key: cord-0764328-pf4zu8p4
authors: Habtewold, Jemaneh; McCarthy, David; McBean, Edward; Law, Ilya; Goodridge, Larry; Habash, Marc; Murphy, Heather M.
title: Passive sampling, a practical method for wastewater-based surveillance of SARS-CoV-2
date: 2021-09-11
journal: Environ Res
DOI: 10.1016/j.envres.2021.112058
sha: b50657e774e51f7478118e98e23eedf4f18e0a2e
doc_id: 764328
cord_uid: pf4zu8p4

In search of practical and affordable tools for wastewater-based surveillance of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), three independent field experiments were conducted using three passive sampler sorbents (electronegative membrane, cotton bud, and gauze) in Guelph, Ontario, Canada. Total daily cases during this study ranged from 2 to 17/100,000 people and 43/54 traditionally collected wastewater samples were positive for SARS-CoV-2 with mean detectable concentrations ranging from 8.4 to 1780 copies/ml. Viral levels on the passive samplers were assessed after 4, 8, 24, 48, 72, and 96 hrs of deployment in the wastewater and 43/54 membrane, 42/54 gauze, and 27/54 cotton bud samples were positive. A linear accumulation rate of SARS-CoV-2 on the membranes was observed up to 48 hours, suggesting the passive sampler could adequately reflect wastewater levels for up to two days of deployment. Due the variability in accumulation observed for the cotton buds and gauzes, and the pre-processing steps required for the gauzes, we recommend membrane filters as a simple cost-effective option for wastewater-based surveillance of SARS-CoV-2.

To track the spread of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), 31 researchers around the globe are using wastewater-based epidemiology (WBE) to understand the 32 temporal and spatial spread of the virus 1 . SARS-CoV-2 infects the absorptive enterocytes of the 33 human gastrointestinal tract 2 and large amounts of the virus are shed into municipal wastewater 34 through feces [3] [4] [5] . Accordingly, studies have quantified the virus RNA in raw wastewater 6 and 35 sludge 7 where its genetic material may persist unaffected for days 8 . Interestingly, the virus has 36 been detected in wastewater days before clinical cases were reported 9 , and include 37 asymptomatic 10 and pre-symptomatic populations. Therefore, monitoring of SARS-CoV-2 RNA 38 is gaining attention as a promising tool in tracking the spread at the community level and serve 39 autosamplers are not always feasible at every sampling site (e.g., insecure sites, no power, deep 48 sewer mains and within buildings). Moreover, use of autosamplers is costly and processing 49 liquid samples involves time-consuming sample processing techniques prior to nucleic acid 50 extraction. 12 Therefore, development of practical and inexpensive sampling tools that can be 51 used at any site and can reduce processing time to provide timely data, are critical in 52 streamlining WBE surveillance. 53 54 Passive sampling, whereby a material is directly deployed in wastewater to sorb the virus over 55 time, may provide an inexpensive and practical alternative to autosampling 13 . Liu et al. 14 used a 56

Moore swab (i.e., pieces of gauzes) tied to a fishing line for surveillance of SARS-CoV-2. 57

Although they deployed the swabs for 24-72 hrs, accumulation of SARS-CoV-2 over time, 58 which is critical in establishing ideal deployment times, was not shown. At larger scales (i.e., lot, 59 suburb, and city), Schang et al. 15 evaluated three passive sampler materials (i.e., sorbents) 60 including electronegative membranes, gauzes, and cotton buds (Q-tips) for detection of SARS-61

CoV-2 in wastewater. They installed these materials in a torpedo-like perforated device that 62 continuously exposes the materials to flowing wastewater. They reported greater sensitivity of 63 SARS-CoV-2 from wastewater over set durations. However, this is precisely the evidence 71 required to demonstrate that passive samplers can adequately represent time-averaged 72 wastewater concentrations 17 . Non-linear uptake may either suggest that the passive sampler has 73 reached its capacity during its exposure period, the adsorption rates are too low to reliably 74 measure SARS-CoV-2 on the sampler, or that adsorption and desorption rates are similar 17 . 75

Consequently, to fill this gap, the goal of the present study was to evaluate whether the three 76 passive sampler materials developed by Schang et al. 15 follow a linear process in taking up 77 SARS-CoV-2 (i.e., follow first-order kinetics). We performed three replicate experiments at a 78 wastewater pilot facility where we deployed the three materials (electronegative membranes, 79 gauzes, and cotton buds) and retrieved them at 4, 8, 24, 48,72 and 96 hrs to assess their ability to 80 accumulate SARS-CoV-2 in raw wastewater. 81

