Occurrence, distribution, and risk assessment of selected pharmaceuticals in municipal biosolids from the four wastewater treatment plants in Gauteng Province, South Africa

Abstract

The presence of pharmaceutical residues in municipal biosolids is an emerging environmental concern due to their potential accumulation in agricultural soils following land application. While parent compounds like sulfamethoxazole and carbamazepine are frequently studied, data on their transformation products in biosolids remain limited. This study assessed the occurrence, distribution, and environmental risks of sulfamethoxazole, carbamazepine, and two of their selected metabolites in biosolids collected from four wastewater treatment plants (WWTPs) in Gauteng Province, South Africa. A single biosolid sampling campaign was conducted across the four WWTPs, each characterized by varying operational parameters. Sulfamethoxazole and its two metabolites were not detected in any of the samples. In contrast, carbamazepine and its metabolites—carbamazepine-10,11-epoxide and carbamazepine diol—were detected, with maximum concentrations of 150.67, 601.95, and 89.18 ng/g dry weight, respectively. Carbamazepine and carbamazepine-10,11-epoxide occurred in all samples, while carbamazepine diol appeared in 50% of the samples. The concentration patterns varied across WWTPs and were significantly correlated with biosolid pH, organic carbon content, and oxidation-reduction potential. Risk quotient analysis indicated that the predicted soil concentrations of carbamazepine (0.50 ng/g dry weight) pose negligible ecological risk (RQ = 0.003). However, due to a lack of ecotoxicological data, environmental risks associated with carbamazepine-10,11-epoxide and carbamazepine diol could not be fully assessed. Overall, the findings highlight the persistence of carbamazepine and its metabolites in biosolids and underscore the importance of optimizing sludge treatment processes and generating toxicity data for pharmaceutical metabolites to support safe land application practices.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-29215-8.

Introduction

The presence of pharmaceuticals in municipal biosolids has garnered significant environmental attention in recent years due to their potential to accumulate in soil systems. Numerous studies conducted worldwide have explored the prevalence of pharmaceuticals within municipal biosolids, highlighting a diverse array of pharmaceutical classes^1–3. Notably, carbamazepine and sulfamethoxazole are among the most commonly detected pharmaceuticals in these biosolids samples⁴. Carbamazepine functions as an anticonvulsant and is prescribed for epilepsy, neuropathic pain, and bipolar disorder management⁵. Conversely, sulfamethoxazole serves primarily as an antibiotic, targeting bacterial infections of the skin, gastrointestinal tract, respiratory system, and urinary tract in humans⁶. For optimal efficacy, sulfamethoxazole is often administered alongside trimethoprim⁶. Upon administration, both sulfamethoxazole and carbamazepine undergo substantial hepatic metabolism, leading to their excretion predominantly in metabolite forms, which subsequently enter wastewater treatment facilities^7,8. Carbamazepine diol, carbamazepine-10,11-epoxide, unchanged carbamazepine, along with other minor hydroxylated and conjugated metabolites, collectively account for up to 30.0%, 5.0%, 3.0%, and 40.0% of the total excretion, respectively⁷. In a similar vein, N4-acetyl sulfamethoxazole, unchanged sulfamethoxazole, and various glucuronide conjugates and minor metabolites represent approximately 60.0%, 30.0%, and 15.0% of the total human excretion, respectively⁸.

Despite substantial research dedicated to the detection of carbamazepine and sulfamethoxazole in municipal biosolids across various regions, studies examining their metabolites within municipal biosolids remain relatively scarce. This gap is notable considering that the prevalent evidence indicates that sulfamethoxazole and carbamazepine are primarily excreted in their metabolite forms^8,9. Once these metabolites enter the soil, they may (1) exhibit higher bioavailability for plant uptake compared to their parent drugs, potentially leading to food contamination¹⁰, and (2) demonstrate increased mobility in soil, thus raising concerns about groundwater contamination through leaching¹¹. Consequently, thorough assessments of pharmaceuticals and their metabolites in municipal biosolids intended for land application are imperative to mitigate potential risks of food and groundwater contamination.

Hence, the objectives of this study were to: (1) investigate the concurrent occurrence of sulfamethoxazole, carbamazepine, and their selected metabolites in municipal biosolids from the four WWTPs in Gauteng province, South Africa; (2) investigate the relative distribution of sulfamethoxazole, carbamazepine, and their selected metabolites in municipal biosolids across the four WWTPs; (3) evaluate the correlations between the concentration levels of sulfamethoxazole, carbamazepine, and their selected metabolites in municipal biosolids; (4) evaluate the correlations between the concentration levels of sulfamethoxazole, carbamazepine, and their selected metabolites with the physicochemical parameters of municipal biosolids; (5) evaluate the correlations between the concentration levels of sulfamethoxazole, carbamazepine, and their selected metabolites with the key operating parameters of the four WWTPs; and (6) estimate the environmental risk associated with the concentration levels of sulfamethoxazole, carbamazepine, and their selected metabolites in the biosolids upon land application. We hypothesized that: (1) sulfamethoxazole, carbamazepine, and their selected metabolites contamination are present in municipal biosolids, irrespective of the treatment processes applied and the origins of the wastewater; (2) the distribution patterns of sulfamethoxazole, carbamazepine, and their metabolites differ among the various WWTPs; (3) If detected, the concentration levels of sulfamethoxazole and carbamazepine in municipal biosolids will positively correlate with those of their metabolites; (4) If detected, the concentration levels of sulfamethoxazole, carbamazepine, and their metabolites will positively correlate with organic carbon of municipal biosolids; (5) If detected, the concentration levels of sulfamethoxazole, carbamazepine, and their selected metabolites in municipal biosolids will positively correlate with the treatment capacity and serving population; and (6) upon land application, the residual concentration levels of sulfamethoxazole, carbamazepine, and their metabolites in the soil will pose negligible risk to the environment, provided that biosolids are applied according to the current maximum allowable rate of 10 tonnes/ha/yr.

