Anal. Chem. 1999, 71, 4237-4244
Automated In-Tube Solid-Phase Microextraction Coupled with Liquid Chromatography/Electrospray Ionization Mass Spectrometry for the Determination of β-Blockers and Metabolites in Urine and Serum Samples Hiroyuki Kataoka,,‡ Shizuo Narimatsu,‡ Heather L. Lord,† and Janusz Pawliszyn*,†
Department of Chemistry, University of Waterloo, Waterloo, Ontario, N2L 3G1, Canada, and Faculty of Pharmaceutical Sciences, Okayama University, Okayama 700-8530, Japan
The technique of automated in-tube solid-phase microextraction (SPME) coupled with liquid chromatography/ electrospray ionization mass spectrometry (LC/ESI-MS) was evaluated for the determination of β-blockers in urine and serum samples. In-tube SPME is an extraction technique for organic compounds in aqueous samples, in which analytes are extracted from the sample directly into an open tubular capillary by repeated draw/eject cycles of sample solution. LC/MS analyses of β-blockers were initially performed by liquid injection onto a LC column. Nine β-blockers tested in this study gave very simple ESI mass spectra, and strong signals corresponding to [M + H]+ were observed for all β-blockers. The β-blockers were separated with a Hypersil BDS C18 column using acetonitrile/methanol/water/acetic acid (15:15:70:1) as a mobile phase. To optimize the extraction of β-blockers, several in-tube SPME parameters were examined. The optimum extraction conditions were 15 draw/eject cycles of 30 µL of sample in 100 mM TrisHCl (pH 8.5) at a flow rate of 100 µL/min using an Omegawax 250 capillary (Supelco, Bellefonte, PA). The β-blockers extracted by the capillary were easily desorbed by mobile-phase flow, and carryover of β-blockers was not observed. Using in-tube SPME/LC/ESI-MS with selected ion monitoring, the calibration curves of β-blockers were linear in the range from 2 to 100 ng/mL with correlation coefficients above 0.9982 (n ) 18) and detection limits (S/N ) 3) of 0.1-1.2 ng/mL. This method was successfully applied to the analysis of biological samples without interference peaks. The recoveries of β-blockers spiked into human urine and serum samples were above 84 and 71%, respectively. A serum sample from a patient administrated propranolol was analyzed using this method and both propranolol and its metabolites were detected. β-Adrenoceptor blocking drugs (β-blockers) are used for the treatment of various cardiovascular disorders such as hypertension, angina pectoris, and cardioarrhythmia.1-3 They are very toxic † ‡
University of Waterloo. Okayama University.
10.1021/ac990356x CCC: $18.00 Published on Web 09/03/1999
© 1999 American Chemical Society
and most have only a narrow therapeutic range. Furthermore, the use of β-blockers has been banned by the International Olympic Committee and International Sports Federations4 because of their sympathomimetic properties, similar to other central nervous system stimulants, and because of their activity as anabolic agents. Therefore, screening and determination of β-blockers in biological samples are required in many circumstances such as clinical control for diagnosis and treatment of diseases, doping control, forensic analysis, and toxicology. Their pharmacological effects are produced by relatively high doses of drugs for most members of the class, and they are excreted in urine as free drug or metabolites or as conjugates with glucuronic acid or sulfate. In a typical oral dose (5-100 mg), maximum blood concentrations of these drugs are below 1 µg/mL and unchanged drug excreted in urine accounts for less than 20% of an oral dose except for pindolol (30-40%).5 Most of the common β-blockers are weakly basic compounds (pKa 8.7-9.7) 5 and structurally have one secondary amino group and one hydroxyl group situated on adjacent carbon atoms (Figure 1). These similarities suggest the possibility of simultaneous analysis of these compounds. Capillary electrophoresis,6 micellar electrokinetic chromatography,7,8 gas chromatography (GC),9 high-performance liquid chromatography (HPLC),10-15 GC/mass spectrometry (GC/MS),16-19and LC/MS20 have been used for the (1) Cruickshank, J. M.; Prichard, B. N. C. Beta-blockers in Clinical Practice; Churchill Livingstone Inc.: New York, 1988. (2) Davies, C. L. J. Chromatogr. 