INTRODUCTION:
In our present investigation three drug combinations containing commercial formulation was selected. This contains ramipril, aspirin and atorvastatin calcium. Ramipril1,2 is chemically [2s, 3aS, 6aS] – 1- [(2S) – 2 – {[(2S) – 1 – ethoxy -1- oxo -4 – phenylbutan – 2 –yl] amino} propanoyl] –octahydrocyclopenta [b] pyrrole – 2- carboxylic acid is an angiotensin II converting enzyme inhibitor. Atorvastatin calcium is chemically (3R, 5R)-7-[2-(4-fluorophenyl)-3-phenyl-4-(phenylcarbamoyl)-5-(propan-2-yl)-1H-pyrrol-1-yl]-3, 5-dihydroxyheptanoic acid calcium salt (2:1) trihydrate. This is a unique hypolipideamic, anti hyper lipoproteinaemic and anti cholesterol agent, acts mainly by inhibiting HMG – Co A reductase enzyme, which is required for the synthesis of mevalonate. Aspirin is 2- acetoxy benzoic acid, a cyclo oxygenase inhibitor.
It is an analgesic, antipyretic, anti-inflammatory and antithrombotic agent. These three drugs are given in combination for the treatment of lipeaemic patients with hypertensions, to reduce the plaques by decreasing the cholesterol synthesis, probably by increasing the blood flow through arteries and also reducing the aggregation of platelets. Commercial formulation containing these three drugs named as “Polytorva” was manufactured by USV (Ltd) – Mumbai. Each capsule contains 2.5, 75 and 10 mg of ramipril, aspirin and atorvastatin calcium respectively, along with other excipients. Few methods have been reported for determination of these drugs either alone or in combination with other drugs3 - 11
MULTIVARIATE CALIBRATION:
Multivariate calibration techniques are widely used in many practical applications, including pharmaceutical industries. Whereas the traditional univariate calibration relies on a single scalar measurement for each sample, multivariate calibration utilize a vector (or matrix) of data for each sample, such as a spectrum, which permits analysis in the presence of multiple interferences. This study focuses on evaluation of one of the multivariate calibration methods for the problem of interest that is CLS method. (In this paper, scalars will be indicated by lower case italic symbols, vectors by boldface lower case symbols, and matrices with boldface upper case symbols. The super script ‘ and -1 will be used to indicate the transpose of a matrix or a vector and inverse of the square matrix respectively).
In most cases, the calibration procedure is as follows, first, a set of samples is obtained that is representative of the range of compositions that are likely to be encountered in practice. Spectra are then recorded for each sample. The samples are then divided into two groups, one that is used to build the calibration model ( the training or calibration set) and another that is used to independently assess the performance of the model (the validation or prediction set).
CLASSICAL LEAST SQUARES REGRESSION 12 - 14
Classical least squares (CLS), also known as the K matrix method, assume that the observed responses are linearly related to the component concentrations and that the concentrations of all components in the mixture are known for the calibration set. The most common example of this is the multivariate form of Beer’s law, which relates a matrix of spectra, A, to a matrix of concentrations, C, through an equation:
matrix that has rows equal to the number of samples (m) and columns equal to the number of wavelengths (n) for which samples are measured, C is an m x p matrix of component concentrations (p is the number of components in the mixture, in our case p is 3), and K is the p x n matrix of pure component spectra coefficients. E is the error matrix of dimensions equal to A (m x n), that are not fit by the model K is normally calculated from the calibration data by solving the system of linear equations applying least squares solution.
then for prediction of unknown samples ,
Statistical evaluation of the developed method includes the usual mean % found, slope, intercept, standard deviation and % RSD and also the RMSEC and RMSEP 15 parameters, which is calculated as follows
m = number of samples (mixtures); y = actual concentration of the calibration or prediction sets; ŷ = concentration found by applying any multivariate method in either calibration or prediction sets.
EXPERIMENTAL:
INSTRUMENTATION, SOFTWARE:
A double beam Shimadzu (Japan) UV – visible spectrophotometer model UV- 2550PC
connected to PC. The bundled software used to collect the signals was UV winlab ver. 2.0 software (shimadzu). The spectral bandwidth was 0.5 nm and the readability was 0.1 nm. The absorption spectra of test and reference solutions were recorded in 1 cm matched quartz cells over the range of 210 -340 nm. CLS calculations were performed using PLS Toolbox 5.0 (Demo version) for use with MATLAB 7.0.
