Modelling insulin adsorption in intravenous infusion sets in the intensive care unit. (30th June 2019)

• Record Type: Journal Article
• Authors: Knopp, Jennifer L.; Hardy, Anna R.; Vergeer, Sarah; Chase, J. Geoffrey
• Is Part Of: IFAC journal of systems and control. Volume 8(2019)
• Publication Date: 2019-06-30
• Subjects: Glucose control -- Insulin adsorption -- Insulin delivery -- Compartment model -- Infusion tubing; Automatic control -- Periodicals; Relay control systems -- Periodicals; Embedded computer systems -- Periodicals; Feedback control systems -- Periodicals; Artificial intelligence -- Periodicals; Artificial intelligence; Automatic control; Embedded computer systems; Feedback control systems; Relay control systems; Electronic journals; Periodicals; 629.89
• DOI: 10.1016/j.ifacsc.2019.100042
• Languages: English
• ISSNs: 2468-6018

Abstract

Abstract: While contentious, some studies suggest regulating blood glucose (BG) improves clinical outcomes. Model-based approaches have been successful at modulating BG in intensive care unit (ICU) setting using insulin. However, insulin adsorption by infusion set materials may reduce the quality of control. Previous studies highlight the occurrence of significant adsorption, but are highly variable in methodology and results. No study considers or accounts for all potential variables affecting insulin adsorption by infusion tubing. This study identifies and provides a first model for insulin adsorption. A two-compartment insulin adsorption model is developed based on conservation of mass. It describes 'free' (F) insulin in solution and 'bound' (B) insulin adsorbed to material surfaces. Experimental data from the literature is used to validate the model based on fitting error and consistency in identified model parameters for the same material. Parameters k 1 and B e q describe the adsorption rates of insulin and the equilibrium for insulin bound to material surface, and are identified for two different infusion set materials (PE and PVC). The model dynamics fit the experimental data well. A steady state analysis shows it converges to the expected final result. Values B e q mostly fell within the expected range, but overall parameter values varied significantly between studies, materials, and flow rates. These results show the need to refine the model, or discretise the infusion set tube into several identical compartments to account for spatial changes in adsorption across the infusion set tube length, to better capture the observed dynamics independent of other variables. The results also show the need for better experimental setup and reporting to provide enough data to assess more complex model structures.