, Volume 82, Issue 1, pp 287–295 | Cite as

Synthesis and Characterization of Molecularly Imprinted Polymers for the Selective Extraction of Carbamazepine and Analogs from Human Urine Samples

  • Audrey Combes
  • Porkodi Kadhirvel
  • Louis Bordron
  • Valerie PichonEmail author
Part of the following topical collections:
  1. 50th Anniversary Commemorative Issue


Two molecularly imprinted polymers (MIPs) were synthesized according to a previous work from our group dealing with the extraction of carbamazepine from environmental water. The potential of these MIPs, which differ in the nature of the monomer used for their synthesis, to selectively extract the drugs carbamazepine and oxcarbazepine and the metabolite 10,11-epoxycarbamazepine was first studied in spiked pure water, and high selectivity was obtained with both MIPs for the three target molecules in this pure medium. This selectivity was maintained when applying one of the MIPs to urine samples. Indeed, extraction recoveries were higher than 82% on the MIP and lower than 20% on the corresponding non-imprinted polymer used as a control. The repeatability of the extraction procedure applied to urine was also demonstrated, with relative standard deviation (RSD) below 20% for extraction recoveries of the three targets at a spiking level of 20 ng L−1. Limits of quantification between 1 and 7 ng L−1 were determined for urine samples using the MIP as extraction sorbent combined with LC–MS analysis. The potential of the MIP was compared to that of the Oasis HLB sorbent. This study shows that the MIP constitutes a powerful tool for avoiding matrix effects encountered in the quantification of the target molecules in urine samples extracted on Oasis HLB.


Molecularly imprinted polymers Carbamazepine Oxcarbazepine Metabolite Urine LC–MS analysis 


Carbamazepine (CBZ) is the most commonly used drug to treat partial epileptic seizures. Oxcarbazepine (OXC) is a structural derivative of carbamazepine, with a ketone on the dibenzazepine ring, and was developed in an effort to achieve effectiveness comparable to carbamazepine, while causing fewer side effects. Because of the high intake of these drugs, they are frequently detected in urban wastewater cycles due to their excretion in urine and feces [1]. Carbamazepine is metabolized in the liver, with the primary metabolic pathway involving the conversion to 10,11-epoxycarbamazepine (epoCBZ), which can also be found in water samples. Indeed, concentration levels in surface water in some locations in Germany have recently reached 1.64, 0.44 and 0.08 µg L−1 for CBZ, OXC and epoCBZ, respectively [1]. As urine constitutes a major route of water contamination, detecting these compounds in this biological fluid constitutes an important task.

Such analysis can be achieved by numerous separation methods including gas, high-performance liquid or thin-layer chromatography and electrokinetic techniques, as recently reviewed [2], yet liquid chromatography coupled with mass spectrometry (LC–MS) remains an inevitable choice because of its high sensitivity and specificity. However, the low concentrations that must be reached, combined with the complexity of urine samples, necessitates the use of a sample pretreatment method to concentrate and purify the samples before their analysis. For this, different approaches have been proposed, including dilution, protein precipitation, liquid–liquid extraction (LLE), dispersive LLE, solid-phase extraction (SPE) and dispersive SPE [3, 4, 5, 6]. To improve the sensitivity of this sample pretreatment step by the removal of interfering compounds, a procedure was recently proposed that combines a first extraction using acetonitrile on a freeze–dried sample of urine with an SPE step on a mixed-mode sorbent [3]. Improved selectivity can also be obtained using molecularly imprinted polymers (MIPs). These polymers are synthesized in the presence of a template molecule, leading to the formation of a polymer that possesses cavities that are complementary in size, shape and chemical function to the template compound and that will determine its selectivity. Indeed, the ability of a MIP to selectively recognize and then retain targeted compounds is directly dependent on the shape and the nature of the chemical function of the cavities, which are both fixed by the conditions of synthesis of the MIPs, i.e. the nature of the template, monomers, cross-linker and porogenic solvent.

