Reproducibility and sensitivity of 36 methods to quantify the SARS-CoV-2 genetic signal in raw wastewater: findings from an interlaboratory methods evaluation in the U.S.

Abstract

In response to COVID-19, the international water community rapidly developed methods to quantify the SARS-CoV-2 genetic signal in untreated wastewater. Wastewater surveillance using such methods has the potential to complement clinical testing in assessing community health. This interlaboratory assessment evaluated the reproducibility and sensitivity of 36 standard operating procedures (SOPs), divided into eight method groups based on sample concentration approach and whether solids were removed. Two raw wastewater samples were collected in August 2020, amended with a matrix spike (betacoronavirus OC43), and distributed to 32 laboratories across the U.S. Replicate samples analyzed in accordance with the project's quality assurance plan showed high reproducibility across the 36 SOPs: 80% of the recovery-corrected results fell within a band of +/- 1.15-log10 genome copies/L with higher reproducibility observed within a single SOP (standard deviation of 0.13-log10). The inclusion of a solids removal step and the selection of a concentration method did not show a clear, systematic impact on the recovery-corrected results. Other methodological variations (e.g., pasteurization, primer set selection, and use of RT-qPCR or RT-dPCR platforms) generally resulted in small differences compared to other sources of variability. These findings suggest that a variety of methods are capable of producing reproducible results, though the same SOP or laboratory should be selected to track SARS-CoV-2 trends at a given facility. The methods showed a 7-log10 range of recovery efficiency and limit of detection highlighting the importance of recovery correction and the need to consider method sensitivity when selecting methods for wastewater surveillance.

Comparison of approaches to quantify SARS-CoV-2 in wastewater using RT-qPCR: Results and implications from a collaborative inter-laboratory study in Canada

“…The notion that currently deployed methods can reliably distinguish lesser from greater amounts of viral surrogates at the conditions investigated is also supported. These inferences are largely consistent with other international studies performed in parallel ( Pecson et al., 2020 ; Pocock et al., 2020 ). Both of these parallel studies concluded that sample preparation methods did not have a clear, systematic impact on results ( Pecson et al., 2020 ; Pocock et al., 2020 ).…”

“…These inferences are largely consistent with other international studies performed in parallel ( Pecson et al., 2020 ; Pocock et al., 2020 ). Both of these parallel studies concluded that sample preparation methods did not have a clear, systematic impact on results ( Pecson et al., 2020 ; Pocock et al., 2020 ). Remarkably, there is substantially less variability in the results amongst the methods deployed in this study than that observed in the parallel study of 36 methods ( Pecson et al., 2020 ).…”

“…Given the lack of evidence that all surrogates—including HCoV-229E that was spiked in—comparably represent the recovery of authentic in situ SARS-CoV-2 viruses from wastewater, the effort required to perform matrix spikes in every sample (that might be desirable to characterize recovery variability attributable to changing wastewater matrix conditions) is likely not warranted. In the parallel study by Pecson et al., 2020 , five of six additional surrogates introduced as matrix spikes exhibited statistically different recoveries than their primary choice of spiked-in recovery surrogate (human coronavirus strain OC43). Therefore, a preferred approach might be to report quantitated SARS-CoV-2 concentrations prior to adjusting for recovery, along with an estimate of the recovery efficiency itself and the matrix spike surrogate used.…”

“…Both of these parallel studies concluded that sample preparation methods did not have a clear, systematic impact on results ( Pecson et al., 2020 ; Pocock et al., 2020 ). Remarkably, there is substantially less variability in the results amongst the methods deployed in this study than that observed in the parallel study of 36 methods ( Pecson et al., 2020 ). This observation might be linked to 1) the limited range of methods captured in this study (i.e., primarily concentration methods involving PEG precipitation and ultrafiltration) amongst various other types of sample preparation methods that exist (e.g., aluminum salt coagulation, skim milk flocculation, charged membrane filtration) and/or 2) the use of known, quantified SARS-CoV-2 surrogate spikes in this work rather than the quantification of in-situ SARS-CoV-2 in the parallel study.…”

“…Second, the choice of surrogates and methods used to assess recovery should be verified (additional details elaborated upon in the recommendations on process controls below). Third, the benefits and weaknesses inherent to the design of these inter-laboratory comparisons are highlighted: spike-and-recovery approaches (such as the one applied herein this study) might provide greater confidence of intended spiking levels but may not be truly representative of authentic in-situ SARS-CoV-2 behavior, while inter-laboratory studies that estimate in-situ SARS-CoV-2 to wastewater samples (e.g., Pecson et al., 2020 ) might be limited by inevitable uncertainty arising from not being able to know the true concentration, but may better reflect the ability to detect authentic in-situ SARS-CoV-2 in wastewater.…”