ANALYST. MAY 1992, VOL. 117 843 Evaluation of Diuretics in Pharmaceuticals by High-performance Liquid Chromatography With a 0.05 mol dm-3 Sodium Dodecyl SuIfate-3% Propanol Mobile Phase Emilio Bonet Domingo, Maria Jose Medina Hernandez, Guillermo Ramis Ramos and Maria Celia Garcia Alvarez-Coque* Departamento de Quimica Analitica, Facultad de Quimica, Universidad de Valencia, Burjassot, Valencia, Spain The use of micellar liquid chromatography for the determination of diuretics in pharmaceutical preparations is studied. Micellar mobile phases containing sodium dodecyl sulfate (SDS), with and without different alcohols, are considered in order to determine the most appropriate combination. The elution behaviour of each diuretic in an unmodified micellar mobile phase is related to the solute-micelle association constants and the stationary phase-water partition coefficients.Diuretics of high, intermediate and low efficacy, contained in several pharmaceutical preparations, are determined using a 0.05 mol dm-3 SDS-3% propanol micellar mobile phase and a CI8 column. Keywords: Diuretic; pharmaceutical analysis; sodium dodecyl sulfate; high-performance liquid chromatography; micellar mobile phase Diuretics enhance renal excretion of water and electrolytes and are among the most extensively used drugs. The action of diuretics is based on interference with the mechanism of ionic transport along the complete length of the nephron. Accord- ing to their action diuretics are classified as having high, intermediate or low efficacy. Numerous procedures have been described for the control of the content of diuretics in pharmaceutical formulations using reversed-phase liquid chromatography with hydro- organic mobile phases.1-12 Most of the reports consider the determination of one or two diuretics. For these analyses, a C18 stationary phase, with methanol-water and acetonitrile- water mobile phases and acetate and phosphate buffers were usually used. Detection was performed in the ultraviolet (UV) region and recoveries and reproducibility were usually high. In the last 5-10 years, reported applications of micellar liquid chromatography have increased. Micellar liquid chro- matography, which employs solutions of surfactants as the mobile phases, is a mode of liquid chromatography that can be considered as an alternative to classical partition chromato- graphy.Some advantages of the technique are the low cost and the non-flammability , non-toxicity and easy disposal of the solvent. Difficult separations of hydrophobic and hydrophilic compounds have been achieved as a result of the large number of interactions of the solutes with the stationary and mobile phases. First, micellar mobile phases without the addition of modifiers were used, however, it was demonstrated that the presence in the mobile phase of a small amount of alcohol, giving rise to the so-called ‘hybrid’ mobile phases, enhances the efficiency of the separation and improves the retention control. 13-15 Khaledi et al. 16.17 demonstrated that the mechan- ism of separation with hybrid micellar eluents more closely resembles that in a purely micellar phase than that in conventional hydro-organic phases, as long as the integrity of the micelles is maintained.In this work, the use of micellar liquid chromatography for the determination of diuretics in pharmaceutical preparations was studied. Different micellar mobile phases were con- sidered in order to determine the most appropriate. A sodium dodecyl sulfate (SDS) micellar solution of increasing concen- * To whom correspondence should be addressed. tration, without any modifier, was first used. Next, the addition of several alcohols to the micellar mobile phase was studied. Nine diuretics, present in several different phar- maceutical preparations commercially available in