Anal. Chem. 1999, 71, 2908-2914
Identification and Quantitation of Despropionyl-bezitramide in Postmortem Samples by Liquid Chromatography Coupled to Electrospray Ionization Tandem Mass Spectrometry Siegrid M. De Baere,† Willy E. Lambert,† Eddy L. Esmans,‡ and Andre´ P. De Leenheer*,†
Laboratory of Toxicology, University of Ghent, Harelbekestraat 72, B-9000 Ghent, Belgium, and Nucleoside Research and Mass Spectrometry Unit, University of Antwerp, Groenenborgerlaan 171, B-2020 Antwerp, Belgium
A sensitive and specific method for the determination and quantitation of (despropionyl) bezitramide in postmortem samples using liquid chromatography combined with electrospray ionization tandem mass spectrometry (LCESI-MS/MS) is presented. The method is the result from a simple methodological transfer of a liquid chromatographic method with fluorescence detection (LC-FL) previously developed in our laboratory. A liquid-liquid back-extraction procedure using n-hexane isoamyl alcohol (93:7, v/v) as the extraction solvent is performed for a basic sample cleanup. N-Methyldespropionyl bezitramide is used as the internal standard. Chromatographic separation of the analytes of interest is achieved on a Hypersil ODS 5-µm column, using a 80:20 (v/v) mixture of 1.0 mM ammonium acetate and methanol/acetonitrile (50: 50, v/v) and 1.0 mM ammonium acetate as the mobile phase. To obtain as high a sensitivity and selectivity as possible, a selected reaction-monitoring mass spectrometric technique is applied. In addition, low-energy collisional-activated dissociation (CAD) product ion spectra are recorded for a few samples. Calibration graphs are prepared for blood and urine, and good linearity is achieved over a concentration range of 1-150 ng/mL. The intra- and interassay coefficients of variation (CV%) for the analysis of quality control samples at 10 and 50 ng/mL concentration levels do not exceed 10.2%, and percent of targets are within 12.1%. Postmortem samples (blood, urine, stomach contents, bile, liver, and kidney) from three fatalities, all suspected victims of drug overdoses, are analyzed, and the results are reported. The results obtained with LC-ESI-MS/MS are in close agreement with those obtained using the LC-FL method. Moreover, the isolates’ identity and structure are confirmed by the CAD product ion spectra, thus allowing to make unequivocal conclusions about the prior intake of bezitramide by the three subjects.
of severe chronic pain (e.g., postoperative pain, cancer pain).1 However, nausea, confusion, and side effects of a psychic nature were reported, 2 resulting in a serious decrease of the clinical application of the drug. Today bezitramide is often abused by drug addicts for its euphoric side effect, and as such it is not inconceivable that an overdose arises as a result of the combined intake of massive amounts of drugs and bezitramide. Hence the compound belongs to the sphere of interests of today’s forensic toxicologists. Bezitramide is considered a prodrug: rapid hydrolysis of the propionyl group in the gastrointestinal tract leads to a hydrolysis product with analgesic properties similar to the parent compound. Further metabolism leads to inactive metabolites which are mainly found in urine.1 A sensitive radioimmunoassay (LOD: 0.5 ng/tube) has been reported for the determination of the parent compound, bezitramide, and its active metabolite despropionyl bezitramide, in biological fluids and tissues. Although other bezitramide metabolites are not determined, a number of compounds show a substantial cross-reactivity in this assay (e.g., piritramide, diphenoxylate, diphenoxin, a bis(thiophene) analogue, and the compounds with a halogen substituent in the 5 or 6 position of the benzimidazole moiety).3 In addition, the antibody for this radioimmunoassay is no longer available, which is a practical disadvantage of this method. Recently, a reversed-phase high-performance liquid chromatographic method with fluorescence detection (LC-FL) was developed in our laboratory for the quantitative analysis of despropionyl bezitramide in biological samples.4 The combination of a liquid chromatographic separation and a detection based on native fluorescence was advantageous in view of the simplicity of the technique and the rather inexpensive instrumentation that is required. As a quantification limit of 1 ng/mL could be achieved, the method was sensitive enough to establish the generally low bezitramide/despropionyl bezitramide levels in blood and urine. However, the method lacks implicit selectivity since some anti-
