Sensitive and Rapid High-Performance Liquid Chromatographic Method for Simultaneous Determination of Antidepressants in Pharmaceutical Formulations 2 0 0 2 , 56, 5 4 5 - 5 5 1

J.J. Berzas / C. Guiberteau/A. M. Contento*/V. Rodrfguez Department of Analytical Chemistry and Food Technology, University of Castilla - La Mancha, 13071 Ciudad Real,Spain; [email protected]

KeyWords Column liquid chromatography Antidepressants Pharmaceutical formulations

Summary An easy, rapid, and sensitive high-performance liquid chromatographic (HPLC)method is proposed for simultaneous determination of five antidepressants (trazodone, citalopram, fluoxetine, fluvoxamine and clomipramine) and norfluoxetine, a major metabohte of fluoxetine. Optimum conditions for the quantitative separation were investigated.The compounds were separated on a Nova Pack Cls column with acetonitrile-pH 2.5 phosphate buffer, 40:60 (v/v), as mobile phase at a flow rate of 1.5 mL min 1..Stability of solutions, hnearity, repeatability and reproducibility, accuracy, specificity, and detection and quantification limits were examined to validate the method. The method is rapid and sensitive and can be applied to the analysis of antidepressants in pharmaceutical formulations and biological fluids; norfluoxetine can also be determined in these matrices. We have tested the method by analysis of several pharmaceutical formulations; recoveries obtained were bef,,veen 102.7% and 98.2% of the labelled values.

Introduction Serotonin (5-HT) has been implicated in the aetiology of many disease states and might be particularly important in mental illnesses such as depression, anxiety, schizophrenia, eating disorders, obsessive compulsive disorder, migraine, panic disorders, bulimia, etc. Indeed, many currently used treatments of these disorders are thought to act by modulating serotoninergic tone [1]. Antidepressant drugs are widely used for treatment of a variety of depressive

Chromatographia 2002, 56, November (No. 9/10)

Original 0009-5893/00/02

states and other psychiatric disorders. Tricyclic antidepressants (TCA), which act mainly on the adrenergic system, are a standard treatment for depression [2, 3] but have many concentration-dependent adverse effects. During the last decade many 5-HT receptor subtypes have been characterized. This has led to the realization of many treatments acting via the serotoninergic system, e.g. selective serotonin re-uptake inhibitors (SSRI) [1] have been introduced which differ from the TCA both in their chemical structure and in their mechan-

ism of action. The SSRI, in contrast with TCA, do not significantly inhibit re-uptake of norepinephrine in the nerve endings in-vitro or in-viva. SSRI are comparable with TCA in their clinical efficacy but, because of their favourable pharmacological profile, are considered to be safe and well tolerated drugs [4, 5]. Several methods have been published for the determination of SSRI and TCA, mainly in biological fluids. Gas chromatography (GC) with electron-capture [6], mass spectrometric [7, 8], or flame ionization [9] detection has been used to determine several of these drugs and their metabolites. Other relevant work includes quantification of the enantiomeric forms of both fluoxetine and norfluoxetine by HPLC [10] and CG-ECD [11]. HPLC methods with ultraviolet [12, 13] or fluorescence [14, 15] detection have been reported mainly for the determination of fluoxetine and its major metabolite (norfluoxetine). Other publications have reported the separation and quantification of several antidepressants and their metabolites in biological fluids with diode-array [16, 17] and fluorescence [18, 19] detection; usually several mobile phase additives or derivation reactions are used to achieve the desired sensitivity or selectivity. Use of these techniques can result in solubility problems and damage to the column. Capillary zone electrophoresis (CZE) also has been used to determine several SSRI in pharmaceutical formulations [19, 21] and fluoxetine and norfluoxetine in biological samples [22, 23]. Given that the availability of a single chromatographic method suitable for the

545- 07

$ 03.00/0

9 2002 Friedr. Vieweg & Sohn Verlagsgesellschaft mbH

545

Table I. Names, abbreviations, and chemical structures of the antidepressants.

