Epidemiological monitoring of sewage sludge and wastewater from an upflow anaerobic sludge blanket reactor using SARS-CoV-2 as a model

DNA

Deoxyribonucleic acid

RNA

ribonucleic acid (Ácido Ribonucleico)

SARS-CoV-2

Severe Acute Respiratory Syndrome Coronavirus 2

COVID-19

Coronavirus Disease 2019

SARS-CoV

Severe Acute Respiratory Syndrome Coronavirus

UASB

upflow anaerobic sludge blanket digestion

WWTP

wastewater treatment plant

ddPCR

digital droplet polymerase chain reaction

The Coronavirus Disease 2019 (COVID-19) pandemic has made it necessary to test large numbers of individuals for Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2). However, the high cost of testing every potentially infected individual in a population and the scarcity of PCR reagents during SARS-CoV-2 outbreaks, in addition to the delay between infection and the appearance of symptoms, have made mass testing impractical or ineffective in some regions.

With over 20 million cases and more than 500 thousand deaths in 2021, Brazil was exceptional in terms of both numbers of cases and confirmed deaths. Yet, until October 2020 less than 7% of the Brazilian population had been tested (Marinho et al. 2021). Even in 2022, despite their reduced cost, rapid tests and PCR testing remained inaccessible to the majority of Brazilians. Thus, the development and implementation of alternative methods of monitoring would be desirable for tracking the true scale of pandemics in this country.

Because SARS-CoV-2 is released into the feces of infected patients, including asymptomatic patients, a more practicable approach to assisting in SARS-CoV-2 monitoring, and thus in governmental decisions aimed at containing and mitigating the COVID-19 epidemic, is to monitor the virus in wastewater. Indeed, wastewater monitoring has been carried out in many countries, such as the Netherlands (Medema et al. 2020), Australia (Ahmed et al. 2020), the United States (Wu et al. 2020a), India (Kumar et al. 2020), and Brazil (Claro et al. 2021; Fongaro et al. 2021; Prado et al. 2021). In addition, SARS-CoV-2 has been detected in sewers since the beginning of the pandemic (Medema et al. 2020; Tang et al. 2020; Wu et al. 2020b; Zhang et al. 2021). Considering that each sewer network is used by tens of thousands of people, monitoring sewers could provide essential epidemiological data at relatively low cost, especially in regions or countries where there is limited testing of the population.

Brazil has the largest number of anaerobic reactors in the world (Chernicharo et al. 2018), and upflow anaerobic sludge blanket digestion (UASB) reactors are the most common type. Other countries with several UASB plants in operation include Colombia, India, and Japan (Arthur et al. 2022). Reactors of this type form a concentrated sludge bed, which may be considered a composite sample and can be easily sampled. Since coronaviruses have a greater affinity for solids than non-enveloped viruses (Ye et al. 2016), extracting the virus directly from the solid, concentrated part of the sewage, i.e. the sludge, could be an effective way to monitor SARS-CoV-2 – or other viruses – dynamics in the population, especially in municipalities where the composite sampling of wastewater does not take place.

A key advantage of sludge-based epidemiology is that it does not require an additional concentration step before viral detection. Because the virus naturally binds to solids and accumulates in the sludge, samples already contain a higher concentration of viral particles compared to raw wastewater, which often requires concentration via methods like ultracentrifugation, filtration, or chemical precipitation (Rashid et al. 2024). This reduces both time and cost, making sludge sampling a simpler and more efficient approach for monitoring viral loads in sewage networks, especially in resource-limited settings. This makes it particularly advantageous for regions with limited access to advanced laboratory equipment needed for concentration techniques. However, despite the potential benefits, while several works monitored SARS-CoV-2 in wastewater (almost 100), to date only seven works did so in sludge (Shah et al. 2022). The feasibility of UASB's sludge-based epidemiology may open the opportunity for developing countries to implement a relatively cheaper and easier monitoring compared to wastewater-based epidemiology. Sludge-based epidemiology could also be extended beyond SARS-CoV-2, offering opportunities for monitoring other viruses of public health concern, such as poliovirus, measles virus, hepatitis A and E, norovirus, and rotavirus. Studies have demonstrated the utility of wastewater-based epidemiology (WBE) in detecting these pathogens (Xagoraraki & O'Brien 2020), and adapting this approach to sludge could provide a cost-effective tool for epidemiological surveillance in diverse settings.

