key: cord-0332625-1yqldvw4
authors: Bowes, D. A.; Driver, E. M.; Kraberger, S.; Fontenele, R. S.; Holland, L. A.; Wright, J.; Johnston, B.; Savic, S.; Newell, M. E.; Adhikari, S.; Kumar, R.; Goetz, H.; Binsfeld, A.; Nessi, K.; Watkins, P.; Mahant, A.; Zevitz, J.; Deitrick, S.; Brown, P.; Dalton, R.; Garcia, C.; Inchausti, R.; Holmes, W.; Tian, X.-J.; Varsani, A. U.; Lim, E.; Scotch, M.; Halden, R. U.
title: Unrestricted Online Sharing of High-frequency, High-resolution Data on SARS-CoV-2 in Wastewater to Inform the COVID-19 Public Health Response in Greater Tempe, Arizona
date: 2021-08-01
journal: nan
DOI: 10.1101/2021.07.29.21261338
sha: 29db5ef70d2e4bedb797bb6fb8cd729fb8215ec8
doc_id: 332625
cord_uid: 1yqldvw4

The COVID-19 pandemic prompted a global integration of wastewater-based epidemiology (WBE) into public health surveillance. Among early pre-COVID practitioners was Greater Tempe (population ~200,000), Arizona, where high-frequency, high-resolution monitoring of opioids began in 2018, leading to unrestricted online data release. Leveraging an existing, neighborhood-level monitoring network, wastewater from eleven contiguous catchment areas was analyzed by RT-qPCR for the SARS-CoV-2 E gene from April 2020 to March 2021 (n=1,556). Wastewater data identified an infection hotspot in a predominantly Hispanic and Native American community, triggering targeted interventions. During the first SARS-CoV-2 wave (June 2020), spikes in virus levels preceded an increase in clinical cases by 8.5+/-2.1 days, providing an early-warning capability that later transitioned into a lagging indicator (-2.0+/-1.4 days) during the December/January 2020-21 wave of clinical cases. Globally representing the first demonstration of immediate, unrestricted WBE data sharing and featuring long-term, innovative, high-frequency, high-resolution sub-catchment monitoring, this successful case study encourages further applications of WBE to inform public health interventions.

21 wave of clinical cases. Globally representing the first demonstration of immediate, unrestricted 48 WBE data sharing and featuring long-term, innovative, high-frequency, high-resolution sub-49 catchment monitoring, this successful case study encourages further applications of WBE to 50 inform public health interventions. 51 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 1, 2021. ; https://doi.org/10.1101/2021.07.29.21261338 doi: medRxiv preprint Triggered by the SARS-CoV-2 pandemic, the use of wastewater-based epidemiology (WBE) 52 as a potentially powerful, rapid, and inexpensive tool to inform public health decision-making has 53 seen a remarkable increase globally. For decades, WBE has been exercised to track chemical and 54 biological threats, with numerous studies underscoring its efficacy and usefulness for 55 understanding and managing community health 1-8 . At the onset of the SARS-CoV-2 pandemic, 56 significant delays in conventional and individualized clinical testing, due in part to an 57 overwhelmed healthcare system and resource limitations 9 , positioned WBE as a promising 58 supplemental tool for assessing SARS-CoV-2 spread at the population-level, a strategy that soon 59 was adopted more broadly [10] [11] [12] [13] . Early data showed SARS-CoV-2 levels in wastewater and sludge 60 as a concomitant or early indicator of clinical confirmed infections, disease and mortality in a 61 community 14, 15 . 62

The City of Tempe, Arizona, residential population ~200,000, had been an early adopter of 63 sharing of monthly wastewater samples, subsequent analysis, and to joint reporting of use-trends 67 of opioids within the community monthly by displaying the obtained collaborative results for 68 oxycodone, codeine, heroin, and fentanyl (and metabolites; µg d -1 per 1000 people) in five urban 69 sewersheds 17 . The City also had established a routine for data analysis and public health response 70 by integrating Tempe Fire Medical Rescue, Human Services (e.g., CARE 7 crisis intervention 71 organization), and others into a workgroup that relied on WBE data as an important and innovative 72 source of information to guide resource deployment by community need (Figure 1) . 73 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted August 1, 2021. ; https://doi.org/10.1101/2021.07.29.21261338 doi: medRxiv preprint With this existing framework in place, Tempe and ASU were in a unique position at the start 74 of the SARS-CoV-2 pandemic to quickly transition into WBE surveillance of SARS-CoV-2. 75

