Development, Optimization, and Application of Molecularly Imprinted Polymers‐Solid Phase Extraction Procedure for the Analysis of Selected Pharmaceuticals in Vegetable Samples

ABSTRACT

The growing presence of pharmaceutical residues in the environment has aroused worries about their possible buildup and accumulation in the food chain, particularly in edible plants. Conventional analytical methods frequently fail to selectively isolate and quantify trace amounts of these chemicals in complex plant matrices. This study introduces an optimized methodology for the analysis of pharmaceuticals in vegetables. The enhanced molecularly imprinted solid phase extraction (MISPE) approach investigates the extraction of selected pharmaceuticals, fenoprofen, naproxen, diclofenac, ibuprofen, and gemfibrozil in vegetables. The extracted compounds were identified both qualitatively and quantitatively using a high‐performance liquid phase chromatographic (HPLC) system coupled with a photodiode array detector. This method was effectively implemented on vegetable samples collected from Durban, South Africa (SA), including lettuce, carrot, cucumber, and green pepper. The recovery rates varied from 45% to 103%, with relative standard deviation (%RSD) ranging from 0.9% to 13%. Fenoprofen was the most prevalent compound, exhibiting high concentrations in pepper and cucumber, with maximum concentrations of 6.44 and 4.99 mg kg⁻¹, respectively. The health index (HI) values for the vegetables ranged from 0.27 to 1.25. The pepper sample (1.25) surpassed the HI threshold value of 1, reflecting the health indicator risk associated with the consumption of peppers available within the area. The health risk assessment (HRI) values spanned from 0.00012 to 0.83 for both adults and children, suggesting no health risk associated with the consumption of these vegetables.

1. Introduction

A myriad of chemicals permeates the air, water, soil, and food, thereby significantly impacting the quality of the ecosystems inhabited by humans (Mancuso 2024). The persistence of these chemicals renders them particularly detrimental, as they can remain in the environments where flora such as plants and vegetables develop for prolonged durations (Mancuso 2024). Pharmaceuticals are frequently synthesized to produce and alter chemical compounds that can be used to treat diseases or medical conditions (Mancuso 2024). Nevertheless, a considerable portion of these pharmaceuticals often infiltrates public wastewater systems. Certain pharmaceuticals that enter wastewater treatment plants (WWTPs) are reportedly not entirely eliminated during treatment processes, subsequently integrating into sewage sludge products and effluent streams (da Silva and de Souza 2019; Hassen et al. 2019; Andrea et al. 2018). These extensively utilized pharmaceuticals primarily originate from land applications, inclusive of animal manure, wastewater biosolids, and crops irrigated with treated water from streams that are impacted by WWTPs. Presently, an extensive variety of drugs are employed by humans and domestic animals globally each year. Naproxen, fenoprofen, ibuprofen, and diclofenac constitute classes of nonsteroidal anti‐inflammatory drugs (NSAIDs) frequently utilized by humans for their analgesic properties. Conversely, gemfibrozil serves as a lipid regulator used to address aberrant blood lipid levels. These pharmaceuticals (Table 1) constitute the most employed class of pharmaceuticals, administered in substantial doses, seldom eliminated during wastewater treatment processes, and consequently manifest in elevated environmental concentrations (Fernandes et al. 2020). These compounds have been identified in WWTP effluents, sewage sludge, and water sources (Zahmatkesh et al. 2022; Camacho‐Muñoz et al. 2012; Sanusi et al. 2023; Rikard et al. 2020; Brunetti et al. 2023). Moreover, the persistence of pharmaceutical application on soils and their accumulation over time, along with the improper disposal of surplus medications in landfills or dumping sites, may culminate in escalated concentrations of these pharmaceuticals, inadvertently posing potential threats to living organisms (Khezami et al. 2024). In the context of sample pre‐treatment, solid‐phase extraction (SPE) has attracted considerable attention in recent years owing to its simplicity and instrumental role in separation chemistry (Wakim et al. 2024; Metwally et al. 2021). A diverse array of SPE sorbents exists, with one of the most promising approaches for creating sorbent material with pronounced selectivity being molecularly imprinted polymerization (MIPs), which are specialized polymeric materials, crafted with cavities that embody the functional groups, size, and shape of the template molecule, thereby possessing distinct molecular recognition capabilities (Janczura et al. 2021; Arreguin‐Campos et al. 2021). Molecularly imprinted solid phase extraction (MISPE) proves exceedingly valuable in extraction techniques and other contexts, encompassing optical isomer separations, sensors, and analytical applications (Anderson 2021; Chen et al. 2021). Additionally, the elevated binding capacity of these sorbents may facilitate the efficient extraction of analytes from a broad spectrum of matrices (Antonio 2022).

TABLE 1.

Physiochemical properties of pharmaceutical compounds in this study.

Compound	Structure	Molecular weight (g mol−1)	Water solubility (mg L−1)	pK a	Uses
Fenoprofen		242.3	30	4.50	Pain reliever, swelling, and stiffness
Diclofenac		296.2	4.52	4.15	Pain reliever, nausea, headaches
Ibuprofen		206.3	41.1	4.85	Chronic arthritis, Bloating, nausea, dizziness
Naproxen		230.3	44.1	4.15	Pain reliever, cramps, Swelling
Gemfibrozil		250.3	8.4	4.42	Stomach pain, blurred vision, joint pain, rash

Note: pK _a is the acid dissociation constant.