In the present work, we analyzed SARS-CoV-2 viral loads in sludge samples as well as wastewater from a UASB reactor of the Serraria sewage treatment plant, Porto Alegre (RS). Serraria is the largest treatment plant in Porto Alegre, receiving the sewage of approximately 760,000 inhabitants (50% of the city's sewage). It is also the only treatment plant offering the possibility of wastewater composite sampling, thus allowing us to compare the viral loads found in sludge with those found in wastewater. In addition to that, we employed probes from the TaqMan SARS-CoV-2 Mutation Panel (ThermoFisher) to identify, by RT–qPCR, variants of concern (VoCs) that could be circulating in the population. Our aim was to assess the potential of sludge-based monitoring using UASB reactors as a complementary method to WBE, with SARS-CoV-2 serving as a model pathogen. Specifically, we investigated the capacity of sludge to capture viral loads and mutations and its potential to provide valuable epidemiological insights in resource-limited contexts.

The wastewater treatment plant (WWTP) Serraria is the largest and thus the most representative WWTP in Porto Alegre; its eight UASB reactors treat the wastewater of approximately 760,000 inhabitants (50% of the city's population), including the wastewater of some of the most important hospitals of Rio Grande do Sul, at an average of 2,710 L of wastewater per second with a maximum capacity of 4,114 L/s. Samples of sewage sludge and wastewater were collected every Tuesday from 19 January 2021 to 31 August 2021. Composite samples from the influent wastewater were obtained using an automatic sampler: 500 mL samples were taken every hour over periods of 24 h, during which the temperature of samples was kept at 4 °C.

The UASB reactor at the Serraria WWTP has a depth of 6 m, a maximum design upflow velocity of 1.24 m/h, and an average upflow velocity of 0.82 m/h. The sampling height was 0.5 m. Sludge samples were obtained through a sampling faucet: 500 mL of sludge was collected directly from the sludge bed of the reactor during the last hour of the composite sampling of wastewater. The sludge sampled represents the solids sedimented over 20–30 days. Samples were then transported at 4 °C to the laboratory and kept at the same temperature during processing.

Wastewater was concentrated by ultracentrifugation (Pina et al. 1998). Briefly, samples of raw wastewater (36 mL) were centrifuged at 110,000 g and 4 °C for 1 h. Supernatants were discarded, and pellets were resuspended in 4 mL of 1 N glycine pH 9.5, followed by 30 min of incubation on ice, with vortexing every 5 min. Four mL of 2× phosphate buffered saline (PBS 2 × , pH 7.4) was added, and samples were clarified at 4,000 g for 20 min to remove in-suspension particles. Supernatants were then centrifuged at 110,000 g for 1 h, and the pellets resuspended in 500 μL of PBS and stored at −80 for RNA extraction. All products and recipients used throughout the process were sterile to prevent contamination.

Total RNA was extracted from concentrated samples of wastewater (200 μL of eluate) using a Maxwell^® 16 Viral Total Nucleic Acid Purification Kit and Maxwell^® RSC automatic extractor (Promega), according to the manufacturer's instructions. Total RNA was extracted from 2.5 mL of sludge samples with an RNeasy PowerSoil Total RNA kit (Qiagen). Purified RNA was resuspended in 50 μL of ribonuclease-free water and stored at −80 °C.

