Development of hollow fiber-supported liquid-phase microextraction and HPLC-DAD method for the determination of pyrethroid metabolites in human and rat urine.

Research article Received: 28 May 2013,

Revised: 24 October 2013,

Accepted: 31 October 2013

Published online in Wiley Online Library: 17 December 2013

(wileyonlinelibrary.com) DOI 10.1002/bmc.3097

Development of hollow fiber-supported liquidphase microextraction and HPLC-DAD method for the determination of pyrethroid metabolites in human and rat urine Wielgomas Bartosz*, Wiśniewski Marcin and Czarnowski Wojciech ABSTRACT: A simple hollow fiber liquid-phase microextraction method for the determination of synthetic pyrethroid metabolites, 3-phenoxybenzoic acid and 4-hydroxy-3-phenoxybenzoic acid, in human and rat urine was developed and validated. A polypropylene hollow fiber tightly fitted onto a Nylon rod and impregnated with organic solvent served as a disposable extraction device. Desorption of analytes was carried out in NaOH solution, analyzed further by gradient HPLC and diode array detection method. Important factors were identified using Taguchi OA16 (45) orthogonal array design and further optimized using univariate approach. The optimum method performance was observed when 1 mL of urine hydrolyzed with 0.2 mL of concentrated HCl was further supplemented with 100 mg of NaCl and extracted for 120 min into dihexyl ether immobilized in the pores of the hollow fiber. Metabolites were desorbed into 0.1 mL of 0.1 M NaOH for another 120 min. Limits of detection and quantitation of 15 and 50 ng/mL were obtained for both analytes. Relative standard deviations of 1.6–12.6% over the linear range (50–10,000 ng/mL, r > 0.9906) were observed. Intra- and inter-day accuracies of the method ranged from 98.3 to 109.5% and from 93.3 to 110.9%, respectively. The optimized method was applied to the analysis of real urine samples collected from rats exposed orally to cypermethrin. Copyright © 2013 John Wiley & Sons, Ltd. Keywords: hollow fiber; HPLC; liquid-phase microextraction; pyrethroids; urine

Introduction

708

Synthetic pyrethroids are actually one of the most common insecticide groups, used all over the world in agriculture, veterinary medicine, households and public health. Human exposure to those compounds has been revealed in numerous results of human biomonitoring studies. Urinary metabolites of pyrethroids serving as biomarkers of exposure are detected at high rates in the general population around the world (Fortin et al., 2008; Sams and Jones, 2012; Heudorf and Angerer, 2001; Riederer et al., 2008; Barr et al., 2010; Kimata et al., 2009a, 2009b; Feo et al., 2012; Wielgomas et al., 2013). Wide use of pyrethroids is based on a high insecticidal activity, relatively low toxicity to humans and also rapid degradation in the environment. The highest exposure to pesticides is attributed to applicators, and urinary levels of metabolites are usually one order of magnitude higher than in nonoccupationally exposed individuals (Kimata et al., 2009a, 2009b). The general population is mostly exposed to pyrethroids used for indoor spraying to eradicate adult mosquitoes or bed bugs. Other exposure sources of minor importance are residues found in food, mainly of plant origin, and also contaminated air, water and house dust (Quirós-Alcalá et al., 2011; Starr et al., 2008). Pyrethroids, when adsorbed via dermal, pulmonary or oral routes, are transported in the bloodstream and undergo oxidative and hydrolytic reactions to produce polar metabolites which are easily eliminated with urine in conjugated forms, mainly glucuronides and sulfates (Eadsforth et al., 1988; Eadsforth and Baldwin, 1983; Wielgomas and Krechniak, 2007). Owing to the effective metabolism in laboratory animals and humans, it is exceptional to detect the parent compound in blood even after high

Biomed. Chromatogr. 2014; 28: 708–716

exposure, thus for biomonitoring purposes it is advisable to measure metabolite concentration in urine. Half-lives for pyrethroid metabolites in urine are longer than for blood and also sampling of urine is simpler and less invasive. Several analytical techniques including ELISA (Chuang et al., 2011; Kim et al., 2009) and liquid (Baker et al., 2004) and gas chromatography (Schettgen et al., 2002; Leng and Gries, 2005; Aprea et al., 1997; Arrebola et al., 1999) have been employed for the determination of various metabolites of pyrethroids in biological samples. Usually the total concentration is measured (free plus conjugated), thus a hydrolysis step is required to liberate metabolites from their conjugates. In addition to the acidic hydrolysis, enzyme-catalyzed hydrolysis can be applied. However the latter one needs longer incubation time (usually overnight) and is more expensive than acidic hydrolysis. Both glucuronidase and sulfatase are needed to obtain quantitative hydrolysis. Isolation of the analytes is usually carried out to concentrate target compounds (extraction) with simultaneous elimination of matrix components