Considering that food is one of the primary sources of nutrition for people, food safety is a serious public health concern around the globe due to rapid population growth and industrialization. Pharmaceutical residues in vegetables can have a variety of negative health consequences, including endocrine disruption, reproductive issues, increased cancer risk, and probable allergic reactions or disease. Long‐term exposure to low levels of these contaminants has been related to these issues, and some research suggests an increased risk of some cancers, including prostate, lung, and liver cancer (Minh‐Ky et al. 2023). The selected pharmaceuticals are widely used in South Africa (SA), both by prescription and over‐the‐counter, for pain and inflammation management and are anticipated to be existing at high concentrations in the SA environment. In recent years, the presence of selected pharmaceuticals in the environment has been reported; however, no studies have been conducted on their uptake on vegetables by MISPE. To the best of our knowledge, there is an absence of published research on the occurrence or uptake of the selected pharmaceuticals in vegetable absorption using MIPs and further assessment of human health risks. Moreover, various investigations have explored the synthesis and utilization of MIPs to selectively identify either naproxen, ibuprofen, or diclofenac as single compounds (Amdany et al. 2014; Sun et al. 2008; Lindqvist et al. 2005). This study aimed to simultaneously assess and detect naproxen, gemfibrozil, diclofenac, ibuprofen, and fenoprofen in vegetables that are often eaten raw, as salads. Additionally, recent research conducted in Africa has primarily concentrated on analysis of naproxen, ibuprofen, and diclofenac, extracted alongside other pharmaceuticals mostly in water matrices. Hence, the lack of African‐based research on the presence of pharmaceutical substances in plants, especially vegetables (Ngubane et al. 2019; Madikizela and Ncube 2021).

This study is pivotal for comprehending the extent of pollution, food security, and potential phytoremediation strategies. Furthermore, only a limited amount of research has been disseminated concerning the analysis of these pharmaceuticals in Durban, Kwa‐Zulu Natal (KZN), the designated study area for this investigation. A considerable portion of the population in SA Development Community (SADC) countries, such as SA, relies on food sourced from open markets. The bulk of food products, including vegetables, are provided by small‐ to medium‐scale farmers who often employ contaminated water for irrigation purposes. Additionally, there exists a scarcity of information regarding the toxicity and environmental impact of certain pollutants. Furthermore, the quantification of these chemicals is challenging due to the high costs associated with current analytical procedures, especially within the realm of sample preparation. Consequently, the validation and optimization of cost‐effective alternative extraction strategies could potentially mitigate some of these challenges. Therefore, this study endeavors to optimize and apply the MISPE approach for the analysis of naproxen, gemfibrozil, diclofenac, ibuprofen, and fenoprofen in vegetable samples.

2. Materials and Methods

2.1. Reagents

Diclofenac sodium salt (99%), naproxen (≥98 %), gemfibrozil (≥98%), fenoprofen (99%), ibuprofen (98%), and high‐performance liquid phase chromatographic (HPLC) grade acetonitrile (99.7%), acetone (≥99.8%), methanol (≥99.8%), dichloromethane (≥99.8%), 2‐vinylpyridine (99%), and 1,1′‐azobis‐(cyclohexanecarbonitrile) (98%) were obtained from Sigma‐Aldrich in Steinheim, Germany. Furthermore, 98% formic acid was obtained from Fluka in Steinheim, Germany.

2.2. Instrumentation

Analytes separation was performed on Shimadzu Corporation, HPLC obtained from Kyoto in Japan. The system consisted an SPD‐20A UV/vis detector, an online degasser unit (mobile phase: DGU‐20A3), an LC‐20AB pump, and a 20 µL sample loop. The mobile phase was made of a mixture of acetonitrile: 0.2% formic acid in water (60:40, v:v) at 1 mL min⁻¹ flow rate. The separation was performed on a C18 Kinetex column, 150 mm × 4.6 mm × 2.6 µm (Phenomenex, California, USA). The detector wavelength for analysis was 210 nm (ibuprofen and diclofenac), whereas naproxen, gemfibrozil, and fenoprofen were detected at 230 nm. Experiments for SPE were performed using a vacuum pump obtained from Pall Corporation in Fribourg, Switzerland. Strata‐X cartridges were obtained from Phenomenex, Aschaffenburg, Germany. Mixtasel—BLT 230v 50/60 hz centrifuge was obtained from Orto Alresa in Madrid, Spain. An RV 10 digital V‐C 4L rotary evaporator was obtained from Heidolph in Schwabach, Germany. Moulinex Blender was obtained from Prestige Laboratory (Durban, SA). The Series 2000 scientific oven was purchased from Scientific in Roodepoort, SA. Automatic Sweep Frequency Laboratory Ultrasonic Bath was obtained from Shalom Laboratory (Durban, SA).

