Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 143 (2015) 281–287
Contents lists available at ScienceDirect
Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy journal homepage: www.elsevier.com/locate/saa
Application of derivative spectrophotometry under orthogonal polynomial at unequal intervals: Determination of metronidazole and nystatin in their pharmaceutical mixture Mohamed A. Korany ⇑, Heba H. Abdine, Marwa A.A. Ragab, Sara I. Aboras Pharmaceutical Analytical Chemistry Department, Faculty of Pharmacy, University of Alexandria, Egypt
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Derivative curves convolution using
orthogonal polynomials for unequal intervals. Specific polynomial order for each drug was selected based on its derivative curve. Solving spectral interferences as spectral overlap and irrelevant absorption. Simultaneous determination of a binary mixture in its pharmaceutical matrix. Specific and selective determination of each drug in mixtures.
a r t i c l e
i n f o
Article history: Received 19 April 2014 Received in revised form 17 December 2014 Accepted 29 January 2015 Available online 7 February 2015 Keywords: Orthogonal polynomials for unequal intervals First derivative curve convolution Spectrophotometric analysis Extensive curve overlap
a b s t r a c t This paper discusses a general method for the use of orthogonal polynomials for unequal intervals (OPUI) to eliminate interferences in two-component spectrophotometric analysis. In this paper, a new approach was developed by using first derivative D1 curve instead of absorbance curve to be convoluted using OPUI method for the determination of metronidazole (MTR) and nystatin (NYS) in their mixture. After applying derivative treatment of the absorption data many maxima and minima points appeared giving characteristic shape for each drug allowing the selection of different number of points for the OPUI method for each drug. This allows the specific and selective determination of each drug in presence of the other and in presence of any matrix interference. The method is particularly useful when the two absorption spectra have considerable overlap. The results obtained are encouraging and suggest that the method can be widely applied to similar problems. Ó 2015 Elsevier B.V. All rights reserved.
Introduction Background UV–VIS absorption spectrophotometry is a rapid, inexpensive and reliable technique for simultaneous determination of several
⇑ Corresponding author. Tel.: +20 100 680 4387. E-mail address: [email protected] (M.A. Korany). http://dx.doi.org/10.1016/j.saa.2015.01.076 1386-1425/Ó 2015 Elsevier B.V. All rights reserved.
analytes without the need of prior separation. However, for multi-component analysis, the presence of spectral interference and/or spectral overlapping, would represent a major problem for spectrophotometric quantitation [1]. Different chemometric methods aiming at solving this problem have been applied. In this respect, using of multivariate analysis for spectrophotometric determination of multi-component mixtures has been extensively used. The most widely used of which are, principal component regression (PCR) [2–5], partial least squares (PLS) [2–5], classical least square (CLS) [3,5], principal component
282
M.A. Korany et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 143 (2015) 281–287
analysis (PCA) [4], Multiple Linear Regression (MLR) [4], inverse least squares (ILS) [3], rank annihilation factor analysis (RAFA) [6], radial basis function-artificial neural network (RBF-ANN) [5] and principal component-radial basis function-artificial neural network (PC-RBF-ANN) [5]. Other methods have also been successfully used, including derivative [7] and derivative ratio spectrum [8]. Vierordt’s method has also been found beneficial [9]. Despite the simplicity and usefulness of these methods in solving certain problems, however, at certain conditions where the two analytes show high overlap and heavy spectral interferences, their applicability may be hindered. In this connection, a new approach was developed in the present work to the use of orthogonal polynomials [1]. The orthogonal functions method has been used to correct irrelevant absorption in multi-component spectrophotometric analysis depending on the use of orthogonal polynomials for the equally spaced intervals [10,11] and with the least-squares approach [10,11]. In the application of orthogonal polynomials for equal intervals to spectrophotometric analysis [10,11], a problem can be arisen where the polynomial integers (polynomial fundamental shapes) especially higher order polynomials may not exactly fit with the fine structure of the absorption curve. In this respect, Korany et. al. [1] introduced the use of orthogonal polynomials for unequal intervals (OPUI). The unequal interval wavelengths were selected according to the shape of a specified segment on the absorption curve. This method was applied for two component analysis where spectra of the analytes were highly overlapped [1]. The method was also extended to the multi-component [1] analysis and was considered to be highly advantageous over the use of orthogonal polynomials for equal intervals. In this paper, a new hybrid technique was developed by using first derivative D1 curve instead of absorbance curve to be convoluted using OPUI method. In OPUI, the use of first derivative technique is advantageous over normal absorbance one as it eliminates any constant interference. Also, it gives different maxima and minima which facilitate choosing different polynomial orders for each drug to be solved in the presence of each other with good accuracy and precision. This new approach is used for the determination of metronidazole (MTR) and nystatin (NYS) in their mixture and comparing the results with those obtained from first derivative method alone. Microbiological and HPLC-based techniques are the common pharmacopeial methods for determination of nystatin and metronidazole, respectively but they are usually very time-consuming relative to spectroscopic techniques. Thus, a method capable of rapidly obtain the content of nystatin and metronidazole, in a reliable and simultaneous manner, is of great interest to pharmaceutical industries. Different methods have been reported for the determination of MTR either in pharmaceutical dosage form [12–15] or in biological fluids [13,15]. Also for NYS, different methods have been reported either in pharmaceutical dosage form [16–18] or biological fluids [19,20]. To the best of our knowledge, only one method is reported for the analysis of NYS and MTR in their mixture. This method is based on diffuse reflectance NIR measurements and partial least squares regression in vaginal cream [21]. In general, NIR technique is characterized by producing spectra with low intensity, highly overlapping and broad bands. Thus, it is difficult to be used in measuring samples with low concentration. As a result, NIR is usually a chemometric based technique. Most of these chemometric methods, consume time as they need large number of training sets [22]. The proposed UV method lacks these disadvantages. Moreover, the chemometric method introduced in the study was specific and selective by fitting the fine structures
of the derivative curves for each drug and it can be easily performed using Excel software. Theory Derivative technique (D method) Application of derivative techniques to spectrophotometric data has become a well-established analytical method [23]. The elimination of interference by the use of derivative techniques depends on the fact that the first derivative of a constant function is zero. Consequently, a first derivative would eliminate constant interferences, thus:
D1 ¼
dA dk
ð1Þ
where, D1 is first derivative. Derivative technique followed by orthogonal polynomials for unequal intervals D1/OPUI (the proposed method) It is well known that the basis of harmonic analysis is that, a given function, for example D1 of absorbance curve, can be expanded in terms of an orthogonal function for unequal intervals. Thus,
D1 ¼ k0 K 0 þ k1 K 1 þ k2 K 2 þ k3 K 3 þ k4 K 4 þ k5 K 5 þ . . . þ kn K n
ð2Þ
where D1 is the first derivative of absorbance at a wavelength k that belongs to a set of n + 1 unequally spaced wavelengths. K0, K1, K2, K3, K4, K5, . . ., Kn are the polynomials corresponding to constant, linear, quadratic, cubic, quartic, quintic etc. and k0 ; k1 ; k2 ; k3 ; k4 ; k5 ; . . . ; kn are their corresponding coefficients. In view of the orthogonality, the polynomials can be calculated according to the method of Grandage [24], based on the following n X K ij ¼ 0
ð3Þ
i¼0 n X
K ij K iu ¼ 0 where j – u
ð4Þ
i¼0
where j, u represent different polynomial orders.Any coefficient, kij, can be calculated from a set of first derivative data by the equation n X kij ¼ D1K ij i¼0
, n X
K 2ij
ð5Þ
i¼0
The denominator of this equation is the sum of the squared individual values of Kij. After the construction of the convoluted polynomial curves of each analyte, the orthogonal polynomial coefficients, kij, at any k are proportional to D1 and concentration, thus
kij ¼ aj ca
ð6Þ
where aj = kij (1%, 1 cm) is a constant analogous to absorptivity of the pure compound, a, and ca is the concentration of the absorbing compound, a. In binary mixture system (providing that each component obeys Beer’s law and no interaction exists between the components), the orthogonal function for unequal interval could be applied for the analysis of binary component mixtures. Thus choosing the orthogonal polynomial order (quadratic, cubic, quartic, quintic etc.), number of points, wavelength intervals, the convoluted first derivative curve (orthogonal function coefficient, k, versus mean wavelength, km) could be derived in order to select the optimum km for a given component while the other component exhibits a zero coefficient at this wavelength and vice versa. The
M.A. Korany et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 143 (2015) 281–287
calculated orthogonal function coefficient of a mixture of M and N components at any km can be estimated by
kij ¼ ciM aMj þ ciN aNj
ð7Þ
In this equation, kij is the orthogonal polynomial coefficient of solution mixture of M and N (i) calculated at the mean wavelength (km), j, ciM or ciN is the concentration of component M or N in solution i and aMj or aNj is the k (1%, 1 cm) of component M or N at mean wavelength (km), j. Convolution of the first derivative curves of each drug using OPUI lead to the appearance of many maxima and minima points in the resulting convoluted curves. This allows the simultaneous determination of each component in the mixture using zero-crossing method by the proper choice of km for each drug. The selected km for each drug based on that, maximum response was obtained for the component under investigation with zero contribution of the other component or background interferences. At this selected km, the values of orthogonal polynomial were correlated to concentration as shown in Eq. (6). Experimental Instrumentation A Specord S600 spectrophotometer, associated with WinAspect software version 2.3, Analytik Jena AG, Germany was used. A KERN & SOHN GmbH balance was used and sonicator, J.P. SELECTA, S.A., was used.
