Vancomycin syringe study shows significant reduction in dosing variability after introducing a revised protocol

Dear Sir, An unexpected variation in plasmatic vancomycin concentration in a newborn admitted to our neonatal intensive care unit (NICU), prompted us to systematically measure the concentration of the antibiotic in the syringes prepared by the nurses. The objective of the study was to determine the causes of dosing errors by systematically assessing each step of the nurses' preparation process. Vancomycin was prescribed for continuous infusion in 20-mL syringes. The drug amount prescribed varied from 20 to 80 mg over a 24-h period, according to the baby's weight, and was infused at a flow rate of 0.3 mL/h. Nurses prepared the vancomycin syringes under a vertical laminar airflow hood. The commercial drug was a vial containing powder for intravenous infusion solution (Vancomycin Mylan 250 mg, Mylan, Saint-Priest, France). The average amount of vancomycin base present in the vials was first checked in 25 vials, using analytical scales. The successive steps for preparing a vancomycin syringe were as follows: We carried out six assessments of each step of the preparation process to see what effect this had on the amount of vancomycin. This assessment comprised: After reviewing the protocol, we came up with revised steps for preparing a vancomycin syringe: We assessed the accuracy of the amount of vancomycin in the syringes prepared for intravenous administration in the NICU before and after revising the preparation protocol. Nurses prepared thirty syringes dosed at 40 mg of vancomycin. Vancomycin amounts were determined just after preparation using a UV spectrophotometric method. This showed that the average amount of vancomycin base was 229 ± 1 mg in marketed vials, which was 91.6% of the expected amount of 250 mg. The amount of vancomycin in the syringes prepared by the nurses before the protocol was revised was significantly lower than the theoretical one (mean ± SD, 33.83 ± 4.58 mg vs. 40 mg; Student's test, p < 0.001). When it came to the method of reconstitution, we observed that the amount of vancomycin was significantly different between water and saline, with or without stirring (Kruskal–Wallis test; p < 0.001), especially between water with stirring and saline without stirring [median (min-max): 34.5 mg (34.2–35.1) vs. 27.6 mg (25.5–33.0); Conover and Iman, p < 0.0001]. The drug transfer method and dilution process were also associated with a significant difference between the amounts of vancomycin (p = 0.043). After revising the protocol, there was still a significant difference between the observed and theoretical amounts (35.80 ± 0.88 mg vs. 40 mg; p < 0.001). The relative standard deviation (Fig. 1) was significantly lower before we revised the protocol (2.5% vs. 13.0%; Fisher's test, p < 0.001). Our main hypothesis is that the dosing variability of vancomycin syringes is mainly due to the solvent used to reconstitute the solution and the nonsystematic practice of stirring to ensure that the powder is completely dissolved. Another way to reduce these risks and errors is to standardize and centralize preparation and reconstitution in the hospital pharmacy 1. The authors wish to thank the members of the NICU care team, University Hospital of Lille and Dr Damien Lannoy for their dedication and commitment to this study and the manufacturers, Safic Alcan and ACS Dobfar, for providing free vancomycin powder samples for our research.