Anal. Chem. 2006, 78, 416-423
Voltammetric Behavior of Nitrazepam and Its Determination in Serum Using Liquid Chromatography with Redox Mode Dual-Electrode Detection Kevin C. Honeychurch, Gemma C. Smith, and John P. Hart* Centre for Research in Analytical, Materials and Sensors Science, Faculty of Applied Sciences, University of the West of England, Bristol, Frenchay Campus, Coldharbour Lane, Bristol, BS16 1QY, U.K.
A method involving high-performance liquid chromatography with dual-electrode electrochemical detection in the redox mode (LC-DED) has been successfully developed for the determination of the benzodiazepine tranquilizer, nitrazepam, in serum. To elucidate the electrochemical mechanism occurring at a glassy carbon electrode, cyclic voltammetry was preformed with 1 mM solutions of nitrazepam at pH values between 2 and 12, using a potential range from -1.5 to +1.5 V. Two reduction peaks were observed over the whole pH range; the first, designated R1, was consistent with the 4e-, 4H+ reduction of the 7-nitro group to a hydroxylamine species; the second more negative peak, designated R2, was shown to be the result of a 2e-, 2H+ reduction of the 4-5 azomethine group. On the reverse anodic scan, an oxidation signal was observed, designated O1, which was considered to result from a 2e-, 2H+ oxidation of the hydroxylamine to a nitroso group. On the second forward scan, a new reduction peak, designated R3, was observed, which was considered to result from reduction of the nitroso species back to the hydroxylamine species. Studies were then undertaken to exploit the hydroxylamine/nitroso redox couple using LC-DED detection for the measurement of nitrazepam in serum. The optimal chromatographic conditions were found to comprise a mobile phase containing 60% methanol, 40% 50 mM pH 4.1 acetate buffer, in conjunction with a Hypersil C18 250 mm × 4.6 mm column. Hydrodynamic voltammetric studies were undertaken to optimize the operating potentials required for dual-electrode detection. It was found that an applied potential of -2.4 V was optimum for the “generator” cell and +0.5 V for the “detector” cell. The proposed method was evaluated by carrying out replicate nitrazepam determinations on spiked bovine and human serum samples. The former evaluation was preformed at a concentration of 11.2 µg mL-1, and the latter at 1670 ng mL-1. For bovine serum, the recovery of nitrazepam was found to be 75.8% and the associated coefficient of variation was 6.1% (n ) 6). For human serum, the recovery was 74.1% with a coefficient of variation of 7.8% (n ) 7). These data suggest that the method holds promise for applications in toxicology and where an alternative reliable method to confirm drug abuse may be required. 416 Analytical Chemistry, Vol. 78, No. 2, January 15, 2006
Nitrazepam (I) (Mogadon), a member of the 1,4-benzodiaz-
epine class (1,3-dihydro-7-nitro-5-phenyl-2H-1,4-benzodiazepin-2one) of tranquilizers, is commonly prescribed for several stressrelated disorders.1 In recent years, it has become a drug of abuse and has been implicated in cases of drug-facilitated assault. Therefore, reliable, sensitive analytical procedures are desired for the trace determination of this drug in biological fluids. One possible approach, which we considered worthy of investigation, is liquid chromatography with dual-electrode electrochemical detection (LC-DED) in the redox mode.2 To our knowledge, this powerful analytical technique has not been exploited for the determination of any of the benzodiazepines. In this technique, two electrochemical cells are arranged in series after the column; the upstream cell is used as the “generator” cell to produce an electroactive species, which is detected at the downstream “detector” cell. The advantage of this approach is that, for a number of species, the electrochemically generated product is much more easily oxidized or reduced than the parent compound;2 this improves the selectivity of the system, and as the detector potential is operated at lower values, the sensitivity is improved. This latter improvement occurs as a result of the lower background currents occurring at the detector electrode compared with those obtained at the high potentials required for direct electrochemical detection. Hart et al.3 have shown these properties of the DED to greatly improve the sensitivity and selectivity for the HPLC assay of the naphthoquinone compound, vitamin K1, in human serum compared to that obtained by HPLC using amperometric detection in the reductive mode.4 * Corresponding author. Tel.: +44 117 3282469. Fax: +44 117 3282904. E-mail:
[email protected]. (1) Rang, H. P.; Dale, M. M.; Ritter, J. M.; Moore, P. K. In Pharmacology, 5th ed.; Hunter, L., Ed.; Churchill Livingstone: London, 2003; pp 515-522. (2) Hart, J. P. Electroanalysis of Biologically Important Compounds; Ellis Horwood: London, 1990. (3) Hart, J. P.; Shearer, M. J.; McCarthy, P. T. Analyst 1985, 110, 1181-1184. 10.1021/ac058035a CCC: $33.50
© 2006 American Chemical Society Published on Web 12/13/2005
Nitrazepam contains an aromatic nitro group, which is known to undergo reduction to a hydroxylamine species at mercury electrodes.5 We considered that this species could be produced in our “generator” cell; then the hydroxylamine could be reoxidized at the “detector” cell to give the analytical response. It should be mentioned that we have demonstrated the principle of utilizing the reoxidation of a hydroxylamine for the analytical response in a previous study. In that study, a planar screen-printed carbon electrode was used in conjunction with stripping voltammetry for the determination of 2,6-dinitrotoluene.6 Our present study was divided into three parts. Initially, we investigated the voltammetric behavior of nitrazepam at a glassy carbon electrode (GCE). Next, we optimized the HPLC conditions needed for the separation on a reversed-phase column and then optimized the conditions required at the “generator” cell and “detector” cell using hydrodynamic voltammetry. Finally, we investigated the possibility of using the optimized system for the determination of nitrazepam in serum samples. EXPERIMENTAL SECTION Chemicals and Reagents. All chemicals were obtained from Fisher (Loughborough, U.K.), unless otherwise stated. Deionized water was obtained from a Purite RO200, Stillplus HP system (Purite Oxon.). Stock solutions of disodium, trisodium, and sodium orthophosphate were made at a concentration of 0.2 M by dissolving the appropriate mass of solid in deionized water. A stock solution of orthophosphoric acid was made by dilution in deionized water to give a concentration of 0.2 M. These were then titrated, to give the desired pH. A primary stock solution of nitrazepam, obtained from Sigma-Aldrich (Dorset, U.K.), was prepared by dissolving the required mass in either ethanol for cyclic voltammetric (CV) studies or methanol for HPLC investigations, both at a concentration of 10 mM. Working standards, for initial voltammetric studies, were prepared by dilution of the primary stock solution with phosphate buffer to give a final concentration of 100 mM phosphate buffer. These were then adjusted with sufficient water to give a 10% ethanol solution. Standards for HPLC analysis were made by dilution of the primary methanol stock solution in the mobile phase. A 50 mM acetate pH 4.1 buffer was prepared by titration of a solution of 50 mM sodium acetate with a 50 mM acetic acid solution. Hydrogen peroxide measurements were undertaken using cobalt-phthalocyanine-modified screenprinted carbon electrodes, supplied by Gwent Electronic Materials Ltd. (GEM, Mamhilad, Gwent, U.K.). Human serum was obtained from BioSera Ltd. (East Sussex, U.K.). Bovine serum, paracetamol, ascorbic acid, and caffeine were obtained from Sigma-Aldrich. 2-Amino-5-nitrobenzophenone was obtained from Acros Organics (Geel, Belgium). Apparatus and Instrumentation. (1) Cyclic Voltammetry. CV was performed with an EG&G Princeton Applied Research (Princeton, NJ) model 263 potentiostat connected to a PC with EG&G Echem electrochemistry software. The voltammetric cell (Metrohm) contained a glass-coated platinum wire auxiliary (4) Hart, J. P.; Shearer, M. J.; McCarthy, P. T.; Rahim, S. Analyst 1984, 109, 477-481. (5) Franklin Smyth, W. Voltammetric Determination of Molecules of Biological Significance; Wiley: Chichester. 1992. (6) Honeychurch, K. C.; Hart, J. P.; Pritchard, P. R. J.; Hawkins, S. J.; Ratcliffe, N. M. Biosen. Bioelectron. 2003, 19, 305-312.
