Artificial Neural Networks and Concentration Residual Augmented Classical Least Squares for the Simultaneous Determination of Diphenhydramine, Benzonatate, Guaifenesin and Phenylephrine in their Quaternary Mixture

Purpose: To develop two multivariate calibration methods for the simultaneous spectrophotometric determination of a quaternary mixture composed of diphenhydramine HCl, benzonatate, guaifenesin and phenylephrine HCl in Bronchofree ™ capsules in the ratio of 2.5 : 10 : 10 : 1, respectively.
Methods: Novel artificial neural networks (ANNs) and concentration residual augmented classical least squares (CRACLS) methods were developed for the quantitative determination of the quaternary mixture. For proper analysis, a four-level, four-factor experimental design was established resulting in a training set of 16 mixtures containing different ratios of the four analytes. A validation set consisting of six mixtures was used to validate the prediction ability of the suggested models.
Results: ANNs and CRACLS methods were successfully applied for the analysis of raw materials and capsules. For ANNs method, % recovery of diphenhydramine HCl, benzonatate, guaifenesin and phenylephrine HCl in the capsules was 102.21 ± 1.34, 100.30 ± 1.17, 99.31 ± 2.00 and 98.50 ± 1.27, respectively. On the other hand, % recovery of the four analytes by CRACLS was 99.84 ± 2.22, 100.07 ± 0.63, 98.37 ± 1.42 and 97.99 ± 0.96, respectively.
Conclusion: The proposed methods can be applied for the quantitative determination of the four components without interference from excipients, thus obviating the need for preliminary extraction of analytes from the pharmaceutical formulation. The ability of the methods to deconvolute the highly overlapped UV spectra of the four components’ mixtures using low-cost and easy-to-handle instruments such as UV spectrophotometer is also an advantage.

Keywords
Artificial neural networks; Concentration residual augmented classical least squares; Quaternary mixture; Simultaneous determination