pH-spectrophotometric titration data were used to determine protonation constants of vardenafil at different ionic strengths I and temperatures of 25°C and 37°C. The use of two different multiwavelength and the multivariate treatment of spectral data, SPECFIT32 and SQUAD(84) nonlinear regression analyses and INDICES factor analysis is presented. The reliability of the protonation constants of the drug was proven with goodness-of-fit tests of the pH-spectra. The thermodynamic protonation constants log K Ti were estimated by a nonlinear regression of (log K, I) data using the Debye-Hückel equation, yielding log K 4T = 3.59(1) and 3.26(1), log K 3T = 5.64(1) and 5.81(1), log K 2T = 9.41(1) and 8.59(2), log K 1T = 10.92(2) and 10.05(1) at 25°C and 37°C, where the figure in brackets is the standard deviation in last significant digit. Concurrently, the experimental determination of four

...

pH-spectrophotometric titration data were used to determine protonation constants of vardenafil at different ionic strengths I and temperatures of 25°C and 37°C. The use of two different multiwavelength and the multivariate treatment of spectral data, SPECFIT32 and SQUAD(84) nonlinear regression analyses and INDICES factor analysis is presented. The reliability of the protonation constants of the drug was proven with goodness-of-fit tests of the pH-spectra. The thermodynamic protonation constants log K Ti were estimated by a nonlinear regression of (log K, I) data using the Debye-Hückel equation, yielding log K 4T = 3.59(1) and 3.26(1), log K 3T = 5.64(1) and 5.81(1), log K 2T = 9.41(1) and 8.59(2), log K 1T = 10.92(2) and 10.05(1) at 25°C and 37°C, where the figure in brackets is the standard deviation in last significant digit. Concurrently, the experimental determination of four thermodynamic protonation constants was combined with the computational prediction of the MARVIN program based on knowledge of the chemical structures of the drug and was in good agreement with its experimental value. The factor analysis of spectra in the INDICES program predicts the correct number of light-absorbing components when the instrument error is known and when the signal-to-error ratio SER is higher than 10. <alternatives> [...] </alternatives>