Sampling site and the passive sampling tools 83

We conducted a controlled experiment at a pilot-scale wastewater facility (Text S1) in the city of 84 Guelph, ON, using an apparatus to simulate conditions in a sewer main (Text S1, Fig. S1 ). The 85 apparatus was fed a continuous flow (7.8L/min) of raw wastewater. Six pairs of torpedo-style 86 samplers were deployed horizontally in the PVC pipe apparatus and remained submerged in the 87 wastewater stream. They were tied to a piece of stiff tubing and placed in the PVC pipe (Fig.  88   S1 ). Each pair of torpedoes contained either three cotton buds and three membranes or three 89 gauzes and three cotton buds. Details on the design of the samplers are described in Schang et 90 al. 15 . In this study, we processed all of the membranes and the gauzes in triplicate for each time 91

point. An autosampler (Sigma 900 MAX) was connected at the inflow to the PVC pipe, 92 composited wastewater samples. These samples were used to assess the relative accumulation of 94 SARS-CoV-2 on the passive samplers over time. Membranes were then stored at -80 °C until RNA extraction. Viral concentrations during the 107 longer deployment times (i.e., 48, 72, and 96 hrs) were calculated by applying time-weighted 108 mean concentration principle. Filtration blanks of tap water were performed on-site at the end of 109 experiment 1 and 3 to assess for contamination during the filtration process of liquid samples. 110

Electronegative membranes (from filtered composite samples collected by the autosampler and 112 torpedoes) and cotton buds were directly used for RNA extraction, whereas gauze samplers were 113 first eluted and then filtered on membranes prior to extraction. Elution was performed per 114 Schang et al. 15 with slight modifications (see Text S2). Extraction recovery efficiency (%) was 115 SARS-CoV-2 stability) 18, 19 , recoveries were calculated as (copies recovered/copies spiked) 117 ×100 (Fig. S5 ). RNA extraction was performed using RNeasy PowerMicrobiome Kit (Qiagen) 118 with some modifications (Text S2). 119

Prior to running TaqMan probe-based reverse transcription quantitative real-time PCR (RT-120 qPCR), tests for potential inhibition were performed for 81/216 samples (37.5%) (details in Text 121 S3). To further evaluate inhibition, dilutions were performed on 200/216 samples. If the sample 122 still showed evidence of inhibition through dilution (i.e. concentrations increased during 123 dilution), the lower dilutions were used in the calculation of copies per well. The CDC 124 emergency use authorization kits (2019-nCov CDC EUA Kit) which contain N1 and N2 primer-125 probe sets (IDT, Kanata, Canada) and Reliance One-Step Multiplex RT-qPCR Supermix 126 (BioRad, Hercules, CA) were used to perform one-step RT-qPCR reaction. Details on the 127 calibration curves and limit of detection can be found in Text S4. To support potential 128 accumulation of SARS-CoV-2 on the passive samplers with deployment time, we also quantified 129 the pepper mild mottle virus (PMMoV), an abundant human fecal marker20. Probes, primers and 130 thermocycling conditions and standard curve efficiencies, R 2 , slope, and Y-intercepts for both 131 fragment was more sensitive (i.e., consistently detectable) in most of the samples when 141 compared to the N2 gene fragment. Therefore, we present the SARS-CoV-2 results based on the 142 N1 gene fragment. There were some non-detects (NDs) in the technical replicates of the RT-143 qPCR assays, thus, to avoid overestimation by taking NDs as blanks, NDs were replaced by half 144 the LOD value (4.15, see text S4) when at least one replicate was positive. Using recovery 145 adjusted copies, we calculated accumulation ratios over time for SARS-CoV-2 and PMMoV, 146 calculated as Cp/Cw, where Cw= concentration of target in the composite wastewater sample for 147 that time frame and Cp = concentration of target in the passive sampler for that same time frame. 148

The recovery data are presented in Fig S5. All data reported for SARS-CoV-2 and PMMoV were 149 adjusted for recovery and reported as geometric means. 150

Performance of passive samplers 152

The study was conducted when daily SARS-CoV-2 incidence rates (i.e., new daily confirmed 153 cases) in Guelph were lower (2 to 17 /100,000 people) than the rates seen in January and late in 154

April 2021 (Fig. S3a) . Considering a 25-day duration of positive signals on feces 21 , the total 155 estimated active COVID-19 cases during the study ranged from 211 to 484 per 100,000 people 156 ( Fig. S3b ). At these caseloads, both the composite and passive samplers, particularly membranes 157 and gauzes, were able to sorb detectable amounts of SARS-CoV-2 RNA within 24 hrs of 158 deployment. Wu et al. 22 , who conducted a large campaign of wastewater surveillance in the US, 159

indicated that detection of SARS-CoV-2 is most likely when daily incidence rates are greater 160 than 13 /100,000 people. However, it is important to note that SARS-CoV-2 titers in wastewater 161 symptomatic individuals may also be significant 4, 24 . 163 164 In all three trials, where 216 passive and liquid wastewater samples were processed, total RNA 165 content on the passive samplers increased with deployment time (S4b, c, d) . While membranes 166 and cotton buds accumulated RNA gradually, gauzes had quicker saturation where the level of 167 RNA at 4 hrs were equivalent to the RNA contents of membranes and cotton buds after 48 hrs 168 ( Fig. S4b, c, d) . 169 170 Extraction recoveries varied between the materials (2 to 24%) ( Fig. S5 ; Text S5). 171