1990, 531, 131-180. (3) Hampton, J. R. Eur. Heart J. 1996, 17 (Suppl. B), 17-20. (4) International Olympic Committee. IOC Medical Code and Explanatory Document; IOC, Lausanne, Switzerland, 1995. (5) Baselt, R. C.; Cravery, R. H. Disposition of Toxic Drugs and Chemicals in Man, 4th ed; Chemical Toxicology Institute: Foster City, CA, 1995. (6) Nguyen, N. T.; Siegler, R. W. J. Chromatogr., A 1996, 735, 123-150. (7) Siren, H.; Saarinen, M.; Hainari, S.; Lukkari, P.; Riekkola, M. L. J. Chromatogr. 1993, 632, 215-227. (8) Lukkari, P.; Nyman, T.; Riekkola, M. L. J. Chromatogr., A 1994, 674, 241246. (9) Hyotylainen, T.; Andersson, T.; Riekkola, M. L. J. Chromatogr. Sci. 1997, 35, 280-286. (10) Musch, G.; Buelens, Y.; Massart, D. L. J. Pharm. Biomed. Anal. 1989, 7, 483-497. (11) Alpertunga, B.; Sungur, S.; Ersoy, L. Manav, S. Y. Arch. Pharm. (Weinheim) 1990, 323, 587-589. (12) Saarinen, M. T.; Siren, H.; Riekkola, M. L. J. Chromatogr., B: Biomed. Appl. 1995, 664, 341-346.
Analytical Chemistry, Vol. 71, No. 19, October 1, 1999 4237
Figure 1. Structures of β-blockers used in this study.
determination of β-blockers in biological samples. Comprehensive summaries of these methods have been given in some reviews.2,21-23 Among these methods, GC/MS is widely accepted as the definitive analytical method because of its sensitivity and selectivity and the easy identification of compounds from mass spectra. However, the required derivatization, poor GC properties, and instability of some derivatives may limit the use of GC/MS. Although HPLC and electrophoresis methods do not necessarily require derivatization, these methods are less sensitive and specific for β-blockers. On the other hand, LC/MS offers the advantages of both HPLC and MS, although it has not yet been applied to complex matrix samples such as urine and serum. Furthermore, most of above methods require laborious cleanup of the biological samples because β-blockers are generally present at low concentration in these complex matrixes. To achieve a more efficient, practical, and reliable method for the analysis of β-blockers in biological samples, sample preparation is very important. Ideally, a sample preparation method should be fast and simple. Several methods such as liquid-liquid extraction and solid-phase extraction have been used for cleanup (13) Maguregui, M. I.; Alonso, R. M.; Jimenez, R. M. J. Chromatogr., B: Biomed. Appl. 1995, 674, 85-91. (14) Ohta, T.; Niida, S.; Nakamura, H. J. Chromatogr., B: Biomed. Appl. 1996, 675, 168-173. (15) Rapado-Martinez, I.; Garcia-Alvarez-Coque, M. C.; Villanueva-Camanas, R. M. Analyst 1996, 121, 1677-1682. (16) Zamecnik, J. J. Anal. Toxicol. 1990, 14, 132-136. (17) Hartonen, K.; Riekkola, M. L. J. Chromatogr., B: Biomed. Appl. 1996, 676, 45-52. (18) Black, S. B.; Stenhous, A. M.; Hansson, R. C. J. Chromatogr., B: Biomed. Appl. 1996, 685, 67-80. (19) Branum, G. D.; Sweeney, S.; Palmeri, A.; Haines, L.; Huber, C. J. Anal. Toxicol. 1998, 22, 135-141. (20) Caldwell, G. W.; Easlick, S. M.; Gunnet, J.; Masucci, J. A.; Demarest, K. J. Mass Spectrom. 1998, 33, 607-614. (21) Soltes, L. Biomed. Chromatogr. 1989, 3, 139-152. (22) Maurer, H. H. J. Chromatogr. 1992, 580, 3-41. (23) Hemmersbach, P.; de la Torre, R. J. Chromatogr., B: Biomed. Appl. 1996, 687, 221-238.
