• 24 Jun 2019
  • 15 min
  • 24 Jun 2019
  • 15 min

Justin Coleman interviews David O’Neal about insulin pumps. What’s new and what do GPs need to know? Read the full article in Australian Prescriber.


Welcome to the Australian Prescriber Podcast. Australian Prescriber, independent, peer-reviewed and free.

I'm Dr Justin Coleman, a GP on the Tiwi Islands in the Northern Territory, unfortunately home to plenty of diabetes and I'm also editor of the Diabetes Management Journal, auspiced by Diabetes Australia.

Today I'm speaking with Professor David O'Neal who's an endocrinologist at the Werribee Mercy Hospital and St Vincent's Hospital, Melbourne, and a researcher at the University of Melbourne. And we're talking today about insulin pumps, a technology which is advancing, particularly as we're standing at the crest of the wave of technological feedback systems which start with a bang... that's real-time glucose measurement... and end with a squirt via the insulin pump.

Professor David O'Neal, welcome to Australian Prescriber.

Thank you very much, Justin, it's my pleasure to speak here.

Professor, insulin pumps have been around since the 1970s, around the same time as the first brick-sized mobile phone. I'm hoping that a lot has changed since those times. Can you describe an insulin pump for us?

An insulin pump is a pager-size device and contained within it is an insulin reservoir which contains rapid-acting insulin. This reservoir is attached to tubing, which is flexible and of varying length and then the tubing ends in a cannula, usually Teflon, which sits under the skin and delivers the rapid-acting insulin subcutaneously.

The pager-size device itself is really a mini computer. It can be programmed to give infusions of subcutaneous rapid-acting insulin at different rates and the background rate of insulin can be varied in almost infinite number of ways, in a pre-programmed sense, throughout the day and night.

So Professor O'Neal, you mentioned a plastic cannula. So is that a bit like a butterfly needle size thing with a plastic sheath over the end and the user inserts the needle under the skin and then withdraws the needle?

Yes it is. Exactly. There's a metal component which is used at the time of insertion. The metal component will be withdrawn, leaving the Teflon component sitting under the skin. And this can be at a varying angle, either 90 degrees or at an angle into the subcutaneous fat. And these are usually inserted into the abdomen and sometimes it's inserted into the buttock area, above or below the belt line.

Okay. And how stable is this set-up... still thinking about the hardware for the moment... does everything need to be taped down including all the tubing and how reliable is that?

The cannula should be changed every three days and certainly one of the issues with these devices, is if the cannula fails because you've only got rapid-acting insulin being infused subcutaneously. It's not like you've got a whole lot of background glargine insulin sitting there if you fail to give an injection. Once the cannula fails, because you're only infusing rapid-acting insulin, the insulin washes out over a period of four hours and there is a risk of ketosis if the line fails and it's not recognised.

Once it's taped in, it's usually fairly robust. The tubing itself is flexible and it does allow a degree of mobility. The pumps themselves are also fairly robust.

So things like during exercise it tends to remain still working well. And how about overnight, people sleeping on it and rolling over? It's all fine?

Overnight is not usually an issue. Men might hook it into their pajama pants or put it in a little pouch around their neck, as may women, or women may hook it into a bra. So rolling around at night isn't usually a major issue. During exercise, depending on the degree of exercise... for example, if it's very vigorous exercise... people may disconnect the pump for a couple of reasons. One, because they want to withhold insulin while they're having that brief period of exercise and we don't suggest that they stop the insulin delivery for any prolonged periods of time. We're just talking about certainly less than two hours. Or if they're sort of going into some very vigorous exercise such as wrestling or white water rafting, we would suggest they disconnect the pump because of the risk of damage.

Okay, so the pump can be detached from the mechanism when you're swimming, showering, white water rafting etc.

Yeah. If you're showering, the pump can be suspended and removed. There's a little locking mechanism where the tubing attaches to the cannula. You disconnect that area.