* Correspondence to: Wielgomas Bartosz, Department of Toxicology, Medical University of Gdańsk, Al. Gen. J. Hallera 107, 80-416 Gdańsk, Poland. Email: [email protected] Department of Toxicology, Medical University of Gdańsk, Gdańsk, Poland Abbreviations used: 2PBA, 2-phenoxybenzoic acid; 3PBA, 3-phenoxybenzoic acid; 4OH3PBA, 4-hydroxy-3-phenoxybenzoic acid; ACN, acetonitrile; DAD, diode array detection; DHE, dihexyl ether; HF-LPME, hollow fiber liquid-phase microextraction; OCT, 1-octanol; TOPO, trioctylphosphine oxide.

Copyright © 2013 John Wiley & Sons, Ltd.

HF-HPLC-DAD method for pyrethroid metabolites in urine (cleanup). This can be done in several ways by employing various sample preparation techniques: liquid–liquid extraction, solidphase extraction and, now gaining popularity, microextraction techniques, where no or almost no organic solvent is used. Among several microextraction concepts presented in the literature, membrane-assisted techniques are especially suitable for extraction of ionizable compounds from complicated matrices. Hollow fibers are used in a variety of setups with two being the most popular: two- and three-phase hollow fiber liquidphase microextraction (HF-LPME). In the first one, extraction solvent fills the pores of the fiber and also its lumen. In the three-phase mode, only the pores of the fiber hold the organic solvent while the lumen is filled with aqueous receiving solution (base or acid depending on the analyte properties). The pH gradient is a driving force for diffusion of analytes through the liquid membrane; therefore it is very effective for extraction and simultaneous cleanup of ionizable analytes from interference-rich matrices. While in a three-phase configuration the extraction and desorption processes occur simultaneously, in the setup proposed in this paper those two processes are separated in time. In the first step an impregnated membrane precisely fitted on the Nylon stick (rod) is placed in the sample vial for the predetermined time. When the extraction is finished the rod is removed from the sample, briefly washed with water and placed in a receiving solution for desorption. Using this concept, high cleanup efficiency is facilitated but the preparation of the fibers and handling is much simpler than in a classical three-phase hollow fiber microextraction. In this work we propose a simple and effective extraction technique for isolation of pyrethroid metabolites from rat and human urine following gradient HPLC with diode array detection (DAD) detection. The method was validated and its suitability was proven on real rat urine samples as well as on spiked human urine samples.

Experimental

and a UVD-340 diode array detector were controlled through a UCI-100 chromatography interface and a PC computer on which Chromeleon version 6.8 (Dionex) was installed for data management. Separation of analytes was achieved on a TosoHaas TSKgel Super ODS HPLC column, 4.6 mm × 10 cm, 2 μm (Montgomeryville, PA, USA) kept at 40 °C. The mobile phase consisted of 30 mM phosphate buffer pH 3.5 (component A) and gradient-grade ACN (component B). A linear gradient elution was programmed from 10 to 40% B in 15 min, at a flow rate of 1.0 mL/min then changed back to 10% and held for 5 min before the next injection. Detection was performed at 210 nm for all analytes. A 20 μL aliquot of the final extract was injected into HPLC system.

Urine samples Pooled blank urine was obtained from six nonoccupationally exposed individuals (3PBA background level in general population in Poland is below 0.3 ng/mL; Wielgomas et al., 2013). Blank urine spiked with both analytes and the internal standard (2PBA) at a level of 1 μg/mL was used for optimization experiments. Both peak heights and background interferences were evaluated to achieve the highest signal-to-noise (S/N) ratio, since high abundance of interferences could decrease the S/N ratio in the low-concentration samples. Rat urine samples from animals exposed orally to α-cypermethrin were used for the comparison of acidic and enzymatic hydrolyses. For acidic hydrolysis, 1 mL of fortified urine was pipetted to a 4 mL glass vial and 0.2 mL of concentrated HCl was added; the vial was capped and heated for 90 min at 95 °C. Afterwards samples were cooled down to room temperature, centrifuged to remove any precipitate and subjected to extraction. For enzymatic hydrolysis 1 mL of fortified urine was placed in a 4 mL glass vial and 0.5 mL of acetate buffer (0.4 M, pH 5.0) and 0.01 mL of β-glucuronidase were added. After mixing the samples were placed in the heating block (37 °C) for 17 h. On the following day the samples after glucuronidase treatment were acidified with 0.04 mL of concentrated HCl and further subjected to microextraction. A preweighted amount of NaCl was added to the sample before extraction.