Coincidentally, Tempe and the ASU community also had one of the first early diagnoses of a 76 positive SARS-CoV-2 patient (26 January 2020) 18 . As the City of Tempe and ASU quickly 77 transitioned into molecular-based monitoring, the immediate goal was to use previously 78 established expertise in sampling, infrastructure access, and WBE-framed public health response 79 to begin quantitative assessments of SARS-CoV-2 levels in wastewater. The ultimate objective 80 was to identify hotspots of infection early and implement interventions including education, 81 outreach, and targeted clinical testing to limit the spread of the virus within the Greater Tempe 82 community. The local health department shared data from clinical testing of individuals only at 83 the zip code level, 19 a policy intended to protect small communities and personal identifiable 84 information, which potentially limited stakeholders' ability to respond to local virus clusters. 85

Unique to the US, zip codes are a series of 5 numbers created by the US postal service to delineate 86 small geographical areas within counties to improve mail service, and are used extensively by local 87 and state agencies, including public health departments 20 . The 5-digit, Tempe, AZ zip codes 88 involved in this study are 85281, 85282, 85283, and 85284, and will be referred to here as ZC-1, 89 ZC-2, ZC-3, and ZC-4. The pre-existing, neighborhood-level wastewater monitoring network 90 offered an opportunity to test the potential of WBE to serve as an early warning system that may 91 reveal virus presence and spread prior to clinical case data reported from testing of individuals 21, 92 22 . Thus, important goals of the work were (i) to compare WBE data to newly reported clinical 93 cases of SARS-CoV-2, related hospitalizations, and associated deaths at a high temporal and 94 geospatial resolution (i.e., county, city, zip code, and neighborhood levels), and (ii) to determine 95 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted August 1, 2021. ; https://doi.org/10.1101/2021.07.29.21261338 doi: medRxiv preprint whether the concurrent pandemic monitoring by WBE produced data and information not available 96 or obvious from clinical testing. 97

Neighborhood-level Sampling 99

At the onset of the pandemic, our team had divided the Greater Tempe area into five sewer 100 catchments (Areas 1-5; Figure 2a) , including two additional, non-published locations that received 101 wastewater from adjacent municipalities, which were necessary to determine the Tempe-102 associated sewage signal where wastewater was comingled. The neighborhood-level sampling 103 methodology was synchronized with reoccurring compliance monitoring of the Sub-Regional 104

Operating Group (SROG), a cohort of five municipalities including Phoenix, Tempe, Mesa, 105

Glendale, and Scottsdale, that jointly own and operate the 91 st Avenue wastewater treatment plant 106 (WWTP) in Phoenix, Arizona. The predefined sampling strategy consisted of 7-consecutive days 107 of sample collection each month, across variable weeks from permanent, sub-surface sampling 108 stations. While this sampling strategy was sufficient for long-term, opioid-related monitoring, 109 tracking of SARS-CoV-2 levels required an increased temporal resolution. Accordingly, we 110 adopted a high-frequency sampling approach consisting of weekly collection on Tuesday, 111 Thursday, and Saturday, in addition to the SROG sampling events. To improve spatial resolution, 112 additional sampling locations were also identified based on ease of collection (Area 6), while three 113 other permanent locations needed slight infrastructure modifications (Area 7) and/or approvals 114 prior to onboarding, including the Town of Guadalupe with strong representation by Native 115

American and Hispanic residents (Figure 2a) , and Tempe St. Luke's Hospital (not displayed on 116 dashboard). Permanent sampling locations outside of the Tempe jurisdiction (necessary for 117 eliminating non-Tempe SARS-CoV-2 signals) were available only during the previously 118 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted August 1, 2021. ; https://doi.org/10.1101/2021.07.29.21261338 doi: medRxiv preprint referenced week of compliance monitoring. City of Tempe personnel began sampling from 119 maintenance holes (also known as manholes) immediately downstream of these locations, within 120 the City's jurisdiction, during the three weeks each month when regular compliance sampling was 121 not performed. Figure S1 . 134

The SARS-CoV-2 viral load was calculated for each sample at a collection point using 135 wastewater flow data provided by Tempe ( Figure S2 ). Flow rates in catchment Areas 1-7 had data 136 recorded at 2-min intervals in real-time using permanent laser flow meters, while the Town of 137 Guadalupe and the Tempe St. Luke's Hospital had only historical flow data available. Flow varied 138 from a maximum in Area 1 of 54.5 ± 6.6 million L day -1 (MLD) to a minimum of 0.106 MLD 139 (historical estimate) for the hospital location; average flow rates across all catchments were 15.1 140

CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 1, 2021. ; https://doi.org/10.1101/2021.07.29.21261338 doi: medRxiv preprint At select collection sites, the corresponding wastewater sample was representative of multiple 142 collection catchment areas due to the comingling of wastewater in the collection system ( Figure  143 S3); this occurred in Areas 1-3. To isolate individual catchments and provide a catchment-specific 144 viral load, a mass balance was performed. Resultant viral loads in each sewer catchment ranged 145 from 6 x 10 10 -1 x 10 13 genome copies d -1 (Figure S4) August 2020 ( Figure S4) . 151

Understanding the number of people contributing to wastewater in any given catchment is 152 critical when working with data generated from within the collection system. Estimated Tempe 153 subpopulations ranged from a low of 8,114 ± 848 in Area 5 to a high of 132,082 ± 7,374 in Area 154 1, the largest geographic catchment area (Table S1 ). Variability in Tempe data was a function of 155 the total numbers of residents, employed individuals, and the number of students in the 156 contributing area. The population of the Town of Guadalupe (6,500) was determined using US 157 census data 23 . The hospital location was omitted from this population analysis since the number 158 of individuals working or serving as patients was unknown. 159

The result of these efforts ultimately culminated in eight SARS-CoV-2 collection locations 161 viewable online by the public (Areas 1-7 of Tempe and Town of Guadalupe) on an interactive 162 dashboard that went live the first week of May 2020 (Figure 2) . The dashboard displays each 163 catchment area overlain on a street-level city map so users can geospatially identify contributing 164 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 1, 2021. ; https://doi.org/10.1101/2021.07.29.21261338 doi: medRxiv preprint locations in the catchment. In response to a request of the impacted communities, the Town of 165

Guadalupe is displayed on a separate tab of the dashboard. Data are shown as the logarithm of 166 genome copies L -1 and are presented as a weekly average consisting of the Tuesday, Thursday, 167

and Saturday collected samples. Since the sewage collection system in Tempe separates 168 stormwater from municipal wastewater and the study site is in an arid climate, the use of 169 concentration was permissible. Users have the ability to control spatial and temporal parameters 170 to their preference; and text and infographics accompany these data, which explain WBE basics, 171 how to properly interpret the data, and how data are created and used by the City. Additionally, 172 the SARS-CoV-2 wastewater dashboard is nested in a Community COVID-19 Health Site that Estimating population size by study area was challenging due to the unique nature of collecting 221 wastewater from within the sewer infrastructure rather than by determining the population by 222 counting the residents of local buildings served by a wastewater treatment plant as performed in 223 traditional WBE studies 27, 28 . As a net importer of people to the City for work, it was important 224 not only to quantify the residents but also the non-resident employed and transient student 225

populations, a task accomplished by using Maricopa Association of Governments (MAG) and on-226 campus student resident data provided by ASU. MAG data needed to be corrected for lockdown 227 activities which closed businesses, for which we used Arizona department of transportation arterial 228 traffic flow data (e.g., 40% decrease in arterial traffic equated to 40% decreased in employment 229 populations). Due to the bulk of student classes moving online, using only campus resident data 230 (based on student housing contracts which were updated monthly) was appropriate to assess 231 temporal changes in student populations. These numbers did not account for changes in resident 232 population during holiday travel or off-campus housing locations; thus, overall percentage changes 233 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 1, 2021. ; https://doi.org/10.1101/2021.07.29.21261338 doi: medRxiv preprint in wastewater flow were also used to estimate population size changes. For instance, wastewater 234 flow from Area 6 increased by 20% and was sustained throughout the academic year; therefore, 235 the population in that area was assumed to increase proportionally. This increase in total flow in 236

Area 6 also coincided with increases in viral load, suggesting that infected students were moving 237 back into Tempe for the start of the academic year. Looking ahead to Fall 2021 when classes are 238 expected to resume in-person, quantifying that transient population will become more important 239 for sewersheds impacted by students. We have therefore begun testing the utility of campus Wi-240

Fi data to better estimate population size as students and faculty return to pre-pandemic campus 241 activities 10 . 242

The measured viral loads per day of SARS-CoV-2 within each catchment area in Tempe were 243 aggregated and partitioned to their respective zip codes (ZC-1 through 4) according to their 244 estimated percent contribution (Figure S5) . Wastewater-derived SARS-CoV-2 peaks in activity 245 correlated with newly detected clinical cases per day in three distinct waves of activity: June 2020, 246