283
Materials and reagents MTR and NYS were supplied by El-Amryia and Pharaonia Pharmaceuticals Co., Alexandria, Egypt, respectively. Amrizole-NÒ vaginal suppositories, batch number: 877466, (El-Amryia Pharmaceuticals Co., Alexandria, Egypt) were used. Each vaginal suppository was labeled to contain 500 mg MTR and 100,000 IU = 20.5 mg NYS [16]. Methanol used was of analytical grade. Standard solution and calibration graphs Stock solutions were prepared by dissolving MTR and NYS in methanol to obtain stock solutions of 0.1 mg mL1. The stock solutions of MTR and NYS were further diluted with methanol to obtain working standard solutions of suitable concentrations in the range of (1–25) and (1–15) lg mL1, respectively. Preparation of vaginal suppository solution Five Amrizole-NÒ vaginal suppositories were accurately weighed and finely powdered in a mortar. An amount of the suppository mass equivalent to one suppository content (500 mg of MTR and 100,000 I.U NYS which is equivalent to 20.5 mg [16]) was dissolved in 60 mL of methanol. After 10 min of sonication at room temperature, the solution was quantitatively transferred to a 100-mL volumetric flask. The mixture was completed to volume with methanol. The excipient – the base of suppository – is Poly Ethylene Glycol, so filtration was not needed as it is completely soluble in methanol. Further dilution was conducted with methanol to produce solution containing MTR 25 lg mL1 and NYS 1.025 lg mL1. Spectrophotometric measurements and data treatment Application of first derivative technique D1 to the spectrophotometric data The absorbance was measured in 1-cm cell using methanol as blank and covering the wavelength range 200–400 nm at 0.5 nm intervals. The obtained absorption curve was highly overlapped thus the first derivative was calculated as shown in Fig. 1a and b. Using D1 method, MTR can be determined at 290 nm which shows minimum contribution of NYS. On the other side NYS can be determined at 261 nm which shows minimum contribution of MTR as shown in Fig. 1b. This method was used as a comparable method with the proposed one. Application of OPUI to the D1 data (D1/OPUI) For first derivative absorption curves in methanol of MTR and NYS, the polynomials can be calculated according to the selected wavelength intervals and the mixture can be resolved by zero crossing as follows: For metronidazole (MTR). The segment abcd (Fig. 2a) shows that a K3 polynomial (cubic) can fit the derivative absorption curve. Therefore, the wavelengths 247 (min), 290 (max), 330 (min) and 360 (shoulder) were selected. There wavelengths were spaced by the intervals 0, 43, 40 and 30 nm as shown in Table 1. The polynomials K1, K2 and K3 can be calculated [24,25] as presented in Table 1.
Fig. 1. Absorption spectra of MTR 25 lg mL1 ( ) and NYS 1 lg mL1 ( methanol (a), and their corresponding first derivative curves (b).
) in
For nystatin (NYS). The segment efghij (Fig. 2a) shows that a K0 5 polynomial (quintic) can fit the derivative absorption curve. Therefore, the wavelengths 288 (max), 294 (min), 300 (max), 308 (min), 316 (max) and 323 (min) were selected. There wavelengths were spaced by the intervals 0, 6, 6, 8, 8 and 7 nm Table 1. The polyno-
284
M.A. Korany et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 143 (2015) 281–287
(Fig. 2b and c). The coefficients were calculated according to Eq. (5) as follows:
k3 ¼ ½D1247 ð12801:4Þ þ D1290 ð42880Þ þ D1330 ð51834:8Þ þD1360 ð21756:2Þ=5162748971
ð8Þ
0
k5 ¼ ½D1288 ð14231:06Þ þ D1294 ð62498:79Þ þD1300 ð94869:4Þ þ D1308 ð74867:93Þ þD1316 ð36485:78Þ þ D1323 ð8219:52Þ=20112804470:27 ð9Þ As mentioned in ‘‘Derivative technique followed by orthogonal polynomials for unequal intervals D1/OPUI (the proposed method)’’, the concentration of MTR and NYS were calculated by zero crossing method. As shown in Figs. 2b and 2c, the selected km 301.5 and 241.5 nm were suitable for determination of MTR and NYS, respectively in their mixtures. Each km is selected on the basis that the drug shows maximum response at this k with zero contribution of the other. Results and discussion Optimization The absorption spectra of MTR and NYS in methanol are presented in (Fig. 1a). The drugs show considerable spectral overlap. In vaginal suppository form, the two drugs are formulated as MTR, major component, (500 mg/vaginal suppository) and NYS, minor component, (20.5 mg/vaginal suppository). Accordingly, determination of NYS (minor component) using the conventional absorbance data will lead to erroneous results. In the present work, the assay of the components in a mixture was investigated using the proposed D1/OPUI method and the comparable D1 method. In D1/OPUI method, the first derivative absorption curves for the two analytes (Fig. 2a) were convoluted using the values of the orthogonal polynomials; K3 and K0 5 with their wavelength intervals, selected from Table 1 for MTR and NYS. Two coefficients; k3 and k0 5 were selected at the corresponding km 301.5 and 241.5 nm, respectively, and the concentrations of the drugs were determined by zero crossing at the selected wavelengths (Fig. 2b and c). In D1 method, the concentrations of the drugs were determined, using first derivative absorption curves, at 290 and 261 nm for MTR and NYS, respectively. Method validation The method was validated according to ICH guidelines [26]. The following validation characteristics were addressed: Fig. 2. First derivative absorption spectra of MTR 25 lg mL1 ( ) and NYS 1 lg mL1 ( ) in methanol (a) and its convolution (b) and (c) derived in the form: 0 k3 and k 5, respectively, using selected orthogonal polynomials for unequal intervals.