MIPs were previously developed for the selective extraction of CBZ [7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18] which involved the systematic use of CBZ as template molecule, with the exception of one recent study from our group [18]. Indeed, as the synthesis process of the MIP requires the introduction of a large amount of template molecule, and in light of the targeted concentrations in samples, this choice was not considered relevant. Even after careful washing of the polymer after its synthesis and rigorous control of this step, the template molecule can still leak from the MIP during subsequent extraction procedures, thus leading to false positives. This can be circumvented by selecting a CBZ analog, methoxycarbamazepine (MCBZ), which can be distinguished from CBZ by its different retention time in LC–MS analysis [18]. In most cases, methacrylic acid (MAA) has been used as monomer in conjunction with ethylene glycol dimethacrylate (EGDMA) [12, 13, 15, 16], trimethylolpropane trimethacrylate (TRIM) [10, 11] or divinylbenzene (DVB) [7, 8, 9, 14, 18] as cross-linker, and mainly in acetonitrile, toluene, chloroform and dichloromethane or in some mixture of these solvents such as porogen. A recent study also reported the synthesis of an imprinted interpenetrating polymer network based on a mixture of styrene and tetramethoxysilane [17]. Some of the CBZ MIPs have been applied to the extraction of CBZ from urine [7, 8, 16, 17], serum [12] or plasma [16] and environmental water [9, 13, 14, 18]. All these MIPs were synthesized for the selective extraction of CBZ, while a few studies reported results concerning the recognition of its structural analog OXC in water [13, 18], serum [12] and urine [8, 12], with highly variable results in terms of selectivity.

The selectivity of a MIP can be evaluated by comparing the extraction recoveries of a target analyte on a MIP with those obtained using a non-imprinted polymer (NIP). The latter is synthesized under the same conditions as the MIP but without the introduction of the template, and the lack of specific cavities in the NIP generate lower extraction recovery on this sorbent than on the MIP. For water samples, selectivity has been demonstrated for the extraction of CBZ, OXC and even epoCBZ (a CBZ metabolite) using a MIP obtained with MAA and DVB for its synthesis [18]; no retention or selectivity was obtained using a MIP prepared with 2-vinylpyridine and EGDMA [13]. For biological fluids, high retention of CBZ in serum [12] and urine [7, 8] has been reported, even for OXC [8, 12], but without providing any results derived from NIP to control the selectivity. The selective extraction of CBZ from urine was reported by Asgari et al., with recoveries of 91% and about 20% on MIP and NIP, respectively [17]; however, none of the data related the extraction of OXC and CBZ metabolites.

Two MIPs previously developed by our group using an analog of CBZ as template and using either TFMAA or MAA as monomer and DVB as cross-linker have shown high potential for the selective extraction of CBZ, OXC and epoCBZ from mineral and river water samples. Therefore, the objective of this work was to synthesize MIPs under the same conditions and apply them to the selective extraction of CBZ, OXC and epoCBZ from urine. After checking the selectivity of these newly synthesized MIPs in pure aqueous media, an extraction procedure was optimized for the treatment of urine samples. The potential of one of the MIPs was then further evaluated by measuring the extraction recovery of the three target analytes and by comparing the performance of this MIP with that of a conventional Oasis HLB polymer.

Materials and Methods


Trifluoromethacrylic acid (TFMAA; 98%) was purchased from Apollo Scientific Ltd (Manchester, UK). Carbamazepine (CBZ; 98%) and oxcarbazepine (OXC; 98%) were procured from ABCR (Karlsruhe, Germany), methoxycarbamazepine (MCBZ; 95%) from Fluorochem Ltd (Derbyshire, UK), and methacrylic acid (MAA; 99%), divinylbenzene (DVB; 80%) and carbamazepine 10,11-epoxide (epoCBZ; 98%) from Sigma-Aldrich (Saint Quentin Fallavier, France). MAA was distilled under a vacuum to remove inhibitors. Azobisisobutyronitrile (AIBN; 98%) was purchased from Acros Organics (Noisy-le-Grand, France). HPLC-grade acetonitrile (ACN), methanol (MeOH), dichloromethane (DCM), hexane and toluene were supplied by Carlo Erba (Val de Reuil, France). Acetic acid (AA; 99.7%) was purchased from VWR (Fontenay-sous-Bois, France). High-purity water was dispensed by a Milli-Q purification system (Millipore, Saint Quentin en Yvelines, France).

