Design of packed-fiber solid-phase extraction device for analysis of the drug and its metabolite in plasma
A mini-column packed with 1 mg electrospun polystyrene nanofibers (about 200∼400 nm in diameter) was designed for simple, fast extraction of drugs, diazepam and its major metabolite, N-desmethyldiazepam for the analysis of them in human and dog plasma. Ttrezodone was selected as internal standard. The drugs adsorbed on the solid phase could be desorpted with 50 μl of the methanol and then monitored by liquid chromatography coupled to an ultraviolet detector. Parameters influencing the extraction efficiency such as fiber packing amount, eluted solvent, and pH of the sample were decided. The time for the pretreatment of 0.5 ml plasma sample was less than 10 min. The detection limits of diazepam and N-desmethyldiazepam in plasma could be as low as 1 μg/L. The intra- and inter-day precision, calculated from quality control (QC) samples, was less than 9.1%. The method was evaluated by its application in determination of dog plasma samples from three beagles after a single dose oral of diazepam. The technique was validated by comparison with conventional plasma analysis. It was observed that the mini-column offers improved limits of detection and reduced sample preparation time as compared to conventional method. For its simplicity and sensitivity, the method may be used in therapeutic drug monitoring and pharmacology study.
KeywordsPacked-fiber solid-phase extraction Preconcentration Nanofiber Diazepam HPLC-UV
In developing an analytical method used in therapeutic drug monitoring, it is important to determine how much time and effort is necessary for sample preparation. The general sample preparation method for detection of low concentration sample is based on the extraction technique (Kataoka 2005). The well-used liquid–liquid extraction technique (LLE) is a multi-stage operation which is both time consuming and labor-intensive. Each step, especially the concentration, can introduce losses of volatile compounds, resulting in a decrease of the sensitivity. Waste disposal of solvents is another problem, adding extra cost and pollution to the environment. Solid-phase extraction (SPE) cartridges partially solved the problems of the classical LLE methods. However, the operation of the conventional SPE is still multistep, and a concentration via solvent evaporation is often necessary.
Recent developments in sample preparation made the sample preparation more convenient and more effective. These newly appeared methods include solid-phase microextraction (SPME; Musteata and Pawliszyn 2007) and related techniques, including in-tube SPME (or capillary microextraction; Eisert and Pawliszyn 1997), solid-phase dynamic extraction (SPDE; Lipinski 2000), microextraction in a packed syringe (MEPS; Abdel-Rehim 2004), stir-bar-sorptive extraction (SBSE; Stopforth et al. 2007), and fiber-in-tube solid-phase extraction (FIT-SPE; Jinno et al. 2007). Commonly, these sample cleanup techniques base on micron scale factures such as the membranes coated to out surface of the fiber (micrometer scale in diameter) or inner wall of the capillary, the particles (micrometer size) or fibers (micrometer scale in diameter) packed into the column or capillary. One potential technique for SPE to increase the efficiency of the solid phase is use of substances in nanometer order as sorbents. Although the submicrometer-sized particles can greatly increase the surface area, impractically high pressures is need for the liquid flow, which limits the application. However, the problem was solved by use of nanofibers as the sorbents in our previous paper (Kang et al. 2007). The novel method is called packed-fiber solid-phase extraction (PFSPE).
The nanofibers have diameters in the nanometer range and are arbitrarily long. If they were used as sorbents, the column pressure would be lower than the column packed with nanometer particles. In addition, the nanofibers possess a high aspect ratio that leads to a larger specific surface. The effective interaction of the target compounds with a number of fine nanofibers could reduce the volume of both solid phase and desorption solvent greatly. The evaporation of solvent and reconstitution of target analytes could be avoided because the sample preparation method delivered the target analytes in a sufficiently small volume of solvent, suitable for direct injection onto an analytical instrument. Therefore, PFSPE should offer advantages over conventional LLE and SPE, simplicity, low cost, and very low solvent consumption. PFSPE should be more sensitive than SPME, because it is a quantitative extraction in which the entire solution is passed through a sorbent cartridge for compound uptake, whereas SPME is an equilibrium technique, which merely samples the solution.