4238 Analytical Chemistry, Vol. 71, No. 19, October 1, 1999
of the samples. However, these methods are time-consuming, require large volumes of samples and solvent, and are not easy to automate. To address these problems, some on-line sample preparation methods such as column-switching techniques,12,14 micellar chromatography,7,8 and hyphenated chromatographic techniques such as LC/GC9 have been developed. Although these methods save preparation time, some problems still exist. For example, in the column-switching techniques, the retention behaviors of β-blockers by the precolumn are dependent on the pH of the mobile phase. Micellar chromatography requires an additional preconcentration procedure to increase sensitivity, and LC/GC is a relatively complicated system. Solid-phase microextraction (SPME), recently developed by Pawliszyn and coworkers,24,25 is an extraction technique using a fused-silica fiber that is coated on the outside with an appropriate stationary phase. The method saves preparation time, solvent purchase, and disposal cost and can improve the detection limits.24-27 It has been used routinely in combination with GC and GC/MS and successfully applied to a wide variety of compounds.25-28 However, these methods are not suitable for weakly volatile or thermally labile compounds such as most drugs. To solve these problems, SPME was recently introduced for direct coupling with HPLC29-34 and LC/MS.35-39 The SPME/LC interface equipped with a special (24) Arthur, C. L.; Pawliszyn, J. Anal. Chem. 1990, 62, 2145-2148. (25) Zhang, Z.; Yang, M. J.; Pawliszyn, J. Anal. Chem. 1994, 66, 844A-853A. (26) Eisert, R.; Levsen, K. J. Chromatogr., A 1996, 733, 143-157. (27) Pawliszyn, J. Solid-Phase Microextraction: Theory and Practice; Wiley-VCH: New York, 1997. (28) Eisert, R.; Pawliszyn, J. Crit. Rev. Anal. Chem. 1997, 27, 103-135. (29) Lord, H. L.; Pawliszyn, J. Current Trends and Developments in Sample Preparation LC-GC 1998, S41-S46. (30) Chen, J.; Pawliszyn, J. Anal. Chem. 1995, 67, 2530-2533. (31) Boyd-Boland, A. A.; Pawliszyn, J. Anal. Chem. 1996, 68, 1521-1529. (32) Jinno, K.; Muramatsu, T.; Saito, Y.; Kiso, Y.; Magdic, S.; Pawliszyn, J. J. Chromatogr., A 1996, 745, 137-144. (33) Liu, Y.; Lee, M. L.; Hageman, K. J.; Yang, Y.; Hawthorne, S. B. Anal. Chem. 1997, 69, 5001-5005. (34) Jia, C.; Luo, Y.; Pawliszyn, J. J. Microcolumn Sep. 1998, 10, 167-173.
Figure 2. Schematic diagram of the in-tube SPME/LC/MS system: (A) load position (extraction phase); (B) injection position (desorption phase).
desorption chamber is utilized for solvent desorption prior to LC analysis instead of thermal desorption in the injection port of the GC. Moreover, a new SPME/LC system known as in-tube SPME, was recently developed using an open tubular fused-silica capillary as the SPME device instead of the SPME fiber.40 In-tube SPME is suitable for automation, and automated sample-handling procedures not only shorten the total analysis time but also usually provide better accuracy and precision relative to manual techniques. In this study, an automated in-tube SPME method coupled with LC/MS was developed for the determination of nine β-blockers. (35) Mioder, M.; Loster, H.; Herzschuh, R. J. Mass Spectrom. 1997, 32, 11951204. (36) Volmer, D. A.; Hui, J. P. M.; Joseph, P. M. Rapid Commun. Mass Spectrom. 1997, 11, 1926-1934. (37) Moder, M.; Popp, P.; Pawliszyn, J. J. Microcolumn Sep. 1998, 10, 225234. (38) Volmer, D. A.; Hui, J. P. M.; Joseph, P. M. Rapid Commun. Mass Spectrom. 1998, 12, 123-129. (39) Volmer, D. A.; Hui, J. P. M. Arch. Environ. Contam. Toxicol. 1998, 35, 1-7.
This has been facilitated by the Hewlett-Packard 1100 LC/MS, as the standard autosampler for this system (ALS 1100) is ideally suited for in-tube SPME. The schematic diagram of the automated in-tube SPME/LC/MS system is illustrated in Figure 2. In this technique, organic compounds in aqueous samples are extracted directly from the sample into the internally coated stationary phase of a capillary. The capillary is placed between the injection loop and the injection needle of the HPLC autosampler. While the injection syringe, under computer control, repeatedly draws and ejects sample from the vial, the analytes partition from the sample matrix into the stationary phase until equilibrium is reached. The extracted analytes are directly desorbed from the capillary coating by mobile-phase flow, transported to the HPLC column, and then detected by the mass-selective detector (MSD). This method was also applied to the analysis of β-blockers in spiked urine and serum samples and of propranolol and its metabolites in a clinical serum sample. (40) Eisert, R.; Pawliszyn, J. Anal. Chem. 1997, 69, 3140-3147.
