Category: Formulation and Quality
Purpose: This study was conducted to optimize an abuse-deterrent composition of three different crosslinked anionic polymers (carboxymethyl starch (CMS), carboxymethyl cellulose (CMC), and polyacrylic acid (PAA)) using the Minitab Experimental Design and Analysis. The three polymers bind to cationic opioid drugs in solutions, minimizing the amount of the free drug available for extraction and subsequent injection. The optimized composition is expected to minimize the opioid drug (dextromethorphan HBr, DEX) extraction in different aqueous solutions intended for intravenous drug abuse, while still providing an intended release profile in the simulated gastric medium (0.1N HCl) when legitimately used.
Methods: Minitab® 18 Software was used to design and randomize the number of experiments, differing in the ratios of each deterrent polymer (CMS, CMC, PAA) in the mixture. The selected experimental design was the Mixture Design. Under the Mixture Design, the Simplex Lattice Design was chosen with a predefined lattice degree of 3. The design was further augmented by center and axial points. The total amount of the components was restricted to 200 mg, with lower and higher limits of 0 mg and 200 mg respectively, for each polymer.
The designed experiments (13 trials) were run in the lab to determine the actual % binding in different solvents (water, 40% v/v ethanol, normal saline, 0.83 M acetic acid, and 0.1N HCl). Based on the binding results, the software was further used to optimize the proportion of the deterrent polymers. The Response Optimizer of the Mixture Design was chosen. The setup values in water, ethanol, normal saline, and acetic acid were 50% and 100% for the lower and target binding respectively, where the goal was to maximize the binding. The setup values in 0.1N HCl were 10% and 0% for the upper and target binding respectively, where the goal was to minimize the binding in this solvent. The optimized composition was experimented in the lab to ensure the accuracy of the Minitab prediction.
The binding studies were conducted by mixing the proposed compositions with 25 mg DEX and 7 mg alkalizing agent in 10 mL solvent (water, 40% ethanol, normal saline, 0.83 M acetic acid, and 0.1N HCl). Followed by vortex mixing (30 sec.) and centrifugation (2500 rpm, 5 min.), the supernatant was filtered through a 0.2 μm syringe filter and diluted using the solvents. The concentration of the drug in each solution was measured by a UV-Vis spectrophotometer @276 nm. Finally, the percentage of binding was calculated from the mass balance. The drug release over one hour in simulated gastric medium (900 mL 0.1N HCl) was also determined using the USP Apparatus II at 50 rpm and 37°C.
Results: The Minitab predicted an optimized composition containing 66.7 mg of each of the three polymers. The expected binding results by Minitab were 94.7% binding in water, 72.2% binding in 40% v/v ethanol, 55.7% binding in normal saline, 53.4% binding in 0.83 M acetic acid, and 32.0% binding in 0.1N HCl. The empirical results were similar to what was predicted by Minitab, taking into consideration the inability to measure the binding in the ethanolic solution due to the high viscosity of the sample and difficulty in its handling for subsequent dilution and UV absorbance measurement. The individual desirability indicating how well the optimized composition satisfies the goal of minimizing the binding in 0.1N HCl, while maximizing the binding in the other solvents varied between 0.00000-0.89475. The drug release from the tablets of the optimized composition was > 80% after 15 min. in 900 mL of 0.1N HCl.
Conclusion: The Minitab experimental design can successfully be used to optimize the composition of the deterrent hybrids that can effectively reduce drug extraction and deter drug abuse by injection. The optimized deterrent hybrids of CMS, CMC, and PAA could prevent drug extraction in most commonly used extracting solvents by >50%, while provided an intended drug release when used as intended.