2.3. Synthesis of the Molecularly Imprinted Polymer

The MIP synthesis was prepared, characterized, and optimized using a modified method developed from our previous work (Nkosi et al. 2022). A two‐stage reaction procedure was employed to complete the bulk polymerization synthesis of a polymer. The first stage involved dissolving all of the templates at 0.1 mmol each in a 250 mL round bottom flask and thereby adding 54 µL of 2‐vinyl pyridine (2‐VP) and a mixture of acetonitrile/toluene (1:3, v:v). The mixture was stirred at room temperature for 30 min. The second step involved the addition of 100 mg of 1,1′‐azobis‐(cyclohexanecarbonitrile) as a radical initiator and 4.77 mL Ethylene glycol dimethylacrylate. The mixture was purged with nitrogen gas for 10 min, taped up, and stirred in an oil bath at 60°C for 16 h to initiate polymerization. Usually, initial polymerization is conducted at temperatures around 60°C in order for sufficient numbers of radicals to form as much higher temperature polymerization during this process destabilizes colloidal structures (Duan et al. 2023). However, in order to achieve a solid polymer once polymerization has been accomplished; the temperature was increased after 16 h to 80°C and maintained for 24 h. The resulting polymer was oven‐dried at 80°C to constant mass. Using an excess functional monomer than the template is a typical tactic in molecular imprinting, especially in non‐covalent imprinting techniques. In order to create the imprinted cavities in the polymer, this method encourages the creation of template‐monomer assemblies.

2.4. Sample Collection and Pre‐Treatment

The edible part of vegetable samples was purchased from fruits and vegetables open markets around Durban, KZN. These included lettuce, cucumber, carrot, and pepper. These were chosen because they have gained popularity among consumers worldwide and represent a good source of minerals, vitamins, and diverse phytochemicals (Noopur et al. 2023). All samples were washed with distilled water before blending and analyzed immediately after blending.

2.4.1. Sample Extraction

A method from the literature that had been modified was used for the extraction (Nkosi et al. 2022; Madikizela and Chimuka 2016). Experimentally, vegetable samples were blended; therefore, 25 mL of methanol was added to a 5 g of blended vegetable sample, resulting in a further stirred solution. The resulting mixture was sonicated in an ultrasonic bath for 20 min at ambient temperature and kept for 5, 10, 15, and 20 min each for extraction; this was done to determine extraction efficiencies at different intervals. The samples were, therefore, centrifuged for 30 min at 2500 revolutions per minute. A 250 mL round bottom flask was used to preserve the liquid extracts. The extraction was repeated, where the extracts were combined, and a rotary evaporator was used to evaporate them to 1 mL. Subsequently, the samples were diluted with 200 mL of deionized water each. Finally, it was filtered, and pH adjusted to 4.0.

2.4.2. Optimization of the Molecularly Imprinted Solid Phase Extraction Method

This was achieved by mounting the empty polypropylene single‐fitted (3cc) cartridges into the manifold; the cartridges were rinsed with the solvent first. Thereafter, the first polypropylene frit was inserted in the bottom of the cartridge and packed with 40 mg of MIP particles; the second frit was placed on top. The packed cartridge was first conditioned with 5 mL of methanol and then equilibrated with 5 mL of deionized water. The vegetable sample (pH 4.0) of 200 mL was loaded and extracted at 1 mL min⁻¹. The cartridge was vacuum‐dried for 30 min, and the retained compounds were successively eluted with 10 mL acetonitrile. The extract was evaporated to dryness under nitrogen at 40°C. The residue obtained was reconstituted with 1 mL of acetonitrile and injected into the HPLC system.

2.4.3. Percentage Recovery Studies

Percentage recovery refers to the fraction of an analyte that is successfully recovered from a sample following a certain analytical technique and is calculated by dividing the total quantity of analyte recovered from the sample by the initial analyte concentration, then multiplying by 100, as shown in Equation (1). Strata‐X empty polypropylene single‐fitted SPE cartridges containing MISPE were employed to analyze selected vegetable samples. The pharmaceutical recovery percentages were assessed on the edible parts of vegetables, specifically roots for carrot, all parts for cucumber, fruit parts for pepper, and leaf parts for lettuce, which were previously spiked with a pharmaceutical solution mixture at ambient temperature. These experiments aimed to evaluate the pharmaceutical recovery percentages in vegetables. Each experiment was performed in triplicate, and recoveries were determined for each vegetable:

$$\mathrm{Percentage}\mathrm{recovery}\left(\%\right)=\frac{\mathrm{Co}}{\mathrm{Ce}}\times 100$$ (1)

In the MISPE analysis, extraction time, sample volume, solvent choice, and flow rate are some of the key parameters that play a significant role during extraction. Effect of time on recoveries was determined by employing samples from 5‐, 10‐, 15‐, and 20‐min extraction times to extract the analyte from the vegetable samples. The effect of sample volume on recoveries was investigated by extracting 50, 100, 200, and 500 mL vegetable samples spiked with a 5 mg kg⁻¹ (pH 4.0) mixture of selected pharmaceuticals. The effect of elution flow rate on recoveries was determined by employing sample flow rate 1–10 mL min⁻¹. Furthermore, the effect of solvent conditioning was conducted using lettuce as a test sample, which was spiked with a pharmaceutical solution concentration of 5 mg kg⁻¹. Equation (1) was then used to determine the percentage recoveries for each time interval, sample volume, elution flow rate, and conditioning solvent.

2.5. Method Validation

The analytical performance of the MISPE method for analyzing the pharmaceuticals of interest was tested using analytical methods. Linearity was performed using standard solutions and matrix‐matched calibrators; the limit of detection (LOD), the limit of quantitation (LOQ), and percentage recovery are parameters considered in this study. The precision was determined in terms of the relative standard deviation (%RSD). This was achieved using lettuce as a chosen sample, spiked with two different concentration levels (0.05 and 5 mg kg⁻¹), and analyzed in the same optimized conditions. In each case, three replicates were analyzed. The method's selectivity was assessed by spiking the lettuce sample with a pharmaceutical standard solution in the vegetable matrix. This is important, as several types of compounds can interfere with the extraction of selected compounds from vegetables, mainly due to their ability to compete for binding sites or disrupt the extraction process. These include other organic acids, plant metabolites, pigments like chlorophyll and carotenoids, proteins, lipids, and even some inorganic molecules (Tegegne et al. 2023). Because of unavailability of analyte free plant samples, lettuce plant leaves (5 g) grown in the laboratory with regular irrigation with analyte free water at the Mangosuthu University of Technology were used as blank samples.

