Guy's Blog

Guy frequently keeps this blog updated with thoughts, challenges, interviews and more!

Skittles beat Watermelon: blood glucose experiment update

Watermelon is worse for me than Skittles.* Who’d have thought?

If you haven’t read my post on testing blood sugar response to foods, you’d better do that before proceeding. Just to recap briefly, here are the assumptions/opinions/beliefs I’m working from:

1) it is better to avoid spiking your blood sugar levels

2) your blood sugar response to specific foods is unique to you. What spikes mine may not spike yours.

And let me re-state for the gadgillionth time: I’m not a medical doctor. I’m not a biochemist or a nutritionist. I’m a martial arts teacher, documenting the results of some experiments I’m conducting on myself for my own reasons and following my own approach, and sharing for your entertainment and/or interest. It’s up to you what you do with your body. 

I have been testing my blood sugar levels before and after meals, to determine what foods I’m eating regularly that I should actually avoid, and in the hopes that there will be foods I avoid for health reasons that I could actually eat without causing damage. By far the best part of this has been testing things like Skittles, FOR SCIENCE.

Let’s start with the testing techniques and process:

First, the finger prick. They say it doesn’t hurt. They lie like bastards. It hurts exactly as much as you would expect jamming a steel spike into your finger would hurt. But it gets much less painful over time, and it is quite subjective. My wife started doing the blood sugar tests and it doesn’t hurt her at all. My younger daughter decided to try too, and it didn’t hurt her much either. And, it’s a skill like any other. Especially for my wife, getting enough blood out to take a reading took some practice, as her skin is apparently quite thick, and her capillaries quite far from the surface. Shaking the hand before testing, and doing some fist clenches, both helped.

I bought an iPhone-compatible GlucoRX HCT blood sugar monitor. But it had a headphone jack on it, and didn’t work through the dongle. There was nothing on the sales page to say that the lightning port version existed for those of us on jackless iPhones, which was very annoying.

So then I bought a lightning-port GlucoRX HCT monitor. And that barely worked either. I kept getting weird error messages, and my first round on the phone to GlucoRX support came to not much- I got told to hold the monitor vertically. When in fact it should be at about 45 degrees, and the problem was a defective monitor, which I found out when I rang them back at lunch time and got not a customer service person, but an actual engineer! He was super-helpful, diagnosed the problem (“that error message ought to be impossible on that monitor as it doesn’t have an internal battery”) and got a new monitor, plus one of the standalone (no-phone-required) monitors added in for free, into the post to me that very day.

If I was to start this all over again, I would go with a continuous blood glucose monitor. It’s more expensive for a diabetic taking maybe 5 readings a day, but it’s about the same price as using the measuring sticks 20+ times a day for a month, but without the damn finger pricking, and with (as the name suggests) actual continuous monitoring. Matching up that data to a food diary would give a very complete record, with much less fuss. 

So armed with a monitor that worked, and with a large supply of very expensive test strips (about 32p per test, plus a few pence for each new lancet, which when you’re doing 20+ tests a day adds up pretty fast), I started taking some readings and recording them. First on the GlucoRX app, which is ok, and then I tried to add them to the Personalized Nutrition app. Oh my goddess, that app is a disaster. 

Here are the functions that that app is supposed to have: 1) record blood sugar readings. 2) record food intake. Those are the two critical ones. 

But it gives you three options for things to record: Exercise, Sleep (which you have to select right before you sleep- you can’t record it after the fact, so it’s 100% useless), and Food. But the much-vaunted massive database of foods to choose from doesn’t include toast. Toast!! 

And can you tell what’s missing? Right. You have to dig through two sub-menus to find the option to record your blood sugar. Every single time you need to record it. That’s 20+ times per day if you’re tracking every meal.

Seriously, somebody at the app design agency needs a beating with a very big stick.

So if you’re going to try this protocol, stay TF away from the Personalized Nutrition app. It’s shit.

Here’s what I’m doing instead:

1) I’m not tracking every meal every day. I did that for a couple of days, and it’s a pain. So I focussed on breakfast as the place to start, and I have already made some changes.

2) I record the time and the blood sugar reading, with a note about what I’ve been eating, in an actual notebook with an actual pen. Old school, baby.

3) I use my phone to photograph each meal I’m tracking. This gives me a time-stamped visual record to flesh out the notes. That way I don’t have to measure anything, and can tell meal sizes and details from the photos. This is important because quantity matters, as does what else you’re eating at the same time.

4) I’m only tracking meals I eat often. There would have been no point (other than curiosity) in tracking my mum’s killer chocolate cake that we ate last weekend, as it was a one-off.

5) I put the numbers into a spreadsheet (I’m on a Mac so using Numbers), and use that to create graphs to show blood glucose levels over time. 