Microextraction setup

Chemicals All chemicals were of analytical grade or higher. Dihexyl ether (DHE), 1-octanol (OCT), trioctylphosphine oxide (TOPO) and 2-phenoxybenzoic acid (2PBA) were purchased from Fluka–Sigma–Aldrich (Buchs, Switzerland). 4-Hydroxy-3-phenoxybenzoic acid (4OH3PBA) was from Roussell-Uclaf (Paris, France) and 3-phenoxybenzoic acid (3PBA) was from Lancaster (Eastgate, England). Acetonitrile (ACN) gradient-grade far-UV was from P.O.CH. (Gliwice, Poland). Stock solutions were prepared by dissolving pure substance in ACN to give a concentration of 1 mg/mL and were stored protected from light at 20 °C. Further dilutions in water were prepared daily from stock solutions.

After inserting the rod into the vials containing hydrolyzed urine the vial opening was sealed with Parafilm so the rod was secured to prevent its movement during shaking. Vials were placed on an orbital shaker for desired time and then the rods were removed, washed shortly with water and transferred to 0.1 mL of NaOH solution in glass autosampler microinserts (0.2 mL nominal volume; Fig. 1B). After a predetermined time the rods were removed from the inserts and the samples were ready for HPLC analysis.

Hollow fiber membranes Hollow fibers made from polypropylene from MTB Technologies (Warsaw, Poland) were used throughout the study. The inner diameter of the fiber was 1.67 mm and outer was 2.71 mm; average pore size was 0.33 μm. The fibers were supplied in 1 m-long tubes and were then cut into 1 cm-long pieces. After washing with acetone in an ultrasonic bath and drying at 60 °C, the fibers were stored in a closed glass jar protected from light. Before use the fibers were fitted on 4 cm-long Nylon rods (we adopted brush cutter Nylon line).

HPLC apparatus and conditions


Figure 1. HF-LPME setup with membrane on the rod device during extraction (A) and desorption (B).


wileyonlinelibrary.com/journal/bmc

709

Chromatography was performed using a Dionex HPLC system. A gradient P-580 pump, STH-585 column thermostat, ASI-100 autosampler

B. Wielgomas et al. Screening design

Desorption and extraction time 5

In the current work, OA16(4 ) orthogonal experimental design based on Taguchi tables was applied to screen for the most significant factors affecting extraction of metabolites. Five factors supposed to affect extraction efficiency were chosen for the screening process: extraction time (30–180 min), molarity of the acceptor (NaOH, 0.01–0.5 M), sample pH (from unadjusted after acidic hydrolysis to 4), impregnation solvent (OCT, DHE alone and with 5% TOPO addition) and salt concentration (NaCl, 50–300 mg/mL). Lord and Pawliszyn (2000) suggested the addition of inorganic salt to urine samples subjected to microextraction to minimize the influence of urine dilution. For that reason concentrations of NaCl in samples were tested in the range given above. The presence of particles in the solution during extraction may affect the boundary between phases, so oversaturated solution of NaCl was not tested. The lowest level of pH range was equal to pH of the hydrolyzed urine (1 mL of urine and 0.2 mL HCl) and the highest level was close to the calculated pKa value of analytes (3PBA, pKa = 3.92; http://www.vcclab.org/). Four levels were assigned to each factor. In the screening process desorption time was set to 180 min, equal to the longest time of extraction to assure equilibrium between phases. It is also known that sample agitation significantly increases mass transfer during extraction. Thus the highest speed was set, which is dependent on the construction and capabilities of available orbital shakers. In our case, the platform with an attached sample tray was rotated at 300 rpm. Sixteen trials were performed in triplicate and average sum of peak areas was used as a measure of analytical response. Design and results are presented in Table 1. ANOVA calculations were performed using Statistica version 10 from StatSoft (Tulsa, OK, USA).

Several fortified (2 μg/mL 3PBA and 4OH3PBA) urine samples were extracted using the procedure described earlier (DHE as an extraction solvent, pH unadjusted, NaCl 100 mg/mL, 120 min extraction time) and the analytes were further desorbed to 100 μL of 0.1 M NaOH for a different length of time (30–180 min) to find an optimum duration of the desorption process. In the next experiment optimum desorption time was kept constant while extraction time was tested in the range 30–180 min.