Analytical Chemistry, Vol. 71, No. 19, October 1, 1999
4239
Table 1. Comparison of Direct Injection and In-Tube SPME on the Extraction of β-Blockers area counts (× 106)
Figure 3. Evaluation of four capillary columns for the in-tube SPME/ LC/MS analysis of β-blockers. Capillary column: 60 cm × 0.25 mm i.d., 0.25-µm film thickness. SPME conditions: β-blockers, 0.5 µg/ mL; sample pH, 8.5 (100 mM Tris-HCl); draw/eject cycles, 15; draw/ eject volume, 30 µL; draw/eject rate, 100 µL/min; desorption, mobile phase.
compounds
direct injectiona
in-tube SPMEb
extraction yieldc (%)
nadolol pindolol acebutolol timolol metoprolol oxprenolol labetalol propranolol alprenolol
0.36 0.62 0.46 0.42 1.02 0.41 0.15 0.39 0.42
0.76 (5 ng) 8.10 (33 ng) 4.63 (25 ng) 1.75 (10 ng) 5.00 (12 ng) 4.02 (25 ng) 2.56 (43 ng) 10.61 (68 ng) 7.38 (44 ng)
1.0 6.6 5.0 2.0 2.4 5.0 8.6 13.6 8.8
a A 5-µL (2.5 ng) sample of β-blockers in 100 mM Tris HCl (pH 8.5) was directly injected. b A 1-mL sample of 0.5 µg/mL β-blockers in 100 mM Tris HCl (pH 8.5) was extracted by in-tube SPME, desorbed with mobile phase, and injected. Extracted amounts were calculated in comparison with area counts of β-blockers in direct injection and in-tube SPME. c Percentages of extracted amount of β-blockers per initial amounts (500 ng) in sample solution using in-tube SPME.
Table 2. Linear Regression Data and Detection Limits for β-Blockers
Figure 4. Effect of sample pH on the extraction efficiency of β-blockers with Omegawax 250 capillary. Buffer: pH 5.5, sodium acetate buffer; pH 7.0, sodium phosphate buffer; pH 8.5, Tris-HCl buffer; pH 10.0, sodium carbonate buffer. Buffer concentrations in all samples were 100 mM. Other conditions are the same as in Figure 3.
Figure 5. Extraction-time profile of β-blockers with Omegawax 250 capillary. Other conditions are the same as in Figure 3.
EXPERIMENTAL SECTION Reagents. Structures of the nine β-blockers used in this study are shown in Figure 1 and include acebutolol hydrochloride, alprenolol hydrochloride, labetalol hydrochloride, metoprolol tartrate, nadolol, oxprenolol hydrochloride, pindolol, (S)-(-)propranolol hydrochloride, and timolol maleate. All β-blockers were purchased from Sigma Chemicals (St. Louis, MO). N4240 Analytical Chemistry, Vol. 71, No. 19, October 1, 1999
compounds
regression linea slope intercept
nadolol pindolol acebutolol timolol metoprolol oxprenolol labetalol propranolol alprenolol
0.1681 1.7622 0.7886 0.3247 1.1191 0.8111 0.5344 1.1766 1.0753
3.551 × 104 2.887 × 105 1.490 × 105 8.928 × 104 2.557 × 105 1.953 × 105 6.123 × 104 -8.976 × 104 6.542 × 104
correl coeff
detection limit (ng/ml)
0.9997 0.9995 0.9992 0.9986 0.9982 0.9990 0.9996 0.9996 0.9995
1.15 0.10 0.32 0.87 0.20 0.43 1.19 0.82 0.82
a Calculated from area counts of each β-blocker; range, 2-100 ng/ mL; number of data points, 6 points (n ) 3 each point).