2.5.1. Linearity, the LOD and Quantification

Linearity was assessed using a matrix matching method with spiked lettuce levels ranging from 0.05 to 5 mg kg⁻¹. To determine the linearity equation and regression coefficient (R ²), the peak areas of each analyte were plotted against the concentrations. For each analyte, the sensitivity of the optimized MISPE method for quantifying selected pharmaceuticals from vegetable samples was evaluated using the LOQ and LOD. LOD was calculated in the following equations, respectively:

$$\mathrm{LOD}=\frac{3.3\sigma }{S}$$ (2)

$$\mathrm{LOQ}=\frac{10\sigma }{S}$$ (3)

where the standard deviation of the response is represented by σ, and the slope of the calibration curve for each analyte is represented by S (Cömert et al. 2020).

2.6. Health Risk Assessment

One important criterion to evaluate the potential damage to ecosystems and public health from exposure to harmful substances is to compare the concentrations of these pollutants in samples to the maximum levels set by various national or international public organizations (Shalaby et al. 2021). A variety of criteria were established for this purpose; in addition to the hazard quotient (HQ) and the health index (HI) were also assessed for health quality control. For instance, HI is a tool that sums and quantifies different aspects of health, delivering a single value or a series of values to represent the overall health of a population or an individual. This parameter evaluates the impact of every target chemical in the food product (Shalaby et al. 2021; Chaikasen and Roiet 2020); in this case, HI was evaluated using HQ. Literature states that these indicators can be computed with the help of the following equations (Chaikasen and Roiet 2020; Zhang et al. 2021):

$$\mathrm{HQ}=\frac{Ci}{\mathrm{MLR}}$$ (4)

$$\mathrm{HI}=\mathrm{\Sigma }\mathrm{HQ}$$ (5)

where Ci represents the compound's concentration (mg kg⁻¹) in the sample, whereas MRL represents the maximum residual limit. The health risk assessment (HRA) with respect to children and adults was estimated using the following equation, which is the ratio of the acceptable daily intake (ADI in mg kg⁻¹) and the estimated daily intake (EDI) (Wang et al. 2022; Shalaby et al. 2021):

$$\mathrm{HRA}=\frac{\mathrm{EDI}}{\mathrm{ADI}}$$ (6)

The following equation was used to calculate the EDI of each target chemical in each vegetable sample (Shalaby et al. 2021; Wang et al. 2022):

$$\mathrm{EDI}=\frac{Cp\times D}{bw}$$ (7)

Cp is the concentration of pharmaceutical residues in vegetables (mg kg⁻¹). D is the average daily consumption of fresh vegetables for the African region, estimated at 38.8 g person⁻¹ day⁻¹ (WHO 2003); bw is the average body weight, which is regarded as 60 kg for an adult and 15 kg for a kid (Shalaby et al. 2021).

3. Results and Discussion

3.1. Optimization of MISPE

3.1.1. Effect of Sample Volume on the Recoveries of SPE

The sample volume in SPE is an important parameter that needs to be carefully tuned to obtain reliable and precise analyte recovery. Understanding the effect of volume of sample on sorbent capability and breakthrough is critical to developing and evaluating effective SPE procedures (Maranata et al. 2021). The obtained results are presented in Figure 1. Sample volumes of 50 and 100 mL resulted in lower percentage recoveries, between 25%–35% and 29%–39%, respectively. The low percentage recoveries could be because most of the analytes are not yet partitioned on the sorbent. Therefore, acceptable recoveries from 80% to 98% were achieved for a 200 mL sample. Hence, a 200 mL sample volume was chosen as the optimum. Recoveries ranging from 38% to 54% were obtained for a 500 mL sample. The decrease in recoveries at a high sample volume (500 mL) suggests that the sample breakthrough volume was exceeded. Elsewhere, emtricitabine, tenofovir, naproxen, diclofenac, ibuprofen, efavirenz, and gemfibrozil were investigated by extracting 500 and 1000 mL deionized water samples. The 500 mL sample volume was chosen as the optimum (Sigonya et al. 2022). When the sample volume exceeds the sorbent's capacity, the analytes may pass through the cartridge without being retained, resulting in breakthrough and lower recoveries (Mandal et al. 2023). This notion corresponds with the statement reported by Bagheri et al., stating that increasing the sample volume may improve the efficiency of binding the analyte to the sorbent, but only up to a certain point, after which there is no further improvement.

FIGURE 1.

The effects of sample volume on percentage recovery.

Other researchers have also observed this concept elsewhere (Sibeko et al. 2019; Madikizela and Chimuka 2017).

