ANALYST, NOVEMBER 1986, VOL. 111 1245 Sensitive Adsorptive Stripping Voltammetric Measurements of Antihypertensive Drugs Joseph Wang,* Timothy Tapia and Mojtaba Bonakdar Department of Chemistry, New Mexico State University Las Cruces, NM 88003, USA Controlled adsorptive accumulation of reserpine, rescinnamine and hydralazine on carbon electrodes provides the basis for sensitive stripping measurement schemes for these compounds. Cyclic voltammetry is used to explore the interfacial and redox behaviours. The detection limits are reserpine, 2 x 10-9 M, rescinnamine, 3 x 10-9 M, and hydralazine, 1 x 10-8 M. The adsorptive stripping response is evaluated with respect to various experimental conditions. Exchange to a blank solution corrects for major interferences. The relative standard deviation at the sub-micromolar level ranges from 5 to 9%.Applicability to measurements in urine is illustrated. Keywords Reserpine; h ydralazine; stripping voltammetry; adsorptive accumulation; antihypertensive drugs Adsorptive stripping voltammetry has been demonstrated to be a highly sensitive electroanalytical method for the determi- nation of a wide range of electroactive compounds.'-3 The method utilises controlled interfacial accumulation of the analyte on to the electrode surface as an effective pre- concentration step, prior to the voltammetric determination of the surface species. This extends the scope of stripping voltammetry towards additional analytes that cannot be pre-concentrated electrolytically. Among these are numerous compounds of pharmaceutical significance; reducible drugs such as diazepam and nitrazepam, 1 digoxin4 and tetracyclines5 have been determined at a hanging mercury drop electrode, whereas oxidisable drugs, e.g., chlorpromazine ,6 adriamycin7 and tricyclic antidepressants8 have been determined at various carbon electrodes.This paper describes sensitive adsorptive stripping pro- cedures for the determination of trace amounts of several antihypertensive agents. The determination of the widely used Rauwolfia alkaloid reserpine, its related drug rescinnamine and the vasodilator hydralazine at various carbon electrodes is described. These compounds are electrochemically active. There have been several earlier studies on the polarographic determination of reserpine, particularly using a.c. polaro- graphy and an aprotic organic solvent system.9-11 Such studies yielded detection limits in the 10-5 M range, thus allowing convenient tablet assays. Reserpine and rescinnamine also exhibit an anodic response. This was exploited recently for amperometric detection following liquid chromatography.12 Similarly, the oxidation of hydralazine was used for amper- ometric detection following liquid chromatography. 1 3 ~ 4 Sen- sitive voltammetric measurements of these drugs, based on their anodic behaviour at carbon electrodes, have not been attempted. As illustrated in this study, such behaviour can be coupled with the interfacial properties of reserpine, rescinn- amine and hydralazine to yield a highly sensitive adsorptive stripping procedure. Hence, detection limits at the nanomolar concentration level are obtained.The characteristics of such procedures are described in this paper. Experimental Apparatus A 10-ml voltammetric cell (Model VC-2, Bioanalytical Systems) was used. The cell was joined to the working electrode, reference electrode (Ag - AgC1, Model RE-1, Bioanalytical Systems) and platinum wire auxiliary electrode through holes in its Teflon cover. A magnetic stirrer and a stirring bar (1.2 cm in length) provided the convective * To whom correspondence should be addressed. transport during the pre-concentration. In experiments involving the medium exchange, a second (measurement) cell was used. The carbon working electrodes included a glassy carbon disc (3 mm in diameter, Bioanalytical Systems) and a laboratory-made carbon paste disc (3 mm in diameter), prepared by mixing graphite powder (Acheson 38) and Nujol oil (40% oil by mass).A fresh carbon paste surface was used daily, with the surface being smoothed on a computer card. The glassy carbon surface was polished daily with a 0.05-pm alumina slurry, rinsed with de-ionised water and allowed to air-dry . Stripping and cyclic voltammograms were recorded with EG & G Princeton Applied Research Models 364 and 264 polarographic analysers, respectively. Reagents Stock solutions (5 x M) of the antihypertensive agents (Sigma) were prepared daily. The hydralazine solution was prepared by dissolving the compound in de-ionised water. Reserpine and rescinnamine were dissolved in acetic acid and ethanol, respectively, and then diluted with de-ionised water.The supporting electrolyte was 0.05 M phosphate buffer, prepared from a 1 + 4 mixture of KH2P04 and K2HP04 and adjusted to pH 4.0 with H3P04. All solutions were prepared from de-ionised water and