Desisopropylpropranolol hydrochloride and 4-hydroxypropranolol hydrochloride were purchased from ICI Pharmaceuticals (Macclesfield, U.K.) and Sumitomo Chemical Co. (Osaka, Japan), respectively. 5-Hydroxypropranolol and 7-hydroxypropranolol were synthesized according to the method of Oatis et al.41 The drugs and metabolites were dissolved in methanol to make a stock solution at a concentration of 1 mg/mL. The solutions were stored at 4 °C and used after dilution with water to the required concentration. All solvents used in this study were of HPLC grade. Water was obtained from a Barnstead/Thermodyne Nano pure ultrapure water system (Dubuque, IA). Instrument and Analytical Conditions. The LC/MS system used was a model 1100 series LC coupled with an atmospheric pressure (AP)-electrospray ionization (ESI) mass spectrometer (Hewlett-Packard, Palo Alto, CA). A Hypersil BDS C18 column (5.0 cm × 2.1 mm i.d., 3-µm particle size) from Hewlett-Packard was used for the LC separation. LC conditions were as follows: column temperature, 25 °C; mobile phase, acetonitrile/methanol/water/ acetic acid (15:15:70:1); flow rate, program from 0.2 to 0.45 mL/ min for a 25-min run. ESI-MS conditions were as follows: nebulizer (41) Oatis, J. E.; Russel, M. P.; Knapp, D. R.; Walle, T. J. Med. Chem. 1981, 24, 309-314.
Figure 6. Total ion and SIM chromatograms obtained from urine samples by in-tube SPME/LC/MS. (A) Total ion chromatograms obtained from urine and spiked urine samples. The spiked urine trace is offset by 10% for clarity. (B) SIM chromatograms obtained from spiked urine samples. Urine samples (10 µL) were diluted 10 times with water and used for analysis after filtration. β-Blockers were spiked at concentrations of 2 µg/mL urine. In-tube SPME/LC/MS conditions see Experimental Section. Peaks: (1) nadolol, (2) pindolol, (3) acebutolol, (4) timolol, (5) metoprolol, (6) oxprenolol, (7) labetalol, (8) propranolol, and (9) alprenolol.
gas, N2 (40 psi); drying gas, N2 (10 L/min, 350 °C); fragmentor voltage, 70 V; ionization mode, positive; mass scan range, 100400 amu; selected ion monitoring (SIM) of the [M + H]+ ions, m/z 218 (N-desisopropylpropranolol), 249 (pindolol), 250 (alprenolol), 260 (propranolol), 266 (oxprenolol), 268 (metoprolol), 276 (hydroxypropranolol), 310 (nadolol), 317 (timolol), 329 (labetalol), and 337 (acebutolol). In-Tube Solid-Phase Microextraction. As shown in Figure 2, a GC capillary (60 cm × 0.25 mm i.d., 0.25-µm film thickness) was used as the in-tube SPME device and placed between the injection loop and injection needle of the autosampler. The injection loop was retained in the system to avoid fouling of the metering pump. Capillary connections were facilitated by the use of a 2.5-cm sleeve of 1/16-in. polyetheretherketone (PEEK) tubing at each end of the capillary. A PEEK tubing internal diameter of 330 µm was found to be suitable to accommodate the capillary used. Normal 1/16-in. stainless steel nuts, ferrules, and connectors
Figure 7. Total ion and SIM chromatograms obtained from serum samples by in-tube SPME/LC/MS. (A) Total ion chromatograms obtained from serum and spiked serum samples. The spiked serum trace is offset by 10% for clarity. (B) SIM chromatograms obtained from spiked serum samples. Serum samples (100 µL) were diluted 5 times with 1% acetic acid and used for analysis after ultrafiltration. β-Blockers were spiked at concentrations of 200 ng/mL serum. Intube SPME/LC/MS conditions, see Experimental Section. Peak numbers are the same as in Figure 6.
