Stability study of a ternary parenteral nutrition admixture after addition of human insulin
1. Univ. Lille, EA 7365 - GRITA - Groupe de Recherche sur les formes Injectables et les Technologies Associées, F-59000 Lille, France
2. CHU Lille, Institut de Pharmacie, F-59000 Lille,France
3. Clinical Nutrition, Baxter R&D Europe, Boulevard d’Angleterre 2-4, B-1420 Braine-l’Alleud, Belgique
Parenteral nutrition admixture (PNA) induces administration of high dextrose amounts. Hyperglycemia is a frequent side effect, with potential deleterious consequences. Consequently, human insulin is sometimes directly added into the PNA bag. Drug mixing with PNA is strongly not recommended, especially in the case of ternary parenteral nutrition (TPN). Indeed, TPN admixtures, corresponding to emulsions, are supposed to be potentially destabilized, even broken by the addition of drugs or other external nutrients. The main goal of this work was to study in vitro stability of a TPN admixture after addition of human insulin into the bag.
Material and methods
Three industrial TPN admixtures (Olimel N7E®, Baxter) supplemented with vitamins (Cernevit®, Baxter) and trace elements (Nutryelt®, Aguettant) were prepared as reference. Human insulin (Umuline rapide®, Lilly) was added at a final concentration of 20 IU/L, in 3 other TPN bags. PFAT5 defined by the US Pharmacopoeia1, represents the percentage of lipidic droplets having a diameter over 5µm; when below 0.05% threshold, admixture is considered stable. Both pH and osmolality were measured in triplicate immediately after insulin addition and over a 24-hour period. TPN bags were stored at 25°C. For each condition, samples were taken from bags at 0, 4 and 24 hours after insulin addition and immediately analyzed. Descriptive analysis of pH, osmolality and PFAT5 values was performed (mean ± standard deviation).
PFAT5 varied from 0.0180 % to 0.0045% without insulin in TPN, vs. 0.0170% to 0.0041% in the presence of insulin. Whatever the medium (with or without insulin), pH values at t0 were comparable (6.35 ± 0.01 vs 6.36 ± 0.01) and evolution was < 0.1 unit. Concerning osmolality, initial mean values were similar: 1700.3 ± 14.0 mOsm/kg without vs. 1718.7 ± 15.9 mOsm/kg with insulin. Finally, osmolalities remained constant between t0 and t24 (relative standard deviation were respectively 0.7% and 0.3% without and with insulin).
Discussion and Conclusion
All PFAT5 values were below the threshold of 0.05%1,with and without insulin. This hypothesis is confirmed by the constance of both pH and osmolality values. The addition of human insulin in TPN admixtures seems a good solution to avoid hyperglycemia in patients treated with TPN. Indeed, both dextrose and insulin are rigorously administered together. However, further studies are needed to prove that admixtures of both TPN and insulin remains stable. In the present work, stability of TPN emulsion has been demonstrated but these results could be strengthened by zeta potential measurements and size distribution of lipid droplets. Moreover, insulin stability studies in real conditions have to be performed to complete these experiments.
1. United Standard Pharmacopoeia, chapter 729.