3.1.2. Effect of Elution Flow Rate on the Recoveries of MISPE

In essence, the flow rate is a key element in SPE that must be precisely managed to enable efficient and dependable analyte extraction. The flow rate and the analytes’ retention capacity are both influenced by the sample solution's flow rate via the SPE column. To shorten the analysis time, the flow rate needs to be sufficiently high. However, it needs to be sufficiently sluggish to carry out an efficient retention of the analytes (Hossein et al. 2016). Hence, optimizing the total quantity of analyte eluted through the SPE sorbent and guaranteeing precise analytical findings depend on optimizing this flow rate (Maranata et al. 2021). In this study, a sample flow rate of 1–10 mL min⁻¹ was examined, and the results are shown in Figure 2. SPE sample percolation and elution in flow rate analysis are essential as this impact how the analytes of interest are retained and eluted from the cartridge (Rimayi et al. 2018). The results reveal that elution flow rate and percentage recovery are in inverse proportion. This is often anticipated as setting up a slow flow rate, which results in more analytes from the desorbed sorbent. Although gemfibrozil and naproxen showed good outcomes (>80%) with a 5 mL min⁻¹ flow rate, ibuprofen, fenoprofen, and diclofenac were recovered below acceptable limits, at 61%, 60%, and 66%, respectively. Hence, to achieve acceptable recoveries for all compounds (80%–110% with a %RSD value <15), the elution flow rate of 1 mL min−¹ was selected. All compounds’ 10 mL min⁻¹ flow rate results were unsatisfactory (<80%).

FIGURE 2.

The effects of elution flow rate on percentage recovery.

3.1.3. Effect of Conditioning Solvent on the Recoveries of MISPE

The conditioning effect of solvent in SPE is critical for effective analyte collection and subsequent elution. Conditioning eliminates contaminants and activates the sorbent, resulting in improved analyte interactions and, eventually, higher recoveries. Choosing the right conditioning solvent is critical for increasing the efficiency of the SPE process (Baciu et al. 2015). The evaluation of solvent conditioning was conducted using lettuce as a test matrix, which was spiked with a pharmaceutical solution at a concentration of 5 mg kg⁻¹. The initial phase of the procedure involves the conditioning stage; an appropriate solvent is introduced through the SPE cartridge to moisten the sorbent bed, thereby activating the sorption functional groups and enhancing the surface area available for analyte binding. The conditioning phase is critical in SPE analysis, as it influences the interaction between the compounds and the SPE sorbent, thereby affecting the yield of recovered analytes. This investigation examined the effect of the conditioning solvent using acetonitrile, dichloromethane, and methanol, given their widespread use in the extraction of acidic pharmaceuticals (Sigonya et al. 2022). An optimal sample volume of 200 mL was utilized. Figure 3 illustrates that recoveries for all compounds surpassed 80% for all solvents tested. Nevertheless, acetonitrile achieved the highest recoveries (105%–112%), in contrast to dichloromethane (81%–92%) and methanol (82%–94%). This may be attributed to acetonitrile's greater polarity, 5.8, and lower viscosity, 0.38, relative to dichloromethane, with a polarity of 3.1 and viscosity of 0.44, and methanol, with a polarity of 5.1 and viscosity of 0.59. This suggests the effectiveness of acetonitrile in activating sorbent functional groups, enhancing solvent–analyte interaction, and facilitating pore opening (Sigonya et al. 2022). Consequently, acetonitrile was determined to be the optimal conditioning solvent. Its greater polarity enables it to dissolve a greater variety of substances, whereas the lower viscosity facilitates quicker diffusion and lessens the system's pressure drop, both of which improve separation and analysis. Acetonitrile recoveries greater than 100% could be due to co‐elution of interfering compounds.

FIGURE 3.

The effects of conditioning solvents on percentage recovery.

These findings are consistent with the study by (Hlengwa and Mahlambi 2020), which concentrated on optimizing and applying ibuprofen, fenoprofen, naproxen, carbamazepine, and diclofenac in river water samples. Their study reported that acetonitrile (89%–120%) yielded higher recoveries compared to methanol (78%–99%) for the majority of the compounds tested.

3.1.4. Effect of Time on Recoveries of MISPE

Given that the extraction time directly affects the quantity of desirable molecules removed from a matrix, it has a substantial impact on efficiency. In order to maximize yield and minimize expenses, extraction time optimization is essential. Extended exposure can degrade the extracted compounds, decreasing overall efficiency, even though extended extraction durations may initially result in higher yields (Che Sulaiman et al. 2017). Figure 4 shows the obtained results, where 10 min extraction time was chosen as an optimum, giving the maximum recoveries (>80%) for all compounds. This symbolizes the MIP having binding sites conforming on its surface, which is suitable for template adsorption. Qwane, reported recoveries greater than 70%, achieved within 10 min of contact between the adsorbents and abacavir (Qwane et al. 2020).

FIGURE 4.

Effect of percentage recovery on time.

3.2. Application and Method Validation Studies

To guarantee the correctness and dependability of analytical procedures and outcomes, recovery and validation studies are essential. These investigations aid in assessing a method's practical performance, including its capacity to precisely measure and extract the target analyte from a sample matrix. Recovery studies evaluate the effectiveness of extracting and measuring the analyte, whereas validation studies specifically confirm that a method satisfies specified performance criteria (Thakur et al. 2022). A statistical metric called RSD is used to express how accurate or consistent a collection of measurements or data points is. In essence, it indicates the degree of variation between your data points and the data's average or mean. The RSD is calculated by dividing the standard deviation by the average value of the calibration or response components. Generally, a threshold of less than 15% or less than 20% will be applied in order to accept the calibration (Burrows and Par 2020). The obtained percentage recoveries and %RSD results are presented in Table 2, showing the satisfactory % RSD < 15. Figure 5 further illustrates trends in MISPE vegetable recoveries, ranging from 45% to 103%. The recovery percentages higher than 100% could be due to minor errors in the analytical method, such as co‐eluting interferences or matrix effects that enhance the signals. The carrot samples reported all compound recoveries below 80%, except for diclofenac. Similarly, the pepper sample observed recoveries below 80% concerning fenoprofen. In some instances, low recoveries could be attributed to the suppression effect of the plant matrix. Literature indicates that vegetables contain numerous colorants and components, which can analytically affect the detection and recovery of a real sample (Tegegne et al. 2023). Tegegne et al. (2023) reported recovery rates from 33.5% to 115% in their optimization method for the extraction and HPLC analysis of carrot, cabbage, and lettuce samples collected in Ethiopia.

