Development and validation of UHPLC-PDA method for simultaneous determination of bioactive polyphenols of horse-chestnut bark using numerical optimization with MS Excel Solver.

A fast and efficient UHPLC-PDA method was developed for quantification of polyphenols of horse-chestnut (Aesculus hippocastanum) bark, an herbal medicinal agent used for the treatment of vascular ailments. To optimize a simple one step-gradient separation, four main factors were selected (initial acetonitrile concentration, gradient slope, flow rate and temperature). Retention times and widths of ten target peaks were modelled using central composite design, and the optimal compromise between the time of the analysis and resolution was found using Derringer's desirability function and numerical optimization procedure with the use of MS Excel software. The optimal separation conditions were as follows: initial acetonitrile concentration of 9.9 %, gradient slope of 2.4 %/min, flow rate of 0.7 mL/min and temperature of 35 °C. Chromatographic separation was carried out on a Titan C18 column (1.9 μm, 100 × 2.1 mm i.d.; Supelco) and was completed within 5 min with minimal resolution between the observed critical pairs of 1.93. The experimental results were in good accordance with the ones calculated from the models. Stability of the analytes in standard solutions and processed plant samples was found acceptable up to 14 days of storage at 4 °C. The validation proved also good linearity (r > 0.99995), precision (RSD < 4 %), accuracy (96.8-99.2 %), and sensitivity (LOQs 0.14-0.61 ng; LODs 0.05-0.20 ng) of the method, and its applicability for quantification of six primary coumarins and flavan-3-ols in A. hippocastanum bark was demonstrated for real commercial samples (different providers and years of collection). Additionally, the good separation of the plant matrix allowed for determination of the approximate contents of three minor constituents. The developed procedure proved to be a useful tool in quality control of the herbal material, and MS Excel software was demonstrated to be valuable for optimization of Derringer's function and visualization of the modelled separation.