electrode, a saturated calomel electrode (Russell, Fife, U.K.), and a GCE (6-mm diameter), as the working electrode. (2) High-Performance Liquid Chromatography. HPLC studies were undertaken using a system consisting of an IsoChrom pump (Spectra Physics), with a 250 mm × 4.6 mm Hypersil Hypurity C18, 5-µm column connected to a 7125 valve manual injector fitted with a 10- or 50-µL sample loop (Rheodyne, Cotati, CA). Sample extracts were analyzed using a mobile phase consisting of 60% methanol (Fischer, far UV, HPLC grade), 40% 50 mM pH 4.1 acetate buffer, at a flow rate of 1.0 mL min-1. Initial HPLC investigations were undertaken using an Agilent 1100 HPLC system with detection at 254 nm. (3) Dual-Electrode Detection. Both the generator and detector cells were obtained from BAS (Congleton, Cheshire, U.K.). The generator cell consisted of a two-piece thin-layer cell, formed from an upper Teflon block containing the GCE (3-mm diameter), and a bottom steel block serving as the pseudoreference/counter electrode. The analytical detector cell consisted of a Teflon two-piece (top and base) thin-layer cell. The detector cell operated as a three-electrode system comprising a GCE (3-mm diameter), a stainless steel counter electrode, and a Ag/AgCl reference electrode (BAS). Teflon gaskets were purchased from BAS. An EG&G Princeton Applied Research (Princeton, NJ) model 362 scanning potentiostat was used to control the potential at the generator cell at -2.4 V versus the pseudoreference/counter steel electrode. The potential at the detector cell was held at +0.5 V versus Ag/AgCl and the current monitored using a BAS LC-4B amperometric detector. Chromatograms were recorded using a Siemens Kompenosograph X-T C1012 chart recorder. Cyclic Voltammetric Studies. Cyclic voltammograms were initially recorded in plain solutions of 0.1 M phosphate buffer, containing 10% ethanol, and then in the same solution containing 1 mM nitrazepam. Degassing was achieved by purging with oxygen-free nitrogen (BOC, Guildford, U.K.) for 5 min to eliminate oxygen reduction waves. A starting potential of 0 V was used, with an initial switching potential of -1.5 V and a second switching potential of +1.5 V. Before each experiment, the GCE was polished manually with slurries prepared from 5-µm aluminum oxide on a smooth polishing mat. Residual polishing material was removed by rinsing with deionized water. Cyclic voltammetry was also performed with the flow cell containing 1 mM nitrazepam in 60% methanol, 40% 50 mM pH 4.1 acetate buffer. Hydrodynamic Voltammetry (HDV). HDV was performed by injecting fixed volumes of a standard solution of nitrazepam and varying the applied potential between -1.7 and -3.7 V, for the down stream generator cell, and between 0.0 and +2.0 V, for the upstream detector cell. HDVs were constructed by plotting the recorded peak current against the applied potential. The optimum potential was determined from the position of the plateau on the hydrodynamic wave. The possible direct oxidation and reduction of nitrazepam were investigated over the ranges 0 to +2.0 V and -1.7 to -3.8 V, respectively. Gas Chromatography/Mass Spectroscopy (GC/MS) and UV-Visible Spectrographic Procedures. GC/MS analysis was carried out using a Hewlett-Packard 5890 Series II Plus gas chromatograph coupled to a 5985B quadrupole MS system (Hewlett-Packard, Palo Alto, CA). Manual injections were made Analytical Chemistry, Vol. 78, No. 2, January 15, 2006
417
Figure 1. Cyclic voltammogram, obtained at a scan rate of 50 mV s-1, for a 1 mM solution of nitrazepam in 10% ethanol, buffered with 0.1 M phosphate at pH 4.1. Starting potential 0.0 V, initial switching potential -1.5 V, and second switching potential +1.5 V.