mials K0 1, K0 2, K0 3, K0 4 and K0 5 can be calculated [24] as presented in Table 1. The polynomials given in Table 1 were calculated, according to the method of Grandage [24], using Excel software. All the generated polynomials fulfill the requirements of Eqs. (2)–(4). Convolution of the D1 curves was done using the values of the orthogonal polynomials; K3 and K0 5 and their corresponding wavelength intervals, selected from Table 1 for MTR and NYS (Fig. 2a). Two coefficients; k3 and k0 5 were selected at the corresponding km 301.5 and 241.5 nm, to solve the MTR and NYS respectively
Linearity Under the above selected parameters of the chemometric method, the linearity of the proposed method was evaluated through the analysis of seven serial concentrations of each drug. The D1 method and the coefficients of the D1/OPUI method were plotted as a function of the corresponding concentration and the calibration equation was calculated using the least squares regression method. Table 2 summarizes the regression and statistical parameters of the studied drugs using the derivative and the proposed chemometric method. As shown in Table 2, comparing the results of the two methods, the regression analysis verifies the good linearity of the proposed chemometric method with slight enhancement in the regression parameters over the derivative method. This slight enhancement is indicated by the increase of correlation coefficient (r) values, the decrease in values of the standard deviation of the intercept (Sa), standard deviation of the
285
M.A. Korany et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 143 (2015) 281–287
Table 1 P P K 2ij M; K 2ij N, for the determination of two component mixtures of metronidazole (MTR) and nystatin Calculation of the polynomials, Kj, and their corresponding sum, (NYS), respectively. MTR
NYS
k Set (I) (nm)
Interval (nm)
247 290 330 360
0 43 40 30
P
K2
K3
k Set (II) (nm)
Interval (nm)
K01
K02
K03
K04
K05
59.75 16.75 23.25 53.25
1313.197 1652.33 1090.95 1430.079
12801.4 42880 51834.8 21756.2
0 6 6 8 8 7
7226.75
7689968
5162748971
288 294 300 308 316 323 P
16.83 10.83 4.83 3.17 11.16 18.16 888.25
159.19 15.35 117.90 142.62 39.34 156.02 85708.07
1107.79 1215.92 1073.91 711.00 1503.42 1032.38 7690543.98
4933.40 12419.73 4836.21 10911.12 12485.20 4224.20 494753680.33
14231.06 62498.79 94869.40 74867.93 36485.78 8219.52 20112804470.27
K1
K 2ij M
K 2ij N
Table 2 Linear regression and statistical parameters for the determination of metronidazole (MTR) and nystatin (NYS) using the derivative method (D1) and the proposed chemometric method (D1/OPUI). ra
ab
|b|c
Sy/xd
6.48 104 3.92 104
2.25 104 6.52 105
1.45 104 4.55 105
1.35 10-5 5.82 10-6
Orthogonal polynomial of unequal interval on derivative data (D1/OPUI method) MTR (k0 3) 301.5 nm 0.9997 9.4 1010 6.21 109 1.46 109 NYS (k0 5) 241.5 nm 0.9998 1.25 108 3.19 108 3.45 109
8.64 1010 2.41 109
6.45 1011 3.08 1010
Derivative technique (D1 method) MTR 290 nm 0.9989 1.48 104 NYS 261 nm 0.9995 3.6 10-5
a b c d e f g h i
S ae
Sbf
Fg
LODh (lg mL1)
LOQi (lg mL1)
2306.89 4547.03
0.416 0.461
1.260 1.396
9264.35 10762.78
0.310 0.073
0.938 0.222
Correlation coefficient. Intercept. Absolute slope. Standard deviation of residuals. Standard deviation of intercept. Standard deviation of slope. Variance ratio, equals the mean of squares due to regression divided by the mean of squares about regression (due to residuals). Limit of detection. Limit of quantitation.
slope (Sb) and standard deviation of the residuals (Sy/x) with high F values [27]. Detection and quantitation limits Limit of detection (LOD) and limit of quantitation (LOQ) are calculated according to ICH guidelines [26] based on the standard deviation of the blank and the Slope as shown in Table 2. As seen from the table, lower values of LOD and LOQ were obtained using the proposed method than the D1 method. Accuracy and precision In order to assess the accuracy, as mean percentage recovery, and the precision, as percentage relative standard deviation (RSD%), MTR and NYS were assayed in different laboratory-prepared mixtures of different ratios including percentage similar to that present in dosage form (25 MTR:1 NYS). The results presented in Table 3 showed that erroneous results for the two analytes were obtained when applying conventional first derivative method. This is attributed to heavy spectral overlapping of MTR during the assay of minor component (NYS). That is why the error relatively increased in mixture number 4, representing the same ratio in the dosage form (Table 3). With the application of orthogonal polynomial for unequal interval (D1/OPUI) method, such error has been corrected (Table 3). The results obtained using the (D1/OPUI) method are not affected by any interferences contributing to coefficients other than those involved in the assay coefficients. Thus in the above assay, the cubic k3 and quintet k0 5 components (here ‘component’ is used in the mathematical sense) are not eliminated, which means in particular that, constant, linear and other components of the irrelevant absorption are all rejected. Nevertheless, in
common, irrelevant absorption contributes much more to the constant and linear components of the total absorption than any others. As shown in Table 3 great enhancement of accuracy and precision when using the proposed (D1/OPUI) method compared with D1 method. The accuracy in Table 3, represented as mean percentage recovery, was enhanced from 107.28 to 100.00 for NYS. The Intra-day precision in Table 3, calculated as RSD%, was improved from 4.25 to 0.67 and from 14.10 to 0.56 for MTR and NYS, respectively. Also, the inter-day precision, calculated as RSD%, was improved from 13.71 to 0.77 and from 13.89 to 0.81 for MTR and NYS, respectively (Table 3). A comparison was done between the derivative method alone and the proposed D1/OPUI method for analyzing each drug in different synthetic mixtures. This was done by applying student’s t-test and variance ratio F-test to check if there is a significant difference between the two methods or not. As can be seen in Table 3, significant difference between the two methods variances was found. This was confirmed as the calculated F-values exceeded the theoretical ones (F-theoretical and calculated values are shown in table for each drug). While no significant difference between the two methods means was noticed. This confirms that the proposed D1/OPUI is valid for analyzing both drugs in different synthetic mixtures ratios. Specificity. Specificity of the method was tested by analyzing each drug in presence of the other (Table 3) and in presence of dosage form excipients (Table 4). Specificity of the proposed D1/OPUI method was confirmed by analyzing MTN and NYS with good accuracy (98–102%) and precision (less than 2%) in presence of each other and the in their dosage form (Tables 3 and 4).
286
M.A. Korany et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 143 (2015) 281–287
Table 3 Evaluation of the accuracy and precision for the determination of two-component mixtures of metronidazole (MTR), nystatin (NYS) in laboratory made mixtures using the proposed D1/OPUI method and D1 method. Mix. No.
Nominal conc.
(a) Accuracy and intraday precision
(a) Accuracy and interday precision
(lg mL1)
MTR
MTR
MTR 1 2 3 4
a b c d e f
NYS
10 10 5 15 15 1 25 1 Mean%a Er%b SDc RSD(%)d t calculatede F calculatedf
NYS
NYS
D1
D1/OPUI
D1
D1/OPUI
D1
D1/OPUI
D1
D1/OPUI
93.31 103.39 97.13 98.64 98.12 1.88 4.17 4.25 0.96 39.13
99.66 99.64 100.22 101.06 100.15 0.15 0.67 0.67
95.20 95.75 111.03 127.12 107.28 7.28 15.13 14.10 0.96 738.55
100.06 100.66 99.98 99.30 100.00 0.00 0.56 0.56
105.40 77.45 104.96 101.15 97.24 2.76 13.33 13.71 0.41 298.35
101.13 99.64 99.42 99.79 99.99 0.01 0.77 0.77
90.11 108.36 111.03 127.12 109.15 9.15 15.16 13.89 1.22 345.66
99.27 101.01 99.98 99.30 99.89 0.11 0.81 0.81
The mean of all recoveries of different concentration in the same method for each drug. Percentage relative error. Standard deviation of the recoveries of different concentration in the same method for each drug. Percentage relative standard deviation. For t-test for unequal variances, the critical t value at 3 degrees of freedom is 3.18 for two tailed test (P = 0.05). For F-test; the critical value F3,3 = 15.44 for two tailed test (P = 0.05).
Table 4 Assay of two-component mixtures, metronidazole (MTR) and nystatin (NYS), in vaginal suppositories using the proposed D1/OPUI method and D1 method. Sample No.
Found (%) MTR
1 2 3 4 5 6 Mean SD RSD (%) t calculateda F calculatedb
NYS
D1 method
D1/OPUI method
D1 method
D1/OPUI method
105.22 94.72 104.60 91.64 105.22 102.13 100.59 5.93 5.89 0.05 17.86
99.77 101.90 101.74 99.30 102.25 99.25 100.70 1.40 1.39
165.92 75.71 53.15 133.19 83.46 108.33 103.29 41.25 39.94 0.17 855.42
100.82 100.07 100.21 98.74 99.55 102.88 100.38 1.41 1.41
a For t-test for unequal variance, there are 6 & 5 degrees of freedom for MTR & NYS respectively, the critical values of t are 2.45 and 2.57 for MTR and NYS respectively at the two tailed test (P = 0.05). b For F-test; the critical value F5,5 = 7.146 for two tailed test (P = 0.05).