Validation Analytical figures of merit were evaluated in optimized conditions. Validation was performed following the European Medicines Agency (2011) guidelines for both human and rat matrix. The selectivity of the method was evaluated by analyzing human and rat urine (six independent samples in both cases analyzed separately) and visual assessment of the baseline at the retention time of analytes. The carryover effect of the autosampler was evaluated by three consecutive injections of the upper limit of quantitation sample and extracted blank urine sample. The autosampler was programmed to wash the syringe twice (with acetonitrile) after each injection. An eight point matrix matched calibration curve was prepared by spiking a blank urine with known amounts of analytes in the range of 50–10,000 ng/mL and then subjected to optimized microextraction to assess linearity of the method. Quality control samples (QC) were prepared in a similar manner but using independently prepared set of stock solutions. Four levels od QC samples were used to cover linear range of the method: lowest limit of quantitation (LLOQ, 50 ng/mL), low QC (150 ng/mL), medium QC (5000 ng/mL) and high QC (7500 ng/mL).

Table 1. Taguchi OA16(45) design with factor assignment and results Trial A 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 r1 r2 r3 r4 d

Responsea

Factor

30 30 30 30 60 60 60 60 120 120 120 120 180 180 180 180 86.7 122.1 447.4c 278.3 360.7

B 0.01 0.1 0.2 0.5 0.01 0.1 0.2 0.5 0.01 0.1 0.2 0.5 0.01 0.1 0.2 0.5 62.3 180.8 231.6 459.8 397.5

C 1b 2 3 4 2 1b 4 3 3 4 1b 2 4 3 2 1b 408.9 160.9 139.8 224.9 269.1

D OCT OCT/TOPO DHE DHE/TOPO DHE DHE/TOPO OCT OCT/TOPO DHE/TOPO DHE OCT/TOPO OCT OCT/TOPO OCT DHE/TOPO DHE 192 186.2 458.9 97.5 361.4

E 50 100 200 300 300 200 100 50 100 50 300 200 200 300 50 100 186.2 405.9 184 158.4 247.5

0.4 21.5 41.6 23.2 7.07 14.7 60.5 39.9 54.6 140.9 124.5 127.4 0.3 3.7 5.0 269.3

710

A, Extraction time (min); B, molarity of acceptor (NaOH); C, sample pH; D, impregnation solvent; E, salt concentration, NaCl (mg/ mL). Desorption time 180 min kept constant, equal to the longest extraction time. a Sum of peak areas of analytes. Average of three replicates for each trial. b Unadjusted pH of the sample after acidic hydrolysis – nonmeasurable using a glass electrode. c The highest response for each factor is in bold. OCT, 1-octanol; TOPO, trioctylphosphine oxide; DHE, dihexyl ether.




HF-HPLC-DAD method for pyrethroid metabolites in urine


DHE/TOPO

3PBA 2PBA 4OH3PBA 3PBA 2PBA 4OH3PBA 3PBA 2PBA 4OH3PBA 3PBA 2PBA 4OH3PBA 0%

25%

50%

75%

100%

Figure 2. Quantitative distribution of analytes in different compartments of HF-LPME setup after 120 min extraction and 120 min desorption to 0.1 mL of 0.1 M NaOH. Light grey, sample; transparent, liquid membrane; dark grey, final extract. The initial concentration of all analytes in the sample was 2 μg/mL.



711

Effort was made in this work to simplify and increase the sample throughput of the hollow fiber-based liquid microextraction. Here we propose a simple, disposable device consisting of a 1 cm-long tubular hollow fiber, tightly fitted on a Nylon rod and then impregnated with an organic solvent. The impregnated fiber is immersed in the vial containing hydrolyzed urine. Most of the hollow fiber-based microextraction methods use three-phase mode with an organic solvent-impregnated membrane separating the sample from the acceptor phase placed in the lumen of the membrane. This approach is effective but the process of preparation of fibers for extraction and manual removal of the acceptor from the inner part after extraction requires some practice. In this work we used a hollow fiber fitted on a Nylon rod as a support for organic solvent in which the membrane was soaked for a few seconds; after rinsing with water the rod with the membrane was placed in a glass vial containing hydrolyzed urine (Fig. 1A). One of the main advantages of the proposed method is that one analyst can prepare a large batch of samples in a relatively short time and the sample extraction/desorption process can be performed almost unattended. To additionally raise the sample throughput, mixing of the samples comprised the use of an orbital platform shaker. The vials were placed in a plastic tray with 60 wells precisely fitted to vials. A tray was mounted on the platform and then the whole tray was shaken for a predetermined time. In this work we used, at the same time, two plastic trays with 60 places each, but it could be modified depending on the actual needs. For example, using this system, one person can manage 120 samples in less than 8 h, including the time required for hydrolysis and the preparation of fibers. The number of simultaneously extracted samples is virtually unlimited because the fibers can be prepared in advance. No significant solvent losses were observed for up to 24 h (weight of the fibers impregnated with DHE and stored in a closed glass vessel at +4 °C). Moreover, extraction efficiency remained unchanged for freshly and fibers prepared in advance.