using a split-splitless technique on to a HP-5MS capillary column (15 m × 0.25 mm i.d., 0.25-µm film thickness, 5% diphenyl-95% dimethylsiloxane phase) interfaced directly into the ion source. The GC oven temperature was maintained at 50 °C for 3 min, then programmed to 250 °C at 12 °C min-1, and finally held isothermally for 10 min at this temperature. The injector and transfer lines were at 330 °C. Source and quadrupole temperatures were 200 and 100 °C, respectively. UV Spectra were obtained on a Perkin-Elmer Lambda 40 UVvisible spectrometer using a Perkin UV Winlab program for instrument control and data processing. Sample Pretreatment Procedure. A 250 µl aliquot of serum was added to a glass vial followed by 500 µL of acetone. The resulting solution was then passed through a 2 µm PTFE syringe filter (Gelman Laboratory, Acrodisc CR 13 mm syringe filter). The filter was then washed with 200 µL of acetone. The combined filtrates were then blown down to dryness under nitrogen. The resulting residue was then reconstituted in either 500 (bovine serum) or 200 µL (human serum) of mobile phase (60% methanol, 40% 50 mM pH 4.1 acetate buffer), and syringe filtered again. Aliquots of the filtrate were then examined by LC-DED. RESULTS AND DISCUSSION Cyclic Voltammetric Investigations. We began our investigations by studying the cyclic voltammetric behavior of nitrazepam at a planar GCE over the pH range 2-12. Figure 1 shows a typical cyclic voltammogram obtained at pH 4.1. On the first forward negative scan, two reduction peaks, R1 and R2, appear at potentials of -0.8 and -1.1 V, respectively; on the first positive reverse scan, a single oxidation peak (O1) appears at a potential of 0 V. On the 418 Analytical Chemistry, Vol. 78, No. 2, January 15, 2006
Figure 2. Proposed mechanism for the electrochemical behavior of nitrazepam.
second negative scan, a new cathodic peak R3 appears at a potential of -0.1 V. We believe that the reduction process giving rise to R1 is the formation of a hydroxylamine species II and that R2 is a result of the reduction of the azomethine moiety to give species III (Figure 2). The relative heights of R1 and R2 are in agreement with these findings and concur with the behavior reported using a mercury electrode.5 The anodic peak O1 is considered to be the result of oxidation of the hydroxylamine group present in the reduced species III to produce IV; the resulting nitroso group IV undergoes a quasireversible reduction reaction as the second negative going scan shows a cathodic peak (R3) at -0.1 V. It should be mentioned that, if the negative scan is switched at -0.8 V, i.e., before the reduction process producing R2, then the quasi-reversible redox couple (O1/R3) is still seen, but this is now considered to result from the couple II and V. It should be mentioned that the potential values of O1 would strongly suggest that it is the azomethine peak giving rise to R2 and not the hydroxylamine to an amine. The latter would be expected to oxidize at higher potentials than found here and result in an irreversible process.5 Figure 3 shows the ip versus pH and Ep versus pH plots obtained for the oxidation peak O1 over the range studied. A pH of 4 appeared to offer a good compromise between sensitivity and selectivity, and compatibility with reversed-phase chromatography, bearing in mind the possible degradation of stationary phase at pH >7 (discussed later).
Figure 3. Plot of (a) ip and (b) Ep vs pH for nitrazepam peak O1. Voltammetric conditions as Figure 1.