Analysis of pharmaceutical preparation In vaginal suppository assay, with regard to accuracy of the major component, MTR, the results did not improve in D1/OPUI compared with D1 method and that may be contributed to its high concentration in dosage form that even interferences from the minor component, NYS, may not be noticed. On the contrary, upon the assessment of precision (calculated as RSD%), high RSD% values were obtained. The RSD% decreased from 5.89 to 1.39 upon applying D1 method to D1/OPUI method, respectively, which indicates the benefit of the proposed chemometric method as shown in Table 4. The improvement of the minor component (NYS) was obvious for the proposed method in both accuracy as mean percentage found (from 103.29% to 100.38%) and precision as RSD% (from 39.94 to 1.41). Table 4 also represents a statistical comparison between the proposed (D1/OPUI) method and the D1 method for the assay of MTR and NYS in their vaginal suppository. This was done using the student’s t- and variance ratio F-tests. The results in Table 4 indicate that there is a significant difference between the two methods as the calculated F-values for each compound exceeded the theoretical ones. As a result, the D1/OPUI is the method of choice for analyzing the drugs in their dosage form (Table 4).
Comparing the results of the proposed D1/OPUI method with the only reported method for the mixture [21], no significant difference between the two methods variances was found. This was confirmed as the calculated F-values (F-calculated values are 1.88 for MTR and 2.90 for NYS) didn’t exceed the theoretical ones (Ftheoretical values for the two tailed test at P = 0.05, F5,4 = 9.36 for MTR and F4,5 = 7.39 for NYS). In contrast, the calculated t-value of NYS exceeded the theoretical one. This indicates significant difference between the two methods in analyzing the minor component (t-calculated values, t = 1.74 for MTR and 2.52 for NYS) (t-theoretical values at df = 9 for the two tailed test at P = 0.05, is 2.26 for MTR & NYS).
Conclusion The proposed chemometric method can be readily applied for simultaneous determination of MTR and NYS in laboratory made mixtures and vaginal suppositories using widely available UV spectrophotometers. The method is specific and there is no interference from any of the matrix components. The proposed method is selective, sensitive. Compared to different chromatographic methods, the proposed method is economic and inexpensive. It
M.A. Korany et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 143 (2015) 281–287
could be used in the routine analysis of this mixture in its pharmaceutical preparation. References [1] M.A. Korany, M. Bedair, F.A. El-Yazbi, Use of orthogonal polynomials for unequal intervals to eliminate interference in spectrophotometric analysis. Simultaneous determination of ephedrine hydrochloride and diphenhydramine hydrochloride in two-component mixtures, Analyst 111 (1) (1986) 41–44. [2] V. Ragupathy, S. Arcot, Simultaneous spectrophotometric determination of diacerein and aceclofenac in tablets by chemometric methods, Int. Res. J. Pharm. 4 (3) (2013) 211–214. [3] P. Bhamre, A. Pathak, S. Rajput, Simultaneous determination of doxylamine succinate, pyridoxine hydrochloride and folic acid by chemometric spectrophotometry, Int. J. Pharma Bio Sci. 4 (1) (2013) 738–749. [4] S. Dinç, Ö.A. Dönmez, B. Asçi, A.E. Bozdogan, Spectrophotometric multicomponent resolution of a tablet formulation containing lisinopril and hydrochlorothiazide by multivariate calibration methods, Asian J. Chem. 25 (2) (2013) 999–1002. [5] Y. Ni, Y. Wang, S. Kokot, Multicomponent kinetic spectrophotometric determination of pefloxacin and norfloxacin in pharmaceutical preparations and human plasma samples with the aid of chemometrics, Spectrochim. Acta Part A Mol. Biomol. Spectrosc. 70 (5) (2008) 1049–1059. [6] M. Bahram, M. Mabhooti, Rank annihilation factor analysis for multicomponent kinetic-spectrophotometric determination using difference spectra, J. Chemom. 23 (5) (2009) 236–247. [7] A. Varghese, L. George, Simultaneous first order derivative spectrophotometric determination of vanadium and zirconium in alloy steels and minerals, Spectrochim. Acta Part A Mol. Biomol. Spectrosc. 95 (2012) 46–52. [8] P.P. Goti, P.B. Patel, Development and validation of ratio-derivative spectrophotometric method for simultaneous estimation of gabapentin, methylcobalamin and alpha lipoic acid in tablet formulation, J. Pharm. Res. 6 (6) (2013) 609–614. [9] P. Giriraj, T. Sivakkumar, Simultaneous estimation of dutasteride and tamsulosin hydrochloride in tablet dosage form by vierordt’s method, Arabian J. Chem. (in press). http://dx.doi.org/10.1016/j.arabjc.2013.07.013 [10] M.A.-H. Elsayed, M.A. Korany, M.M. Bedair, H. Mahgoub, E.A. Korany, Computer assisted spectrophotometric analysis of three component mixtures (caffeine mixtures and codeine mixtures), Anal. Lett. 25 (2) (1992) 239–252. [11] M.A. Korany, M.A.-H. Elsayed, M. Bedair, H. Mahgoub, E.A. Korany, An interactive computer program for multicomponent spectrophotometric analysis using orthogonal polynomials, Anal. Lett. 23 (3) (1990) 507–527. [12] H.M. Maher, R.M. Youssef, Development of validated chromatographic methods for the simultaneous determination of metronidazole and spiramycin in tablets, Chromatographia 69 (3–4) (2009) 345–350.
287
[13] M.M. Issa, R.a.M. Nejem, A.M.A. Shanab, N.T. Shaat, Kinetic spectrophotometric H-point standard addition method for the simultaneous determination of diloxanide furoate and metronidazole in binary mixtures and biological fluids, Spectrochim. Acta Part A Mol. Biomol. Spectrosc. 114 (2013) 592–598. [14] M.A. Mahrouse, E.F. Elkady, Validated spectrophotometric methods for the simultaneous determination of ciprofloxacin hydrochloride and metronidazole in tablets, Chem. Pharm. Bull. 59 (12) (2011) 1485–1493. [15] A.A. Salem, H.A. Mossa, Method validation and determinations of levofloxacin, metronidazole and sulfamethoxazole in an aqueous pharmaceutical, urine and blood plasma samples using quantitative nuclear magnetic resonance spectrometry, Talanta 88 (2012) 104–114. [16] H.M. Heneedak, I. Salama, S. Mostafa, M. El-Sadek, HPLC and chemometric methods for the simultaneous determination of miconazole nitrate and nystatin, J. Chromatogr. Sci. 50 (10) (2012) 855–861. [17] M. Pendela, G. Kahsay, I. Baekelandt, A. Van Schepdael, E. Adams, Simultaneous determination of lidocaine hydrochloride, hydrocortisone and nystatin in a pharmaceutical preparation by RP-LC, J. Pharm. Biomed. Anal. 56 (3) (2011) 641–644. [18] J.M. Lemus Gallego, J. Pérez Arroyo, Micellar electrokinetic capillary chromatography as an alternative method for determination of hydrocortisone and its most important associated compounds in local pharmaceutical preparations, Chromatographia 56 (7–8) (2002) 455–462. [19] E. Scheuch, T. Giessmann, W. Siegmund, Quantitative determination of nystatin in human plasma using LC–MS after inhalative administration in healthy subjects, J. Chromatogr. B 844 (1) (2006) 84–88. [20] J.M. Llabot, D.A. Allemandi, R.H. Manzo, M.R. Longhi, HPLC method for the determination of nystatin in saliva for application in clinical studies, J. Pharm. Biomed. Anal. 45 (3) (2007) 526–530. [21] S.C. Baratieri, J.M. Barbosa, M.P. Freitas, J.A. Martins, Multivariate analysis of nystatin and metronidazole in a semi-solid matrix by means of diffuse reflectance NIR spectroscopy and PLS regression, J. Pharm. Biomed. Anal. 40 (1) (2006) 51–55. [22] Q. Wang, D. Ma, J.P. Higgins, Analytical method selection for drug product dissolution testing, Dissolution Technol. (2006). [23] M.A. Korany, A.M. Wahbi, S. Mandour, M.A. Elsayed, Determination of certain drugs in multicomponent formulations by first derivative ultraviolet spectrophotometry, Anal. Lett. 18 (1) (1985) 21–34. [24] A. Grandage, Orthogonal coefficients for unequal intervals, Biometrics 14 (2) (1958) 287–289. [25] K.C. Peng, The Design and Analysis of Scientific Experiments, Addison-Wesley, London, 1967. [26] ICH, Q2 (R1), Validation of analytical procedures. Text and methodology, International Conference on Harmonization, Geneva, 2005. [27] P. Armitage, G. Berry, Statistical Methods in Medical Research, Blackwell Science Ltd., Oxford, England, 1994. pp. 283–285.
