Managing glycaemic variability by eating low carb. Does it work?

As I mentioned on Twitter not so long ago, I’ve been undertaking a low carb diet for a few weeks as a great way to cut my weight before a summer holiday which will involve beaches and sunshine. Wearing a continuous glucose monitoring sensor while doing so also enables me to show the effects of this diet very effectively, and the three week averages (worked out on a per day basis) are shown below:

Now for any type one, that’s an incredible data summary. Spending 93% of time in range, 3% below and 4% over is superb. It’s worth digging deeper into the data and seeing what this looks like on a per day basis over the period it’s been going on, and comparing to a normal period.

Can you manage glycaemic variability?

Before we go into that though, it’s worth bearing in mind that even when eating non-low carb, my last Hba1C (done on the 4th July, so just as this was getting underway) was 5.9%/40mmol/mol so I’ve not been getting large volumes of high readings.

The starting date for this period of low carb eating was 3rd July. The amount of carbs per day is between 18g and 35g, dependent on which week we are looking at, split <30g in week 1, <20g in week two and <40g in week three. Saturdays are a refeed day and I visit the gym two or three times per week to undertake weight training. My daily calorie intake is roundabout 2000 calories, and the entire aim of this program was to lose weight.

From a weight management perspective, this is also proving to be incredibly successful. On the July 3rd, I registered my highest weight in a year, at just over 15st. This morning, I registered 14st 7lbs. That’s a half stone (3.2kg) weight loss in 16 days, which is exactly what I wanted it to be. It looks like the below:

The refeeds show up clearly as weight gain spikes, mainly driven by fluid gain as a result of eating carbs, and then drop off again fairly quickly. I’m (obviously) fairly pleased with this graph.

So let’s dig into the glucose numbers. The last three weeks can be seen in the three graphs below:

What’s striking about all three of these is the consistency and closeness of the trace lines throughout the centre of the range. In the second there is a large spike that was Friday/Saturday, and I’m suspicious that this was caused by that bugbear, hidden sugar in food. I ate out in the evening, and in spite of a “low carb” meal, saw the spike. Something had sugar in that I didn’t know about.

What’s great across all of these is the Standard Deviation number, which on the best week was 1.1 and the worst, 1.8. Given the average blood glucose levels, this is 19% in the best case and 29% in the worst. Both numbers are normally considered very much in target. There are also not a lot of lows. As I’ve mentioned, in terms of life, I’ve not been doing anything different. Purely eating lower carb. Exercise is still in there, a couple of glasses of wine are in there. There’s a whole load of things in there.

This has a secondary side effect, which is obviously that with less carbohydrate, less insulin is required. In the summary chart below, you can see the last occasion where a higher carb diet was eaten and the transition into low carb. It’s very clear that both my glycemic variability and the amount of insulin I need is reduced.

When shown like this, the difference is night and day. The first part of these two graphs includes the period where I was at Glastonbury. I was very pleased wit the result of this, but in comparison to the last three weeks it looks dramatically awful!

So how did I achieve these numbers? For the most part, I split my macronutritional elements in the below range:

• Protein: 720 calories – 180g per day
• Fats: 1160 calories – 130g per day
• Carbohydrate: 120 calories – 30g per day

This works for me. For obviously for anyone else, you need to experiment, and it goes without saying that I’m doing this myself. If you have any doubts, then you should consult your favourite medical professional. Mine are fairly good in this respect and let me get on with it. Some may be less encouraging.

An example daily diet looks a lot like the below:

To get to the macro levels required, pretty much everything is homemade. The lasagne replaces pasta sheets with oven baked Aubergine slices. The chicken burgers are home made. We bulk cook and then I take food into the office and take advantage of the provided microwave.

The macros on the above menu are:

As can be seen, this is basically in line with what I stated earlier. I do eat higher protein than a classic Low Carb diet, but this is in response to resistance training and not wanting to lose muscle mass.

For the most part I find that I can modify my insulin for protein by doubling my insulin:carb ratio, for example, if I am 1:8 on carbs, I am 1:16 on protein. This seems to work well. The pump manages the dawn phenomenon to great effect, and if I have a particularly large meal, I may use a square wave bolus, but I’ve not found a huge amount of need for it.

I buy low carb “treats” from an online source that makes great brownies and biscuits, which means I can have a “low carb” brownie for 5g of carbs and stay within my numbers. I love cake, so look for alternatives! For example, Fathead Pizza, is a great alternative, and deliciously filling:

Glycaemic Variability – why is it important?

That was the how. It clearly demonstrates that, in answer to the original question, it works. It also raises a different one, namely, why is it important?

The answer to this seems to lie in presentations given at the Diabetes UK professional conference, where Nick Oliver presented evidence that micro-vascular diabetic complications seem to have more to do with the oxidative stress caused by fluctuating glucose levels than with specifically long running high levels. In theory, at least, by reducing the variability, one should reduce the risk of retinopathy, neuropathy and nephropathy.

Variability is often blamed on a lot of factors. Stress, hormones, exercise, the weather, which way is the wind blowing, etc. And yet. I’d suggest that the biggest factor in variability is purely the food you eat. If you don’t eat carbs to any great extent, your variation has to drop.

That can be seen in my summary data. And if a better example is needed, this is a week eating “normally”, and by that I mean around 180g of carbs a day:

It shows a much more up and down set of data. The variation is still considered in professional circles to be relatively good, but the consistency is much worse (and my highs are generally not exceedingly high). It reflects my experiences perfectly, in that with a higher carb content, I have far more every day ups and downs.

Some conclusions, and for the T1 GRIT community, teaching people to suck eggs