An artificial endocrine pancreas (AP) consists of a real-time glucose sensor, an insulin infusion pump and a computer control algorithm (controller). The integrated system can automatically determine the appropriate dose and timing of insulin based upon the absolute glucose measurement and recent glucose trends, integrating the rate of increase or decrease, and the recent dose of delivered insulin. Predictive and adaptive control algorithms use a model of patient physiology, insulin pharmacokinetics and insulin dynamics to determine the appropriate dose of insulin for subsequent delivery (51). Closed-loop control algorithms
FIGURE 8 Omni-Pod insulin pump manufactured by Insulet Corporation has separate patient-worn insulin delivery system (right) connected wirelessly to hand-held glucose meter/controller (left). The insulin delivery system is able to deliver a variety of basal and bolus doses of insulin, according to finger-stick SMBG measurements (www.insulet.com).
have been developed that automatically regulate the delivery of insulin without patient intervention. It is difficult for closed-loop systems to achieve normal BG control because of time delays in the CGM recognizing an increase in glucose due to food, and delays in insulin absorption from the sc tissue. Semi-closed-loop algorithms overcome this limitation by utilizing data input from the patient regarding meal timing/composition, and the onset of exercise. Inadequate glucose sensor accuracy and robustness remains the major obstacle to routine application of an AP system in the clinical setting (28-30).
Was this article helpful?