Instrumentation and Analytical Conditions

The LC–MS/MS analyses were performed using a liquid chromatograph (UltiMate 3000®, Thermo Scientific, Illkirch, France) coupled with a triple-stage quadrupole mass spectrometer (TSQ Quantum Access MAX, Thermo Scientific) equipped with a heated electrospray ionization source (HESI-II). The LC/MS acquisitions were controlled using Thermo Xcalibur Control software version 2.2. The chromatographic separation was performed on a Varian C18 Omnispher column (150 × 2.1 mm, 5 µm) maintained at 35 °C. The mobile phase, MeOH/ACN/H2O (38/20/42, v/v/v), was flowed through the column at 0.2 mL min−1, and the injection volume was set at 5 µL. MS was operated in positive ion mode with multiple reaction monitoring (MRM) detection using an electrospray voltage of 3000 V and a tube lens offset of 60 V for CBZ and 72 V for epoCBZ and OXC. Capillary and vaporizer temperatures were set at 350 °C and 300 °C, respectively. Nitrogen was used as desolvation gas and argon as collision gas at a pressure of 1.5 mTorr. Two transitions were monitored for each compound (the first transition for each compound gave the highest signal and was used for quantification), and the collision energy was optimized and indicated in brackets: (1) CBZ: 237 > 194 (19 V), 237 > 179 (34 V); (2) epoCBZ: 253 > 236 (14 V), 253 > 180 (28 V); (3) OXC: 253 > 208 (19 V), 253 > 236 (12 V).

Synthesis of the MIPs

Two MIPs were synthesized using the procedure previously described by our group for the extraction of CBZ from environmental water [18]. Briefly, MCBZ was used as template, with a ratio 1:4:20 between the template, monomer and cross-linker. A 1-mmol sample of template and 4 mmol of monomer (TFMAA for MIP 1 and MAA for MIP 2) were mixed with 1.8 mL of a toluene/DCM (0.64/0.46, v/v) mixture. DVB (20 mmol) was then added and the mixture was purged for 10 min with an N2 stream to remove the dissolved oxygen and the initiator. The initiator (AIBN, 0.2 mmol) was then added into the above mixture, and the vial was sealed and placed in a water bath at 60 °C for 48 h. Then, the polymers were crushed, ground automatically in a mixer mill (MM 301; Retsch®) and sieved in a vibratory sieve shaker (Retsch®). Particles 25–36 µm in size were collected and sedimented four times with 5 mL of MeOH/water (80/20, v/v) to remove the thin particles and then dried at room temperature. Non-imprinted polymers (NIPs) were synthesized by performing the overall procedure, but in the absence of template.

Cartridges (3 mL) were packed with 55 mg each of MIP or NIP. The polymers were washed with 30 mL of MeOH/AA (90/10, v/v) to remove the template molecules. The washing step was continued until the template could no longer be detected in the washing fraction by LC/MS. Finally, the cartridges were washed with 10 mL of MeOH to remove residual AA. The same washing procedure was applied to the NIPs.

Repeatability of the Extraction Procedure in Aqueous Media

The retention capacity of the MIPs was first checked under conditions similar to those previously described [18] which ensured good selectivity for the extraction of CBZ from both MIPs. For the extraction of CBZ, epoCBZ and OXC from pure water, all the cartridges were conditioned with 2.5 mL of MeOH, 1.5 mL of HCl 0.1 M and then 2.5 mL of water before percolation of 1 mL of pure water spiked at 5 µg L−1 with CBZ, epoCBZ and OXC. A washing step with 300 µL of HCl 0.1 M followed by the same volume of water was performed prior to drying of the cartridge for 20 min under a vacuum. A second washing step with 1 mL of a mixture of DCM/hexane (40/60, v/v) was then carried out for the MIP/NIP 2, and finally the analytes were eluted with 1 mL of MeOH. The second washing and elution fractions were evaporated under an N2 stream and resuspended in 100 µL of MeOH/ACN/H2O (38/20/42, v/v/v) before injection into LC/MS-MS.

