Optimization of the sterile compounding preparation process: example of a filling compounder and its qualification

Introduction
Nutritional care for preterm newborns is a challenge in clinical practice. We produce and dispense to the pediatric wards an admixture of vitamins and lipids in syringes, according to the French Guidelines for Neonatal Parenteral Nutrition, composed of water and fatsoluble vitamins. Firstly, a bulk solution is manually prepared, then filled into syringes using a peristaltic pump (Repeater^®). Batches of 60 units are produced weekly, involving time-consuming work and increases musculoskeletal disorders risks. The improvement of productivity and work environment is required. A new compounder device was purchased, to prepare the solution and to fill the syringes. Before using it, qualification is mandatory.

Material & Method
To perform the qualification of SLB Mibmix^® C1 compounder device (Hemedis, Germany), French Good Manufacturing Practices (GMPs) were followed. According to the GMPs, three qualifications must be done: Installation Qualification, Operational Qualification (OQ) and Performance Qualification (PQ). OQ was assessed by measuring volumetric accuracy, repeatability, reliability and by evaluating media fill tests process. Three filling speeds were tried, compared (slowest, medium, fastest speed), on five different theoretical volumes, five times each with several components (saline solution, water for injection (WFI), 50% dextrose solution and Clinoleic^® emulsion). The volumes were verified gravimetrically with an analytical balance (Sartorius Stedim, France). Aseptic filling tests were realized three times on batches of 60 syringes with trypticase soja and thioglycolate media (Merck, France). PQ was performed during a real-life batch.

Results
No volumetric difference was found between the three speeds and the five fluids. The volumes calculated were accurate (Bias<5%), compared to the theoretical volume. Accuracy profiles showed a trend towards improved accuracy with higher volumes, with α=5% (e.g. for WFI, relative error: 2.2% for smallest volumes, 0.3% for highest volumes), still acceptable. For every solutes, speed enabled an adequate repeatability (RSD<1%), its reliability was compliant with a bias ⩽ 1%. Regarding the media fill tests, no contamination was observed after seven days at 20-25°C followed by seven days at 30-35°C. To fill 30 syringes, four minutes at a medium speed were required. No time difference was found on small volumes, between the slowest and fastest speed. On larger volumes twice the time was needed between the slowest and highest speed.

Conclusion
We used to use the Repeater^® compounder device to split the bulk solution into the syringes. However a manual step was required. SLB Mibmix^® seems to be a good option for current parenteral nutrition activity. It permits to be more efficient and is less traumatic. This device allows the fully automation of the process.

See the poster