Spain, were determined using a 0.05 mol dm-3 SDS-3% propanol micellar mobile phase and a C18 column.Experimental Reagents Aqueous SDS (9970, Merck, Darmstadt, Germany) solutions were used as the mobile phases. Micellar mobile phases with modifier were prepared by mixing the surfactant solution with a small amount of the alcohol to obtain the working concentration (v/v). Methanol [for high-performance liquid chromatography (HPLC)] and propanol (analytical-reagent grade) were from Panreac (Barcelona, Spain), and pentan-l- 01 (analytical-reagent grade) was from Merck. Nanopure de-ionized water (Barnstead Sybron, MA, USA) was used throughout. The mobile phases were vacuum-filtered through 0.47 pm nylon membranes from Micron-Scharlau (Barcelona, Spain). Stock solutions of 100 pgml-1 of the diuretics were prepared.Most of the compounds were soluble in 0.1 mol dm-3 SDS, but for some it was necessary to dissolve them in a small volume of methanol prior to the addition of the SDS solution. Most of the compounds were kindly donated by several Spanish pharmaceutical laboratories: acetazolamide (Lederle, Madrid, Spain), amiloride and atenolol (ICI Farma, Madrid, Spain), bendrofluazide (Davur, Madrid, Spain), bumetanide (Boehringer Ingelheim, Barcelona, Spain), chlor- thalidone (Ciba-Geigy, Barcelona, Spain), ethacrynic acid (Merck Sharp and Dohme, Madrid, Spain), frusemide (Lasa, Barcelona, Spain), hydrochlorothiazide (Galloso Wellcome, Madrid, Spain), spironolactone (Searle, Madrid, Spain), and xipamide (Lacer, Barcelona, Spain). Probenecid and triam- terene were purchased from Sigma (Buchs, Switzerland). No decomposition was observed in the diuretic stock solutions for between 15 d and 1 month; except for bendroflu- azide, which should be prepared every 2-3 d.The decomposi- tion was evident by the appearance of a peak during the dead volume of the chromatographic column and another with a retention time shorter than that of bendrofluazide, which increased with the age of the solution.844 ANALYST, MAY 1992, VOL. 117 Table 1 Influence of the concentration of SDS in the mobile phase on the values of capacity factor, k ' , efficiency, N . and asymmetry, BIA, of the peaks* [SDS]/mol dm-3 Diuretic k' Ace tazolamide 1.9 Amiloride 67.6 Bendrofluazide 28.3 Chlorthalidone 18.1 Ethacrynic acid 4.1 Frusemide 2.0 Hydrochlorothiazide 2.2 Probenecid 5.2 Triamterene 149 Xipamide 38.0 Spironolactone >150 * Asymmetry factors.18 N 1315 1780 1688 204 600 632 42 96 - - - BIA 1.10 1.09 1.21 3.00 0.92 1.89 4.70 2.14 - - - 0.03 k' 1.8 33.7 15.4 10.7 6.2 3.0 1.8 6.9 78.0 40.6 > 80 0.05 0.075 N 920 111 978 1287 65 147 1100 52 677 - - BIA 1.43 3.69 1.36 1 .oo 4.43 2.17 1.20 3.90 1 .oo - - k' 1.6 20.8 9.9 7.1 5.6 2.9 1.5 7.1 44.9 12.5 >50 N BIA 995 1.50 131 3.77 798 1.40 920 1.27 115 3.67 460 0.67 1019 1.31 59 3.80 96 3.47 148 2.93 - - k' 1.5 14.1 7.3 5.3 4.6 2.5 1.3 6.0 30.4 20.7 >35 0.1 N 838 201 844 812 102 739 865 41 123 879 - BIA 1.56 3.25 0.95 1.30 2.90 0.43 1.22 3.30 2.41 0.71 - k' 1.3 9.6 4.9 3.7 3.9 2.5 1.0 6.2 21.2 9.6 >25 0.15 N 772 147 605 555 139 494 792 85 50 273 - BIA 1 S O 3.20 1.21 1.17 2.82 1.22 1.45 2.90 3.29 2.00 - Apparatus Absorption spectra were obtained using a Hewlett-Packard 8452A diode-array spectrophotometer (Avondale, PA, USA).The HPLC system consisted of a Hewlett-Packard HP 1050 chromatograph, with an isocratic pump, a programmable UV visible detector and an HP 3396A integrator. The sample was injected through a Rheodyne valve (Cotati, CA, USA), with a 20 p1 loop. A Spherisorb octadecyl-silane (ODs)-2 CI8 analytical column (5 pm particle size, 12.5 cm x 4 mm) and a C18 pre-column of similar characteristics (2 cm X 4 mm) both from Hewlett-Packard were used. The dead volume (tM = 0.77 min) was determined from 10 replicate injections of an aqueous solution of potassium