Structures

Abbreviation

Compounds

Fluoxetine

FLX

/CH3

H

Norfluoxetine

NFLX

~ Trazodone

NH2

~--'N

TRA

CI

Citalopram

CN.~?

CIT

~CH3 N4..CH ]

~ F

Fluvoxamine

F3C~

FLV

Clomipramine

~

Retention time (RT, min) and resolution ( R s ) at pH: 7

5.5

3.5

2.5

RT

Rs

RT

Rs

RT

Rs

RT

Rs

TRA CIT FLV NFLX FLX CLO

3.2 3.9 3.9 4.2 4.2 6.7

2.5 <1 <1 <1 <1 4.8

1.7 2.1 2.3 2.7 3.4 6.1

1.6 3.6 0.5 1.5 1.8 3.4

1.3 1.8 2.0 2.8 3.4 4.5

1.5 1.5 1.9 1.9 1.9 2.9

1.1 1.6 2.1 2.6 3.1 4.3

1.6 1.6 2.0 2.0 1.9 3.0

Rs=2(t2

tl)](W2 +

546

H3

CLO

Table II. Effect of mobile phase pH.

Drug

~,~O,'~'C N\ O~NH a

g/l).

C h r o m a t o g r a p h i a 2002, 56, N o v e m b e r (No. 9/10)

simultaneous d e t e r m i n a t i o n of several drugs would be very useful in a n y laboratory, because m a i n t e n a n c e of i n d e p e n d e n t analytical m e t h o d s for each drug is more expensive, we have developed a very simple, rapid, a n d sensitive liquid c h r o m a t o graphic m e t h o d , w i t h o u t use of mobilephase additives, t h a t enables determination of four SSRI (trazadone, T R A , citalopram, CIT, fluoxetine, F L X , a n d fluvoxamine, FLV), clomipramine (CLO, a T C A antidepressant), a n d norfluoxetine ( N F L X , a metabolite of fluoxetine). This

Original

method can be applied to the analysis of these compounds in pharmaceutical formulations and biological fluids. To demonstrate its usefulness the procedure was applied to the determination of these antidepressants in pharmaceutical preparations, commercially available in Spain, which contain one of these compounds with other components.

Experimental

Nova Pack Cls reversed-phase column from Waters Millipore (Milford, MA, USA) with 40:60 acetonitrile-75 mM phosphate buffer, pH 2.5, as mobile phase. Isocratic elution was performed at a flow rate of 1.5 mL min 1. The volume injected (samples and standards) was 20 ixL. pH was measured with a Crison model 2001 pH meter with a combined glass electrode.

Procedures

Reagents All solvents and reagents were of analytical grade unless indicated otherwise. Solutions were prepared in deionized water (Milli-Q quality). Fluoxetine (hydrochloride), fluvoxamine (maleate), and clomipramine (hydrochloride) were obtained from Tocris Coolson and norfluoxetine (hydrochloride) from Sigma-RBI. Trazodone (hydrochloride) and citalopram (bromide) were kindly supplied by Farma-Leporl and Lundbeck Laboratories, respectively. Excipients were purchased from Acofarma. Stock solutions of all the drugs (200 mg L 1) were prepared by dissolving pure samples in Milli-Q water; they were stored under refrigeration at 4 ~ Working standard solutions were prepared daily by dilution of the stock standard solution with purified water. Mobile phases and buffer solution were prepared from analytical-reagentgrade Na2HPO4, NaH2PO4, and H3PO 4 from Merck (Germany) and HPLC-grade acetonitrile from Sharlab. The phosphate buffer solutions were filtered through 0.45-1xm filters (MF membrane filters) and acetonitrile through 0.5-1xm filters (Fluoropore membrane filters). Both types of membrane filter were purchased from Millipore.