Bezitramide (Burgodin) is a potent, long-acting, orally active narcotic analgesic, which was used since 1970 for the treatment
(1) Meijer, D. K. F.; Hovinga, G.; Versluis, A.; Bro ¨ring, J.; Van Aken, K.; Moolenaar, F.; Wesseling, H. Eur. J. Clin. Pharmacol. 1984, 27, 615-618. (2) Van Der Linden, C.; Hamersma-Van Der Linden, E.; Van Dam, F.; Engelsman, E. J. Drug Res. 1979, 1, 344-351. (3) Hendriks, R.; Michiels, M.; Heykants, J. Preclinical research report no. R4845/3. Janssen Pharmaceutica, Beerse, Belgium, 1977. (4) De Baere, S.; Lambert, W.; De Leenheer, A. Anal. Chem. 1997, 69, 51865192.
* To whom correspondence should be addressed. Phone: ++32 9 264 81 31. Fax: ++32 9 264 81 97. E-mail:
[email protected]. † University of Ghent. ‡ University of Antwerp.
2908 Analytical Chemistry, Vol. 71, No. 14, July 15, 1999
10.1021/ac981310r CCC: $18.00
© 1999 American Chemical Society Published on Web 06/10/1999
depressive drugs (e.g., butriptyline, trimipramine, lofepramine, nortriptyline, and mianserine) interfered with despropionyl bezitramide. For forensic purposes, which could be postmortem investigations or analyses of samples from living persons (e.g., drug abusers, impaired drivers, etc.), the confirmation of the identities and structures of the analytes of interest is of major importance. To combine both high sensitivity and selectivity, the coupling of HPLC to MS was found suitable. The purpose of the present study was to measure the levels of despropionyl bezitramide in the various postmortem samples of fatally intoxicated drug addicts by means of LC-ESI-MS/MS. The study also gave the opportunity to compare the present results with those obtained with the LC-FL method. EXPERIMENTAL SECTION Apparatus. LC-FL System. The system was composed of a Pye Unicam isocratic LC3-XP pump (Cambridge, U.K.) equipped with an injection valve with a 100-µL sample loop. The HPLC unit was coupled to a Perkin-Elmer LS-4 fluorescence spectrometer with a Xenon lamp (Buckinghamshire, U.K.). The detector was linked to a Hewlett-Packard 3396 series II reporting integrator (Palo Alto, CA). LC-ESI-MS/MS System. The HPLC unit was composed of a ternary low-pressure gradient pump (model 325) and an autosampler (model 460) with a 100-µL loop (both from Kontron Instruments, Milano, Italy). Solvents were continuously degassed using Helium. The mobile phase flow (0.5 mL/min) was split 1/20 after the column using a LC-Packings Accurate splitter (LC Packings, San Francisco, USA). Twenty five µL/min entered the ESI source, the rest was diverted to the waste. The HPLC unit was coupled to a VG Quattro II triple quadrupole system (Micromass, UK) equipped with a Mass Lynx data system. Reagents and Standards. All solvents used for the mobile phase (water, methanol, acetonitrile) were of HPLC grade and were purchased from Prosan (Gent, Belgium). All products (ammonium acetate, sodium hydroxide, sulfuric acid, concentrated ammonia) and solvents used for the extraction procedure (nhexane, isoamyl alcohol) were analytical grade and were obtained from UCB (Leuven, Belgium). The ethereal diazomethane solution was synthesized in our laboratory. Elution solvent A consisted of a 1.0 mM aqueous solution of ammonium acetate. Elution solvent B was prepared by adding to a mixture of methanol-acetonitrile (50:50, v/v) a 1 M aqueous solution of ammonium acetate in a 999:1 (v/v) ratio. After appropriate mixing of both elution solvents (20% of A, 80% of B), the resulting mobile phase was filtered through a Nylon 66 (0.2µm pore size) filter and degassed. The standard of despropionyl bezitramide was a gift from Janssen Pharmaceutica (Beerse, Belgium). A stock solution of 0.1 mg/mL was prepared in methanol. Dilution of the stock solution with methanol yielded the working solutions at concentrations of 1.0, 0.5, 0.2, 0.1, and 0.02 µg/mL. The internal standard (Nmethyldespropionyl bezitramide) was synthesized following our earlier described procedure.4 An internal standard working solution at a concentration of 0.2 µg/mL was obtained after appropriate dilution of the stock solution with methanol. All stock and working solutions were stored in the refrigerator (6 °C) and were stable for at least 6 months.