The purpose of the present work was to investigate the applicability of PFSPE in the plasma analysis of drug and its metabolite. Diazepam and one of its principal metabolite, N-desmethyldiazepam were selected as the model drugs. Diazepam is frequently used in clinical practice such as tranquilizer, sleep inducer, antiepileptic hypnotic, anticonvulsant, and muscle relaxant. N-desmethyldiazepam is also pharmacologically active. These drugs, however, can cause intoxication due to accidental overdosage or intentional abuse. The extensive use and potential abuse of this class of compound demands an accurate and rapid method for analysis (Mullett and Pawliszyn 2001). There were numerous papers on extraction of diazepam for biological samples prior to chromatographic-determination, such as LLE (Atta-Politou et al. 1999; Kinani et al. 2007), SPE (Soriano et al. 2001; Liu et al. 2001; He et al. 2005) and SPME (Reubsaet et al. 1998; Musteata et al. 2006). Except for the SPME methods, other extraction procedures possessed evaporation step which should last at least ten minutes to concentrate the target compounds, therefore the total time required to prepare a single sample for them was longer than our method (10 min). In addition, most methods were coupled to GC-MS or LC-MS to enhance sensitivity and selectivity.
In this paper, a PFSPE coupled to HPLC was evaluated by monitoring two target compounds in spiked human and dog plasma samples. The technique was validated by comparison with conventional plasma analysis (a liquid–liquid extraction method; Atta-Politou et al. 1999).
2.1 Chemicals and reagents
Diazepam (DZ) was obtained from the National Instituite for the Control of Pharmaceutical and Biological Products (Beijing, China). N-desmethyldiazepam (DMDZ), DZ tablet (2.5 mg) and trezodone (IS) were kindly supplied by China Pharmaceutical University. LC-grade methanol was from Jiangsu Huaiyin Fine Chemical Research Institute (Huaian, China). Polystyrene (PS, Mw = 185,000) and other reagents (analytical reagent grade) were from Shanghai Chemical Agents Institute (Shanghai, China). Double distilled water was used throughout the experiments. The pipette tip (200 μl) was from Yonghua glasswork (Haimen, China). Heparinized blank (drug free) human plasma was provided by Nanjing Blood Donor Service (Nanjing, China) and was stored at −20°C. Heparinized blank (drug free) dog plasma and heparin-coated tubes were from Southeast University Laboratory Animal Center.
2.2 Standards and calibration curves
Standard stock solutions (1 g/L) of drugs were prepared by dissolving an appropriate amount of the compounds in methanol. Working solutions were prepared daily by an appropriate dilution of the stock solutions with water. The internal standard solution was prepared by dilution of standard stock solutions to the working concentration (20 mg/L) with water. All solutions were stored at 4°C in the dark before using. The stock solutions were found to be stable for 3 weeks. The mobile phase for determination of DZ was composed of aqueous mobile phase (20 mM sodium acetate—methanol, 40:60, v/v) at pH 6.0 adjusted with 2 M sodium hydroxide and acetic acid.
Calibrators were prepared by adding 10 μl of the appropriate working solutions of DZ and DMDZ into 0.5 ml of drug-free dog plasma containing 10 μl of IS solution. The drug concentrations were 1, 5, 20, 100, 500, and 1,000 μg/L. QC samples of DZ and DMDZ used in the validation were prepared in the same way as the calibrators (three different concentrations, 10, 200, and 800 μg/L).
The chromatographic system consisted of a Shimadzu LC-10A pump and a Shimadzu SPD10-Avp UV detector (Kyoto, Japan). Sample injection was performed via a Rheodyne 7125 injection valve (Rheodyne, Cotati, California, USA) with a 20 μL loop. A Tailihua TL9000 HPLC software package (Beijing, China) was used for the data analysis. An Eka Chemicals AB C18 column (Bohus, Sweden) was used for HPLC separation (particle size 5 μm; column dimensions 150 × 4.6 mm). The wavelength used for detection was 230 nm. The HPLC flow rate was 1 ml/min. Samples and standards were injected duplicated. The injection volume was 20 μL. All the measurement operations were performed at room temperature.