iodide and measurement of the absorbance at 254 nm.Efficiencies were calculated as theoretical plates, according to the equa- tion of Foley and Dorseylg for skewed peaks. Analysis of Pharmaceutical Formulations The pharmaceuticals analysed in this work were presented as tablets, except one containing probenecid, which was a powder for oral suspensions. In order to perform the determinations, the tablets were pulverized and an adequate amount weighed out. A 0.05 rnol dm-3 SDS solution was added and the sample immersed for 5 min in an ultrasonic bath to facilitate dissolution. Dilutions were made with 0.05 rnol dm-3 SDS. It was unnecessary to add methanol for solution. In order to eliminate any solid particles the sample was filtered through sintered glass and vacuum-filtered through the 0.47 pm membrane. Results and Discussion Mobile Phases Without Modifier Table 1 shows capacity factors, k ' , and efficiencies, N , for various SDS concentrations, together with the asymmetry factors of the chromatographic peaks.Detection was carried out at 254nm. As observed, the capacity factors usually decreased at increasing SDS concentrations. In 0.03 and 0.05 rnol dm-3 SDS, retention of acetazolamide, ethacrynic acid, frusemide, hydrochlorothiazide and probenecid was in the 1 < k' < 10 range. For the other diuretics, k' > 10. At larger SDS concentrations, the k' value of more diuretics is in the 1 < k' < 10 range. However, the value of k' for triamterene was still too large even in 0.15 rnol dm-3 SDS; retention of spironolactone was even longer. At increasing SDS concentrations, a decrease in efficiency was observed for acetazolamide, bendrofluazide and chlor- thalidone.For the other compounds, efficiency did not show a clear trend, although for hydrochlorothiazide and xipamide an important increase in efficiency was observed in 0.05 rnol dm-3 SDS compared with the 0.03 rnol dm-3 mobile phase. Efficiencies were extremely low in all the unmodified SDS mobile phases for amiloride, ethacrynic acid, probenecid and triamterene. There was no clear trend in the values of the asymmetry factors either. Very asymmetric peaks were obtained for amiloride , ethacrynic acid, probenecid, triam- terene and xipamide. These results indicated that an aqueous micellar mobile phase of SDS is not appropriate for the chromatographic determination of diuretics, partially due to the poor ef- ficiencies obtained.Partitioning Behaviour of the Diuretics Armstrong and Nome19 and Arunyanart and Cline Love20 proposed equivalent equations to describe the behaviour of a solute in micellar liquid chromatography as the micelle concentration is changed. The equations can be re-written as: where K A M is the solute-micelle association constant; [MI is the micelle concentration, i. e . , the surfactant concentration minus the critical micellization concentration (c.m.c.); and +fsw is the stationary phase-water partition coefficient multiplied by the phase ratio, @, where @ = (Vs/VM) and Vs is the volume of the stationary phase and VM the volume of the mobile phase. The value of Psw was not calculated because of the difficulty in determining the volume of the stationary phase.Table 2 shows the values of +fsw and K A M , calculated from eqn. (l), for several diuretics. For the most retained of the diuretics, the intercepts (1/@Psw) were almost zero (ami- loride, -1.2 x 10-3; bendrofluazide, 5.7 x 10-3; and triamterene, -7.2 x 10-4). Borgerding etal.21 indicated that a zero intercept requires the reciprocal of the phase ratio to be zero, which is not physically possible, or that Psw must be very large. This behaviour has been observed with sparingly soluble solutes. Bendrofluazide and triamterene were not soluble in water, but amiloride was soluble. On the other hand, as expected, for diuretics with low values of +fsw (acetazolamide, frusemide and hydrochloro- thiazide), retention times were very short.Retention of probenecid was not appreciably modifed by varying the SDS concentration and should be considered as a non-bindingANALYST, MAY 1992. VOL. 117 845 Table 2 Values of QPsw and KAM for some diuretics obtained from the llk' versus cSDS - c.m.c. linear plots [eqn. (l)] Diuretic w s w KAM Acetazolamide 2.1 4.5 Bendrofluazide 177 249 Chlorthalidone 56 101 Frusemide 3.9 5.8 Ethacrynic acid 8.2 7.9 Hydrochlorothiazide 2.6 10.8 solute22 ( KAM being close to zero). Xipamide had a variable behaviour when any condition was changed. Order of Elution It is of interest to relate the elution behaviour of each diuretic to the parameters obtained from eqn. ( l ) , KAM and @Psw. At a low micelle concentration, the system resembles conven- tional reversed-phase chromatography.If the compounds are ordered according to their elution with a 0.03 rnol dm-3 SDS mobile phase, it is observed that Psw controls the retention: the less retained solutes had lower +Psw values and the most retained had the largest ( e . g . , acetazolamide, 2.1; hydro- chlorothiazide, 2.6; frusemide, 3.9; ethacrynic acid, 8.2; chlorthalidone, 56; and bendrofluazide, 177). The values of @Psw for amiloride, triamterene and xipamide were also large. In micellar liquid chromatography, when log k' is plotted versus log cSDS (surfactant concentration), for solutes of different character, the linear plots intersect one another, which leads to a reversal in the elution order.23 This occurs as a result of the concurrence of two competitive equilibria: solute-micelle association and solute-stationary phase inter- action.An increasing micellar concentration brings the solute into the micellar phase, whereas it has no effect, or a small effect on the stationary phase equilibria. For solutes with high KAM values, the modification in surfactant concentration leads to important changes in retention, and the elution order is altered. Such elution order reversals were also observed for diuret- ics. The following compounds altered their elution order at the SDS concentration indicated: probenecid-ethacrynic acid, 0.03; acetazolamide-hydrochlorothiazide, 0.06; probenecid- chlorthalidone, 0.08; xipamide-amiloride, 0.09; bendrofluaz- ide-probenecid, 0.1; and chlorthalidone-ethacrynic acid, 0.13 rnol dm-3.Table 3 shows that the diuretics with the largest changes in capacity factors (the slope of the log k' versus log cSDS plot was larger) were triamterene, amiloride, bendrofluazide, and chlorthalidone. The latter three reversed their elution order. Triamterene, because of its long retention, although being affected, did not produce any reversal. If the diuretics are ordered according to the value of KAM, the order in Table 3 is observed: bendrofluazide, 249; chlorthalidone, 101; hydrochlorothiazide, 10.8; ethacrynic acid, 7.9; frusemide, 5.8; and acetazolamide, 4.5. Elution order reversals occurred between diuretics with sufficiently different K A M values, such as bendrofluazide and probenecid, chlorthalidone and probenecid, and ethacrynic acid and chlorthalidone.Addition of Modifiers to the Micellar Mobile Phase The addition of short-chained alcohols to the micellar mobile phase reduces the thickness of the film of surfactant molecules covering the stationary phase and thus, produces an enhance- ment in efficiency.21 The presence of the alcohol in the micellar mobile phase also alters the retention mechanism by shifting the equilibria of the solutes from the stationary phase Table 3 Regression line for the log k' Diuretic Triamterene Amiloride Bendrofluazide Chlorthalidone Hydrochlorothiazide Ethacrynic acid Acetazolamide Frusemide Slope -1.2 -1.2 -1.1 -1.0 -0.45 -0.43 -0.25 -0.22 versus log cSDS plot Intercept 0.29 -0.05 -0.22 -0.26 -0.34 0.24 -0.09 0.20 Coefficient of regression 0.997 0.998 0.999 0.9995 0.994 0.986 0.981 0.913 (4 Table 4 Influence of the modifier on the values of capacity factor, k ' , efficiency, N , and asymmetry, BIA, of the peaks 5% methanol 3% propanol 1% pentanol Diuretic k' N BIA k' N BIA k' N BIA Acetazolamide Amiloride Bendrofluazide Bumetanide Chlorthalidone Ethacrynic acid Frusemide Hydrochlorothiaz Probenecid Spironolactone Triamterene Xipamide 1.3 1302 1.46 1.1 - - 0.8 1654 1.13 29.5 237 3.74 22.4 241 4.76 5.7 207 3.40 12.5 780 2.57 9.9 2552 0.87 4.0 983 0.68 - - - 1.4 939 1.10 2.0 1080 1.00 8.7 2145 1.07 6.0 2240 1.00 2.2 1345 1.13 2.9 270 2.92 1.4 1240 1.12 1.2 593 1.54 1.5 1165 1.83 0.6 475 0.75 0.4 1740 0.67 idel.3 1136 1.67 1.1 380 4.83 0.8 1093 1.28 3.4 59 - 0.4 896 1.75 1.1 320 1.46 - 55.5 - - 10.8 - - - - 60.0 - - 37.1 336 2.05 9.9 452 0.94 18.9 94 3.52 1.1 36 3.25 3.8 220 3.00 and the micelle toward the bulk aqueous phase.This leads to a reduction in the capacity factors.16,17 A comparative study was performed to observe the effect of different alcohols added to the SDS micellar mobile phase, on the retention of the diuretics, and on the efficiency and asymmetry of the chromatographic peaks. For the preparation of these mobile phases a 0.05 rnol dm-3 SDS solution was selected. This concentration is not high and the values for efficiency were large compared with other SDS concentra- tions. The alcohols used were 5% methanol, 3% propanol and 1% pentanol. Table 4 gives the chromatographic parameters. Addition of 5% methanol produced the smallest modifications of the capacity factors; all were lower than with a 0.05 rnol dm-3 SDS mobile phase without modifier. However, only acetazol- amide, ethacrynic acid, frusemide, hydrochlorothiazide and probenecid shifted to shorter retention times as compared with a 0.075 rnol dm-3 SDS phase.When compared with a 0.15 rnol dm-3 SDS phase, only ethacrynic acid, frusemide and probenecid were less retained in a 0.05 rnol dm-3 + 5% methanol mobile phase. The addition of 3% propanol to a 0.05 rnol dm-3 SDS mobile phase had a similar effect. Compared with a 0.15 rnol dm-3 SDS mobile phase, acetazolamide, ethacrynic acid, frusemide, probenecid and xipamide were less retained. A 0.05 rnol dm-3 SDS + 1% pentanol mobile phase gave the largest eluent strength, and with the elution being largely enhanced compared with the 0.15 rnol dm-3 SDS mobile phase.The behaviour of bumetanide, probenecid and xipamide was different, as retention was decreased when propanol was used as modifier compared with pentanol as modifier. Spironolactone underwent important changes in retention: with a 0.05 rnol dm-3 SDS mobile phase without an alcohol modifier, and with a 5% methanol modifier, its retention time was >60 min, whereas with a 3% propanol modifier it was reduced to 43.5 min and with 1% pentanol it was 9 min. Other diuretics for which the retention times suffered an important diminution when using a 1% pentanol modifier in the mobile846 ANALYST, MAY 1992, VOL. 117 Table 5 Detection wavelength and limits of detection (LOD] for use with a 0.05 rnol dm-3 SDS-3% propanol mobile phase Compound Acetazolamide Amiloride Atenolol Bendrofluazide Bumetanide Chlorthalidone Frusemide Hydrochlorothiazide Probenecid Spironolactone Triamterene Xipamide Unm 224 220 220 274 224 274 224 224 224 242 242 224 LOD/pg ml-1 0.037 0.057 0.59 0.019 0.0032 0.0052 0.0029 0.0036 0.0051 0.55 0.10 0.0040 Table 6 Nominal contents, recoveries and reproducibility for the drugs in the pharmaceuticals Recovery Pharmaceutical Content (”/I Aldactone-A (Searle) 25 mg spironolactone 102.5 BlCnox (Farma) 1 g probenecid 96.1 2.5 g amoxycillin 40 mg sodic saccharin Diamox (Cyanamid I bCrica) 250 mg acetazolamide 117.0 Diurex (Lacer) 20 mg xipamide 98.3 Fordiuran (Boehringer lactose