Samples. The samples (blood, urine, stomach contents, bile, liver, and kidney) were from three fatalities that were due to the combined intake of several drugs, including bezitramide. The samples were frozen at -20 °C immediately after receipt. They were thawed just prior to extraction and analysis. Isolation of the Compounds. Preanalysis. The concentrations of despropionyl bezitramide in tissues (liver, kidney), stomach contents, and bile are usually much higher than the linear range of the fluorescence detector (0-50 ng/mL or ng/g). Hence, the samples need to be diluted in order to not exceed the linear range of the detector. The appropriate dilution factor is established by performing a preanalysis. The samples are extracted undiluted as previously described,4 following which the dry extract is reconstituted in 400 µL of a 50:50 (v/v) mixture of the elution solvents A and B. Serial dilutions of the reconstituted extract are injected onto the LC-FL until the signal for the despropionyl bezitramide peak falls within the linear range of the fluorescence detector. Once the appopriate dilution factor is known, each sample is reanalyzed as described below. Blood and Tissue Homogenates. To 2 mL of blood or 2 g of the appropriate dilution of tissue homogenate (liver and kidney) was added 100 µL of the internal standard working solution, 1 mL of 1 M NaOH, and 3 mL of water. After the sample was vortexmixed for 30 s, it was extracted by rotation for 10 min with 6 mL of n-hexane:isoamyl alcohol (93:7, v/v), followed by a 15-min centrifugation step. The organic layer was transferred to another extraction tube containing 2 mL of 0.05 M sulfuric acid. After extraction (10 min) and centrifugation (5 min), the organic phase was rejected. The aqueous phase was made alkaline with 300 µL of concentrated ammonia and extracted again for 10 min with 4 mL of the organic solvent mixture. Finally, the organic layer was transferred to a glass test tube and evaporated to dryness at 40 °C under a gentle stream of nitrogen. The residue was redissolved in 200 µL of a 50:50 (v/v) mixture of the elution solvents A and B and divided in two parts of 100 µL each. One part was transferred to a sealed reaction vial suitable for analysis by the autosampler of the LC-ESI-MS system. To the other part was added 100 µL of the 50:50 (v/v) mixture of the elution solvents A and B, followed by analysis with the LC-FL system. Urine, Stomach Contents, and Bile. To 2 mL of urine or 2 mL of the appropriate dilution of stomach contents and bile was added 100 µL of the internal standard working solution and 1 mL of 1 M NaOH. After the sample was vortex-mixed, it was analyzed following the same extraction procedure as that for blood. Calibration Samples. Samples for the calibration curves were prepared by spiking blank blood or urine from healthy individuals not taking any medication with bezitramide. The addition of 100 µL of each working standard resulted in despropionyl bezitramide concentrations of 1, 5, 10, 25, and 50 ng/mL. For the high calibration standards, 100 and 150 ng/mL, 200 and 300 µL of the 1 µg/mL working standard were added, respectively. The calibration curves for the LC-FL and the LC-ESI-MS/MS methods ranged from 0 to 50 ng/mL and from 0 to 150 ng/mL, respectively. The calibration samples were treated in a way similar to the unknown samples’ treatment. Chromatography. One-hundred and 50-µL aliquots of each reconstituted extract were injected onto the LC-FL (manual injection) and the LC-ESI-MS (autosampler injection) systems, Analytical Chemistry, Vol. 71, No. 14, July 15, 1999
2909
Figure 1. LC-FL (A) and LC-ESI-MS/MS (B) chromatogram of the blood sample of subject one (despropionyl bezitramide concentration, 106 ng/mL).
