No abstract

Calibration data of LC-MS/MS rarely fit the pure least square regression model, especially for large concentration intervals. The response function of the MS instrument is corrected by weighted regression models or logarithms. The choice of a response linearization method is based on results produced through back-interpolation of experimental data and/or evaluation of correlation coefficients. Two bioequivalence studies carried out for pharmaceutical formulations containing metformin gave us the opportunity to appreciate the impact of the MS response linearization method (logarithm and 1/x weighted linear regression) on method quality characteristics. The sample preparation was based on protein precipitation with acetonitrile. Chromatographic separation was achieved on a Zorbax CN column (mobile phase acetonitrile and aqueous 10 m m ammonium acetate solution, pH 3.5). Tandem MS detection was performed on a triple quadrupole spectrometer equipped with an electrospray source, operated under positive-ion mode. The method was validated and used for evaluation of the bioequivalence of formulations containing 500 and 1000 mg metformin. The 500 mg metformin study used logarithms for linearization of the detector response, while the 1000 mg metformin study was based on 1/x linear weighted regression. Data resulting from validations and studies completion were compared with evaluate the impact of the response linearization on the method quality characteristics.