Quantifying community-wide antibiotic usage via urban water fingerprinting: Focus on contrasting resource settings in South Africa

There has been a significant increase in antimicrobial agents (AAs) usage, globally - however the relative consumption is unevenly distributed between nations. Inappropriate use of antibiotics can harbour inherent antimicrobial resistance (AMR); therefore, it is important to understand and monitor community-wide prescribing and consumption behaviours throughout different communities around the world. Wastewater-Based Epidemiology (WBE) is a novel tool enabling low cost and large scale studies focussed on AA usage patterns. The back-calculation of community antimicrobial intake was performed from quantities measured in municipal wastewater and informal settlement discharge in the city of Stellenbosch, utilising WBE. Seventeen antimicrobials, and their human metabolites, were evaluated, in concordance with prescription records corresponding to the catchment region. The proportional excretion, biological/chemical stability, and method recovery of each analyte were all crucial factors in the efficacy of the calculation. Mass per day measurements were normalised to the catchment area via population estimates. Municipal wastewater treatment plant population estimates were used to normalise the wastewater samples and prescription data (mg/day/1000 inhabitants). Population estimates for the informal settlements were less accurate due to a lack of reliable sources that were relevant to the sampling time period. Both mass loads and normalised loads suggested higher than average usage throughout the settlements, relative to municipal wastewater. This was seen most prominently in emtricitabine and lamivudine; but also, sulfamethoxazole, trimethoprim, sulfadiazine, clindamycin, ciprofloxacin, ofloxacin, and doxycycline. Urban water fingerprinting (UWF) data triangulation with prescription datasets showed good correlations for several antimicrobial agents (AAs) (e.g., clindamycin, clarithromycin, ofloxacin, and doxycycline). It also revealed discrepancies in usage for some compounds (e.g., tetracycline and sulfapyridine). This might be linked with a lack of pharma compliance in prescription datasets; erroneous association of prescription boundaries with the sewerage catchment; and/or uncertainties within the sewerage catchment (e.g., population estimations). The UWF tool provided a comprehensive overview of multiclass AAs usage, both prescription and over-the counter. For example, tetracycline was not reported in available prescription statistics, but was detected at an average of 18.4 mg/day/1000inh; and no antiviral prescriptions were obtained, but emtricitabine and lamivudine were quantified at 2415.4 and 144.4 mg/day/1000inh, respectively. A lack of clarity regarding prescriptions and a lack of inclusion of several critical (often over-the-counter) medications in public health databases makes WBE a useful and comprehensive epidemiology tool for tracking pharma usage within a catchment.