Artificial Neural Networks to Optimize Formulation Components of a Fixed- Dose Combination of Rifampicin, Isoniazid and Pyrazinamide in a Microemulsion

The aim of this study to design a stable microemulsion formulation to deliver a combination of rifampicin, isoniazid and pyrazinamide in quantities suitable for administration to a paediatric population. The chemical stability of rifampicin, isoniazid and pyrazinamide alone and in various combinations was investigated in different solvents, solubilizing agents and surfactants. An artificial neural network was used to model data from the stability studies and a sensitivity analysis was applied to optimize the selection of the formulation components. Imwitor 308 and Crillet 3, exhibiting the highest overall positive sensitivity were selected to formulate the stable microemulsion. Due to drug dose specifications and solubility limitations, the final formulation contained only rifampicin and isoniazid, since the solubility of pyrazinamide in the lipid and aqueous components of the microemulsion did not achieve the required dose. The stability and solubility of rifampicin were improved in the formulation. Solubilization of the rifampicin in the lipid droplets of the internal phase and lipophilic chains of the surfactants increased the quantity of rifampicin that can be incorporated, while protecting it from oxidative degradation and also limited its contact with isoniazid, which has been shown to affect its stability. The results of this study indicate that the Artificial Neural Network can be successfully used to optimize the choice of solvents, solubilizing agents and surfactants prior to formulation of the microemulsion, limiting the amount of experiments, thus reducing the costs during the preformulation study.