Materials and methods

Study site

The research was carried out at the four wastewater treatment plants (WWTPs) situated within the Ekurhuleni metropolitan municipality in Gauteng province, South Africa. This municipal area is located in the eastern portion of Gauteng and ranks as the second most densely populated region in the country, following Johannesburg¹². These four WWTPs were selected for this study due to their production of municipal biosolids with varying quality levels, particularly regarding microbiological characteristics, stability, and pollutant classifications (refer to Table 1). Consequently, the municipal biosolids produced at these facilities is either stored, applied to designated lands, utilized on non-agricultural land, or distributed to agricultural land¹³. An overview of the key details pertaining to the four WWTPs examined in this study is provided in Table 1.

Table 1.

Treatment capacity, wastewater sources, types of industries within the catchment area, sludge stabilization process, and biosolids drying process of the four WWTPs under investigation¹³.

Treatment plant	Capacity (Ml/d)	Sources of wastewater	Type of industries	Sludge stabilization	Sludge drying	Sludge class
WWTP 1	5.4	Residential (90%)	Plastics and tiles	Anaerobic	Drying beds	A1aa
		Industrial (10%)
WWTP 2	6.1	Residential (85%)	Tannery	Aerobic	Paddies	B1bb
		Industrial (15%)
WWTP3	16.0	Residential (70%)	Batteries	Anaerobic	Drying beds	A1bc
		Industrial (30%)	Dairy items
WWTP 4	15.0	Residential (70%)	Food, beverages, and pulp	Anaerobic	Lagoons	A1aa
		Industrial (30%)	Paper and abattoir

Ml/d means megalitre per day.

^aSludge class with unrestricted use, including agriculture (food crops), landscaping, and composting¹⁴.

^bSludge class with restricted agriculture use (non-food crops), forestry, and land rehabilitation. Not suitable for food crops or public land¹⁴.

^cSludge class with restricted use for forestry, mine dumpers, and energy crops. Not suitable for agriculture or public areas¹⁴.

Sample collection, preparation, and preservation

Samples of municipal biosolids were gathered from the paddies, drying beds, and lagoons at the four WWTPs using shovels and transferred into pre-cleaned glass amber jars. These municipal biosolids samples were collected during the winter months, a period characterized by low to no rainfall, to reduce the potential for dilution─ thereby representing the worst-case scenario for the levels of target compounds. In total, three representative samples from each WWTP were collected and transported to the laboratory in a CAMP MASTER cooler box (Model CB46) for analysis of the target pharmaceuticals. Upon arrival at the laboratory, the municipal biosolids samples underwent freeze-drying, followed by grinding and sieving through a 2 mm sieve, after which they were stored at -20 degrees Celsius until extraction.

Chemicals, reagents, and solvents

The analytical standards and certified reference materials listed below were procured from Sigma-Aldrich, South Africa: Vetranal^® sulfamethoxazole powder, cerilliant^® carbamazepine solution, cerilliant^® carbamazepine-10,11-epoxide solution, and N4-acetyl sulfamethoxazole powder. Additional chemicals including sodium hydroxide, ethylenediaminetetraacetic acid disodium salt dehydrate, hydrochloric acid, and phosphate buffer were also acquired from Sigma-Aldrich, South Africa. Sulfamethoxazole-N1-glucuronide powder and carbamazepine diol solution were sourced from Industrial Analytical, South Africa. For chromatographic analysis, ultra-purity solvents (Romil-UpS™,Microsep, South Africa) and formic acid (99+% purity) (Thermo Scientific, South Africa) were used. Lastly, Waters Oasis^® HLB 6 cc (6 mL, 200 mg, waters) cartridges were acquired from Microsep, South Africa. LC MS grade water was produced using a water purification system (PURELAB^® Chorus 1, Labotec). An overview of the physicochemical parameters pertaining to the targeted pharmaceuticals is provided in Table 2.

Table 2.

Physicochemical parameters of targeted pharmaceuticals^15–17.

Compound	Molecular weight (g/mol)	Water solubility (g/L)	Log Kow	pKa
CBZ	236.27	0.02	2.77	13.9
CBZ-EP	252.27	0.08	1.97	15.9
CBZ-DiOH	270.28	0.10	0.80	12.8
SMX	253.28	0.61	0.89	5.6
Ac-SMX	295.31	0.30	1.20	5.9
SMX-N1-Glu	429.40	2.55	0.22	3.5

CBZ = carbamazepine, CBZ-EP = carbamazepine-10,11-epoxide, CBZ-DiOH = carbamazepine diol, SMX = sulfamethoxazole, Ac-SMX = N4-acetyl sulfamethoxazole, and SMX-N1-Glu = sulfamethoxazole-N1-glucuronide.

