Estimation of antitubercular drugs combination in pharmaceutical formulations using multivariate calibration

The use of multivariate calibration methods, principal component regression (PCR) and partial least square (PLS) regression, applied on the raw and first derivative spectra, for the estimation of pyrazinamide (PYZ), isoniazid (INH) and rifampicin (RIF) in antitubercular combination formulation is presented. First derivative spectra were computed using Savitzky-Golay transformation. The three drugs show considerable degree of spectral overlap. The calibration set exhibiting central composite design was used to develop models able to predict the concentration of unknown samples containing three drugs. The calibration model was optimized by an appropriate selection of number of factors and wavelength regions to be used for building data matrix, based on minimization of predicted error sum of square (PRESS) using leave-one-out cross-validation strategy. The calibration model was validated by a separate set of 15 mixture solutions. The PCR and PLS models on the raw and first derivative spectra were statistically compared by ANOVA. All methodologies gave statistically similar results. The methodologies were successfully applied to the simultaneous estimation of PYZ, INH and RIF in pharmaceutical formulation.