MATERIALS AND REAGENTS:
Pharmaceutical grades of ramipril, aspirin and atorvastatin calcium were kindly gifted by Lyka pharmaceuticals Mumbai. Methanol used was of analytical grade. The commercial formulation ‘Polytorva capsule’ was procured from the local market.
STANDARD SOLUTIONS:
To produce standard solutions of the drugs under investigation, 100 mg of each were weighed accurately, transferred to 100 ml volumetric flask, dissolved in methanol and made up to the volume with the same solvent to get 1 mg mL-1. From this further dilutions were made appropriately with methanol, to produce working standards for calibration and prediction set.
CALIBRATION AND PREDICTION SETS:
A five and four level factorial design was used to produce the calibration and prediction set, respectively. Five levels considered for the calibration set was as follows 1.0 - 5.0, 10.0 – 50.0 and 2.0 – 10.0 µg mL-1 and four level design for the prediction set was 1.2 – 4.8, 12.0 – 48.0 and 2.5 – 10.0 µg mL-1 of ramipril, aspirin and atorvastatin calcium respectively. UV spectra of the prepared mixtures were recorded in the wavelength range of 210 .0 – 340.0 nm verses a solvent blank, and digitized absorbance was sampled at 0.1 nm intervals. The data were then exported to MS office EXCEL ®. The computations were made in PLS Toolbox 5.0 (Demo version) for use with MATLAB 7.0. The multivariate calibration models were applied to both the validation and prediction set mixtures to calculate the concentrations of each component.
COMMERCIAL SAMPLE PREPARATION:
Twenty capsule contents were emptied, accurately weighed and mixed well in a mortar.
An amount of the powder equivalent to one capsule content was extracted with methanol. After 30’ of shaking, the solution was filtered into a 100.0 ml volumetric flask. The residue was washed thrice with methanol and the volume was made up to 100.0 ml with the same solvent. From this further dilutions were made as required to produce final concentration of 2.0, 30.0 and 4.0 µgmL-1 of ramipril, aspirin and atorvastatin calcium respectively. UV spectra were recorded as for the calibration and prediction set.
RECOVERY STUDIES:
Accuracy of the method was studied by performing recovery studies at three levels. This study was done by adding known concentrations of standard raw materials to the pre-analyzed formulation and reanalyzed. Results obtained for the recovery solutions were statistically evaluated.
Figure i: Overlain spectrum of ramipril, aspirin and atorvastatin calcium in
Methanol. _____ Ramipril, ……… Aspirin, - - - - Atorvastatin calcium
RESULTS AND DISCUSSION:
Full spectrum methods are usually significantly more precise than methods restricted to a small number of wavelengths. A convenient method for resolving mixtures, which can be applied in this work, is least squares analysis. The simplest such method is classical least squares analysis (CLS). This method is preferred when selection of the variables is simple. The overlain absorption spectra for the drugs under study were shown in fig 1. It is apparent there is substantial spectral overlap is seen. Multivariate CLS method was applied to the UV absorption spectra. For application of CLS method, a training set (calibration set) of 25 synthetic mixtures with different concentrations in the range mentioned in Table 1 was prepared in the laboratory. The root mean squared error of calibration as well as prediction has been applied for the selection of wavelengths. The Satisfactory recoveries with small standard deviations were obtained only for the selected wavelength region for that particular drug component. Wavelengths ranging between 207 – 312 nm for ramipril and aspirin and 207 – 320 nm for atorvastatin calcium (Table 3), indicating an adequate model for CLS method applied on UV absorption spectra. To test the validity and applicability of the CLS chemometric method, the developed model was challenged with the spectra of a prediction set, made of sixteen samples different than those of the calibration set, the composition of which is shown in table 2, was prepared and the recorded spectral values were used to calculate the concentration. Table 2 summarizes the results obtained. The mean recovery percent of component concentrations in the prediction set was 99.78, 99.55 and 99.64 for RA, AS and AT respectively, (Table 3). The predicted concentrations of RA, AS and AT in each sample of the prediction set were compared with their known concentrations, and the root mean squared error of prediction (RMSEP) was calculated. The RMSEP was used as a diagnostic test for examining the errors in predicted concentrations. It indicates both precision and accuracy of predictions, is shown in Table 3. Also a linear relationship with a slope approaches one when plotting the predicted concentration against the true ones indicates the precision of the method (Table 3). CLS model built for RA, AS and AT were now applied to the commercial tablets containing these three drugs in combination. Satisfactory results were obtained for all the three drug components in good agreement with the label claim, 99.33 ± 1.17, 98.23 ± 2.28 and 99.83 ± 3.23 were obtained for RA, AS and AT respectively are shown in table 4. To assess the validity of the proposed CLS method, a standard addition technique was applied for all the three drugs at three levels of concentration to the previously analyzed tablets dosage form. Recovery of each drug was calculated by comparing the spiked concentrations and amount recovered. The results of the analysis of the recovery data were shown in table 5. From the results, (% RSD values for the three drugs) it was clearly understood that there were no interferences from the tablet excipients.