2.4 Design and preparation of packed-fiber solid-phase extraction (PFSPE) column
2.5 Solid-phase extraction procedure
Solid-phase extraction of drugs from plasma samples was performed as follows: The columns were conditioned by washing with 50 μl of ethanol and 50 μl of water. A 0.5 ml of spiked plasma sample was added to a 1.5 ml eppendorf centrifuge tube containing 0.1 ml of 40% perchloric acid and IS (10 μl, 20 mg/L in water). The mixture was centrifuged at 10,000×g for 5 min. Because the drugs have pKa values (3.3 DZ, 3.5 DMDZ), so the weak alkaline or neutral sample medium exhibits positive influence in the extraction. Therefore, the supernatants were adjusted to neutral with 2 M sodium hydroxide (about 70 μl added), and then the mixtures were pumped through the columns using a 12-ports SPE Vacuum Manifold (Beijing, China). The columns were then washed with 50 μl of water. Drugs were quantitatively eluted with 50 μl of methanol by the air pressure forced by a gas tight plastic syringe (2 ml), which was fitted to the pipette tip. Finally, 20 μl of elution was analyzed by high performance liquid chromatography.
2.6 Data analysis
For determination of DZ and DMDZ, peak area ratios (drug/IS) were plotted as a function of the known amounts of analytes, and the results analysed by least-squares linear regression. Analyzing QC samples at the above-mentioned three concentrations assessed both intra- and inter-assay imprecision. For intra-assay studies, replicate analyses (n = 6) of each QC sample were performed on the same day. For inter-assay studies, each of the three QC samples were analyzed once a day on different days (n = 6). Imprecision is given as the relative standard deviation (RSDs of the concentrations calculated). The extraction efficiency was expressed as extraction recoveries which were calculated by comparing the peak areas from extracted samples with those obtained from a direct injection of the corresponding unextracted standards dissolved in methanol. The extraction recoveries of DZ and DMDZ by use of QC samples at the above-mentioned three concentrations (n = 3) were determined. The limit of detection (LOD) and limit of quantification (LOQ) for each compound in plasma were determined from six runs of procedure plasma blanks.
2.7 Conventional plasma analysis
For comparative studies, the classical liquid-liquid phase extraction was selected as references. Plasma was prepared and analyzed as described previously (Atta-Politou et al. 1999), aside from the internal standard (trezodone, 10 μl, 20 mg/L in water) and 0.5 ml of sample. The linear range was 10–500 μg/L.
2.8 Sampling in pharmacokinetic study
The method was used to quantify DZ and DMDZ in beagle plasma after oral administration of 0.5 mg/kg of DZ tablet. The animal study was approved by the local Animal Use Committee at Medical School of Southeast University. Blood samples were collected from three male beagles weighing 9–10 kg. The dogs were feed with no food all night before administration of the drug. Plasma samples were collected prior to dosing and at 10, 20, 30 min, 1, 2, 3, 4, 6, and 8 h post dose. About 3.0 ml blood was taken from dog forelimb vein. The blood was processed into Heparin-coated tubes and centrifuged at 4,000 rpm for 10 min at room temperature to obtain plasma (supernatant). The supernatants were stored at –20°C before using.
3 Results and discussion
3.1 Optimization of extraction
The distinctive design in our method was use of mini-column with nanofibers as sorbents. Saito et al. have validated the applicability of fibers with microscale structures as sorbents. It has been demonstrated that the fiber-packed SPE device offers a reduced pressure drop during the extraction and desorption compared to a conventional particle-packed SPE cartridge (2002, 2004; Imaizumi et al. 2003). In order to investigate the factors affecting extraction efficiency, optimization experiments had been carried out in the same way as that in our previous paper (Kang et al. 2007). Fiber diameter, packing quantity, medium pH, and desorption solution were decided as 200∼400 nm, 1 mg, pH 6.0, 50 μl of methanol, respectively.