Extraction Procedure Applied to Urine Samples

Before applying urine samples on the MIP/NIP, samples were filtered using a 0.2-µm syringe filter (Millipore®, Merck, Ireland), and the urine samples were further spiked at 20 ng L−1 and then diluted five times with pure water before percolation. The diluted urine samples were percolated through a conventional sorbent, an Oasis HLB polymer (3 cc, 60 mg, Waters), and on the MIP 2/NIP 2. The procedure applied to the MIP2/NIP2 was identical to that described previously for spiked pure water. For the Oasis HLB sorbent, the extraction procedure was adapted from [19]. Briefly, the cartridge was conditioned with 2 mL of MeOH and 2 mL of water. After this conditioning step, 1 mL of the diluted spiked urine samples was percolated through the cartridge, followed by 2 mL of a water/MeOH (90/10, v/v) mixture as a washing step. The sorbent was dried for 20 min under a vacuum before elution of the target analytes with 1 mL of MeOH. The elution fractions for both sorbents (Oasis and MIP) were evaporated under an N2 stream and resuspended in 100 µL of MeOH/ACN/H2O (38/20/42, v/v/v) to be analyzed in LC/MS-MS.

The elution fraction was injected twice, first in MRM mode to quantify CBZ, OXC and epoCBZ in the fraction and to determine the recovery yields, and second in scan mode (m/z = 100–1100) in order to visualize the clean-up effect of the extraction on MIP when comparing with results obtained using Oasis HLB.

Evaluation of Matrix Effects During LC/MS-MS Analysis

Matrix effects were evaluated by comparing the slopes of the calibration curves constructed in pure water and in the elution fraction resulting from MIP 2 or from the Oasis HLB. For this, a blank urine sample was diluted five times with water, and 1 mL was percolated through the Oasis HLB cartridge or MIP 2. The resulting elution fraction for each support (1 mL for both supports) was divided into four equal parts and evaporated to dryness under an N2 stream. The residue was reconstituted in 20 µL of mobile phase [MeOH/ACN/H2O (38/20/42, v/v/v)] containing 0.1, 0.2, 0.5 and 2 pg of CBZ, epoCBZ and OXC. Each calibration point was analyzed in MRM mode for the quantification of CBZ, OXC and EPOCBZ.


Development of the LC/MS-MS Method

To ensure sensitive detection of CBZ and its two analogs epoCBZ and OXC (see Table 1 for structures and log P values) in biological samples, it was first necessary to optimize the MS detection parameters (MRM mode) in order to ensure a highly sensitive and specific method. The optimization was performed for each compound by the direct infusion of a standard solution in MS, and led to the choice of two transitions per compound; the most intense ion was used for quantification and the second most intense for confirmation. Taking into account the hydrophobicity of the analytes, and as previously reported for environmental water analysis [18], the LC separation was performed on a Varian C18 Omnispher column in isocratic mode. The limit of detection (LOD) and limit of quantification (LOQ), defined as concentration values producing signal-to-noise (S/N) ratios of 3 and 10, respectively, were calculated using the MRM chromatogram obtained for the lowest concentration injected and which gave an S/N ratio greater than or equal to 10 (concentration level of 5 ng L−1). The LOD and LOQ values are reported in Table 2, and range from 3 to 5 ng L−1 for the three analytes. The calibration curves were linear, with concentrations ranging from LOQ to 1 mg L−1 (see Table 2).

Table 1

Structure and physicochemical properties of target compounds

Table 2

Equation of calibration curves, LOD and LOQ in pure medium during LC/MS–MS analysis in MRM mode for CBZ, epoCBZ and OXC


Quantification transition

Calibration curve

LOD (ng L−1)

LOQ (ng L−1)


