Exercise and Type 1 Diabetes: part 2

A picture I found on the Interwebs that amused me

In part 1 I tried to set out the problem of exercising with Type 1 Diabetes (T1D) - in brief, keeping blood glucose levels within reasonable bounds while hormones stimulated by activity are doing their best to frustrate your efforts. Here is part 2, which contains a few things that might help to manage the situation.

Managing blood glucose and insulin

Let’s start with the basic theory. Usually, with low or moderate intensity exercise and some active insulin on board, blood glucose will fall steadily and relatively predictably, and insulin will work more effectively. In order to avoid a hypo then, you would need either to reduce your mealtime insulin at the meal beforehand, or consume carbohydrate during the activity, or both. It is estimated that between 30g and 60g of carbohydrate is needed per hour to fuel moderate exercise.

So you could measure your blood glucose level before and after your activity and see how much it drops – say, from 11 to 6 mmol/L over 30 minutes fast walking 2 hours after a meal containing 60g carbohydrate for which you took half your usual dose of rapid insulin. If on another occasion your blood glucose was only 8 mmol/L before the same activity in the same circumstances, you could predict that carbohydrate would probably be needed to prevent a hypo.

To be able to reduce your rapid insulin dose at the previous meal, the activity needs to be planned or anticipated. Often activity is not planned, in which case there is no option but to eat or drink carbohydrate, unless your blood glucose happens to be high anyway. You can see that this makes it much more difficult to lose weight by exercising than for someone without diabetes. So another tactic that was suggested to help weight loss was to do the activity when insulin levels are at their lowest, usually first thing in the morning, although clearly this also requires an element of planning. But I can’t see how that would work if blood glucose is also at its lowest, because that’s just asking for a hypo, so maybe you’d have to reduce your overnight background insulin so that fasting blood glucose levels are a bit higher than usual. I’m not a fan of messing with background insulin on a day-to-day basis, which I will outline later on in this huge essay.

Blood glucose doesn’t always drop with exercise. If the activity is anaerobic (sprint, weight lifting, resistance exercise at the gym) then blood glucose tends to rise because those other hormones (especially adrenaline) stimulate the release of glucose and increase insulin resistance. In this situation extra insulin may be needed to take blood glucose levels down rather than extra carbohydrate to prevent hypos. A stressful or competitive situation like a football match where adrenaline is a factor may have a different impact on blood glucose compared with regular football training, and may need a different insulin dosing strategy.

This effect can be used to your advantage. If blood glucose before an exercise session is between 4 and 7 mmol/L, then starting with anaerobic or high intensity/stressful exercise may raise blood glucose enough to allow you to carry out some aerobic exercise without the need for insulin or carbohydrate adjustment ahead of time.

So we can start to imagine types and duration of activity and the likelihood of blood glucose rising and falling so that insulin and carbohydrate can be managed before and during exercise. Then comes the aftermath.

There are two effects of exercise on blood glucose after the activity is completed. The first is that glycogen stores in muscles and the liver have been depleted and need to be restocked, which makes blood glucose drop in the hours following the exercise. The other is that activity makes muscles more sensitive to insulin (less resistant) particularly in the period between 7 and 11 hours after exercise – the stress hormones released during activity induce insulin resistance for about 7 hours afterwards. For exercise in the afternoon or evening, this period of greatest hypo potential occurs during the night. Exercising first thing in the morning means the period of maximum hypo risk occurs during the day rather than overnight, which may be helpful.

Ways to manage this hypo risk after exercise include taking carbs on board immediately after exercising, and/or reducing the amount of insulin given for subsequent meals and corrections by about 50%, and possibly also reducing overnight basal insulin (but see below). Another option uses adrenaline to raise blood glucose levels by incorporating a 10-second sprint at maximum exertion level at the end of the period of exercise.

Blood glucose monitoring is the key to managing the amount of carb/insulin to maintain good control after exercise. Some experimentation is likely to be needed, while bearing in mind the poor reproducibility mentioned earlier. Perfection is unlikely to be achieved.

Background insulin adjustment

So far, all the insulin adjustment has been with the rapid insulin that works with carbohydrates that are eaten or drunk. But it is possible to adjust the background (basal) insulin too, and it was at this point that our practice and the recommendations within the study day diverged.

Background insulin works over long periods – from 12 to 72 hours depending on the type. Reducing the long-acting insulin will reduce the hypo risk overnight, so the advice on the course included routinely reducing this insulin both before and particularly after exercise. Doing this will certainly reduce the hypo risk, but on the other hand calculations of rapid insulin will be thrown out of kilter if background insulin is being adjusted day to day, especially if you exercise some days but not others. We didn’t reach any consensus on this point, so I suppose I’d have to look in the research literature to see if there’s anything relevant there.

I can, however, see the point of a basal adjustment for a short continuous period of daily exercise like an activity holiday – skiing, watersports or walking holidays being the most common examples. And I had not considered the pros and cons of different background insulins before – the newer, very long lasting insulins being less flexible if background insulin is to be adjusted. It’s also true that adopting a more active lifestyle will probably reduce the need for total background (and rapid) insulin, but injecting different amounts of long-acting insulin on a daily basis might be problematic.