analytical-reagent grade chemicals. The urine samples were obtained from a healthy volunteer and diluted (1 + 4) with the supporting electrolyte prior to use. Procedure The pre-concentration step was performed by immersing the working carbon electrode into a stirred (ca. 400 rev. min-1) 10-ml sample solution for a given time period. During this period the electrode was held at 0.0 V (reserpine, rescinn- amine) or -0.3 V (hydralazine). The stirring was then stopped and the surface species were determined by applying an anodic potential scan (a differential pulse waveform for reserpine and rescinnamine and a d.c.ramp for hydralazine). In experiments involving medium exchange, the pre-concentration proceeded at an open circuit; the electrode was then transferred into an electrolytic blank solution where an anodic potential scan was applied. To clean the surface of the remaining accumulated species, the electrode was held at +1.4 V (reserpine, rescinnamine) and +1.3 V (hydralazine) for 60 s; a subsequent scan was used to indicate the absence of memory effects. Results and Discussion Determination of Reserpine and Rescinnamine The accumulation of organic compounds at carbon paste electrodes usually proceeds via a mixed (adsorptive - extrac-1246 ANALYST, NOVEMBER 1986, VOL. 111 I 400 nA I B I I I 1 I I 1.0 0.8 0.6 0.4 0.2 0 E N Fig. 1.Repetitive cyclic voltammograms for 5 x 10-6 M reserpine (A) and rescinnamine (B), following 3- and 2-min stirring periods, respectively, at 0.0 V. Scan rate, 100 mV s-l; electrolyte, phosphate buffer (pH 4); electrode, carbon paste (a, T I I I 0.4 0.6 0.8 EIV Fig. 2. Differential pulse voltammograms for 6 x 10-8 M reserpine (a) and rescinnarnine (b) following (A) 0 and (B) 3 min pre- concentration. Differential pulse ramp with 50 mV amplitude and 5 mV s-l scan rate. Other conditions as in Fig. 1 tive) process. Fig. 1 shows repetitive cyclic voltammograms at a carbon paste electrode for 5 x 10-6 M reserpine (A) and rescinnamine (B). Stirring the solution prior to the first scan (designated as 1) results in large anodic peaks [at 0.71 V (A) and 0.75 V (B)] because of the oxidation of the accumulated drugs.Substantially smaller peaks are observed on continued scanning, indicating desorption (from the surface) and “back- extraction” (from the electrode interior) of the reaction product. Eventually, a stable response, corresponding to the contribution of the solution species alone, is observed. No peaks are observed in the cathodic branch, indicating an irreversible redox process. According to Adams,lS the anodic oxidation of methoxy-substituted indole alkaloids, including reserpine-like compounds, involves the introduction of a hydroxy group at the indole 5-position. By the use of 5 x 10-6 M solutions, surface saturation was observed after 120 s (reserpine) and 180 s (rescinnamine).The response for the surface-attached drugs under these condi- tions was used to determine the surface coverage and scan rate dependence. For example, the amounts of charge consumed during the cyclic voltammetry experiment by the redox 12 0 8 .-a . .-! 4 0 8oo . . P 400 I 0 I I I ( a ) 2 4 6 rimin Fig. 3. Dependence of the reserpine (A) and rescinnamine (B) peak current (a) and peak current enhancement (b) on the pre-concentra- tion time. Other conditions as in Fig. 2 process at saturation, i.e., 0.79 pC (reserpine) and 0.59 yC (rescinnamine), correspond to surface coverages of 3.2 X 10-11 and 2.5 x 10-11 mol cm-2, respectively. Such values reflect the similar molecular structures of the drugs. Graphs of log (peak current) vs. log (scan rate) for the surface-attached compounds over the 5-500 mV s-1 range were linear. These graphs had slopes of 0.76 (correlation coefficient 0.997) and 0.72 (correlation coefficient 0.998) for reserpine and rescinn- amine, respectively.Thus, a deviation from an ideal behavi- our of surface species (slope 1.00) is observed. The change of scan rate from 50 to 500 mV s-1 resulted in positive shifts in peak potential from 0.71 to 0.80 V (reserpine), and from 0.76 to 0.82 V (rescinnamine). Graphs of peak current vs. scan rate were linear over this range. No change in peak potentials was observed over the 5-50 mV s-1 scan rate range. The spontaneous adsorption of reserpine and rescinnamine can be used as an effective pre-concentration step prior to the voltammetric determination of these drugs.The resulting adsorptive stripping procedure offers a convenient determina- tion at the sub-micromolar and nanomolar concentration levels. For example, Fig. 2B shows differential pulse voltam- mograms at a carbon paste electrode that was dipped in stirred 6 x 10-8 M solutions of reserpine (a) and rescinnamine ( b ) for a 3-min pre-concentration period. Also shown (A) is the corresponding response without accumulation. For the short pre-concentration period used, a significant