Training for the ultra – preparation

The chart shows how my blood sugar was affected by exercise and the carbs I ate during the run. The green line represents the time I was running for.

The chart shows how my blood sugar was affected by exercise and the carbs I ate during the run. The green line represents the time I was running for.

I’ve said before that by character, I’m pretty happy-go-lucky. But for anyone doing an ultra, I’m told that preparation is key. This is doubly true for someone with diabetes.

To that end I’m making headway with a few things on the preparation list. Firstly, I don’t want to get injured with all this running I’m doing, so I made an appointment to see a physio today to get myself checked out and get advice on injury prevention. The session started out well (“you’ve got strong quads”) but quickly turned to character assassination (“you don’t use your calves or bum muscles and you have no flexibility”) and ended in daylight robbery (“I can solve all your problems in just five expensive one-to-one sessions”). But seriously, he gave me really good advice, and after coming to terms with the heavy investment I’m going to have to make in both time (daily exercises – I’m starting off with kneeling practice – it’s harder than it sounds!) and money, I’m hoping that it will reduce my chances of injury.

The next thing is a plan for nutrition when running. This is the really complicated bit for me. And it’s why I’m so looking forward to all my long runs: I can collect more data!

On Saturday, I went down to the South Downs to do a 20 mile run. I decided getting the train out of london was a good idea, because at least the Downs have some hills which is good practice. I ran to my aunt and uncle’s house, and was greeted by another aunt, two cousins and a massive spaghetti bolognaise, which was a really nice surprise.

The run went really well (except having to stop ALL THE TIME to look at the map and figure out where I was going – oh, and to do those pesky blood tests). But my 20 miles took 3.5 hours, and apparently – according to Strava – I was actually moving for 2.5 hours. If that’s true that’s pretty quick. I didn’t take any short acting insulin (I want to avoid doing that whilst running because it’s too complicated and risky) and ate on average 30 grams of carbs per hour. This is good information about how much my body can process whilst exercising.

There are two problems though. Firstly, according to running websites, I need to eat more like 70g or carbs per hour whilst running an ultra. Secondly, my blood sugar went down to around 4 (or just under) a couple of times during the run. I was totally fine, and didn’t suffer symptoms of a hypo, but I probably need to give myself a bit more of a safety buffer if I don’t want to risk having problems at some point in the future. So next time I’m going to try to eat slightly more per hour – my theory is that if I eat more, my body might start producing it’s own insulin (I’m in the honey moon phase, so still produce some insulin) which will allow me to digest the extra carbs and eating more will give me a higher safety margin from hypos.

The other thing I’m finding with my training, is that I’m absolutely loving all this running! I enjoy the feeling of running along, and it’s great to be able to cover long distances in amazing scenery. I suppose this is part of the “gift of diabetes” – I’ve always enjoyed running, but I wouldn’t have attempted to do an ultra marathon this year if I hadn’t been diagnosed with type 1.

Data Running

Will information liberate me?

“Knowledge is power. Information is liberating.” Kofi Annan.

Not a bad place for a hypo! The photo doesn't do it justice, but we came round the corner of the hill, the clouds parted and we had a spectacular view of the Argentiere Glacier. By testing regularly during exercise, I hope to gain a better understanding of how to manage my blood sugar and minimise risk.

Not a bad place for a hypo! The photo doesn’t do it justice, but we came round the corner of the hill, the clouds parted and we had a spectacular view of the Argentiere Glacier. By testing regularly during exercise, I hope to gain a better understanding of how to manage my blood sugar and minimise risk.

When I was diagnosed, one of my first questions to Dr Powrie was “can I still climb mountains and run long distances?” He told me that it would be complicated, and that I should maybe reign in my ambitions.*

I’m not going to reign in my ambitions.

Having diabetes makes doing prolonged physical activity a more risky because of the risk of experiencing a hypo (low blood sugar). It is particularly important to bear this in mind when half way up a mountain, because it is hard to get a paramedic up a cliff, and it can endanger the diabetic and his companions.

Fortunately, mountaineering is already a risky business involving lots of kit. When climbing a mountain, or descending a snowy slope on skis, one has to constantly assess risk of falling, of weather, of avalanches etc. So I already have transferrable skills I can use to manage diabetes – it’s just an additional risk factor to manage, and it requires another load of kit.