237 > 194

y = 987566 x, r2 = 0.9983




253 > 236

y = 18456 x, r2 = 0.9935




253 > 208

y = 1567 x, r2 = 0.9995



Repeatability of the MIP Synthesis and the Extraction Procedure in Pure Water

In a previous study from our group, several conditions for the synthesis of MIPs were screened by varying the nature of the monomers, cross-linker and porogenic solvent. Two MIPs enabled the selective extraction of CBZ from pure organic media with similar performance. These two MIPs, synthesized using TFMAA or MAA as monomer and DVB as cross-linker in a toluene/dichloromethane mixture, were then studied directly in greater detail in environmental water by optimizing the washing step for each to ensure high retention of CBZ on MIP and lower retention on NIP after percolation of 25 mL of spiked tap water. The MIP produced with MAA appeared slightly more selective towards CBZ than that produced with TFMAA. Indeed, the selectivity of these MIPs for CBZ, determined by comparing the ratios between the recoveries obtained on MIP and NIP for a given molecule, were 3.07 for the TFMAA-based MIP and 4.41 for the MAA-based MIP. Therefore, the latter MIP was applied for the extraction of CBZ, OXC and epoCBZ from mineral and surface water without real optimization of the extraction conditions for these two compounds. Nevertheless, high recoveries were obtained for the three compounds in a real environmental water sample (with recovery of 60–69% on the MIP and only 19–27% on the NIP) [18]. Here, the purpose was slightly different. Indeed, the extraction procedure had to be optimized in order to enable not only the selective extraction of CBZ, but also the extraction of two of its analogs, epoCBZ and OXC, from urine, which is a more complex matrix than the mineral and surface water previously studied and is available in smaller volumes than environmental water. These new objectives could mean that the MIP previously identified as best for the selective trapping of CBZ in environmental samples may not be the best for the trapping of CBZ and its analogs from urine samples. The two most promising MIPs previously synthesized using the MCBZ as template and TFMAA or MAA as monomer (MIP 1 and MIP 2, respectively) were then again synthesized as their corresponding NIPs. To assess their selectivity, the retention of the three target compounds in pure water was first studied by applying an extraction procedure very similar to that previously applied to surface water, but to a reduced volume of water (1 mL instead of 25 mL) and by again optimizing the washing conditions (different DCM/hexane ratios were assayed) to ensure high selectivity for the three targets. The highest recovery on the MIPs while maintaining low recovery on the NIPs was obtained using a DCM/hexane mixture of 70/30 (v/v), instead of 80/20 previously fixed for MIP 1 and 60/40 for MIP 2. As shown in Fig. 1, after percolation of pure water spiked with CBZ, epoCBZ and OXC at 5 µg mL−1, both MIPs provided (1) high retention not only for CBZ but also for epoCBZ and OXC, with high extraction recovery in the elution fraction of 79–82% for the three compounds, and (2) high selectivity, illustrated by extraction recovery below 20% on both corresponding NIPs. Figure 1 also illustrates the contribution of the washing solution on the removal of nonspecific interactions: more than 80% of the target analytes were lost from the NIP during this step. The high extraction recoveries in the elution fraction and the selectivity for CBZ observed on these two supports are in good agreement with those described previously for CBZ extracted from aqueous media [18], thus indicating the reliability of the synthesis of these MIP/NIPs as well, even if the evaluation conditions were not identical. High repeatability of the extraction procedure can also be observed for both MIPs, with relative standard deviation (RSD) values lower than 8% for recoveries in MIPs (n = 3).

Fig. 1

Extraction profiles obtained when percolating pure water spiked at 5 µg L−1 with CBZ, epoCBZ and OXC on both synthesized MIPs and NIPs. Extraction procedure: percolation of 1 mL spiked pure water, washing with 1.5 mL of a mixture DCM/Hexane, 70/30 (v/v) for MIP/NIP 1 and 60/40 (v/v) for MIP/NIP 2; elution with 1 mL MeOH. The gray and black bars correspond to the recoveries in the washing and elution fractions, respectively

The recoveries obtained using the two pairs of MIP/NIPs for CBZ, epoCBZ and OXC are very similar, thus rendering the selection of one or the other difficult. However, MIP 2 (synthesized with MAA as monomer) had previously shown the highest selectivity towards CBZ in real environmental samples [18], and slightly higher recovery for epoCBZ was observed, in conjunction with lower recovery on NIP 2 than on NIP 1, during this study. Thus MIP 2 was selected for the remainder of the study.

