Normalization of wastewater-based epidemiology data for pathogen surveillance: a case study of campus-wide SARS-CoV-2 surveillance at a South African university

Abstract

Background

Wastewater-based epidemiology (WBE) has emerged as a valuable tool for monitoring community-level SARS-CoV-2 exposure during the COVID-19 pandemic. However, several limitations of WBE have been identified, which have hindered its wider application. One key challenge is interpreting pathogen incidence data meaningfully, considering that measured pathogen concentrations can be influenced by external factors such as dilution by greywater in combined sewer systems and the sampling method used. This study aimed to evaluate data normalization strategies for viral copy numbers measured in samples collected passively from sewer lines at a South African university.

Methods

Wastewater samples were collected weekly over a one-year period from sewer lines at seven on-campus sites at the Nelson Mandela University. Passive, 3D-printed, torpedo-style samplers containing standard medical gauze as adsorbent were deployed directly into sewer lines for 9 hours to interact with wastewater. The gauze was then retrieved from the samplers and solids were eluted into PBS with 0.05% Tween 80. Total RNA was subsequently extracted from the samples using the Qiagen RNeasy Powersoil kit, followed by quantification of the RNA concentration using a NanoDrop spectrophotometer. SARS-CoV-2 copy numbers were determined using RT-qPCR with primers and probes targeting two regions in the nucleocapsid gene, namely N1 and N2. RT-qPCR was also employed to quantify the copy numbers of two commonly used viral normalizers, namely aichi virus (AiV) and pepper mild mottle virus (PMMoV). SARS-CoV-2 copy numbers were then normalized against AiV and PMMoV copy numbers and against extracted RNA concentration. Normalized and unnormalized SARS-CoV-2 data were evaluated against clinical numbers using Spearman correlation to determine the most effective normalization strategy.

Results

Normalization against AiV showed weak correlations with clinical case numbers (r=0.29), and AiV was not consistently detected in all samples. Normalizing SARS-CoV-2 data against PMMoV data improved correlations significantly when compared with unnormalized SARS-CoV-2 (r=0.67 vs r=0.44; P≤0.05). The strongest correlation with clinical case data was obtained when SARS-CoV-2 copy numbers were normalized against initial RNA concentrations (r=0.81; P≤0.05).

Discussion

When employing passive sampling to collect wastewater samples for the quantification of pathogens for epidemiology, the traditionally used normalization strategies that apply community and physicochemical parameters and flow rates cannot be employed, especially in mixed grey- and blackwater systems. Normalizing against extracted RNA concentration is not affected by diet, takes into account dilution of pathogens by greywater and the variability in RNA extraction between samples, and improved the correlation between wastewater pathogen concentrations and clinical case numbers.

Conclusion

Normalizing SARS-CoV-2 data from passively collected wastewater samples against extracted RNA concentrations enhances the reliability of the data. This normalization strategy should be further evaluated for WBE of other pathogens and for different sampling methodologies.