Table :1 Composition and results of calibration set in µg mL-1
|
S. No: |
mixture composition µg mL-1 |
results in µg mL-1 |
||||
|
RA |
AS |
AT |
RA |
AS |
AT |
|
|
1 |
3 |
30 |
6 |
2.77 |
29.45 |
5.86 |
|
2 |
3 |
10 |
2 |
3.00 |
9.53 |
2.03 |
|
3 |
1 |
10 |
10 |
1.05 |
9.65 |
9.82 |
|
4 |
1 |
50 |
4 |
0.84 |
48.92 |
3.84 |
|
5 |
5 |
20 |
10 |
4.77 |
20.02 |
10.07 |
|
6 |
2 |
50 |
6 |
2.18 |
50.75 |
6.30 |
|
7 |
5 |
30 |
4 |
5.22 |
30.62 |
3.98 |
|
8 |
3 |
20 |
4 |
3.06 |
19.86 |
4.04 |
|
9 |
2 |
20 |
8 |
2.10 |
20.01 |
7.88 |
|
10 |
2 |
40 |
10 |
2.03 |
40.15 |
9.77 |
|
11 |
4 |
50 |
8 |
4.06 |
51.18 |
7.75 |
|
12 |
5 |
40 |
6 |
5.01 |
39.89 |
6.03 |
|
13 |
4 |
30 |
10 |
3.99 |
30.28 |
9.94 |
|
14 |
3 |
50 |
10 |
2.94 |
49.96 |
10.65 |
|
15 |
5 |
50 |
2 |
4.87 |
49.59 |
1.93 |
|
16 |
5 |
10 |
8 |
5.13 |
10.15 |
8.05 |
|
17 |
1 |
40 |
2 |
1.14 |
40.33 |
1.95 |
|
18 |
4 |
10 |
6 |
4.10 |
9.87 |
6.08 |
|
19 |
1 |
30 |
8 |
0.98 |
29.59 |
7.95 |
|
20 |
3 |
40 |
8 |
3.11 |
40.65 |
7.72 |
|
21 |
4 |
40 |
4 |
3.91 |
39.51 |
3.99 |
|
22 |
4 |
20 |
2 |
3.95 |
19.72 |
2.01 |
|
23 |
2 |
10 |
4 |
1.84 |
9.37 |
3.96 |
|
24 |
1 |
20 |
6 |
0.89 |
19.29 |
6.34 |
|
25 |
2 |
30 |
2 |
2.06 |
29.90 |
2.01 |
RA : Ramipril AS : Aspirin AT : Atorvastatin calcium
CONCLUSION:
Simple, rapid, accurate, and sensitive measurements with CLS multivariate calibration analysis have been developed for analysis of the content of ramipril, aspirin and atorvastatin calcium in several synthetic mixtures and commercial tablets. The good recoveries obtained from the above procedures proved that the proposed method could be applied efficiently for determination of ternary mixtures containing ramipril, aspirin and atorvastatin calcium. Precision was satisfactory and the method could be easily used in a quality-control laboratory for analysis of these drugs.
ACKNOWLEDGEMENT:
One of the authors, A.S.K. Sankar, gratefully acknowledges the support from the management of Adhiparasakthi Charitable, Medical, Educational and Cultural Trust, Melmaruvathur, for providing necessary facilities to carry out this research work.