Reliability assessment of gravimetric "in process" control of an IV compounding robot
2 ARNA Laboratory-ChemBioPharm U1212 INSERM - UMR5320, CNRS - University of Bordeaux, France
Automation of injectable drug manufacturing improve safety regarding manual process both for protection of operator and quality of final product. Our hospital purchased a robot, the KIRO Oncology®, equipped of an « in process » qualitative (datamatrix codes and photographic recognition of vials) and quantitative (gravimetric method) control. The aim of this study is to assess the reliability of this gravimetric control in order to know if it can replace an analytical control for batch release.
Material and method
In order to assess the entire process of preparation, we used a powder to be reconstituted as model drug. 24 bags of aciclovir were prepared. For each preparation, theoretical concentration at each step (after reconstitution in the vial, after dilution in the final bag if the reconstitution was fully accurate and after dilution in the final bag considering the entire process) were calculated according to the theoretical amount in the vial (500 mg), the theoretical volume of reconstitution (20 ml), the theoretical amount added in the preparation (400 mg equivalent to 16 ml) and the initial volume of each NaCl bag determined by gravimetry (107,41 to 109,79 ml). Simultaneously, we measured aciclovir concentration in vial and final bag by analytical method HPLC/UV-VIS and we corrected values according to purity of aciclovir batch announced by supplier (100,5%). Finally, concentrations in the reconstituted vial, in the final bag if reconstitution was fully accurate and in final bag considering the entire process were compared to theoretical expected concentrations for the 24 preparations.
Settings of admissible accuracies (-2%/+5% for reconstitution and -5%/+5% for dose withdrawal) allow to achieve a mean accuracy on the final product in the range of -4% with a maximum deviation of -6,67%. These results are in accordance with the accepted incertitude for a manual process (±10% to ±15%). To conclude, "in process" gravimetric control combined to products identification and traceability of each step can replace an analytical control on the final product joining industrial approach. However, in the context of quality approach, we plan to periodically and randomly introduce an analytical control on final product in order to predict and detect any deviation in comparison to initial process validation.
1. European Medicines Agency. Guideline on process validation for finished products - information and data to be provided in regulatory submissions. 2016.