Spatially resolved qualified sewage spot sampling to track SARS-CoV-2 dynamics in Munich - One year of experience

Abstract

Wastewater-based epidemiology (WBE) is a tool now increasingly proposed to monitor the SARS-CoV-2 burden in populations without the need for individual mass testing. It is especially interesting in metropolitan areas where spread can be very fast, and proper sewage systems are available for sampling with short flow times and thus little decay of the virus. We started in March 2020 to set up a once-a-week qualified spot sampling protocol in six different locations in Munich carefully chosen to contain primarily wastewater of permanent residential areas, rather than industry or hospitals. We used RT-PCR and sequencing to track the spread of SARS-CoV-2 in the Munich population with temporo-spatial resolution. The study became fully operational in mid-April 2020 and has been tracking SARS-CoV-2 RNA load weekly for one year. Sequencing of the isolated viral RNA was performed to obtain information about the presence and abundance of variants of concern in the Munich area over time. We demonstrate that the evolution of SARS-CoV-2 RNA loads (between <7.5 and 3874/ml) in these different areas within Munich correlates well with official seven day incidence notification data (between 0.0 and 327 per 100,000) obtained from the authorities within the respective region. Wastewater viral loads predicted the dynamic of SARS-CoV-2 local incidence about 3 weeks in advance of data based on respiratory swab analyses. Aligning with multiple different point-mutations characteristic for certain variants of concern, we could demonstrate the gradual increase of variant of concern B.1.1.7 in the Munich population beginning in January 2021, weeks before it became apparent in sequencing results of swabs samples taken from patients living in Munich. Overall, the study highlights the potential of WBE to monitor the SARS-CoV-2 pandemic, including the introduction of variants of concern in a local population.

Keywords: B.1.1.7; COVID-19; PCR; Sequencing; Surveillance; Wastewater.

Graphical abstract

Fig. 1

City map of Munich with the sampling areas highlighted. Sampled neighbourhoods are spread across the city and include about 1/3rd of the total population. The small region 2 includes the largest university clinics of Munich. For details of the respective sampling areas see Table 1; Geographic North is indicated by the arrow; size bar represents 10 km.

Fig. 2

Incidence rates (cumulative seven day incidence per 100,000 inhabitants) over the last year for the drainage areas under investigation (areas 1–6, black). Copy numbers of the SARS-CoV-2 N1-gene target (CDC protocol) are indicated in red, expressed as the number of copies per 100 ml of sewage. The lower limit of detection (LOD) of the PCR reaction (equivalent to 7.58 copies per 100 ml sewage) was used to plot negative samples. The solid line represents the LOESS (locally estimated scatterplot smoothing or local regression) modelling the incidence rates (in black) and the viral loads (in red). The grey regions represent the 95% CI of the LOESS estimate. The viral load in sewage precedes the incidence based on notifications by roughly 3 weeks. For reference the cumulative seven day incidence per 100,000 inhabitants for the same timeframe for the whole of Munich (MUC) and for the non-sampled areas (MUC-R), interpolated with a LOESS. The dashed vertical line separates 2020 from 2021. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 3

Cross correlation values, for identifying lag time (in weeks) of the viral load-variable predictors of incidence rate. Regions where the most dominant cross correlations occur above or below horizontal lines. A lag of roughly three weeks (dashed vertical line) is observed, between sewage viral load and notification data. The solid line represents the LOESS (locally estimated scatterplot smoothing or local regression) modelling of the cross correlation values over the different sampling areas. The grey regions represent the 95% CI of the LOESS estimate. See Supplemental Fig. 1B for specific sampling site information. Cross correlation values, for identifying lag time (in weeks) of the viral load-variable predictors of incidence rate. Regions where the most dominant cross correlations occur above or below horizontal lines. A lag of roughly three weeks (dashed vertical line) is observed, between sewage viral load and notification data. The solid line represents the LOESS (locally estimated scatterplot smoothing or local regression) modelling of the cross correlation values over the different sampling areas. The grey regions represent the 95% CI of the LOESS estimate. See Supplemental Fig. 1B for specific sampling site information.

Fig. 4

A: Selected signature mutations for the variant of concern lineage B.1.1.7 are exemplarily shown over time in a heatmap for the sampling site 6 (Savitsstr.) between July 2020 and March 2021. Percentages (starting at 0) represent the fraction of the indicated single nucleotide polymorphisms detected at the respective time point. Grey fields (NA) depict no coverage (less than 20 mapped reads) at the respective genome position in the sample. No sustained signature mutation signals were detected before mid of January (19th). Subsequently, the proportion of key mutations increased rapidly and reached high levels by the beginning of March 2021. Signature mutations for P1 and B.1.351 were not detected over the study period. B: Baggtitr chart of reported percentage of B.1.1.7 in sequenced SARS-CoV-2 swabs between calendar weeks 1 and 9 of 2021 in Munich. Dots represent frequencies of B.1.1.7 mutations detected in sewage as plotted in A. The solid line represents the LOESS (locally estimated scatterplot smoothing or local regression) modelling the different signature mutations. The grey region represents the 95% CI of the LOESS estimate. Blue bars represent confirmed sequenced B.1.1.7 cases, black and yellow represent B.1.351 and P1 respectively. Red represents other variants than the three aforementioned variants of concern; green bar is identification of S1 mutants by hybridization assays without definitive confirmation by sequencing. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)