DHE

Setup

When more than one factor affects any process that has to be optimized, one should consider using a design of experiments approach, which has advantages over the one-factor-at-a-time protocol (Roy, 2001). Several multivariate designs were proposed and applied for microextraction optimization: Taguchi orthogonal arrays (Saleh et al., 2009), central composite designs (Ebrahimzadeh et al., 2011), Box–Boehnken (Martendal et al., 2007; Batlle et al., 2008) and Plackett–Burmann (Vera-Candioti et al., 2008), thoroughly reviewed by Stalikas et al. (2009). In this work Taguchi orthogonal design was used to evaluate the influence of various factors on extraction efficiency. Since five variables were selected for this study (extraction time, molarity of the acceptor, sample pH, impregnation solvent and salt concentration), the OA16(45) design was chosen. With only 16 trials, five variables at four levels were studied and analyzed. All trials were performed in triplicate and a response, the sum of the peak areas of all analytes (including internal standard), was calculated. The experimental design matrix and results are presented in Table 1. We preselected two of the most often used organic solvents for hollow fiber-based microextraction: OCT and DHE. Both exhibit low vapor pressure, which is desirable for preparation of the fibers in advance (negligible evaporation of the solvent during the storage of impregnated fibers). The solubility in water is also minimal. For ionizable compounds it was shown that addition of TOPO significantly increases extraction efficiency; however, in this study 5% TOPO in the solvent reduced extraction efficiency in combination with both solvents. The distribution of analytes was studied by measuring concentration after 120 min extraction and 120 min desorption in donor and acceptor solutions. The mass of the analytes trapped in membrane solvent was calculated as the difference between initial concentration in the sample and determined concentration in the final extract. The results are presented in Fig. 2. DHE exhibited lower than OCT capability in extraction of analytes from the sample solution but simultaneously desorption of analytes from DHE to NaOH was more efficient. Addition of TOPO to both solvents (DHE and OCT) resulted in higher extraction rates from the sample solution but simultaneously reduced the efficiency of the process of desorption. In general

OCT/TOPO

Results and discussion

Experimental design

OCT

The concentration of the internal standard (2PBA) in the calibration samples was kept constant at 2 μg/mL. The calibration curve was plotted based on analyte to internal standard relationship. Intra- and interday accuracy and precision were assessed by replicate (n = 5) analyses of QC samples on the same and on three different days, respectively. The QC samples were analyzed against the freshly prepared calibration curve, and the obtained concentrations were compared with the nominal value. The accuracy is reported as percent of the nominal value. Precision is expressed as the coefficient of variation and should not exceed 15% for the QC samples, except for the LLOQ, which should not exceed 20%. Limits of detection and quantitation were calculated based on S/N ratio of 3 and 10, respectively. Absolute recovery represents the amount of analyte found in the final extract against post-extraction spiked amount expressed as a percentage. Evaluation of stability were carried out to ensure that the steps taken during sample preparation and sample analysis, as well as the storage conditions used, do not affect the concentration of the analyte. The stability study of the analytes in urine matrix was performed using low (150 ng/mL) and high (7500 ng/mL) human and rat QC samples analyzed immediately after preparation and after: three freeze–thaw cycles (from 20 °C to room temperature); 2 months’ storage at 20 °C; and 6 h storage at 22 °C. Processed samples were assayed immediately after preparation and following 24 h in the autosampler (cooled tray, 4 °C).

B. Wielgomas et al.

Relative peak area [%]

120

the lowest extraction rates were observed for 4OH3PBA since it is the most polar among those three studied compounds. Desorption of ionizable analytes is pH dependent, with increasing molarity of NaOH the extraction efficiency also increases for acidic compounds. The highest response was observed when 0.5 M NaOH was used as a receiving solution. Other authors observed higher extraction rates for acidic pesticides using 0.5 M NaOH (Zhu et al., 2002; Wu et al., 2005); however, strongly alkaline samples in this case have to be neutralized before injection. Since silica-based column was used for HPLC analyses, injection of strongly alkaline samples should be avoided. Additionally, disturbed peak shapes (peak fronting) were observed for 0.2 and 0.5 M acceptor solutions owing to the limited buffering capacity of the mobile phase (0.03 M phosphate buffer/ACN). Samples in this case should be neutralized with concentrated acid but an extra step is required in the analytical protocol, thus we decided to use 0.1 M NaOH without the need for pH adjustment before HPLC analysis. Desorption time was optimized using an univariate approach and plateau on the plot (sum of the peak areas vs desorption time) was reached after 120 min with 0.1 M NaOH as a receiving phase (Fig. 3). The influence of the extraction time on the analytical response was studied between 30 and 180 min. During the first 90 min, significant mass transfer was observed and then the system reached equilibrium (Fig. 4); thus the extraction time was set to 120 min. Although the total time needed for extraction and desorption seems long, those two processes are carried out unattended. Based on the visual observations (Table 1), calculated r1–4 (sum of the responses for each level of the factor) and d (difference between the highest and the lowest response for the respective

100 80 60 40 4OH3PBA 20

2PBA 3PBA

0 0

50

100

150

200

Desorption time [min] Figure 3. Optimization of the desorption time.

