Feasibility study: Semi-automated preparation and filling of continuous 5-fluorouracil (5-FU) administration devices

Context/Objective
In a context of constant increase in the number of chemotherapy preparations, the unit has equipped itself in 2021 with a preparation robot for cytotoxics and monoclonal antibodies (RIVA™, ARxIUM). However, some preparations, such as MINIFULL SET^® (MF) continuous 5-FU administration devices (Micrel Medical), cannot be robotized. In order to optimize the MF preparation process, a feasibility study of several preparation methods was carried out. The aim is to determine the best MF preparation process in terms of work ergonomics, precision and control of chemical and microbiological risks.

Methods
Several filling procedures were tested: - P1: 22800mg/570mL 5-FU Stock Solution (SS) by robot/MF filling by Baxa Repeater^® pump (BAXTER) under a laminar flow hood; - P2: PM by robot/MF filling by Medimix Vigo® pump (IMF) under a laminar flow hood; - P3 : SS and filling by Medimix Vigo^® pump under isolator; - P4: MF filling from 5-FU vials and a 0.9% NaCl bag by Medimix Vigo® pump into an isolator. The BAXA^® and Medimix Vigo^® pumps were used at flow rates of 165mL/min and 135mL/min respectively. A grid was drawn up to record preparation time. Gravimetric controls were carried out for the different processes (BAXA, MEDIMIX, RIVA). Chemical sampling and a Media Fill Test (MFT) with Tryptone-Soy Broth (TSB) were carried out for each process. TSB MFs were filtered, inoculated onto 4 culture media and incubated for 5 days at 33°C (Chocolate + PolyViteX, Meat Yeast, Blood agar) and 7 days at 22°C (Sabouraud).

Results
For each process, 5 MF were filled for 5-FU (4400mg/110mL) and 4 MF for TSB (25mL). With regard to gravimetric control, all MFs were compliant with an average error rate of +3.94% ± 0.45% for the Baxa® pump and +0.09% ± 0.00 for the Medimix Vigo^® pump. Preparation time was 1h07 for P1, 1h13 for P2 and 50min for P3 and P4, with human resources time of 24min, 30min and 38min respectively. For the chemical tests, traces of 5-FU were detected on the septa of the Chemolock^® for PM and 5 MF for P1. Samples from P2, P3 and P4 are currently being analyzed.
For MFT, no microbial growth was detected for all processes.

Discussion/Conclusion
The Medimix Vigo^® pump with integrated balance gives better results in terms of accuracy. The results of the MFT show that the microbiological risk is under control regardless of the filling location (hood adjacent to the RIVA or isolator). Preparation according to P3 and P4 is faster, but P2 would reduce the time spent by the preparation operator by 26%. The results of the chemical sample will be used to determine the most optimized process.

See the poster