respectively. Chromatographic separation was achieved on a Hypersil ODS column (150 × 4.6 mm i.d., 5-µm particle size) from Alltech Europe (Laarne, Belgium). The mobile phase consisted of a 20/80 (v/v) mixture of elution solvents A and B. The flow rate was 0.5 mL/min. Under these conditions, despropionyl bezitramide and the internal standard eluted within 10 min on both HPLC systems (Figure 1A, B). However, to avoid the interference of late-eluting peaks in a following run, the next sample was only injected after 30 min using the LC-FL system. In the case of the LC-ESI-MS/MS system, the column was rinsed with methanol for 3 min after an isocratic run of 10 min. The pump was programmed to regain its initial conditions within 1 min, and a 3-min reconditioning time was allowed. Hence, the total run time was reduced from 30 min in the LC-FL method to 17 min in the LC-ESI-MS. Detection. LC-FL. The excitation and emission wavelengths of the fluorescence detector were fixed at 280 and 310 nm, respectively (bandwidth was 10 nm for both excitation and emission slits). LC-ESI-MS/MS. Low-energy Collisional-Activated Dissociation (CAD) Product Ion Spectra. The tube lens and capillary voltages were optimized for maximum despropionyl bezitramide and internal-standard signal. Tuning was performed using 100-µL loop injections of a standard mixture of each analyte in the mobile phase (respective despropionyl bezitramide and internal standard 2910 Analytical Chemistry, Vol. 71, No. 14, July 15, 1999
concentrations: 10 and 1 µg/mL). Optimal conditions were the following: ESI capillary voltage 4.05 kV, HV lens voltage 0.2 kV, cone voltage 45 V, skimmer offset and skimmer 5 and 1.5 V, respectively, and source temperature 80 °C. Nitrogen was used as nebulizer gas at a flow rate of 20 L/h, and ultrapure nitrogen was used as curtain gas at 250 L/h. The MS was operated in the MS-MS mode with a collision energy of 30 eV and a 3 × 10-3 mbar Argon collision gas pressure. Q1 was set to monitor ions at m/z 437.2 and 451.2. Scanning in Q3 was performed over a mass range from m/z 20 to 500 in 1 s. Selected Reaction Monitoring (SRM). The source parameters, collision gas pressure, and collision energy were the same as described for the collection of the CAD product-ion spectra. Q1 and Q3 were set to monitor ions at m/z 437.2 and 451.2, and 111.2, respectively. Method Validation. As sample pretreatment and chromatographic conditions were slightly modified compared with the LCFL method recently described,4 the validation of some parameters proved to be necessary. The decision regarding which parameters require validation was based on logical consideration of the specific validation parameters likely to be affected by the changes.5 The following parameters were validated: linearity, precision, and accuracy. Linearity. Standard curves were prepared over a concentration range of 1-150 ng/mL for despropionyl bezitramide in blood and urine. For each curve, eight different concentration levels were used. Peak-area ratios of despropionyl bezitramide and the internal standard were plotted against the corresponding despropionyl bezitramide concentration. Weighted linear regression was used in an effort to compensate for data heteroscedasticy. Precision. Precision was evaluated by analyzing blank blood samples spiked with despropionyl bezitramide at two different concentrations (10 and 50 ng/mL) on the same day (five replicates, repeatability) and over 5 consecutive days (reproducibility). Accuracy. To assess the accuracy of the method, two positive control samples were independently prepared, using volumetric material different from that employed for the preparation of the calibration samples. They were both extracted and analyzed with each batch