Profiling Pharmaceutical Residues in South African Aquatic Ecosystems: A Dual Approach Using Grab Sampling and Passive Sampling Techniques

Abstract

Pharmaceutical pollution in aquatic ecosystems is a growing environmental concern due to its persistence as a result of the continuous output of pharmaceuticals, their potential to bioaccumulate, and the ecotoxicological risks they pose. This study employed a dual sampling approach, that is grab sampling and passive sampling to assess the occurrence, transport, and distribution of pharmaceutical contaminants in South African freshwater and estuarine systems. Ultra-high-performance liquid chromatography coupled with high-resolution mass spectrometry (UHPLC-Q-Orbitrap MS) was used to analyse pharmaceutical residues across three case studies focusing on urban rivers and estuarine waters. Chapters 1 and 2 provide a comprehensive review of the existing literature related to pharmaceutical residues in aquatic environments, with a particular focus on the South African context. Chapter 1 provides an overview of pharmaceutical pollution in aquatic ecosystems, highlighting its emergence as a critical environmental issue in South Africa. It discusses the sources, persistence, and ecological impacts of pharmaceutical contaminants, emphasizing their ability to remain biologically active at trace and ultra-trace concentrations. The chapter highlights the importance of effective sampling strategies in monitoring these pollutants, comparing traditional grab sampling with passive sampling techniques. It reviews the limitations of current approaches and identifies knowledge gaps, particularly in South African estuarine systems. The chapter concludes by advocating for an integrated sampling methodology to improve the accuracy and reliability of pharmaceutical pollution assessments, thereby informing better environmental management and policy development. Chapter 2 begins by exploring the global and local occurrence of pharmaceutical contaminants, outlining their sources, pathways, and environmental persistence. The chapter then examines various sampling techniques employed for monitoring these contaminants, specifically highlighting the use of grab sampling and passive sampling methods such as the Chemcatcher device. The strengths, limitations, and comparative efficiencies of these techniques are discussed. Analytical techniques for detecting pharmaceutical residues, especially high-resolution methods UHPLC-MS, are also reviewed. Finally, the chapter addresses the regulatory landscape, potential ecological and identifies existing knowledge gaps that justify the need for further research. This review lays the groundwork for the study’s objectives by contextualizing the environmental threat posed by pharmaceutical pollutants and the methodologies suitable for their detection and monitoring. Chapter 4 explored pharmaceutical pollution in the Hennops River and Hartbeespoort Dam, a heavily polluted system influenced by wastewater treatment plant (WWTP) effluents, agricultural run-off and untreated sewage discharges. In the study, a UHPLC-Q-Orbitrap MS method was developed and optimized for the analysis of carbamazepine, dexamethasone, etilefrine, methocarbamol, nevirapine and venlafaxine. Method optimization yielded SPE recoveries ranging from 55.18 – 80.52 % and 55.73 – 83.92 % for ultrapure and river water, respectively. Method LODs and LOQs ranged from 0.15 – 2.03 ng L-1 and 0.49 – 6.75 ng L-1, respectively. A Chemcatcher passive sampler was calibrated for the uptake of the target pharmaceutical compounds. Five of the six pharmaceutical compounds showed good linearity (R2 > 0.97) for the duration of the calibration experiment. Etilefrine, however, was only linear until day eight. As such, the sampling rate for etilefrine could not be determined. For the five compounds that showed good linearity over the 16-day calibration experiment, sampling rates ranging from 0.360 (dexamethasone) to 0.447 (carbamazepine) L day-1 were obtained. Applying the sampling rates, time-weighted average (TWA) concentrations in the ranges 0.50 (methocarbamol) – 17.87 (dexamethasone) µg L-1 were obtained. These were compared to SPE extract concentrations, which ranged from <LOD (venlafaxine) – 84.71 (dexamethasone) µg L-1. Although the SPE maximum extract