Chemometric Assisted Spectrophotometric Methods for Simultaneous Determination of Paracetamol and Tolperisone Hydrochloride in Pharmaceutical Dosage Form

Three simple Thermometric assisted UVVisible Spectrophotometric methods, Classical Least Square (CLS), Partial Least Square (PLS) and Principal Component Regression (PCR) were developed for simultaneous estimation of PCM and TOL in pharmaceutical dosage form without any chemical separation and any graphical treatment of the overlapping spectra of two drugs. The UV absorption spectra of the drugs studied in the range of 220 – 280 nm. Beer’s law was obeyed for both drugs in the concentration ranges of 5 – 25 μg/ml for Paracetamol and 1.5 – 7.5 μg/ml for Tolperisone Hydrochloride. Twenty five (25) mixed solutions were prepared for the chemo metric calibration as training set and sixteen mixed solutions were prepared as validation set. The absorbency data matrix was obtained by measuring the absorbency at twenty-one wavelength points, from 220 to 280nm with the interval of 3nm. The developed methods can be applied in simultaneous determination of the selected drugs from the pharmaceutical formulation in routine analysis.

Literature review revels that numbers of analytical methods like RP-HPLC [7][8][9] and UV Spectrophotometry [10][11][12][13][14] are available for determination of PCM and TOL in combination and individual dosage forms and combination with other drugs .The reported UV spectrophotometric methods are based on multicomponent analytical methods viz.simultaneous equation, absorbance ratio and first derivative methods.These methods are based on the univariate or bivariate calibration methods.
In the present work, the chemometric (multivariate calibration methods) approach was applied for the multicomponent analysis of drug substances with a spectrophotometric method.Multivariate calibration is a chemometrics method which has been employed for determination of drugs in combined dosage [15][16][17][18][19], this study aims to introduce an alternative analytical procedure based on the chemometric-assisted spectrophotometric methods for the analysis of PCM and TOL in tablet dosage form and three multivariate calibration methods like principle component regression (PCR), partial least square regression (PLS) and classical least square (CLS) were developed and validated for the determination of PCM and TOL combination in tablet dosage form.

Materials and reagents
Reference standard of PCM and TOL were procured as gratis sample from Alembic Pharmaceuticals Ltd. (Baroda, Gujarat, India) and Zydus Cadila Healthcare Ltd. (Ahmedabad, Gujarat) respectively.Distilled water was prepared using Millipore SmartPak ® DQ5 Purification Pack system.

Instrumentation and software
Shimadzu AUX 220, electronic balance was used for weighing the samples.The Shimadzu UV-1800, UV-Visible double beam spectrophotometer with a matching pair of 1 cm quartz cuvettes (Shimadzu Corporation, Kyoto, Japan), connected to a computer loaded with Shimadzu UVPC version 3.42 software was used to record UV spectra of solutions.The spectral band width was 0.5 nm.Unscrambler ® 10.3.0.80,MATLAB (R2009A) 7.8.0.347 and Microsoft excel were used for PCR, PLS and CLS model development and data analysis.

Preparation of standard stock solution
Accurately weighed and transferred 10 mg of PCM and TOL working standard into two different 100 mL volumetric flask respectively, and volume was made up to 100 mL with distilled water.The final concentration of PCM and TOL were 100 µg/mL of each.

Preparation of working stock solution
Standard stock solution of PCM and TOL were used as a working solution.

Construction of Calibration set
The samples for calibration set were prepared in distilled water by using mutually orthogonal design.Total 25 mixtures were prepared for calibration set which are shown in Table 1.Absorbance of UV spectrums was recorded in the wavelength range 220-280 nm at 3 nm wavelength interval.

Construction of Validation set
The samples for validation set were prepared in distilled water.Total 16 mixtures were prepared for validation set which are shown in Table 2. Absorbance of UV spectrums was recorded in the wavelength range 220-280 nm at 3 nm wavelength interval.

Calibration matrix and selection of spectral zones for analysis by CLS, PCR and PLS
Figure 2 has shown the overlain zero-order spectra for PCM and TOL individually and in mixed in distilled water.As shown in figure PCM exhibit absorption maxima at 243 nm and TOL exhibit absorption maxima at 261 nm.Two Chemometric calibrations, using the zeroorder spectra, were separately applied to simultaneous determination of these drugs in mixtures.The absorbance matrix was obtained by measuring the zero-order absorbance in the wavelength range between 220 and 280 nm, as shown.The quality of multicomponent analysis is dependent on the wavelength range and the spectral mode used.Original and reconstructed spectra of the calibration matrix were compared in order to select the range of wavelengths along with using the root mean squared error of cross validation (RMSECV) and root mean squared error of prediction (RMSEP) values.The wavelength range 220 -280 nm with 3 nm intervals was selected, since this range was providing the greatest amount of information about the two components.

