Comparison of three automated compounding devices for parenteral nutrition according to four key technical tests.

OBJECTIVES: Automation of parenteral nutrition (PN) preparation is nowadays a recommended practice in order to reduce human errors and thus improve the safety and accuracy of the finished product. Other benefits of automation may include full documentation of preparation and a reduction in personnel requirements or staff injuries. The market of automation compounding presents different automated compounding devices (ACDs). The aim of this study is to compare the technical characteristics of ACDs by carrying out four specific challenges. METHODS: Three ACDs: Two piston pumps with ACD 1: MediMix Multi 4120R (Impromediform) and ACD 2: Mibmix Compounder C12 (Hemedis), and one peristaltic pump ACD 3: ExactaMix Compounder EM2400 (Baxter) were assessed in a pharmaceutical manufacturing unit within a controlled atmosphere area, under horizontal laminar flow hood (LFH) according to four tests: volumetric accuracy, flush volume, smoke test, and a production test with three configurations of PN bags. For this test, a PN bag was considered accepted when all quality controls (weight, molar concentration of sodium, potassium and calcium) were fulfilled. RESULTS: The maximum relative biases found for the different ACDs were heterogeneous. ACD 1 had the best volumetric accuracy with respect to supplier specifications and for extreme volumes (0.2 mL). Evaluation of the flushing volume allowed the validation of 50 mL volumes for ACD 1 and ACD 3. The smoke test was only conclusive for ACD 1 under a horizontal LFH. The percentage of PN bags accepted were 98.8% for ACD 1, 70% for ACD 2%, and 95.5% for ACD 3. CONCLUSION: This study compared three ACDs according to four relevant and specific tests. Based on the data acquired, we conclude that ACD 1 is the most accurate, has the lowest flushing volume, is suitable for use in a LFH, and achieves the best results in the production test.