Evaluation of the Performance of MIP for the Extraction of CBZ and Its Analogs from Urine

After these promising results obtained for the selective extraction of the three target analytes from spiked pure water, the performance of MIP 2 was evaluated for the extraction of the target molecules from urine samples. The optimized extraction procedure was first applied to a non-spiked urine sample, and no target analytes were detected in the eluate. The procedure was then applied to this urine sample spiked at 20 ng L−1 with each compound. The extraction recoveries in the elution fraction corresponding to the MIP and the NIP are reported in the Table 3. The recoveries on the elution fraction of the MIP were higher than 82% for the three compounds and lower than 20% for the NIP, thus highlighting the selectivity of the extraction procedure on MIP applied to a urine sample. The RSD values (n = 3) describing the repeatability of the extraction procedure applied to urine samples were acceptable for this concentration level of 20 ng L−1 and in the range of 13–23%.

Table 3

Extraction recoveries (% ± RSD values, n = 3) of CBZ, epoCBZ and OXC from urine on the MIP/NIP 2 and on Oasis HLB






82 ± 23

20 ± 7

74 ± 11


87 ± 17


88 ± 30


106 ± 13

4 ± 10

31 ± 8

Extraction procedure: percolation of 1 mL of diluted urine sample (urine spiked at 20 ng L−1 and diluted by a factor of 5 with pure water); washing with 1.5 mL of a mixture DCM/hexane, 60/40 (v/v) for MIP/NIP2 or with 1.5 mL of H2O/MeOH, 90/10, v/v for Oasis HLB; elution with 1 mL MeOH

ND not detected

In order to highlight the potential of MIP 2, its performance in terms of recovery yield and cleaning capacity was compared to that obtained with a conventional Oasis HLB sorbent commonly applied in the extraction of drugs from biological samples. For this purpose, the same urine sample spiked at 20 ng L−1 was percolated on an Oasis HLB cartridge following a procedure previously reported [19]. The extraction recoveries obtained using Oasis HLB for CBZ, epoCBZ and OXC are also reported in Table 3. Recoveries for CBZ and epoCBZ obtained using Oasis HLB were similar to those obtained on MIP 2, with recoveries of 74 and 88%, respectively, using Oasis HLB, and 82 and 87%, respectively, using MIP 2. For OXC, recovery of only 31 ± 8% was obtained using Oasis HLB, compared with 106 ± 13% on MIP 2. As this compound is less polar than epoCBZ (see log P value, Table 1), it must be better retained than epoCBZ in this reversed-phase mode retention process. Therefore, a matrix effect during LC/MS quantification of this compound was suspected. The comparison of the LC/MS chromatogram in scan mode and in MRM mode resulting from the use of the MIP (Fig. 2a) and the Oasis HLB sorbent (Fig. 2b) for epoCBZ and OXC clearly illustrates that the elution fraction resulting from the Oasis HLB contained more interfering compounds than that of the MIP: (1) higher baseline signal and presence of many peaks in scan mode, and (2) an important noise in MRM mode, particularly visible for the transition of the OXC (Fig. 2b). The cleaner extract obtained using the MIP (Fig. 2a) should provide more repeatable quantification, as demonstrated by the lower RSD values observed for the quantification in the elution fraction of the target analytes on the MIP versus Oasis HLB, which reached 30% for epoCBZ (Table 3). The S/N ratio observed in MRM mode for CBZ, OXC and epoCBZ (191, 33 and 54, respectively) highlighted the ability of MIP 2 coupled with LC/MS-MS analysis to quantify the CBZ, OXC and epoCBZ at concentrations of only 1, 7 and 4 ng L−1 , respectively, using only 1 mL of diluted urine, i.e. 0.2 mL of urine, applied to the MIP.

Fig. 2

LC/MS analysis in scan mode and MRM mode of the elution fraction obtained after the percolation of urine spiked at 20 ng L−1 on MIP 2 (a) and on Oasis HLB (b). From top to bottom: MRM chromatograms (CBZ: 237 > 194, epoCBZ: 253 → 236 and OXC: 253 > 208) and LC/MS chromatograms in scan mode (m/z = 100–1100)