were then used to complete the connections. Omegawax 250, SPB5, SPB-1, and retention gap capillary (no coating) (Supelco, Bellefonte, PA) were tested for comparison of extraction efficiency. The total internal volume of each capillary was 29.4 µL. The autosampler software was programmed to control the in-tube SPME extraction, desorption, and injection. Vials (2 mL) were filed with 1 mL of sample in 100 mM Tris-HCl (pH 8.5) for the absorption and set into the autosampler programmed to control the SPME absorption and desorption technique. In addition, 1.5 mL each of methanol and mobile phase in 2-mL vials were set on the autosampler. The capillary column was washed and conditioned by two repeated draw/eject cycles (40 µL each) of these solvents prior to extraction. The extraction of β-blockers into the capillary coating was performed by 15 repeated draw/eject cycles of 30 µL of sample at a flow rate of 100 µL/min with the six-port valve in the LOAD position. After washing the tip of the injection needle by a draw/eject cycle (2 µL) of methanol, the six-port valve was switched to the INJECT position and the extracted β-blockers were desorbed from the capillary coating with mobile-phase flow Analytical Chemistry, Vol. 71, No. 19, October 1, 1999
4241
Table 3. Recoveries of 9 β-Blockers Spiked into Urine and Serum Samples recovery (%), mean ( SDa (n ) 3) urineb
serumc
compound
0.5 µg/mL
5 µg/mL
50 ng/mL
500 ng/mL
nadolol pindolol acebutolol timolol metoprolol oxprenolol labetalol propranolol alprenolol
98.4 ( 7.5 (7.6) 93.7 ( 3.9 (4.2) 93.5 ( 3.7 (4.0) 85.5 ( 1.6 (1.9) 85.1 ( 4.5 (5.3) 95.6 ( 2.2 (2.3) 103.7 ( 0 (0) 113.0 ( 4.2 (3.7) 106.0 ( 2.5 (2.4)
89.7 ( 3.1 (3.5) 92.6 ( 1.4 (1.5) 88.3 ( 2.7 (3.1) 84.9 ( 2.1 (2.5) 84.1 ( 3.2 (3.8) 87.9 ( 4.0 (4.6) 96.0 ( 0.8 (0.8) 98.4 ( 2.3 (2.3) 94.2 ( 3.2 (3.4)
71.0 ( 1.3 (1.8) 86.4 ( 2.0 (2.3) 89.0 ( 3.8 (4.3) 81.9 ( 1.7 (2.1) 76.4 ( 1.6 (2.1) 90.0 ( 0.9 (1.0) 103.0 ( 3.7 (3.6) 105.5 ( 4.4 (4.1) 110.3 ( 3.4 (3.1)
79.2 ( 1.5 (1.9) 78.7 ( 2.3 (2.9) 81.9 ( 0.5 (0.6) 75.4 ( 1.5 (2.0) 72.1 ( 1.0 (1.4) 73.3 ( 1.3 (1.8) 103.9 ( 2.0 (1.9) 111.7 ( 1.4 (1.3) 90.7 ( 1.9 (2.1)
a Relative standard deviations (%) are shown in parentheses. b β-Blockers were spiked into 10 µL of urine; the mixture was diluted 10 times with water and filtered using a syringe microfilter before analysis. c β-Blockers were spiked into 100 µL of serum, and the mixture was diluted 5 times with 1% acetic acid and ultrafiltered using a Nanosep centrifugal microconcentrator model 3K before analysis.
and then transported to the LC column. The sample was transferred to the MSD detector by means of the mobile phase flow. Calibration curves of β-blockers were constructed by in-tube SPME/LC/MS-SIM of six point calibration solutions ranging from 2 to 100 ng/mL. Preparation of Urine and Serum Samples. Drug-free urine and serum samples were collected from a healthy volunteer. A clinical serum sample was aquired from a patient taking 4-60mg doses/day in chronic oral therapy. Urine samples were diluted 10 times with water, and 100 µL of this solution (corresponding to 10 µL of urine) was used after filtration (syringe microfilter, 0.45 µm, Gelman Science). Serum samples were first diluted 5 times with 1% acetic acid and then ultrafiltered using a Nanosep centrifugal microconcentrator model 3K (Pall Filtron, Northborough, MA) at 10000g for 20 min; 500 µL of the filtrate (corresponding to 100 µL of serum) was used for the analysis. An aliquot of each sample was pipetted into a 2-mL vial, and 0.2 (for urine samples) or 0.5 mL (for serum samples) of 0.5 M Tris-HCl buffer (pH 8.5) was added. After the total volume was made up to 1 mL with water, the vials were set on the autosampler. Recoveries of β-blockers from urine and serum samples were calculated on the basis of calibration curves mentioned above after in-tube SPME/LC/MS-SIM of nonspiked and spiked (5 or 50 ng of β-blockers) samples in urine (10 µL) and serum (100 