Statistical analysis in selecting the number of principal components or factors
The ability of a calibration can be define in several ways.In this sub-section, we were calculated the standard variation of chemometric calibrations in the case of investigated mixtures.An appropriate choice of the number of principal components or factors is necessary for PCR and PLS calibrations.A cross-validation method leaving out one sample at a time was employed.With a calibration set of 25 calibration spectra, PLS, CLS and PCR calibrations on 24 calibration spectra were performed, and using this calibration the concentration of the sample left out during the calibration process was determined.This process was repeated 25 times until each calibration sample had been left once.The predicted concentrations were compared with the known concentrations of the compounds in each calibration sample.
To validate the model, both RMSECV and RMSEP were considered; they must be as low as possible for a particular model.RMSECV and RMSEP were calculated for each method as per equation 1 and 2.  The predicted concentrations of the components in each sample were compared with the actual concentrations of the components in each validation samples and the root mean square error of cross validation (RMSECV) was calculated for each method.The RMSECV was used for examining the error in the predicted concentrations.The model is a key to achieving correct quantitation in PLS, CLS and PCR calibrations.The resulted models were also validated by prediction of the concentration of analytes in separate validation set which was not used in the model development.The results of prediction and the % recoveries are represented in Table 3 to 5. The predictive abilities of the models were evaluated by plotting the actual known concentrations against the predicted concentrations which are shown in Figure 3. Figure has shown there was good agreement between the predicted (calculated) and actual concentration of drugs.The means recoveries and the relative standard deviation of our proposed methods were computed and indicated in Table 3 to 5 for PCM and TOL, respectively.Another diagnostic test was carried out by plotting the concentration residuals against the predicted concentrations.Figure 4 has shown the residuals appear randomly distributed around zero, indicating adequate models building.Satisfactory correlation coefficient (r 2 ) and slope values were obtained for each compound in the validation set by PLS, CLS and PCR optimized models indicating good predictive abilities of the models.

Assay of Marketed formulation
Twenty tablets were accurately weighed and finely powdered.Tablets powder equivalent to about 500 mg of PCM and 150 mg of TOL accurately weighed and transferred to into 100 mL amber colored volumetric flask and 70 mL of distilled water was added.The mixture was sonicated for 20 min and diluted up to the mark with distilled water and filtered through a whatman filter paper no.41.From this solution 1 mL aliquot was withdrawn into a 100 mL amber colored volumetric flask and diluted up to the mark with water to get the solution containing 50 μg/mL of PCM and 15 μg/mL of TOL.From this solution 3 mL aliquot was withdrawn into a 10 mL amber colored volumetric flask and diluted up to mark with water.So solution contains 15 μg/mL of PCM and 4.5 μg/mL of TOL.The analysis procedure was repeated six times for tablet formulation.The result was shown in Table 6.

Accuracy Study
The accuracy of the method was carried out at three levels 80 %, 100 % and 120 % of the working concentration of the sample.The calculated amount of a standard solution of PCM and TOL was spiked with added sample solution to prepare level 80 %, 100 % and 120 % of the working concentration.The analysis procedure was repeated for three times.The result was shown in Table 7 and 8.

SUMMARY AND CONCLUSION
Three chemometric methods were applied successfully to simultaneous determination of PCM and TOL in pharmaceutical dosage form.The summary parameters of all three chemometric methods were shown in Table 9. Model that gave lowest RMSECV values when used for predicting the unknown samples, predicted well by giving lowest RMSEP values.On the other hand, the fundamental advantages of investigated methods are the simultaneous analysis of the mixture of the subject drugs, without chemical pre-treatment, speed of analysis

Figure 1 .
Figure 1.(a) Chemical Structure of PCM, (b) Chemical Structure of TOL

Figure 2 .
Figure 2. Overlay Spectra of PCM, TOL and Mixture

Table 2 .
Composition of validation set data

Table 3 .
Recovery studies of PCM and TOL by PCR method Yact= Actual concentration of the prediction set samples Ypred= Predicted concentration of the prediction set samples Ip= Total number of prediction set samples.

Table 4 .
Recovery studies of PCM and TOL by PLS method

Table 5 .
Recovery studies of PCM and TOL by CLS method

Table 6 .
Assay results of PCM and TOL by developed PCR, PLS and CLS methods (n=6)

Table 7 .
Accuracy data of PCM by PCR, PLS and CLS methods

Table 8 .
Accuracy data of TOL by PCR, PLS and CLS methods

Table 9 .
Summary parameters of chemometric methods