are being utilized for the normalization of SARS-CoV-2. Some recent studies have demonstrated that biomarker normalized SARS-CoV-2 viral signal may in fact track be capable of tracking the incidence of COVID-19 infections in the community (Gerrity et al., 2021; Kitamura et al., 2021) more effectively than non-biomarker-normalized viral signal (Jafferali et al., 2020; Wu et al., 2020), even if accounting for corrections due to flow or solids mass flux (D'Aoust et al., 2021).

To appropriately utilize epidemiological data obtained by measuring SARS-CoV-2 viral signal in wastewater, PHUs require an understanding of the predictive ability of this epidemiological metric to appropriately action obtained data (Hart and Halden, 2020; Hill et al., 2020; Thompson et al., 2020). Recent studies appear to demonstrate that WBE was able to identify the onset of community infection prior to clinical testing during the first wave, (D'Aoust et al., 2021; Kaplan et al., 2020; La Rosa et al., 2020; Medema et al., 2020b; Nemudryi et al., 2020; Peccia et al., 2020; Sherchan et al., 2020; Vallejo et al., 2020; Wu et al., 2020) however, there isn't a clear consensus on how early WBE can detect resurgences of the disease. Due to reporting limitations or absence of widespread clinical PCR or serological testing at the onset of the first wave of the COVID-19 pandemic, the predictive qualities of SARS-CoV-2 viral signal tracking in wastewater may not have been fully assessed, creating a critical and fundamental lack of understanding of intrinsic predictive qualities of wastewater-based SARS-CoV-2 surveillance, particularly for subsequent resurgences of COVID-19 infections in the community. As such, it remains unclear whether WBE can have a role in identifying COVID-19 resurgences before other, more traditional epidemiological metrics, such as clinical and serological testing results. As a result, the objective of this study is to establish a relationship between SARS-CoV-2 RNA signal in wastewater during a resurgence in a community with established clinical testing and (i) increases in SARS-CoV-2 RNA signal in wastewater, (ii) increases in the number of new COVID-19 positive patient cases in the community, and (iii) increases in the number of new hospitalizations of COVID-19 positive patients in the community.

2. Materials and methods

2.1. Characteristics of the City of Ottawa's water resource recovery facility

Primary clarified sludge samples were harvested from the City of Ottawa's Robert O. Pickard Environmental Centre, in Ottawa, ON, which services approximately 1 M residents of the national capital region. The hydraulic residence time of the Ottawa sewershed ranges from 2 h to 35 h, with an average residence time of approximately 12 h. The Ottawa water resource recovery facility (WRRF) has an average daily flow of 435,000 m³ per day. The facility is comprised of preliminary treatment consisting of coarse and fine screening and grit chambers for the removal of larger particles. The primary treatment units consist of rectangular primary clarifiers, where samples from this study are harvested (Fig. 1 ), and the secondary treatment is a conventional activated sludge system operating without nitrification. Finally, chlorination is used as a disinfection method followed by dechlorination prior to discharge to receiving waters.

Open in a separate window

Fig. 1

Bird's-eye view of the Ottawa WRRF, with the primary clarifiers identified.

3. Results, discussion and implications

3.1. Trend and peak of wastewater SARS-CoV-2 RNA viral level in wastewater compared to other clinical COVID-19 surveillance metrics

This study compares PMMoV normalized SARS-CoV-2 viral copies, a metric outlined in previous SARS-CoV-2 WBE work (D'Aoust et al., 2021; Kitamura et al., 2021; Wu et al., 2020) with clinical epidemiological metrics. This WBE-based metric normalizes SARS-CoV-2 viral copies to the number of copies of PMMoV in wastewater to account for variations inherent to wastewater samples; in particular variations related to the wastewaters water quality, solids and fecal matter concentrations. In addition, PMMoV normalizes for variations associated with sampling, storage, and testing. PMMoV is ubiquitous in both the liquid and solid fraction of wastewater and is effective due to its low temporal variability in wastewater (Fig. 2 ). The lower temporal variability of PMMoV is likely to be linked to the stability and hardiness of its virions (Sassi et al., 2018).

Open in a separate window

Fig. 2

Ct values of the SARS-CoV-2 N1 and N2 gene regions, and the PMMoV normalization gene (1:10 dilution shown), outlining the low variability of PMMoV, justifying its use as a normalization gene. The 10th and 90th percentile are displayed, along with the median (dotted line inside shapes). The variance of the different targets is also shown at the top.

All conventional COVID-19 surveillance measures (# of hospitalized cases, 7-day mid-point floating average percent positivity, clinical test percent positivity and number of new daily cases; Fig. 3a–d) increase during the study period from a low level at the beginning of the study. The number of current hospitalized patient cases are at their highest on July 31-Aug 2 with a 168% increase compared to the week of July 13 (Fig. 3a–d). Meanwhile, percent positive cases and daily cases are at their highest on July 19, with increases of 215% and 326% respectively, relative to the preceding 7-day period (Fig. 3b–d).