The desire to keep on doing all this stuff has given me the motivation to learn as much as I can about the condition, and part of that is to collect a load of data. I’m suffering from a bit of “computer programmers block” at the moment and can’t quite decide how best to organise the mountain of data I’m creating every day. The crux of the problem is as follows: a normal person has to eat the right food to fuel their body whilst spending a day running or climbing. I need to do that, but also take the right mix of long and short acting insulin, and the right amount of carbs to stop me from experiencing hypos.

I’ve already found, for instance that if I’m running fast (for me, I’m defining a “fast run” as any distance up to half marathon) my blood sugar generally goes up for the first 45 minutes of exercise. If I’m doing less intense exercise, my blood sugar will go down. If I’ve taken short acting insulin before exercising (for instance if I’ve gone for a walk after lunch) my blood sugar will drop fairly quickly. It’s all very complicated and whilst my intuition is improving, I think I need to analyse the data more formally as well to give me the best chance of optimal blood sugar control, and decrease the probability of bad hypos.

I’ve created some charts to help me. Here’s an example of two runs I did – one half marathon (fast) and one run over two and a half hours which involved a climb of 1200m (slow).

The chart shows how my blood sugar changed depending on exercise and carb intake. I was running during the period between the green boxes. The blue diamonds show my blood sugar at different times, and the red boxes show how many grams of carbohydrate I ate at different times. Running fast (I have defined my half marathon pace as "fast") seems to mean that my liver releases glycogen into my blood stream at a quicker rate than I can absorb the glucose for the first 45 minutes of a run. After an hour I started eating jelly babies to prevent my blood sugar from falling too much.

The chart shows how my blood sugar changed depending on exercise and carb intake. I was running during the period between the green boxes. The blue diamonds show my blood sugar at different times, and the red boxes show how many grams of carbohydrate I ate at different times. Running fast (I have defined my half marathon pace as “fast”) seems to mean that my liver releases glycogen into my blood stream at a quicker rate than I can absorb the glucose for the first 45 minutes of a run. After an hour I started eating jelly babies to prevent my blood sugar from falling too much.

This is the same kind of chart, but it was a very different run. I ran for almost two and a half hours, and climbed 1200m. It's impossible for me to run fast doing that, and from the limited data I have (I didn't test after half an hour which would have showed the initial response from my liver) I would say that at this slower pace my liver does not release so much glycogen into my system.

This is the same kind of chart, but it was a very different run. I ran for almost two and a half hours, and climbed 1200m. It’s impossible for me to run fast doing that, and from the limited data I have (I didn’t test after half an hour which would have showed the initial response from my liver) I would say that at this slower pace my liver does not release so much glycogen into my system.

The data collection will continue, and I’m really looking forward to doing more long runs to find out about my insulin and carbohydrate requirements.

*I don’t want to make Dr Powrie sound like a killjoy. It was good of him to manage my expectations. I’m very lucky to have found such a good doctor and I have found him very supportive over the past two months.

Data

Insul-independent

I am of course exaggerating with the title of this blog post. I have type-1 diabetes, therefore I am insulin dependent.

However, since being diagnosed seven weeks ago, I have become much less dependent on insulin than at first. In my opinion, this is down to two things. Firstly, when newly diagnosed, diabetics often experience the “honeymoon effect”. I don’t think this is fully understood, but it seems that before treatment starts, the remaining insulin producing beta cells are working flat-out and are totally knackered. (A bit like the shortly-to-be diagnosed diabetic!) When treatment starts, these cells recover a bit and can produce more insulin which helps newly diagnosed diabetics to control their blood sugar fairly easily.

Data What is type 1?

Will I go blind? HbA1c can help measure the risk

THIS IS NOT TO SCALE! Illustrative chart to show how average blood sugar (HbA1c) is linked to the risk of developing complications.

THIS IS NOT TO SCALE! Illustrative chart to show how average blood sugar (HbA1c) is linked to the risk of developing complications.

Before I start, I’d like to remind you that I’m not a doctor and all my knowledge about diabetes comes from conversations with my doctor (which I may misremember) and Wikipedia. Oh, and personal experience!

