Clinical Assessment of a Closed-loop Insulin Delivery System

Background:

Type 1 diabetes is caused by antibodies attacking insulin-producing β-cells in the pancreas. Treatment is usually by regular insulin injections, informed by glucose measurements from fingerprick blood samples. However, injections do not mimic the normal behaviour of the β-cell and this leads to suboptimal blood glucose control and complications including kidney failure, blindness, nerve damage and heart disease. Aggressive treatment can help but may lead to potentially-dangerous low blood glucose levels (hypoglycaemia). Glucose control is measured by HbA1c (normal range 4 to 6%), a measure of the amount of haemoglobin exposed to glucose over a period of around 3 months.

Current regimens for treating Type 1 diabetes in clinical practice are mainly based on injections of subcutaneous insulin several times daily in dosages determined by intermittent blood glucose measurements. The DCCT (Diabetes Control and Complications Trial) demonstrated that intensive management using these principles reduced complications by 50-76%. This was at the expense of increased hypoglycaemia, especially at HbA1c levels <7.5%. In other studies, intensive management resulted in people spending 30% of the day with glucose values >10mM and >2 hours/day in hypoglycaemia, often at night.

A closed loop system provides the potential to improve HbA1c while avoiding hypoglycaemia. It requires continuous glucose measurement, a control device and a pump for insulin delivery. The subject has been extensively reviewed. Intelligent control devices have been developed by others, using the principles of feedback control or predictive modelling. These were initially cumbersome e.g. the 'biostator' but more recent systems have been miniaturised and are capable of achieving blood glucose control in the fasting state, when provided with an input of interstitial glucose levels. They have not yet proven robust, may be associated with hypoglycaemia and are not capable of achieving adequate post-prandial control nor of coping with intercurrent illness outside hospital environments.

This clinical trial protocol assesses the Imperial College closed loop insulin delivery system. The closed loop insulin delivery comprises 3 main components: the glucose sensor, the control algorithm and the insulin delivery system.

The glucose sensor that will be used throughout the clinical validation studies is a CE marked, MHRA approved device manufactured by Medtronic. It is a subcutaneous sensor which sits just under the skin and samples interstitial fluid using an enzyme electrode. A small voltage is applied across the sensor and a current is fed back to the sensor instrumentation. This current is proportional to the glucose concentration in interstitial fluid and is calibrated against blood glucose a minimum of 12 hourly.

The control algorithm is derived from physiological experiments carried out by other groups which have demonstrated how the beta cells in the pancreas produce insulin in people without diabetes. Utilising the data from these experiments it has been possible to implement the behaviour of the beta cell in software and we have used a simulator with 200 virtual patients to demonstrate the safety and efficacy of the software. The data from the simulator is attached to this application as an appendix. The simulator was developed from human data and takes into account sensor errors, sensor placement, route of insulin administration and meal-time glucose absorption. It has been approved by the FDA in the United States as a step in the pathway of developing an artificial pancreas and has been validated against human data. In the clinical validation device the control algorithm is implemented on a printed circuit board using a programmable micro-controller.

The insulin pump device used throughout the clinical validation is the Roche Accu-Check Combo Spirit. This is a CE marked MHRA approved device and will be supplied by Roche with capability for direct communication from the motor so that we can verify the pump is doing what the software commands and with a license to use the communications protocol for research purposes. This ensures safe communication between the control algorithm and the pump and provides a fail-safe to ensure that the pump motor is responding appropriately to the control algorithm.

Clinical validation of the closed loop insulin delivery device follows a path of incremental challenges to the algorithm and hardware, starting with a fasting basal study in advisory mode and progressing to ambulatory, meal studies in full closed loop.

The aim of this trial is to assess the safety and efficacy of the closed loop device by applying the technology to participants with type 1 diabetes in a variety of scenarios, starting with a fasting test and progressing to overnight control, mealtime control and, finally, an ambulatory test.

Brief outline of each of the 5 visits within the trial period:

• Visit 1: Screening including clinical examination, fasting blood tests, completion of diabetes quality of life questionnaire, continuous glucose monitor attached to subject
• Visit 2: Review of continuous glucose monitoring results after 5 days
• Visit 3: Short Duration Fasting Closed Loop (6 hours of closed-loop assessment)
• Visit 4: Long Duration/ Overnight Fasting Closed Loop and Standard Meal Challenge (13 hours of closed-loop assessment)
• Visit 5: 24 Hour Ambulatory Automatic Closed Loop

During visits 3-5 blood sampling for capillary glucose \& ketones, venous glucose and insulin levels will take place every 15-30 minutes while the closed-loop insulin delivery system is running.