of samples, and their quantitative results were compared to the spiked concentrations (10 and 50 ng/mL). Limit of Detection, Quantification. Conclusions about the limit of detection (LOD) and the limit of quantification (LOQ) of the method were drawn by inspecting the chromatograms obtained after the analysis of the lowest calibration sample (despropionyl bezitramide concentration: 1 ng/mL). Specificity. The specificty of the method with regard to the interference of endogenous compounds was demonstrated by analyzing blank blood and urine samples. Conclusions about the specificity of the method in relation to other therapeutic and abused drugs could be drawn by relying on data from the literature.4,6,7 (5) Bressolle, F.; Bromet-Petit, M.; Audran, M. J. Chromatogr., B 1996, 686, 3-10. (6) Pfleger, K.; Maurer H. H.; Weber, A. Mass spectral and GC data; VCH: Weinheim, Germany, 1992; pp 829, 1254, 1259, 1455. (7) Ardrey, R. E.; Allan, A. R.; Bal, T. S.; Joyce, J. R.; and Moffat, A. C. Pharmaceutical Mass Spectra; The Pharmaceutical Press: London, U.K., 1985; p 87.
Figure 2. Full-scan ESI mass spectrum of despropionyl bezitramide.
Safety Considerations. General guidelines for work with organic solvents, acids, and bases were respected. All operations with diazomethane (preparation of the reagent, methylation of despropionyl bezitramide) were performed under a fume hood and in a well-ventilated room. Blood and urine and other samples of human origin were handled as potentially infectious. RESULTS AND DISCUSSION Isolation of the Compounds. We prepared 2 mL or 2 g of (the appropriate dilution) of each sample matrix compared with 1 g in the procedure described by De Baere et al.4 Hence, it was possible to divide the final reconstituted extract in two equal portions which were equivalent with 1 mL or 1 g of the original sample. Each portion was then analyzed using the LC-FL or the LC-ESI-MS/MS system. By comparing the obtained quantitative results, conclusions about the compliance between or the advantages and/or disadvantages of both analytical techniques could be drawn. Chromatography. The methodological crossover between LCFL detection and LC-ESI-MS/MS was simple since only minor modifications proved necessary.4 The same column was used, and the samples could be injected on both systems after one common extraction. The mobile-phase composition and flow rate were changed slightly; in order to prevent contamination of the API chamber8 the ammonium acetate buffer concentration was lowered from 0.1 to 0.001 M. This could be performed without deterioration of peak symmetry. For LC-FL detection a mobile-phase flow rate of 1 mL/min was initially used. As this flow rate is not compatible with LC-ESI MS/MS, splitting of the column effluent is necessary. Since splitting 40:1 would negatively affect the sensitivity of the method, we decided to use a 20:1 split and to lower the flow rate from 1.0 to 0.5 mL/min, while increasing the mobile phase strength (A:B, v/v from 31:69 to 20:80). These changes resulted in a lower capacity factor for both despropionyl bezitramide (from 3.65 to 2.11) and the internal standard (from 5.47 to 2.73) and in a smaller, but still acceptable, resolution between both compounds (from 1.50 to 1.29). The peak shape, however, was not affected (asymmetry factors for despropionyl bezitramide and the internal standard: 1.0 and 1.1, respectively). Mass-Spectrometry. Identification. The full-scan ESI mass spectra are devoid of much structural information since they contain a strong [M + H]+ signal (at m/z values of 437 and 451 for despropionyl bezitramide and the internal standard, N-methyl(8) Voyksner, R. Electrospray Ionization Mass Spectrometry; John Wiley & Sons, Inc: New York, 1997; p 324-327.