Ingelheim) 1 mg bumetanide 103.7 lactose Hidrosaluretil (Gayoso Wellcome) 50 mg hydrochlorothiazide 104.4 100.2 Triniagar (Farmasines) 50 mg triamterene Aldoleo (Leo) Ameride (Merck Sharp & Dohme) Neatenol (Fides) Normopresil (Semar) Spirometon (Davur) 50 mg mebuticine lactose 50 mg spironolactone 99.8 50 mg chlorthalidone 97.5 50 mg hydrochlorothiazide 100.7 5 mg amiloride chlorhydrate lactose 5 mg bendrofluazide 101.4 100 mg atenolol 107.5 25 mg chlorthalidone 102.9 100 mg atenolol 108.0 2.5 mg bendrofluazide 103.2 50 mg spironolactone 99.4 RSD 4.2 2.8 (Yo 1 2.6 0.8 0.5 0.5 3.6 4.3 6.8 0.3 1.5 1.4 4.5 2.1 0.4 2.8 phase were amiloride, bendrofluazide, chlorthalidone, triam- terene and xipamide.When elution was performed with 1% pentanol in the mobile phase, k’ < 10 for all diuretics, except spironolactone. With respect to the efficiencies, the behaviour was variable, but frequently the best efficiencies corresponded to a micellar mobile phase with 3% propanol, as has been indicated by other workers.24.25 The improvement was most marked for bendrofluazide, ethacrynic acid and probenecid.However, with the 3% propanol in the mobile phase, efficiency was decreased compared with the other two mobile phases for frusemide, hydrochlorothiazide and xipamide. When reten- tion times were either extremely short (acetazolamide) or extremely long (triamterene and spironolactone), background noise was excessive and sometimes the efficiencies could not be calculated, as the peak width at 1/10 peak height, Wo.l, was needed. For many diuretics, the most symmetrical peaks were obtained with a ‘hybrid’ 1% pentanol mobile phase. In conventional liquid chromatography with hydro-organic mobile phases, as the elution strength of the solvent increases there is a systematic decrease in selectivity, expressed as a = (k’2/kli) (V1 and k’2 are the capacity factors of two solutes, with kI2 > k‘,).17 In contrast, in micellar liquid chromato- graphy, selectivity might increase or decrease with micelle concentration depending on the nature of the compounds, that is, on the electrostatic and hydrophobic interactions with micelles. Selectivity modifications were observed with different pairs of diuretics eluting close together.Elution order reversals occur when the mobile phase is changed. Thus, in order to compare selectivity among different mobile phases, a set order of elution should be taken to obtain the selectivity factors, such as the order of elution in 0.05 mol dm-3 SDS solution, in the absence of modifier.With the values of k‘ given in Tables 1 and 4, it can be calculated whether the selectivity increases or decreases for different pairs of compounds. Analysis of Pharmaceutical Formulations Retention of diuretics in a purely SDS micellar mobile phase should be decreased in order to perform the analyses. With methanol as modifier, elution was still too slow, however, when using pentanol, elution of the diuretics appearing at the beginning of the chromatogram was markedly accelerated and resolution deteriorated. Propanol showed an intermediate behaviour and sometimes gave the best efficiencies. There- fore, a 0.05 mol dm-3 SDS-3% propanol mobile phase was chosen.Table 5 indicates the wavelengths of detection used for each diuretic, which was close to a maximum wavelength. A calibration curve was obtained for each diuretic with five 0.05 mol dm-3 SDS solutions at different concentrations. Usually, the coefficients of regression were >0.99. The highest sensitivities corresponded to bumetanide, frusemide, hydrochlorothiazide, probenecid and xipamide. The limits of detection were calculated from the background noise in the nearest of the chromatographic peaks (30 criterion, 10 replicates). The lowest limits of detection corresponded to bumetanide, frusemide, hydrochlorothiazide and xipamide (Table 5 ) , i.e., the diuretics showing the highest sensitivity. The poorest limits of detection corresponded to spironolac- tone and triamterene, owing to their long retention times.Table 6 shows the pharmaceuticals analysed that contained one or two diuretics; when two diuretics were together, one was of high or intermediate efficacy and the other of low efficacy. The peaks were always well resolved. Some prepara- tions also contained another drug, such as a &blocker or a stimulant, which did not interfere with the analyses. The determination of atenolol was also considered. When the solutions of the preparations were injected into the column, a peak was always observed in the dead volume of the system, which probably corresponded to the excipient . Amiloride, triamterene, spironolactone and atenolol had long retention times. The determination of these compounds is more appropriate with a mobile phase of high eluent strength, e.g., with an SDS-pentanol mobile phase.The determination of some of these compounds with the SDS- propanol mobile phase was also examined. Table 6 also shows the recoveries with respect to the composition given by the manufacturers. These were usually in the 96104% range. Relative standard deviations (RSDs) of five replicate injections were usually in the 0.3-4.270 range. The results indicate that micellar liquid chromatography is adequate for the determination of diuretics in pharmaceutical preparations. This work was supported by the Comisi6n Interministerial de Ciencia y Technologia (CICyT), Project DEP89-0429. References 1 2 Honigberg, I . L.. Stewart. J . T.. Smith. A. P.. and Hester. D. W., J . Pharm.Sci., 1975, 64, 1201. Menon, G. N.. and White. L. B., J. Pharm. Sci.. 1981.70,1083.ANALYST, MAY 1992, VOL. 117 847 3 4 5 6 7 8 9 1 0 11 12 13 14 Roth, J . , Rapaka, R. S., and Prasad, V. K., Anal. Lett., 1981, 14, 1013. Walters, S. M., and Stonys, D. B., J . Chromatogr. Sci., 1983, 21,43. de Croo, F., van den Bossche. W., and de Moerloose, P., Chromatographia, 1985. 20, 477. Fogel, J., Sisco, J . , and Hess, F., J . Assoc. Of5 Anal. Chem.. 1985, 68, 96. de Croo. F., van den Bossche, W., and de Moerloose, P., J . Chromatogr.. 1985, 329, 422. Yarwood, R. J., Moore, W. D., and Collett, J. H., J . Pharm. Sci., 1985. 74, 220. Sane, R. T.. Sadana, G. S . , Bhounsule, G. J . , Gaonkar, M. V., Nadkarni, A. D., and Nayak, G.. J . Chromatogr., 1986, 356, 468. de Croo, F., van den Bossche, W., and de Moerloose. P., J . Chromatogr., 1986,354, 367. Hitscherich, M. E., Rydberg, E. M., Tsilifonis, D. C., and Daly. R . E., J . Liq. Chromatogr., 1987, 10, 1011. Bachman. W. J., and Stewart, J . T., J . Chromatogr. Sci.. 1990, 28, 123. Dorsey. J., Khaledi, M. G., Landy. J . S . , and Lin, J.-L.. J. Chromatogr., 1984, 316, 183. Borgerding, M. F., Hinze, W. L., Stafford, L. D., Fulp, G. W., Jr., and Hamlin, W. C., Jr., Anal. Chem.. 1989, 61, 1353. 15 16 17 18 19 20 21 22 23 24 25 Berthod, A., Girard. I.. and Gonnet, C., Anal. Chem.. 1986, 58, 1362. Khaledi, M. G., Anal. Chem., 1988, 60, 876. Khaledi, M. G., Strasters, J . K., Rodgers, A. H., and Breyer, E. D., Anal. Chem.. 1990, 62, 130. Foley, J . P., and Dorsey, J. G., Anal. Chem., 1983, 55. 730. Armstrong, D. W., and Nome, F., Anal. Chem., 1981,53,1662. Arunyanart, M., and Cline Love, L. J., Anal. Chem., 1984,56, 1557. Borgerding, M. F., Quina, F. H . , Hinze, W. L., Bowermaster, J . , and McNair, H. M., Anal. Chem.. 1988, 60,2520. Armstrong, D. W., and Stine, G. Y., Anal. Chem., 1983, 55, 2317. Yarmchuk, P., Weinberger, R., Hirsch. R. F., and Cline Love, L. J., Anal. Chem., 1982, 54, 2233. Dorsey, J . G., DeEchegaray, M. T., and Landy, J . S., Anal. Chem., 1983,55, 924. Berthod, A., and Roussel, A., J . Chromatogr., 1988,449, 349. Paper 1f04774D Received September 16, 1991 Accepted November 11, 1991