| Summary: | In this study we used sparse data to predict optimum pH for cefotaxime sodium aqueous solutions stability using design of experiment (DoE) and artificial neural network (ANN). Cefotaxime aqueous solutions of pH 1, 4, 7 were prepared, separately in final drug concentration of 250 μg/mL and incubated at 37 °C under light protection. Samples of 1 mL were collected from 0.5 to 8.0 h and analyzed with validated HPLC method. Percent drug remaining in samples was analyzed by Design-Expert®, quadratic model and on INFORM®, software designed for ANN. DoE predicted pH was 5.49 at which drug remained above 98% for 3.9 h. ANN algorithm predicted pH was 4.7 for maximum stability up to 2.2 h. Predicted pH values from sparse data were within reported pH range of 4.3-6.5 for maximum drug stability. Finally, in validation experiments, DoE and ANN approaches successfully predicted pH 5.5, with sparse data at which drug remained stable for 4 h. © 2018, Colegio de Farmaceuticos de la Provincia de Buenos Aires. All rights reserved.
|
|---|