SARS-CoV-2 was quantified by a one-step quantitative reverse transcriptase–polymerase chain reaction (qRT–PCR), using an AgPath-ID™ One-Step RT-PCR (ThermoFisher) kit and the N1 primer set targeting the virus nucleocapsid recognized by the US Centers for Disease Control (Vogels et al. 2020). To avoid inhibition by contaminants, a 10-fold dilution of sludge RNA was performed before the RT-PCR assay. Five μL of the diluted RNA was used in 20 μL reactions that were run under the following conditions: 50 °C for 15 min, 95 °C for 10 min followed by 45 cycles of 95 °C for 15 s and 55 °C for 30 s. Reactions were considered positive when the cycle threshold was below 40 cycles. Positive samples were quantified using a standard curve, run in parallel, employing serial dilutions (1:10) of a SARS-CoV-2 positive control containing a known concentration of genomic copies previously quantified by digital droplet PCR (ddPCR). The detection limit considered was five genomic copies (gc)/μL. Each qPCR run included SARS-CoV-2 RNA positive controls and a negative control to ensure the accuracy and reliability of the results. The negative control consisted of ultrapure Milli-Q water only, without any template RNA, to confirm that there was no contamination or false positives during the amplification process. The final copies/L calculations were adjusted by the volume assayed.

Numbers of adult patients hospitalized for COVID-19 in the intensive care units of public and private hospitals were retrieved from the official website of the city of Porto Alegre, where a panel of epidemiological data was updated daily over the period analyzed. The numbers of COVID-19 cases in the regions supplying the WWTP were retrieved from the official COVID-19 website of the government of Rio Grande do Sul (Secretaria Estadual da Saúde do Rio Grande do Sul, n.d.); when a neighborhood was only partially served by the WWTP, the number of cases from that neighborhood was adjusted accordingly.

We employed the Shapiro–Wilk test to evaluate if the viral loads fit a normal distribution. The Shapiro–Wilk test results indicated that the viral load per liter in both sludge and wastewater does not follow a normal distribution with 95% confidence. The viral load in sludge had a Shapiro–Wilk value of 0.808 and a p-value of 5.88 × 105, while the viral load in wastewater had a Shapiro–Wilk value of 0.794 and a p-value of 3.10 × 10⁵, confirming the non-normality of the data in both cases. We then employed the Spearman rank correlation (non-parametric) test to measure degrees of association. An alpha of 5% was used in both tests.

We also performed cross-correlation analysis to examine the temporal relationships between viral loads in sludge and wastewater and epidemiological data (new cases and hospitalizations). Calculated Spearman rank correlation values while shifting epidemiological data to account for lags ranging from −20 to +20 days. The results were visualized to identify potential lead-lag associations, contributing to insights into the timing of changes in viral loads relative to epidemiological trends. All tests were performed using the PANDAS – version 01.5.1 Python package and plots where made using Matplotlib – version 3.9.2 and Seaborn – version 0.13.2.

We used primers and probes approved by the TaqMan SARS-CoV-2 Mutation Panel (ThermoFisher), to identify by RT-PCR the following nine mutations: N501Y, E484K, K417T, K417N, delH69V70, P681H, L452R, E484Q, and P681Ri in 15 samples with cycle thresholds from 27.9 to 34.3 obtained in different months between December 2020 and August 2021. The nine mutation variants selected for this study were chosen based on their prevalence during the sampling period and their association with VoCs identified in clinical samples from the region (Salvato & Gregianini 2021).

During the period from January to August 2021, 32 sludge samples were analyzed. Thirty (93%) were positive for SARS-CoV-2, and copy numbers varied from 0.147 to 2.314 × 10⁶ copies/L. In the wastewater samples collected on the same days, 31 (97%) were positive for SARS-CoV-2, and copy numbers varied from 0.058 to 3.014 × 10⁶ copies/L.

From May to August, the number of new RT–qPCR-positive individuals remained as low as 50 per day; nevertheless, it was still possible to detect and quantify SARS-CoV-2 in both wastewater and sludge samples (Figure 1). Furthermore, we were also able to identify mutations in viruses recovered from the sludge. This shows that this kind of sampling may be suitable for long-term monitoring of SARS-CoV-2, since it was able to provide information about the virus in the population even in the absence of outbreaks.