despropionyl bezitramide, respectively) and drastically reduced signals from fragment ions (Figure 2). However, tandem MS (daughter-ion scanning mode) affords product ion spectra that contain ample information for identification of the analyte in question. In Figure 3 A, B the low-energy CAD product-ion spectra of the despropionyl bezitramide and the internal standard peak in the chromatogram of a spiked blood sample (analyte concentration: 50 and 20 ng/mL) are shown, respectively. As can be seen, the main product ions at m/z values of 83, 111, 134 (IS: 148), 173 (IS: 187), 192, 201 (IS: 215), 217 (IS: 231), and 303, together with the [M + H]+ ions, confirm the structure of both compounds (Figure 4). Mass spectral data were obtained, not only for standards, but equally for spiked calibration samples (Figure 3 A,B) and, more importantly, also for all the forensic samples of real drug abusers. By comparing the mass spectrum of the unknown peak in the chromatogram of a real sample (Figure 3C) with that of the despropionyl bezitramide peak in a calibration sample (Figure 3A), bezitramide abuse could be established. Hence we can say that LC-MS inevitably affords much more analytical information compared with LC-FL. Nevertheless, in the forensic samples investigated, the qualitative results obtained with both techniques corresponded almost perfectly. None of the peaks in the LC-FL chromatograms were erroneously identified, substantiating the worthiness of this LC-FL technique. Quantitation. For our quantitative purposes, we preferred to operate in the selected reaction monitoring (SRM) mode, the equivalent of selected ion monitoring (SIM) in single-stage MS.9,10 Using this technique, an abundant analyte ion (m/z 437 and 451 for despropionyl bezitramide and the internal standard, respectively) is selectively transmitted through Q1 and collisionally dissociated to product ions in Q2. Then only selected product ions (m/z 111 for both despropionyl bezitramide and the internal standard) are scanned in Q3. This process greatly reduces the chemical noise reaching the detector and, therefore, results in a higher sensitivity than that achievable by single-stage mass analysis. In addition, operating in the SRM mode allies excellent sensitivity with high specificity, as only ions which correspond to a specific fragmentation route are monitored. Method Validation. Linearity. The calibration curves for despropionyl bezitramide in blood and urine were linear over the (9) Hoja, H.; Marquet, P.; Verneuil, B.; Hayat, L.; Pe´nicaut, B.; Lachaˆtre, G. J. Anal. Toxicol. 1997, 21, 116-126. (10) Moody, D. E.; Laycock, J. D.; Spanbauer, A. C.; Crouch, D. J.; Foltz, R. L.; Josephs, J. J.; Amass, L.; Bickel, W. K. J. Anal. Toxicol. 1997, 21, 406-414.
Analytical Chemistry, Vol. 71, No. 14, July 15, 1999
2911
Figure 3. Low-energy CAD product-ion spectra of the protonated molecules of despropionyl bezitramide (A), the internal standard (B), and the unknown peak in the chromatogram of the blood sample of subject one (despropionyl bezitramide concentration, 106 ng/mL) (C).
range tested (1-150 ng/mL). Correlation coefficients of 0.9994 or higher were obtained for the relationship between the peakarea ratios (despropionyl bezitramide/internal standard) and the corresponding calibration concentrations. For both calibration curves, y-intercepts (b) which were virtually zero and similar regression slopes (a) were observed (b ) 0.0180 and 0.0100, a ) 0.0333 and 0.0349 for blood and urine, respectively). Compared with the previously described LC-FL method,4 our LC-MS method shows linearity over a much broader concentration range. Hence, tissue samples which usually contain high concentrations of despropionyl bezitramide (Table 1), do not need such a drastic dilution step prior to LC-MS as they do prior to LC-FL analysis. Precision and Accuracy. Precision and accuracy were thoroughly investigated for the analysis of blood. Table 2 presents the repeatability and reproducibility data obtained (n ) 5) for the different concentrations tested. The coefficients of variation (CV%) ranged from 1.6 to a maximum of 10.2%. As all values for coefficients of variation were within 10%, the precision of the method can be considered acceptable11 and undoubtedly meets with our objective of a routinely applicable analysis for despropionyl bezitramide in forensic (blood) samples. (11) Mehta, A. J. Clin. Pharm. Ther. 1989, 14, 465-473.