µL). RESULTS AND DISCUSSION Liquid Chromatography/Electrospray Ionization Mass Spectrometry. To select the monitoring ion for each of the β-blockers, ESI mass spectra were initially analyzed by direct liquid injection. Each drug gave a very simple spectrum in the scan mass range m/z 100-400. All drugs gave [M + H]+ ions as the base ion. Each base ion accounted for more than 70% of the total ion current except for labetalol (56%). These results indicate that these base ions are useful for quantification in selected ion monitoring of each β-blocker. Optimal fragmentor voltage was 70 V for all β-blockers except for labetalol (50 V). LC separation of β-blockers was performed with a Hypersil BDS C18 column using acetonitrile/methanol/water/acetic acid as a mobile phase. Although a decrease of the content of water in 4242 Analytical Chemistry, Vol. 71, No. 19, October 1, 1999
the mobile phase shortened the retention time of β-blockers, the separation was worsened. Especially, oxyprenolol and labetalol coeluted completely with acetonitrile/methanol/water/acetic acid (20:20:60:1). Nine β-blockers were well separated under the following conditions: column, Hypersil BDS C18 (5.0 cm × 2.1 mm i.d., 3-µm particle size); mobile phase, acetonitrile/methanol/ water/acetic acid (15:15:70:1, pH 4); flow rate, program from 0.2 to 0.45 mL/min for a 25-min run. A higher water content in the mobile phase required an inconveniently long analysis time. Optimization of In-Tube Solid-Phase Microextraction and Desorption. To optimize the extraction of β-blockers by in-tube SPME, several parameters such as stationary phase of capillary, extraction pH and number, and volume for draw/eject cycles were investigated. Four different capillaries were evaluated for in-tube SPME/LC/MS-SIM. As expected, the relatively polar Omegawax 250 capillary gave superior extraction efficiency as compared to the less polar SPB-5, SPB-1, and uncoated columns (Figure 3). The effect of sample matrix pH on the extraction of β-blockers by in-tube SPME was examined using several buffer solutions at pH 5.5-10.0. As shown in Figure 4, Tris-HCl (pH 8.5) was most effective, and the optimal concentration of this buffer was 100 mM. To monitor the extraction-time profiles of β-blockers by in-tube SPME, the number of draw/eject cycles was varied from 0 to 20. After 20 draw/eject cycles, equilibrium conditions were not obtained for the extraction of several β-blockers (Figure 5). An increase of the draw/eject volume to 40 µL did not change these results. However, above 30 µL or 20 cycles, peak broadening was observed, although sensitivity also increases. The peak broadening was considered to be caused by broadening of the bandwidth of analytes extracted into the capillary. Best overall results were obtained using 15 draw/eject cycles of 30 µL of sample. The absolute amounts of β-blockers extracted by the capillary column under optimal conditions were calculated by comparison with the corresponding direct injection of the sample solution onto the LC column. As shown in Table 1, 5-68 ng (1.0-13.6%) was extracted by in-tube SPME of β-blockers at 0.5 µg/mL. The low extraction yields of drugs containing tertiary butyl groups such as nadolol and timolol were considered to be caused by their low affinities to the capillary coating (partition coefficients).
Figure 8. Total ion and SIM chromatograms obtained from standard propranolol and its metabolites, and a clinical serum sample by intube SPME/LC/MS. (A) Standard solution containing 200 ng/mL propranolol, 50 ng/mL 4-hydroxypropranolol and 7-hydroxypropranolol, and 20 ng/mL 5-hydroxypropranolol and N-desisopropylpropranolol. (B) Clinical serum sample (100 µL). In-tube SPME/LC/MS conditions are given in the Experimental Section except for LC flow rate, program from 0.25 to 0.45 mL/min for 20-min run. Peaks: (1) 5-hydroxypropranolol, (2) 4-hydroxypropranolol, (3) 7-hydroxypropranolol, (4) N-desisopropylpropranolol, and (5) propranolol.
