Another insulin pump

Krakow Botanical Gardens, July 2016

There are a number of different insulin pumps available on the UK scene - unfortunately one fewer than there were last week, as one company (Animas) has withdrawn from the pump market. We support three different pumps in our service, although the majority of people use one particular brand (Roche). A training day was offered by Medtronic, one of our less used suppliers, and all three of our nurses, one doctor and I attended to learn more about how the pump is used and about some other associated products.

It's a 'tubed' pump, which means that a tube carries the insulin from the reservoir in the pump down to a cannula inserted subcutaneously. [Tubeless or 'patch' pumps (like the Omnipod) are mounted directly onto the skin with the cannula on the underside of the pump.] A blood glucose monitor communicates with the pump, and an insulin calculator is integrated into the pump rather than the monitor. The main consequence of this is that the pump has to be accessible for anything other than a preset bolus amount - the other brands we use can stay hidden under clothing as all their functions can be operated from the blood glucose monitor alone.

Aside from this one drawback, I like the Medtronic pump. It has potential beyond just being an insulin pump because it can be hooked up with an integrated Continuous Glucose Monitoring (CGM) system. This allows for anticipation of low blood glucose (hypoglycaemia) and the pump can actually suspend insulin delivery for up to 2 hours ('low glucose suspend'). It can do this automatically or with user input, with or without alarms, day or night. This has been shown to prevent hypos without leading to high blood glucose afterwards.

'Low glucose suspend' kicks in when blood glucose is within 3.9 mmol/L of a user-selected 'Low Limit' and predicted to drop to less than 1.1 mmol/L above the low limit within 30 minutes. So if the user selects a low limit of 3.4 mmol/L, insulin can be automatically suspended if 4.5 mmol/L is predicted when blood glucose is 7.3 mmol/L, as measured by the CGM sensor which is monitoring glucose levels through another cannula. Insulin is restarted either when the user intervenes, or when blood glucose is predicted to rise 2.2 mmol/L above the low limit within 30 minutes, or after a maximum of two hours. I started to argue about why the user-specified low limit couldn't be the actual prediction (4.5 mmol/L) rather than 1.1 mmol/L below it (3.4 mmol/L), but our tutor wasn't following my argument so I had to abandon that line.

We looked through the pump's menus, set up two basal patterns, set temporary basal rates, entered blood glucose levels, delivered different types of bolus and generally learned about all the different settings. We were shown different cannulas and inserters and stabbed ourselves with those, and then looked at the reports that can be produced, both with and without the associated CGM trace. They left this bit until the end, and I would have preferred a bit more time to work out what the reports featured, where the most useful and important information could be found, and a bit more practice at interpreting the tables and graphs.

The chap leading the training day was a healthcare professional who works for Medtronic and goes out to patients to start them on pumps, review, or troubleshoot. He also has Type 1 Diabetes himself, and uses a Medtronic pump. Occasionally he would throw in some personal experience about how he would manage a situation himself. Nobody, and I mean nobody that I have met through our diabetes service (or elsewhere) treats their diabetes in the way that he does. "But if I don't do it like this I'll have a higher HbA1c [i.e. worse control], so why wouldn't I?" he said. True, but he is in a bit of a unique position, understanding his diabetes and how the pump works as he does. And he seems more motivated to manage his diabetes than almost anyone I've met.

[For those who know about the numbers, he tested his blood glucose about an hour before lunch, it was 7.1 mmol/L with no active insulin on board and he gave himself a bolus of less than one unit of insulin. "My target is 5.5, I've got the technology, so why not?" he said.]

This company is aiming at development of the technology beyond the 'low glucose suspend' option. They have plans for a 'hybrid' system that will adjust basal insulin all the time according to CGM results, not just when blood glucose is dropping. Bolus insulin for food will still be down to the user, which distinguishes the hybrid from the 'artificial pancreas' level of technology (which does exist but is still too experimental for mainstream use).

At the end of the day we were also reminded of another bit of kit that the company produces, which consists of a cannula that can be worn for up to three days (like a pump or CGM cannula). This one, however, is designed for manual injections, so its function is to enable someone to give all their multiple daily injections for three days through the one device. The obvious benefit of this device is fewer skin-piercing injections; the downside is that by the end of three days the effectiveness of insulin injected in the same spot can fade a little, and the risk of developing lipohypertrophy ('lipos', or fatty lumps) is higher.

The sales pitch suggested that this is a device that would benefit everyone, but would it benefit someone who is doing well and doesn't mind multiple injections? Such a person would be moving their injection site around, wouldn't risk lipos and wouldn't experience the decrease in the effectiveness of their insulin. Balanced against that is the benefit of reducing the unpleasantness of multiple injections. It's difficult to judge the psychological strain of many injections compared with one cannula for three days, but so far I haven't started pushing this device to patients.