Open in a separate window

Fig. 3

Epidemiological metrics and SARS-CoV-2 viral concentrations over the study period; a) COVID-19-caused hospitalizations, b) percent test positivity (7-day mid-point floating average), c) percent test positivity and d) number of new daily cases, along with e) wastewater SARS-CoV-2 N1 and N2 gene copies/PMMoV gene copies. Open data points signify points which were below the limit of quantification. The dotted line in the background represents the number of daily clinical tests/100 K pop. Performed.

As observed with other epidemiological metrics, the PMMoV normalized viral signal in Ottawa's wastewater is also shown to be low at the beginning of the study period through to the week of July 13 with both N1 and N2 regions below 1.00 × 10⁻⁴ copies/copies PMMoV (Fig. 3e). There is a 450% relative increase (p < 0.05) in N1 and 440% increase (p < 0.05) in N2 viral signal on July 17 to 19 compared to July 13 to 15. An absolute increase of 2.03 × 10⁻⁴ and 3.01 × 10⁻⁴ copies/copies PMMoV is reported during this period of increase for N1 and N2 gene regions, respectively (Fig. 3e). The PMMoV normalized viral signal then decreases again following this peak, to levels higher than the previous baseline, but notably lower than the peak outlined between July 17 to 19.

Throughout the study, a good agreement is observed between N1 and N2 SARS-CoV-2 gene region targets (R = 0.852, r² = 0.726), implying that some portions of the N1 and N2 gene regions may be preserved. This potentially indicates that RNA extracted from primary sludge and wastewater could be sequenced to track the transmission of new, and potentially more infectious mutations of the disease (Li et al., 2020a; Oude Munnink et al., 2020; Rockett et al., 2020; Volz et al., 2020)

This study also suggests that the PMMoV normalized viral signal peaks for a relatively short period of time before showing a subsequent rapid decrease in signal. This short peaking period may be an artifact of the inherent variation in wastewaters or associated with variations inherent to the assay. The short peak period may also be associated to fecal shedding patterns of the population contributing to the wastewater. It is noted that several current studies investigating fecal shedding timelines and intensity for SARS-CoV-2 may in fact be investigating the sickest of patients in a hospital setting (Chen et al., 2020; Fontana et al., 2020), and so patterns may not reflect occurrences in the general population, where a significant quantity of patients will likely suffer from milder symptoms, or no symptoms at all (Pan et al., 2020; Park et al., 2020). Furthermore, sicker or elderly patients may in fact “fall off” the wastewater monitoring efforts if bed pans or adult diapers are utilized as part of their treatment and convalescence in a hospital or retirement home (Coelho et al., 2015; Tsang et al., 2017).

Previous work on PMMoV identified that it is a useful environmental metric to track in surface waters and drinking waters as it correlated strongly with fecal contamination of said waters (Hamza et al., 2011; Haramoto et al., 2013; Rosiles-González et al., 2017). As a result, the use of PMMoV normalized data in SARS-CoV-2 is seeing increased utilization in several locales due to the metrics' ability to correct for the true quantity of fecal matter in wastewater (D'Aoust et al., 2021; Wu et al., 2020). It is hypothesized that due to the fact PMMoV signal in wastewater is likely diet based (due to consumption peppers, some tomatoes, or related foodstuff infected with PMMoV) (Jarret et al., 2008; Peng et al., 2015; Zhou et al., 2020), this approach will work best when performing a survey of a larger group of individuals or geographical area, where enough survey participants with varied enough diets will average out the signal.

3.3. Temporal association between surveillance metrics

A visual comparison of clinical testing percent positivity and the number of new daily positive COVID-19 cases compared to the PMMoV-normalized RNA signal shows that the viral RNA signal predates both the clinical testing percent positivity and new daily positive COVID-19 cases epidemiological data.

A times-step correlation analysis suggests that increases in SARS-CoV-2 signal in wastewater precedes increases in new daily positive COVID-19 cases and clinical testing percent positivity by two days (N1: R = 0.703, p < 0.001; N2: R = 0.721, p < 0.001; N1: R = 0.703, p < 0.001; N2: R = 0.714, p < 0.001) (Table 1 ). Furthermore, the same time-step correlation analysis also suggests that increases in SARS-CoV-2 signal in wastewater precedes increases in the number of hospitalized cases by four days (N1: R = 0.741, p < 0.001; N2: R = 0.767, p < 0.001). This is further confirmed visually (Fig. 4 ). The precedence of viral RNA signal vs. other epidemiological metrics has been reported elsewhere (Chavarria-Miró et al., 2020; Kumar et al., 2020; Peccia et al., 2020).

Table 1

Time-step analyses of correlations (Pearson's R) between normalized SARS-CoV-2 viral RNA signal (copies/copies PMMoV) and 7-day rolling average percent positivity, test percent positivity and daily new cases epidemiological metrics.