What about insulin pumps?

So far all the discussion has been based on multiple daily injections of rapid-acting and long-acting insulin. Pumps are a bit different, because they only use rapid-acting insulin, and basal rates can be adjusted hour by hour. So with a pump there’s no problem about reducing background insulin as well as rapid mealtime insulin to avoid the need for extra carbs or to reduce the risk of hypos. This raises the chances of better control as well as being an advantage if weight loss is one of the aims of doing the activity. Reducing insulin is usually preferable to increasing carbohydrate for the ‘ordinary’ person. Proper athletes will want the carbohydrate, though.

The reduction suggested on the course was to set a temporary basal rate (TBR) of 50% for an hour before and up to an hour after aerobic exercise. If extra insulin is needed for anaerobic exercise, the course recommended raising the basal rate by only 10% starting 30 minutes before and lasting until 60 minutes after the activity. The TBR might be reduced again by 10% in that crucial period 7 to 12 hours after the exercise. There are more complicated formulae for calculating TBRs but I will leave those to the serious competitors.

The main downside to a pump is that it needs to be attached to you, and most types are not waterproof. So the pump would need to be disconnected completely for contact sports or watersports, which is really only safe to do for an hour or so. Some pumps can’t be disconnected temporarily, like the tubeless pumps which are actually attached to the skin. This type is usually waterproof for bathing or swimming up to an hour or so, although it clearly wouldn’t be suitable for scuba diving, and might be dislodged in a rugby scrum or during martial arts.

For situations where the pump has to be disconnected for longer than an hour, competitive athletes sometimes connect up with the pump from time to time to give themselves a quick bolus, or revert to the use of basal and bolus injections from a pen to maintain insulin levels on those occasions. When the pump is reconnected then there may be a need for a correction, which could take one of several forms. You could increase the basal rate by 50% for up to an hour, or give 50% of a correction bolus, or even work out how much basal insulin was missed and bolus half this amount. Then, of course, be a bit more rigorous about monitoring and correcting blood glucose levels.

What else?

There are a whole lot more factors that affect management of T1D with exercise, some of which I haven’t mentioned up to now because they are routine, like the need for fluids. Dehydration not only affects athletic performance but can make the blood glucose level appear higher than it really is.

Heat and cold also affect the uptake of insulin from the injection site as they do at any time. The location of the injection site matters because if you’ve injected near a muscle that will be used for the exercise (usually leg or buttock/lower back) then the insulin will reach your bloodstream faster than if you injected in a non-exercising part of the body.

Keen exercisers may use Continuous Glucose Monitoring (CGM) either standalone or in conjunction with an insulin pump. The main point to highlight with CGM is that there is a delay between the readings they give for the glucose in interstitial fluid and the level of blood glucose, which may not matter if you’re in an office and it’s coming up to lunchtime, but may be critical if you’re just reaching the summit of a mountain.

Carb intake: it has been established that the requirement for carbohydrate during moderate intensity exercise is around 1g per kg body weight per hour, i.e. for a 70kg person that would be around 70g per hour. It has also been established that the gut can only absorb dietary carbohydrates at the rate of 60g per hour, so there is no point trying to increase intake beyond this as it will just cause gastro-intestinal discomfort. The difference is made up by the use of stored glucose and fat as fuel.

All foods are not equal, but the question of which carbs to have at what time was not covered in the course. Of course hypos associated with exercise have to be treated with fast-acting carbohydrate as at any other time, and it would make sense to have slow-acting carbohydrate to sustain any prolonged period of activity. Beyond that, I suppose it has to be trial and error with plenty of blood glucose monitoring to find out which foods before, during and after exercise have the best effect on blood glucose levels. Aside from diabetes, the prevailing view is that a mixture of protein and carbohydrate such as cereal+milk, yogurt or meat/cheese sandwich is a good idea post-exercise to replenish glycogen stores and supply material for muscle regeneration and repair.

The overall message I took away from the study day was that exercising with Type 1 Diabetes is very, very complicated if you want to do anything more exciting than up to an hour of moderate intensity exercise in a regular controlled environment like road cycling, a run around the park or an hour in the gym. Competitive athletes need much more insight into their own physiology, but it is possible to compete at the highest level, and one of the diabetes pharmaceutical companies sponsors competitive cycling with the Team Novo Nordisk.

I have had a couple of patients asking me questions about serious exercise, and we have very quickly reached the limits of my knowledge. I don't see that changing much as a result of this course, but perhaps over time I will absorb more on this subject alongside my greater experience in diabetes as a whole.

Pumping insulin

October 2012

In the early days of insulin-treated diabetes, people with Type 1 diabetes used to draw up insulin from a vial and inject using a hypodermic syringe. Nowadays, most people use 'pen' devices. These have a pre-filled insulin cartridge, a way of 'dialling' the number of units of insulin you want to inject, and a very tiny thin needle, which can be as little as 4 mm long. [Note: the linked website made me chuckle with its description of one of the insulin pens as "designed for people who don't like needles and children."]