Apparatus HPLC was performed with Shimadzu (Kyoto, Japan) model LC-10AD solvent delivery pumps, a Rheodyne injection valve with 20-1xL sample loop, and a model SPD-M10A diode-array detector. The system was monitored by means of a computer equipped with Class-LC 10 software, which was used for all measurements and data treatment. Compounds were separated on at room temperature on a 3.9 mm i.d. x 150 mm, 5 ixm particle, Original

Use of diode-array detection enabled extraction of chromatograms at different wavelengths. In the optimisation process we monitored the analytes at 230 nm. In the validation procedure we monitored each antidepressant at its wavelength of maximum absorbance. Because all the pharmaceutical preparations analysed contained only one of the antidepressants studied, any of the other five drugs could be used as internal standard for quantification by use of internal normalization criteria [24] in these pharmaceutical applications. FLX was used as internal standard for determination of TRA, CIT, FLV, and CLO; TRA was used as internal standard for determination of FLX and NFX. Each solution was injected in duplicate and average relative peak areas and peak heights were used for quantification.

Results and Discussion Optimization of the ChromatographicProcedure As has already been mentioned we used reversed phase LC to study the separation of five antidepressants and the major metabolite of fluoxetine on a Nova-Pak Cls column. A standard solution containing 5 mg L 1 of each compound (solution A) was prepared by diluting the stock solutions of the six drugs. This solution was used throughout the optimization procedure. Initial experiments with the LC system using methanol or acetonitrile as organic modifier in the buffered mobile phase (70 mM phosphate buffer, pH 3) resulted in better separation than when acetonitrile was used. The chromatographic conditions optimized were pH, ionic strength, concentration of the organic solvent, and mobile phase flow rate. Optimization was Chromatographia 2002, 56, November (No. 9/10)

performed by modifying the value of the conditions studied while maintaining other conditions constant. The optimum value of the condition was chosen on the basis of sensitivity, peak resolution, and analysis time.

Effectof Mobile-Phase pH One means of varying peak resolution (Rs) and retention times (RT) is by changing mobile-phase pH, especially when the samples include ionizable acid or bases. The antidepressants studied in this work are compounds with basic properties in the usual working pH range of a Cls column (2.5 7.5). In this pH range the drugs were in the cationic form, because of the protonation of the amine groups in the molecules (Table I). To establish adequate selectivity we therefore varied the pH of the mobile phase and examined the resulting changes in RT, capacity factors (k), and Rs. Chromatograms of solution A were acquired by isocratic elution using mobile phases comprising 60:40 phosphate buffer (pH from 2.5 to 7.5)-acetonitrile at a flow-rate 1.5 mL min 1. Table II shows the effect of mobile phase pH on retention time R T a n d Rs. It is apparent that RT increases when the pH of the mobile phase is increased. This is because the compounds are less ionized at high pH and thus have more affinity for the stationary phase. Rs was always < 1.0 for pH > 5.5; pH 2.5 was selected as optimum because Rs was > 1.5 and analysis times were short.

Effectof the Ionic Strength of the Mobile Phase The ionic strength of aqueous mobile phases affects solute retention and selectivity, because this property affects sample distribution between the mobile and stationary phases. With this in mind the effect on k and Rs of the concentration of the phosphate buffer (from 30 to 200 mM) in the mobile phase was studied. The effect of ionic strength on k values is shown in Figure 1. It is apparent that for all the compounds studied k decreases quickly when the ionic strength of the mobile phase is increased, probably because when the buffer concentration is increased the pH is fixed more reliably and so, therefore, is the form of the ionized drug also. The resolution between peaks (Rs) decreased when the ionic strength of the mobile phase was increased Rs was > 1.5 only when the ionic strength was < 75 mM. 547

Clomipramine Fluoxetine x Norfluoxetine * Fluvoxamine 9 Citalopram 9

m

6-

LTrazodone

5o

4-

3, m

~ t~

D.

2

O

1 0

I

4

0 Figure

'

I

'

I

'

I

'

8 12 16 Ionic strength ( mM )

I

20

Effect of Mobile Phase Flow-Rate The optimum mobile phase flow-rate depends on the internal diameter of the column, the packing, the composition of the mobile phase, and the goal of the separation. To determine the optimum mobilephase flow-rate under our conditions the effect on RT, peak height and peak width, and Rs was studied. As expected when the mobile-phase flow-rate was increased RT and Rs of all the compounds decreased. A flow-rate of 1.5 mL min 1 was chosen as a compromise between Rs and analysis time, because this value also maintains good peak shape.