TABLE 2.

Percentage recovery and relative standard deviation (%RSD) for each selected pharmaceuticals in vegetable (n = 3).

	Gemfibrozil		Naproxen		Ibuprofen		Diclofenac		Fenoprofen
Vegetable sample	%R	%RSD	%R	%RSD	%R	%RSD	%R	%RSD	%R	%RSD
Lettuce	95	6.7	97	10.2	96	12.0	95	6.2	103	6.4
Carrot	66	12.4	72	13.0	57	9.8	88	11.7	45	11.5
Pepper	96	2.8	85	1.3	90	7.6	94	2.5	62	3.8
Cucumber	99	7.9	91	1.1	98	3.4	98	7.5	99	6.9

Abbreviations: molecularly imprinted solid phased extraction (MISPE); %R, percentage recovery; RSD, relative standard deviation.

FIGURE 5.

Recoveries of MISPE on vegetables.

Selectivity is known as the capacity of a technique to identify a particular analyte free from interference from other compounds in the sample (Tegegne et al. 2023). The optimal MIPSE method's good selectivity towards the selected pharmaceuticals is demonstrated by the chromatograms in Figure 6A–C, and this bodes well for the analysis of vegetable samples in the future.

FIGURE 6.

(A) HPLC chromatogram for non‐spiked vegetable extract. (B) Typical HPLC chromatogram for spiked vegetable extract (1 = Naproxen, 2 = Fenoprofen, 3 = Diclofenac, 4 = Ibuprofen, 5 = Gemfibrozil). (C) HPLC chromatogram for the pharmaceutical's standard mixture (1 = Naproxen, 2 = Fenoprofen, 3 = Diclofenac, 4 = Ibuprofen, 5 = Gemfibrozil).

To determine the linearity equation and regression coefficient (R ²), the peak areas of each analyte were plotted against the concentrations, and the data (Table 3) demonstrated a good linearity, ranging from 0.9502 to 0.9985 between the analytical signal and analyte concentration, which indicates good precision and accuracy of the optimized method. The LOD and the LOQ were calculated as 3 and 10 times the signal‐to‐noise ratio, respectively. Table 3 shows that the LOQs ranged from 1.0 to 2.5 mg kg⁻¹ and the LODs from 0.1 to 0.8 mg kg⁻¹. The method's good sensitivity is indicated by the low values for LODs and LOQs, which may enable the detection of pharmaceuticals at trace quantities. The study's LOD and LOQ are slightly higher than those found in previous studies that used HPLC in conjunction with a photodiode array detector for the analysis of similar compounds, which were 0.02–0.05 and 0.05–0.18 mg L⁻¹, respectively (Montagner and Jardim 2011). Reproducibility and repeatability were investigated in terms of intra‐day and inter‐day, respectively. This was achieved by calculating the %RSD, using lettuce sample spiked with two different concentration levels. In each case, three replicate spiked lettuce samples were evaluated under the same optimal conditions. The precision findings in repeatability and reproducibility were satisfactory for all tested pharmaceutical compounds, with a value of less than 15% (Table 4).

TABLE 3.

Analytical performance of molecularly imprinted solid phase extraction (MISPE) for the determination of selected pharmaceuticals in vegetable using matrices matches analysis.

Compound	Calibration range (mg kg−1)	R 2	LOD (mg kg−1)	LOQ (mg kg−1)
Diclofenac	0.05–5	0.9502	0.8	2.5
Fenoprofen	0.05–5	0.9861	0.4	1.3
Gemfibrozil	0.05–5	0.9985	0.1	1.0
Ibuprofen	0.05–5	0.9666	0.5	1.5
Naproxen	0.05–5	0.9611	0.6	1.8

Note: R ² is the linear correlation.

Abbreviations: LOD, limit of detection; LOQ, limit of quantitation.

TABLE 4.

Analytical performance on reproducibility (intra‐day) and repeatability (inter‐day), % relative standard deviation (RSD).

Compounds	Intra‐day, %RSD (n = 3)		Inter‐day, %RSD (n = 6)
	Level 1	Level 2	Level 1	Level 2
Diclofenac	2.3	7.2	6.3	8.7
Fenoprofen	3.5	5.5	5.7	5.9
Gemfibrozil	2.9	6.4	7.8	8.3
Ibuprofen	3.0	4.8	5.2	7.0
Naproxen	1.2	1.9	10.1	11.2

Note: Level 1 = 0.05 mg kg⁻¹, Level 2 = 5 mg kg⁻¹.