However, today's blog is about one of the newer developments in insulin delivery, the insulin pump. This consists of an insulin reservoir containing rapid-acting insulin that provides continuous subcutaneous insulin infusion, 24 hours a day, via a cannula inserted into the skin. Most insulin pumps have a tube between the reservoir and the cannula, but there are tubeless types as well. The cannula can stay in place for two to three days. You can safely disconnect for about an hour, for showering, sports, trying on clothes etc.

For those of us who don't have Type 1 diabetes, our insulin level is adjusted minute by minute in response to all sorts of hormonal signals, such as levels of adrenaline, cortisol and growth hormones as well as the amount of blood-borne glucose. To imitate the insulin-secreting action of the normal pancreas, people on multiple daily injections (MDI) inject two types of insulin: long-acting once or twice a day, which is designed to release slowly into the circulation over a period of around 24 hours, and quick-acting insulin that lasts about 4 hours, injected at mealtimes to deal with the glucose entering the circulation from the carbohydrates that are eaten.

Like the human pancreas, an insulin pump only has one type of insulin, but two separate ways that it is delivered. There is a background or 'basal' rate that delivers the constant low-level drip of insulin day and night. At mealtimes, the pump user still has to decide how much 'bolus' insulin to give and at what rate - the pump doesn't remove the need to test blood glucose and calculate the carbohydrate content of food, and the injected insulin still doesn't act as quickly as insulin secreted by the pancreas directly into the blood circulation. But there are advantages to having more control over the amount of insulin going in.

For example, the basal rate can be adjusted on an hourly basis, and some pump manufacturers set up their pumps with a variable basal rate to match the general circadian pattern found in most people. A pump can also be of great benefit when exercising or drinking alcohol, because both of these activities tend to lower blood glucose, and it can be frustrating to have to take extra carbohydrate in order to avoid a hypo, especially if you're trying to lose weight. With a pump, a temporary lower basal rate can be set. [Note: exercise with diabetes is a fascinating physiological puzzle that I've been trying to figure out for the past few weeks, and I'm sure a blog post will appear on that subject very soon.] If you're on a once-daily injection of background insulin, then you can only adjust the background rate over a period of days.

There's lots of flexibility with the bolus insulin too (which could almost be done with MDI if you were prepared to give several extra injections). Many people find particular foods take ages to digest, especially if there is a lot of fat mixed with the carbohydrate (pizza is a typical example). If they inject just before or just after the meal, the rapid-acting insulin deals with the first surge of glucose but then its period of action is over before the food is fully digested, and their blood glucose level drifts upwards. With a pump, you can program the meal bolus to be delivered either in multiple bursts (just after the meal and then 30/60 minutes after or whatever) or at a constant rate over a period of time. This also works for really long meals, like a posh dinner with big gaps between courses, or for buffets and parties where you might be grazing over a long period of time.

Other advantages with the pump: it can help you with the sums, so you can be more accurate with your insulin dosage. For example, if you have calculated that your meal contains 48 grammes of carbohydrate and your ratio is 1.8 units of insulin per 10 grammes of carbs, the pump can do the calculation without blinking and deliver fractions of a unit (8.64 units in this case). On MDI, everything would be rounded off because we don't hold 1.8 times tables in our heads, and most pens can only deliver whole units of insulin (50 g carbs at 2 units/10g = 10 units of insulin). The ratio of insulin to carbs may change through a day (many people are more insulin resistant in the morning) and that can be programmed in to the pump software. And if you tell it your blood glucose reading, it can also suggest a correction dose, which can even take account of 'insulin on board', i.e. any insulin previously injected that may still be having an effect.

Of course there are disadvantages, not least the cost of the pump, which is available on the NHS only for those who meet the funding criteria. There's the inconvenience of wearing the pump at all times, which can place limits on clothing: at the beach, or on special occasions (e.g. with a party dress). The most worrying to me, based on what I've seen of patient care for diabetes in hospital, is that with a pump you have no long-acting insulin on board, and if your pump is disconnected for any reason your blood glucose level will start to rise after just an hour or two. If this happens at home, perhaps because of a blocked tube or kinked cannula, you should spot it and be able to deal with it. In hospital, the general awareness of diabetes is fairly minimal, and knowledge of insulin pumps is non-existent. There might be a risk of not being given insulin because it is assumed that you will have some level of background insulin on board, and that could have serious consequences.

Where I work, there are special clinics for pump patients, and I attended one of the monthly education sessions where pumpers can drop in to catch up on particular aspects of treatment. My colleague RSB is covering the pump clinic at the moment, so I don't yet have much contact with this group of patients, except for Mr M, who is a pump user, and who is always happy to enhance my knowledge of the issues.

[Update: Mr M informs me that contrary to my assumption, the first wearable insulin infusion pump was invented by Dean Kamen in 1970, while the first insulin injection pen device was introduced and marketed by Novo Nordisk in 1985. Thank you, Mr M!]