Open in a separate window

Open in a separate window

Fig. 4

Figure of the new daily cases epidemiological metric (in red) with SARS-CoV-2 viral RNA signal in Ottawa's primary clarified sludge samples, with the precedence of SARS-COV-2 viral signal in wastewater clearly outlined. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Recent studies proposed that wastewater viral SARS-CoV-2 RNA signal may in fact precede hospital admissions by 1 to 5 days in primary sludge (Fig. 4) (Kaplan et al., 2020; Peccia et al., 2020), and 6–14 days in raw wastewater (Kumar et al., 2020). This supports findings by Peccia et al. (2020), identifying that hospitalizations may begin peaking four days after an increased viral RNA signal, with the caveat that hospitalizations rates may be significantly higher when viral RNA signal is the result from shedding of an at-risk population group (young infants, elderly, immunocompromised, etc.). In addition, the RNA signal post-resurgence indicates that the resurgence events in viral RNA signal serves to weaken the correlation of subsequent peaks with clinical testing metrics. It is hypothesized that early and later infections and overlapping shedding events of duration and intensity convolute peaks within a period of time following the initial new peaks/surges in COVID-19 infections. Nonetheless, the RNA epidemiological metric of copies/copies PMMoV may become extremely valuable to public health units in their preparation for the next COVID-19 resurgences.

Earlier studies investigating the predictive capabilities of SARS-CoV-2 WBE (Kumar et al., 2020; Peccia et al., 2020) may have been performed at times where clinical metrics were dampened due to limited clinical testing resources, which may in turn have impacted the conclusions regarding predictive ability. This present study was performed during a period of relatively higher testing intensity in the community (avg. 80.9 tests per/100 K pop. performed) and confirms previous findings of early detection of SARS-CoV-2 WBE as compared to clinical metrics such as hospitalizations, new daily positives cases or percent positivity.

4. Conclusions

This study demonstrates that due to the general stability of SARS-CoV-2 RNA measurements in PCS, this fraction of wastewater can reliably be utilized to predict rapid increases or resurgences in COVID-19 cases in the community at large (Alpaslan-Kocamemi et al., 2020; D'Aoust et al., 2021; Peccia et al., 2020), and the results strongly suggest that PCS can be utilized as part of sentinel-type wastewater based epidemiology endeavors at WRRFs.

Advantages of this approach include: i) the analyses in water resource recovery facilities are high-enrollment surveys, which capture a majority of the urban population, regardless of individual's willingness to get tested clinically; ii) tests are anonymous and do not pose ethical challenges as opposed to mass clinical testing or targeted sewershed monitoring efforts; iii) data can be used as a confirmation to clinical tests, increases the effectiveness of local public health units and data remains independent of official testing strategies or media sentiment, and iv) provides an exit strategy/path forward for public health units as something to eventually transition to once immunization is available, to cut down on clinical testing costs, while maintaining broad surveillance capabilities. However, limitations to this WBE approach includes the following: i) data obtained only gives high-altitude view of the situation in the community, and it is currently not possible to correlate these data to an actual number of cases in the community; ii) testing protocols must be optimized and retain sensitivity at very-low levels of disease incidence to remain relevant; iii) widespread adoption of WBE requires scale-up of public health units' current capabilities and/or partnership with private laboratories or research institutions, and iv) increases in resolution (more localized, upstream of water resource recovery facilities, for example) of tests in watersheds may lead to ethical questions for public health unit if it wishes to act upon the data, due to the potential risk for identifying or singling out small subgroups of a population. It is noted however that these issues are not dissimilar to other ethical issues currently existing with other COVID-19 public health endeavors, such as COVID-19 potential exposure notification smartphone applications.

Funding

This research was supported by a CHEO (Children's Hospital of Eastern Ontario) CHAMO (Children's Hospital Academic Medical Organization) grant, awarded to Dr. Alex E. MacKenzie.

CRediT authorship contribution statement

Patrick M. D'Aoust: Investigation, Formal analysis, Writing – original draft. Tyson E. Graber: Validation, Supervision, Writing – review & editing. Elisabeth Mercier: Investigation, Formal analysis, Writing – original draft. Danika Montpetit: Investigation. Ilya Alexandrov: Validation, Writing – review & editing. Nafisa Neault: Validation, Writing – review & editing. Aiman Tariq Baig: Validation, Writing – review & editing. Janice Mayne: Validation, Writing – review & editing. Xu Zhang: Validation, Writing – review & editing. Tommy Alain: Validation, Writing – review & editing. Mark R. Servos: Methodology, Validation, Writing – review & editing. Nivetha Srikanthan: Methodology, Validation, Writing – review & editing. Malcolm MacKenzie: Validation, Writing – review & editing. Daniel Figeys: Validation, Writing – review & editing. Douglas Manuel: Validation, Writing – review & editing. Peter Jüni: Validation, Writing – review & editing. Alex E. MacKenzie: Methodology, Validation, Supervision, Writing – review & editing, Funding acquisition. Robert Delatolla: Methodology, Validation, Supervision, Writing – review & editing, Funding acquisition.

Declaration of competing interest

The authors declare that no known competing financial interests or personal relationships could appear to influence the work reported in this manuscript.

Acknowledgements

Notes

Editor: Thomas Kevin V

Footnotes

References