3.2.1. Matrix Effects Evaluation

Matrix effects are effects on an analytical test caused by all sample constituents other than the specific compound (analyte) being analyzed (Tegegne et al. 2023). These effects are often observed within the field of chemical analysis, causing analytical signals to be suppressed or amplified (Dural et al. 2015). This occurrence happens because some target chemicals and matrix compounds elute simultaneously (Tisler et al. 2021). This also depends on the type of matrix and the efficiency of the sample preparation step. Therefore, the sample preparation phase should eliminate interfering substances while retaining the target analytes (Yifeng et al. 2023). To study this occurrence, solvents without matrix and with matrix were spiked by the addition of 1 mg kg⁻¹ of the pharmaceuticals, and the blank without spike was analyzed using the optimized MISPE methods. This effect was calculated using the following equation, A (matrix), A (blank), and A (solvent), which represents the area in the spiked matrix, the area in the blank matrix, and the area in the spiked solvent without the sample matrix, respectively:

$$\mathrm{Matrix}\mathrm{effect}\left(\%\right)=\frac{A\left(\mathrm{Spiked}\right)-A\left(\mathrm{Blank}\right)}{A\left(\mathrm{Solvent}\right)}-1\times 100$$ (8)

Figure 7 represents the calculated matrix effect results showing variability between the targeted compounds and types of vegetable samples analyzed. The signal suppression matrix effect has been higher in the lettuce and pepper than in other vegetables. Soft matrix effects values ±37% to ±1% were achieved for studied compounds in vegetable samples. Matrix effect between <20%, 20%–50%, and >50% is considered soft, medium, and firm, respectively (Miossec et al. 2018).

FIGURE 7.

Matrix effect (%) calculated for all target compounds in vegetable sampled on MISPE.

3.3. Concentrations of Pharmaceuticals in Vegetables

The optimized and validated analytical method was applied to monitor the levels of gemfibrozil, naproxen, ibuprofen, diclofenac, and fenoprofen in vegetables. The concentrations of these compounds in vegetable samples are presented in Table 5. The vegetables examined included lettuce, carrot, pepper, and cucumber. Several studies using various extraction methods have been conducted in different countries to investigate pharmaceuticals in vegetables (Mravcová et al. 2024; Zhou et al. 2017). However, limited data exist on the compounds investigated in the current study in vegetables, particularly through MIPs sample preparation. All target compounds were quantified in all analyzed vegetables except for ibuprofen, which was not detected in a cucumber sample (Table 5). On the basis of the average concentrations, fenoprofen was the most abundant compound, with high concentrations in pepper and cucumber, reaching a maximum average concentration of 6.44 (ww) and 4.99 mg kg⁻¹ (ww), respectively. The highest average concentrations observed were in the pepper sample for naproxen, 2.82 mg kg⁻¹ (ww) and ibuprofen, 2.06 mg kg⁻¹ (ww), in the carrot sample for gemfibrozil, 2.51 mg kg⁻¹ (ww), and in the cucumber sample for naproxen, 2.92 mg kg⁻¹ (ww). The elevated concentrations of fenoprofen, ibuprofen, and naproxen could be attributed to the excretion rates of these compounds, which exceed 90% in urine (Ponce‐Robles et al. 2022; Ramadan et al. 2020; Ahmed et al. 2019). Consequently, significant amounts of these compounds reach WWTPs; thus, high traces may be present in effluent water (used for irrigation) and sludge (used as fertilizer), ultimately ending up in vegetables. These findings demonstrate the method's applicability. The detected levels of pharmaceutical compounds in vegetables were either lower than or comparable to those reported by other researchers. For instance, Calderón‐Preciado et al. (2009) reported a concentration of 2.90 mg kg⁻¹ for ibuprofen in lettuce. Dong et al. (2013) reported naproxen concentrations ranging from 0.2 to 2.4 mg kg⁻¹ (fresh weight) in plant samples. To the best of our knowledge, this represents the first instance where MIPs have been utilized to extract concentration levels of the targeted compounds in vegetables and to further evaluate the human health risk assessment.

TABLE 5.

Average concentrations (mg kg⁻¹) of selected pharmaceuticals in vegetables (n = 3).

Vegetable samples	Gemfibrozil	Naproxen	Ibuprofen	Diclofenac	Fenoprofen
Lettuce	0.22 ± 1.9	0.81 ± 1.1	0.45 ± 1.2	0.88 ± 1.0	0.89 ± 0.4
Carrot	0.10 ± 2.1	2.51 ± 2.0	0.24 ± 1.5	1.45 ± 2.2	0.24 ± 0.4
Pepper	1.11 ± 2.3	2.82 ± 0.1	2.06 ± 2.3	0.71 ± 1.8	6.44 ± 2.3
Cucumber	0.01 ± 1.0	2.92 ± 1.8	Nd	1.93 ± 1.3	4.99 ± 1.3
MRL	600	500	400	150	200

Abbreviations: MRL, maximum residual limit; Nd, not detected.