Summary of Optimum Conditions

1. Effect of ionic strength on capacity factor (k). 25-

CLO

2o

'

FLX

x, NFLX FLV 'A CIT

15-

TRA "~

the mobile phase was selected as 40%, because this always afforded Rs > 1.5 and short analysis times.

10 -

From the studies described above the chromatographic conditions used were Nova-Pack Cls column with 75 mM phosphate buffer (pH 2.5)-acetonitrile, 60:40, as mobile phase at a flow-rate of 1.5 mL min 1. The injection volume was 20 ixL and detection was performed at 230 nm. A chromatogram obtained from a standard mixture of the six antidepressants under these conditions is shown in Figure 3. It is apparent that good peak resolution was achieved in a short run time.

g O

Validation of the Procedure

5-

I

'

I

30

40

'

T

'

50

I

'

60

Acetonitrile % Figure

2. Effect of mobile-phase acetonitrile content on capacity factor (k).

A buffer concentration value of 75 mM was therefore selected as suitable, because it is certain to maintain pH 2.5 and furnishes good Rs, short analysis times, and good peak shape for all the compounds.

Effect of the Organic Solvent in the Mobile Phase One reason for the popularity of reversedphase HPLC is the relative simplicity of mobile phases that will separate a wide range of analytes of very different polarity. The mobile phases used in reversedphase chromatography are mixtures of 548

water or an aqueous buffer and an organic solvent methanol, acetonitrile, or tetrahydrofuran. In our work we used acetonitrile as organic mobile-phase modifier. In reversed-phase HPLC increasing the proportion of the organic component increases eluent strength and reduces retention time. The effect on solute retention and Rs of different amounts of acetonitrile (from 30 to 60%) in the mobile phase was studied. As expected, k for all the compounds decreased when the amount of acetonitrile was increased (Figure 2) but loss of peak resolution also occurred. The optimum amount of acetonitrile in Chromatographia 2002, 56, November (No. 9/10)

Validation was performed by measuring peak area and height at the wavelength of maximum absorbance of each of the antidepressants, to maximize sensitivity. The wavelengths used were 230 nm for FLX and NFLX, 210 nm for TRA, 238 nm for CIT, 250 nm for FLV, and 220 nm for CLO.

Stabili~/ of the Solutions The stability of standards solutions of TRA, CIT, FLV, FLX, NFLX, and CLO was determined by comparing the response factors (concentration/average peak area) of solutions stored at 4 ~ with those of freshly prepared solutions. The difference between the concentrations of freshly prepared solutions and those aged for 15 days was <0.2%, and the absorption spectra of the solutions were found to be unchanged after this period. The solution can therefore be used within this periOriginal

25t

od without the results being affected. Stock standard solutions were checked in the same way and found to be stable for a period of at least one month.

1-Solvent peak 2-Trazodone 3-Citalopram 4-Fluvoxamine 5-Norfluoxetine 6-Fluoxetine 7-Clomipramine

2O

Linearity Detector response (peak area and height) was linearly dependent on sample concentration over the range 1 11 mg L 1 (r/ = 11) for all the antidepressants studied. The linearity was determined from results from replicate injection of eleven different concentrations of each antidepressant over three different days. An analysis of variance ( A N O V A ) test was performed to compare the different regression lines obtained, to determine whether the data could be combined to enable estimation of the appropriate quantities by use of a comprehensive regression line [25, 26]. Because experimental values of F (between 0.2 and 1.6) were always lower than the theoretical values, it can be concluded than there were no significant differences between the regression lines compared, and that the data could be combined for estimation of the appropriate quantities by use of a comprehensive regression line. The equations of the calibration graphs proposed are listed in Table III, as also are the confidence intervals calculated for p = 0.05.