Sample extraction and clean up procedure

Municipal biosolids samples were processed utilizing the EPA method 1694, incorporating adaptations based on the methodology established by previous studies¹⁸. The extraction process consisted of three successive extractions. For the initial extraction, 20 mL of acetonitrile was added to the sample, which was then sonicated for 30 min. The second extraction involved the incorporation of 15 mL of phosphate buffer solution (pH 2) to the previously extracted sample, followed by vortexing to resuspend the solids. The third extraction required adding 15 mL of acetonitrile to the subsequent sample and again sonicated for 30 min. After each extraction, the mixture underwent centrifugation at 3000 rpm for 5 min, and the supernatant was carefully transferred into a clean test tube. The combined extracts from the three sequential extractions were consolidated in a Buchner flask, and the solvents were evaporated using a rotary evaporator until a final volume of approximately 20 mL was achieved. Subsequently, these extracts were moved to a 100 mL volumetric flask, where 250 mg of Na₂EDTA·2H₂O was introduced. Upon the addition of Na₂EDTA·2H₂O, the flask was filled to the 100 mL mark with ultrapure water. The sample was then loaded into Oasis HLB cartridges, which had been sequentially conditioned with 6 mL of methanol, 6 mL of ultrapure water, and 6 mL of acidified ultrapure water (pH = 2). Following the loading of the sample, any interfering substances were removed with a 20 mL rinse of ultrapure water. The cartridges were subsequently dried under vacuum at 5 bar pressure for 5 min. Target pharmaceuticals were eluted with three 4 mL aliquots of methanol, which were later evaporated using a GeneVac (EZ-2 Plus series) at a heating temperature of 45 °C and a pressure of ≤ 200 mbar. Finally, the pharmaceuticals were reconstituted in a 1 mL solvent mixture of methanol and water (50:50, v/v). The reconstituted solutions were vortexed from the Polytop glass vials (20 mL, with a push in stopper cap) for 1 min, transferred into an injection amber glass vial, and the pharmaceuticals were promptly quantified using the Ultra Performance Liquid Chromatography Quadrupole Time-of-Flight (UPLC-QTOF) instrument.

Instrument analysis

Instrument and sample injection

Target pharmaceuticals and their corresponding metabolites were quantified using a Waters ^® Synapt G2 high-definition mass spectrometry (HDMS) system (Waters Inc., Milford, Massachusetts, USA). The apparatus was integrated with a Waters Acquity Ultra Performance Liquid Chromatography (UPLC) system linked to a quadrupole time-of-flight (QTOF) instrument. Quantitative data-independent acquisition (DIA) was done using two simultaneous acquisition functions with low and high collision energy (MS^E approach) with a QTOF instrument. Fragmentation was performed using high energy collision induced dissociation (CID). The fragmentation energy was set at 4 V for low energy acquisition and the high energy ramp transfer collision energy set from 20 to 40 V. The source conditions were as follows: the capillary voltage for ESI was 2.6 kV and 2.0 kV for positive and negative mode ionisation, respectively. The source temperature was set at 120 °C, the sampling cone voltage at 30 V and 25 V for positive and negative mode ionisation, respectively. The extraction cone voltage at 4.0 V and cone gas (nitrogen) flow at 15.0 L/Hr. The desolvation temperature was set at 350 °C with a gas (nitrogen) flow of 600.0 L/Hr. Mass spectral scans were collected every 0.3 s. The raw data was collected in the form of a continuous profile. Mass to charge ratios (m/z) between 50 and 1 200 Da were recorded.

The instrument incorporated a Kinetex^® 1.7 μm EVO C18 100 Å (2.1 mm ID × 100 mm length) column to facilitate the chromatographic separation of the target pharmaceuticals. The temperature of the column was maintained consistently at 50 °C, and the flow rate was established at 0.3 mL/min throughout the duration of the analysis, resulting in a total runtime of 20 min. The injection volume for the samples, blanks, solvents, and standards was exactly 5 µL and was performed in duplicate to measure the precision of the instrument. Separation was executed using a reverse-phase gradient elution protocol commencing from 97% H₂O (Romil-UpS™, Microsep, South Africa) (augmented with 0.1% formic acid (99+% purity, Thermo Scientific, South Africa)) to 100% methanol (Romil-UpS™, Microsep, South Africa) (also with 0.1% formic acid). The volume of injection was 5 µl. Data acquisition and subsequent processing were conducted employing MassLynx™ (version 4.1) software (Waters Inc., Milford, Massachusetts, USA). Additional parameters of the instrument and details regarding chromatographic separation are encapsulated in Supplementary Table S 1.

