From waste to wisdom: The utility of wastewater surveillance for community-level health information

The success of monitoring SARS-CoV-2 virus in wastewater to understand disease dynamics has been documented throughout the pandemic; however, often missing is an incorporation of the flow and population dynamics into the analysis, as well as the impact of various potential medical interventions. This study compares the wastewater surveillance from two small central Pennsylvania towns: (i) Town A with a relatively stable population except for an active summer tourist season; and (ii) Town B, a college town, with population fluctuating at one WWTP based the percent of in-person class offerings. Methodology Town A's WWTP serves 2 small towns, plus a major hospital, an amusement park, and 2 food processors. Wastewater was sampled once per week starting on November 30, 2020. In Town A, the data were separated into three distinct flow/medical intervention categories (Table 1) to remove flow as a variable. The transition of flow regimes in March 2021 coincided with 5% of the county residents being fully vaccinated. The separation into Summer flow occurred the week after the local schools closed. Town B has two WWTPs. One WWTP is owned by the university (Plant 1), which collects the wastewater generated by the university. The second WWTP (Plant 2) serves the area surrounding Plant 1's service area. Wastewater was collected daily at both plants during much of the academic year starting in August 2020, twice weekly over Summer 2021, and at least three times per week starting in late August 2021. In Town B, the data were separated into the following categories (Table 1): (i) Low Flow when the university was wholly remote or between semesters; (ii) Fall 2020 semester (limited in-person classes); (iii) Spring 2021 (very limited in-person); (iv) Summer 2021 (transition to in-person classes); and (v) Fall 2021 (90+% in-person). Town B's full vaccination rate occurred during the Spring 2021 semester. For each sampling location, virus concentrations from 24-hour composite samples of the influent were quantified using RT-PCR. Samples collected from the WWTP in Town A and both plants in Town B were also analyzed for over-the-counter and prescription medications. Health statistics were obtained from the Pennsylvania's or the university's COVID-19 dashboard, or from the CDC's hospital weekly data. A Spearman correlation was performed using the flow periods above. Loads were correlated but their correlations were weaker than the concentration, likely because flow became a variable. Because of the limited sample size for some parameters, statistically-significant correlation coefficients (p ≤ 0.05) are highlighted in yellow, while correlation coefficients (p ≤ 0.10) are highlighted in light orange. Results and Discussion For Town A, the parameters that were correlated with the virus concentration are in Table 2. Very little was correlated with the virus concentration while the vaccine rollout was less than 5%, except dexamethasone and emergency department visits, which were negatively correlated. Pediatric hospitalizations were positively correlated, although adult hospitalizations were not. As vaccination increased, wastewater viral concentrations steadily declined. As viral concentrations declined, the use of naproxen declined, as did the 7-day new case rate, hospitalization rate, intensive care unit use rate, and emergency department use. In the summer season, however, the virus concentration was positively correlated with vaccine percentage. The town sees an influx of tourists of unknown disease and vaccination status and the potential for asymptomatic diseases/breakthrough cases were documented in the news media. Adult hospitalizations and adult ICU use increased as viral concentrations increased. The use of sulfamethoxazole and trimethoprim decreased as viral concentrations increased; prior research shows that antibiotics generally exhibit seasonal trends in wastewater. For Town B, the university's operational status was important (Table 3), especially at Plant 1. During low-flow, with small campus populations, Plant 1's virus concentration was positively correlated with all measures of the university's COVID cases, as well as the county's 7-day average case rate. For Plant 2, the university's COVID case rates were positively correlated with virus concentrations, but the county's rates were weakly correlated. During the Fall 2020 semester when no vaccines were available, all residence hall rooms were single occupancy, with students also attending classes from off-campus apartments, resulting in none of the university's COVID-19 case measurements being correlated with the virus concentration at Plant 1. At Plant 2, the university's case metrics were positively correlated. The virus concentrations also were positively correlated with county case rates. In the Spring 2021 semester, again with few classes being or in person, all university and county measures of COVID-19 cases were positively correlated with virus concentrations. This trend continued in summer 2021 when the university began offering more courses in person. In fall 2021, with the return to mostly in-person classes and no vaccine mandate, COVID case measures at the university and county were positively correlated with virus concentrations at Plant 1. Positivity rates were not correlated for Plant 2, although the total university cases were correlated. Very few pharmaceuticals were correlated with virus concentration with more correlations occurring early in the pandemic before vaccines were readily available. Pediatric hospitalization data was not available. Prior to vaccine rollout, hospitalizations, ICU use, and death were negatively correlated with virus concentrations. This may be because viral shedding occurs primarily prior to hospitalization, when the case has not yet accessed in-patient medical services. In spring and summer 2021, as vaccines were available, deaths decreased as viral concentrations decreased at Plant 2, likely due to growing immunity. When vaccination coverage was less than 5%, vaccine percent increases were positively correlated to increases in virus concentration. Similar to Town A, vaccine distribution was insufficient to reduce disease spread. During Spring 2021 when the vaccine was available, increasing vaccine uptake correlated with decreasing virus concentrations, except at Plant 1, the university's WWTP. This may be because student vaccinations were recorded at their permanent address, and that students may not have gotten vaccinated during the semester. The negative correlation between vaccination uptake and virus concentration continued into summer 2021 for Plant 2. At Plant 1, virus concentration and vaccine uptake were positively correlated, which may reflect the increasing student numbers and potentially delayed student vaccinations. For Fall 2021, virus concentrations were positively correlated with vaccine uptake. Similar to Town A, only a very small increase in vaccine uptake was seen in Summer and Fall 2021. The increasing case rate may be due to breakthrough cases of the vaccinated or due to the return of students to campus, with those not reporting vaccination status being tested weekly Limited data is available so far for Fall 2021, but the first six weeks saw an initial uptick in cases and then decreasing to a steady case rate and virus concentration. Conclusions Flow rate and the timing of the deployment of various medical interventions are significant variables in wastewater-based epidemiology studies. In this study, the SARS-CoV-2 virus concentrations in the wastewater of two small towns were split into various flow phases, which also coincided with the rollout of vaccines and the transient nature of the population of the college town. For both towns, the rollout of vaccines to the wider population resulted in the decrease in viral concentrations in wastewater, with the exception of the wastewater at the university treatment plant. This reduction of viral concentrations was positively correlated with decreases in cases and hospitalizations for COVID-19. Pharmaceuticals were not regularly correlated positively or negatively with viral concentrations in the wastewater. This more detailed analysis, by tying medical interventions and population size to the flow segment, allows for an improved understanding of how wastewater can monitor the progress of the pandemic.