Positivity rates for SARS-CoV-2 were comparable among composite, membrane, and gauze 172 samples, which had 43/54, 43/54, and 42/54 positives, respectively (Table S1 ). Consistent with 173 the results of Schang et al 15 , the lowest positivity rate was observed for the cotton buds where 174 27/54 samples were negative, although in our study positivity increased with time (i.e., 2/9 at 4 175 hrs to 7/9 at 96 hrs). The cotton buds were significantly smaller than the other materials which 176 may have contributed to this observation. Like the cotton buds, there was an increase in 177 positivity with time for the membrane and gauze samplers. Interestingly, after 24 hrs of that changes in the geometric means of SARS-CoV-2 gene copies (gc) in the wastewater were 187 not significant between the three experimental trials (p =0.45). When detected, mean copies 188 recovered on the membrane, cotton bud and gauze passives samplers ranged from 270 to 1640 189 gc/ membrane, 379 to 2230 gc/ cotton bud and 609 to 3600 gc/ gauze, respectively (Table S3) . 190

The detection capabilities of the passive materials were also assessed by measuring the quantities 191 of PMMoV (Table S4 ). Concentrations of PMMoV in the wastewater ranged from 1.58+E05 to 192 4.06E+07 gc/ mL. Mean copies recovered on the membrane, cotton bud and gauze passives 193 samplers ranged from 9.50E+04 to 1.31E+08 gc/ membrane, 3.66E+05 to 3.81E+09 gc/ cotton 194 bud and 1.13E+06 to 5.36E+09 gc/ gauze, respectively. Interestingly, by 8 hours, PMMoV 195 copies in gauzes reach capacity that membranes and cotton could accumulate after 96 hrs of 196 deployment, (Table S4) . 197

To compare the relative accumulation of SARS-CoV-2 and PMMoV over time on the different 200 passive samplers, we calculated accumulation ratios whereby we divided the concentration on 201 the passive sampler by the time weighted concentration from the wastewater for the same time-202 period (Fig. 1, 2) . In the three trials, the accumulation of SARS-CoV-2 on the membrane 203 samplers was linear up to 48 hours and then was variable between 48 and 96 hours (Fig. 1) , 204 Accumulation ratios were low and variable for cotton buds, except at 96 hrs where the 205 accumulation ratio spiked to 23 (Fig. 1) . The accumulation ratios for gauze samplers were 206 variable and did not follow a trend. The current gold standard for WBE is processing of liquid composite samples often over a 24-hr 253 period. Passive samplers provide a viable and possibly superior alternative as they are 254 continuously exposed to the wastewater stream and, as a result, may capture variabilities in the 255 wastewater missed by composite methods. 256 257 From a methodological standpoint, processing of composite and gauze samples involve at least 258 one additional step prior to extraction. Both require filtration on electronegative membranes after 259 preparation of the time-composited wastewater samples and gauze-elution, respectively. The 260 point. However, the additional processing steps (i.e., elution and filtration) required for the gauze 262 are time consuming and might require the use of process controls. Cotton buds, which showed 263 mediocre detection of SARS-CoV-2 RNA after 24 hrs of deployment, did not require additional 264 sample processing or manipulation but technical challenges during nucleic acid extractions were 265 greater for cotton buds than the membrane filters (Text S6). 

Prevalence of Gastrointestinal Symptoms and Fecal Viral Shedding in Patients With 286

Coronavirus Disease 2019: A Systematic Review and Meta-analysis

Fecal viral 289 shedding in COVID-19 patients: Clinical significance, viral load dynamics and survival analysis

Viral load dynamics and disease severity in patients infected with 295

retrospective cohort study. 296 BMJ 2020

Coronavirus-2 RNA in Sewage and Correlation with Reported COVID-19 Prevalence in the 310

Early Stage of the Epidemic in The Netherlands

Clinical Course and Molecular Viral

Shedding Among Asymptomatic and Symptomatic Patients With SARS-CoV-2 Infection in a 315

Community Treatment Center in the Republic of Korea

Surveillance of Wastewater for Early Epidemic Prediction (SWEEP): Environmental and 319 health security perspectives in the post COVID-19 Anthropocene

SARS-CoV-2 in wastewater: State of the knowledge and research 323 needs

Passive and grab sampling methods to assess 325 pesticide residues in water. A review

A 327 Novel COVID-19 Early Warning Tool: Moore Swab Method for Wastewater Surveillance at an 328 Institutional Level

Passive sampling of 332 viruses for wastewater surveillance

A novel passive sampling approach for

SARS-CoV-2 in wastewater in a Canadian province with low prevalence of COVID-19

Passive sampling

Persistence and Suitability as an Enveloped Virus Surrogate

bacteriophage Phi6 (a surrogate for SARS-CoV-2) in evaporated saliva microdroplets deposited 343 on glass surfaces

SARS-CoV-2 Titers in Wastewater Are 356

Higher than Expected from Clinically Confirmed Cases

viral load dynamics, duration of viral shedding, and infectiousness: 359 a systematic review and meta-analysis. The Lancet Microbe

Reverse 365 transcriptase droplet digital PCR shows high resilience to PCR inhibitors from plant, soil and 366 water samples

The authors would like to acknowledge the following sources of funding for this research, the 274 National Science and Engineering Council of Canada (Grant # 401655), and the Canada 275Research Chairs Program (Grant # 950-232787). 276