Calibration, quantification, and quality controls

Six standard solutions ranging from 1 to 1000 ng mL^{− 1} were used to obtain calibration curves for the targeted pharmaceuticals (Supplementary Table S2). The correlation coefficients (R²) of the calibration curves were greater than 0.99, indicating the validity of the calibration curves for the quantification the target pharmaceuticals. Procedural blanks were spiked with known concentrations of the pharmaceuticals and extracted to determine the percentage recoveries and relative standard deviations. The percentage recoveries for the target pharmaceuticals ranged from 62 to 105% (Supplementary Table 3), indicating that this method is reliable and reproducible. The limits of quantification (LOQs) and limits of detection (LODs) were determined from the calibration curves as concentrations that give a signal-to-noise (S/N) ratio of 10 and 3, respectively¹⁹. LODs and LOQs for the target pharmaceuticals are given in the supplementary Table S3.

Statistical data analysis

Descriptive statistical analysis, including frequency distribution, mean, range, and standard deviation, were performed to determine the occurrence and concentration levels of the target pharmaceuticals in municipal biosolid samples. To evaluate potential associations, Pearson correlation analysis was conducted using Origin-Pro software (Student version: 10.2.0.188). This analysis explored the relationship between (i) the concentration of the target pharmaceuticals; (ii) the analyte concentrations and the physicochemical properties of the biosolids; and (iii) the analyte concentrations and key operating parameters of the four WWTPs involved in the study.

Environmental risk assessment

Predicted concentration of the target pharmaceuticals in the soil

The predicted concentrations of the target pharmaceuticals within the soil matrix were calculated in accordance to the formula proposed in the European Technical Guidance Document on Risk Assessment EUR 20,418 EN/2 (Eq. 1)²⁰:

1

Where, PECsoil denotes the predicted environmental concentration of the target pharmaceuticals within the soil subsequent to the application of biosolids (ng/kg), Cbiosolids represents the quantified maximum concentration of the target compound present in the municipal biosolids (ng/kg), APPbiosolids signifies the rate of biosolids application (kg/ha), BDsoil refers to the bulk density of the soil (kg/m³), and DEPTHsoil indicates the depth of application (m). The concentrations of the target pharmaceuticals derived from the municipal biosolids originating from WWTP 4 were selected for the present investigation, as the current municipal biosolids is employed in the cultivation of human food crops (Table 1). The biosolids application rate will be established at the prevailing upper limit rate of 10 tonnes/ha/yr as prescribed by the South African sludge guideline volume 2¹⁴. Sandy clay loam soils were used in the present study due to their representation as the predominant soil textural class in the Gauteng province, where sludge from WWTP 4 is utilised²¹. Consequently, the bulk density of the soil was designated at 1500 kg/m³, which is the characteristic of the sandy clay loam²². Municipal biosolids are frequently applied to the surface soils and incorporated into the top 0.10 to 0.2 m layer, a parameter that was adopted for the present investigation.

Predicted ecological risk of the target pharmaceuticals in the soil

The ecological risk assessment serves as a methodological framework to evaluate the potential hazards posed to the environment and its constituent organisms. This assessment employs the risk quotient ratio of the projected concentration of the target pharmaceuticals in the soil to the predicted no-effect concentration in the soil²³. The PNECsoil can be determined by multiplying the predicted no-effect concentration in water with the soil partitioning coefficient²³. The predicted no-effect concentration in water was derived from the Canadian Water Quality Guidelines designed for the Protection of Aquatic Life²⁴. The soil partitioning coefficient values pertinent to the target pharmaceuticals in municipal biosolids-amended soils were sourced from existing literature²⁵.

Results and discussion

Occurrence and concentration of the target pharmaceuticals in municipal biosolids

Occurrence of the target pharmaceuticals in municipal biosolids

In the present study, three out of the six target pharmaceuticals were identified within the analysed municipal biosolids samples (Fig. 1). Sulfamethoxazole and its two metabolites were not identified in any of the biosolids samples (Fig. 1), suggesting their absence and/or expediated degradation within WWTPs. Among the pharmaceuticals that were detected, carbamazepine and its metabolite carbamazepine-10,11-epoxide were present in all analysed biosolids samples (Fig. 1), signifying their considerable prevalence in municipal biosolids. Conversely, carbamazepine diol was identified in only half of the biosolids samples (Fig. 1), reflecting its relatively low prevalence in municipal biosolids. The greater prevalence of carbamazepine and carbamazepine-10,11-epoxide compared to carbamazepine diol in municipal biosolids can be elucidated by their elevated octanol-water partitioning coefficient (Log K_ow) (Table 2), which may have facilitated their adsorption onto organic matter present in the raw sludge, ultimately resulting in their presence in municipal biosolids²⁶.

Fig. 1.

Detection frequency of target pharmaceuticals in municipal biosolids samples. CBZ = carbamazepine, CBZ-EP = carbamazepine-10,11-epoxide, CBZ-DiOH = carbamazepine diol, SMX = sulfamethoxazole, Ac-SMX = N4-acetyl sulfamethoxazole, and SMX-N1-Glu = sulfamethoxazole-N1-glucuronide.

