Application of the variability budget approach to the Dissolution test

Abstract

The aim. This study aimed to evaluate the completeness of our knowledge about the sources of variation in the Dissolution test with 100 % release by compiling a variability budget.
 Materials and methods. The study was performed on 500 mg metformin tablets, using pharmacopoeial quality reagents, State Pharmacopoeia of Ukraine (SPhU) Metformin HCl reference standard, Pharmatest DT70 Dissolution apparatus, Perkin Elmer Lambda 35 spectrophotometer, Mettler Toledo XP 204 analytical balance, and ISO class A volumetric glassware. The SPhU metrological approach was employed.
 Results and discussion. The variability budget was compiled based on the comparison of uncertainty estimates obtained from the requirements for maximum permissible variation in normal analytical practice (UNAP, bottom-up estimation) and experimental data (Uexp). This involved characterizing Metformin content in tablets using the Uniformity of Dosage Units (UDU) test as an independent method. The 100 % release of Metformin in the Dissolution test (infinity point) was proved by increasing the dissolution time. Having optimized Dissolution and UDU analytical procedures for variability budget compiling, we achieved insignificance of Uexp compared to the target uncertainty (Utg) for the Dissolution test in compliance testing. The differences in UDU and Dissolution mean results did not exceed UNAP for the release time of 45 and 60 min, i.e. uncertainty budget was proven. Uexp for the Dissolution test indicated the presence of an unknown statistically significant source of random variation, which, however, was less than Utg; therefore, the procedure is suitable for compliance testing.
 Conclusion. Experimental results confirmed the completeness of our knowledge about sources of variation (absence of bias) for the Dissolution test with 100 % release. An essential condition for compiling the budget was the optimization of uncertainty of analytical procedures. For UDU, all significant sources of variation were within the expected range. Yet, there is a need for additional research to identify and manage an unknown source of practically significant random variation for the Dissolution test