Repeatabilityand Reproducibi/ily Two different aqueous mixtures containing 5 mg L 1 of each of the six compounds were prepared and analysed on two different days by use of the optimized method. Repeatability was studied by performing a series of eight separations of one of these mixtures. Eight injections of standards were performed sequentially. The reproducibility was studied by performing eight separations of the other mixture 24 h later than analysis of the first, under the same conditions, and comparing the averages from the two series. In experiments performed to determine repeatability it is remarkable that relative standard deviations were <3.5% and < 1.7% for peak area and retention time, respectively, for all the compounds. In the reproducibility study the Snedecor F-test was used to compare two series of analyses performed on differences days. The experimental value of F is calculated for standard deviation ( [ e x p = SD12/ SD22= 1) and theoretical values of F are tabulated with degrees of freedom (n 1) for every day. If the experimental value of Original

2

15-6

.6 E

10--

0 '

0

I'1'1'1

I

1

2

3

7 4 5 6 Time (minutes)

8

9

10

Figure 3. Chromatogram obtained from antidepressants by use of the proposed method. Table III. Calibration graphs for peak area.

Drug

Peak area

TRA (210 nm) CIT (238rim) FLV (250 rim) NFLX (225 nm) FLX (230 nm) CLO (220 nm)

Equation

0-2

Y= 107909 (d- 10136) X+ 61307 (d- 61823) Y= 64081.7 (d- 5220) X+ 30947.4 (d- 31069.6) Y= 26252.34 (d- 1711) X+ 10148.1 (d- 7495) Y= 46427.3 (d- 4839) X+ 28035.8 (d- 28049) Y = 41968.2 (d- 4088) X+ 24434.2 (d- 24431) Y= 99630 (d- 7575) X+ 44519.3 (d- 57293)

0.9984 0.9982 0.9979 0.9963 0.9943 0.9959

Table IV. The composition of the solutions used to determine recovery.

Solution

TRA

CIT

FLV

NFL

FLX

CLO

M1 M2 M3 M4 M5 M6

1.00 2.98 4.98 6.98 8.95 11.08

3.00 5.00 7.00 8.96 11.04 1.00

5.00 7.00 9.00 11.00 1.00 3.00

7.00 9.00 11.00 1.00 3.00 5.00

9.01 10.97 0.99 2.99 4.98 6.98

11.00 1.00 3.00 5.00 7.00 9.00

F is less than the theoretical value there are no significant differences between series whereas if ftheor is less than the experimental value of F there are significant differences between series. Experimental values of F were between 1.2 and 2.5 for peak area and between 2.3 and 3.9 for retention time. Reproducibility was shown to be adequate because comparison of the averages by means of the Snedecor F-test did not reveal significant differences between both series for the significance level 0.05 (. = 8).

Chromatographia 2002,

56, N o v e m b e r (No.

9/10)

Accuracy To test the accuracy of the method different amounts several aliquots of the drugs were added to an analytical placebo prepared from the excipients manitol, maize starch, pregelatinized starch, anhydrous colloidal silica, sodium esterase fumarate, methyl hydroxypropylcellulose, poly(ethylene glycol) 6000, talcum powder, and titanium dioxide. The results obtained, by use of relative peak area, are summarized in Tables IV and V. It is apparent that excellent recoveries were always obtained. G o o d results were also obtained when peak height was used as analytical signal. 549

RT: 4.2 'frninl

R

210r~

01

Limit of Detection (LOD) and Limit of Quantification (LOG))

Deprax

Anafranil

I'

I T-f T--

207r~ I ....... I

----T--

I

~1

I T-"

320 nm

!-

0

I I I

4.[12

( .06

3213 nm

510

1.48 Time (rain)

Time (min)

Figure 4. Purity or homogeneity of peaks obtained from application of the method to two different pharmaceutical products. Table V. Recoveries (%) of the antidepressants (by measurement of peak area). Solution

Trazodone

Citalopram

Fluvoxamine

Obtained (mgL 1)

Recovery (%)

Obtained (mgL 1)

Recovery (%)

Obtained (mgL 1)

Recovery (%)

M5 M6

1.04 2.97 4.90 7.08 9.05 10.81

103.9 98.91 98.01 101.2 100.5 98.24

3.01 4.96 6.95 9.11 10.88 0.99

100.3 99.13 99.31 101.2 98.90 99.90

4.98 6.99 9.01 10.82 1.03 3.07

99.56 99.95 100.1 98.35 102.8 102.4

Solution

Norfluoxetine

Fluoxetine

Obtained (mgL 1)