Mobile phase (pH 4) was found to be most suitable for desorption of β-blockers absorbed into the stationary phase of the capillary. Desorption of β-blockers from the capillary with mobile phase was achieved by simply switching the six-port valve to the INJECT position. The desorbed β-blockers were easily transported to the LC column with mobile-phase flow. The entire in-tube SPME extraction and desorption of β-blockers was accomplished automatically in 15 min, and carryover of β-blockers was not observed. The capillary was very stable for analyses of at least 500 samples and deterioration was not observed for at least 3 months. Detection Limits and Calibration Curves of β-Blockers. β-Blockers provided excellent response in SIM, and the detection limits (S/N ) 3) under our LC/MS conditions were 0.1-1.2 ng/ mL (Table 2). The in-tube SPME method gave 2-27 times higher sensitivity than the direct injection method (Table 1). These
sensitivities are similar or higher than those obtained previously. To test the linearities of the calibration curves, various concentrations of β-blockers ranging from 2 to 100 ng/mL were analyzed. The calibration curves were constructed from peak area counts using the SIM mode. As shown in Table 2, a linear relationship was obtained for each drug in this range (six-point calibration). The correlation coefficients were 0.9982-0.9997, and relative standard deviations were 0.4-8.9% (n ) 3). Application to the Analysis of Urine and Serum Samples. Urine samples (10 µL) were used after a 10 times dilution with water and filtration by syringe microfilter (0.45 µm). Serum samples (100 µL) were used after adjustment to pH 4 with 1% acetic acid and ultrafiltration by a centrifugal microconcentrator. No further sample pretreatment was performed. As shown in Figures 6A and 7A, no interference peaks were observed in nonspiked urine and serum samples. β-Blockers (each 5 or 50 ng) were spiked to 10 µL of urine and 100 µL of serum, and analyzed by in-tube SPME/LC/MS after filtration. As shown in Figures 6B and 7B, each β-blocker in urine and serum samples could be selectively detected in the SIM mode. In these samples, however, the peaks of β-blockers were slightly broadened in comparison with direct liquid injection. As shown in Table 3, the recoveries from urine and serum samples were in the range of 84-113 and 71-112%, respectively. Recoveries of β-blockers from serum samples by ultrafiltration in water were low but were improved by ultrafiltration under acidic conditions. β-Blockers are likely bound to serum components such as albumin, at neutral pH, because β-blockers were spiked to serum before ultrafiltration. β-Blockers are basic compounds and are thus more soluble in acidic solutions. In the analysis of biological samples, the quantitation limits of β-blockers were 13-79 ng/mL urine and 3-49 ng/mL serum, and coefficients of variation of three replicate analyses were below 7.6%. Analysis of a Clinical Serum Sample. The method developed for β-blockers was applied to the analysis of a serum sample from a patient administrated propranolol. The major metabolites of propranolol in humans arise from initial oxidative reactions, ring hydroxylations, and side-chain N-desisopropylation, catalyzed by P-450 enzymes.42 Propranolol and most of the metabolites are excreted in urine both as glucronide conjugates and in free form.5 In this study, free propranolol and its major metabolites, 4-, 5-, and 7-hydroxypropranolol and N-desisopropylpropranolol, were analyzed by SIM mode detection. As shown in Figure 8, propranolol (134 ng/mL), 4-hydroxypropranolol (7.0 ng/mL), 7-hydroxypropranolol (1.5 ng/mL), and N-desisopropylpropranolol (2.3 ng/ mL) were detected. The concentration of propranolol in the patient serum was within the range reported previously.5 CONCLUSION In-tube SPME is an ideal sample preparation technique because of fast operation, simple automation, lack of solvent requirement, and low expense. The automated in-tube SPME/LC/MS method demonstrated in this study can continuously perform extraction of β-blockers from aqueous samples, followed by LC/MS analysis. This method is simple, rapid, selective, and sensitive for β-blocker (42) Von Bahr, C.; Hermansson, J.; Lind, M. J. Pharmacol. Exp. Ther. 1982, 222, 458-462.
Analytical Chemistry, Vol. 71, No. 19, October 1, 1999
4243
analysis and is applied directly to the analysis of biological fluid samples after a minimal cleanup procedure (filtration only). It also has the benefit of removal of matrix interferences, such as salt ions, prior to injection.29 We believe that this method provides a useful tool for the screening and determination of β-blockers in clinical control, doping control, and forensic analysis, and the application range of automated in-tube SPME/LC/MS can be easily extended for nonvolatile and thermally labile compounds by selection of appropriate capillary coatings.
4244 Analytical Chemistry, Vol. 71, No. 19, October 1, 1999
ACKNOWLEDGMENT This work was supported by the National Science and Engineering Research Council, Supelco Inc., Varian Associates, and Hewlett-Packard.
Received for review April 6, 1999. Accepted June 25, 1999. AC990356X
