Application of PLS regression to fluorimetric data for the determination of furosemide and triamterene in pharmaceutical preparations and triamterene in urine

Multivariate calibration methods that use fluorescence data for the simultaneous determination of furosemide and trianaterene were developed. One of the most salient advantages of them is that the vast amount of information provided by the whole spectrum of the sample is not required. This makes analyses simple and fast. The methods require selecting chemometric parameters such as the specific spectral region and number of factors to be used. Both spectral region and number of factors are selected, simultaneously, by minimising the prediction residual error sum of squares (PRESS). The proposed methods were used for the simultaneous determination of the two drugs in real samples (pharmaceutical preparations) with no excipient separation pre-treatment, with furosemide and triamterene contents of 1.68E-3 to 4.31E-2 and 1.03E-3 to 3.12E-2 mug ml(-1), respectively; as well as that of triamterene at concentrations of 5.00E-4 to 5.80E-3 mug ml(-1) in urine samples. The ability to construct the calibration validation sets directly from the urine samples itself avoids the need to consider matrix interferences or to pre-treat the sample and/or separate some analytes The results were quite good in all cases. (C) 2003 Elsevier B.V. All rights reserved.