The effect of scan rate was investigated for each of the principle peaks (R1, R2, O1). Between pH 2 and 8, all the peaks were found to be diffusion controlled, i.e., ip was linearly dependent on the square root of scan rate (v1/2). Above pH 8, the solutions were found to change from clear to yellow upon pH adjustment. This we believe is due to the base hydrolysis of nitrazepam to give 2-amino-5-nitrobenzophenone, as shown by Davidsons et al.7 To investigate this further, we undertook several studies using UV spectroscopy and GC/MS. An aliquot of nitrazepam was initially adjusted to pH 10 to obtain the yellow solution. This was then adjusted to pH 2 and extracted with dichloromethane (DCM). Examination of the extract by UVvisible spectroscopy revealed an adsorption maximum of 360 nm, compared to 260 nm for nitrazepam in DCM. GC/MS analysis showed this extract to be predominantly 2-amino-5-nitrobenzophenone. Liquid Chromatography of Nitrazepam Using a ReversedPhase Column. The variation of chromatographic capacity factor (k′) of nitrazepam with the percentage of methanol in the mobile phase containing acetate buffer pH 4.0 is shown in Figure 4. The optimum percentage of methanol was found to be 60% (v/v), as this gave a retention time of only 6.4 min (k′ ) 2.2) without compromising the chromatographic performance. Therefore, a mobile phase of 60% methanol, 40% 50 mM pH 4.1 acetate buffer was used in further studies. Cyclic Voltammetry and Hydrodynamic Voltammetry in Mobile Phase. To ascertain whether the optimized mobile phase was suitable for the electrochemical determination of nitrazepam, we undertook a second cyclic voltammetric investigation with a 1 mM solution in this medium. Figure 5a shows the cyclic voltammogram obtained with the same three-electrode cell used in the previous investigations. Clearly, the voltammogram demonstrates the same characteristics as that seen in the previous study (Figure 1). It should be mentioned that a steel pseudoreference/counter electrode was to be used in our LC-DED system; therefore, we undertook a further cyclic voltammetric study on the same solution. The resulting cyclic voltammogram is shown in Figure 5b. All three principle peaks are still observable, with the Eps of R1 and R2 shifted negatively by ∼1.0 V. The peak potentials did (7) Davidsons, A. G.; Chee, S.-M.; Millar, F. M.; Watson, D. Int. J. Pharm. 1990, 63, 29-34.
Figure 4. Variation of capacity factor (k′) with percentage of methanol in mobile phase.
Figure 5. Cyclic voltammograms obtained at a scan rate of 100 mV s-1, for a 1 mM solution of nitrazepam in 60% methanol, 40% 50 mM pH 4.1 acetate buffer. Voltammetric conditions: (a) Starting potential 0.0 V, initial switching potential -1.5 V, and second switching potential +1.0 V. (b) Starting potential 0.0 V, initial switching potential -3.0 V, and second switching potential +1.0 V. (Note that potentials in (b) are versus the stainless steel counter/pseudoreference electrode).
not shift when the same solution was repeatedly subjected to cyclic voltammetry using the cell containing the stainless steel electrode; therefore, we considered that this would provide a stable halfcell. Hydrodynamic Voltammetry. Figure 6 shows the cathodic HDV obtained with the generator cell over the potential range -1.7 to -3.2 V. The HDV exhibits two waves with E1/2 values of -1.95 and -2.7 V. These are similar to the Ep values obtained using cyclic voltammetry in quiescent mobile phase (Figure 5b) using the same cell. Consequently, the mechanism of reduction Analytical Chemistry, Vol. 78, No. 2, January 15, 2006
419
Figure 6. Cathodic hydrodynamic voltammogram for 1.4 µg injections of nitrazepam.
Figure 7. Anodic hydrodynamic voltammogram for 1.4-µg injections of nitrazepam. (a) Generator cell held at -2.4 V; (b) direct oxidation, generator cell switched off.