Relative peak area [%]

120 100 80 60 40 4OH3PBA 20

2PBA 3PBA

0 0

50

100

150

200

Extraction time [min] Figure 4. Optimization of the extraction time.

Table 2. ANOVA table of experimental results Factor

DOF

SS

MS

F

PC (%)

Extraction time (A) Molarity of NaOH (B) Sample pH (C)a Impregnation solvent (D) NaCl concentration (E) Pooled error

3 3 — 3 3 32

1410.3 1876.7 — 524.3 643.6 33.1

470.1 625.6 — 174.8 214.5 11.0

42.7 56.8 — 15.9 19.5

31.4 41.8 — 11.7 14.4 0.7

—

—, Pooled to the error calculation. DOF, Degrees of freedom; SS, sum of squares; MS, mean squares; F, critical value; PC, percentage contribution.

a

Table 3. Validation parameters of HF-LPME combined with HPLC-DAD for determination of synthetic pyrethroid metabolites in human and rat urine Human urine

Retention time (min) Correlation coefficient (r) Linear range (ng/mL) LOD (ng/mL) LOQ (ng/mL) Absolute recovery (%)

Rat urine

4OH3PBA

3PBA

4OH3PBA

3PBA

11.02 0.9906 50–10,000 15 50 49.8

15.90 0.9917 50–10,000 15 50 54.3

11.03 0.9912 50–10,000 15 50 48.8

15.90 0.9945 50–10,000 15 50 55.1

4OH3PBA, 4-Hydroxy-3-phenoxybenzoic acid; 3PBA, 3-phenoxybenzoic acid.

712 wileyonlinelibrary.com/journal/bmc



HF-HPLC-DAD method for pyrethroid metabolites in urine Table 4. Intra- and inter-day accuracy and precision of 4OH3PBA and 3PBA in human and rat urine Spiked concentration (ng/mL)

Human urine Concentration found

Relative standard deviation (%)

Intra-day (n = 6 for each level, one day) 4OH3PBA 50 50.2 150 149.7 5000 5475.3 7500 7571.8 3PBA 50 49.7 150 154.3 5000 5146.2 7500 7584.2 Inter-day (n = 6 for each level, three days) 4OH3PBA 50 53.3 150 144.0 5000 5124.0 7500 7476.5 3PBA 50 53.2 150 152.8 5000 4913.7 7500 7583.7

Rat urine Accuracy (%)

Concentration found

Relative standard deviation (%)

Accuracy (%)

9.7 5.5 2.5 3.0 8.0 7.5 4.5 3.9

100.3 99.8 109.5 101.0 99.3 102.9 102.9 101.1

52.2 150.8 5125.8 7602.7 49.2 154.5 5030.2 7630.2

11.3 7.1 2.4 2.6 8.7 7.5 3.9 2.3

104.3 100.6 102.5 101.4 98.3 103.0 100.6 101.7

12.6 7.0 5.7 4.4 10.7 6.7 5.7 2.8

106.7 96.0 102.5 99.7 106.3 101.9 98.3 101.1

46.7 147.7 5542.8 7607.0 53.5 152.2 4934.5 7388.0

10.3 6.1 2.4 3.2 11.3 5.1 4.8 1.6

93.3 98.4 110.9 101.4 107.0 101.4 98.7 98.5

Table 5. Results of stability study Storage conditions

Analyte

Human urine Concentration (ng/mL) Added

Freeze–thaw cycles, three

4OH3PBA 3PBA

20 °C, 2 months

4OH3PBA 3PBA

Short-term, 6 h, 22 °C

4OH3PBA 3PBA

Autosampler, 24 h, 4 °C, processed sample

4OH3PBA 3PBA

150 7500 150 7500 150 7500 150 7500 150 7500 150 7500 150 7500 150 7500

Accuracy (%)

Found 155.0 7741.0 151.3 7588.7 153.0 7671.7 148.0 7651.0 149.0 7567.3 152.0 7671.7 156.3 7453.7 140.7 7562.0

