Просмотр новости
Найдите то, что Вас интересует
Спутники
РН
SpaceX
Роскосмос
NASA
СпутникиРНSpaceXРоскосмосNASABackground. GML-3 is a promising non benzodiazepine anxiolytic with antidepressant activity, but its poor water solubility and flowability complicate the development of an oral solid dosage form. Modern regulatory requirements (ICH Q8–Q10) mandate systematic risk assessment throughout development, making a risk based approach to excipient selection highly relevant.Objective. To develop a GML 3 capsule formulation that provides satisfactory granule rheological properties, acceptable particle size distribution, and active pharmaceutical ingredient (API) release of at least 75 % in aqueous medium, robust to manufacturing variability, using a Taguchi orthogonal array, probabilistic modeling, and multicriteria optimization.Methods. A Taguchi L16(4⁴) orthogonal array was used to screen the filler, binder, disintegrant, and solubilizer, allowing the main effects of the four factors to be estimated in 16 runs, whereas a full factorial experiment (4⁴), which additionally captures all factor interactions, would require 256 combinations. Linear regression models were built for eight critical quality attributes (flowability, Carr and Hausner ratios, angle of repose, d₅₀, span, dust fraction, API release). Monte Carlo simulation (100,000 iterations per combination) was applied to estimate empirical probabilities of exceeding design space boundaries, accounting for parametric uncertainty of model coefficients and residual process variability. Multicriteria optimization used an overall desirability function combining technological properties and risk probability estimates.Results. The insoluble filler (dicalcium phosphate) and the absence of disintegrant consistently led to unacceptable attributes. Several combinations with formally acceptable point predictions showed high probabilities of violating design space limits under process fluctuations (e.g., release failure probability up to 0.43). The optimal formulation was dextrose, polyvinylpyrrolidone 25, sodium starch glycolate, and poloxamer F68, with predicted API release of 95.6% at 45 min and near zero violation probabilities for all controlled attributes.Conclusion. The risk based approach with probabilistic assessment enables early elimination of formulations that are not robust to process variability and allows rational selection of an excipient combination ensuring the required quality of GML 3 capsules with high reliability