Stability study of several insulins in a binary parenteral nutrition mixture.

Introduction
Parenteral nutrition (PN) infusion often induces hyperglycemia, increasing patient morbidity and mortality through the risk of infection. To limit this side effect, insulin is administered at the same time as the PN. Because subcutaneous administration of insulin increases the risk of hypoglycemia (unexpected discontinuation of PN), adding rapid insulin directly to the PN mixture could represent a safer alternative. However, as recommended by the SFNCM, a stability study must be carried out beforehand. A previous study showed a 50% reduction in human insulin concentration starting 6 hours after its addition to an industrial binary mixture (binary part of an Olimel^® bag), related to a protein glycation reaction.
The aim of this work is to evaluate the stability of several insulins in an industrial binary PN mixture associated with the formation of a glycated form of these insulins, in order to establish a recommendation for the use of a preferred insulin in this situation.

Methods
Human (Umuline Rapide® (UR)), aspart (Asparte Sanofi^® (ASP)) and glulisine (Apidra^® (API)) insulins were each dissolved (n = 3; 0.07 mg/L) in an industrial binary NP mixture (Smofkabiven^®), supplemented with vitamins and trace elements. These solutions were stored at 25°C for 24 hours. After optimization of insulin detection parameters and their glycated form, an internal HPLC-MS/MS calibration method was validated. Insulins were quantified at different times from t0 to t24h to express the rate of insulin disappearance over time. In the meantime, the ratio of the areas of the glycated forms over the area of the internal standard was monitored.

Results
A rapid decrease in insulin concentration was observed starting t1h for UR and ASP, while for API it was more gradual. At t6h, the decrease was respectively -42 ± 11%, -50 ± 4% and -14 ± 4% for UR, ASP and API. Then, a plateau phase is reached showing a decrease at t24h of -37 ± 6%; -52 ± 7% and -19 ± 11%. In contrast to these decreases, the areas of the glycated forms over the area of the internal standard increase in a correlated manner.

Discussion/Conclusion
The decrease in UR concentration, in the NP mixture used, confirms the results of the previous study, and in comparison, it is higher for ASP and much lower for API. The proportion of glycation in insulins seems to differ from one speciality to another (different structures and compositions), and the use of API in PN seems to be the preferred choice. It would be interesting to test the stability of other insulins in PN and assess the bioactivity of glycated versus non-glycated forms.

See the poster