Evaluation of the Matrix Effects During LC/MS-MS Analysis

Potential matrix effects were evaluated in greater detail for epoCBZ and OXC by comparing the slopes of three different calibration curves. The first calibration curve, used as reference, was constructed by injecting solutions corresponding to the composition of the mobile phase spiked with epoCBZ and OXC. The second and third curves were constructed by injecting the elution fraction of a non-spiked and spiked (with different amounts of the target molecules) new urine sample into the LC/MS-MS system, after percolation on an Oasis HLB or on the MIP. The resulting calibration curves are reported in Fig. 3. The slope of the calibration curves for epoCBZ and OXC after applying the urine sample on the MIP are very close to those obtained for spiked pure media, thus indicating that the selectivity provided by MIP makes it an effective tool for matrix removal. On the other hand, the large difference between the slope obtained for spiked pure water and spiked extract of urine passed through the Oasis HLB highlights a strong matrix effect when analyzing this new urine sample, which could lead to overestimation of the amount of those two compounds in urine. Indeed, the signal was exhausted for both compounds (around 56 and 23% for epoCBZ and OXC, respectively). This quantification problem was easily circumvented by using the MIP.

Fig. 3

Calibration curves for epoCBZ and OXC obtained for spiked pure water (solid line), spiked extract of urine sample obtained after SPE on MIP (dotted line) or on Oasis HLB (dashed line)


After a first study focusing on the screening of synthesis conditions for MIP/NIP capable of selectively extracting CBZ from environmental samples [18], the two most promising MIPs were assayed for their ability to selectively extract the two drugs CBZ and OXC and the metabolite epoCBZ from human urine. High retention and high selectivity were demonstrated for the three structural analogs for a concentration of only a few ng L−1 in urine. The sample treatment with the MIP also demonstrated more efficient clean-up of human urine compared with a conventional Oasis HLB sorbent. It thus appears to be crucial to readjust (re-optimize) the extraction procedure applied to the MIP for each matrix. However, an application for the quantification of the three targeted compounds from other biological fluids such as plasma or saliva (noninvasive sampling), which is now considered an alternative for the determination of drug intake [20], could be considered in the future.



This work was supported by the French National Research Agency (ANR Program: ANR-15-CE04-0012, project MIP_WQT).

Compliance with Ethical Standards

Conflict of interest

The authors declare that they have no conflict of interest.

Research involving human participants and/or animals

Informed consent was obtained from all individual participants included in this study.