Concentration (ng/mL) Added

10.1 2.3 10.0 3.5 9.8 2.8 9.5 3.4 5.9 2.2 6.3 1.5 4.3 3.9 3.4 3.2

103.3 103.2 100.9 101.2 102.0 102.3 98.7 102.0 99.3 100.9 101.3 102.3 104.2 99.4 93.8 100.8

150 7500 150 7500 150 7500 150 7500 150 7500 150 7500 150 7500 150 7500

RSD (%)

Accuracy (%)

Found 159.0 7457.3 142.7 7439.0 156.7 7543.7 139.3 7406.7 157.0 7480.0 145.7 7723.3 150.7 7559.3 159.0 7659.0

7.1 5.8 6.1 4.1 7.0 5.1 7.1 2.6 7.3 5.5 7.3 4.4 4.5 0.9 3.9 1.8

97.3 101.5 94.7 100.3 97.3 102.0 94.0 99.9 96.7 101.4 105.3 103.3 95.3 99.7 109.3 104.2

The results of the sums of squares (SS) and F values for particular factors were calculated. Since no dummy column was present in the orthogonal array design, the factor with the least influence was pooled to the error calculation. As a rule, pooling is recommended for factors with an influence of 10% or lower than the most significant factor (Roy, 2001). In this case sample pH was pooled to the error calculation since its MS (mean square) was 11.0 while for the most significant factor the calculated MS was 625.6 before pooling. Results of ANOVA after pooling are presented in Table 2.



713

level of the factor) values for the studied parameters, some conclusions can be drawn. The most significant factor is the molarity of NaOH (d = 397.5) and the least significant sample pH and salt concentration (d = 269.1 and 247.5 respectively). Both extraction time and extraction solvent seems to be also significant (d = 360.7 and 361.4 respectively). The best extraction efficiency should be obtained when factors at the level with the highest r values are applied (A-3, B-4, C-1, D-3 and E-2). Furthermore, ANOVA calculations were performed to verify visual observations.


RSD (%)

Rat urine

5

10

15

20

Time (min)

714

Figure 5. HPLC-DAD (210 nm) chromatograms of blank (lower trace) and fortified at 50 ng/mL (LLOQ – middle trace) and 200 ng/mL (upper trace) human urine sample processed using proposed HF-LPME method in optimized conditions. Concentration of 2PBA in fortified samples was 2000 ng/mL.



Lin et al. (2011) Current method + +++

+ +

12 EI: 4 NCI: 6 LLE: 4.5 SPE: 21 — 1 1

0.05 EI: 0.02 NCI: 0.01 LOD: 1.8 50 10 2

GC-MS GC-HRMS, EI and NCI GC-ECD HPLC-DAD

++ 7

LOQ, limit of quantitation; SPE, solid-phase extraction; HPLC-UV, high-performance liquid chromatography ultraviolet detection; LLE, liquid–liquid extraction; GC-MS, gas chromatography mass spectrometry; GC-HRMS, gas chromatography high-resolution mass spectrometry; EI, electron ionization; NCI, negative chemical ionization; GC-ECD, gas chromatography electron capture detection; LOD, limit of detection; HF-LPME, hollow fiber liquid phase microextraction; HPLC-DAD, high-performance liquid chromatography diode array detection.

0

3PBA 4OH3PBA 3PBA

0

LLE HF-LPME

500

3PBA 3PBA

3PBA

LLE LLE

4OH3PBA

3PBA

1000

Offline-SPE

1500

Analyte

2PBA

Table 6. Comparison with other methods

Absorbance (mAU)

2000

Sample preparation technique

2500

Analytical method

Sample volume (mL)

LOQ (ng/mL)

Application to real samples. The concept of a new setup for hollow fiber-based microextraction presented in this paper was verified by its application to the analysis of real samples. For this purpose we used rat urine from a previous experiment (Wielgomas and Krechniak, 2007), in which animals were treated orally with α-cypermethrin. Using those samples the efficiencies of acidic and enzymatic hydrolyses were compared. No significant differences in the level of deconjugated metabolites were observed between acidic and enzymatic hydrolysis. Thus for all experiments acidic hydrolysis was employed as a simpler and faster technique. Six rat urine samples from animals receiving orally cypermethrin at a dose of 10 mg/kg of body weight daily

100

Organic solvent consumption per sample (mL)

Under optimized conditions the method was validated according to European Medicines Agency (2011) guidelines. Relatively high absolute recovery was observed for analytes (49.8–55.1 %). Using only 1 mL of urine, a limit of detection at the level of 15 ng/mL is achievable (S/N ratio = 3). The method is reproducible as expressed by the relative standard deviation (RSD, 1.6–12.6 %) in a wide range of concentrations (50–10,000 ng/mL) (Table 3). Table 4 shows a summary of the accuracy and precision results of QC samples. Intra- and inter-day accuracies of the method ranged from 98.3 to 109.5% and from 93.3 to 110.9%, respectively. Results of the stability study (Table 5) revealed the concentrations of the analytes in urine to be stable under different storage conditions ( 20 °C for 2 months; 22 °C for 6 h; and after three freeze–thaw cycles). Processed samples were shown to be stable when stored in the autosampler tray at 4 °C for 24 h.

