Emerging Pharmaceutical Pollutants in Urban South Africa: Environmental and Public Health Impacts

1. Introduction

The expansion of the urban population and increased pharmaceutical consumption especially in metropolitan regions, have led to rising levels of pharmaceutical contaminants, commonly referred to as Emerging pharmaceutical pollutants (EPPs) [1,2,3]. These pollutants enter aquatic systems primarily through wastewater discharge and inadequate treatment infrastructure, posing significant risk to environmental integrity and public health [4]. EPPs include antibiotics, hormones, and anti-inflammatory drugs, as well as personal care products like parabens and triclosan [5,6,7]. Their persistence and ability to induce antibiotic resistance and endocrine disruption make them particularly concerning for both aquatic organisms and human populations [7,8].

Globally, more than 5,000 pharmaceutical products are approved for use, where surface waters are routinely monitored [9]. In contrast, South Africa has prioritized monitoring for only about 52 EPPs, despite widespread reliance on rivers and dams for water supply in communities lacking adequate sanitation [7]. This gap in monitoring, combined with poor wastewater infrastructure, contributes to elevated contamination in metropolitan areas such as eThekwini.

The World Health Organization, emphasizes that safe water environments are critical for disease prevention [10]. However, the Department of Water and Sanitation [11] reported that only 0.3% of microbiological tests, 3.1% of chemical tests, and 7.5% of physical compliance tests from effluent assessments met required standards. More recent reports showed widespread non-compliance, in with KwaZulu-Natal (WWTWs) failing the green drop certification [12]. These results highlight the urgent need for enhanced monitoring and treatment systems to safeguard both environmental and public health.

Rivers such as the uMlazi, Mbokodweni, and iSipingo, are heavily impacted by industrial effluents, domestic wastewater, and illegal sewage connections [13]. Additional pressures include improper waste disposal, such as nappies dumped directly into streams. Climatic factors further exacerbate pollution loads. intense rainfall results established that in increased flow rates and elevated EPP concentrations [14],while flooding in the has been linked to heightened diseases incidence. These combined pressures contribute to declining water quality and biodiversity loss, particularly in urban areas with inadequate infrastructure such as leaking sewer pipes [5,15].

Amanzimtoti, located in the eThekwini Municipality, had a population of roughly 14,000 in 2011, within a compact area of about 9 km², leading to a high population density [16]. Significantly, over 17% of its inhabitants were senior citizens (aged 60+), a percentage above that of numerous South African regions, indicating an above-average elderly demography for this "Main Place." Several studies have reported pharmaceutical pollution in eThekwini’s rivers and WWTWs. NSAIDs such as Diclofenac (DIC), Ibuprofen (IBU), Naproxen (NAP) and Fenoprofen (FEN), have been reported in the Mbokodweni river and Amanzimtoti WWTW [3,6,17,18]. Psychotropic drugs, including carbamazepine, were detected at Amanzimtoti WWTW [15,19]. While antiretroviral drugs such as efavirenz and emtricitabine were consistently found in surface waters and WWTWs across Durban [20].

South Africa has the highest global HIV burden, with approximately 8.0 million people on antiretroviral therapy [5,21,22]. In eThekwini, HIV prevalence among individuals aged 15–49 is estimated at 16.6% [23,24]. ARV drugs such as emtricitabine, tenofovir disoproxil, and efavirenz have been frequently detected in surface waters [20]. Their presence poses risks to aquatic ecosystems and public health, with concerns about biodiversity impacts and the emergence of drug-resistant pathogens [25,26]. A Souza-Silva, Bispo [27] study demonstrated that ARVs can alter the behaviour and reproduction of aquatic organisms, reinforcing the need for catchment-based monitoring strategies.

Passive sampling such as polar organic chemical integrative samplers (POCIS), and Chemcatcher® are widely used to monitor polar organic pollutants [28]. In this study, bond elut plexa cartridges were selected for solid-phase extraction (SPE) methods due to theory proven efficiency in capturing a broad spectrum of semi-polar and polar pharmaceuticals with reported recoveries ranging from 69% to 101% for 16 PPCPs in surface and drinking Pedrouzo, Borrull [29] and recoveries of 62–110% (RSD 0.56–4.68%) for NSAIDs, ARVs, and a lipid regulator across different seasons [20].

By focusing on township rivers such as iSipingo, uMlazi, and KwaMashu, densely populated areas with high pharmaceutical consumption that discharge into the Indian Ocean yet remain under-researched, this study addresses critical monitoring gaps. The results aim to strengthen understanding of EPP dynamics in urban and peri-urban South African contexts, support improved environmental management, and inform water resource protection strategies. Ultimately, these insights will safeguard aquatic ecosystems, promote public health, and contribute to socio-economic resilience in the face of climate-related extremes.