The detection frequencies of carbamazepine documented in this study exhibit a remarkable similarity to those observed in municipal biosolids samples from Spain (100%)²⁷, the Czech Republic (100%)⁴, Greece (100%)²⁸, and Turkey (92%)¹. It is pertinent to note that the comparative analysis of the detection frequencies of the two metabolites of carbamazepine with respect to other prior studies proved to be challenging due to the paucity of information in the literature regarding the occurrence of these metabolites in municipal biosolids. Nonetheless, the detection frequencies of carbamazepine-10,11-epoxide reported herein surpass those documented in Spain (0%)²⁹. The absence of carbamazepine-10,11-epoxide in the municipal sludge of Spain may be elucidated by the potential for further transformation or the complete degradation of carbamazepine-10,11-epoxide into alternative metabolites²⁶. Notably, the research conducted by²⁹acknowledged the presence of various metabolites of carbamazepine-10,11-epoxide, despite its absence in municipal sludge, thereby indicating a likely progression of degradation into other metabolites.

Concentrations of the target pharmaceuticals in municipal biosolids

The quantification of carbamazepine, carbamazepine-10,11-epoxide, and carbamazepine diol within municipal biosolids samples exhibited a range from non-detectable levels to 601.95 ng/g (Table 3), thereby reflecting the variability in the quality of the analysed municipal biosolids samples. Notably, carbamazepine-10,11-epoxide demonstrated the highest concentration, which ranged from 33.13 (in samples from WWTP 3) to 601.95 ng/g (in samples from WWTP 2) (Table 3). Conversely, carbamazepine diol, exhibited the lowest concentration, oscillating between non-detectable (in samples from WWTP 1 and 4) and 89 ng/g (in samples from WWTP 2) (Table 3). The mean concentrations of carbamazepine documented in this study are markedly lower than those previously reported in municipal biosolids from Canada (183 ng/g)³⁰, United States (163 ng/g)³¹, Sweden (130 ng/g)³², United Kingdom (121 ng/g)³³, and China (11060 ng/g)³⁴. The comparatively lower concentration of carbamazepine in municipal biosolids reported in this study, when juxtaposed with those from the aforementioned developed nations, may be elucidated by the anticipated low levels of consumption attributed to restricted access to mental health care, as well as the diminished rates of diagnosis and treatment for mental health disorders³⁵. According to³⁵, over 90% of individuals in South Africa diagnosed with mental health disorders do not adhere to their treatment protocols, thereby resulting in frequent hospital readmissions. In contrast, the average concentrations of carbamazepine-10,11-epoxide reported in this analysis are higher than those documented in various stabilized municipal biosolids from Spain (0 ng/g)²⁹.

Table 3.

Concentration levels of target pharmaceuticals in municipal biosolids samples.

Compounds	Range (ng/g dry weight)	Mean ± SD (ng/g dry weight)	Median (ng/g dry weight)
CBZ	7.25–150.67	44.34 ± 70.92	9.72
CBZ-EP	29.41–601.95	182.92 ± 279.87	50.17
CBZ-DiOH	n.d – 89.18	28.56 ± 42.10	12.54
SMX	n.d	n.d	n.d
Ac-SMX	n.d	n.d	n.d
SMX-N1-Glu	n.d	n.d	n.d

SD = standard deviation, n.d = not detected, CBZ = carbamazepine, CBZ-EP = carbamazepine-10,11-epoxide, CBZ-DiOH = carbamazepine diol, SMX = sulfamethoxazole, Ac-SMX = N4-acetyl sulfamethoxazole, and SMX-N1-Glu = sulfamethoxazole-N1-glucuronide.

Relative distribution of the target pharmaceuticals in municipal biosolids across the four wastewater treatment plants

The stacked bar chart reveals substantial variability in the relative distribution of the three target pharmaceuticals, carbamazepine, carbamazepine-10,11-epoxide, and carbamazepine diol, across the four WWTPs (Fig. 2). This variability likely reflects differences in influent pharmaceutical loads, microbial communities, redox conditions, retention times and enzymatic activities associated with each treatment process. WWTP 1 exhibited a clear dominance of the parent carbamazepine (69.15%), with carbamazepine-10,11-epoxide comprising the remaining 30.85% and no detectable carbamazepine diol (Fig. 2). This suggests limited transformation of carbamazepine into its metabolites. One possible explanation is the low pH of the municipal biosolids from this plant (pH = 5.6; supplementary Table S 4), which may have suppressed oxidative enzyme activities required for the biotransformation of carbamazepine. He et al.³⁶, demonstrated that acidic biosolids (pH < 6.5) inhibit enzyme systems, particularly those involved in oxidative transformation such as cytochrome P450 monooxygenases, which are crucial for converting carbamazepine to carbamazepine-10,11-epoxide. In contrast, WWTP 2 showed an overwhelming dominance of carbamazepine-10,11-epoxide (86.18%) a moderate level of carbamazepine diol (12.77%), and minimal residual carbamazepine (1.06%) (Fig. 2). This distribution indicates extensive transformation of the parent compound, likely facilitated by more favourable conditions for microbial metabolism. Notably, biosolids from this plant had a pH > 6.5 (Supplementary Table S 4), which may have supported more active enzyme-mediated oxidation of carbamazepine, in line with the findings of²⁹. The presence of carbamazepine diol further suggests either continued microbial degradation of carbamazepine-10,11-epoxide or enhanced co-metabolic processes during sludge treatment³⁷. WWTP 3 showed similar pattern, with carbamazepine-10,11-epoxide accounting for 50.16%, carbamazepine diol for 38.31%, and only 11.08% of the total being parent carbamazepine (Fig. 2). The high proportion of carbamazepine diol in this plant implies not only effective transformation of carbamazepine but also further degradation of its epoxide form. This may point to longer sludge retention time³⁸, possibly enhancing microbial transformation of carbamazepine and its epoxide metabolites. WWTP 4 also showed a high dominance of carbamazepine-10,11-spoxide (71%) with notable fraction of parent carbamazepine (29%) and no detectable carbamazepine diol (Fig. 2). This distribution suggests an intermediate levels of transformation, greater than in WWTP 1 but less than in WWTPs 2 and 3. The absence of carbamazepine diol could be attributed to either a shorter sludge retention time or limited microbial populations capable of further metabolising carabamazepine-10,11-epoxide. Again, the pH at WWTP 4 (> 6.5) (Supplementary Table S 4) was within a range favourable to enzyme activity³⁶, supporting carbamazepine oxidation but possibly not sufficient for complete degradation to carbamazepine diol.