0.5

Validation

HPLC-UV

High throughput

Reference

According to analysis of variance, the highest component of variance is related to molarity of NaOH and extraction time. Those two factors generated over 70% of the total variance in optimization design. Since the pH of the sample showed little effect on the extraction efficiency in the range between strongly acidic and pH 4, samples after hydrolysis (1 mL of urine and 0.2 mL of concentrated HCl) can be extracted without any modification. An optimum extraction and desorption time of 120 min was chosen, with DHE as an extraction solvent and NaCl concentration of 100 mg/mL.

Abu-Qare and Abou-Donia (2001) Schettgen et al. (2002) Leng and Gries (2005)

B. Wielgomas et al.


HF-HPLC-DAD method for pyrethroid metabolites in urine for 14 days were analyzed using optimized HF-LPME method. The results obtained using the method presented in this work were in good agreement with those after SPE-HPLC-DAD analysis (Wielgomas and Krechniak, 2007). Chromatograms of the blank and fortified human urine samples are presented in Fig. 5. In addition to the apparent advantages of HF-LMPE proposed in this work, the fiber-on-the-rod concept exhibited higher cleanup efficiency than SPE. It was not an objective of this study to achieve the highest sensitivity, allowing for detection of concentrations usually found in nonexposed humans (

Determination of low levels of benzodiazepines and their metabolites in urine by hollow-fiber liquid-phase microextraction (LPME) and gas chromatography-mass spectrometry (GC-MS).

MS method for the determination of cortisol, cortisone and tetrahydro-metabolites in human urine: assay development, validation and application in depression patients.

Multi-residue method for the determination of chlorinated phenol metabolites in urine.

Metabolic studies of the nonnutritive sweeteners cyclopentylmethylsulfamate and cyclopentylsulfamate: determination of metabolites in rat urine.

Three-phase hollow-fiber microextraction combined with ion-pair high-performance liquid chromatography for the simultaneous determination of five components of compound α-ketoacid tablets in human urine.

Gas chromatographic method for the determination of dextropropoxyphene and nordextropropoxyphene in human plasma, serum and urine.

[Determination of 2 vinylchloride metabolites in urine].

Development of a dispersive liquid-liquid microextraction technique for the extraction and spectrofluorimetric determination of fluoxetine in pharmaceutical formulations and human urine.

Method for Evaluation of the Requirements of B-group Vitamins Using Tryptophan Metabolites in Human Urine.

Development of a method for the determination of total bisphenol a at trace levels in human blood and urine and elucidation of factors influencing method accuracy and sensitivity.

Magnetic graphene oxide as adsorbent for the determination of polycyclic aromatic hydrocarbon metabolites in human urine.

Double isotope derivative dilution method for the determination of prostaglandin F and E type metabolites in urine.

Determination of aromatic hydrocarbons and their metabolites in human blood and urine.

High-performance liquid chromatographic determination of lansoprazole and its metabolites in human serum and urine.

Solid phase extraction method for rapid isolation and clean-up of some synthetic pyrethroid insecticides from human urine and plasma.

Development of Hollow-Fiber Liquid-Phase Microextraction Method for Determination of Urinary trans,trans-Muconic Acid as a Biomarker of Benzene Exposure.

Binary Solvents Dispersive Liquid-Liquid Microextraction (BS-DLLME) Method for Determination of Tramadol in Urine Using High-Performance Liquid Chromatography.

Determination of DDT and metabolites, including DDA, in human urine by gas chromatography.

Determination of the metabolites of terfenadine in human urine by thermospray liquid chromatography-mass spectrometry.

Capillary gas chromatographic method for the determination of the thromboxane A2 receptor antagonist S-1452 and its metabolites in human urine.

Sequential hollow-fiber liquid phase microextraction for the determination of rosiglitazone and metformin hydrochloride (anti-diabetic drugs) in biological fluids.

An automated method for the determination of total trichloro-compounds in human urine.

Studies on digitalis. IV. A method for thin-layer chromatographic separation and determination of digitoxin and cardioactive metabolites in human blood and urine.

A simple high performance liquid chromatography method for determination of rebamipide in rat urine.