Recovery (%)

Obtained (mgL 1)

Recovery (%)

Obtained (mgL 1)

Recovery (%)

7.10 8.90 11.32 0.98 2.95 4.95

101.5 98.91 102.8 98.43 98.34 98.97

8.95 10.99 1.02 2.93 4.94 7.04

99.45 99.95 101.9 97.81 98.78 100.6

11.08 1.04 3.08 4.88 7.06 9.20

100.7 104.1 102.6 97.63 100.8 102.3

M 1

M2 M3 M4

M 1

M2 M3 M4

M5 M6

Clomipramine

Table VI. Results from application of the method to pharmaceutical formulations. Preparation (presentation)

Active principle

Obtained (mg)*

Deprax, 50 mg (injectable) Deprax, 100 mg (tablets)

Trazodone Trazodone

50.19 -k 1.92 99.58 -k 0.39

Prisdal, 20mg (tablets) Seropram 20mg (tablets)

Citalopram Citalopram

19.91 -k 0.14 20.15 -k 0.21

Dumirox, 50mg (tablets) Dumirox, 100mg (tablets)

Fluvoxamine Fluvoxamine

49.08 -k 1.15 99.51 -k 0.21

Anafrani125mg (blister) Anafranil 10mg (pills) Anafrani125mg (pills) Anafrani175mg (pills)

Clomipramine Clomipramine Clomipramine Clomipramine

25.03 -k 2.32 10.71 -k 1.82 25.19 -k 0.84 76.45 -k 2.18

Adofen, 20mg (enveloped) Adofen, 20mg (tablets) Adofen, 20mg (capsules) Adofen, 20mg (solution) Reneuron, 20mg (solution) Reneuron 20mg (tablets) Reneuron 20mg, (capsules) Prozac 20mg (solution) Prozac 20mg (tablets) Prozac 20mg (capsules) Astrin 20mg (solution) Astrin 20mg (capsules)

Fluoxetine Fluoxetine Fluoxetine Fluoxetine Fluoxetine Fluoxetine Fluoxetine Fluoxetine Fluoxetine Fluoxetine Fluoxetine Fluoxetine

20.39 -k 1.29 19.85 -k 2.13 20.07 -k 2.30 20.17 -k 0.42 20.51 -k 1.06 20.18 -k 1.36 20.41 -k 1.44 20.20 -k 1.11 19.77 -k 1.01 20.16 -k 1.27 20.47 -k 0.48 20.51 -k 1.73

Mean value d- SD (n = 3).

550

Limits of detection a n d quantification (LOD a n d LOQ) were estimated in terms of the baseline noise. The LOD was defined as the sample c o n c e n t r a t i o n furnishing a peak of height three times the baseline noise a n d the LOQ was defined as t h a t furnishing a peak of height ten times the baseline noise. LOD a n d LOQ estim a t e d in this way were between 1.0 a n d 101xg L 1 a n d between 3.3 a n d 33.3 ixg L 1, respectively, for all the drugs. The LOQ was subsequently validated separately by analysis of six standards prep a r e d at the respective c o n c e n t r a t i o n s for T R A , CIT, FLV, N F L X , F L X , a n d CLO.

C h r o m a t o g r a p h i a 2002,

56, N o v e m b e r (No. 9/10)

Appficotions To d e m o n s t r a t e the usefulness of this c h r o m a t o g r a p h i c procedure the m e t h o d was used to determine the c o n c e n t r a t i o n s of the analytes in p h a r m a c e u t i c a l preparations, commercially available in Spain, which contain a n a n t i d e p r e s s a n t a n d other c o m p o n e n t s (excipients). Different procedures were used to prepare the different types of sample.