Fig. 2.

Percentage relative distribution of the target pharmaceuticals in municipal biosolids among different WWTPs. Concentrations were normalized to % of total load of the target pharmaceuticals detected per WWTP. CBZ = carbamazepine, CBZ-EP = carbamazepine-10,11-epoxide, and CBZ-DiOH = carbamazepine diol.

Overall, the results demonstrate that the extent of carbamazepine transformation in biosolids varies considerably across WWTPs, influenced by treatment conditions, particularly pH, and possibly by other operational parameters such as sludge age, aeration efficiency, and microbial community structure^38,39. These findings highlight the importance of optimizing treatment conditions to minimize the persistence of parent pharmaceuticals in biosolids intended for reuse.

Correlation analysis between the concentration levels of the target pharmaceuticals and their metabolites in municipal biosolids

The Pearson correlation analysis revealed significant positive relationships between carbamazepine and its two transformation products—carbamazepine-10,11-epoxide and carbamazepine diol—in municipal biosolid samples (Table 4). These correlations suggest that the three pharmaceuticals are closely linked in terms of their origin, environmental persistence, and metabolic pathways within wastewater treatment systems. Specifically, a very strong correlation was observed between carbamazepine and carbamazepine-10,11-epoxide (r = 0.98, p < 0.01) (Table 4), supporting the role of carbamzepine-10,11-epoxide as the primary oxidative metabolite of carbamazepine, typically formed via enzymatic epoxidation during biological treatment processes⁴⁰.

Table 4.

Pearson correlations coefficients (r) between target pharmaceuticals and target pharmaceuticals concentrations and physiochemical parameters of municipal biosolids.

Variables	pH	OPR	%OC	CBZ	CBZ-EP	CBZ-DiOH
pH	1.0
ORP	− 0.99***	1.0
% OC	n.s	n.s	1.0
CBZ	n.s	n.s	0.94*	1.0
CBZ-EP	n.s	n.s	0.90*	0.98**	1.0
CBZ-DiOH	0.88*	− 0.89*	0.89*	0.89*	0.95*	1.0

*Significant correlation at the level p < 0.05 (2-tailed).

**Significant correlation at the level p < 0.01 (2-tailed).

***Significant correlation at the level p < 0.001 (2-tailed).

OPR = Oxidation Redox Potential; OC = organic carbon; CBZ = carbamazepine; CBZ-EP = carbamazepine-10,11-epoxide; CBZ-DiOH = carbamazepine diol; n.s = not significant at p < 0.05.

Furthermore, significant correlations between carbamazepine and carbamazepine diol (r = 0.89, p < 0.05) (Table 4), as well as between carbamazepine-10,11-epoxide and carbamazepine diol (r = 0.95, p < 0.05) (Table 4), imply a sequential transformation pathway, where carbamazepine is first converted to carbamazepine-10,11-epoxide and then further to carbamazepine diol. These findings reinforce the metabolic linkage among the three compounds and align with earlier studies documenting their persistence and biotransformation under wastewater treatment conditions^40,41.

The presence and strong correlations of these pharmaceuticals can also be attributed to shared sources. All three pharmaceuticals are commonly introduced into wastewater systems via human excreta²⁶ and are known to resist degradation during conventional sludge stabilization processes, resulting in their accumulation in biosolids⁴¹.

Correlation analysis between the concentration of target pharmaceuticals and the physicochemical parameters of municipal biosolids

A significant positive correlation was observed between organic carbon and the concentration levels of the three identified target pharmaceuticals —carbamazepine, carbamazepine-10,11-epoxide, and carbamazepine diol—in municipal biosolids (p < 0.05) (Table 4), suggesting that increasing organic matter content enhances the retention of these pharmaceuticals. This finding is consistent with previous reports that pharmaceuticals, particularly those with hydrophobic characteristics like carbamazepine and its metabolites, tend to adsorb strongly onto organic material from sludge^23,42. The high affinity of carbamazepine and its transformation products for organic matter results in reduced mobility and increased accumulation in biosolids, due to processes such as hydrophobic partitioning and non-specific binding to humic substances. Furthermore, the sorption capacity of organic matter not only retains these pharmaceuticals but may also shield them from further degradation, thereby promoting their persistence during sludge treatment and storage⁴³. The strong correlations observed in this study reinforce the importance of considering organic matter quality and quantity when assessing the environmental fate of pharmaceuticals in sludge-amended soils or land-applied biosolids.

