Background: The United States Pharmacopeia (USP) has proposed large-globule-size limits to ensure the physical stability of lipid injectable emulsions, expressed as the percent fat >5 microm, or PFAT(5), not exceeding 0.05%. Visibly obvious phase separation as free oil has been shown to occur in some samples if PFAT(5) is >0.4%. We recently found that lipids, newly packaged in plastic (P), exceed the proposed USP limits and seem to produce less stable total nutrient admixtures compared with those made from conventional glass (G), which do meet proposed USP standards. We tested the possible stability differences between 20% lipid injectable emulsions in either P or G in a simulated neonatal syringe infusion study.
Methods: Eighteen individual syringes were prepared from each 20% lipid injectable emulsion product (n = 36) and attached to a syringe pump set at an infusion rate of 0.5 mL/hour. The starting PFAT(5) levels were measured at time 0 and after 24 hours of infusion, using a laser-based light obscuration technique as described by the USP Chapter <729>. The data were assessed by a 2-way analysis of variance (ANOVA) with Container (G vs P) and Time as the independent variables and PFAT as the dependent variable.
Results: At time 0, the starting PFAT(5) level for lipids packaged in G was 0.006% +/- 0.001% vs 0.162% +/- 0.026% for P, whereas at the end of the infusion they were 0.013% +/- 0.003% and 0.328% +/- 0.046%, respectively. Significant differences were noted overall between groups for Container, Time, and Container-Time interaction (all p < .001). Bonferroni tests showed significant differences in PFAT(5) levels between Containers at time 0 (T-0; p < .001) and T-0 vs T-24 for P-based lipids (p < .001), whereas no such differences were noted for Time for the G-based lipids. Similar results were noted for PFAT(10) levels.
Conclusions: We confirm that presently available lipid injectable emulsions packaged in newly introduced plastic containers exceed the proposed USP <729> PFAT(5) limits and subsequently become significantly less stable during a simulated syringe-based infusion. Although modest growth (p = NS) in large-diameter fat globules was observed for the glass-based lipids, they remained within proposed USP globule size limits throughout the study. Glass-based lipids seem to be a more stable dosage form and potentially a safer way to deliver lipids via syringe infusion to critically ill neonates.