Capsules and Envelopes The c o n t e n t of a capsule or a n envelope was placed in a beaker, water (approx. 90 m L ) was added, a n d the mixture was shaken mechanically for 5 min. The suspension was placed a 100-mL volumetric flask a n d diluted to volume with water. A n sample of the s u p e r n a t a n t was diluted 1:40 (v/v) with water, to give a final concentration of the studied a n t i d e p r e s s a n t of approximately 5 m g L 1, by a d d i n g a k n o w n a m o u n t of a n aqueous stock solution of the drug being used as internal standard.

Tablets A tablet was g r o u n d in a m o r t a r , the solid was transferred to a beaker, water (approx. 90 m L ) was added, a n d the mixture was s h a k e n mechanically for 5 min. The suspension was placed in a 100-mL volumetric flask, diluted to volume with water, a n d centrifuged at 5000 rev m i n 1 for 5 min. A sample of the s u p e r n a t a n t was diluted 1:40 (v/v) with w a t e r to give a final c o n c e n t r a t i o n of the studied antidepressant of approximately 5 m g L 1, by add-

Original

ing a known amount of the aqueous stock solution of the drug being used as internal standard.

Solutionsand Injections The pharmaceutical solution (5 mL) was diluted to volume with water in a 100-mL volumetric flask. A sample from this solution was diluted 1:40 (v/v) with water to give a final concentration of the studied antidepressant of 5 mg L 1, by adding a known amount of the aqueous stock solution of the drug being used as internal standard. Determination of the amount of each antidepressant in the pharmaceutical formulations was performed in triplicate. As is apparent from Table VI the results were reproducible and recoveries were in the range 102.7 98.1% of the values declared by the manufacturers.

at the wavelength selected. The absorbance ratio for a pure compound is constant. A plot of this ratio against R T will give a square wave function of amplitude K. Any co-eluting impurity peak leads to a deviation from the flat ratio plot. The wavelengths were selected at the highest absorbance difference. For example, Figure 4 shows the absorbance ratio plot corresponding to a sample of two pharmaceutical formulations analysed (Anafranil and Deprax). In this example the absorbance ratio was investigated using 210/ 248 nm and 207/245 nm, respectively, as the selected wavelengths. The square wave function obtained demonstrates the purity of the peaks obtained. In all the applications of the proposed method similar results are obtained. There was, therefore, no interference from the excipients in the formulations.

Conclusions Peak Homogeneity

It is useful to investigate the purity of separated peaks because co-migration of peaks is possible in HPLC as in any other separation technique. Several techniques for validating peak purity have been proposed in the literature [27, 28]. In this work two techniques were used to validate the peak purity of the antidepressants in each pharmaceutical formulation studied.

Normalizing and Comparing Spectra from DifferentPeak Sections Absorption spectra acquired at the apex and the ascending and descending parts of the peak corresponding to the drugs present in each product were studied. The spectra were normalized and overlaid for graphical presentation. The similarity of the spectra demonstrated the high purity of the peaks obtained for the antidepressants in each product (Figure 4).

Absorbance at Two Wavelengths For a pure compound, the molar absorptivity, A, at a wavelength 21 is directly proportional to that at any other wavelength 22: A (21) = K A (22) where K characterizes the pure compound Original

In this work an easy, rapid, and sensitive HPLC method has been developed for determination of six antidepressants. The experimental results obtained indicate that the procedure proposed is specific and enables sensitive and accurate determination of the antidepressants in pharmaceutical formulations. It could be a valuable alternative to current official methods established by the European Pharmacopoeia.

Acknowledgements The authors thank the DGICYT of the 'Ministerio de Educacion y Ciencia' for supporting this study (project PB-970431).

References [1] Berzas, J.J.; Villasefior, M.J.; Contento, A.M.; Aguas, E. J. Chromatogr. Sci. 2000, 38, 200. [2] Holister, L.E. Drugs 1981,22, 129. [3] Dugas, M.; Zarifian, E.; Leheuzey, M.F.; Rovei, V.; Durand, G.; Morselli, P.L. Ther. Drug Monit. 1980,2,307. [4] Kasper, S.; H6flich, G.; Scholl, H.P.; M61ler, H.J. Hum. Psychopharmacol. 1994, 9, 1. [5] Preskorn, S.H.J. Clin. Psychiatry 1994, 55(Suppl. A), 6. [6] Lautz, R.I.; Farid, J.; Koons, K.Z.; Tenbargue J.B.; Bopp. R.I.J. Chromatogr. B. Biomed. Appl. 1993, 614(1), 175 179. Chromatographia 2002, 56, November (No. 9/10)

