High-Frequency Blood Glucose Testing

I’m going to talk about some unexplained blood glucose spikes at night. I tracked these using a little sensor that has a tiny filament on it, about 3mm long and it goes on with adhesive. I’ve ripped the adhesive off so you can see what the back looks like and there’s a little applicator that pushes it onto the skin.

And then you read it with an NCE Reader, just very little power and you click get sensor data and it loads data from the button. And then to get the data, you load some proprietary software on your computer and you connect it with the yellow cord. And you fill out a patient ID and a name, any name. I put “Why do I have to do this” and then you get a file, and the files really interesting and records blood glucose every 15 minutes.

And you can see my glucose in my case is quite variable. The lowest numbers are just random, but from 55 to I think there’s a 130 down there.

So what did I do? I had some questions, of course I’m interested in new forms of sensing, but also blood glucose data is super relevant to metabolic conditions and like many other people I had some health related questions about my metabolism. Although I’ve gone and checked, I’ve had suspicions that I’ve you know kind of diabetes type II and or pre-diabetes or something and I’ve had the spot checks, fasting glucose checks and I don’t really have diabetes.

The test is you do a fasting blood sugar. A blood sample will be taken after an overnight fast. If the fasting blood sugar has less than 100mg/dl or whatever that is it’s normal. Mine is normal.

So I don’t have a right to have this device. This is the doctors, but I’ll just say thank you doctor, without a last name and say I had questions and she collaborated with me in getting this device.

So okay, there we go, I get my time and see it’s every 15 minutes, and I get a number and I put it into a spreadsheet. And I start to ask myself what is the cause of my blood sugar going up and down.

Now here’s what’s interesting. You can see that my spot checks. So I go in and this is taken at six sixteen in the morning for a week in the morning. Imagine I go into the doctor for my spot check right. It’s 92, I have no problem. I have 144, I scream diabetes. I’ve 105, so that’s kind of borderline but probably something to worry about. On Thursday I go in and it’s 92, I’m fine, 72. And this is my next week, no sorry tis is half an hour later, maybe there’s (time ? 03:09) effects. Half an hour you see, oh, if I went along those first days I would have had that diagnosis. But if I had gone in on the other days at a different time I would have had no diagnosis. So here you see spot checks don’t give the right answer.

This is what I’ve found as a normal. Blood glucose rise and falls for me after eating during the day, I’ve about 112 and it goes right up to 150, but then it comes right back down again in an nice curve. That’s just kind of a baseline measurement and the shape is a baseline.

And here’s a typical night. It starts at 100, it has some lows during the night and then as I wake up there’s a thing called the dawn phenomena and it starts to go up. And then I eat breakfast and then did spots.

So the first thing I decided to track was what re my peaks, okay, so let’s just take a look at the peaks. I define the peaks at over 150, and I’ll show you how I kind of found the peaks.

I simple sorted it in a spreadsheet from high to low and then most of them clumped together but not always so I had to kind of move them around and put little dots next to the incidents where I had a peak. And I had 24 peaks.

So what are the cause of these peaks? And one of the main causes is special eating like eating lots of sweet things and things like blood glucose go way up.

Of course I had to find that out and I use a variation from what I learned from Shultz in a case crossover design. I didn’t do some complex correlation, I just found the peaks and went back to all my data sources and tried to find, in my photos for instance that day, so I found evidence of what I did.

So this we call dinner with Italians. These are close friends, and you can see we had pasta and bread, and she always makes cookies and the kids were there and talking about life and eating a lot of that and then you get a glucose spike.

So here you see it’s all very clear. you have some lunch, you have beer on an empty stomach when you come to Amsterdam and have stuff with your friends and then you get these really high numbers.

It’s useful to really to really be able to track it to a source, and what became especially helpful to me was an app developed by some Quantified Self people who you probably know the name (unclear name 05:31) and (unclear name 05:34), allowed me to take a photo upside down and it pulls the metadata off the photo and identifies that as a photo of food. So you don’t have to sort your photos, then you can put it on a timeline with your glucose values.

The workflow is hard to kind of get going but they’re very helpful. And here you can see something kind of unexpected. You see a big spike of 168 on the slide, and you see well I had normal food and I just had fruit which actually doesn’t spike my blood glucose. But for some reason it spiked my blood glucose. So that’s interesting, and gets that personal glycemic index here.

And now you see my question, and this is what I gained from this project was my question. Look at those question marks. That’s the middle of the night.

So I had four peaks in the middle of the night and they’re very dramatic. And they’re hours and hours after eating. I’m fast asleep and don’t even know it. And they’re not just small rises, and when you look at the graph, they have a very well defined shape. So I have a low. This is 63 and then I have a very big spike to 168. And then it falls down and starts to kind of like normalize. And then when I get and if I went to the doctor and get tested, I’d have 96. So there’s a hidden phenomenon in there somewhere.

What is the cause of this high variation at night? If you look in the literature which I have, and you talk to your friends, it’s not that obvious. So there’s not an obvious answer you can look up on Google to find that.

One possible suggestion is its measurement error. We don’t know. Another possibility which is very interesting, I think that it’s pre-diabetes, signaled by variability rather than via fasting blood glucose number. And variability in blood sugar isn’t used as a diagnostic category for people who are at risk of diabetes type II. So it’s quite simple that you would think like before you crash your car maybe you kind of warm up a little bit, and that’s what I take into my next steps so thank you.