A significant positive correlation was observed between pH and the concentration of carbamazepine diol in municipal biosolids (p < 0.05) (Table 4). This relationship suggests that higher (more alkaline) pH conditions favour the formation or stability of carbamazepine diol. This is supported by the known pH sensitivity of the oxidative enzymes involved in pharmaceutical transformation; alkaline environments (pH > 6.5) typically enhance microbial enzymatic activity that facilitates oxidative and hydrolytic transformation of carbamazepine into its diol metabolite³⁶. In contrast, acidic conditions can suppress enzyme activity and limit transformation efficiency, leading to the persistence of the parent compound.

A significant negative correlation was observed between oxidation-reduction potential and carbamazepine diol (r = − 0.89; p < 0.05) (Table 4), indicating that anaerobic (reducing) conditions promote the formation or retention of carbamazepine diol in biosolids. This finding aligns with the known microbial and hydrolytic degradation pathways of carbamazepine, in which carbamazepine diol forms under low oxidation-reduction potential conditions where oxidative enzymes are less active^36,44. The persistence of carbamazepine diol under reducing conditions may be attributed to limited further degradation in the absence of oxygen or oxidative cofactors. In contrast, oxidation-reduction potential was not significantly correlated with either carbamazepine or carbamazepine-10,11-epoxide (Table 4), although carbamazepine-10,11-epoxide is primarily formed via oxidative epoxidation processes⁴⁰. The lack of significant correlation suggests that while oxidation processes influence carbamazepine epoxidation, other environmental or operational variables (e.g., pH, organic matter content, microbial diversity) may moderate this effect in biosolids matrices²³.

The interplay between pH and oxidation-reduction potential is also critical. A significant positive correlation was observed between pH and carbamazepine diol (r = 0.88; p < 0.05), indicating that alkaline conditions support carbamazepine diol formation, possibly by enhancing microbial reductive reactions and hydrolysis pathways³⁶. Given that increasing pH can reduce oxidation-reduction potential in organic-rich environments, the co-occurrence of high pH and low oxidation-reduction potential likely creates favourable conditions for carbamazepine diol accumulation in WWTP biosolids. This highlights the synergistic influence of physicochemical parameters on pharmaceutical transformation and persistence.

Taken together, the observed correlations suggest that both high organic carbon content and alkaline pH contribute synergistically to the retention and transformation of carbamazepine and its metabolites in municipal biosolids. These findings underscore the importance of understanding sludge physicochemical characteristics when evaluating the environmental risks associated with land application of biosolids containing trace organic contaminants.

Correlation analysis between the concentration of target pharmaceuticals in municipal biosolids and the key operating parameters of the four wastewater treatment plants

There were no significant relationships between the absolute treatment capacity or population served and the concentration of any of the three target pharmaceuticals (p > 0.05) (Table 5), supporting previous findings that the total size or throughput of a WWTP does not reliably predict pharmaceutical concentrations in biosolids⁴⁵. However, the ratio of treatment capacity to population served—an indicator of per capita hydraulic loading (m³/person/day)—was significantly correlated with carbamazepine (r = 0.96, p < 0.05) and carbamazepine-10,11-epoxide (r = 0.95, p < 0.05) (Table 5). This implies that higher per capita loading is associated with greater retention of these compounds in biosolids. This relationship may be explained by the dilution of biological activity or shorter hydraulic retention times in systems with high daily input volumes per capita, potentially limiting microbial degradation and enhancing the persistence of carbamazepine and carbamazepine-10,11-epoxide⁴⁶. In contrast, carbamazepine diol showed no significant correlation with the per capita loading ratio (p > 0.05) (Table 5), further supporting the notion that its formation is more dependent on microbial redox dynamics than on operational loading rates.

Table 5.

Pearson correlations coefficients (r) between the concentration of target pharmaceuticals in municipal biosolids and the key operating parameters of the four WWTPs.

Variables	Treatment capacity	Population	Ratio	CBZ	CBZ-EP	CBZ-DiOH
Treatment capacity	1.0
Population	n.s	1.0
Ratio	n.s	n.s	1.0
CBZ	n.s	n.s	0.96*	1.0
CBZ-EP	n.s	n.s	0.95*	0.98**	1.0
CBZ-DiOH	n.s	n.s	n.s	0.89*	0.95*	1.0

*Significant correlation at the level P < 0.05 (2-tailed).

**Significant correlation at the level P < 0.01 (2-tailed).

***Significant correlation at the level P < 0.001 (2-tailed).

CBZ = carbamazepine; CBZ-EP = carbamazepine-10,11-epoxide; CBZ-DiOH = carbamazepine diol; n.s = not significant at p < 0.05. Ratio = Treatment capacity/Population. Treatment capacity is in units m³/day.

Together, these findings reinforce the idea that sludge transformation of pharmaceuticals is more strongly governed by environmental conditions (e.g., pH, oxidation-reduction potential, organic content) than by system size or throughput. However, per capita loading still emerges as an important operational proxy influencing the persistence of certain pharmaceuticals, especially parent compounds like carbamazepine and their primary oxidative metabolites.