Contents lists available at ScienceDirect
Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy journal homepage: www.elsevier.com/locate/saa
Application of derivative spectrophotometry under orthogonal polynomial at unequal intervals: Determination of metronidazole and nystatin in their pharmaceutical mixture Mohamed A. Korany ⇑, Heba H. Abdine, Marwa A.A. Ragab, Sara I. Aboras Pharmaceutical Analytical Chemistry Department, Faculty of Pharmacy, University of Alexandria, Egypt
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Derivative curves convolution using
orthogonal polynomials for unequal intervals. Specific polynomial order for each drug was selected based on its derivative curve. Solving spectral interferences as spectral overlap and irrelevant absorption. Simultaneous determination of a binary mixture in its pharmaceutical matrix. Specific and selective determination of each drug in mixtures.
a r t i c l e
i n f o
Article history: Received 19 April 2014 Received in revised form 17 December 2014 Accepted 29 January 2015 Available online 7 February 2015 Keywords: Orthogonal polynomials for unequal intervals First derivative curve convolution Spectrophotometric analysis Extensive curve overlap
a b s t r a c t This paper discusses a general method for the use of orthogonal polynomials for unequal intervals (OPUI) to eliminate interferences in two-component spectrophotometric analysis. In this paper, a new approach was developed by using first derivative D1 curve instead of absorbance curve to be convoluted using OPUI method for the determination of metronidazole (MTR) and nystatin (NYS) in their mixture. After applying derivative treatment of the absorption data many maxima and minima points appeared giving characteristic shape for each drug allowing the selection of different number of points for the OPUI method for each drug. This allows the specific and selective determination of each drug in presence of the other and in presence of any matrix interference. The method is particularly useful when the two absorption spectra have considerable overlap. The results obtained are encouraging and suggest that the method can be widely applied to similar problems. Ó 2015 Elsevier B.V. All rights reserved.
Introduction Background UV–VIS absorption spectrophotometry is a rapid, inexpensive and reliable technique for simultaneous determination of several
⇑ Corresponding author. Tel.: +20 100 680 4387. E-mail address: [email protected] (M.A. Korany). http://dx.doi.org/10.1016/j.saa.2015.01.076 1386-1425/Ó 2015 Elsevier B.V. All rights reserved.
analytes without the need of prior separation. However, for multi-component analysis, the presence of spectral interference and/or spectral overlapping, would represent a major problem for spectrophotometric quantitation [1]. Different chemometric methods aiming at solving this problem have been applied. In this respect, using of multivariate analysis for spectrophotometric determination of multi-component mixtures has been extensively used. The most widely used of which are, principal component regression (PCR) [2–5], partial least squares (PLS) [2–5], classical least square (CLS) [3,5], principal component
282
M.A. Korany et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 143 (2015) 281–287
analysis (PCA) [4], Multiple Linear Regression (MLR) [4], inverse least squares (ILS) [3], rank annihilation factor analysis (RAFA) [6], radial basis function-artificial neural network (RBF-ANN) [5] and principal component-radial basis function-artificial neural network (PC-RBF-ANN) [5]. Other methods have also been successfully used, including derivative [7] and derivative ratio spectrum [8]. Vierordt’s method has also been found beneficial [9]. Despite the simplicity and usefulness of these methods in solving certain problems, however, at certain conditions where the two analytes show high overlap and heavy spectral interferences, their applicability may be hindered. In this connection, a new approach was developed in the present work to the use of orthogonal polynomials [1]. The orthogonal functions method has been used to correct irrelevant absorption in multi-component spectrophotometric analysis depending on the use of orthogonal polynomials for the equally spaced intervals [10,11] and with the least-squares approach [10,11]. In the application of orthogonal polynomials for equal intervals to spectrophotometric analysis [10,11], a problem can be arisen where the polynomial integers (polynomial fundamental shapes) especially higher order polynomials may not exactly fit with the fine structure of the absorption curve. In this respect, Korany et. al. [1] introduced the use of orthogonal polynomials for unequal intervals (OPUI). The unequal interval wavelengths were selected according to the shape of a specified segment on the absorption curve. This method was applied for two component analysis where spectra of the analytes were highly overlapped [1]. The method was also extended to the multi-component [1] analysis and was considered to be highly advantageous over the use of orthogonal polynomials for equal intervals. In this paper, a new hybrid technique was developed by using first derivative D1 curve instead of absorbance curve to be convoluted using OPUI method. In OPUI, the use of first derivative technique is advantageous over normal absorbance one as it eliminates any constant interference. Also, it gives different maxima and minima which facilitate choosing different polynomial orders for each drug to be solved in the presence of each other with good accuracy and precision. This new approach is used for the determination of metronidazole (MTR) and nystatin (NYS) in their mixture and comparing the results with those obtained from first derivative method alone. Microbiological and HPLC-based techniques are the common pharmacopeial methods for determination of nystatin and metronidazole, respectively but they are usually very time-consuming relative to spectroscopic techniques. Thus, a method capable of rapidly obtain the content of nystatin and metronidazole, in a reliable and simultaneous manner, is of great interest to pharmaceutical industries. Different methods have been reported for the determination of MTR either in pharmaceutical dosage form [12–15] or in biological fluids [13,15]. Also for NYS, different methods have been reported either in pharmaceutical dosage form [16–18] or biological fluids [19,20]. To the best of our knowledge, only one method is reported for the analysis of NYS and MTR in their mixture. This method is based on diffuse reflectance NIR measurements and partial least squares regression in vaginal cream [21]. In general, NIR technique is characterized by producing spectra with low intensity, highly overlapping and broad bands. Thus, it is difficult to be used in measuring samples with low concentration. As a result, NIR is usually a chemometric based technique. Most of these chemometric methods, consume time as they need large number of training sets [22]. The proposed UV method lacks these disadvantages. Moreover, the chemometric method introduced in the study was specific and selective by fitting the fine structures
of the derivative curves for each drug and it can be easily performed using Excel software. Theory Derivative technique (D method) Application of derivative techniques to spectrophotometric data has become a well-established analytical method [23]. The elimination of interference by the use of derivative techniques depends on the fact that the first derivative of a constant function is zero. Consequently, a first derivative would eliminate constant interferences, thus:
D1 ¼
dA dk
ð1Þ
where, D1 is first derivative. Derivative technique followed by orthogonal polynomials for unequal intervals D1/OPUI (the proposed method) It is well known that the basis of harmonic analysis is that, a given function, for example D1 of absorbance curve, can be expanded in terms of an orthogonal function for unequal intervals. Thus,
D1 ¼ k0 K 0 þ k1 K 1 þ k2 K 2 þ k3 K 3 þ k4 K 4 þ k5 K 5 þ . . . þ kn K n
ð2Þ
where D1 is the first derivative of absorbance at a wavelength k that belongs to a set of n + 1 unequally spaced wavelengths. K0, K1, K2, K3, K4, K5, . . ., Kn are the polynomials corresponding to constant, linear, quadratic, cubic, quartic, quintic etc. and k0 ; k1 ; k2 ; k3 ; k4 ; k5 ; . . . ; kn are their corresponding coefficients. In view of the orthogonality, the polynomials can be calculated according to the method of Grandage [24], based on the following n X K ij ¼ 0
ð3Þ
i¼0 n X
K ij K iu ¼ 0 where j – u
ð4Þ
i¼0
where j, u represent different polynomial orders.Any coefficient, kij, can be calculated from a set of first derivative data by the equation n X kij ¼ D1K ij i¼0
, n X
K 2ij
ð5Þ
i¼0
The denominator of this equation is the sum of the squared individual values of Kij. After the construction of the convoluted polynomial curves of each analyte, the orthogonal polynomial coefficients, kij, at any k are proportional to D1 and concentration, thus
kij ¼ aj ca
ð6Þ
where aj = kij (1%, 1 cm) is a constant analogous to absorptivity of the pure compound, a, and ca is the concentration of the absorbing compound, a. In binary mixture system (providing that each component obeys Beer’s law and no interaction exists between the components), the orthogonal function for unequal interval could be applied for the analysis of binary component mixtures. Thus choosing the orthogonal polynomial order (quadratic, cubic, quartic, quintic etc.), number of points, wavelength intervals, the convoluted first derivative curve (orthogonal function coefficient, k, versus mean wavelength, km) could be derived in order to select the optimum km for a given component while the other component exhibits a zero coefficient at this wavelength and vice versa. The
M.A. Korany et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 143 (2015) 281–287
calculated orthogonal function coefficient of a mixture of M and N components at any km can be estimated by
kij ¼ ciM aMj þ ciN aNj
ð7Þ
In this equation, kij is the orthogonal polynomial coefficient of solution mixture of M and N (i) calculated at the mean wavelength (km), j, ciM or ciN is the concentration of component M or N in solution i and aMj or aNj is the k (1%, 1 cm) of component M or N at mean wavelength (km), j. Convolution of the first derivative curves of each drug using OPUI lead to the appearance of many maxima and minima points in the resulting convoluted curves. This allows the simultaneous determination of each component in the mixture using zero-crossing method by the proper choice of km for each drug. The selected km for each drug based on that, maximum response was obtained for the component under investigation with zero contribution of the other component or background interferences. At this selected km, the values of orthogonal polynomial were correlated to concentration as shown in Eq. (6). Experimental Instrumentation A Specord S600 spectrophotometer, associated with WinAspect software version 2.3, Analytik Jena AG, Germany was used. A KERN & SOHN GmbH balance was used and sonicator, J.P. SELECTA, S.A., was used.