While these correlations provide useful insights, it is essential to interpret them cautiously due to inherent limitations in lag analysis. Testing multiple lag spans increases the risk of identifying spurious or coincidental correlations that may not reflect true temporal relationships. Despite these caveats, the observed trends in Figure 3 highlight important dynamics. For instance, the close alignment of wastewater viral loads with new cases reinforces its value for timely outbreak detection and response. Similarly, the moderate correlation between sludge viral loads and hospitalizations suggests that sludge monitoring could provide early indications of severe disease burden, even if it is less effective for tracking the overall trajectory of infections.

All 15 tested samples were positive for the presence of the reference alleles, and from March 2021 we also identified the mutations N501Y, E484K, and K417T, corresponding to the Gamma variant (Table 1). Mutations P681R and L452R, corresponding to the Delta variant, were detected from the beginning of August 2021.

The implementation of alternative ways to monitor the SARS-CoV-2 pandemic is particularly desirable in countries with limited resources for testing a large number of individuals. Since the beginning of the COVID-19 pandemic, wastewater monitoring has been employed in many studies (Medema et al. 2020) and has shown that this is a promising way to monitor levels of circulating viruses within a population (Peccia et al. 2020). However, the best results in wastewater monitoring require automatic composite sampling and sample concentration by ultracentrifugation, which in turn require expensive devices not widely available in developing countries. Other approaches might therefore be useful.

Considering that coronaviruses have a greater affinity for solids (Ye et al. 2016), they would be expected to end up in the sludge formed from wastewater. Sludge is easily sampled and can serve as a composite sample. Therefore, our aim was to test the viability of using sludge samples from a UASB reactor, the most common reactor used in WWTPs in Brazil, as an alternative to wastewater for monitoring SARS-CoV-2 in the population.

The viral copy numbers found in sludge were similar to those found in wastewater. However, while the viral loads detected in wastewater closely paralleled the rise and fall of new COVID-19 cases and both measurements anticipated 2–4 weeks in the number of hospitalizations in the regions served by the WWTP, the peak of the viral load in sludge was delayed by 3–4 weeks compared to the peak in new cases (Figure 1), and by 1–2 weeks compared to the number of hospitalizations (Figure 2). Sludge data also had fluctuations in estimates of viral loads.

Thus, according to our results, monitoring the sludge, unlike monitoring the wastewater, would not be an effective surrogate for measurements of actual numbers of new cases. The delayed kinetics of the appearance of the virus in the sludge could be explained by the slow sedimentation of the sludge in the UASB reactor (20–30 days). The fluctuations in estimates of viral loads in sludge, in turn, could be explained by the use of manual sampling and RNA extraction, in contrast to the automated sampling and RNA extraction used for wastewater. Peccia et al. (2020) have previously quantified SARS-CoV-2 copies in the sludge from a gravity thickener and were able to detect an increase in viral loads ahead of both new cases and hospitalizations in the city. However, the WWTP they studied employed a different sludge treatment method that is not common in Brazil, in which the solid residence time is only 4 h. Daily sampling and duplicate extractions were also conducted in Peccia's work; this can significantly increase the precision of viral load estimates, but it is not economically feasible for long-term monitoring.

Modifications in the methodology need to be tested; for example, sampling different heights in the sludge bed of the UASB reactor may help to explain the delay in detecting viral copies. Also, additional controls for RNA extraction and for organic deposits in samples might reduce the fluctuations seen in estimates of viral copies in sludge. Further improving SARS-CoV-2 monitoring via the sludge of UASB reactors could be important since this could provide a significant alternative source of epidemiological data that would be more widely available in developing countries, notably Brazil. Moreover, this methodology could be used to monitor different pathogens, such as poliovirus, measles virus, hepatitis A and E viruses, noroviruses, and rotaviruses, that are already monitored via wastewater-based epidemiology (Xagoraraki & O'Brien 2020).

Although sludge monitoring does not seem to be an effective way of anticipating infection numbers, it proved possible to identify known VoCs in sludge using a relatively cheap PCR analysis even in samples from periods with a low number of cases. The early identifications in the sludge of both the Gamma and Delta variants matched in time the predominance of these variants in Rio Grande do Sul as communicated in the weekly genomic bulletins from the state surveillance organ (Salvato & Gregianini 2021) that monitors clinical samples in the state. Our results show that monitoring sludge has great potential for identifying VoCs since it was possible to identify predominant variants circulating in the population by PCR analysis. Moreover, with a more robust meta genomic analysis, it might even be possible to identify variants not yet identified in the clinic, as has been done with wastewater samples (Crits-Christoph et al. 2021).