is considered to be the same under hydrodynamic conditions as obtained under quiescent conditions. Figure 7a shows the anodic HDV obtained with the detector cell while keeping the generator cell at a potential of -2.4 V. The current response at the latter cell was found to increase with increasing potential, until a value of +0.45 V was reached. Consequently, further studies were undertaken using an applied potential of -2.4 V at the generator cell and +0.5 V at the detector cell. Smyth and co-workers8-10 have postulated direct oxidation can occur for some benzodiazepines at the 1-N group. However, in the specific case of nitrazepam, the electron delocalization caused by the 7-NO2 was shown to stabilize the 1-N group against oxidation. As the main focus of this present study was to investigate the oxidation of the hydroxylamine, produced at the generator cell, we wanted to deduce whether the direct oxidation of nitrazepam could occur in the same potential range as that of the hydroxylamine. To investigate whether the response we (8) Ivaska, A.; Franklin Smyth, W. Anal. Chim. Acta 1980, 114, 283-291. (9) Franklin Smyth, W.; Burmicz, J. S.; Ivaska, A. Analyst 1982, 107, 10191025. (10) Franklin Smyth, W.; Ivaska, A. Analyst 1985, 110, 1377-1379.
420
Analytical Chemistry, Vol. 78, No. 2, January 15, 2006
obtained in the previous study was due to the direct oxidation of nitrazepam itself, we undertook a further HDV study at the detector cell, over the potential range 0 to +2.0 V, with the generator cell switched off. As can be seen from Figure 7b, the cathodic response reached a maximum value at +1.9 V; clearly, there is no direct oxidation occurring in the range observed for the hydroxylamine species. Consequently, it was decided to exploit only the oxidation of the hydroxylamine for analytical purposes. Effect of Dissolved Oxygen. A 10-fold enhancement in sensitivity was seen after degassing the mobile phase. We believe that this is due to electrochemically generated hydrogen peroxide chemically oxidizing the electrochemically generated hydroxylamine to the corresponding nitroso species. Obviously, this chemical oxidation decreases the amount of hydroxylamine available for the subsequent electrochemical oxidative signal produced at the second downstream detector cell, hence decreasing the sensitivity of the assay. To investigate whether H2O2 might be formed at the generator electrode, we examined aliquots of the mobile phase before and after its passage through the generator cell at an applied potential of -2.4 V, using an electrochemical technique previously described by the present authors.11 In the presence of oxygen, H2O2 concentrations of 0.13 mM were present in the mobile phase after passage through the generator cell at -2.4 V. Such concentrations are well in excess of the concentrations expected for hydroxylamine and, consequently, would be expected to oxidize this latter species. We consider that this reaction explains the decrease in oxidative signal when oxygen is not removed from the mobile phase. Studies of Possible Interferences. Paracetamol, caffeine, ascorbic acid, and 2-amino-5-nitrobenzophenone (chemical degradation product of nitrazepam) were investigated as possible interferences under the optimized electrochemical parameters. Both paracetamol and ascorbic acid were found to give oxidative responses, at concentrations of 1 mM, but did not interfere as they appeared as part of the unretained fraction. Only a small peak for caffeine was seen, with a retention time of 3.0 min. This was far removed from that of nitrazepam and so did not interfere. Additions of 2-amino-5-nitrobenzophenone produced well-defined chromatographic peaks but were resolved from that of nitrazepam. Calibration Plot, Limit of Detection, and Precision. Standard solutions containing nitrazepam in the concentration range 0.0-1.0 mM were prepared in mobile phase and determined by the optimized LC-DED procedure. The calibration plot was found to be linear from 0.070 to 11.24 µg, injected on column with a slope of 4.53 nA µg-1, with an R2 value of 0.9906. The coefficient of variation was determined by performing five replicate measurements using 140 ng and was calculated to be 4.2%. The limit of detection was calculated by making replicate current measurements at 6.2 min (n ) 5) for a blank solution; the detection limit based on three times the mean of these measurements gave a value of 5 ng of nitrazepam. Comparison of Redox Mode with Reductive Mode Liquid Chromatography Electrochemical Detection. It has been shown that liquid chromatography with electrochemical detection can be performed in the reductive mode for species that can readily undergo reduction processes.4 Therefore, we were inter(11) Gilmartin, M. A. T.; Ewen, R. J.; Hart, J. P.; Honeybourne, C. L. Electroanalysis 1995, 7, 547-555.