2. Experimental

2.1. Chemicals and Materials

High-purity standards (>98%) of target pharmaceutical compounds, including DIC, IBU, NAP, FEN, Gemfibrozil (GEM), Carbamazepine (CAR), and Efavirenz (EFA), were obtained from Sigma-Aldrich (Steinheim, Germany). HPLC-grade solvents (acetonitrile, methanol, acetic acid) were sourced from Merck (Darmstadt, Germany). Ultrapure water was generated using a Purite Select HP 40 system (Purite Ltd, UK). Solid-phase extraction was performed using Bond Elut Plexa cartridges (3 mL, 200 mg styrene divinyl benzyl, Phenomenex, USA). Sodium chloride (NaCl) was supplied by Associated Chemical Enterprise (Johannesburg, South Africa). Water samples were filtered through 0.45 μm hydrophilic polypropylene membrane filter papers and Whatman 70 mm × 100 circle filter papers, purchased from GE Healthcare UK Ltd Supplies (South Africa). For the extraction of the target compounds from the water samples, a vacuum SPE manifold sourced from Phenomenex (California, USA) attached to a vacuum pump purchased from Pall Corporations (Fribourg, Switzerland).

A stock solution of 100 mg/L was prepared in acetonitrile and working standards (0.001–10 mg/L) were formulated in acetonitrile/0.2% formic acid (60:40 v/v). Table 1 summarizes the physicochemical properties (molecular weight, solubility, pKa) relevant to environmental persistence and mobility.

2.2. Instrumentation

Chromatographic analysis was performed using a Shimadzu HPLC system with an online mobile phase degasser unit (DGU-20A3), and a ternary pump (LC-20AB). All samples and standards were injected onto a rheodyne 7010 injector equipped with a 20 µL sample loop. The compounds were separated using a C18 Kinetex column (150 mm × 4.60 mm × 2.65 μm). Detection was achieved using a PDA detector (SPD-M20A) at 220nm (NSAIDs) and 250 nm (ARVs) as shown in Figure 1 below.

2.3. Sample Collection and Pre-Treatment

Water samples were collected from 12 distinct locations across Durban (KwaZulu-Natal, South Africa) during winter (June–Aug 2024). Sites include influent and effluent streams of Kingsburgh and Amanzimtoti WWTWs, as well as upstream and downstream points along the uMlazi, iSipingo, Mbokodweni, and uMgeni rivers. GPS coordinates of sampling sites are listed Table 2. Samples were collected at 0.3–0.5 m depth in pre-cleaned 1 L clear glass bottles and transported to the laboratory in a cooler box and processed within 24 hours.

Following collection, samples were immediately placed on ice (4°C) in light-protected coolers and transported to the laboratory within 4 hours. Upon receipt, samples were filtered through 0.45 µm polypropylene membranes, acidified to pH 2 with HCl (for NSAID stability), and stored at –20°C until extraction. All samples were processed within 7 days of collection, in accordance with US EPA guidelines for pharmaceutical analysis in water. Stability tests confirmed <10% degradation of target analytes under these conditions over a 14-day period. The study was conducted in four WWTW catchments within the eThekwini municipality, varying in size, treatment capacity, and surrounding land use as shown in Figure 2 below.

2.4. Solid Phase Extraction (SPE) of Water Samples

SPE was performed using Bond Elut Plexa cartridges following [20]. Cartridges were conditioned with 5 mL methanol and 5 mL ultrapure water. Samples (500 mL) were loaded under vacuum, washed with 5 mL ultrapure water, and eluted with 5 mL acetonitrile. Extracts were evaporated under nitrogen and reconstituted in 1 mL acetonitrile/formic acid (60:40 v/v) prior to HPLC analysis.

To assess background contamination and method integrity, procedural blanks were prepared using laboratory tap water and ultra-pure water processed through the entire solid-phase extraction (SPE) workflow. In accordance with the environmental sample protocol, 500 mL of tap water was loaded onto Bond Elut Plexa cartridges (3 mL, 200 mg styrene divinyl benzyl; Phenomenex, USA) following [20], these cartridges were conditioned with 5 mL methanol and 5 mL ultrapure water where extracts were evaporated under nitrogen and reconstituted in 1 mL acetonitrile/formic acid (60:40 v/v) prior to HPLC analysis to ensure data reliability.