[7] Eap, C.B.; Gaillard, N.; Powell, K.; Baumann, P. J. Chromatogr. B. Biomed. Appl. 1996,682(2),265 272. [8] Eap, C.B.; Bouchoux, G.; Amey, M.; Cochard, N.; Savary, L.; Baumann, P. J. Chromatogr. Sci. 1998,36. [9] Meineke, I.; Schreeb, K.; Kress, I.; Gundert-Remy, U. Ther. Drug Monit. 1998, 20(1), 14 19. [10] Pichini, S.; Pacifini, R.; Altieri, I.; Pellegrini, M.; Zuccaro, P. J. Liq. Chromatogr. RelatedTechnol. 1996, 19(12), 1927 1935. [11] Torok-Both, G.A.; Baker, G.B.; Coutts, R.T.; Mackenna K. F.; Aspeslet. L.I.J. Chromatogr. Biomed. Appl. 1992, 579(1), 99 106. [12] Alvarez, J.C.; Bothua. D.; Collignon, I.; Advenier, C.; Spreux-Varoquaux, O. J. Chromatogr. B. Biomed. Appl. 1998, 707(1/2), 175 180. [13] Wong, S.H.Y.; Dellafera, S.S.; Fernandes, R.; Kranzler, H. J. Chromatogr. 1990, 499, 601. [14] Raggi, M.A.; Mandrioli, R.; Casamenti, G.; Bugamelli, F.; Volterra, V. J. Pharm. Biomed. Anal. 1998,18(1/2), 193 199. [15] Raggi, M.A.; Mandrioli, R.; Casamenti, G.; Volterra, V.; Desiderio, C.; Fanali, S. Chromatographia 1999, 50(7/8), 423 427. [16] Aymard, G.; Livi, P.; Pham, Y.T.; Diquet, B. J. Chromatogr. B 1997, 700, 183. [17] Kristoffersen, L.; Bugge, A.; Lundanes, E.; Slordal, L. J. Chromatography B 1999, 734, 229. [18] Lacassie, E.; Gaulier, J.M.; Marquet, P.; Rabatel, J.F.; Lachatre, G. J. Chromatogr. B 2000, 742,229. [19] Berzas, J.J.; Contento, A.M.; Villasefior, M.J.; Aguas, E. Anal. Chim. Acta 2000, 417, 169. [20] Desiderio, C.; Rudaz, S.; Raggli, M.A.; Fanali, S. Electrophoresis 1999, 20(17), 3432. [21] Berzas, J.J.; Contento, A.M. Chromatographia 2002,55, 369 373. [22] Goodnough, D.B.; Baker, G.B.; Coutts, R.T. J. Pharmacol. Toxicol. Methods 1995, 34, 143. [23] Berzas, J.J.; Contento, A.M.; Villasefior, M.J.J. Chromatogr. B. 2002, 769, 261 268. [24] Storch de Gracia, J.J.; Ed. Alhambra, 1987, pp. 122. [25] Lark, P.D.; Craven, B.R.; Bosworth, R.C.L. The Handling of Chemical Data, Pergamon Press, Oxford, 1968, Chapter 4, pp. 136 206. [26] Massart, D.L.; Vandeginste, B.G.M.; Deming, S.N.; Kaufmann, L. In: Chemometrics: A Textbook, 1st edn, Vandeginste, B.G.M.; Kaufmann, L., Eds, Oxford University Press, Oxford, 1988. [27] Fabre, H.; Le Bris, A.; Balchin, M.D.J. Chromatogr. A 1995, 697, 81. [28] Huber, L. Applications of Diode-Array Experience, Hewlett-Packard Publication no. 12-5953-2330, 1989. Received: Mar 5, 2002 Revised manuscripts received: May 7 and Jun 12, 2002 Accepted: Jul 8, 2002

551