3.2 Analytical performance
Results o f DZ and DMDZ exraction recovery, accuracy, intra- and inter-day precision in dog plasma
Extraction recovery (%) (n = 6)
Intraday accuracy (%) (n = 6)
Interday accuracy (%) (n = 6)
Intraday CV (%) (n = 6)
Interday CV (%) (n = 6)
83.2 ± 9.3
96.7 ± 8.2
95.8 ± 8.9
56.3 ± 3.1
99.5 ± 3.2
98.7 ± 4.0
51.7 ± 2.6
101.1 ± 2.2
100.4 ± 2.5
82.1 ± 7.8
96.2 ± 8.7
103.6 ± 11.3
61.5 ± 2.9
98.8 ± 2.6
97.9 ± 3.2
57.2 ± 2.2
99.3 ± 2.3
98.5 ± 2.7
3.3 Real sample analysis
3.4 Comparison to liquid–liquid extraction
Since liquid–liquid extraction is a well-established method for the determination of drugs in plasma before SPE is introduced into use in drugs analysis, we decided to compare the performance of the PFSPE (1 mg of fibers packed) method developed here with that of LLE, with which lower limit of detection was 1 μg/L, and the lower limit of quantification was 2 μg/L for both compounds. In addition to the better clearning effect of PFSPE apparently shown in chromatograms, PFSPE possesses the advantage of speed, only 10 min needed for adsorption and desorption in stead of spending more times (at least 30 min) in LLE. This makes PFSPE practical when large-scale human studies are required which involve the analysis of a great number of specimens in clinical studies.
Although both two methods possessed equivalent LOD, when using LLE to determine DZ plasma concentrations, most of data in elimination phase had been lost because of insufficient sensitivity of LLE method at low DZ plasma concentration level. We attribute poor sensitivity of LLE at low concentration level to its multi-step operation such as evaporation and reconstituting.
As a result only PFSPE performance (y) for analysis of DMDZ was evaluated by comparing with the LLE method (x) utilizing plasma concentration of three dog plasma samples. The regression equation, correlation coefficient and standard error for the comparison were \(y = 1.1269x + 2.6203\), γ = 0.988 and Sy/x = 11.2 μg/L respectively. The results seem to show that the developed method is suitable for pharmacokinetic studies of drug and its metabolite in dogs.
3.5 Comparison to other methods
Comparison between the reported methods and the proposed method
Exhausted solvent (ml)
Sample volume (ml)
Linear range (μg/L)
Extraction recovery (%)
Evaporating solvent and reconstituting residue
In contrast to other methods, the volume of organic solvent used in PFSPE is not more than 100 μL, which obviously illustrates the environmental friendliness. It is apparent that with the exception of a few methods, PFSPE is the more sensitive technique for the determination of analytes, and it needs less sample volume and simpler equipment than those of methods. The advantages thus conferred by PFSPE allow for decreased operation procedures, along with concomitant decreases in both the expense and the amount of waste generated. Therefore, it could be conclude that PFSPE is a viable technique for the determination of analytes in complex samples.
In conclusion, a new mini-column based on electrospun nanofibers was developed and was successfully used for the extraction of drugs. The overall ease with which PFSPE was used to accurately quantify DZ and DMDZ in plasma clearly demonstrates that PFSPE is a viable technique for overcoming the matrix effects that plague LC analyses of plasma samples. Thus, PFSPE could serve as a valuable tool to provide accurate knowledge of blood drugs levels. The distinct advantages of this pretreatment format over traditional method are in terms of simplicity, rapid speed, low cost and environmental benignity. The development of new, efficient, and automated sample cleanup techniques based on functional nanofibers will be very useful in pharmaceutical and biomedical analysis.
This study was supported in part by the grants from Jiangsu Science and Technology Department (Grant No. BG2007044), and in part by a grant from “111” Project (B08024). We thank Professor Juan Li from China Pharmaceutical University for the supply of DZ tablets and DMDZ standard in this work.