ABSTRACT
Wastewater-based epidemiology (WBE) has been suggested as a useful tool to predict the emergence and investigate the extent of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). In this study, we screened appropriate population biomarkers for wastewater SARS-CoV-2 normalization and compared the normalized SARS-CoV-2 values across locations with different demographic characteristics in southeastern Michigan. Wastewater samples were collected between December 2020 and October 2021 from nine neighborhood sewersheds in the Detroit Tri-County area. Using reverse transcriptase droplet digital polymerase chain reaction (RT-ddPCR), concentrations of N1 and N2 genes in the studied sites were quantified, with N1 values ranging from 1.92 × 102 genomic copies/L to 6.87 × 103 gc/L and N2 values ranging from 1.91 × 102 gc/L to 6.45 × 103 gc/L. The strongest correlations were observed with between cumulative COVID-19 cases per capita (referred as COVID-19 incidences thereafter), and SARS-CoV-2 concentrations normalized by total Kjeldahl nitrogen (TKN), creatinine, 5-hydroxyindoleacetic acid (5-HIAA) and xanthine when correlating the per capita SARS-CoV-2 and COVID-19 incidences. When SARS-CoV-2 concentrations in wastewater were normalized and compared with COVID-19 incidences, the differences between neighborhoods of varying demographics were reduced as compared to differences observed when comparing non-normalized SARS-CoV-2 with COVID-19 cases. This indicates when studying the disease burden in communities of different demographics, accurate per capita estimation is of great importance. The study suggests that monitoring selected water quality parameters or biomarkers, along with RNA concentrations in wastewater, will allow adequate data normalization for spatial comparisons, especially in areas where detailed sanitary sewage flows and contributing populations in the catchment areas are not available. This opens the possibility of using WBE to assess community infections in rural areas or the developing world where the contributing population of a sample could be unknown.