Very good. Thank you. Well look, enough talk about the hardware. Let's look at the liquid inside and talk insulin dosing. So you mentioned pumps only ever use rapid-acting insulin. Could you run us through the basal dosing system, which is the regular background rate of insulin?

Yeah. Background insulin requirements will vary according to the time of day, both according to the diurnal rhythm, with cortisol going up in the early hours of the morning, but also with physical activity as well and stress. So we can vary the background insulin, the rate at which it's infused in a pre-programmed way during the day and night. And usually the average patient will require about somewhere between three and five different background rates over the period of the 24-hour cycle.

And how frequently does the insulin actually enter the body?

It's literally every five minutes.

Okay, so it's five minutes throughout the 24-hour period.

Yeah, little micro-boluses and it'll vary depending upon the manufacturer of the pump.

Okay. And that continues overnight, just at a lower amount of insulin each time?

It may be lower or it may be higher, depending on what the patient's requirements are. If the patient's extremely active during the day, you may actually have a higher basal rate at night and certainly with the cortisol surge in the early hours of the morning, you may require one basal rate for the first half of the night and a second basal rate as the night goes on after 3:00 am.

Okay, thank you. So that's the background basal doses, which I take it the user doesn't need to worry about so much, certainly on a day-to-day basis?

Certainly the patient doesn't need to automatically adjust the insulin delivery according to those time blocks because they are pre-programmed, but they do need to set them and they do need to adjust them in response to what they're seeing with their glucose level. So in that sense they do need to worry about them.

Let's talk now about the things that would concern the user in the short-term, in the immediate term, which is the magic, I guess, the secret sauce, which is bolus dosing and this is where the user can adjust for a meal or perhaps for exercise. And to do that they need to take two things into account. The current blood glucose level and also how that level might be expected to increase with the... I don't know... hamburger they're about to eat. So could you take us through the bolus dosing things?

When a patient boluses and uses their bolus calculator in their pump, they will need to perform a finger-prick, measure their glucose levels. A lot of meters will actually automatically transmit that information to the pump so they don't have to manually enter it in. The patient will also have to be able to look at, if they are going to eat, what the carbohydrate content of the meal is and they will have to manually enter that into the pump.

Pre-programmed into the pump, what we call an insulin-to-carbohydrate ratio, basically how much insulin is required for each gram of carbohydrate, and this can vary according to the time of day. Typically you need more insulin in the morning and less insulin in the evening per gram of carbohydrate. Also programmed into the pump is an insulin action time, because if they've given a bolus before, how much of that is still around so it doesn't end up stacking. So you give one bolus and then another bolus and it prevents you getting this tidal wave of insulin coming in. If the pump feels that there's still a lot of insulin on board from an earlier bolus, it will modify that dose or it won't let you bolus the second time.

So the insulin action time is the amount of time insulin is going to be active in your bloodstream.

Yeah, and usually that's about three hours. Some people have four, some people have two. And finally we have a target glucose because if your glucose levels are high, we set a target. Maybe one target during the day, which is a little bit lower and one target at night, which may be a little bit higher... you tend to be a little bit more cautious at night... and there's a correction factor thrown in as well where you program into the pump how much your sugar is expected to drop per unit of insulin, and it may be one unit will drop... you say it's 3 mmol/L... it's a calculation we do.

So as a general rule, the basal dosing influences the pre-meal glucose levels, which would include the overnight level, I guess, the early morning, while a bolus dosing influences the post-meal glucose levels. And post-exercise as well, I guess. What do people do about exercise?

This is extremely complex. Usually what we suggest to patients pre-exercise, is they can drop their basal rates in preparation for it, usually one-to-two hours beforehand. If their glucose levels are below a certain amount, say less than 7 mmol/L, we suggest they have some carbohydrate before the exercise. Exercise isn't really one big homogenous intervention. High-intensity exercise can actually push your sugar glucose levels up because of the counter-regulatory hormone, while moderate-intensity exercise will tend to drop it. So it also depends on the exercise and the mix of high-intensity and moderate intensity.