7) I keep track of which meals don’t spike my blood sugar, and which ones do, and the overall shape of the spikes.

8) For the ones that do spike me, I try the meal again but removing the most likely culprit, and test again. Sadly, my breakfast oranges have to go 🙁

9) I put those graphs into a Pages file with the photos and notes, so I can see, for example, the effects of:

  • my usual breakfast; 
  • the same meal minus the orange; 
  • the same meal minus the toast but with the orange 
  • the same meal minus the orange and minus the toast; 
  • and so on.

The critical thing is to change only one thing at a time, so I can be sure what is having the effect.

Here are three breakfasts, and their results:

Breakfast 1: toast with smoked salmon; toast with peanut butter and blueberries; orange; coffee; crossword.

Breakfast 2: toast with smoked salmon; toast with peanut butter and blueberries; no orange; coffee; crossword.

Breakfast 3: smoked salmon with lettuce, peanut butter and blueberries, coffee, crossword.

And the results from those three versions:

In general, I can predict the effects of most foods. Eating Skittles after dinner sent my blood sugar predictably up to 10.7 mmol/L (about 194 mg/dL for my American friends). There was no immediate crash though, it took about two hours to get gently back to baseline. I don’t usually eat Skittles at all, but I love them, so had to try…. Bye bye Skittles 🙁

But eating watermelon after a vegetarian chilli with sweet potato… that got me over 11.1mmol/L 202 mg/dL, and I was back to baseline in an hour. My poor pancreas. What a trooper. (This one result is my entire basis for the somewhat misleading blog post title.)

The chilli by itself put me up over 8mmol/L (145 mg/dL), the springboard from which the watermelon leapt into action, but salmon with white rice and vegetables (which preceded the Skittles) got me only up to 6.9 (125 mg/dL). White rice! I was amazed- I was very much expecting it to be a metabolic hand-grenade.

Some meals push me up to over 8mmol/L, and keep me there for over two hours (such as my daughter’s favourite gluten-free pumpkin pasta). With others I stay under 7, and get back to baseline in an hour. Incidentally, it’s very clear that I’m in no way diabetic or pre-diabetic (I wasn’t concerned, but it’s nice to know anyway).

I am not planning to share my data here because it would take me hours and hours to make it presentable, and indeed most of it is still in the notebook. I can read the numbers just fine off the page- the handy graph visualisations are unnecessary for me at this point. Besides, spreadsheets and I do not get along well. Also, while this protocol may be useful to you, my data is not: the whole point of this exercise is that your blood sugar response is unique. Knowing what’s bad for me doesn’t help you.

Now that I’m familiar with the system and the effects of some foods, I can cut some corners and am taking fewer readings (which further reduces the usefulness of the data to an actual scientist). Having established the ranges of my sugar spikes, I have a general goal of keeping my level at 30 minutes after eating (timed from the beginning of the meal) to below 7.0. This is quite easy to accomplish. I would also like to drop my fasting blood glucose level to the middle of the normal range. At the moment, it’s hovering a safe margin below the top of the normal range. I’m already seeing it trend in the right direction, now that I’m able to predict and therefore avoid sugar spikes.

And of course, I have a lot of foods left to try. Including Nutella. I couldn’t quite bring myself to face the awful truth… 

*I am well aware that blood glucose response is not the only measure of a food’s healthiness, and that watermelon may have components that are helpful, and Skittles may have components that are actually harmful, beyond the sugar issue. Adding cyanide to food completely prevents a blood sugar spike- because you’re dead before the sugar hits your system! Also, I massively overstated the difference between watermelon and skittles, and haven't taken the pre-existing rise from the dinner into account, and not discussed the time taken to recover back to baseline into account. So it's not objectively true, I am taking massive licence for rhetorical effect. But this is not a scientific paper, it’s a blog post. M’kay?

I'm sure you have an opinion: do share!

3 Responses

  1. Well done. In general I’m not surprised by what you show in you’re graph. I ama little surprised by the skittles/watermelon results. Like you say though the meal prior to eating those things also effects the results.
    The orange results blow a giant hole in the “fiber slows the sugar uptake” belief. The bread appears to be a better moderator than the orange. I can understand about your the difficulties in posting your “data” .:) I considered showing some of mine but without all the circumstancial info – which would take pages, and hours for me to hunt and peck down, it would be more confusing than informative. Keep up the good work. I’m looking forward to further updates. Cheers

    1. Of course! but does it spike the blood glucose, or suppress it? I’ll have to do some further tests…

Leave a Reply

Your email address will not be published.

You May Also Like

Recent Posts

Jaegerstock, part 3

Now that we have a working Jaegerstock, let’s take a look at lessons two and