283
Materials and reagents MTR and NYS were supplied by El-Amryia and Pharaonia Pharmaceuticals Co., Alexandria, Egypt, respectively. Amrizole-NÒ vaginal suppositories, batch number: 877466, (El-Amryia Pharmaceuticals Co., Alexandria, Egypt) were used. Each vaginal suppository was labeled to contain 500 mg MTR and 100,000 IU = 20.5 mg NYS [16]. Methanol used was of analytical grade. Standard solution and calibration graphs Stock solutions were prepared by dissolving MTR and NYS in methanol to obtain stock solutions of 0.1 mg mL1. The stock solutions of MTR and NYS were further diluted with methanol to obtain working standard solutions of suitable concentrations in the range of (1–25) and (1–15) lg mL1, respectively. Preparation of vaginal suppository solution Five Amrizole-NÒ vaginal suppositories were accurately weighed and finely powdered in a mortar. An amount of the suppository mass equivalent to one suppository content (500 mg of MTR and 100,000 I.U NYS which is equivalent to 20.5 mg [16]) was dissolved in 60 mL of methanol. After 10 min of sonication at room temperature, the solution was quantitatively transferred to a 100-mL volumetric flask. The mixture was completed to volume with methanol. The excipient – the base of suppository – is Poly Ethylene Glycol, so filtration was not needed as it is completely soluble in methanol. Further dilution was conducted with methanol to produce solution containing MTR 25 lg mL1 and NYS 1.025 lg mL1. Spectrophotometric measurements and data treatment Application of first derivative technique D1 to the spectrophotometric data The absorbance was measured in 1-cm cell using methanol as blank and covering the wavelength range 200–400 nm at 0.5 nm intervals. The obtained absorption curve was highly overlapped thus the first derivative was calculated as shown in Fig. 1a and b. Using D1 method, MTR can be determined at 290 nm which shows minimum contribution of NYS. On the other side NYS can be determined at 261 nm which shows minimum contribution of MTR as shown in Fig. 1b. This method was used as a comparable method with the proposed one. Application of OPUI to the D1 data (D1/OPUI) For first derivative absorption curves in methanol of MTR and NYS, the polynomials can be calculated according to the selected wavelength intervals and the mixture can be resolved by zero crossing as follows: For metronidazole (MTR). The segment abcd (Fig. 2a) shows that a K3 polynomial (cubic) can fit the derivative absorption curve. Therefore, the wavelengths 247 (min), 290 (max), 330 (min) and 360 (shoulder) were selected. There wavelengths were spaced by the intervals 0, 43, 40 and 30 nm as shown in Table 1. The polynomials K1, K2 and K3 can be calculated [24,25] as presented in Table 1.
Fig. 1. Absorption spectra of MTR 25 lg mL1 ( ) and NYS 1 lg mL1 ( methanol (a), and their corresponding first derivative curves (b).
) in
For nystatin (NYS). The segment efghij (Fig. 2a) shows that a K0 5 polynomial (quintic) can fit the derivative absorption curve. Therefore, the wavelengths 288 (max), 294 (min), 300 (max), 308 (min), 316 (max) and 323 (min) were selected. There wavelengths were spaced by the intervals 0, 6, 6, 8, 8 and 7 nm Table 1. The polyno-
284
M.A. Korany et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 143 (2015) 281–287
(Fig. 2b and c). The coefficients were calculated according to Eq. (5) as follows:
k3 ¼ ½D1247 ð12801:4Þ þ D1290 ð42880Þ þ D1330 ð51834:8Þ þD1360 ð21756:2Þ=5162748971
ð8Þ
0
k5 ¼ ½D1288 ð14231:06Þ þ D1294 ð62498:79Þ þD1300 ð94869:4Þ þ D1308 ð74867:93Þ þD1316 ð36485:78Þ þ D1323 ð8219:52Þ=20112804470:27 ð9Þ As mentioned in ‘‘Derivative technique followed by orthogonal polynomials for unequal intervals D1/OPUI (the proposed method)’’, the concentration of MTR and NYS were calculated by zero crossing method. As shown in Figs. 2b and 2c, the selected km 301.5 and 241.5 nm were suitable for determination of MTR and NYS, respectively in their mixtures. Each km is selected on the basis that the drug shows maximum response at this k with zero contribution of the other. Results and discussion Optimization The absorption spectra of MTR and NYS in methanol are presented in (Fig. 1a). The drugs show considerable spectral overlap. In vaginal suppository form, the two drugs are formulated as MTR, major component, (500 mg/vaginal suppository) and NYS, minor component, (20.5 mg/vaginal suppository). Accordingly, determination of NYS (minor component) using the conventional absorbance data will lead to erroneous results. In the present work, the assay of the components in a mixture was investigated using the proposed D1/OPUI method and the comparable D1 method. In D1/OPUI method, the first derivative absorption curves for the two analytes (Fig. 2a) were convoluted using the values of the orthogonal polynomials; K3 and K0 5 with their wavelength intervals, selected from Table 1 for MTR and NYS. Two coefficients; k3 and k0 5 were selected at the corresponding km 301.5 and 241.5 nm, respectively, and the concentrations of the drugs were determined by zero crossing at the selected wavelengths (Fig. 2b and c). In D1 method, the concentrations of the drugs were determined, using first derivative absorption curves, at 290 and 261 nm for MTR and NYS, respectively. Method validation The method was validated according to ICH guidelines [26]. The following validation characteristics were addressed: Fig. 2. First derivative absorption spectra of MTR 25 lg mL1 ( ) and NYS 1 lg mL1 ( ) in methanol (a) and its convolution (b) and (c) derived in the form: 0 k3 and k 5, respectively, using selected orthogonal polynomials for unequal intervals.