Beyond SARS-CoV-S, wastewater surveillance has demonstrated significant utility in monitoring a range of pathogens, including SARS-CoV-2, norovirus, influenza A, RSV, and Mpox. Studies highlight their complementary roles in public health, offering early warning of outbreaks and insights into community transmission trends. For instance, norovirus monitoring showed high correlations between wastewater viral loads and syndromic data, suggesting its effectiveness in predicting outbreaks earlier than traditional surveillance methods (Ammerman et al. 2023). Similarly, wastewater tracking of influenza A and RSV revealed viral concentrations preceding emergency department visits, highlighting its value in seasonal respiratory virus surveillance (DeJonge et al. 2023).

Wastewater monitoring also proved effective for detecting Mpox in both high- and low-prevalence areas. In Chile, Mpox deoxyribonucleic acid (DNA) was detected in wastewater, correlating with case increases and even identifying viable viruses (Ampuero et al. 2023). In El Paso, Texas, wastewater signals appeared weeks before the single reported Mpox case, emphasizing its capacity to uncover unreported or asymptomatic infections (Oghuan et al. 2023). Sludge surveillance, due to its practicality and cost-efficiency, emerges as a promising alternative for long-term pathogen monitoring. Its slower accumulation dynamics and lower sampling frequency requirements make it particularly suitable for tracking trends in resource-limited settings. By leveraging these findings, sludge with slight changes in methodology can be tested for diverse pathogens, enhancing public health preparedness and response efforts.

We acknowledge that many characteristics of UASB sludge-based epidemiology remain to be explored. Despite that, this study represents a pioneering effort to showcase the feasibility of this approach and inspire further research in this domain. We believe this work will encourage additional studies to elucidate critical aspects, such as detailed mechanisms of virion enrichment, variability across reactor types, and broader applications for monitoring other pathogens.

Our findings suggest that sludge monitoring is a promising and cost-effective methodology for tracking SARS-CoV-2 dynamics and identifying circulating variants in urban populations. While the delayed viral kinetics and fluctuations in viral load estimates highlight some limitations, these may be addressed with methodological improvements, such as testing new sampling methods. With further refinements, sludge monitoring could yield reliable viral load measurements that correlate more closely with active case numbers, offering a valuable epidemiological tool in resource-limited settings. Additionally, sludge-based epidemiology has broader applicability beyond SARS-CoV-2. Its low cost and practicality may make it particularly suitable for monitoring other pathogens, such as poliovirus, hepatitis A and E, noroviruses, and rotaviruses in middle- and low-income countries. The natural viral concentration in sludge eliminates the need for additional processing steps, reducing technical barriers and costs. Moreover, the ability to detect VoC underscores its potential for identifying novel pathogens and guiding public health interventions, making it a critical complement to global infectious disease surveillance systems.

We are most grateful to Dr Paulo Robinson da Silva Samuel (DMAE, Porto Alegre) for insights into the sewer systems, which were crucial for the design of experiments; Adriana Cechin, Lucimara Comunello, and Eric Farias De Souza for their invaluable help with sludge sampling; and Martin Cancela for his instruction and help with the RT-qPCRs; Professor Rodrigo de Freitas Bueno (UFABC, Santo André) for reviewing the statistical analysis; and Professor Julian Gross from the University of Oxford for reviewing the text and giving excellent suggestions.

This work was partially funded by FAPERGS/MS/CNPq 08/2020 – PPSUS and the Coordination of Improvement of Higher Education Personnel (CAPES). Funding sources had no involvement in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; and in the decision to submit the article for publication.

J.P.S.W. wrote the original draft preparation, investigated the project. M.F.R. contributed to resources. B.A.P. investigated the work. A.C.F. wrote the reviewed and edited the article. C.R. and F.H. supervised the work, contributed to project administration and funding acquisition.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.