Chemometric assisted solid-phase microextraction for the determination of anti-inflammatory and antiepileptic drugs in river water by liquid chromatography–diode array detection

In the present work, an analytical method for the simultaneous determination of seven non steroidal anti-inflammatory drugs (naproxen, ketoprofen, diclofenac, piroxicam, indomethacin, sulindac and diflunisal) and the anticonvulsant carbamazepine is reported. The method involves preconcentration and clean-up by solid-phase microextraction using polydimethylsiloxane/divinylbenzene fibers, followed by liquid chromatography with diode array detection analysis. Parameters that affect the efficiency of the solid-phase microextraction step such as soaking solvent, soaking period, desorption period, stirring rate, extraction time, sample pH, ionic strength, organic solvent and temperature were investigated using a Plackett–Burman screening design. Then, the factors presenting significant positive effects on the analytical response (soaking period, stirring rate, stirring time) were considered in a further central composite design to optimize the operational conditions for the solid phase microextraction procedure. Additionally, multiple response simultaneous optimization by using the desirability function was used to find the optimum experimental conditions for the on-line solid-phase microextraction of analytes in river water samples coupled to liquid chromatography and diode array detection. The best results were obtained using a soaking period of 5 min, stirring rate of 1400 rpm and stirring time of 44 min. The use of solid-phase microextraction technique avoided matrix effect and allowed to quantify the analytes in river water samples by using Milli-Q based calibration graphs. Recoveries ranging from 71.6% to 122.8% for all pharmaceuticals proved the accuracy of the proposed method in river water samples. Method detection limits were in the range of 0.5–3.0 μg L−1 and limits of quantitation (LOQs) were between 1.0 and 4.0 μg L−1 for pharmaceutical compounds in river water samples. The expanded uncertainty associated to the measurement of the concentration ranged between 8.5% and 29.0% for 20 μg L−1 of each analyte and between 9.0% and 29.5% for the average of different concentration levels. The main source of uncertainty was the calibration step in both cases.