Environmental risk assessments

The predicted environmental concentrations of carbamazepine, carbamazepine-10,11-epoxide, and carbamazepine diol in soil following biosolid application were estimated at 0.50, 2.0, and 0.30 ng/g, respectively. These values reflect the residual presence of parent and metabolite compounds in municipal biosolids after wastewater treatment, highlighting their persistence and potential entry into terrestrial ecosystems via land application. For carbamazepine, the predicted no-effect concentration in soil was determined to be 198 ng/g, based on existing ecotoxicological data⁴⁷. The resulting risk quotient was 0.003, indicating a very low ecological risk. According to international guidelines, a risk quotient value below 0.1 is generally considered to reflect negligible ecological concern^20,39.

Due to the lack of available toxicological and ecotoxicological data for carbamazepine-10,11-epoxide and carbamazepine diol in soil ecosystems, the predicted no-effect concentration values could not be calculated for these metabolites. This data gap is consistent with other reports emphasizing the limited knowledge on the environmental behaviour and toxicity thresholds of pharmaceutical metabolites compared to their parent compounds^42,44. However, it is worth noting that carbamazepine-10,11-epoxide is pharmacologically active and possess slightly similar higher persistence in the soil compared to carbamazepine⁴⁰, which warrants further investigation into its long-term ecotoxicological impact. Similarly, carbamazepine diol is a secondary metabolite formed under anaerobic and alkaline conditions and may persist in soil for extended periods³⁶. Future studies should prioritize experimental toxicity testing and modelling approaches (e.g., QSAR, read-across) to establish provisional thresholds. In the interim, assessments may consider conservative comparisons with the parent compound, though such extrapolations carry uncertainty. This includes cases where metabolites differ significantly from the parent in terms of toxicity, persistence, mobility, and bioavailability.

Despite the conservative predicted environmental concentrations estimates, the actual risk in agricultural soils is expected to be even lower in practice. This is because biosolids are often applied based on the nutrient requirements of specific crops rather than conventional agronomic rates, thereby reducing the total loading of pharmaceutical residues per unit area⁴⁸. Furthermore, natural attenuation processes such as microbial degradation, photolysis, and adsorption to organic matter further reduce the bioavailability of these compounds over time²³. These findings are consistent with several studies reporting minimal environmental risk from carbamazepine residues in soils treated with municipal sludge^39,47,48. Nonetheless, the absence of the predicted no-effect concentration data for key transformation products such as carbamazepine-10,11-epoxide and carbamazepine diol highlights the need for further toxicological studies to accurately characterize their environmental risks.

Conclusions

This study provides a comprehensive assessment of the occurrence, concentration, transformation, and environmental risk of carbamazepine, sulfamethoxazole, and their metabolites in municipal biosolids from four wastewater treatment plants. Carbamazepine and carbamazepine-10,11-epoxide were consistently detected in all samples, while carbamazepine diol appeared less frequently, reflecting differential transformation dynamics. However, these findings should be interpreted with caution taking into account that a single sampling campaign reflecting one season was conducted, hence there might be seasonal variations in the occurrence of the target pharmaceuticals in municipal biosolids. Future studies should focus on the occurrence of these target pharmaceuticals across varying seasons to evaluate the influence of seasonal variations. The variability in analyte concentrations and relative distributions across WWTPs was strongly influenced by physicochemical conditions, particularly pH, organic matter content, and oxidation-reduction potential, rather than treatment capacity or population served. Correlation analyses confirmed strong metabolic linkages among the three compounds, with carbamazepine undergoing sequential biotransformation to its epoxide and diol forms. High organic carbon and alkaline pH were positively associated with analyte retention and transformation. It would be important to sample more biosolids with a wider range of variability of these parameters to confirm the results of this study. Despite their persistence in biosolids, predicted environmental concentrations of carbamazepine in the soil were significantly below ecotoxicological thresholds, indicating negligible risk. However, the lack of toxicological data for carbamazepine metabolites limits a complete environmental risk assessment. Future studies should prioritize experimental toxicity testing and modelling approaches to establish provisional thresholds. In the interim, assessments may consider conservative comparisons with the parent compound, though such extrapolations carry uncertainty. These findings emphasize the importance of optimizing sludge treatment conditions to reduce pharmaceutical residues in biosolids and underscore the need for further research into the environmental fate and toxicity of transformation products. Overall, land application of biosolids poses minimal ecological risk from carbamazepine, provided regulatory guidelines and application rates are properly managed.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Matome Peter Ngoetjana did the investigation, field and laboratory work, and wrote the original draft. Eyob Habte Tesfamariam supervised the research, organised funding, and made review and editing. Sally Brown co-supervised the research and made review and editing. Madelien Wooding helped with laboratory analysis, data curation, and made a review and editing.

Funding

The first author acknowledges the National Research Foundation (NRF), South Africa for their scholarship. This study was fully funded by Water Research Commission (Project number: C2022/2023 − 00880).

Data availability

All data generated or analysed during this study are included in this published article [and its supplementary information files].

Declarations

Competing interests

The authors declare no competing interests.

Consent to participate

All authors have given consent to their contribution.