A potential dosage form for intranasal administration of diazepam (DZP) in the treatment of epileptic seizure emergencies was investigated.

DZP could be formulated as a supersaturated state in glycofurol/water (GF/water) cosolvent mixtures, and such supersaturated solutions remained sufficiently stable as to prevent the drug from precipitation before crossing the nasal mucosa membrane. Permeation of DZP through polydimethylsiloxane (PDMS) membranes, chosen as a model membrane for nasal mucosa, was significantly enhanced from supersaturated solutions, and this enhancement was not due to interaction of the vehicle with the membrane.

Permeation of DZP from supersaturated formulations was then investigated using Madin-Darby canine kidney (MDCK) epithelial cell monolayers as a nasal mucosa model. MDCK cell viability was assessed under various GF/assay buffer cosolvent mixtures and over different incubation time. Cell monolayer integrity was monitored under various GF and DZP concentrations by measuring transepithelial electrical resistance and permeability of [14C]-mannitol. Results show that cell monolayers exhibited effective barrier function under the probed experimental conditions. The effects of GF and DZP on permeation properties of MDCK cell monolayers were determined before examining permeation of DZP across cell monolayers from supersaturated solutions. Only GF had significant effect on permeation properties of MDCK cell monolayers. Steady state flux of DZP across MDCK monolayers from supersaturated DZP solutions increased proportionally with increasing degree of saturation, as described by Theeuwes’s transference equation, demonstrating the ability to enhance transport using supersaturation. These results support the potential use of supersaturated DZP solutions, formulated at point of administration, for intranasal delivery.

A model based on Flory-Huggins theory was developed to predict the solubility behavior of DZP in GF/water binary solvent mixtures. By comparison with experimental data, the capability of this model to predict DZP solubility was verified. In addition, various cosolvency models to predict solute solubility in mixed solvent systems were presented. The advantages and drawbacks of these models were evaluated.