2912 Analytical Chemistry, Vol. 71, No. 14, July 15, 1999
The accuracy for despropionyl bezitramide was determined at two separate concentration levels. At 10 ng/mL and 50 ng/mL levels (n ) 5) values of 93.4 and 91.8% were found, respectively. Furthermore, two urine quality control samples (concentration levels of 10 and 50 ng/mL) were also analyzed. For these samples, concentrations of 10.1 and 48.5 ng/mL were found, showing percents of target of within 5%. Limit of Detection and Quantification. Inspection of the despropionyl bezitramide peak in the chromatograms of the lowest calibration samples (concentration in blood and urine: 1 ng/mL) reveals S/N ratios of ∼20. Hence, we can conclude that, using the LC-MS technique, despropionyl bezitramide can be quantified at concentrations below 1 ng/mL, which is the LOQ of the previously described LC-FL method.4 Consequently, analogous conclusions can be drawn regarding the LODs of both methods. Specificity. The method proved to be specific for despropionyl bezitramide and the internal standard. There were no interfering peaks from endogenous compounds at the elution positions of both compounds. The method can also be considered specific with regard to other therapeutic and abused drugs. As earlier described,4 a few antidepressive drugs (butriptyline, trimipramine, nortriptyline, mianserine, and lofepramine) eluted within 0.5 min
Table 1. Despropionyl Bezitramide Concentrations in Postmortem Samples from Patients Suspected of Bezitramide Abuse other drugs demonstrated
subject
gender
age
1
M
28
cocaine, opiates, benzodiazepines
2
M
33
opiates, barbiturates, benzodiazepines, tricyclic antidepressants
3
F
34
barbiturates, hydroxyzine
matrix
LC-MS result
LC-FL result
blood urine stomach contents bile liver kidney blood urine stomach contents liver kidney liver kidney
106.1 ng/mL 2.5 ng/mL 5.3 µg/mL 621.5 ng/mL 2.0 µg/g 1.7 µg/g 5.6 ng/mL < LOQ 3.5 µg/g 297.0 ng/g 203.0 ng/g 2.5 µg/g 274.0 ng/g
108.5 ng/mL 1.7 ng/mL 5.1 µg/mL 616.0 ng/mL 1.2 µg/g 1.0 µg/g 5.7 ng/mL < LOQ 3.5 µg/g 291.0 ng/g 136.2 ng/g 1.0 µg/g 160.2 ng/g
Table 2. Precision and Accuracy Evaluation for the Analysis of Despropionyl Bezitramide Using LC-ESI-MS/MS concn added (ng/mL)
concn found (ng/mL)
10
9.3 8.8
50
45.4 47.4
Figure 4. Suggested fragmentation pattern of [M + H]+ of despropionyl bezitramide and the internal standard under low-energy CAD conditions.