3.4. Evaluation of Health Risk Exposure

Consumption of contaminated food results in exposure to all inherent contaminants. Table 6 presents the estimated HQ and HI values derived from samples of lettuce, carrot, pepper, and cucumber. The target compounds were collectively identified in the vegetables, with HI values for lettuce, carrot, pepper, and cucumber determined to be 0.27, 0.42, 1.25, and 0.90, respectively. The HI calculated for the pepper sample, aggregating all chemical constituents, was 1.25, which exceeds the value of 1, suggesting a potential risk associated with the consumption of these foods available in the Durban open markets. The literature suggests that an HI value exceeding 1 indicates the presence of pollutants at a hazardous level (Madilonga et al. 2021). Elsewhere, HI values surpassing the permissible limit for pharmaceuticals in cabbage and kale samples, with values of 23.22 and 30.01, respectively, have been reported (Dong et al. 2013). Further, Table 6 contains HRA values, which assess the nature and likelihood of adverse health effects in humans potentially exposed to pollutants in contaminated environmental media, presently or in the future (Silvestre et al. 2023). HRA values have been calculated for both children and adults exposed to the target compounds in vegetable samples. Consumers of vegetables from Durban's open markets experience equivalent levels of exposure regardless of age (Table 7). The computed HRA values for lettuce consumers vary between 0.00024 and 0.0073 for adults and 0.00095 to 0.015 for children. The HRA values for carrot consumers range from 0.00011 to 0.00063 for adults and from 0.00043 to 0.025 for children, whereas for pepper, values span from 0.0012 to 0.021 for adults and from 0.0048 to 0.083 for children. Cucumber consumers exhibit HRA values from 0.000012 to 0.016 for adults and from 0.000043 to 0.065 for children. Because all these values are below 1, there is no associated health risk from the consumption of cucumber, lettuce, carrot, or pepper, irrespective of age. This analysis indicates no potential health impacts from consuming these vegetables. Literature indicates a concentration of pollutants at toxic levels is denoted when the HRA exceeds 1 (Akoto et al. 2016; Nuapia et al. 2016). These findings align with a study by Dorde, on the human health risk assessment of pharmaceuticals in vegetables (lettuce, tomato, cauliflower, and bean seeds), reporting HRA values ranging from 0.000032 to 0.0022 across all tested compounds (Đorđe et al. 2021). HRA is a technique for identifying and evaluating potential health hazards in individuals or communities, taking into account characteristics such as age, habits, and environmental exposure. However, HI is a tool that summarizes and quantifies different aspects of health, providing a single value or a set of values to represent the overall health of a population or an individual (Đorđe et al. 2021). HRA assessments are determined for both children and adults, as children are more vulnerable to environmental dangers due to their developmental physiology and distinct behaviors. Although the same quantity of a toxin may represent little or no harm to an adult, it can have serious developmental consequences for children. Furthermore, children are more likely to be exposed to environmental contaminants because they breathe more air, consume more food and water, and have a larger skin surface area compared to body weight (Taghavi et al. 2025).

TABLE 6.

Hazard quotient and health index of the selected pharmaceuticals.

	Gemfibrozil	Naproxen	Ibuprofen	Diclofenac	Fenoprofen	HI
Lettuce	0.01	0.08	0.05	0.04	0.09	0.27
Carrot	0.007	0.3	0.02	0.07	0.02	0.42
Pepper	0.07	0.3	0.2	0.04	0.64	1.25
Cucumber	0.00067	0.3	Nd	0.10	0.5	0.90

Abbreviation: HI, health index.

TABLE 7.

Health risk index of the selected pharmaceuticals.

Adults
	Gemfibrozil	Naproxen	Ibuprofen	Diclofenac	Fenoprofen
Lettuce	0.00024	0.0010	0.0073	0.0038	0.0023
Carrot	0,00011	0.0032	0.0004	0.0063	0.00078
Pepper	0.0012	0.0036	0.0035	0.0063	0.021
Cucumber	0.000012	0.0037	Nd	0.0083	0.016

Children
	Gemfibrozil	Naproxen	Ibuprofen	Diclofenac	Fenoprofen
Lettuce	0.00095	0.042	0.0029	0.015	0.012
Carrot	0.00043	0.012	0.0016	0.025	0.0031
Pepper	0.0048	0.015	0.013	0.012	0.083
Cucumber	0.000043	0.015	Nd	0.033	0.065

4. Conclusion

Our current study for the first time developed and validated a MISPE method coupled with HPLC for simultaneous determination of naproxen, gemfibrozil, fenoprofen, ibuprofen, and diclofenac and effectively implemented on various vegetable samples, namely, lettuce, carrot, cucumber, and pepper, collected from Durban market place to assess the human health risk. Recoveries of the method spanned from 45% to 103%, with %RSDs maintained below 15%. Soft matrix effects were achieved for the studied compounds in vegetables samples. Future applications of this method will be directed towards hydroponic investigations of pharmaceutical uptake in vegetables to elucidate their bioaccumulation and translocation metrics within leaves, stems, and roots. The findings of this investigation reveal the occurrence of gemfibrozil, naproxen, ibuprofen, fenoprofen, and diclofenac in lettuce, carrot, pepper, and cucumber available in Durban's open markets. HI suggests that the calculated values are below hazardous thresholds, except for the pepper sample, which could present a potential health risk to the public with respect to the amount of all the tested pharmaceuticals in pepper. The HRA values were all below acceptable limit, which implies no human health risk for both children and adults.

Author Contributions

S'busiso M. Nkosi: conceptualization, writing – original draft, investigation, methodology, validation, visualization, data curation. Njabulo J. Gumede: conceptualization, writing – review and editing, project administration, supervision. Precious N. Mahlambi: conceptualization, writing – review and editing, supervision, resources, project administration.

Conflicts of Interest

The authors declare no conflicts of interest.

Nkosi, S. M. , Gumede N. J., and Mahlambi P. N.. 2025. “Development, Optimization, and Application of Molecularly Imprinted Polymers‐Solid Phase Extraction Procedure for the Analysis of Selected Pharmaceuticals in Vegetable Samples.” Journal of Food Science 90, no. 10: e70556. 10.1111/1750-3841.70556