When one is diagnosed with diabetes, it’s not long before the word “complications” is encountered. Diabetics are more likely to develop heart problems, eye problems (including blindness), kidney failure and ulcers in the feet. It is my understanding that consistently high blood sugar is a causal factor in all of these. So an important reason to measure blood sugar is that it allows a diabetic to assess how successful their blood sugar control is and whether they need to change anything.

An important test for this is the HbA1c test, first used in the Seventies. When the glucose content of blood is high (this can occur in a healthy person immediately after a glass of coke for example), glucose molecules attach to hemoglobin in the red blood cells. Red blood cells live for up to three months, so it is possible to find out how much glucose has stuck to the hemoglobin and therefore find a measure of average blood glucose levels over the past couple of months.

HbA1c can be expressed in different ways, but many people use a percentage. A healthy person will have an HbA1c of between 4% and 5.9%.

According to my doctor, the risks of developing complications rises exponentially as HbA1c rises. To understand what this means, just look at the graph on the top of this post. You can see that as HbA1c goes from 7 to 5.9, the risk of complications goes down by the amount in the lower shaded area on the left hand side. So the risk decreases but not by very much! If a diabetic has higher average blood sugar though, and their HbA1c goes from 11 to 10, then the risk of complications goes down by the higher shaded area on the left hand side of the graph.

What this means, is that if blood sugar is high, there are really big gains in terms of long-term health by controlling it better. If blood sugar is low (say HbA1c is 6.5%) then whilst risk does decrease by bringing blood sugar down, it doesn’t decrease by much. At these levels, other lifestyle factors such as smoking are much more important. (Another reason why diabetics have to eat a super healthy diet. Keeping cholesterol low, for example, is important in reducing the risk of heart disease.) So at low levels of HbA1c the costs of getting average blood sugar down further (increased risk of more hypos) probably outweigh the benefits. I will be advised by my medical team what balance to aim for, and other diabetics will be too – it depends on personal circumstances, sensitivity to hypos etc.

My doctor and I discussed all this in the context of heart problems, so I’m not 100% sure that the graph is the same for other complication such as eye problems.

I can’t remember what my HbA1c levels are. They are still too high. When I was diagnosed they were through the roof, and on my most recent visit last week they were much lower. (Dr Powrie was very complimentary saying how impressed he was with how much it had come down!) Anyway, I’m not going to worry about them yet. I’m sure (I hope) that a few months of high blood sugar around my diagnosis won’t make much difference to my risk of complications, and as I get things under control, my HbA1c should come down by itself.

Data What is type 1?

I have quantified myself

Blood glucose chart. This shows the range my blood sugar has been in over time. I'm aiming to get the coloured bars in between 4 and 8.

Blood glucose chart. This shows the range my blood sugar has been in over time. I’m aiming to get the coloured bars in between 4 and 8.

I’m a bit behind with my blog posts, so I’m going to catch up by writing a load over the next few days. At the end of August I attended the “Quantified Self” meet up in London. I said before that these people were a load of “geeks” who quantify a load of stuff about themselves, either for interest or to help them understand themselves or train for a goal. Of course, I mean geek in a good way – since getting diabetes I have become even more of a geek than I already was.

The evening was really good fun. I met a load of interesting people both before and after the formal session. I met someone who was attempting to create tailored career advice from a huge database of personal traits and job satisfaction for different jobs, an ultra marathon running cell biologist (a very useful person for a diabetic who likes running to talk to), someone who is an expert in virtual currency (bitcoin), a programming expert, a PhD student etc. Meeting all these people and having a load of stimulating discussions wouldn’t have happened before I received the “gift of diabetes”. Haha.

During the main part of the session, people can give presentations on how they are “quantifying themselves”. I saw three entertaining and informative presentations. The most relevant to me was probably the cell biologist who is training for an ultra marathon. He is quantifying a load of personal data to track progress. Running distance and speeds, of course, but also weight, VO2 Max, waist line etc. I was particularly impressed that one of his criteria for his quantified self programme was that he would spend no more than ten minutes a day tracking his personal data. Given that diabetes is so time-consuming (I spend well over an hour a day thinking about it, doing maths, changing needles etc.) minimising the time impact on my day is a worthy aim!