Figure 8. Chromatograms obtained for 300 ng of nitrazepam in (a) reductive mode and (b) and (c) redox mode. Detector currents ranges: 10 nA FSD for (a) and (b); 2 nA FSD for (c).
ested in comparing the achievable detection limits for nitrazepam by reductive and redox mode liquid chromatography. Figure 8 shows the chromatograms obtained for fixed 300-ng injections of nitrazepam using the two techniques. The current ranges were simply altered so that in Figure 8a and b full scale was 10 nA and Figure 8c was 2 nA. Clearly, the best signal-to-noise ratio was obtained by using the redox mode. It should be noted that it was not possible to operate the detector at 2-nA full-scale deflection (FSD) using the reductive mode, due to baseline drift; this was possibly due to the ingress of oxygen into the mobile phase even though degassing was performed continuously. In addition, an unresolved prepeak was seen in the reductive mode (Figure 8a). This would result in significant errors when calculating nitrazepam levels in serum samples. These results are similar to those shown for the study reported previously, which involved the naphthoquinone vitamin K1.3 Analytical Application. To assess the performance of the LCDED, six replicate determinations of nitrazepam in spiked and unspiked bovine serum samples were undertaken. Aliquots of the serum were extracted using the procedure described in section on Sample Pretreatment Procedure, and quantification was achieved by external calibration. Figure 9 shows representative chromatograms for bovine serum extracts, for (a) fortified with 11.2 µg mL-1 nitrazepam and (b) unadulterated serum. Recoveries and precision gained for a 11.2 µg mL-1 nitrazepam fortification of the original sample are summarized in Table 1. The results
obtained at this stage of our study demonstrated the feasibility of using a simple sample preparation procedure in conjunction with LC-DED for the analysis of a biological fluid. To demonstrate the application of the LC-DED assay to human overdose cases, we spiked human serum with 1670 ng mL-1 nitrazepam. This level was selected as previous workers have reported blood samples taken from overdose victims to contain 1720 ng mL-1.12 Figure 10 shows the LC-DED chromatograms for extracts of (a) human serum spiked with 1670 ng mL-1 and (b) for unspiked human serum. Clearly, the extracts showed welldefined signals for nitrazepam (peak N, Figure 10) under the conditions described. It should be noted that in this case the mobile-phase flow rate is 0.5 mLmin-1, which gives good resolution from other electroactive components of the sample; this still results in a retention time of only 14.5 min. The recovery and precision data obtained for seven replicate samples is shown in Table 2. The method can be seen to give reliable data at the concentrations investigated here, which is relevant in overdose cases.12 CONCLUSIONS An assay involving LC-DED, in the redox mode, has been successfully developed for the determination of trace levels of nitrazepam in both human and bovine serum, the former concen(12) Tracqui, A.; Kintz, P.; Mangin, P.; Lugnier, A. A. J.; Chaumont, A. J. Am. J. Forensic Med. Pathol. 1989, 10, 2, 130-133.
Analytical Chemistry, Vol. 78, No. 2, January 15, 2006
421
Figure 10. Representative chromatograms of human serum samples obtained by LC-DED in the redox mode for (a) fortified at 1760 ng mL-1 and (b) unadulterated. U1, unknown 1; U2, unknown 2; N, nitrazepam. Table 2. Recovery and Precision Data for Nitrazepam Obtained on Human Serum Samples
Figure 9. Representative chromatograms of bovine serum samples obtained by LC-DED in the redox mode for (a) fortified at 11.2 µg mL-1 and (b) unadulterated. Table 1. Recovery and Precision Data for Nitrazepam Obtained on Bovine Serum Sample
1 2 3 4 5 6
original concn (µg mL-1)
added (µg mL-1)
found (µg mL-1)
% recoveryb
nda nd nd nd nd nd
11.2 11.2 11.2 11.2 11.2 11.2
9.05 8.38 8.38 8.38 7.68 9.08
80.8 74.8 74.8 74.8 68.6 81.1
a nd, none detected. b Mean recovery, 75.8%; coefficient of variation, 6.1%.