Most people, because they're awake, will adjust to this or address this. Probably the biggest risk is there tends to be an enhancement of insulin sensitivity overnight after you exercise. And you also need to advise the patients that they may need to drop their basal rates overnight after they go to sleep if they've done some fairly vigorous exercise in the evening.

So it sounds like the whole insulin pump is a very individualised thing. The user gets to know their own body and how their body adjusts and then their health professional advising them also has to very much individualise it for each person.

That is correct. But I preface this conversation by talking about a conventional pump. Very recently we've had a glucose sensor talking to the pump, giving continuous glucose information from the interstitial space under the skin. And that basal rate, which previously we had to adjust manually, is now adjusted every five minutes based on glucose levels. And this simplifies the process considerably because an individual's insulin requirements, even overnight, can vary between 200% and 400% without us actually knowing why this is the case. It can be stress, can be exercise, can be the meal. So when we program these basal rates in conventionally, it's sort of like trying to pick a tee shirt and one-size-fits-all. Most nights you'll probably get it right, but some nights it'll be way too much, it'll be too big a tee shirt and some nights it'll be way too little. It'll be too small a tee shirt.

And having it automatically adjusted in response to measured glucose levels every five minutes helps to basically address this issue, which is an important issue with people with type 1 diabetes, and it does allow them, for example, to sleep well overnight and it does improve glucose control. And now we've got what we really call a closed-loop system, a glucose sensor that's measuring the insulin continuously and talking to the pump, adjusting those basal rates.

Could we finish now with just some other things that GPs might need to know about insulin pumps? I imagine some doctors might get a bit worried about what does this button do? For the doctor who can't work their TV remote control, what degree of mastery is a GP or another doctor expected to have over an individual's insulin pump when they come into the surgery?

I think the fairest thing to say is the patients usually know their pumps. I think humility on the part of the health professional in sort of letting the patient explain to us how the pump works. They live with these devices every day.

That equanimity with my ignorance is of a great relief to me, Professor O'Neal. In terms of prescriptions, what sort of prescriptions are we writing for people using insulin pumps?

It'll be overwhelmingly rapid-acting insulin and the patient will know which one they're on. Usually they prefer the insulin in cartridges, some prefer the vials so they can load the reservoirs in their pumps.

They also may need some backup insulin for if the pump fails or if they're travelling etc?

Yeah. Certainly if there is a pump failure they should have some long-acting insulin such as glargine or Levemir. If they're prone to hypoglycaemia, do they have a glucagon script that's current, because some of them will have a glucagon hypokit at home but it's out-of-date, and does their significant other know how to give it if necessary?

And these aren't cheap items, of course. How much does a pump cost?

About $10,000 and they've got a warranty for four years. In addition, there are the sensors. They are reimbursed in children and there's recent legislation come out where they will be reimbursed to certain vulnerable groups. If you ever pay for those yourself and insurance doesn't cover it, about $75 a week.

Looking at adjusting doses now, I gather the users can upload their results onto a software program. Is this likely to be feasible on a GP's computer or is it just health professionals who specialise in adjusting these things?

The patients can upload it to the web. The site they upload it to obviously will depend upon the manufacturer of the pump. Get the patient to upload the pump at home and when they come to see the GP they can log on, on the GP's computer, with a username and password that is specific to the patient and look at their reports. The reports are comprehensive. Care of these patients is never dependent upon one individual healthcare professional. We really view this as a team approach where a group of us are all helping to support this very vulnerable cohort.

Thank you. That all makes sense to me. I enjoy being a team player. Professor David O'Neal, it has been an education talking to you about insulin pumps for Australian Prescriber today. Thank you very much for coming along.

It's my pleasure, Justin.


Professor O'Neal's conflicts of interest include receiving income on advisory boards for Medtronic, producer of insulin pumps, and Novo Nordisk and Sanofi, producers of insulin. The guest's views are his own, not NPS's and I just make up things as I go. I'm Dr Justin Coleman. Thanks for listening to another Australian Prescriber Podcast.