The quantify self presentations can be seen here. The most amazing one (unfortunately I wasn’t there to see it) was given by Dr Ian Clements, who had bladder cancer and decided to track a load of personal data about himself because he was frustrated with the infrequency of medical check ups to diagnose the progress of his cancer. Amazingly he found a statistically relevant correlation between the difference of fat in each leg and the progress of his cancer.

So Dr Clements has joined a long list of people who inspire me. (Having watched the Athletics World Championships recently, Mo Farah is another one. What a legend!)

So back to someone less impressive: me! I’m now some of the way through my computer programme which I use to analyse the data generated by my diabetes app. I’ve produced loads of charts which I’ll probably post on here over time. The one at the top of this post is my favourite (at the moment anyway) because it shows my blood sugar control. I take many glucose readings at different times of the day, but I always take one when I wake up, before each meal, and when I go to sleep. So at least five a day. To give a graphical representation of my blood sugar range, I’ve calculated the average of my five readings for each day and the “standard deviation” of the five readings (the standard deviation is a measure of variability – if it’s high it means the readings a very variable, if it’s low it means the readings are all at very similar levels). I can then do a bit of maths which calculates some ranges.

In the chart above, 90% of my blood glucose readings should be somewhere in the coloured bars. 50% of my readings should be between the two darker blue bars in the middle. I don’t know for sure, but I think the blood glucose range I’ve had isn’t that far from what a normal person would have (Matthew Beard who gave the QS presentation on running did a blood glucose test on himself – he’s not a diabetic – and after drinking a bottle of lucozade his blood sugar varied from just above 4 to just below 8). If I can keep it in this range, my risk of heart disease shouldn’t be any different from a normal person (more on this later).

So a month and a half into having diabetes – well done me! I’m sure it won’t be possible to control blood sugar this consistently at all times, but it’s a good start.

Data

Never ending data

I’m discovering that being a diabetic is a mammoth exercise in data collection.

This evening I am going to eat dinner. Sounds easy? Well it’s not! In order to eat dinner without sending my blood sugar levels off the charts, I need to take enough insulin to allow my body to absorb the carbohydrate I’m about to put into it.

So this evening (with the help of the carbs and cals app on my phone), I have discovered that the amount of carbohydrate in the following foods is: 70g per 100g of noodles; 5g per 100g of carrot; 7.5g per nectarine; 1g per avocado. So I work out that my dinner will have 100g of carbohydrate. (I am now committed, and can’t just help myself to another slice of bread or two after taking my insulin.)

I then have to work out how much insulin to inject. This, I am discovering, is complicated. It depends on what my current blood sugar reading is, what I would like my reading to be by my next meal, how many carbs I’m eating and what exercise I’ve done or am likely to do. Dr Powrie has told me to take 0.75 units of insulin for every 10g of carbs I eat. But I think I only need to take 0.6 units, so that’s what I’m doing.

How do I know that? I’ve been recording carb consumption, insulin consumption, exercise and blood sugar levels (luckily I have an app for that too). And I’m making an educated guess at how much insulin I need to take based on those variables.

I am measuring my blood sugar level obsessively. I want to know what it is when I wake up, before and after exercise, before meals, two hours after meals, and when I go to bed. And maybe even in between. I find myself constantly looking forward to my next blood sugar test. What will the reading be? Can I guess? I also really look forward to my trips to the hospital to see Dr Powrie – they are opportunities to learn more about the condition and more about how it is affecting my body.

My new app can send data files to excel, so I’m trying to manipulate them to learn more about how my blood sugar levels change. I have a hair-brained scheme that after I have collected enough data, I can perform a statistical method called “multivariate regression analysis” which will allow me to predict my required carbohydrate consumption to maintain safe blood sugar levels whilst running or climbing.

By coincidence, this weekend in the Times I read about a bunch of geeks called “self quantifiers” who measure all kinds of things about themselves like sleep patterns, steps taken, heart rate, calorie consumption and a whole load of other things. They use a whole host of gadgets which usually speak to their smart phone, and then they use, or often create, apps to analyse all the data. This all started in San Francisco (of course) but there is a London group who meet in Google’s office near Old Street. Maybe I can learn something from them that will help me manage my diabetes. I’ve joined and plan to attend their (our!) next meeting.

I’m sure I will be blogging more about how exercise affects blood sugar therefore carbohydrate intake and insulin doses as I discover more about it.

Data Diagnosis