concentrations were generally higher than Chemcatcher maximum TWA concentrations, Chemcatcher demonstrated enhanced sensitivity, and in some cases captured TWA concentrations that were up to five times higher than SPE. In suspect screening, 25 pesticides and 61 pharmaceutical compounds were detected with the detection frequency higher for Chemcatcher compared to SPE extracts. Chapter 5 investigated the occurrence of pharmaceutical compounds in the Durban Marina estuary using both sampling approaches. The study was expanded to nine pharmaceutical compounds with the addition of lopinavir, metformin and trimethoprim, however, these compounds were not included in the laboratory calibration of the Chemcatcher sampler in seawater. In the Chemcatcher calibration experiment, five compounds showed good linearity (R2 > 0.981). The exception was etilefrine, which was linear only until day eight. The seawater laboratory calibration experiment yielded sampling rates from 0.34 (dexamethasone) – 0.42 (carbamazepine & venlafaxine) L day-1. The analytical method was optimized and SPE recoveries were in the ranges 74.82–99.45% in seawater, and 72.24–103.42% in ultrapure water. Method LODs and LOQs were found to range from 0.15 to 2.18, and 0.50 to 7.27 ng L-1, for venlafaxine and methocarbamol, respectively. Chemcatcher consistently shows a higher detection frequency (88.89%) across all sampling sites, whereas SPE detection frequencies varied more across sites (55.56 – 77.78%). Extract concentration sums per site were found to be higher in Chemcatcher (up to 2574.52 ng L-1 at S2) than SPE (up to 43.81 ngL-1 at S1). Applying the experimental sampling rates, TWA concentrations up to 290.18 ng L-1 (dexamethasone) at S5 were determined. These were compared to the grab water sample concentrations in the sampling medium, which reached a maximum of 0.1594 ng L-1 (nevirapine) at S2. Passive samplers revealed the persistent presence of these contaminants, while grab sampling showed fluctuating concentrations. The study also highlighted the influence of tidal dynamics on pharmaceutical distribution, emphasising the limitations of grab sampling in estuarine environments. Chapter 6 expanded on Chapter 4 by examining the transport and distribution of the nine target pharmaceutical compounds in the Crocodile River system after it exits Hartbeespoort Dam, using the dual sampling approach. The Crocodile River system is a critical water source, which serves mostly agricultural, domestic and industrial activities. The study identified significant pharmaceutical loads, with Chemcatcher consistently demonstrating a higher detection frequency, reaching 100% at S2 and S3, while SPE exhibited greater variability, with detection frequencies ranging from 44.44% (S5) to 88.89% (S2 & S3). As with detection frequency, pharmaceutical concentration sums per site in the extracts were found to be higher with Chemcatcher (up to 1191.05 ng L-1 at S2), confirming passive sampling advantages in long term monitoring. The major contributors to Chemcatcher concentrations were found to be methocarbamol and etilefrine. S3 and S5 exhibited the lowest concentration sums (447.05 ng L⁻¹ and 388.60 ng L⁻¹, respectively), despite S3 having the highest detection frequency. Time weighted average (TWA) concentrations confirmed methocarbamol as the most persistent compound (68.34 ng L⁻¹ at S1), followed by trimethoprim (29.52 ng L⁻¹ at S3) and lopinavir (30.13 ng L⁻¹ at S1). Contaminant transport was influenced by anthropogenic inputs, dilution, and tributary contributions, particularly from the Gwatlhe, Hex and Elands Rivers, which increased contaminant loads at S7 (518.10 ng L⁻¹). The transport and distribution of pharmaceutical contaminants in the Crocodile River system exhibited spatial variability influenced by multiple factors, including anthropogenic inputs, hydrological processes, and the complexity of tributary networks. Overall, this research highlights the widespread occurrence of pharmaceuticals in South African aquatic environments and emphasises the importance of incorporating passive sampling to capturing low-concentration contamination trends. The study advocates for integrating passive sampling into routine monitoring programs to enhance the detection, management, and mitigation of pharmaceutical pollution.