tration being previously reported in overdose cases.12,13 The detection method is based on the reduction of nitrazepam to the hydroxylamine derivative in the generator cell, which then undergoes oxidation in the detector cell to produce the analytical response. The chromatographic separation is achieved using an octadecyl reversed-phase column in conjunction with methanolic acetate buffer as mobile phase. A convenient and simple sample pretreatment procedure was developed, which merely involved the addition of acetone to the serum sample followed by filtration. Therefore, this assay should be readily applicable to emergency toxicology screening, to drug of abuse testing, and in forensic medicine examinations. It should be mentioned that the detection of the drug at sub-ppb levels should be feasible by a simple (13) Loveland, M. R. Bull. Int. Assoc. Forensic Toxicol. 1974, 10, 16-18.
422 Analytical Chemistry, Vol. 78, No. 2, January 15, 2006
1 2 3 4 5 6 7
original concn (ng mL-1)
added (ng mL-1)
found (ng mL-1)
recoveryb
nd nd nd nd nd nd nd
1670 1670 1670 1670 1670 1670 1670
1129 1098 1192 1325 1325 1274 1318
67.6 65.7 71.4 79.3 79.3 76.3 78.9
a nd, none detected. b Mean recovery, 74.1%; coefficient of variation, 7.8%.
modification of the generator cell. Preliminary studies with a higher surface area working electrode are showing promise, and more detailed studies with this system are under way. It should be emphasized that our method requires only a simple preliminary pretreatment procedure and relies on the high selectivity of the dual-electrode detection method to eliminate interferences. Conversely, other workers14 have reported an HPLC-UV assay for nitrazepam, but this required a more rigorous serum extraction procedure involving several steps with solidphase extraction columns. Therefore, the overall time of the assay is somewhat longer than the proposed LC-DED assay, which is an important consideration in overdose investigations. In addition, the use of a UV detector might be expected to give a response to other benzodiazepines, which have similar retention times. Moriya and Hashimoto15 have also reported a method for the analysis of nitrazepam in biological materials. This involved a more elaborate pretreatment procedure and employed a complicated timeconsuming extraction step followed by gas chromatography/mass spectrometry. This approach would not readily lend itself to rapid (14) Akerman, K. K.; Jolkkonen, J.; Parviainen, Penttila, M. I. Clin. Chem. 1996, 42, 11412-1416.
diagnosis of an overdose of the drug. Although other workers have also reported on the detection of nitrazepam in biological fluids, it does not appear that they have the required speed and simplicity needed for rapid toxicological analysis.16-19 As far as we are aware, our report is the first to describe the use of an LC-DED assay for the detection of any nitro-containing (15) Moriya, F.; Hashimoto Y. Forensic Sci. Int. 2003, 131, 108-112. (16) Louter, A. J. H.; Bosma,.E.; Schipperen, J. C. A.; Vreuls, J. J.; Brinkman, U. A. Th. J. Chromatogr., B 1997, 689, 35-43. (17) Inoue, H.; Maeno, Y.; Iwasa, M.; Matoba, R.; Nagao, M. Forensic Sci. Int. 2000, 113, 367-373. (18) Kunicki, P. K. J. Chromatogr., B 2001, 750, 41-49. (19) He, H.; Sun, C.; Wang, X.-R.; Pham-Huy, C.; Chikhi-Chorfi, N.; Galons, H.; Thevenin, M.; Claude, J. R.; Warnet, J. M. J. Chromatogr., B 2005, 814, 385-391.
drug and any benzodiazepine. However, we believe that the approach developed here could form the basis of a generic approach for the analysis of other nitro-containing compounds, and in future studies, we plan to investigate this further. ACKNOWLEDGMENT We are grateful to the HEFCE for financial support and to Mike Norman, Geoff Smith, and Jim Smith for their technical assistance. Gwent Electronic Materials Ltd. are thanked for the cobaltphthalocyanine-modified screen-printed carbon electrodes. Received for review July 5, 2005. Accepted November 1, 2005. AC058035A
Analytical Chemistry, Vol. 78, No. 2, January 15, 2006
423