The tap-water blank results showed 18 peaks with several concerning observations, early eluting peaks (1.5-2.3 min) - Very large peaks, Peak 1: RT 1.594 min, Area = 3,551,994 mAU·s; Peak 2: RT 1.732 min, Area = 4,387,584 mAU·s; Peak 3: RT 1.834 min, Area = 4,893,175 mAU·s. Target pharmaceutical compounds detected (concerning for a blank): FEN (Fenoprofen): RT 3.807 min, Area = 9,007; DIC (Diclofenac): RT 4.721 min, Area = 15,269; IBU (Ibuprofen): RT 4.897 min, Area = 17,114; EFA (Efavirenz): RT 5.676 min, Area = 8,206; GEM (Gemfibrozil): RT 6.789 min, Area = 80,420.

The tap-water blank fails QC criteria as LOD was lower than LOD for all target compounds. The massive early peaks indicated significant matrix interference. Moreover, quantification was compromised as the sample results cannot be reliably quantified if blanks contain analytes.

While, the deionized water blank showed excellent analytical cleanliness, that is no detection of target pharmaceuticals, minimal, non-interfering background signals and stable baseline across both detection wavelengths. Pharmaceuticals detected in Kingsburgh, Amanzimtoti, Umlazi, and Isipingo samples represent true environmental occurrences, not analytical artifacts. Reported concentrations (e.g., GEM: 5.13–10.86 µg/L; CAR: 4.45 µg/L) are reliable above the method LOQ. The low blank response supports the validity of recovery calculations (70–120%) and precision data (%RSD <10%).

2.5. Chromatographic Analysis Procedure

Isocratic mobile phase of acetonitrile/0.2 formic acid a 60:40 (v/v) ratio was used at 0.8 mL/min. Baseline separation of NSAIDs and ARVs was achieved within 8 min. Retention times were consistent across replicates, confirming method robustness. While isocratic elution was effective for NSAIDs, gradient elution may be required for more polar compounds such as ARVs.

3. Results and Discussion

3.1. Method Validation

Quality Assurance Parameters of the Methods

Calibration curves for all analytes showed excellent linearity (R² > 0.99). Limits of detection (LOD) ranged from 0.63–0.66 µg/L, and limits of quantification (LOQ) from 1.91–2.00 µg/L, confirming sensitivity for trace detection. Recovery studies using spiked ultrapure and tap water samples showed recoveries between 89-124%, with some exceeding 100% due to matrix enhancement effect. Precision assessed through replicate injections (n = 3) yielded (%RSD < 10%) for all compounds showed robust reproducibility. These results as presented in Table 3 confirm that the SPE-HPLC method is reliable for quantifying low concentrations of pharmaceuticals in water.

Calibration curves for each pharmaceutical compound was optimized by plotting analyte concentration against its respective HPLC peak area, ensuring accurate quantification [36]. Linearity was confirmed by the coefficient of determination (R²), between 0.9979 and 0.9991 for all analytes (Table 3). The LOD and LOQ were calculated using the formulas: LOD = 3.3 × (σ/slope) and LOQ = 10 × (σ/slope), where σ is the standard deviation of blank measurements and the slope is derived from the calibration curve. Recovery experiments were performed by spiking deionized and tap water samples 2.5 mg L⁻¹ of each analyte, followed by SPE and HPLC analysis [20]. Recoveries in deionized and tap water were consistently elevated and comparable for all analytes, showing no matrix effect. The use of SPE significantly improved by pre-concentrating the analytes, resulting in lower LOD and LOQ values compared to direct injection.

3.2. Real Sample Results

Pharmaceuticals were detected at all river and wastewater sites, confirming widespread contamination across both direct sources (WWTW discharge) and diffuse inputs (urban runoff, informal settlements) [18,20,28,37]. This consistent presence highlights the persistence of pharmaceuticals in aquatic environments and the limitations of current treatment infrastructure, with implications for both ecosystem health and human exposure as shown in Figure 3.

3.3. Kingsburgh WWTW

At the Kingsburgh WWTW, GEM was identified as the predominant pharmaceutical, exhibiting concentrations of 1.89 μg/L in the influent and 0.49 μg/L downstream. The observed levels correspond with established environmental persistence and are consistent with prior research by [20], which indicated a concentration of 7.25 μg/L during the summer season at the same location. In the current study, there is a decrease in concentration continuous identification of GEM in downstream ecosystems and this highlights the improvement of the standard wastewater treatment methods and its potential to remain in receiving waters. CAR (4.45 µg/L in the influent) and FEN (7.36 µg/L in the influent; 0.92 µg/L downstream) exhibited negligible attenuation, consistent with their recognized stability in aquatic ecosystems. DIC was consistently detected across all treatment phases (1.77–2.21 µg/L), suggesting either persistent intake or inadequate clearance. NAP and IBU levels were reduced downstream, indicating partial removal, most likely due to biodegradation or sorption mechanisms. EFA was identified at low concentrations (0.27–2.42 µg/L), indicating continuous discharge across sampling locations. The CAR concentrations observed correspond with prior research findings [30], corroborating the notion of limited treatment efficacy for certain high-utilization drugs. The findings highlight the necessity of assessing advanced treatment methods or source reduction strategies for enduring pharmaceutical contaminants in urban wastewater systems.

