Population diet and stress estimates using wastewater based epidemiology

Abstract

Consumption or exposure to chemicals by a population contributing to a wastewater catchment can be measured using wastewater-based epidemiology (WBE). Ever since its beginnings, WBE has largely focussed on monitoring illicit and licit drug biomarkers in populations. It has also been suggested as a method for monitoring metabolites arising from food consumption and population stress. However, progress in these areas were limited. The present project aimed to expand the WBE repertoire to be able to measure population stress and diet composition, and to use the resulting methods to gain insights into drivers of diet and stress in Australian populations. A literature search was conducted to identify small molecule biomarkers which could be used to measure aspects of diet (e.g. meat consumption, fruit consumption) and stress (e.g. oxidative stress, acrolein exposure) using WBE. A direct injection liquid chromatography tandem mass spectrometry (LC-MS/MS) method was developed to measure these compounds from wastewater. The suitability of each biomarker for WBE was assessed based on (i) stability of biomarkers under sewer conditions, and (ii) comparison of biomarker loads in Australian wastewater with urinalysis measurements in the literature, and (iii) variations in daily per capita loads of the biomarkers. Almost all biomarkers failed criteria (i), as they were unstable in sewer reactors, where stability is defined as degrading less than 10% in a 12-hour period. However, two plant polyphenol consumption biomarkers, three vitamin B consumption biomarkers and a citrus biomarker may be suitable as non-quantitative biomarkers for WBE as they satisfied criteria (ii) and (iii). The remaining diet and stress biomarkers could not be recommended for use in WBE studies using direct injection LC-MS/MS, as they failed to satisfy the three criteria, or were not detected using the analytical method employed. A spatial study of Australian wastewater found that loads of a histamine turnover biomarker examined the correlations between antihistamines and 1,4-methylimidazoleacetic acid (MIAA), an endogenously produced urinary biomarker of histamine burden. Strong correlations were found between MIAA and fexofenadine (R2 = 0.68, p<0.0001) and cetirizine (R2 = 0.25, p=0.03), but not ranitidine. While per capita loads of MIAA measured using WBE were in agreement with the literature, it should not be used in a strictly quantitative manner due to its potential to degrade rapidly under sewer conditions. This study was the first published study that used WBE to find associations between a biomarker of illness and biomarker of treatment of illness. It demonstrated that WBE could be used to measure histamine turnover in populations sing a simple direct injection LC-MS/MS method, and contributes towards expanding WBE applications beyond monitoring. Next, I aimed to understand the socioeconomic and sociodemographic factors associated with the diet biomarkers (as well as the drug, pharmaceutical and personal care product biomarkers). This was achieved by using Australian wastewater samples and sociodemographic data collected as a part of the Australian Census in 2016. Using georeferencing software, we computed 42 social, economic and demographic attributes of the populations served by 22 wastewater treatment plants (WWTPs). We looked for correlations between this dataset and the daily per capita loads of diet biomarkers, as well as various pharmaceuticals, personal care products, licit drugs and illicit drugs from each WWTP. Many associations were evident between the consumption of different chemicals, and socioeconomic status, occupation, education and other socioeconomic variables. Dietary fibre, citrus, vitamin B and caffeine consumption biomarkers had significant correlations with socioeconomic index. Education, occupation and high rent payments were consistently strongly correlated with the diet biomarkers; incomplete high schooling was negatively associated with the dietary fibre biomarkers enterodiol (R = -0.720) and enterolactone (R = -0.737), high rent payments were positively associated with two vitamin B3 biomarkers (R = 0.662, R = 0.608) and the vitamin B6 biomarker pyridoxic acid (R = 0.534). These results demonstrated the utility of WBE understanding socioeconomic disparities in chemical consumption patterns, as opposed to the traditional role of reporting chemical consumption rates. Overall, this thesis demonstrates that WBE can be used to measure certain diet and stress biomarkers. It is important to note, however, that the diet and stress biomarkers should not be used quantitatively without careful consideration, as they degrade significantly under sewer conditions. Instead, by triangulating WBE results of diet and stress biomarkers with other data, such as other WBE biomarkers or socioeconomic data, new insights into population consumption and exposure behaviour may be found.