mials K0 1, K0 2, K0 3, K0 4 and K0 5 can be calculated [24] as presented in Table 1. The polynomials given in Table 1 were calculated, according to the method of Grandage [24], using Excel software. All the generated polynomials fulfill the requirements of Eqs. (2)–(4). Convolution of the D1 curves was done using the values of the orthogonal polynomials; K3 and K0 5 and their corresponding wavelength intervals, selected from Table 1 for MTR and NYS (Fig. 2a). Two coefficients; k3 and k0 5 were selected at the corresponding km 301.5 and 241.5 nm, to solve the MTR and NYS respectively
Linearity Under the above selected parameters of the chemometric method, the linearity of the proposed method was evaluated through the analysis of seven serial concentrations of each drug. The D1 method and the coefficients of the D1/OPUI method were plotted as a function of the corresponding concentration and the calibration equation was calculated using the least squares regression method. Table 2 summarizes the regression and statistical parameters of the studied drugs using the derivative and the proposed chemometric method. As shown in Table 2, comparing the results of the two methods, the regression analysis verifies the good linearity of the proposed chemometric method with slight enhancement in the regression parameters over the derivative method. This slight enhancement is indicated by the increase of correlation coefficient (r) values, the decrease in values of the standard deviation of the intercept (Sa), standard deviation of the
285
M.A. Korany et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 143 (2015) 281–287
Table 1 P P K 2ij M; K 2ij N, for the determination of two component mixtures of metronidazole (MTR) and nystatin Calculation of the polynomials, Kj, and their corresponding sum, (NYS), respectively. MTR
NYS
k Set (I) (nm)
Interval (nm)
247 290 330 360
0 43 40 30
P
K2
K3
k Set (II) (nm)
Interval (nm)
K01
K02
K03
K04
K05
59.75 16.75 23.25 53.25
1313.197 1652.33 1090.95 1430.079
12801.4 42880 51834.8 21756.2
0 6 6 8 8 7
7226.75
7689968
5162748971
288 294 300 308 316 323 P
16.83 10.83 4.83 3.17 11.16 18.16 888.25
159.19 15.35 117.90 142.62 39.34 156.02 85708.07
1107.79 1215.92 1073.91 711.00 1503.42 1032.38 7690543.98
4933.40 12419.73 4836.21 10911.12 12485.20 4224.20 494753680.33
14231.06 62498.79 94869.40 74867.93 36485.78 8219.52 20112804470.27
K1
K 2ij M
K 2ij N
Table 2 Linear regression and statistical parameters for the determination of metronidazole (MTR) and nystatin (NYS) using the derivative method (D1) and the proposed chemometric method (D1/OPUI). ra
ab
|b|c
Sy/xd
6.48 104 3.92 104
2.25 104 6.52 105
1.45 104 4.55 105
1.35 10-5 5.82 10-6
Orthogonal polynomial of unequal interval on derivative data (D1/OPUI method) MTR (k0 3) 301.5 nm 0.9997 9.4 1010 6.21 109 1.46 109 NYS (k0 5) 241.5 nm 0.9998 1.25 108 3.19 108 3.45 109
8.64 1010 2.41 109
6.45 1011 3.08 1010
Derivative technique (D1 method) MTR 290 nm 0.9989 1.48 104 NYS 261 nm 0.9995 3.6 10-5
a b c d e f g h i
S ae
Sbf
Fg
LODh (lg mL1)
LOQi (lg mL1)
2306.89 4547.03
0.416 0.461
1.260 1.396
9264.35 10762.78
0.310 0.073
0.938 0.222
Correlation coefficient. Intercept. Absolute slope. Standard deviation of residuals. Standard deviation of intercept. Standard deviation of slope. Variance ratio, equals the mean of squares due to regression divided by the mean of squares about regression (due to residuals). Limit of detection. Limit of quantitation.
slope (Sb) and standard deviation of the residuals (Sy/x) with high F values [27]. Detection and quantitation limits Limit of detection (LOD) and limit of quantitation (LOQ) are calculated according to ICH guidelines [26] based on the standard deviation of the blank and the Slope as shown in Table 2. As seen from the table, lower values of LOD and LOQ were obtained using the proposed method than the D1 method. Accuracy and precision In order to assess the accuracy, as mean percentage recovery, and the precision, as percentage relative standard deviation (RSD%), MTR and NYS were assayed in different laboratory-prepared mixtures of different ratios including percentage similar to that present in dosage form (25 MTR:1 NYS). The results presented in Table 3 showed that erroneous results for the two analytes were obtained when applying conventional first derivative method. This is attributed to heavy spectral overlapping of MTR during the assay of minor component (NYS). That is why the error relatively increased in mixture number 4, representing the same ratio in the dosage form (Table 3). With the application of orthogonal polynomial for unequal interval (D1/OPUI) method, such error has been corrected (Table 3). The results obtained using the (D1/OPUI) method are not affected by any interferences contributing to coefficients other than those involved in the assay coefficients. Thus in the above assay, the cubic k3 and quintet k0 5 components (here ‘component’ is used in the mathematical sense) are not eliminated, which means in particular that, constant, linear and other components of the irrelevant absorption are all rejected. Nevertheless, in
common, irrelevant absorption contributes much more to the constant and linear components of the total absorption than any others. As shown in Table 3 great enhancement of accuracy and precision when using the proposed (D1/OPUI) method compared with D1 method. The accuracy in Table 3, represented as mean percentage recovery, was enhanced from 107.28 to 100.00 for NYS. The Intra-day precision in Table 3, calculated as RSD%, was improved from 4.25 to 0.67 and from 14.10 to 0.56 for MTR and NYS, respectively. Also, the inter-day precision, calculated as RSD%, was improved from 13.71 to 0.77 and from 13.89 to 0.81 for MTR and NYS, respectively (Table 3). A comparison was done between the derivative method alone and the proposed D1/OPUI method for analyzing each drug in different synthetic mixtures. This was done by applying student’s t-test and variance ratio F-test to check if there is a significant difference between the two methods or not. As can be seen in Table 3, significant difference between the two methods variances was found. This was confirmed as the calculated F-values exceeded the theoretical ones (F-theoretical and calculated values are shown in table for each drug). While no significant difference between the two methods means was noticed. This confirms that the proposed D1/OPUI is valid for analyzing both drugs in different synthetic mixtures ratios. Specificity. Specificity of the method was tested by analyzing each drug in presence of the other (Table 3) and in presence of dosage form excipients (Table 4). Specificity of the proposed D1/OPUI method was confirmed by analyzing MTN and NYS with good accuracy (98–102%) and precision (less than 2%) in presence of each other and the in their dosage form (Tables 3 and 4).
286
M.A. Korany et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 143 (2015) 281–287
Table 3 Evaluation of the accuracy and precision for the determination of two-component mixtures of metronidazole (MTR), nystatin (NYS) in laboratory made mixtures using the proposed D1/OPUI method and D1 method. Mix. No.
Nominal conc.
(a) Accuracy and intraday precision
(a) Accuracy and interday precision
(lg mL1)
MTR
MTR
MTR 1 2 3 4
a b c d e f
NYS
10 10 5 15 15 1 25 1 Mean%a Er%b SDc RSD(%)d t calculatede F calculatedf
NYS
NYS
D1
D1/OPUI
D1
D1/OPUI
D1
D1/OPUI
D1
D1/OPUI
93.31 103.39 97.13 98.64 98.12 1.88 4.17 4.25 0.96 39.13
99.66 99.64 100.22 101.06 100.15 0.15 0.67 0.67
95.20 95.75 111.03 127.12 107.28 7.28 15.13 14.10 0.96 738.55
100.06 100.66 99.98 99.30 100.00 0.00 0.56 0.56
105.40 77.45 104.96 101.15 97.24 2.76 13.33 13.71 0.41 298.35
101.13 99.64 99.42 99.79 99.99 0.01 0.77 0.77
90.11 108.36 111.03 127.12 109.15 9.15 15.16 13.89 1.22 345.66
99.27 101.01 99.98 99.30 99.89 0.11 0.81 0.81
The mean of all recoveries of different concentration in the same method for each drug. Percentage relative error. Standard deviation of the recoveries of different concentration in the same method for each drug. Percentage relative standard deviation. For t-test for unequal variances, the critical t value at 3 degrees of freedom is 3.18 for two tailed test (P = 0.05). For F-test; the critical value F3,3 = 15.44 for two tailed test (P = 0.05).
Table 4 Assay of two-component mixtures, metronidazole (MTR) and nystatin (NYS), in vaginal suppositories using the proposed D1/OPUI method and D1 method. Sample No.
Found (%) MTR
1 2 3 4 5 6 Mean SD RSD (%) t calculateda F calculatedb
NYS
D1 method
D1/OPUI method
D1 method
D1/OPUI method
105.22 94.72 104.60 91.64 105.22 102.13 100.59 5.93 5.89 0.05 17.86
99.77 101.90 101.74 99.30 102.25 99.25 100.70 1.40 1.39
165.92 75.71 53.15 133.19 83.46 108.33 103.29 41.25 39.94 0.17 855.42
100.82 100.07 100.21 98.74 99.55 102.88 100.38 1.41 1.41
a For t-test for unequal variance, there are 6 & 5 degrees of freedom for MTR & NYS respectively, the critical values of t are 2.45 and 2.57 for MTR and NYS respectively at the two tailed test (P = 0.05). b For F-test; the critical value F5,5 = 7.146 for two tailed test (P = 0.05).