improvement in the sensitivity is observed. Detection limits of about 2 X 10-9 M for reserpine and 3 X 10-9 M for rescinnamine are estimated based on the signal to noise characteristics (SIN = 3) of the data shown in Fig. 2B. ‘Thus, 12 and 19 ng of reserpine and rescinnamine, respec- tively, can be detected in the 10 ml of solution used, i.e., 1.2-1.9 ng ml-1. Compared with the previous a.c.polaro- graphic procedures for reserpine the adsorptive stripping approach lowers the detection limit by four orders of magnitude. Attempts to perform adsorptive stripping measurements, based on the reduction process at the hanging mercury drop electrode, yielded a substantially inferior adsorptive stripping response. The extent of accumulation, and accordingly the resultingANALYST, NOVEMBER 1986, VOL. 111 ( a ) 1247 ( b) response, is affected by the solution conditions. The response was examined in the presence of various supporting electro- lytes, e.g. , sodium hydroxide, phosphate buffers (pH 4.0, 7.4 and 9.0), ammonium chloride and hydrochloric acid. The best results (with respect to peak enhancement and reproducibil- ity) were obtained with a phosphate buffer (pH 4.0); this electrolyte was used in subsequent work.Fig. 3(a) shows the dependence of the peak current on the pre-concentration time for reserpine (A) and rescinnamine (B). Both compounds exhibit similar current - time profiles; the peaks increase rapidly with time at first and then level off. Full surface coverage is approached for pre-concentration times longer than 4 min. At this point, the peak current enhancements, relative to direct measurements, are 12 (reserpine) and 8 (rescinnamine) [Fig. 3(b)]. Obviously, a trade-off would be required when optimising the pre-concen- tration. For convenient determination at the 5 x 10-8 M level, a pre-concentration time of 3 min is usually sufficient (e.g., Fig.2). Fig. 4 shows the dependence of the peak current on the reserpine (A) and rescinnamine (B) concentration using a 1 min pre-concentration. A curvature in the calibration graphs is observed over the 1.25 x 10-8 - 7.5 x 10-8 M range tested. Such behaviour is consistent with a process that is limited by the interfacial accumulation of the analyte. Accordingly, determination should be based on the use of calibration graphs and not the method of standard additions. The reciprocal plots, lli vs. llc, yielded straight lines over the concentration range examined in Fig. 4 (not shown). Such linearity is expected on the basis of the Langmuir adsorption model and may also be useful for the determination.Replicate peaks observed for the same sample solution illustrate the precision of the method. A series of six measurements of 6 x 10-8 M reserpine yielded a mean peak current of 0.67 PA, a range of 0.59-0.70 pA and a relative standard deviation of 9% (2 min pre-concentration; other conditions as in Fig. 2). High selectivity is yet another important advantage of the voltammetric determination of reserpine and rescinnamine based on interfacial accumulation. The selectivity advantage is obtained by incorporating the “medium-exchange” proce- dure716J7 i.e., the transfer of the electrode (with the accumu- lated drug) to a blank solution prior to the voltammetric scan. In this way, effective discrimination against interferences due to dissolved electroactive species is achieved.For example, Fig. 5(a) demonstrates the direct measurement of 2.5 x 10-7 M reserpine in a diluted (1 + 4) urine sample. Determination is not feasible by performing the voltammetric measurement in 0 2.5 5 7.5 Concentration ( x 10-8 M) Fig. 4. Dependence of the peak current on reserpine (A) and rescinnamine (B) concentration using 1 min pre-concentration. Other conditions as in Fig. 2 0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 E N 5. Measurements of reserpine in (a) diluted (1 + 4) urine and M (a); in the presence of 1 x 10-4 M ascorbic acid with (B) and without 1.25 X 1 0 - 7 ~ ( b ) . Pre-concentration time: 2 min. Other conditions as in Fig. 2 medium exchange. Reserpine concentration: 2.5 x 0.8 0.6 0.4 0.2 0.0 -0.2 EIV Fig. 6. Repetitive cyclic voltammograms for 1 x 10-5 M hydralazine at a glassy carbon electrode following 1 min stirring at -0.3 V.Scan rate. 20 mV s-l: electrolvte. ohosohate buffer (DH 4)1248 ANALYST, NOVEMBER 1986, VOL. 111 -0.2 0.0 0.2 0.4 0.6 E N Fig. 7. M hydralazine at a glassy carbon electrode after different pre-concentration times: A, 0; B, 30; .C, 120; and D, 300 s. Pre-concentration at -0.3 V with 400 rev. min-1 stirring. Scan rate, 50 mV s-l; electrolyte, phosphate buffer (pH 4.0) Determination of Hydralazine The interfacial accumulation of the antihypertensive agent hydralazine is indicated from repetitive cyclic voltammograms recorded after a 1 min stirring period at -0.3 V (Fig. 6). The drug exhibits an irreversible oxidation peak at +0.25 V. The first scan (designated as 1) yields a large current response owing to the oxidation of the adsorbed species.Subsequent voltammograms, recorded on continued scanning, yielded much smaller peaks, indicating the rapid desorption of the product. The current enhancement, associated with the interfacial accumulation, allows the determination of trace amounts of hydralazine based on the adsorptive stripping approach. Fig. 7 shows linear scan voltammograms for 8 x 10-7 M hydralazine after different pre-concentration periods. Determination at this level is not feasible without pre-concentration (A). The peak height increases rapidly with increasing pre-concentra- tion time, indicating enhancement of the hydralazine concen- tration on the glassy carbon surface. For example, 30 and 300 s pre-concentration periods yield 10- and 40-fold enhancements of the peak current, respectively, relative to that attained without pre-concentration. As a result, hydralazine can be easily determined at the sub-micromolar concentration level; a detection limit of about 1 x 10-8 M is estimated based on the signal to noise characteristics (SIN = 3) of the 300 s pre-concentration voltammogram (D) .The adsorptive stripping hydralazine response is affected by the pre-concentration potential and the solution conditions. For example, a gradual increase of the peak (up to 70%) was observed on changing the pre-concentration potential from 0.0 to -0.3 V; no further change in current was observed at -0.4 and -0.5 V (1 x 10-6 M hydralazine, 2 min pre- concentration). Strong accumulation and well defined peaks were observed in phosphate buffer solutions of pH 4.0 and 7.4.A pH 9.0 phosphate buffer solution yielded an inferior performance. A -0.3 V pre-concentration potential and pH 4.0 phosphate buffer were used for the determination of hydralazine. Fig. 8 shows the dependence of the hydralazine peak current on the pre-concentration time (A) and the bulk concentration of the drug (B). The peak increases rapidly with time at first and then levels off (with some decrease for times Linear scan voltammograms for 8 X Timeimin 0 2 4 6 8 10 I 1 I I 0 5 10 15 20 Concentrationil 0-7 M Fig. 8. Dependence of the hydralazine peak current on pre-concen- tration time (A) and concentration (B). Pre-concentration time: B, 120 s. Concentration: A, 8 x 10-7 M. Other conditions as in Fig.7 longer than 5 min). A well defined concentration dependence is indicated from the calibration graph, with a curvature that represents the corresponding adsorption isotherm. The adsorptive stripping response of hydralazine is highly reproducible. Ten successive measurements yielded a mean peak current of 0.76 PA, a range of 0.71-0.85 yA and a relative standard deviation of 5% (5 x 10-7 M hydralazine, 2 min pre-concentration). In conclusion, highly sensitive voltammetric measurements of several antihypertensive agents are feasible after their interfacial accumulation and oxidation at various carbon electrodes. Improved selectivity is obtained using the medium-exchange procedure. The entire pre-concentration - medium-exchange - voltammetric scheme can be easily accomplished using a flow injection system,17 as is desirable in the clinical laboratory, Besides the analytical utility, the new knowledge of the interfacial behaviour may offer better understanding of the pharmaceutical activity of these drugs, particularly of their interaction with biosurfaces.This work was supported by the American Heart Association and the National Institutes of Health (Grants No. RR08136-12 and GM 30913-03). 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. References Kalvoda, R., Anal. Chim. Acta, 1982, 138, 11. Wang, J., Am. Lab., 1985, 17, 41. Wang, J., “Stripping Analysis: Principles, Instrumentation and Applications,” VCH, Deerfield Beach, FL, 1985. Wang, J., Mahmoud, J. S . , and Farias, P. A. M., Analyst, 1985, 110, 855. Wang, J., Peng, T., and Lin, M. S . , J . Bioelectrochem. Bioenerg., 1986, 15, 147. Jarbawi, T. R., and Heineman, W. R., Anal. Chim. Acta, 1982, 154, 359. Chaney, C. E., and Baldwin, R. P., Anal. Chem., 1982, 54, 2556. Wang, J., Bonakdar, M., and Morgan, C., Anal. Chem., 1986, 58, 1024. Schaar, J. C., and Smith, D. E., J . Electroanal. Chem., 1979, 100, 145. Woodson, A. L., and Smith, D. E., Anal. Chem., 1970, 42, 242. Taira, A., and Smith, D. E., J . Assoc. Off. Anal. Chem., 1978, 61, 941. Wang, J., and Bonakdar, M., J . Chromatogr., in the press. Shah, M. H., and Stewarts, J. T., J. Pharm. Sci., 1984,72,989. Ravichandran, K., and Baldurin, R. P., J . Chromatogr., 1985, 343, 99. Adams, R. N., “Electrochemistry at Solid Electrodes,” Marcel Dekker, New York, 1979, p. 323. Wang, J., and Freiha, B., Anal. Chim. Acta, 1983, 148, 79. Wang, J., and Freiha, B., Anal. Chem., 1983, 55, 1285. Paper A61169 Received May 30th, 1986 Accepted June 27th, 1986