ABSTRACT
Background: The World Health Organisation recommends wastewater based epidemiology (WBE) for SARS-CoV-2 as a complementary tool for monitoring population-level epidemiological features of the COVID-19 pandemic. Yet, uptake of WBE in low-to-middle income countries (LMIC) is low. We report on findings from SARS-CoV-2 WBE surveillance network in South Africa, and make recommendations regarding implementation of WBE in LMICs Methods: Seven laboratories using different test methodology, quantified influent wastewater collected from 87 wastewater treatment plants (WWTPs) located in all nine South African provinces for SARS-CoV-2 from 01 June 2021 to 31 May 2022 inclusive, during the 3rd and 4th waves of COVID-19. Regression analysis with district laboratory confirmed SARS-CoV-2 case loads, controlling for district, size of plant and testing frequency was determined. The sensitivity and specificity of rules based on WBE data to predict an epidemic wave based on SARS-CoV-2 wastewater levels were determined. Results: Among 2158 wastewater samples, 543/648 (85%) samples taken during a wave tested positive for SARS-CoV-2 compared with 842 positive tests from 1512 (55%) samples taken during the interwave period. Overall, the regression-co-efficient was 0,66 (95% confidence interval=0,6-0,72, R squared=0.59), but ranged from 0.14 to 0.87 by testing laboratory. Early warning of the 4th wave of SARS-CoV-2 in Gauteng Province in November-December 2021 was demonstrated. A 50% increase in log-copies SARS-CoV-2 compared with a rolling mean over the previous 5 weeks was the most sensitive predictive rule (58%) to predict a new wave. Conclusion: Variation in the strength of correlation across testing laboratories, and redundancy of findings across co-located testing plants, suggests that test methodology should be standardised and that surveillance networks may utilise a sentinel site model without compromising the value of WBE findings for public health decision-making. Further research is needed to identify optimal test frequency and the need for normalisation to population size, so as to identify predictive and interpretive rules to support early warning and public health action. Our findings support investment in WBE for SARS-CoV-2 surveillance in low and middle-income countries.