The Technology Jigsaw for the Artificial Pancreas

Let’s have a look at what scientists and manufacturers had to consider in their development of an artificial pancreas and the current technologies available on the market for those living with Type 1 diabetes.

As nature intended

An artificial pancreas needs to mimic a normal functioning human pancreas. Usually, the pancreas acts as a glucose sensor, monitoring any rise or fall in blood glucose; it produces both insulin and glucagon, which work to keep blood glucose levels within very tight limits regardless of what we eat or the activity we undertake.

What does an artificial pancreas have to do and how does it do it?

First of all, here is an image of the basic mechanism of blood glucose level control in the pancreas:

bgl diagram

But what happens with Type 1 Diabetes?

  • Insulin isn’t produced, therefore blood glucose continues to rise.
  • Glucose can’t enter the cells without insulin and so the body thinks there is no glucose available for energy.
  • Glucagon is released so glycogen in the liver is broken down to glucose.
  • This causes the blood glucose levels to rise further; to potentially dangerous levels which could lead to life threatening consequences such as diabetic ketoacidosis.

People with Type 1 diabetes manually monitor their blood glucose levels by pricking their finger or by using a continuous glucose monitor (CGM), a device that senses glucose levels via a needle inserted under the skin.

According to these measurements, they must inject insulin, often multiple times a day, adjusting the dose as necessary or by continually infusing insulin with a pump to keep blood glucose levels within a normal range. This requires a tremendous amount of manual diligence.

An artificial pancreas could automate this process and therefore has the potential to transform the lives of people with Type 1 diabetes.

Components of the artificial pancreas:

Glucose sensor Insulin Glucagon System
Pancreas Beta cells Insulin Glucagon Brain’s homeostatic systems
What we need… A glucose sensor that can tell immediately if glucose levels are rising or falling and by how much


A supply of insulin that can be instantly started and stopped


A supply of glucagon that can be instantly started and stopped


A system that automatically responds to changes rapidly and can decide on what the body requires to maintain equilibrium.


What we started with Up until the 1980s, most people only had access to urine testing. Animal insulin first extracted and used on a patient in 1922 Glucagon is unstable in its liquid state. Its powder form is mixed with a liquid before use.
What we have currently CGM has only been available people with diabetes since 2005 Analogue insulins are still relatively slow and long acting Progress yet to be made… As technology develops  so do the systems used for diabetes


Glucose sensors

The latest glucose sensors need to be calibrated with a blood test twice a day, but don’t need to do a blood test to confirm results. The Abbott Freestyle Libre, in its current model, could not be part of an artificial pancreas as it doesn’t sense 24/7. Although many of the current sensors can last for 7-14 days, longer duration sensors are needed to bring down cost and increase accessibility. Several products that last 3-6 months are in development.

Currently, CGM are the most automated. CGM doesn’t directly measure blood glucose but glucose in interstitial fluid, which lags behind what is happening in the blood by 10-20 minutes. While little can be done about the time lag, work is needed to make sensors more accurate and reliable.

NICE haven’t done a Health Technology Appraisal on CGM, so there is no requirement for the NHS to fund it and only around 100 people get CGM on the NHS in the UK with NICE recommending its use in adults and children only if certain criteria are met. Nevertheless, without an Appraisal showing cost effectiveness, it is unlikely localities will provide it, especially given costs of about £5,000.

It’s a chicken and egg situation; we need more people on CGM to lower the price. However, we need to make CGM more affordable which can only happen if more people are on CGM.


Glucagon kits are available for emergency use but not routinely used. This is because Glucagon is unstable in its liquid form, therefore not viable to be used in an artificial pancreas. Manufacturers are in the process of developing liquid based forms which would provide a perfect solution.  An experimental dual chambered pump is being trialled although there are similar caveats to overcome.

Current artificial pancreas models rely on regulating insulin and allow the body to naturally respond to a hypo. However within five years of diagnosis, those with Type 1 diabetes fail to produce an adequate glucagon response to hypoglycaemia. We need to develop a stable, liquid glucagon that can be used to mimic the natural response.


Analogue insulins are faster and shorter acting than any we’ve had previously. However, once insulin is injected, it continues to work – we can’t switch it off

For an optimum artificial pancreas to work we need to give small amounts of very fast and relatively short duration insulin to mimic the body.  But if the duration is too short, a bolus dose might not match the food absorption leading to later blood glucose rises.

Approaches are being taken to speed up the action of insulin:

  • Mechanical – warming the site to increase blood flow, increasing the area of a single injection.
  • Different absorption sites – intra-dermally rather than subcutaneously, inhaling insulin
  • Changing the molecule – adding excipients to modify how insulin is broken down or absorbed, reformulating analogue insulins

Where are we now commercially with the artificial pancreas?


Figure 1 JDRF Artificial Pancreas Step Generations diagram

The earliest commercial successes have been with sensor augmented pumps which are a combination of a pump with CGM.  These were designed to stop administering insulin if CGM readings fell below a set level, to prevent hypos, particularly overnight. The Medtronic Minimed 640G with SmartGuard enables users to set a target level at which the pump will be stop insulin delivery to prevent a hypo.


SmartGuard uses data from the CGM and an algorithm which helps it predict when blood glucose levels are falling, where they are likely to fall to, based on insulin on board. It then restarts the insulin as quickly as possible, once it reaches the target level. The first people using this technology avoided hypos over 80% of the time with less rebound hyperglycaemia.



The latest model, the Minimed 670G, hopes to manage day and night blood glucose, with the exception of meal times. It’s described as a hybrid closed loop system and has been tested in real world situations, such as exercise, unannounced meals, false sensor calibrations, lost transmission between the CGM and pump etc. Although not perfect it manages these issues safely.

Following a 120 person study the FDA allowed these people to continue to use the system, suggesting they could be open to approval of a hybrid closed loop system. This could be the first form of artificial pancreas to market, perhaps as soon as 2018.

What next?

We are seeing the use of combined technologies. However, the systems needs to be user-friendly, integrated rather than multiple devices, and more information on the safety, efficacy and reliability is needed.

Important questions remain to be asked.

  • What do people with Type 1 diabetes really want?
  • Having diligently managed their blood sugar levels, how will they feel handing control over to the artificial pancreas?
  • Could health services afford an artificial pancreas when currently there is a financial struggle for CGM on the NHS?




Comments disabled.