Analysis of pharmaceutical preparation In vaginal suppository assay, with regard to accuracy of the major component, MTR, the results did not improve in D1/OPUI compared with D1 method and that may be contributed to its high concentration in dosage form that even interferences from the minor component, NYS, may not be noticed. On the contrary, upon the assessment of precision (calculated as RSD%), high RSD% values were obtained. The RSD% decreased from 5.89 to 1.39 upon applying D1 method to D1/OPUI method, respectively, which indicates the benefit of the proposed chemometric method as shown in Table 4. The improvement of the minor component (NYS) was obvious for the proposed method in both accuracy as mean percentage found (from 103.29% to 100.38%) and precision as RSD% (from 39.94 to 1.41). Table 4 also represents a statistical comparison between the proposed (D1/OPUI) method and the D1 method for the assay of MTR and NYS in their vaginal suppository. This was done using the student’s t- and variance ratio F-tests. The results in Table 4 indicate that there is a significant difference between the two methods as the calculated F-values for each compound exceeded the theoretical ones. As a result, the D1/OPUI is the method of choice for analyzing the drugs in their dosage form (Table 4).
Comparing the results of the proposed D1/OPUI method with the only reported method for the mixture [21], no significant difference between the two methods variances was found. This was confirmed as the calculated F-values (F-calculated values are 1.88 for MTR and 2.90 for NYS) didn’t exceed the theoretical ones (Ftheoretical values for the two tailed test at P = 0.05, F5,4 = 9.36 for MTR and F4,5 = 7.39 for NYS). In contrast, the calculated t-value of NYS exceeded the theoretical one. This indicates significant difference between the two methods in analyzing the minor component (t-calculated values, t = 1.74 for MTR and 2.52 for NYS) (t-theoretical values at df = 9 for the two tailed test at P = 0.05, is 2.26 for MTR & NYS).
Conclusion The proposed chemometric method can be readily applied for simultaneous determination of MTR and NYS in laboratory made mixtures and vaginal suppositories using widely available UV spectrophotometers. The method is specific and there is no interference from any of the matrix components. The proposed method is selective, sensitive. Compared to different chromatographic methods, the proposed method is economic and inexpensive. It
M.A. Korany et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 143 (2015) 281–287
could be used in the routine analysis of this mixture in its pharmaceutical preparation. References [1] M.A. Korany, M. Bedair, F.A. El-Yazbi, Use of orthogonal polynomials for unequal intervals to eliminate interference in spectrophotometric analysis. Simultaneous determination of ephedrine hydrochloride and diphenhydramine hydrochloride in two-component mixtures, Analyst 111 (1) (1986) 41–44. [2] V. Ragupathy, S. Arcot, Simultaneous spectrophotometric determination of diacerein and aceclofenac in tablets by chemometric methods, Int. Res. J. Pharm. 4 (3) (2013) 211–214. [3] P. Bhamre, A. Pathak, S. Rajput, Simultaneous determination of doxylamine succinate, pyridoxine hydrochloride and folic acid by chemometric spectrophotometry, Int. J. Pharma Bio Sci. 4 (1) (2013) 738–749. [4] S. Dinç, Ö.A. Dönmez, B. Asçi, A.E. Bozdogan, Spectrophotometric multicomponent resolution of a tablet formulation containing lisinopril and hydrochlorothiazide by multivariate calibration methods, Asian J. Chem. 25 (2) (2013) 999–1002. [5] Y. Ni, Y. Wang, S. Kokot, Multicomponent kinetic spectrophotometric determination of pefloxacin and norfloxacin in pharmaceutical preparations and human plasma samples with the aid of chemometrics, Spectrochim. Acta Part A Mol. Biomol. Spectrosc. 70 (5) (2008) 1049–1059. [6] M. Bahram, M. Mabhooti, Rank annihilation factor analysis for multicomponent kinetic-spectrophotometric determination using difference spectra, J. Chemom. 23 (5) (2009) 236–247. [7] A. Varghese, L. George, Simultaneous first order derivative spectrophotometric determination of vanadium and zirconium in alloy steels and minerals, Spectrochim. Acta Part A Mol. Biomol. Spectrosc. 95 (2012) 46–52. [8] P.P. Goti, P.B. Patel, Development and validation of ratio-derivative spectrophotometric method for simultaneous estimation of gabapentin, methylcobalamin and alpha lipoic acid in tablet formulation, J. Pharm. Res. 6 (6) (2013) 609–614. [9] P. Giriraj, T. Sivakkumar, Simultaneous estimation of dutasteride and tamsulosin hydrochloride in tablet dosage form by vierordt’s method, Arabian J. Chem. (in press). http://dx.doi.org/10.1016/j.arabjc.2013.07.013 [10] M.A.-H. Elsayed, M.A. Korany, M.M. Bedair, H. Mahgoub, E.A. Korany, Computer assisted spectrophotometric analysis of three component mixtures (caffeine mixtures and codeine mixtures), Anal. Lett. 25 (2) (1992) 239–252. [11] M.A. Korany, M.A.-H. Elsayed, M. Bedair, H. Mahgoub, E.A. Korany, An interactive computer program for multicomponent spectrophotometric analysis using orthogonal polynomials, Anal. Lett. 23 (3) (1990) 507–527. [12] H.M. Maher, R.M. Youssef, Development of validated chromatographic methods for the simultaneous determination of metronidazole and spiramycin in tablets, Chromatographia 69 (3–4) (2009) 345–350.
287
[13] M.M. Issa, R.a.M. Nejem, A.M.A. Shanab, N.T. Shaat, Kinetic spectrophotometric H-point standard addition method for the simultaneous determination of diloxanide furoate and metronidazole in binary mixtures and biological fluids, Spectrochim. Acta Part A Mol. Biomol. Spectrosc. 114 (2013) 592–598. [14] M.A. Mahrouse, E.F. Elkady, Validated spectrophotometric methods for the simultaneous determination of ciprofloxacin hydrochloride and metronidazole in tablets, Chem. Pharm. Bull. 59 (12) (2011) 1485–1493. [15] A.A. Salem, H.A. Mossa, Method validation and determinations of levofloxacin, metronidazole and sulfamethoxazole in an aqueous pharmaceutical, urine and blood plasma samples using quantitative nuclear magnetic resonance spectrometry, Talanta 88 (2012) 104–114. [16] H.M. Heneedak, I. Salama, S. Mostafa, M. El-Sadek, HPLC and chemometric methods for the simultaneous determination of miconazole nitrate and nystatin, J. Chromatogr. Sci. 50 (10) (2012) 855–861. [17] M. Pendela, G. Kahsay, I. Baekelandt, A. Van Schepdael, E. Adams, Simultaneous determination of lidocaine hydrochloride, hydrocortisone and nystatin in a pharmaceutical preparation by RP-LC, J. Pharm. Biomed. Anal. 56 (3) (2011) 641–644. [18] J.M. Lemus Gallego, J. Pérez Arroyo, Micellar electrokinetic capillary chromatography as an alternative method for determination of hydrocortisone and its most important associated compounds in local pharmaceutical preparations, Chromatographia 56 (7–8) (2002) 455–462. [19] E. Scheuch, T. Giessmann, W. Siegmund, Quantitative determination of nystatin in human plasma using LC–MS after inhalative administration in healthy subjects, J. Chromatogr. B 844 (1) (2006) 84–88. [20] J.M. Llabot, D.A. Allemandi, R.H. Manzo, M.R. Longhi, HPLC method for the determination of nystatin in saliva for application in clinical studies, J. Pharm. Biomed. Anal. 45 (3) (2007) 526–530. [21] S.C. Baratieri, J.M. Barbosa, M.P. Freitas, J.A. Martins, Multivariate analysis of nystatin and metronidazole in a semi-solid matrix by means of diffuse reflectance NIR spectroscopy and PLS regression, J. Pharm. Biomed. Anal. 40 (1) (2006) 51–55. [22] Q. Wang, D. Ma, J.P. Higgins, Analytical method selection for drug product dissolution testing, Dissolution Technol. (2006). [23] M.A. Korany, A.M. Wahbi, S. Mandour, M.A. Elsayed, Determination of certain drugs in multicomponent formulations by first derivative ultraviolet spectrophotometry, Anal. Lett. 18 (1) (1985) 21–34. [24] A. Grandage, Orthogonal coefficients for unequal intervals, Biometrics 14 (2) (1958) 287–289. [25] K.C. Peng, The Design and Analysis of Scientific Experiments, Addison-Wesley, London, 1967. [26] ICH, Q2 (R1), Validation of analytical procedures. Text and methodology, International Conference on Harmonization, Geneva, 2005. [27] P. Armitage, G. Berry, Statistical Methods in Medical Research, Blackwell Science Ltd., Oxford, England, 1994. pp. 283–285.