August 2020, and December/January 2020-21. ZC-1, home to ASU, was the only zip code that 247 showed viral increases in August 2020. Contributions to viral load within a given community by 248 university students, however, is not an event isolated to Tempe 29-31 . Comparisons between spikes 249 of coronavirus levels in wastewater and clinical case data showed that peaks in wastewater 250 preceded positive clinical cases by 7, 6, 11, and 10 days for ZC-1 through 4 (average of 8.5 ± 2.1 251 days), during the first wave of the pandemic, and again during the isolated university-associated 252 wave, this time by 6 days (Figure 3) . These results align with preliminary assessments of 253 wastewater and clinical case data that suggested monitoring wastewater provided an early-warning 254 capacity ranging between 2 and 21 days 15, 25, 32 . Tempe aggregated viral loads were also compared 255

to Maricopa County Public Health data (Figure 4) . Results again showed peaks in wastewater 256 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 1, 2021. County shows that access to testing was extremely limited during the early stages of the pandemic 272 33, 34 but increased dramatically subsequently with the continuous onboarding of commercial, 273 hospital, and university laboratories, largely driven by ASU biomedical screening. Thus, these data 274 strongly suggest that the greatest benefits of WBE are to be verified early on during the detection 275 of disease outbreaks before health care providers can mount a response. Similar benefits may be 276 reaped late into an epidemic, when clinical testing of individuals becomes cost-prohibitive and 277 may appear unproductive when generating mostly negative results. 278

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 1, 2021. access, actionable data that were shown here to directly help inform and shape the public health 280 response. High-throughput monitoring of the E gene of SARS-CoV-2 in Tempe sewage showed 281 WBE to provide an early-warning benefit, particularly in smaller subpopulations, with a temporal 282 and spatial data resolution that exceeded that of clinical healthcare data, which are shared only to 283 a limited degree with local stakeholders. Use of WBE may also be important for communities with 284 barriers to testing (e.g., lack of access, deficit of testing locations, cost), and testing fear (disbelief), is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 1, 2021. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 1, 2021. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted August 1, 2021. ; https://doi.org/10.1101/2021.07.29.21261338 doi: medRxiv preprint

Measuring biomarkers in wastewater as a new source of 340 epidemiological information: Current state and future perspectives

Evaluation of uncertainties associated with the determination of 343 community drug use through the measurement of sewage drug biomarkers

The spatial epidemiology of cocaine, methamphetamine and 3, 4-346 methylenedioxymethamphetamine (MDMA) use: a demonstration using a population 347 measure of community drug load derived from municipal wastewater

Changes in illicit drug consumption patterns in 2009 detected by 350 wastewater analysis. Drug and alcohol dependence

Sewage epidemiology -A real-time approach to estimate the 352 consumption of illicit drugs in Brussels

Wastewater-based epidemiology biomarkers: Past, present and future

Tracking harmful chemicals and pathogens using the Human Health 357 Observatory at ASU

A new paradigm in public health assessment: Water 359 fingerprinting for protein markers of public health using mass spectrometry. TrAC, 360 Trends in analytical chemistry

High Contagiousness and Rapid Spread of Severe Acute Respiratory 362

363 10. Monitoring SARS-CoV-2 Through Wastewater-Based Epidemiology

First detection of SARS-CoV-2 RNA in wastewater in North 366

America: A study in Louisiana, USA. The Science of the total environment

Early-pandemic wastewater 369 surveillance of SARS-CoV-2 in Southern Nevada: Methodology, occurrence, and 370 incidence/prevalence considerations

COVID-19 surveillance in Southeastern Virginia using wastewater-372 based epidemiology

Measurement of SARS-CoV-2 RNA in wastewater tracks community 374 infection dynamics

SARS-CoV-2 Titers in Wastewater Are Higher than Expected from 376 Clinically Confirmed Cases

SARS-CoV-2 RNA monitoring in wastewater as a potential early 383 warning system for COVID-19 transmission in the community: A temporal case study

RNA in Sewage and Correlation with Reported COVID-19 Prevalence in 387 the Early Stage of the Epidemic in The Netherlands. Environmental science & technology

Computational analysis of SARS-CoV-2/COVID-19 391 surveillance by wastewater-based epidemiology locally and globally: Feasibility, 392 economy, opportunities and challenges

Temporal Detection and Phylogenetic Assessment of SARS-CoV-2 394 in Municipal Wastewater

Evaluation of lockdown effect on SARS-CoV-2 dynamics through viral 396 genome quantification in waste water

Alcohol and nicotine consumption trends in 399 three U.S. communities determined by wastewater-based epidemiology

Wastewater analysis for nicotine, cocaine, amphetamines, opioids and cannabis in New 403

Multiple COVID-19 Clusters on a University Campus -North Carolina

Response to a COVID-19 407 Outbreak on a University Campus -Indiana

SARS-CoV-2 RNA in wastewater anticipated COVID-19 occurrence 411 in a low prevalence area

Substantial underestimation of SARS-CoV-2 infection in the United 413 States

The Need for More and Better Testing 415 for COVID-19