Subsewershed analyses of the impacts of inflow and infiltration on viral pathogens and antibiotic resistance markers across a rural sewer system

Abstract

As wastewater-based surveillance (WBS) is increasingly used to track community-level disease trends, it is important to understand how pathogen signals can be altered by phenomena that occur within sewersheds such as inflow and infiltration (I&I). Our objectives were to characterize I&I across a rural sewershed and assess potential impacts on viral (rotavirus, norovirus GII, and SARS-CoV-2), fecal indicator (HF183, the hCYTB484 gene specific to the human mitochondrial genome, and crAssphage), and antimicrobial resistance (intI1, bla_CTX−M-1) targets. In a small town in Virginia (USA), we collected 107 wastewater samples at monthly intervals over a 12-month period (2022–2023) at the wastewater treatment plant (WWTP) influent and 11 up-sewer sites. Viral, fecal indicator, and antimicrobial resistance targets were enumerated using ddPCR. Physicochemical proxies for organics and nutrient levels in sewage (chemical oxygen demand (COD), total suspended solids (TSS), and NH₃(aq)) and genetic markers of anthropogenic impact were used to characterize I&I across the sewershed. Overall, precipitation was negatively associated (Spearman test; ρ < 0; p < 0.01) with physicochemical markers (TSS, COD, K, PO₄³⁻-P, NH₃(aq)) in the WWTP influent. We observed the highest concentrations of human fecal markers and a measure anthropogenic pollution and antibiotic resistance (intI1) in up-sewer sites with limited I&I. However, median viral gene copy concentrations were highest at the WWTP, compared to 100 % (n = 11), 90 % (n = 10), and 55 % (n = 6) of up-sewer sites for rotavirus, norovirus GII, and SARS-CoV-2, respectively. After adjusting for covariates (Ba, COD, dissolved oxygen, groundwater depth, precipitation, sampling date) using generalized linear models, moderate to high I&I was associated with statistically significant reductions in log₁₀-transformed rotavirus and norovirus GII concentrations across the sewershed (coefficients = -0.7 and -0.9, p < 0.001, n = 95), though not for SARS-CoV-2 (coefficient = -0.2, p = 0.181, n = 95). Overall, we found that while I&I can diminish biomarker signals throughout a sewershed, including at the WWTP influent, I&I impacts vary depending on the target, and pathogen biomarker signals were, on average, higher and less variable over time at the WWTP compared to most up-sewer sites. As far as we are aware, this is the first study to assess in situ I&I impacts on multiple WBS targets. Taken together, our findings highlight challenges and tradeoffs associated with different sampling strategies for different WBS targets in heavily I&I impacted systems.