To date, no study has been reported on the CAR EPP in this area, so these results correspond with data reported by Sigonya, Onwubu [20], who reported a maximum of 2.14 μg/L IBU in the same study site for these compounds around the Durban communities suggesting continued persistence of these drugs in the water system [18,20,28] at Kingsburgh WWTW [17]. The environmental persistence and treatment inefficiency, with removal rates in typical WWTPs frequently below 20%, resulting in buildup in effluents and receiving waters [30,38]. These results confirm that compounds with low biodegradability (gemfibrozil, carbamazepine) persist through treatment, whereas more labile NSAIDs show partial removal [2,39,40].

3.4. Ezimbokodweni River/Amanzimtoti WWTW

Amanzimtoti is a suburban area characterized by a lower population density relative to peri-urban regions like Umlazi. Nonetheless, its proximity to industrial and commercial operations, along with periodic surges in population from tourism and coastal recreation, may influence pollutant levels and strain wastewater systems. At the Ezimbokodweni River/Amanzimtoti WWTW, GEM had high concentrations, measured at 0.79 µg/L upstream and 1.30 µg/L downstream, indicating persistence through the treatment process. FEN was detected at 1.19 µg/L upstream and 1.61 µg/L downstream, confirming limited removal efficiency. IBU concentrations increased from 0.80 µg/L upstream to 3.79 µg/L in effluent, showing poor attenuation during wastewater treatment. DIC was also present at quantifiable levels, rising from 1.06 µg/L upstream to 2.01 µg/L downstream.

These concentrations provide baseline evidence for pharmaceutical contamination at this site, particularly for gemfibrozil, which has not previously been reported in the Ezimbokodweni River system. The persistence of lipid regulators and NSAIDs reflects both limited biodegradation and possible contributions from anthropogenic activities, including industrial effluents and recreational inputs [17].

3.5. Isipingo and Umlazi Rivers

Compared to Amanzimtoti and Kingsburgh, the Isipingo WWTW presented a more complex pattern of pollution. That is, the varied land use subjects the river to a combination of industrial effluents, domestic wastewater, and WWTWs discharge In the Isipingo River, efavirenz was detected at consistently high concentrations (1.96–1.99 µg/L), confirming its recalcitrant nature and minimal removal in surface waters. In the Umlazi River, FEN reached 6.99 µg/L upstream, while GEM and DIC were quantified at 0.23 µg/L and 1.01 µg/L, respectively. These results show significant pharmaceutical loading in peri-urban catchments, where informal settlements and industrial discharges contribute to diffuse pollution sources. For example, Thavarayan and Moodley [37] measured carbamazepine at 0.28 μg/L, diclofenac at 2.44 μg/L, and IBU at 1.26 μg/L during the spring season. In contrast, these compounds were not detected in the present study. Seasonal variations impact persistence, with lower degradation in winter due to low sunlight and cooler temperatures. No systematic investigations on pharmaceutical pollution in the Umlazi river have been conducted so far, rendering this dataset a significant contribution.

4. Conclusions

This study quantified pharmaceutical pollutants in surface waters and wastewater treatment works across the eThekwini Metropolitan Municipality, providing the first dataset for township river systems such as Umlazi and Isipingo. Compounds including gemfibrozil, carbamazepine, and fenoprofen, were consistently detected at high concentrations, reflecting their recalcitrant nature and limited removal efficiency in conventional treatment processes. NSAIDs such as Ibuprofen and naproxen exhibited partial attenuation, while antiretroviral drugs remained detectable at environmentally relevant levels. These results highlight the combined influence of diffuse pollution sources and insufficient wastewater treatment infrastructure in shaping pharmaceutical contamination profiles. Limitations of this study include seasonal restrictions on winter sampling and the focus on a limited set of pharmaceutical classes. Future work should expand monitoring to additional compounds, incorporate seasonal variation, and evaluate advanced treatment technologies such as membrane filtration and advanced oxidation processes. This study provides the first dataset on pharmaceutical pollutants in township river systems of eThekwini, highlighting urgent needs for improved wastewater treatment and mixture-based risk assessment frameworks.by enduring pharmaceutical combinations in urban and peri-urban water environments.