  1. 1.
    Brezina E, Prasse C, Meyer J et al (2017) Investigation and risk evaluation of the occurrence of carbamazepine, oxcarbazepine, their human metabolites and transformation products in the urban water cycle. Environ Pollut 225:261–269. CrossRefGoogle Scholar
  2. 2.
    Datar PA (2015) Quantitative bioanalytical and analytical method development of dibenzazepine derivative, carbamazepine: a review. J Pharm Anal 5:213–222. CrossRefGoogle Scholar
  3. 3.
    Fedorova G, Ben Ari J, Tadmor G et al (2016) Environmental exposure to pharmaceuticals: a new technique for trace analysis of carbamazepine and its metabolites in human urine. Environ Pollut 213:308–313. CrossRefGoogle Scholar
  4. 4.
    Chen S-Y, Wen-Chi C, Sarah C Y (2018) Cyclodextrin-assisted dispersive liquid–liquid microextraction for the preconcentration of carbamazepine and clobazam with subsequent sweeping micellar electrokinetic chromatography. J Sep Sci 41:1871–1879. CrossRefGoogle Scholar
  5. 5.
    Zhang R, Wang S, Yang Y et al (2018) Modification of polydopamine-coated Fe3O4 nanoparticles with multi-walled carbon nanotubes for magnetic-mu-dispersive solid-phase extraction of antiepileptic drugs in biological matrices. Anal Bioanal Chem 410:3779–3788. CrossRefGoogle Scholar
  6. 6.
    Behbahani M, Najafi F, Bagheri S et al (2013) Application of surfactant assisted dispersive liquid–liquid microextraction as an efficient sample treatment technique for preconcentration and trace detection of zonisamide and carbamazepine in urine and plasma samples. J Chromatogr A 1308:25–31. CrossRefGoogle Scholar
  7. 7.
    Beltran A, Caro E, Marce RM et al (2007) Synthesis and application of a carbamazepine-imprinted polymer for solid-phase extraction from urine and wastewater. Anal Chim Acta 597:6–11CrossRefGoogle Scholar
  8. 8.
    Beltran A, Marcé RM, Cormack PAG, Borrull F (2009) Synthesis by precipitation polymerisation of molecularly imprinted polymer microspheres for the selective extraction of carbamazepine and oxcarbazepine from human urine. J Chromatogr A 1216:2248–2253CrossRefGoogle Scholar
  9. 9.
    Dai C, Geissen S-U, Zhang Y et al (2010) Performance evaluation and application of molecularly imprinted polymer for separation of carbamazepine in aqueous solution. J Hazard Mater 184:156–163. CrossRefGoogle Scholar
  10. 10.
    Esfandyari-Manesh M, Mehran J, Fatemeh A et al (2011) Effect of porogenic solvent on the morphology, recognition and release properties of carbamazepine-molecularly imprinted polymer nanospheres. J Appl Polym Sci 121:1118–1126. CrossRefGoogle Scholar
  11. 11.
    Mehran Esfandyari-ManeshMJavanbakht, Fatemeh A, Dinarvand Rassoul (2012) Synthesis and evaluation of uniformly sized carbamazepine-imprinted microspheres and nanospheres prepared with different mole ratios of methacrylic acid to methyl methacrylate for analytical and biomedical applications. J Appl Polym Sci 125:1804–1813. CrossRefGoogle Scholar
  12. 12.
    Esfandyari-Manesh M, Javanbakht M, Dinarvand R, Atyabi F (2012) Molecularly imprinted nanoparticles prepared by miniemulsion polymerization as selective receptors and new carriers for the sustained release of carbamazepine. J Mater Sci Mater Med 23:963–972. CrossRefGoogle Scholar
  13. 13.
    Dai C, Zhang J, Zhang Y et al (2013) Removal of carbamazepine and clofibric acid from water using double templates–molecularly imprinted polymers. Environ Sci Pollut Res 20:5492–5501. CrossRefGoogle Scholar
  14. 14.
    Zhang Y-L, Zhang J, Dai C-M et al (2013) Sorption of carbamazepine from water by magnetic molecularly imprinted polymers based on chitosan-Fe3O4. Carbohydrate Polym 97:809–816. CrossRefGoogle Scholar
  15. 15.
    Lee J-C, Kim C-R, Byun H-S (2014) Synthesis and adsorption properties of carbamazepine imprinted polymer by dispersion polymerization in supercritical carbon dioxide. Korean J Chem Eng 31:2266–2273. CrossRefGoogle Scholar
  16. 16.
    Khalilian Faezeh A Setareh (2016) Molecularly imprinted polymer on a SiO2-coated graphene oxide surface for the fast and selective dispersive solid-phase extraction of carbamazepine from biological samples. J Sep Sci 39:1500–1508. CrossRefGoogle Scholar
  17. 17.
    Asgari S, Bagheri H, Es-haghi A, AminiTabrizi R (2017) An imprinted interpenetrating polymer network for microextraction in packed syringe of carbamazepine. J Chromatogr A 1491:1–8. CrossRefGoogle Scholar
  18. 18.
    Khadirvel P, Combès A, Bordron L, Pichon V (2018) Rational designing and development of water compatible molecularly imprinted polymers for the selective extraction of carbamazepine from environmental waters (submitted)Google Scholar
  19. 19.
    Fortuna A, Sousa J, Alves G et al (2010) Development and validation of an HPLC-UV method for the simultaneous quantification of carbamazepine, oxcarbazepine, eslicarbazepine acetate and their main metabolites in human plasma. Anal Bioanal Chem 397:1605–1615. CrossRefGoogle Scholar
  20. 20.
    Neumann J, Beck O, Dahmen N, Böttcher M (2018) Potential of oral fluid as a clinical specimen for compliance monitoring of psychopharmacotherapy. Ther Drug Monit 40:245. CrossRefGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany, part of Springer Nature 2019

Authors and Affiliations

  1. 1.Department of Analytical, Bioanalytical Sciences and Miniaturization (LSABM)CNRS (UMR CBI 8231), ESPCI Paris, PSL UniversityParis Cedex 05France
  2. 2.Sorbonne Université, UPMCParisFrance

Personalised recommendations