of the eluting position of despropionyl bezitramide. LC-MS experiments with these antidepressants show, however, that their fragmentation pattern differs from that of despropionyl bezitramide. Since these compounds do not show ions with m/z values of 111 (quantifier ion) or 83 and 192 (possible qualifier ions), it can be concluded that they will not interfere during the described LC-MS/MS analysis.6,7 Analysis of Real Samples. To evaluate the applicability of the method, postmortem samples (blood, urine, stomach contents, bile, liver, and kidney) from three fatalities that were due to the combined intake of several drugs, including bezitramide, were analyzed. In addition, all postmortem samples were analyzed using the previously described LC-FL technique. The results of the sample analyses are presented in Table 1. As can be seen, despropionyl bezitramide concentrations were lowest in urine. This could be expected, since the cumulative output of bezitramide
6.4 6.6 10.2 12.0 1.6 9.2 7.4 5.1
repeatability (CV%, n ) 5) accuracy (%) reproducibility (CV%, n ) 5) accuracy (%) repeatability (CV%, n ) 5) accuracy (%) reproducibility (CV%, n ) 5) accuracy (%)
and despropionyl bezitramide over 48 h after oral administration varies from 0.1 to 0.3% of the administered dose.1 The quantitative results obtained for the urine samples also demonstrate that the LC-ESI-MS/MS is at least as sensitive as the LC-FL technique. Moreover, a good correlation between both techniques is seen for this matrix. In the blood sample of subject one, a concentration of 106108 ng/mL despropionyl bezitramide is found. After a single oral dose of 5 mg bezitramide maximal plasma concentrations of 4.28.2 ng/mL are reached. Since the recommended maximal daily intake of the drug amounts to 30 mg, the concentration found suggests the previous intake of supratherapeutic amounts of bezitramide.1 The analysis of the other matrixes from subjects 1 to 3 revealed the presence of much higher despropionyl bezitramide concentrations as compared with those of blood and urine. The highest concentrations were found in stomach contents and liver tissues (Table 1). As a result of the limited linear range (1-50 ng/mL) of the fluorescence detection technique,4 these samples had to be diluted appropriately prior to extraction and analysis. Using the LC-MS technique, a less drastic dilution of the samples proved necessary, since the linear range of the MS detector was at least 3 times higher (g1-150 ng/mL) compared with that of the FL detector. As can be seen from Table 1, a good correlation between the LC-MS and the LC-FL technique is obtained for the analyses of blood, urine, stomach contents, and bile. However, for the liver and kidney samples, much lower concentrations are found using the LC-FL technique. This phenomenon can possibly be explained Analytical Chemistry, Vol. 71, No. 14, July 15, 1999
2913
by the inner filter effect of the fluorescence detector. Because of this effect, which occurs if concentrated or even “dirty” samples such as liver and kidney tissues are analyzed, the fluorescence emission is reduced or “quenched”. An analogous phenomenon is not expected for the LC-ESI-MS/MS technique since, by using this instrument in the SRM mode, a high degree of selectivity with respect to endogenous interferences is reached. Hence, these results demonstrate the extreme usefulness of the LC-ESI-MS/ MS in the field of forensic toxicology. CONCLUSIONS We succeeded in a simple methodological crossover between a previously described LC-FL method and a LC method using ESIMS/MS detection. Using the LC-MS/MS in the SCAN mode gave us the opportunity to make a complementary identification of the analyte in question on the basis of its unique MS/MS spectrum. Moreover, for our quantitative purposes, a SRM mass spectrometric technique was applied. The method was (re)validated with respect to linearity, precision, and accuracy, and good results were obtained for blood and urine, which were the matrixes investigated.
2914 Analytical Chemistry, Vol. 71, No. 14, July 15, 1999
Postmortem samples from three fatalities, all suspected drug abusers, were quantitatively analyzed using both techniques. From the obtained results it is clear that the LC-MS/MS technique is as sensitive as the LC-FL method. In addition, a good correlation between both techniques is demonstrated for the analysis of matrixes such as blood, urine, stomach contents, and bile. For the analysis of “dirty” samples such as liver and kidney tissues, the LC-MS/MS is not subject to a “quenching” effect and hence proves to be, with respect to specificity, the superior of both techniques. In conclusion, this report demonstrates that LC-MS is a technique which is fully suitable for application in the field of toxicology. ACKNOWLEDGMENT The authors thank W. Van Dongen for assistance with the experiments. Received for review November 30, 1998. Accepted April 26, 1999. AC981310R
